Botanical Journal of the Linnean Society

Pederneiras, Leandro Cardoso;Gaglioti, André Luiz;Romaniuc-Neto, Sergio;de Freitas Mansano, Vidal

doi: 10.1093/botlinnean/boy036pmid: N/A

Abstract Studies that place the origin of the stem node of Ficus in Gondwana or Eurasia at c. 136–134 Mya are not congruent with the estimated origin of Moraceae in the Late Cretaceous. To elucidate these origins, molecular analyses were performed using three calibration points in BEAST with 268 terminal taxa based on ITS, ETS, G3pdh, ncpGS and GBSSI sequences. S-DIVA, DEC and S-DEC were used to estimate ancestral areas. The arrangement of staminate flowers within the syconium supports the current phylogenetic framework of Ficus. Our results suggest that the stem lineage of Ficus originated at the beginning of the Cenozoic among the boreotropical flora around the Tethys Seaway, and that the crown lineage radiated during the Eocene. Ficus reached the Americas via two routes: from Europe to North America in the Early Eocene, giving rise to F. section Pharmacosycea, and from Africa to South America in the Late Eocene via the trans-Atlantic route, giving rise to F. section Americanae. East Asia appears to have played an important role in the evolution of species with staminate flowers grouped in the ostiole, here named the Androstiole clade. In the Oligocene, these species probably migrated to South-East Asia, reuniting with lineages of F. subgenus Spherosuke. Bayesian inference, biogeography, Cretaceous, dispersal, divergence, Oligocene, vicariance INTRODUCTION Moraceae are a monophyletic family of 37 genera and c. 1150 species (Berg, 2005; Clement & Weiblen, 2009) classified into six tribes. Ficeae (including Ficus L.) and Castilleae form a clade that is sister to Dorstenieae (Clement & Weiblen, 2009). This family has been intensively investigated in molecular dating studies in the last 20 years, and penalized likelihood approaches have dated Moraceae to the Late Cretaceous in both an angiosperm-wide study [stem node: 78.9–70.4 Mya confidence interval (CI); Magallón et al., 2015] and a family-wide study [crown node: 110.0–72.6 Mya highest posterior density interval (HPD), Zerega et al., 2005]. This timing renders it difficult to determine conclusively whether the ancestral area of Moraceae was on Gondwana or the Eurasian palaeocontinent (Zerega et al., 2005). Ficus, the major genus of Moraceae, is an important globally distributed group, the biogeographical history of which provides insights beyond the genus itself. It is a relatively ancient genus with > 750 extant species distributed mainly in tropical regions (Berg & Wiebes, 1992; Zhekun & Gilbert, 2003; Berg & Corner, 2005). The genus is divided into six subgenera and 19 sections (Berg & Corner, 2005; Pederneiras et al., 2015b), concentrated in four main geographical areas: (1) South-East Asia: ‘F. subsection Urostigma (Endl.) C.C.Berg’, F. subsection Conosycea (Miq.) Corner, F. subsection Stilpnophyllum (Endl.) C.C.Berg, F. subgenus Sycomorus Raf., F. subgenus Terega Raf., F. subgenus Synoecia (Miq.) Miq. and F. subgenus Ficus; (2) Australasia: F. subsection Malvanthera (Corner) C.C.Berg and F. section Oreosycea (Miq.) Corner; (3) Africa: F. section Platyphyllae Mildbr. & Burret; and (4) the Neotropics: F. section Americanae (Miq.) Corner and F. section Pharmacosycea (Miq.) Griseb. According to Berg (2005), the northern Andes, the adjacent parts of Central America, the upper Amazon basin, Central Africa, eastern and northern Africa, the Sino-Himalayan region, western Malesia and eastern Malesia are the primary centres of diversity for Ficus. Two biogeographical scenarios have been proposed for Ficus based on molecular dating, using c. 200 species and the following sequences as markers: nuclear ribosomal internal and external transcribed spacers (ITS and ETS), glyceraldehyde-3-phosphate dehydrogenase (G3pdh), plastid-expressed glutamine synthetase (ncpGS) and granule-bound starch synthase (GBSSI). Xu et al. (2011) hypothesized that the fragmentation of Gondwana led to the diversification of the group, whereas India (including Madagascar) played an important role in the early evolution of Ficus. Cruaud et al. (2012) hypothesized that the origin of the Ficus lineage was in Eurasia, with subsequent dispersals to the Americas, Africa, India, Madagascar and Australia. Other studies hypothesized the origin of ‘F. subsection Urostigma’ to be in Madagascar, with subsequent dispersal to Africa and, by rafting on India, to Asia (Chantarasuwan et al., 2016), and the origin of F. section Americanae and Pharmacosycea to be in Atlantic Forest, Brazil (Machado et al., 2018). Each of these interpretations of the origin of the Ficus lineage is plausible, and each hypothesis should be tested as new evidence is discovered. In the present study, we introduce an important addition to the current knowledge base by adjusting the molecular clock to achieve a more appropriate age estimation that concurs with the origin of flowering plants. Previous molecular dating analyses of Ficus have calibrated the age of the crown group of Ficus to a minimum of 60 Mya, based on fossil achenes from the Eocene, and a maximum of 190 or 196 Mya, using a uniform prior distribution (Xu et al., 2011; Cruaud et al., 2012). These analyses have estimated the stem age of Ficus (the divergence of Ficeae and Castilleae) to between 195 and 85 Mya. However, this age estimate is too old, as it implies that the lineage evolved before Moraceae (Zerega et al., 2005; Magallón et al., 2015; Gardner et al., 2017). There are three possible reasons that might explain the anomalous estimation of the age of Ficus: (1) the calibration of the crown group of Ficus did not account for the possibility of assigning Ficus fossils to the stem group, and therefore overestimated the age; (2) Ficus was calibrated using the maximum age of c. 190 Mya to initialize BEAST analysis in previous versions (Xu et al., 2011; Cruaud et al., 2012), which is much older than angiosperm diversification (139.5 Mya; Magallón et al., 2015); and (3) a uniform distribution prior is used for calibration instead of the lognormal distribution. Crown Moraceae, being in the core eudicots, are unlikely to have been present at the time of the presumed stem angiosperm radiation (Magallón et al., 2015). The present study builds on previous work and includes the largest number of Ficus spp. (249) to date. Using ITS, ETS, G3pdh, ncpGS and GBSSI sequences, including recent molecular data for F. section Pharmacosycea (Pederneiras, Romaniuc-Neto & Mansano, 2015a) and ‘F. subsection Urostigma’ (Chantarasuwan et al., 2015), we aimed to achieve a better understanding of the historical biogeography of the pantropical genus Ficus, and to assess the role of biogeographical barriers and bridges in determining the divergence of lineages. Based on the constant incongruence between the previous molecular phylogenetic results and the classification of Ficus, we also aim to provide new morphological data to strengthen the historical interpretation of Ficus. MATERIAL AND METHODS Taxon sampling, markers and evolutionary models Molecular data from 268 specimens of Moraceae (249 Ficus spp.) were selected from GenBank (Supporting Information, Fig. S1), representing all sections of Ficus recognized by Berg & Corner (2005). The following sequences were used in the analysis: ITS (891 bp), ETS (528 bp), G3pdh (769 bp), ncpGS (1630 bp) and GBSSI (1734 bp). These markers have been widely used in phylogenetic analyses of Ficus (e.g. Weiblen, 2000; Jousselin, Rasplus & Kjellberg, 2003; Rønsted et al., 2005, 2008b; Rønsted, Salvo & Savolainen, 2007; Xu et al., 2011; Cruaud et al., 2012; Harrison et al., 2012; Chantarasuwan et al., 2015; Pederneiras et al., 2015a). The sequences were aligned using Muscle 3.7 under default settings (Edgar, 2004) in the CIPRES Science Gateway (Miller, Pfeiffer & Schwartz, 2010). PartitionFinder v1.0.1 (Lanfear et al., 2012) was used to find the best-fit partitioning schemes and models of molecular evolution for DNA sequences with heuristic search strategy (‘greedy’) and branch lengths linked. The best-fit models for the partitions, based on the Bayesian information criterion (BIC), were GTR+G (ITS1 and ITS2), TrN+G (18S), TrN+I+G (G3pdh), K80+I+G (26S and 5.8S) and HKY+G (ETS, ncpGS and GBSSI). Each codon position in each gene was assigned a separate substitution rate in BEAST. The phylogenetic results of each of the five molecular markers were analysed using BEAST. As the topologies inferred from the five markers were congruent in resolving the Pharmacosycea as sister to the subgenus Spherosuke Raf. + Androstiole clade (Androstiole is defined in the Results), and in confirming the monophyly of several groups (e.g. Americanae, Conosyceae, Malvanthera, Oreosycea, Synoecia and Terega), we combined the datasets. Sauquet (2013) argued that significant increases in both precision and accuracy can be gained by using as many genes and taxa as possible in molecular dating, although he did not address concatenation. Divergence time A Bayesian analysis was conducted using the Markov chain Monte Carlo (MCMC) method, and divergence times of the Ficus lineages were inferred by setting the molecular clock prior to the uncorrelated lognormal relaxed clock (UCLN) in BEAST v1.8 (Drummond & Rambaut, 2007). The birth–death prior (Gernhard, Hartmann & Steel, 2008) was applied. The Yule prior was also tested, but as the results were similar to those obtained by the birth–death prior, they are not presented here. BEAST, via the CIPRES Science Gateway, was used to run a Markov chain of 170 million generations, sampling every 17000 steps. Convergence and the choice of appropriate ‘burn-in’ were evaluated in TRACER 1.6.0 (Rambaut et al., 2013). Multiple runs were conducted to check that the chains converged on the same stationary distribution using TRACER 1.6.0. The tree with maximum clade credibility, the calculation of the median ages (node heights) and the highest posterior density intervals (95% HPD) were summarized using TREEANNOTATOR (Rambaut & Drummond, 2007). Previous estimates (Xu et al., 2011; Cruaud et al., 2012) used a uniform prior distribution to calibrate the deepest nodes of the phylogenetic tree for Ficus (i.e. they considered equal probabilities for all values within the limits of the distribution); however, this approach can be considered overly conservative. According to Ho (2007), the lognormal distribution is perhaps the most adequate approach for modelling palaeontological information. It explicitly assumes that the actual divergence event is more likely to have occurred sometime before the earliest appearance of fossil evidence. In the case of Ficus, a lognormal distribution of three calibration points was set with 95% probability distributed between the fossil age and before the origin of Moraceae (78–70 Mya; Magallón et al., 2015). Node 1 (stem node of Ficus) was calibrated based on the age of the fossil achene of Ficus tethyca Doweld, which has been estimated to date back to the early Eocene (Collinson, 1989; Doweld, 2016), at 48 Mya (offset), 2.1 Mya (mean) and 0.5 Mya (SD). This prior assumes that 95% of the probability of the age of Ficus lies at c. 69–51 Mya, which is younger than the Moraceae crown node according to Magallón et al. (2015) and Zerega et al. (2005). The lack of clear characters in the fossil F. tethyca does not offer sufficient data on whether this fossil belongs to the stem or crown nodes. We followed the recommendation of Forest (2009) and placed it at the stem node to avoid the over-estimation of age. The fossil achene of F. tethyca has a lateral style and an orthogonal reticulate surface, similar to extant Ficus achenes. Thus, the identification of this fossil was assumed correct for the purposes of this study. The fossils of Ficus that were identified by Anzótegui & Herbst (2004; Mid-Miocene), Srivastava, Srivastava & Mehrotra (2011; Early Miocene), Bozukov, Ivanov & Utescher (2013; Lower Oligocene) and several other authors (see PaleoBioDB, http://paleobiodb.org/) generally include leaves with few morphological features, which enable the specific determination of any subdivision of Ficus. Morphological synapomorphies that support subgenera and sections of Ficus mainly include growth habit (trees, shrubs, hemi-epiphytes or climbers) and reproductive characteristics (i.e. plant reproductive strategies: monoecious or gynodioecious), position of the syconia (axillary, ramiflorous, cauliflorous or stoloniflorous), number of stamens (one to five), ostiole form (circular or slit-shaped) and basal bracts of the receptacle. Because these characteristics are difficult to interpret in the fossil record, we calibrated molecular phylogenetic trees using only the oldest fossil. Nodes 3 (crown node of F. section Pharmacosycea) and 7 (crown node of F. section Americanae) were calibrated based on fossil agaonid wasps (Tetrapus and Pegoscapus, respectively, both 16 Mya; Peñalver, Engel & Grimaldi, 2006), the exclusive pollinators of Ficus, with life-cycles that are dependent on the fig (Kjellberg et al., 2005). Both nodes were calibrated at 16 Mya (offset), 2.8 Mya (mean) and 0.6 Mya (SD). This prior assumes that 95% of the probability of the crown node is at c. 69–21 Mya. These standard deviations were wide enough to reflect the origin of sections closer to the fossil age than to the origin of Moraceae (78.9–70.4 Mya; Magallón et al., 2015). The root (crown node of Dorstenieae, and Ficeae + Castilleae) was calibrated using normal distribution at 63.0 Mya (mean) and 3.5 Mya (SD). This prior assumes that 95% of the probability of the age of Ficus lies at c. 69–56 Mya (the upper limits on the fossil Ficeae and below the age of the family estimated by Magallón et al., 2015). Biogeographical analysis A biogeographical analysis of Ficus was conducted using Reconstruct Ancestral State in Phylogenies (RASP) 3.0 (Yu, Harris & He, 2014), which implements a statistical dispersal-vicariance analysis (S-DIVA; Yu, Harris & He, 2010), the dispersal-extinction-cladogenesis (DEC) model of Lagrange (Ree & Smith, 2008), and a statistical DEC model (S-DEC; Yu et al., 2015). Ten thousand post-burnin trees, and the maximum clade credibility tree generated by BEAST were loaded into RASP, and node height was set to four, as extant Ficus spp. are distributed in at least four areas. In the DEC analyses, two time slices were used (0–34 and 34–74 Mya) based on the retraction of tropical forests in the Oligocene (Morley, 2000). Species distribution was based on the floristic regions proposed by Takhtajan (1986) – A: North America; B: western South America (Andes); C: eastern South America; D: Africa; E: South-East Asia and India; F: West to East Asia (Turkey to China); G: Australasia; and H: Madagascar. The discussion of divergence was based on 26 nodes and their median age, HPD intervals, posterior probability and the distribution of the ancestor. Most of the discussion narratively links the evolution of Ficus to geological events in Earth’s history based on temporal coincidence. Although this is common practice in this type of study, these hypotheses can change and improve with advancements in earth sciences (Renner, 2016). For example, it has been hypothesized that the dispersal between the Australian continental crust and SE Asia possibly occurred > 20 Mya, although these dispersal events were presumably rare (Hall, 2002). This hypothesis was strongly questioned and infrequently implemented to explain species divergence until the study by Crayn, Costion & Harrington (2015), who provided further evidence to support it. Morphological data We reviewed material from the herbaria B, BG, BM, F, G, K, NY, P, RB and US, and Ficus bibliographies (e.g. King, 1888; Sata, 1944; Corner, 1960a, b, 1965; Carauta, 1989; Berg & Wiebes, 1992; Berg, 2003; Berg & Corner, 2005; Chaudhary et al., 2012; Chantarasuwan, Berg & van Welzen, 2013). Based on the examined material a spreadsheet was built containing collection information and morphological characters such as sexual system of the plant and the arrangement of staminate flowers, among others. The database was analysed verifying the similarities and dissimilarities between the main subdivisions of Ficus comparing the morphology of the different clades. RESULTS Phylogenetic relationships The BEAST maximum clade credibility tree with 26 selected nodes is presented in Figure 1, and posterior probabilities are listed in Table 1. The phylogenetic relationships exhibited high support for all 26 nodes (≥ 0.90 posterior probability). Of the infrageneric taxa of Ficus, ten groups are monophyletic, including F. section Pharmacosycea, F. section Americanae, F. section Platyphyllae, F. subsection Conosycea, F. subsection Malvanthera, F. subgenus Terega, F. subgenus Synoecia, ‘F. subsection Urostigma’, F. section Oreosycea and F. subgenus Sycomorus. Paraphyletic groups were identified in Ficus subgenus Pharmacosycea, F. subgenus Ficus and F. subgenus Spherosuke, as noted by Rønsted et al. (2008b), Xu et al. (2011) and Cruaud et al. (2012). The relationships of some taxa were not previously recovered: node 13 with F. subsection Ficus, F. subsection Frutescentiae, F. section Eriosycea, F. subgenus Terega and F. subgenus Synoecia; and node 22 with ‘F. subsection Urostigma’, F. subgenus Sycomorus and F. section Oreosycea. Figure 1. View largeDownload slide Chronogram with summarized topology of Ficus (249 species) inferred by BEAST using ITS, ETS, G3pdh, ncpGS and GBSSI sequence data. For details of node support, divergence times and ancestral area reconstruction results at nodes of interest (1–26) see Table 1. Blue bars indicate the 95% highest posterior distribution (HPD) intervals of the divergence time. Ancestral areas and dispersal and vicariance events inferred by the S-DEC analyses are mapped on the chronogram. The coloured squares with letters indicate ancestral areas of the node according to the map. A black diamond indicates the occurrence of a vicariance event. Biogeographical regions: A, North America; B, western South America (Andes); C, eastern South America; D, Africa; E, South-East Asia and India; F, West to East Asia (Turkey to China); G, Australasia; H, Madagascar. Figure 1. View largeDownload slide Chronogram with summarized topology of Ficus (249 species) inferred by BEAST using ITS, ETS, G3pdh, ncpGS and GBSSI sequence data. For details of node support, divergence times and ancestral area reconstruction results at nodes of interest (1–26) see Table 1. Blue bars indicate the 95% highest posterior distribution (HPD) intervals of the divergence time. Ancestral areas and dispersal and vicariance events inferred by the S-DEC analyses are mapped on the chronogram. The coloured squares with letters indicate ancestral areas of the node according to the map. A black diamond indicates the occurrence of a vicariance event. Biogeographical regions: A, North America; B, western South America (Andes); C, eastern South America; D, Africa; E, South-East Asia and India; F, West to East Asia (Turkey to China); G, Australasia; H, Madagascar. Table 1. Clade support (posterior probability, PP), median age, limits of highest posterior distribution (HPD), ancestral area reconstruction (the most likely state only) constrained up to four areas for the 26 nodes of Ficus (Fig. 1) generated using S-DIVA, DEC and S-DEC methods after BEAST, and biogeographical barriers and bridges in determining divergence of lineages Node Name Clade support (PP) Median age (Ma) Earliest (HPD) age (Ma) Youngest (HPD) age (Ma) S-DIVA DEC S-DEC Biogeographical barriers and bridges in determining divergence of lineages 1 Stem Ficus 1 59.3 66.2 52.7 ABE E E 2 Crown Ficus 1 42.8 50.6 34.9 ABE AE AE North Atlantic land bridge 3 Crown Pharmacosycea 1 28.13 36.2 20.6 AB A A GAARlandia land bridge 4 Stem Spherosuke 0.98 39.0 46.3 31.7 E E E Drier route in the Turgai Strait 5 Crown Spherosuke 1 32.2 38.9 25.9 DE CDE CDE Decreased temperatures in middle latitude regions in the early Oligocene 6 Stem Americanae 1 23.4 28.2 19.4 CD CD CD Increased distance between Africa and South America 7 Crown Americanae 0.96 20.1 23.5 17.6 C C C 8 Crown Platyphyllae 0.96 18.6 23.8 13.3 D D D 9 Stem Malvanthera 1 27.1 33.7 21.1 E E E Contact between Australian continental crust and SE Asia 10 Crown Conosycea 1 22.9 29.1 17.5 E E E 11 Crown Malvanthera 1 19.8 26.2 13.2 E EG EG 12 Crown Androstiole clade 1 33.8 41.4 27.3 EF EF EF Dry climate barrier emerged between China and South- East Asia 13 Crown Caricoide clade 0.92 30.4 37.6 23.3 EF F F 14 Stem Terega 1 27.3 34.1 20.8 E EF EF 15 Crown Terega 1 22.3 28.7 16.4 EF EF EF 16 Stem Frutescentiae 1 23.9 29.8 17.6 E EF EF 17 Stem Synoecia 0.97 21.8 27.9 16.1 E E E 18 Crown Synoecia 1 14.7 19.8 9.9 E E E 19 Crown Eriosycea 1 12.5 16.5 6.4 E E E 20 Crown Frutescentiae 1 18.8 24.7 12.9 F F F 21 Crown Ficus 1 2.9 7.1 0.5 F F F 22 Stem Urostigma 0.96 31.7 38.7 25.1 E E E 23 Crown Urostigma 1 22.9 30.3 16.1 EF EF EF 24 Stem Oreosycea 0.97 29.5 36.4 23.3 E E E 25 Crown Oreosycea 1 17.7 26.1 11.0 E EG EG 26 Crown Sycomorus 1 26.3 32.6 20.6 E E E Node Name Clade support (PP) Median age (Ma) Earliest (HPD) age (Ma) Youngest (HPD) age (Ma) S-DIVA DEC S-DEC Biogeographical barriers and bridges in determining divergence of lineages 1 Stem Ficus 1 59.3 66.2 52.7 ABE E E 2 Crown Ficus 1 42.8 50.6 34.9 ABE AE AE North Atlantic land bridge 3 Crown Pharmacosycea 1 28.13 36.2 20.6 AB A A GAARlandia land bridge 4 Stem Spherosuke 0.98 39.0 46.3 31.7 E E E Drier route in the Turgai Strait 5 Crown Spherosuke 1 32.2 38.9 25.9 DE CDE CDE Decreased temperatures in middle latitude regions in the early Oligocene 6 Stem Americanae 1 23.4 28.2 19.4 CD CD CD Increased distance between Africa and South America 7 Crown Americanae 0.96 20.1 23.5 17.6 C C C 8 Crown Platyphyllae 0.96 18.6 23.8 13.3 D D D 9 Stem Malvanthera 1 27.1 33.7 21.1 E E E Contact between Australian continental crust and SE Asia 10 Crown Conosycea 1 22.9 29.1 17.5 E E E 11 Crown Malvanthera 1 19.8 26.2 13.2 E EG EG 12 Crown Androstiole clade 1 33.8 41.4 27.3 EF EF EF Dry climate barrier emerged between China and South- East Asia 13 Crown Caricoide clade 0.92 30.4 37.6 23.3 EF F F 14 Stem Terega 1 27.3 34.1 20.8 E EF EF 15 Crown Terega 1 22.3 28.7 16.4 EF EF EF 16 Stem Frutescentiae 1 23.9 29.8 17.6 E EF EF 17 Stem Synoecia 0.97 21.8 27.9 16.1 E E E 18 Crown Synoecia 1 14.7 19.8 9.9 E E E 19 Crown Eriosycea 1 12.5 16.5 6.4 E E E 20 Crown Frutescentiae 1 18.8 24.7 12.9 F F F 21 Crown Ficus 1 2.9 7.1 0.5 F F F 22 Stem Urostigma 0.96 31.7 38.7 25.1 E E E 23 Crown Urostigma 1 22.9 30.3 16.1 EF EF EF 24 Stem Oreosycea 0.97 29.5 36.4 23.3 E E E 25 Crown Oreosycea 1 17.7 26.1 11.0 E EG EG 26 Crown Sycomorus 1 26.3 32.6 20.6 E E E A, North America; B, western South America (Andes); C, eastern South America; D, Africa; E, South-East Asia and India; F, West to East Asia (Turkey to China); G, Australasia; H, Madagascar. View Large Table 1. Clade support (posterior probability, PP), median age, limits of highest posterior distribution (HPD), ancestral area reconstruction (the most likely state only) constrained up to four areas for the 26 nodes of Ficus (Fig. 1) generated using S-DIVA, DEC and S-DEC methods after BEAST, and biogeographical barriers and bridges in determining divergence of lineages Node Name Clade support (PP) Median age (Ma) Earliest (HPD) age (Ma) Youngest (HPD) age (Ma) S-DIVA DEC S-DEC Biogeographical barriers and bridges in determining divergence of lineages 1 Stem Ficus 1 59.3 66.2 52.7 ABE E E 2 Crown Ficus 1 42.8 50.6 34.9 ABE AE AE North Atlantic land bridge 3 Crown Pharmacosycea 1 28.13 36.2 20.6 AB A A GAARlandia land bridge 4 Stem Spherosuke 0.98 39.0 46.3 31.7 E E E Drier route in the Turgai Strait 5 Crown Spherosuke 1 32.2 38.9 25.9 DE CDE CDE Decreased temperatures in middle latitude regions in the early Oligocene 6 Stem Americanae 1 23.4 28.2 19.4 CD CD CD Increased distance between Africa and South America 7 Crown Americanae 0.96 20.1 23.5 17.6 C C C 8 Crown Platyphyllae 0.96 18.6 23.8 13.3 D D D 9 Stem Malvanthera 1 27.1 33.7 21.1 E E E Contact between Australian continental crust and SE Asia 10 Crown Conosycea 1 22.9 29.1 17.5 E E E 11 Crown Malvanthera 1 19.8 26.2 13.2 E EG EG 12 Crown Androstiole clade 1 33.8 41.4 27.3 EF EF EF Dry climate barrier emerged between China and South- East Asia 13 Crown Caricoide clade 0.92 30.4 37.6 23.3 EF F F 14 Stem Terega 1 27.3 34.1 20.8 E EF EF 15 Crown Terega 1 22.3 28.7 16.4 EF EF EF 16 Stem Frutescentiae 1 23.9 29.8 17.6 E EF EF 17 Stem Synoecia 0.97 21.8 27.9 16.1 E E E 18 Crown Synoecia 1 14.7 19.8 9.9 E E E 19 Crown Eriosycea 1 12.5 16.5 6.4 E E E 20 Crown Frutescentiae 1 18.8 24.7 12.9 F F F 21 Crown Ficus 1 2.9 7.1 0.5 F F F 22 Stem Urostigma 0.96 31.7 38.7 25.1 E E E 23 Crown Urostigma 1 22.9 30.3 16.1 EF EF EF 24 Stem Oreosycea 0.97 29.5 36.4 23.3 E E E 25 Crown Oreosycea 1 17.7 26.1 11.0 E EG EG 26 Crown Sycomorus 1 26.3 32.6 20.6 E E E Node Name Clade support (PP) Median age (Ma) Earliest (HPD) age (Ma) Youngest (HPD) age (Ma) S-DIVA DEC S-DEC Biogeographical barriers and bridges in determining divergence of lineages 1 Stem Ficus 1 59.3 66.2 52.7 ABE E E 2 Crown Ficus 1 42.8 50.6 34.9 ABE AE AE North Atlantic land bridge 3 Crown Pharmacosycea 1 28.13 36.2 20.6 AB A A GAARlandia land bridge 4 Stem Spherosuke 0.98 39.0 46.3 31.7 E E E Drier route in the Turgai Strait 5 Crown Spherosuke 1 32.2 38.9 25.9 DE CDE CDE Decreased temperatures in middle latitude regions in the early Oligocene 6 Stem Americanae 1 23.4 28.2 19.4 CD CD CD Increased distance between Africa and South America 7 Crown Americanae 0.96 20.1 23.5 17.6 C C C 8 Crown Platyphyllae 0.96 18.6 23.8 13.3 D D D 9 Stem Malvanthera 1 27.1 33.7 21.1 E E E Contact between Australian continental crust and SE Asia 10 Crown Conosycea 1 22.9 29.1 17.5 E E E 11 Crown Malvanthera 1 19.8 26.2 13.2 E EG EG 12 Crown Androstiole clade 1 33.8 41.4 27.3 EF EF EF Dry climate barrier emerged between China and South- East Asia 13 Crown Caricoide clade 0.92 30.4 37.6 23.3 EF F F 14 Stem Terega 1 27.3 34.1 20.8 E EF EF 15 Crown Terega 1 22.3 28.7 16.4 EF EF EF 16 Stem Frutescentiae 1 23.9 29.8 17.6 E EF EF 17 Stem Synoecia 0.97 21.8 27.9 16.1 E E E 18 Crown Synoecia 1 14.7 19.8 9.9 E E E 19 Crown Eriosycea 1 12.5 16.5 6.4 E E E 20 Crown Frutescentiae 1 18.8 24.7 12.9 F F F 21 Crown Ficus 1 2.9 7.1 0.5 F F F 22 Stem Urostigma 0.96 31.7 38.7 25.1 E E E 23 Crown Urostigma 1 22.9 30.3 16.1 EF EF EF 24 Stem Oreosycea 0.97 29.5 36.4 23.3 E E E 25 Crown Oreosycea 1 17.7 26.1 11.0 E EG EG 26 Crown Sycomorus 1 26.3 32.6 20.6 E E E A, North America; B, western South America (Andes); C, eastern South America; D, Africa; E, South-East Asia and India; F, West to East Asia (Turkey to China); G, Australasia; H, Madagascar. View Large We named clade 12 the Androstiole group (Ficus subgenus Synoecia, F. subgenus Sycomorus, F. subgenus Terega, F. subgenus Ficus, F. section Oreosycea and ‘F. subsection Urostigma’), and clade 13 the Caricoide group (Ficus subgenus Synoecia, F. subgenus Terega and F. subgenus Ficus). Molecular divergence time estimates Mean divergence age estimates and 95% HPDs at nodes of interest are included in Figure 1 and Table 1. MCMC analysis reached the stationary phase in almost all parameters after c. 17 million generations. The log files from two independent and identical analyses converged on the same stationary distribution for all parameters. Effective sample size values were > 200 in all traces after discarding the initial trees (burn-in 18%). This provided useful information on the posterior probability. Considering only the mean age of the Ficus divergences of the 26 selected nodes, the Oligocene was indicated as the period of greatest diversification of the Ficus lineage (13 nodes), followed by the Neogene (ten nodes), the Eocene (two nodes) and the Palaeocene (one node). The crown nodes of all infrageneric taxa in Ficus, except for F. subsection Ficus (node 21), are dated to > 12 Mya (Fig. 1; Table 1; Appendix S1). Ancestral area reconstruction The S-DEC reconstruction is summarized in Figure 1, providing an overview of inferred dispersal and vicariance events at nodes of interest. The ancestral areas inferred by S-DIVA, DEC and S-DEC analyses for selected nodes are listed in Table 1 (Figs S2–S4). Most ancestral areas were similar in the three analyses (S-DIVA, DEC and S-DEC), except for nodes 1, 2, 3, 5, 11, 13, 14, 16 and 25, although they always shared a common area. We chose the results of the S-DEC analysis to represent the graph and discuss the biogeographical history of Ficus, as West to East Asia was the most represented in the reconstruction of ancestral areas. This region appears to be an important evolutionary area for Ficus, as it has several endemic species (e.g. 16 in China; Zhekun & Gilbert, 2003). Morphological analysis The clades of Ficus that resulted from our analyses (clades 3, 5 and 12) are characterized by the occurrence of a singlet type of staminate flower arrangement. The presence of other characters did not show equally clear, clade-specific patterns. There are two basic types of staminate flower arrangements: (1) scattered-staminate flowers dispersed among pistillate flowers (F. section Pharmacosycea, F. section Americanae, F. section Platyphyllae, F. subsection Conosycea, F. subsection Malvanthera) (Fig. 2A); and (2) clustered (clade 12, ‘the group of clustered staminate flowers’) -staminate flowers grouped in the ostiolar region in monoecious or gynodioecious species (F. subgenus Terega, F. subgenus Synoecia, F. subgenus Ficus, ‘F. subsection Urostigma’, F. subgenus Sycomorus and F. section Oreosycea) (Fig. 2B–D). Subsection Urostigma’ may have staminate flowers grouped in the ostiolar region and few irregularly dispersed among the pistillate ones (Chatarasuwan et al., 2013). Figure 2. View largeDownload slide Two types of flower arrangements inside the syconium of Ficus. A, staminate flowers only dispersed among pistillate ones (staminate, pistillate and gall flowers; Pharmacosycea and Spherosuke p.p. lineages). B–D, staminate flowers grouped at the ostiolar region sometimes with few irregularly dispersed among the pistillate ones; B, monoecious syconium with staminate, pistillate and gall flowers (Sycomorus, Oreosycea and ‘F. subsection Urostigma’ lineages); C, functional male syconium of gynodioecious species (staminate and gall flowers; Caricoide lineage); D, functional female syconium of gynodioecious species (pistillate and neuter flowers, or only pistillate flowers; Caricoide lineage). Figure 2. View largeDownload slide Two types of flower arrangements inside the syconium of Ficus. A, staminate flowers only dispersed among pistillate ones (staminate, pistillate and gall flowers; Pharmacosycea and Spherosuke p.p. lineages). B–D, staminate flowers grouped at the ostiolar region sometimes with few irregularly dispersed among the pistillate ones; B, monoecious syconium with staminate, pistillate and gall flowers (Sycomorus, Oreosycea and ‘F. subsection Urostigma’ lineages); C, functional male syconium of gynodioecious species (staminate and gall flowers; Caricoide lineage); D, functional female syconium of gynodioecious species (pistillate and neuter flowers, or only pistillate flowers; Caricoide lineage). DISCUSSION Molecular dating analyses The confidence interval of the stem age of Ficus was estimated to be 7 My older than the age of the Moraceae stem node according to a previous molecular dating study, with the maximum boundaries reaching the Jurassic (Cruaud et al., 2012). To better fit the dating approach with current knowledge, we modified the methodology mainly according to the three incongruence hypotheses highlighted in the Introduction. We calibrated the stem node of Ficus with the fossil achene of F. tethyca and a minimum age of 46 Mya to correspond to the most recent age in the Early Eocene interval, instead of using the crown node and an age of 60 Mya. We used the lognormal distribution to enforce a strict minimum age (i.e. without hard maximum constraints of the uniform distribution used by Xu et al., 2011, and Cruaud et al., 2012), while specifying that the age of the node is unlikely to be much older (diminishing tail of probability) than the set median age (Sauquet, 2013). In addition, we used two unpublished points to calibrate the internal sections (Americanae and Pharmacosycea). In doing this, we observed a significant contradiction between the ages of Ficus between molecular dating studies (Fig. 3). The confidence interval of the stem nodes of the present study is 19 My more recent than the previous studies and 10 My between the crown nodes. We obtained narrower confidence intervals, with 14 Mya for the Ficus stem node (instead of 110 Mya by Cruaud et al., 2012) and 16 Mya for the crown node (instead of 76 and 41 Mya by Xu et al., 2011 and Cruaud et al., 2012, respectively). Unlike previous studies that inferred the median age of the genus to be in the Early Cretaceous (Xu et al., 2011; Cruaud et al., 2012), the new methodology implemented here deduced the confidence interval (66–52 Mya) of the genus in accordance with Moraceae in studies of all angiosperms (CI 78–70 Mya; Magallón et al., 2015) and the whole family (HPD 110–72 Mya; Zerega et al., 2005). This result is also broadly synchronized with the age of the Ficus stem node estimated by Gardner et al. (2017; HPD 65.8–50.1 Mya). In the discussion that follows, we interpret the evolutionary history of Ficus based on palaeobiogeographical events. Figure 3. View largeDownload slide Comparison of the molecular dating analysis in the present study to those of previous studies. Vertical lines indicate confidence intervals, and horizontal lines indicate the median values. The horizontal grey shading represents the age of Moraceae (Magallon et al., 2015) and indicates that the origin of their genera should be below this margin. Figure 3. View largeDownload slide Comparison of the molecular dating analysis in the present study to those of previous studies. Vertical lines indicate confidence intervals, and horizontal lines indicate the median values. The horizontal grey shading represents the age of Moraceae (Magallon et al., 2015) and indicates that the origin of their genera should be below this margin. Origin of Ficus Our results suggest that the stem lineage of Ficus (node 1) arose in the boreotropical region along the Tethys Seaway (Europe to South-East Asia) at the beginning of the Cenozoic, with a later dispersal to North America (Fig. 4A). The boreotropical region was proposed by Wolfe (1975) and has played a major role in elucidating the geographical evolution of seed plant lineages in the Northern Hemisphere (e.g. Tiffney, 1985; Renner, Clausing & Meyer, 2001; Davis et al., 2002; Antonelli et al., 2009; Couvreur et al., 2011; Baker & Couvreur, 2013). The discovery of Ficus fossils from eastern England to western Russia (Collinson, 1989) indicates that Ficus was present along the Tethys Seaway in the early Tertiary, which is compatible with the current (generally tropical) climatic affinities of the genus. The fossil of F. tethyca was found in Hamworthy, southern England, as part of the Eocene boreotropical flora. Other more recent examples are found in Bulgaria, Germany, the Czech Republic and Russia (Collinson, 1989). Figure 4. View largeDownload slide Hypothetical ancestral distribution and the routes of Ficus dispersal plotted on maps of palaeotropical forests (modified from Morley, 2000). A, from the boreotropical region to North America via the North Atlantic land bridge, Africa and South America via the trans-Atlantic route, India and to East Asia (China) via Turgai Strait. B, from East Asia (China) to South-East Asia. C, between North America and South America (Pharmacosycea and Americanae) via the GAARlandia land bridge, from India to Africa (Sycomorus, Oreosycea and ‘subsection Urostigma’ lineages), and from North Africa to Madagascar (Platyphyllae). D, from India/Arabia to Africa (Caricoide clade); from Central Africa to Madagascar and from South-East Asia to the Australian Plate (Sycomorus, Oreosycea and ‘subsection Urostigma’). E, from India and South-East Asia to Africa and Madagascar (Conosycea) and Australian Plate (Malvanthera). F, from the Andean region to Amazon basin and eastern Brazil (Pharmacosycea). The geographical boundaries are based on the ancestral forest areas that were probably inhabited by Ficus and are not exact. Arrows indicate future lineage dispersals. Colour of the polygon represents different lineages on the same map. In Morley (2000), dark grey areas are megathermal rain forests, probably the ancestral habitat of Ficus, and light grey areas are megathermal monsoon forests. According to Scotese (2001), islands were present between Europe and South Asia in the Cenozoic. Figure 4. View largeDownload slide Hypothetical ancestral distribution and the routes of Ficus dispersal plotted on maps of palaeotropical forests (modified from Morley, 2000). A, from the boreotropical region to North America via the North Atlantic land bridge, Africa and South America via the trans-Atlantic route, India and to East Asia (China) via Turgai Strait. B, from East Asia (China) to South-East Asia. C, between North America and South America (Pharmacosycea and Americanae) via the GAARlandia land bridge, from India to Africa (Sycomorus, Oreosycea and ‘subsection Urostigma’ lineages), and from North Africa to Madagascar (Platyphyllae). D, from India/Arabia to Africa (Caricoide clade); from Central Africa to Madagascar and from South-East Asia to the Australian Plate (Sycomorus, Oreosycea and ‘subsection Urostigma’). E, from India and South-East Asia to Africa and Madagascar (Conosycea) and Australian Plate (Malvanthera). F, from the Andean region to Amazon basin and eastern Brazil (Pharmacosycea). The geographical boundaries are based on the ancestral forest areas that were probably inhabited by Ficus and are not exact. Arrows indicate future lineage dispersals. Colour of the polygon represents different lineages on the same map. In Morley (2000), dark grey areas are megathermal rain forests, probably the ancestral habitat of Ficus, and light grey areas are megathermal monsoon forests. According to Scotese (2001), islands were present between Europe and South Asia in the Cenozoic. In North America, there are fossils from the Palaeocene that are presumably leaves of Ficus ancestors (e.g. Dorf, 1938; Bell, 1962; Robison, Hunt & Wolberg, 1982; Collinson, 1989; Johnson, 2002), providing further proof that the genus could have reached this continent early in its evolution. If ancestors of Ficus lived on these continents, they were probably part of the boreotropical flora proposed by Wolfe (1975) and that have been cited for other genera (e.g. Tiffney, 1985; Renner et al., 2001; Davis et al., 2002; Antonelli et al., 2009; Couvreur et al., 2011; Baker & Couvreur, 2013). Pharmacosycea lineage Our results suggest that the first divergence of Ficus (node 2) occurred via a vicariance event between the North American and the boreotropical region along the Tethys Seaway (Europe to South-East Asia) in the Eocene (c. 42 Mya), giving rise to the Pharmacosycea lineage in North America and the remaining Ficus lineages in Eurasia (Fig. 4B). This divergence coincides with the end of the optimal warm conditions necessary for Ficus migration across the North Atlantic land bridge c. 45 Mya (Tiffney, 1985). The North Atlantic land bridge facilitated the exchange between the biota of Europe and North America at the end of the Early Eocene (Tiffney & Manchester, 2001; Pennington & Dick, 2004; Erkens, Maas & Couvreur, 2009), as hypothesized by Cruaud et al. (2012) for Ficus, although with older ages. Ficus section Pharmacosycea is a Neotropical endemic group that retains plesiomorphic characteristics such as monoecious and terrestrial trees without aerial roots, and syconia with staminate flowers scattered among the pistillate flowers (Fig. 2A). This suggests that North America was the place of origin of Ficus section Pharmacosycea. The remaining species in North America (Pharmacosycea lineage, node 3) probably moved to tropical regions of southern North America and northern South America in the Oligocene (c. 28 Mya; Fig. 4C, F). Today, the Andean region is characterized by a high diversity of Pharmacosycea and retains species with probable plesiomorphies, such as cordiform leaves and large syconia (Pederneiras et al., 2015a). The lineage could have migrated southward through the GAARlandia (Greater Antilles + Aves Ridge) land bridge. Iturralde-Vinent (2006) showed that this route enabled exchanges of biota between the American continents c. 32–28 Mya. In addition, Morley (2000) and Zachos et al. (2001) showed that lower temperatures of the mid-latitudes caused a retraction of the remaining North American lineages (probably sometimes becoming extinct) and forced the boreotropical flora to migrate to the southern portions of North America at the Eocene/Oligocene boundary. With this divergence, the lineage became isolated in southern North America and northern South America (Andean region), thereby giving rise to the lineage of F. section Pharmacosycea. Androstiole lineage Our results suggest that the boreotropical region along the Tethys Seaway (Europe to South-East Asia) served as a source of Ficus dispersal to Africa and East Asia (China) in the Eocene (c. 39 Mya; node 4: stem node of the Androstiole lineage), giving rise to the Spherosuke and Androstiole lineages, respectively (Fig. 4A, B). Some Ficus fossils found in Europe from the Eocene corroborate this theory (Collinson, 1989), although the group association of those fossils remains unknown. Some members of the Androstiole clade (F. subsection Urostigma and F. section Oreosycea) are closely associated with drier environments (Berg, 2005), a less common feature in Ficus. This could be an indication that the Androstiole clade crossed via a drier route in the Turgai Strait (Akhmetiev & Beniamovski, 2009), inhabiting the megathermal monsoonal forest that existed in that area (Morley, 2000; Sun & Wang, 2005). The Turgai Strait could have been the causative agent of the vicariance that gave rise to the Androstiole clade and Spherosuke clade in the Eocene. The dispersal from Europe to Africa probably occurred by way of the Iberian Peninsula, which, according to Morley (2003), was a route in the Late Eocene that allowed for dispersal between these continents (node 5). The Androstiole clade (node 12) had possibly undergone a vicariance event in the Eocene, giving rise to the Caricoide clade in China (node 13) and clade 22 (stem node of F. subsection Urostigma) in South-East Asia (Fig. 4C). A dry climate barrier emerged in the Eocene between China and South-East Asia, hindering gene flow between these areas (Tiffney & Manchester, 2001). This theory may explain the vicariance revealed by our results. In addition, the current distribution of Androstiole clade species agrees with the destinations proposed by this vicariance. In China, the Caricoide clade is more species-rich and has more endemic species (62 species with 14 endemics) compared to node 22 (22 species with one endemic) (Zhekun & Gilbert, 2003). Fossils of Ficus from Japan show that some ancestral lineages were present in East Asia in the Eocene (Huzioka & Takahasi, 1970). Climatic changes in the Early Oligocene lowered temperatures in the middle latitudes and probably resulted in southward displacement of the Androstiole lineage species (Fig. 4D). Spherosuke lineage Our analysis suggests that the ancestor of Spherosuke (node 5) was distributed in South-East Asia, Africa and South America in the Eocene (Fig. 4B). However, as previously discussed, South-East Asia is treated as a relic of the boreotropical flora along the Tethys Seaway (Morley, 2000). According to Tiffney (1985), there was a probable immigration route between Europe, North Africa and South-East Asia (including the coast of the Tethys Seaway) in the Early to Middle Eocene. Therefore, the ancestor of Spherosuke was assumed to be in these regions. The Spherosuke lineage probably reached South America from Africa via the Atlantic Ocean, maintaining gene flow through the closest approximation between these continents in the Eocene, facilitated by a series of islands acting as stepping stones postulated by Theide, (1977), Parrish (1993) and Morley (2003), and assisted by animals (e.g. Bauer & Schreiber, 1997; Houle, 1998; Gamble et al., 2011; Mayr, Alvarenga & Mourer-Chauviré, 2011; Antoine et al., 2012; Bond et al., 2015). In addition, the arrival of the Americanae lineage in the Americas is hypothesized to have occurred via South America, as there are many more species on this continent (c. 60–70 species) than in Central and North America (20–30 species) (Berg, 2012). Machado et al. (2018) hypothesized that F. section Americanae originated in the Atlantic forest, i.e. the American biome closest to Africa. With the increased distance between South America and Africa, an event of vicariance probably occurred in the Oligocene, giving rise to the Americanae lineage (node 7) in South America and the Platyphyllae lineage in Africa (node 8; Fig. 4E). A vicariance event gave rise to the Americanae/ Platyphyllae clade (node 6) in Africa/South America and the Conosycea/Malvanthera clade (node 9) in South-East Asia in the Oligocene (c. 32 Mya; Fig. 1: node 5; Fig. 4C). Climate change that resulted in decreased temperatures in middle latitude regions in the early Oligocene caused migration and disjunction of megathermal rain forests (Morley, 2000). Davis et al. (2002) suggested that cooling during the Oligocene resulted in southward retreat in some lineages of Malpighiaceae. The same probably occurred with Ficus, with Ficus becoming extinct in Europe at this time. Following this vicariance, two centres of diversity are suggested based on the distribution and endemism of current species lineages: one centred in Africa/South America (node 6; Fig. 1) and the other in South-East Asia (lineage 9: stem node of Malvanthera). Of the 105 species occurring in Africa, 68 are endemic and belong to the Americanae/Platyphyllae clade (node 6) (Berg & Wiebes, 1992). In India/South-East Asia, 61 species are endemic and belong to the Conosycea/Malvanthera clade (node 9) (Berg & Corner, 2005), without any evidence of endemism in China (Zhekun & Gilbert, 2003). Dispersal from Africa to India during the Eocene could not have occurred given the drifting of these continents closer to each other because of a filter that prevented most of the dispersal after the Maastrichtian (Morley, 2003). The Malvanthera lineage (node 11) probably originated via dispersal from South-East Asia to Australasia in the Miocene (Fig. 4E). Hall & Sevastjanova (2012) showed that the first contact between the Australian continental crust and South-East Asia occurred c. 25 Mya as the two plates collided. Although only a small number of species may have migrated along this route before 12 Mya, some cases of dispersal via volcanic islands probably occurred (Hall, 2013; Crayn et al., 2015). With this, a dispersal event gave rise to the group of Malvanthera in the Australian Plate. To better understand these processes, it is necessary to study species in a more restricted geographical area as indicated by Chantarasuwan et al. (2016). The current distribution of Ficus subsection Malvanthera is centred in Australasia (Berg, 2005), which indicates a strong relationship to that region (Rønsted et al., 2008a). Ficus subsection Conosycea demonstrates a greater affinity with India and South-East Asia (Berg, Pattharahirantricin & Chantarasuwan, 2011; Chaudhary et al., 2012), which led Corner (1985) to suspect a Eurasian origin for this subsection. Our data support the hypothesis that the species in Madagascar arrived from India (as suggested by Berg, 2005) and not from Africa (as suggested by Xu et al., 2011). According to Buerki et al. (2013) and Janssens et al. (2016), the most likely explanation for the high affinity of Madagascar flora to Asia and Australasia is long-distance dispersal via the Indian Ocean in the Early Tertiary (Fig. 4E). Origin of ‘Ficus subsection Urostigma’ Our molecular results suggest that the lineage of ‘F. subsection Urostigma’ originated in South-East Asia (node 22) with geodispersal to China (crown group, node 23) in the Oligocene and Miocene (Fig. 4E). This assumption differs from the proposal by Chantarasuwan et al. (2016), who suggested an origin in Madagascar. The discrepancy between these two results is probably a result of differences in the methods implemented in the two studies. We have included 249 terminals and have applied fossil-based dating in contrast to 27 terminals (without any Androstiole species) and secondary dating implemented based on Xu et al. (2011). According to Magallón et al. (2015), Moraceae originated between 78.9 and 70.4 Mya (stem node), indicating that Xu et al. (2011), who projected that Ficus originated c. 195–85 Mya, may have over-estimated the age of this genus. Of the 27 recognized species of ‘F. subsection Urostigma’, only two occur in Madagascar and five in Africa, whereas the remaining species are in Asia–Australasia (Chantarasuwan et al., 2013). Considering such a distribution, Berg (2005) postulated that the centre of origin of the group should be Asia–Australasia, which is supported by our results. The origin of Ficus: alternative hypotheses Ficus could have originated in North America with dispersal first to Europe and later to China and Africa. This hypothesis agrees with our analysis and with probable fossils found in North America and dated to the Palaeocene (e.g. Dorf, 1938; Bell, 1962; Robison et al., 1982; Johnson, 2002). The Bering Strait could have been a possible route of dispersal from North America to Asia in the Eocene (Tiffney & Manchester, 2001), but this hypothesis is contradicted by the presence of Ficus fossils in Europe dated to the Eocene (Collinson, 1989). Just as Antonelli et al. (2009) did not consider the Bering Strait to be a likely route of dispersal of Rubiaceae (another large tropical woody family) because of the higher palaeolatitude, a similar scenario is expected for Ficus. The whole of Eurasia is a possible centre of origin for the Androstiole clade and, with the retraction of forests in the Oligocene, the members of this clade would have been restricted to South-East Asia with some relics in Europe (Mai & Walther, 1978; Mai, 1997). Assuming that this is the case, F. carica L., the only extant species in Europe, is then a relict taxon of that time. If Androstiole had originated in Eurasia, some lineages would have reached the Americas via the routes present in the past (as indicated for F. sections Americanae and Pharmacosycea in our study); however, there are no extant species of Androstiole distributed in the Americas (Carauta, 1989; Berg, 2005). If the Androstiole clade had originated in Eurasia, the Spherosuke clade might have originated in Africa and then dispersed to South America and India. In this scenario, the ancestral area of the Pharmacosycea lineage is in North America, that of Androstiole is in Eurasia and that of the Spherosuke lineage is in the Southern Hemisphere. However, the areas of Pharmacosycea and Spherosuke should be more closely related than those of Pharmacosycea and Androstiole because the first two share important characteristics such as monoecy and staminate flowers scattered among the pistillate flowers. If India was the centre of origin for the Conosycea–Malvanthera clade (node 9), the ancestor of the Spherosuke lineage dispersed to Africa in the Early Eocene and, with the separation of the plates, a vicariance event gave rise to the clade. As India approached South-East Asia, the flora of India dispersed there. According to Morley (2003), these routes were probable and agree with the current distribution of the species. Although our data failed to capture the importance of India in the evolutionary history of Ficus because of the lack of endemic taxa among our samples, this territory is certainly a great source of speciation and an important route of dispersal. Morphological and phylogenetic coherence The systematics of Ficus has been controversial because phylogenetic results have not corroborated the morphological classification (e.g. Cruaud et al., 2012). Pharmacosycea and Spherosuke are subgenera that have always been resolved as paraphyletic, the present study included. What could be the underlying cause of this observation? Is it possible that the molecular markers used thus far are not accurately conveying the phylogenetic history of Ficus subdivisions? Alternatively, perhaps the morphological classification of Ficus is in need of revision. We believe that the phylogenetic trees presented here and those in previous studies provide a reliable evolutionary history of Ficus. The arrangement of staminate flowers in the syconium (Fig. 2) supports these phylogenetic reconstructions. The state ‘flowers grouped in the apical region (around the ostiole)’ is as an apomorphic state (Fig. 2B–D) that supports the entire clade 12. Section Oreosycea and the ‘subsection Urostigma’, both of which are part of clade 12 based on molecular analysis, were the only two groups that had this character but did not belong to any taxon in clade 12 according to the previous morphological classification (subgenera Pharmacosycea and Spherosuke, respectively). Staminate flowers grouped around the ostiole is present in 62% of extant Ficus spp. and represents an important step in the evolution of the genus. This arrangement was probably the intermediate step in the evolution of a more specialized condition (i.e. dioecy) in Ficus (species of the Caricoide clade and some species of F. subgenus Sycomorus; Fig. 1). Origin of Moraceae Our results concerning Ficus strengthen the hypotheses of a Eurasian origin of Moraceae. Zerega et al. (2005) used molecular dating to test the hypotheses of a Gondwanan vs. Laurasian origin of Moraceae. Although their results were inconclusive, fossil evidence, palaeoclimate data and lineage ages led them to suggest that a Eurasian origin with multiple opportunities for migration into the Southern Hemisphere was more plausible. Our data strengthen the Eurasian origin of Moraceae, which could have been the plausible scenario for all other tribes of Moraceae, and support the arguments presented by Zerega et al. (2005). The ancestral areas of Castilleae and Dorstenieae (the stem and crown nodes), and those of Ficeae, were placed in Eurasia; however, these results are based on partial data of tribes and should be further explored to elucidate the origin of Moraceae. Future research directions Molecular analyses of Ficus over the past 20 years have refined our understanding of phylogenetic relationships and their implications for systematics and biogeography. Future studies should increase the number of sampled taxa to include > 50% of species (mainly Ficus section Oreosycea, F. subgenus Ficus, F. subgenus Terega and F. subgenus Synoecia) as that will help to further explain the cladogenesis of Ficus in China, India and Australasia. Endemic species play an important role in the analysis of biogeography. In the present work, we obtained endemic species for all areas sampled, with only India and Madagascar being represented by few taxa (< 30). This may have affected the reconstruction of the ancestral areas. The inclusion of these taxa will further test the hypothesis regarding the biogeographical history of these regions. Finally, taxonomic revisions and a new classification of Ficus are necessary, especially for those seeking new morphological evidence, to explain the current clades of Ficus. SUPPORTING INFORMATION Additional Supporting Information may be found in the online version of this article at the publisher’s web-site. Figure S1. Chronogram of Ficus (249 species) inferred by BEAST using ITS, ETS, G3pdh, ncpGS and GBSSI sequence data. Blue bars indicate the 95% high posterior distribution (HPD) intervals of the divergence time. Posterior probabilities are given above the branch. Figure S2. Chronogram of Ficus with ancestral area reconstruction inferred by S-DIVA analyses. Ancestral areas: A, North America; B, western South America (Andes); C, eastern South America; D, Africa; E, South-East Asia; F, India; G, Australasia; H, Madagascar; I, West to East Asia (Turkey to China). Figure S3. Chronogram of Ficus with ancestral area reconstruction inferred by DEC analyses. Ancestral areas: A, North America; B, western South America (Andes); C, eastern South America; D, Africa; E, South-East Asia; F, India; G, Australasia; H, Madagascar; I, West to East Asia (Turkey to China). Figure S4. Chronogram of Ficus with ancestral area reconstruction inferred by S-DEC analyses. Ancestral areas: A, North America; B, western South America (Andes); C, eastern South America; D, Africa; E, South-East Asia; F, India; G, Australasia; H, Madagascar; I, West to East Asia (Turkey to China). ACKNOWLEDGEMENTS The authors thank Fundação de Amparo à Pesquisa do Estado do Rio de Janeiro (FAPERJ, E-26/202.862/2016, E-26/202.863/2016 and 202.411/2017), Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq, 440502/2015-2; 88882.156814/2017-01), and Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP, 2016/50385-4) for funding this research. REFERENCES Akhmetiev MA , Beniamovski VN . 2009 . Paleogene floral assemblages around epicontinental seas and straits in Northern Central Eurasia: proxies for climatic and paleogeographic evolution . Geologica Acta 7 : 297 – 309 . Antoine PO , Marivaux L , Croft DA , Billet G , Ganerød M , Jaramillo C , Martin T , Orliac MJ , Tejada J , Altamirano AJ , Duranthon F , Fanjat G , Rousse S , Duranthon F . 2012 . Middle Eocene rodents from Peruvian Amazonia reveal the pattern and timing of caviomorph origins and biogeography . Proceedings of the Royal Society B 279 : 1319 – 1326 . Antonelli A , Nylander JA , Persson C , Sanmartín I . 2009 . Tracing the impact of the Andean uplift on Neotropical plant evolution . Proceedings of the National Academy of Sciences of the United States of America 106 : 9749 – 9754 . Anzótegui LM , Herbst R . 2004 . Megaflora (hojas y frutos) de la Formación San José (Mioceno Medio) en río Seco, departamento Santa María, provincia de Catamarca, Argentina . Ameghiniana 41 : 423 – 436 . Baker WJ , Couvreur TLP . 2013 . Global biogeography and diversification of palms sheds light on the evolution of tropical lineages. I. Historical biogeography . Journal of Biogeography 40 : 274 – 285 . Bauer K , Schreiber A . 1997 . Double invasion of Tertiary island South America by ancestral New World monkeys ? Biological Journal of the Linnean Society 60 : 1 – 20 . Bell WA . 1962 . Upper Cretaceous floras of the Dunvegan, Bad Heart, and Milk River Formations of Western Canada . Geological Survey of Canada 94 : 1 – 76 . Berg CC . 2003 . Flora Malesiana precursor for the treatment of Moraceae 1: the main subdivision of Ficus: the subgenera . Blumea-Biodiversity, Evolution and Biogeography of Plants 48 : 166 – 177 . Berg CC . 2005 . Moraceae diversity in a global perspective . Biologiska Skrifter 55 : 423 – 440 . Berg CC . 2012 . Moraceae . Flora Mesoamericana 2 : 1 – 91 . Berg CC , Corner EJH . 2005 . Moraceae (Ficus) . Flora Malesiana, Series I, Seed Plants 17 , part. 2: 1 – 730 . Berg CC , Wiebes JT . 1992 . African fig trees and fig wasps . Verhandelingen der Koninklijke Nederlandse Akademie van Wetenschappen, 2de reeks 89 : 1 – 298 . Berg CC , Pattharahirantricin N , Chantarasuwan B . 2011 . Moraceae . Flora of Thailand 10 : 475 – 675 . Bond M , Tejedor MF , Campbell KE Jr , Chornogubsky L , Novo N , Goin F . 2015 . Eocene primates of South America and the African origins of New World monkeys . Nature 520 : 538 – 541 . Bozukov V , Ivanov D , Utescher T . 2013 . Ficus palamarevii sp. nov. (Moraceae), a new subtropical element in the Bulgarian Paleogene flora . Comptes rendus de l’Académie bulgare des Sciences 66 : 1733 – 1738 . Buerki S , Devey DS , Callmander MW , Phillipson PB , Forest F . 2013 . Spatio‐temporal history of the endemic genera of Madagascar . Botanical Journal of the Linnean Society 171 : 304 – 329 . Carauta JPP . 1989 . Ficus (Moraceae) no Brasil: conservação e taxonomia . Albertoa 2 : 1 – 365 . Chantarasuwan B , Berg CC , van Welzen PC . 2013 . A revision of Ficus subsection Urostigma (Moraceae) . Systematic Botany 38 : 653 – 686 . Chantarasuwan B , Berg CC , Kjellberg F , Rønsted N , Garcia M , Baider C , van Welzen PC . 2015 . A new classification of Ficus subsection Urostigma (Moraceae) based on four nuclear DNA markers (ITS, ETS, G3pdh, and ncpGS), morphology and leaf anatomy . PLoS One 10 : e0128289 . Chantarasuwan B , Rønsted N , Kjellberg F , Sungkaew S , Welzen PC . 2016 . Palaeotropical intercontinental disjunctions revisited using a dated phylogenetic hypothesis with nearly complete species level sampling of Ficus subsect. Urostigma (Moraceae) . Journal of Biogeography 43 : 384 – 397 . Chaudhary LB , Sudhakar JV , Kumar A , Bajpai O , Tiwari R , Murthy GVS . 2012 . Synopsis of the genus Ficus L. (Moraceae) in India . Taiwania 57 : 193 – 216 . Clement WL , Weiblen GD . 2009 . Morphological evolution in the mulberry family (Moraceae) . Systematic Botany 34 : 530 – 552 . Collinson ME . 1989 . The fossil history of the Moraceae, Urticaceae (including Cecropiaceae), and Cannabaceae . In: Crane PR , Blackmore S , eds. Evolution, systematics, and fossil history of the Hamamelidae, Vol. 2, ‘Higher’ Hamamelidae, Systematics Ass., Special Volume 40B . Oxford : Clarendon Press , 319 – 339 . Corner EJH . 1960a . Taxonomic notes on Ficus Linn., Asia and Australasia . The Gardens’ Bulletin, Singapore 17 : 368 – 485 . Corner EJH . 1960b . Taxonomic notes on Ficus Linn., Asia and Australasia . The Gardens’ Bulletin, Singapore 18 : 1 – 69 . Corner EJH . 1965 . Check-list of Ficus in Asia and Australasia with keys to identification . The Gardens’ Bulletin, Singapore 21 : 1 – 186 . Corner EJH . 1985 . The distribution and origin of strangling figs (subgen . Urostigma). Allertonia 4 : 138 – 153 . Couvreur TLP , Pirie MD , Chatrou LW , Saunders RMK , Su YCF , Richardson JE , Erkens RHJ . 2011 . Early evolutionary history of the flowering plant family Annonaceae: steady diversification and boreotropical geodispersal . Journal of Biogeography 38 : 664 – 680 . Crayn DM , Costion C , Harrington MG . 2015 . The Sahul–Sunda floristic exchange: dated molecular phylogenies document Cenozoic intercontinental dispersal dynamics . Journal of Biogeography 42 : 11 – 24 . Cruaud A , Rønsted N , Chantarasuwan B , Chou LS , Clement WL , Couloux A , Cousins B , Genson G , Harrison RD , Hanson PE , Hossaert-McKey M , Jabbour-Zahab R , Jousselin E , Kerdelhué C , Kjellberg F , Lopez-Vaamonde C , Peebles J , Peng YQ , Pereira RA , Schramm T , Ubaidillah R , van Noort S , Weiblen GD , Yang DR , Yodpinyanee A , Libeskind-Hadas R , Cook JM , Rasplus JY , Savolainen V . 2012 . An extreme case of plant–insect codiversification: figs and fig-pollinating wasps . Systematic Biology 61 : 1029 – 1047 . Davis CC , Bell CD , Mathews S , Donoghue MJ . 2002 . Laurasian migration explains Gondwanan disjunctions: evidence from Malpighiaceae . Proceedings of the National Academy of Sciences of the United States of America 99 : 6833 – 6837 . Dorf E . 1938 . Upper Cretaceous floras of the Rocky Mountain region . Carnegie Institution of Washington Publication 508 : 1 – 78 . Doweld AB . 2016 . Ficus changii, Ficus fengkaiensis, Ficus obtusatoides and Ficus tethyca (Moraceae), new replacement names . Kew Bulletin 71 : 29 . Drummond AJ , Rambaut A . 2007 . BEAST: Bayesian evolutionary analysis by sampling trees . BMC Evolutionary Biology 7 : 214 . Edgar RC . 2004 . MUSCLE: multiple sequence alignment with high accuracy and high throughput . Nucleic Acids Research 32 : 1792 – 97 . Erkens RHJ , Maas JW , Couvreur TLP . 2009 . From Africa via Europe to South America: migrational route of a species-rich genus of Neotropical lowland rain forest trees (Guatteria, Annonaceae) . Journal of Biogeography 36 : 2338 – 2352 . Forest F . 2009 . Calibrating the Tree of Life: fossils, molecules and evolutionary timescales . Annals of Botany 104 : 789 – 794 . Gamble T , Bauer AM , Colli GR , Greenbaum E , Jackman TR , Vitt LJ , Simons AM . 2011 . Coming to America: multiple origins of New World geckos . Journal of Evolutionary Biology 24 : 231 – 244 . Gardner EM , Sarraf P , Williams EW , Zerega NJC . 2017 . Phylogeny and biogeography of Maclura (Moraceae) and the origin of an anachronistic fruit . Molecular Phylogenetics and Evolution 117 : 49 – 59 . Gernhard T , Hartmann K , Steel M . 2008 . Stochastic properties of generalised Yule models, with biodiversity applications . Journal of Mathematical Biology 57 : 713 – 735 . Hall R . 2002 . Cenozoic geological and plate tectonic evolution of SE Asia and the SW Pacific: computer-based reconstructions, model and animations . Journal of Asian Earth Sciences 20 : 353 – 431 . Hall R . 2013 . The palaeogeography of Sundaland and Wallacea since the Late Jurassic . Journal of Limnology 72 : 1 – 17 . Hall R , Sevastjanova I . 2012 . Australian crust in Indonesia . Australian Journal of Earth Sciences 59 : 827 – 844 . Harrison RD , Rønsted N , Xu L , Rasplus JY , Cruaud A . 2012 . Evolution of fruit traits in Ficus subgenus Sycomorus (Moraceae): to what extent do frugivores determine seed dispersal mode ? PLoS One 7 : e38432 . Ho SYW 2007 . Calibrating molecular estimates of substitution rates and divergence times in birds . Journal of Avian Biology 38 : 409 – 414 . Houle A . 1998 . Floating islands: a mode of long-distance dispersal for small and medium-sized terrestrial vertebrates . Diversity and Distributions 4 : 201 – 216 . Huzioka K , Takahasi E . 1970 . The Eocene flora of the Ube coal-field, southwest Honshu, Japan . Journal of the Mining College, Akita University, Series A: Mining Geology 4 : 1 – 88 . Iturralde-Vinent MA . 2006 . Meso-Cenozoic Caribbean paleogeography: implications for the historical biogeography of the region . International Geology Review 48 : 791 – 827 . Janssens SB , Groeninckx I , De Block PJ , Verstraete B , Smets EF , Dessein S . 2016 . Dispersing towards Madagascar: biogeography and evolution of the Madagascan endemics of the Spermacoceae tribe (Rubiaceae) . Molecular Phylogenetics and Evolution 95 : 58 – 66 . Johnson KR . 2002 . Megaflora of the Hell Creek and lower Fort Union Formations in the western Dakotas: vegetational response to climate change, the Cretaceous–Tertiary boundary event, and rapid marine transgression . Geological Society of America Special Paper 361 : 329 – 391 . Jousselin E , Rasplus JY , Kjellberg F . 2003 . Convergence and coevolution in a mutualism: evidence from a molecular phylogeny of Ficus . Evolution 57 : 1255 – 1269 . King G . 1888 . The species of Ficus of the Indo-Malayan and Chinese countries . Annals of the Royal Botanical Garden, Calcutta 1 : 1 – 185 . Kjellberg F , Jousselin E , Hossaert-McKey M , Rasplus JY . 2005 . Biology, ecology, and evolution of fig-pollinating wasps (Chalcidoidea, Agaonidae) . In: Raman A , Schaefer CW , Withers TM , eds. Biology, ecology, and evolution of gall-inducing arthropods, Vol. 2 . Enfield : Science Publishers , 539 – 572 . Lanfear R , Calcott B , Ho SYW , Guindon S . 2012 . PartitionFinder: combined selection of partitioning schemes and substitution models for phylogenetic analyses . Molecular Biological Evolution 29 : 1695 – 1701 . Machado AFP , Rønsted N , Pereira RAS , Bruun-Lund S , Queiroz LP . 2018 . Atlantic forest to the all Americas: biogeographical history and divergence times of Neotropical Ficus (Moraceae) . Molecular Phylogenetics and Evolution 122 : 46 – 58 . Magallón S , Gómez‐Acevedo S , Sánchez‐Reyes LL , Hernández‐Hernández T . 2015 . A metacalibrated time‐tree documents the early rise of flowering plant phylogenetic diversity . New Phytologist 207 : 437 – 453 . Mai DH . 1997 . Die oberoligozänen Floren am Nordrand der Sächsischen Lausitz . Palaeontographica Abteilung B 244 : 1 – 124 . Mai DH , Walther H . 1978 . Die Floren des Haselbacher Serie im Weisselster-Becken (Bezirk Leipzig, DDR) . Abhandlungen des Staatlichen Museums fuer Mineralogie und Geologie zu Dresden 28 : 1 – 200 . Mayr G , Alvarenga H , Mourer-Chauviré C . 2011 . Out of Africa: fossils shed light on the origin of the hoatzin, an iconic Neotropic bird . Naturwissenschaften 98 : 961 – 966 . Miller MA , Pfeiffer W , Schwartz T . 2010 . Creating the CIPRES Science Gateway for inference of large phylogenetic trees . In: Gateway Computing Environments Workshop (GCE) . Piscataway: IEEE , 1 – 8 . Morley RJ . 2000 . Cretaceous and Tertiary climate change and the past distribution of megathermal rainforests . In: Bush MB , Flenley JR , Gosling WD , eds. Tropical rainforest responses to climatic change, 2nd edn . Berlin : Springer-Verlag , 1 – 34 . Morley RJ . 2003 . Interplate dispersal paths for megathermal angiosperms . Perspectives in Plant Ecology, Evolution and Systematics 6 : 5 – 20 . Parrish JT . 1993 . The paleogeography of the opening South Atlantic . In: George W , Lavocat R , eds. The Africa–South America connection . Oxford : Clarendon Press , 8 – 41 . Pederneiras LC , Romaniuc-Neto S , Mansano VF . 2015a . Molecular phylogenetics of Ficus section Pharmacosycea and the description of Ficus subsection Carautaea (Moraceae) . Systematic Botany 40 : 504 – 509 . Pederneiras LC , Carauta JPP , Romaniuc-Neto S , Mansano VF . 2015b . An overview of the infrageneric nomenclature of Ficus . Taxon 64 : 589 – 594 . Peñalver E , Engel MS , Grimaldi DA , 2006 . Two wasp families rare in the fossil record (Hymenoptera): Perilampidae and Megaspilidae from the Miocene of Spain . American Museum Novitates 3540 : 1 – 12 . Pennington RT , Dick CW . 2004 . The role of immigrants in the assembly of the South American rainforest tree flora . Philosophical Transactions of the Royal Society B: Biological Sciences 359 : 1611 – 1622 . Rambaut A , Drummond AJ . 2007 . TreeAnnotator . Available at: http://beast.community/treeannotator. Rambaut A , Suchard MA , Xie W , Drummond AJ . 2013 . Tracer version 1.6.0. Available at: http://BEAST.bio.ed.ac.uk/tracer. Ree RH , Smith SA , 2008 . Maximum likelihood inference of geographic range evolution by dispersal, local extinction, and cladogenesis . Systematic Biology 57 : 4 – 14 . Renner SS . 2016 . Available data point to a 4‐km‐high Tibetan Plateau by 40 Ma, but 100 molecular‐clock papers have linked supposed recent uplift to young node ages . Journal of Biogeography 43 : 1479 – 1487 . Renner SS , Clausing G , Meyer K . 2001 . Historical biogeography of Melastomataceae: the roles of Tertiary migration and long-distance dispersal . American Journal of Botany 88 : 1290 – 1300 . Robison CR , Hunt A , Wolberg DL . 1982 . New Late Cretaceous leaf locality from lower Kirtland Shale member, Bisti area, San Juan Basin, New Mexico . New Mexico Geology 4 : 42 – 45 . Rønsted N , Salvo G , Savolainen V . 2007 . Biogeographical and phylogenetic origins of African fig species (Ficus section Galoglychia) . Molecular Phylogenetics and Evolution 43 : 190 – 201 . Rønsted N , Weiblen GD , Cook JM , Salamin N , Machado CA , Savolainen V . 2005 . 60 million years of co-divergence in the fig–wasp symbiosis . Proceedings of the Royal Society B 272 : 2593 – 2599 . Rønsted N , Weiblen GD , Clement WL , Zerega NJC , Savolainen V . 2008b . Reconstructing the phylogeny of figs (Ficus, Moraceae) to reveal the history of the fig pollination mutualism . Symbiosis 45 : 45 – 55 . Rønsted N , Weiblen GD , Savolainen V , Cook JM . 2008a . Phylogeny, biogeography, and ecology of Ficus section Malvanthera (Moraceae) . Molecular Phylogenetics and Evolution 48 : 12 – 22 . Sata N . 1944 . A monographic study of the genus Ficus from the point of view of economic botany . Contributions from the Institute of Horticulture and Economic Botany, Faculty of Agriculture, Taihoku Imperial University, 32 . Sauquet H . 2013 . A practical guide to molecular dating . Comptes Rendus Palevol 12 : 355 – 367 . Scotese CR . 2001 . Atlas of earth history . Arlington : University of Texas at Arlington , Department of Geology, PALEOMAP Project. Srivastava G , Srivastava R , Mehrotra RC . 2011 . Ficus palaeoracemosa sp. nov. – a new fossil leaf from the Kasauli Formation of Himachal Pradesh and its palaeoclimatic significance . Journal of Earth System Science 120 : 253 – 262 . Sun X , Wang P . 2005 . How old is the Asian monsoon system? Palaeobotanical records from China . Palaeogeography, Palaeoclimatology, Palaeoecology 222 : 181 – 222 . Takhtajan A . 1986 . Floristic regions of the world . Berkeley : University of California Press . Theide J . 1977 . Subsidence of aseismic ridges: evidence from sediments on Rio Grande Rise (southwest Atlantic Ocean) . American Association of Petroleum Geologists Bulletin 61 : 929 – 940 . Tiffney BH . 1985 . The Eocene North Atlantic land bridge: its importance in Tertiary and modern phytogeography of the Northern Hemisphere . Journal of the Arnold Arboretum 66 : 243 – 273 . Tiffney BH , Manchester SR . 2001 . The use of geological and paleontological evidence in evaluating plant phylogeographic hypotheses in the Northern Hemisphere Tertiary . International Journal of Plant Sciences 162 : S3 – S17 . Weiblen GD . 2000 . Phylogenetic relationships of functionally dioecious Ficus (Moraceae) based on ribosomal DNA sequences and morphology . American Journal of Botany 87 : 1342 – 1357 . Wolfe JA . 1975 . Some aspects of plant geography of the Northern Hemisphere during the late Cretaceous and Tertiary . Annals of the Missouri Botanical Garden 62 : 264 – 279 . Xu L , Harrison RD , Yang P , Yang DR . 2011 . New insight into the phylogenetic and biogeographic history of genus Ficus: vicariance played a relatively minor role compared with ecological opportunity and dispersal . Journal of Systematics and Evolution 49 : 546 – 557 . Yu Y , Harris AJ , He XJ . 2010 . S-DIVA (statistical dispersal-vicariance analysis): a tool for inferring biogeographic histories . Molecular Phylogenetics and Evolution 56 : 848 – 850 . Yu Y , Harris AJ , He XJ . 2014 . RASP, Reconstruct Ancestral State in Phylogenies, version 3.0 . Available at: http//mnh.scu.edu.cn/soft/blog/RASP. Yu Y , Harris AJ , Blair C , He XJ . 2015 . RASP (reconstruct ancestral state in phylogenies): a tool for historical biogeography . Molecular Phylogenetics and Evolution 87 : 46 – 49 . Zachos J , Pagani M , Sloan L , Thomas E , Billups K . 2001 . Trends, rhythms, and aberrations in global climate 65 Ma to present . Science 292 : 686 – 693 . Zerega NJC , Clement WL , Datwyler SL , Weiblen GD . 2005 . Biogeography and divergence times in the mulberry family (Moraceae) . Molecular Phylogenetics and Evolution 37 : 402 – 416 . Zhekun Z , Gilbert M . 2003 . Moraceae . Flora of China 5 : 21 – 73 . APPENDIX Species of Ficus and outgroups included in the study, with voucher information and GenBank accession numbers for ITS, ETS, G3pdh, ncpGS and GBSSI sequences. X = no sequence obtained. Antiaropsis decipiens K.Schum., Weiblen 1706 (MIN), AY730142, EU084403, X, X, X; Broussonetia papyrifera (L.) L’Hér. ex Vent., unknown, BQUCH0114, HM623778, X, X, X, X; Castilla elastica Sessé, Weiblen 1433 (MIN), FJ916997, AY730232, EF092327, X, EU084352; Ficus abutilifolia (Miq.) Miq., unknown, AY730091, AY730180, EF092348, X, X; F. adelpha K.Schum., Lauterb., Weiblen 1689 (MIN), DQ367656, EU084450, DQ367615, X, X; F. adenosperma Miq., Weiblen 1764 (MIN), AF165374, EF092321, EF092374, X, X; F. adhatodifolia Schott, E. Zardini 12836 (NY), KM186256, X, KM186205, X, X; F. adhatodifolia, L.C. Pederneiras 752 (SP), KM186257, KM186226, KM186204, X, X; F. adhatodifolia, L.C. Pederneiras 643 (SP), KM186258, KM186227, KM186203, X, X; F. albert-smithii Standl., unknown, AY730069, AY730157, X, X, X; F. alongensis Gagnep., Steward and Cheo 1187 (P), KJ845962, KJ845902, KJ846015, X, X; F. altissima Blume, XL23, JN117617, JN117654, X, X, EU084363; F. americana Aubl., Rønsted 154 (K), AY730070, AY730158, EF092339, DQ455613, X; F. ampelas K.D.Koenig ex Roxb., Takeuchi 14565 (K), EU091659, X, X, X, X; F. andicola Standl., unknown, AY730071, AY730159, EF092340, X, X; F. annulata Blume, Shine/Rønsted 293 (HITBC), EU091578, EU084417, EU087622, X, X; F. arbuscula K.Schum., Lauterb., Weiblen 500 (MIN), EU091617, X, X, X, EU084375; F. arfakensis King, Weiblen 1726 (MIN), DQ367657, EU084451, DQ367617, X, X; F. asperifolia Miq., Compton (LDS), EU091661, EU084484, EF092394, X, X; F. augusta Corner, unknown, EF545651, EF538767, EF538787, X, X; F. aurata Miq., Jousselin s.n., EU091642, EU084469, X, X, X; F. aurea Nutt., Rønsted 130 (K), EU091598, EU084431, EU087636, X, X; F. auriculata Lour., Rønsted 264 (HITBC), AF165376, FJ812281, EU087653, X, X; F. baeuerlenii King, Isua B121 (K), AF165377, EU084474, X, X, X; F. beipeiensis S.S.Chang, XL30, JN117619, X, JN117686, X, X; F. benghalensis L., unknown, AY730065, AY730153, X, X, X; F. benjamina L., Rønsted 81 (C)/Rønsted 179 (AAU/K), AY063559, AY063520, EF092333, EU084305, EU084364; F. bernaysii King, unknown, AF165378, X, DQ367618, X, DQ367638; F. binnendijkii (Miq.) Miq., unknown, AY063561, AY063522, EF092334, X, X; F. bizanae Hutch. & Burtt Davy, unknown, DQ455636, DQ455670, X, X, X; F. botryocarpa Miq., Weiblen 2101 (MIN), AF165379, EU084452, DQ367619, X, DQ367639; F. botryoides Baker, unknown, AF165380, X, X, X, X; F. brachypoda (Miq.) Miq., Dixon (DNA), EF545652, EF538768, EF538788, EU084309, X; F. broadwayi Urb., unknown, AY730072, AY730160, EF092341, X, X; F. bubu Warb., Forest 339 (NBG), DQ455637, DQ455671, EU087642, X, X; F. bullenei I.M.Johnst., unknown, X, X, EU089833, X, X; F. burkei Miq., Forest 345 (NBG), AY730095, AY730184, X, DQ455621, X; F. burtt-davyi Hutch., Forest 328 (NBG), DQ455647, DQ455675, EU087643, X, X; F. callophylla Blume, Harrison 601 (PUH), EU091582, X, X, X, X; F. callosa Willd., unknown, AY063565, AY063526, EF092367, X, X; F. carchiana C.C.Berg, B. Stahl 6385 (BG), KM186255, X, X, X, X; F. carica L., Rønsted 96 (C), EU091637, EU084464, X, X, EU084382; F. caulocarpa (Miq.) Miq. 1, Weiblen 2384 (MIN), EU091573, EU084413, EU087619, X, X; F. caulocarpa (Miq.) Miq. 2, Chantarasuwan 261111-1 (L), KJ845954, KJ845894, KJ846009, X, X; F. cestrifolia Schott ex Spreng., unknown, AY730076, EF092342, AY730164, X, X; F. chiapensis Lundell, Rønsted 301 (HITBC), EU091638, EU084465, EU087671, X, X; F. chrysolepis Miq., Weiblen 2353 (MIN), EU091583, EU084420, EU087624, X, X; F. citrifolia Mill., Rønsted 112 (K), AY730077, AY730165, AY967955, DQ455615, X; F. colubrinae Standl., unknown, X, EU089848, X, X, X; F. concinna (Miq.) Miq., Chantarasuwan 120910–5 (L), KJ845991, KJ846037, KJ845930, KJ846071, X; F. condensa King, Jousselin, AY063577, AY063538, X, EU084325, X; F. congesta Roxb., unknown, AY730136, AY730225, DQ367620, X, DQ367640; F. conocephalifolia Ridl., Weiblen 1754 (MIN), AF165381, EU084486, X, X, X; F. consociata Blume, unknown, AY063558, AY063519, X, X, X; F. copiosa Steud., Weiblen 57 (A), AF165382, EF092324, EF092395, X, EU084390; F. cordata Thunb., Theson 3363 (WAG), KJ845975, KJ845914, KJ846022, KJ846063, X; F. cordatula Merr., Harrison 606 (PUH), EU091584, EU084421, EU087625, EU084307, X; F. coronata Spin., unknown, AY730131, AY730218, EF092396, X, X; F. costaricana (Liebm.) Miq., Oyama, UNAM, Mexico, EU091602, EU084435, AY967952, X, X; F. crassipes F.M. Bailey, unknown, AY730112, AY730201, EF538789, X, X; F. crassivenosa W.C.Burger, L.C. Pederneiras 680 (SP), X, KM186228, KM186206, X, X; F. crassivenosa, J.S. Barreto-Silva 2156 (SP), KM186254, X, KM186207, X, X; F. craterostoma Mildbr., Burret, Forest 340 (NBG), AY730097, AY730186, EF092349, DQ455622, X; F. crocata (Miq.) Miq., Rønsted 92 (C), DQ455667, DQ455686, EF092343, DQ455618, X; F. curtipes Corner, XL28, JN117627, JN117657, JN117693, X, X; F. cyathistipula Warb., unknown, DQ455657, DQ455679, X, X, X; F. cyathistipuloides De Wild., Rønsted 136 (K), AY063563, AY063524, EU087645, X, X; F. cyrtophylla Miq., Rønsted 124 (K), EU091664, EU084488, JN117694, X, X; F. dammaropsis Diels, Weiblen 1744 (MIN), AF165383, EU084445, DQ367621, X, DQ367641; F. deltoidea Jack., unknown, AY063579, AY063540, EF092378, X, X; F. densifolia Miq., Baider CB2421 (L), KJ845983, KJ845922, KJ846030, KJ846068, X; F. destruens F.Muell. ex C.T.White, unknown, EF545653, EF538790, EF538769, X, X; F. dewolfii Pederneiras & Romaniuc-Neto, J. Betancur 7420 (SP), KM186253, X, KM186208, X, X; F. diversiformis Miq., Samuel 3100 (K), AY730128, AY730215, EF092392, X, X; F. drupacea Thunb., unknown, AY730066, AY730154, EF092335, X, X; F. dugandii Standl., unknown, X, X, AY967957, X, X; F. edelfeltii King, unknown, AF165385, AY730209, X, X, X; F. elastica Roxb. ex Hornem., unknown, AY063555, AY063516, EF092338, X, X; F. elasticoides De Wild., unknown, AY730103, AY730192, EF092354, X, X; F. erecta Thunb., Rønsted 134 (K), AY730121, AY730211, EF092379, EU084330, X; F. esquiroliana H.Lév., Wang, Chen, Xiong 2011007, KX055726, KX055618, KF811032, X, X; F. eugeniifolia (Liebm.) Hemsl., unknown, AY730078, AY730166, X, X, X; F. exasperata Vahl, Rønsted 217 (K)/Rønsted 97 (K), EU091665, EU084489, EU087683, X, EU084392; F. eximia Schott, unknown, AY730079, AY730167, EF092344, X, X; F. fistulosa Reinw. ex Blume, XL3, JN117629, JN117659, EF092375, X, EU084379; F. formosana Maxim., ZHS TW070/ZHS TW011, HQ890711, X, HQ890585, X, X; F. forstenii Miq., Harrison 632 (PUH), EU091587, X, EU087626, X, X; F. fulva Reinw. ex Blume, Li 2009341, KX055695, KX055664, EU087675, EU084335, X; F. geniculata Kurz, Chantarasuwan 150910–1 (L), KJ845940, KJ845882, KJ845999, KJ846044, X; F. gigantosyce Dugand, K. Young 4117 (F), KM186252, X, KM186209, X, X; F. glabella Blume, unknown, KJ845960, KJ845900, KJ846013, KJ846055, X; F. glaberrima Blume, Chantarasuwan 110910–2 (L), KJ845996, KJ845935, KJ846041, KJ846076, X; F. glabrata Kunth, J.S. Barreto-Silva 2150 (SP), KM186250, X, KM186211, X, X; F. glabrata, L.C. Pederneiras 671 (SP), KM186251, KM186229, KM186210, X, X; F. glandifera Summerh., unknown, AY730113, AY730202, EF092361, X, X; F. glumosa Delile, Chase 19873 (K), AY063562, AY063523, X, EU084316, X; F. grossularioides Burm. f., Jousselin, AY063591, AY063548, EF092385, EU084336, EU084386; F. gul Laut. & K.Schum., Takeuchi 15019 (K), AY730132, AY730219, EF092397, EU084349, X; F. habrophylla Seem., Weiblen 1224 (MIN), EU091567, EU084408, EU087612, X, X; F. henneana Miq., J.R. Maconochie 2208 (L), KJ845967, KJ845907, KJ846016, KJ846058, X; F. henryi Diels, Chase 19876 (K)/Rønsted 289 (HITBC), EU091639, EU084466, EU087672, EU084331, X; F. hesperidiiformis King, unknown, EF545655, EF538770, EF092362, X, X; F. heteromeka Corner, unknown, EF545656, EF538772, EF538791, X, X; F. heteropleura Blume, unknown, AY730133, AY730220, EF092400, X, X; F. hirta Vahl, Li, Chen, Wang, Xiong 2009189, KX055668, KX055575, X, X, X; F. hirta, Rønsted 265 (HITBC), AY730127, EU084473, EF092386, X, X; F. hispida L.f., XL 24/Rønsted 171 (K/AAU)/Rønsted 88 (C), JN117634, EU084454, JN117700, EU084326, X; F. hispidioides S.Moore, unknown, AF165388, AY730227, DQ367622, X, DQ367642; F. hookeriana Corner, Hooker & T.Thomson 120 (L), KJ845988, KJ845927, X, X, X; F. ingens (Miq.) Miq. 1, Jongkind 4317 (WAG), KJ845966, KJ845906, X, KJ846057, X; F. ingens 2, Rønsted 106 (K), AY730061, X, X, EU084303, X; F. insipida Willd., G. Aguilar 902 (NY), KM186249, X, KM186212, X, X; F. ischnopoda Miq., Rønsted 175 (AAU/K), AY730122, AY730212, EF092380, X, EU084383; F. itoana Diels, Weiblen 622 (A), AF165391, EU084446, EU087655, X, EU084376; F. jimiensis C.C.Berg, unknown, AY730129, AY730216, EF092388, X, X; F. johannis Boiss., unknown, AY730123, AY730213, EF092381, X, X; F. kiloneura Hornby, unknown, AY730098, AY730187, EF092350, X, X; F. krugiana Warb., E. Stijfhoorn 779 (NY), KM186247, X, KM186214, X, X; F. krugiana, S. Barrier 2434 (NY), KM186248, X, KM186213, X, X; F. lapathifolia (Liebm.) Miq., Oyama (UNAM), EU091564, EU084405, EU087609, X, X; F. lateriflora Vahl, unknown, AY063585, AY063546, EF092398, X, X; F. lecardii Warb., Harris 2136 (WAG), KJ845971, KJ845910, KJ846018, KJ846061, X; F. lepicarpa Blume, unknown, AY730138, X, EF092376, X, X; F. lilliputiana D.J.Dixon, unknown, EF545657, EF538773, X, X, X; F. lingua DeWild, Durrand, Rønsted 208 (K), AY730099, AY730188, EF092351, X, X; F. longifolia Schott, Rønsted 140 (K), EU091604, X, X, X, X; F. luschnathiana (Miq.) Miq., unknown, AY730082, AY730170, EF092345, X, X; F. lutea Vahl, Rønsted 87 (C), AY063564, AY063525, EF092347, X, X; F. lyrata Warb., unknown, AY730104, AY730193, X, X, X; F. maclellandi King, XL 56/Rønsted 281 (HITBC), JN117639, JN117704, JN117664, X, EU084365; F. macrophylla Desf. ex Pers., unknown, EF545658, AY063532, EF538792, X, X; F. macrosyce Pittier, E. Contreras 2182 (NY), KM186244, X, KM186216, X, X; F. macrosyce, N.V.L. Brokaw 198 (NY), KM186245, X, KM186215, X, X; F. mauritiana Lam., unknown, AY063570, AY063531, EF092371, X, X; F. mauritiana Lam., unknown, AY063570, X, X, X, X; F. maxima Mill., J.S. Barreto-Silva 2152 (SP), KM186242, X, KM186218, X, X; F. megaleia Corner, Harrison 643 (SAR/MIN), EU091625, X, EU087661, X, X; F. melinocarpa Blume, Weiblen 1705 (MIN), EU091669, X, EU087685, X, X; F. menabeensis H.Perrier, unknown, AY730067, AY730155, X, X, X; F. mexicana (Miq.) Miq., V.W. Steinmann 858 (NY), KM186241, KM186230, KM186219, X, X; F. microcarpa L.f, XL 35, JN117640, JN117665, JN117705, X, X; F. microdictya Diels, Weiblen 954 (MIN), AF165394, EU084447, EU087656, X, EU084377; F. middletonii Chantaras., Chantarasuwan 051010-2 (L), KJ845952, KJ845893, KJ846008, KJ846052, X; F. mollior F.Muell. ex Benth., DQ367643 without, AY063573, AY063534, X, X, DQ367643; F. morobensis C.C.Berg, Weiblen 2228 (MIN), DQ367659, EU084455, DQ367624, X, X; F. mucuso Ficalho, Rønsted 129 (K), AY730120, AY730210, EF092372, EU084317, X; F. mutisii Dugand, L.A.F. Alzate 250 (F)/R. Callejas 10731 (VEN), KM186240, X, KM186220, X, X; F. natalensis Hochst., Forest 333 (NBG), AY730100, AY730189, EF092352, X, X; F. nervosa Roth, Shine (NR255), EU091570, EU084410, EU087615, X, X; F. nodosa Teijsm., Binn., unknown, AF165395, DQ367625, X, X, DQ367645; F. nota (Blanco) Merr., Weiblen 2284 (MIN), EU091626, X, EU087663, X, X; F. nymphaeifolia Mill., unknown, AY063566, AY063527, EU089843, X, X; F. obliqua G.Forst., Cook 2003–6, EF545660, EF538775, EF538793, X, X; F. obscura Blume, Harrison 206 (PUH), EU091676, X, EU087689, X, X; F. obtusifolia Kunth., unknown, AY730084, AY730172, AY967949, X, X; F. obtusiuscula (Miq.) Miq., G. Pereira-Silva 10590 (BG), KM186239, X, KM186221, X, X; F. ochrochlora Ridl., Weiblen 735 (A), AF165396, EU084448, X, X, EU084378; F. odoardi King, Weiblen 708 (A), AF165397, EF092389, X, X, X; F. oleifolia King, Weiblen 2287 (MIN), AY730124, EF092322, EF092382, EU084332, EU084384; F. oligodon Miq., XL55, JN117632, X, JN117698, X, X; F. opposita Miq., Cook 2003/Weiblen 1102 (MIN), EU091670, X, EU087686, X, X; F. orthoneura H.Lév. & Vaniot, Chantarasuwan 231111-1 (L), KJ845987, KJ845926, KJ846034, X, X; F. ottoniifolia (Miq.) Miq., Rønsted 117 (K), AY730109, AY730198, EF092358, X, X; F. ovata Vahl, unknown, DQ455640, DQ455672, X, X, X; F. pachyrrhachis K.Schum. & Lauterb, Weiblen 2377 (MIN), EU091628, EU084456, DG367626, X, DQ367646; F. padana Burm.f., unknown, AF165398, X, EF092387, X, X; F. palmata Forssk., unknown, AY730125, AY730214, EF092383, X, X; F. palmeri S. Watson, unknown, AY730085, AY730173, X, X, X; F. pantoniana King, Weiblen 2079 (MIN), EU091649, X, X, X, X; F. paracamptophylla Corner, Jousselin, EU091592, EU084426, X, X, X; F. paraensis Miq., unknown, AY730086, AY730174, AY967954, X, X; F. parietalis Bl., unknown, AY063583, AY063544, EF092401, X, X; F. pellucidopunctata Griff., Shine, AF165399, EU084427, X, X, X; F. pertusa L., unknown, AF165400, AY730176, AY967950, X, X; F. petiolaris Kunth, unknown, AY730088, AY730177, X, X, X; F. phaeosyce K.Schum. & Lauterb., unknown, AF165401, X, X, X, X; F. pisocarpa Blume, XL27, JN117643, JN117667, JN117707, X, X; F. platypoda A.Cunn. ex Miq., unknown, AY730116, AY730206, EF538794, X, X; F. pleurocarpa F.Muell., Cook 9812/CLV441, EF545661, EF538776, EF538795, DQ455634, X; F. polita Vahl, unknown, DQ455642, DQ455673, X, X, X; F. politoria Lam., Maurin 74 (K), EU091671, X, EU087687, X, X; F. polyantha Warb., Weiblen 2174 (MIN), EU091571, X, EU087616, X, X; F. popenoei Standl., unknown, X, EU089842, X, X, X; F. prasinicarpa Elmer ex C.C.Berg, Nagari 7309 (L), KJ845948, KJ845889, X, X, X; F. preussii Warb., Rønsted 138 (K), AY730105, AY730194, EF092355, DQ455625, X; F. prolixa G.Forst., Gillett 2206 (L), KJ845949, KJ845890, KJ846051, KJ846005, X; F. prostrata (Wall. ex Miq.) Miq., XL21, JN117644, X, JN117708, X, X; F. pseudoconcinna Chantaras., Soenarko 355 (L), KJ845946, KJ845888, KJ846004, KJ846050, X; F. pseudojaca Corner, unknown, EF092317, EF092320, EF092370, X, X; F. punctata Thunb., FK-1997-23 (Montpellier), AY063584, AY063545, X, X, X; F. pungens Reinw. ex Blume, unknown, AF165404, X, DQ367627, X, DQ367647; F. pygmaea Welw. ex Hiern, Forest 332 (NBG), AY730134, AY730221, EF092399, EU084350, X; F. racemigera Burm.f., unknown, AY063587, AY063554, X, X, X; F. racemosa L., Rønsted 116 (K)/Weiblen 940 (A) /XL4, AF165405, X, JN126051, X, EU084371; F. reflexa Thunb., Maurin 76 (K), DQ455650, X, EU087646, X, X; F. religiosa L., BG 04 (L), KJ845980, KJ845919, KJ846027, KJ846066, X; F. ribes Reinw. ex Blume, Weiblen 2108 (MIN), EU091630, EU084458, EU087665, X, X; F. robusta Corner, Weiblen 1541 (MIN), AF165406, EU084442, X, X, DQ367648; F. rubiginosa Desv. ex Vent., Rønsted 89 (C)/Chase 19870 (K, EF545664, EF538777, EF092363, DQ455635, EU084366; F. ruficaulis Merr., unknown, EU091647, KP407001, X, X, X; F. ruginerva Corner, Weiblen 854 (A/MIN), AF165407, EF092323, EF092393, X, X; F. rumphii Blume, Chantarasuwan 120910–4 (L), KJ845993, KJ845932, KJ846039, KJ846073, X; F. sagittata J.König ex Vahl, Harrison 595 (PUH)/Rønsted 266 (HITBC), EU091652, EU084477, EU087678, EU084339, X; F. sagittifolia Mildbr. & Burret, Chase 19852 (K), AY730106, AY730195, EF092356, DQ455626, X; F. santanderana Dugand, H. Garcia-Barriga 17584 (NY), KM186237, KM186231, KM186222, X, X; F. sarmentosa Buch-Ham. ex Sm., Rønsted 263 (HITBC), EU091653, EU084478, EU087679, X, X; F. saussureana DC., unknown, AY730090, AY730179, X, X, X; F. scassellatii Pamp., unknown, AY730107, AY730196, EF092357, X, X; F. schumacheri (Liebm.) Griseb., unknown, AY063567, AY063528, EF092346, X, X; F. schwarzii Koord., Rønsted 160 (K/AAU), EU091633, X, X, X, X; F. scortechinii King, unknown, AY730139, AY730228, EF092377, X, X; F. segoviae Miq., J.L. Clark 4412 (NY)/J.B. Walker 1514 (NY), KM186236, KM186232, KM186223, X, X; F. semicordata Buch.-Ham. ex Sm., XL2/Rønsted 267 (HITBC), JN117646, EU084441, JN117710, EU084322, X; F. semivestita Corner, Weiblen 2380 (MIN), EU091616, EU084443, DQ367629, X, DQ367649; F. septica Burm.f., unknown, AF165409, AY730229, DQ367630, X, DQ367650; F. simplicissima Lour., Li, Lu J., Zhang Z., Zhang L.F 2014062, KX055711, KX055624, X, X, X; F. sinuata Thunb., unknown, AY730135, AY730222, EF092402, X, X; F. spathulifolia Corner, Weiblen 929 (MIN), EU091594, EU084428, EU087631, X, X; F. squamosa Roxb., Rønsted 262 (HITBC), EU091634, X, X, X, X; F. stenophylla Hemsl., Rønsted 277 (HITBC), EU091640, EU084467, X, X, X; F. sterrocarpa Diels, unknown, EF545666, EF538780, EF538796, X, X; F. stolonifera King, Harrison 644 (SAR/MIN), EU091635, X, X, X, X; F. stricta (Miq.) Miq., XL41/Rønsted 288 (HITBC), JN117647, EU084429, JN117711, X, X; F. subcuneata Miq., Weiblen 700 (A)/Weiblen 2166 (MIN), EU091620, EU084449, DQ367631, X, DQ367651; F. subgelderi Corner, unknown, AY063556, AY063517, EF092336, X, X; F. subincisa J.E.Sm. 1, 2009440, JQ774000, X, X, X, X; F. subincisa 2, GBOWS1068, JQ774001, X, X, X, X; F. subtrinervia Laut. & K.Schum., Weiblen 1543 (MIN), AF165389, EU084411, EU087617, X, X; F. subulata Blume, Takeuchi 14266 (K)/Rønsted 299 (HITBC), EU091677, EU084495, EU087690, X, X; F. sumatrana Miq., Harrison 629 (PUH), EU091597, X, EU087634, X, X; F. sundaica Blume, unknown, AY730068, AY730156, EF092337, X, X; F. superba (Miq.) Miq. 1, Rønsted 63 (C), AF165410, AY730149, EF092332, X, X; F. superba 2, C. Friedberg 138 (L), KJ845944, KJ845886, KJ846002, KJ846048, X; F. sur Forssk., Dewsnap/Rønsted 76 (C), AY063572, AY063533, EU087649, EU084319, EU084372; F. sycomorus L., Rønsted 72 (C), AY063575, AY063536, X, EU084320, X; F. tesselata Warb., Rønsted 143 (K), DQ455662, DQ455682, EU087647, X, X; F. tettensis Hutch., Forest 337 (NBG), DQ455665, DQ455683, X, DQ455620, X; F. thonningii Bl., Forest 341 (NBG), AY730102, AY730191, EF092353, X, EU084369; F. tiliifolia Baker, Maurin 87 (K), EU091609, EU084439, X, X, X; F. tinctoria Forst.f., XL19, JN117649, JN117680, JN117713, X, X; F. tonduzii Standl., FB/S3752 (BR), AY730140, AY730230, EU087611, EU084297, X; F. torresiana Standl., J.L. Clark 3445 (NY), KM186235, X, KM186224, X, X; F. trachypison K.Schum., Weiblen 2378 (MIN), EU091674, EU084493, EU087688, X, X; F. tremula Warb., unknown, AY730111, AY730200, X, X, X; F. treubii King, Weiblen 2283 (MIN), EU091636, EU084463, EU087668, X, X; F. trichopoda Baker, Rønsted 118 (K), DQ455666, DQ455684, EU087648, X, X; F. trigonata L., Oyama (UNAM, Mexico), EU091607, X, AY967956, X, X; F. triradiata Corner, unknown, AY730117, AY730207, EF092364, X, X; F. tsiangii Merr. ex Corner, XL46, JN117650, JN117676, JN117714, X, X; F. tsjakela Burm.f., Kostermans 27682 (L), KJ845951, KJ845892, KJ846007, X, X; F. tuerckheimii Standl., Oyama (UNAM, Mexico), EU091608, EU084438, EU087640, X, X; F. tuphapensis Drake, Li, Chen, Wang, Xiong 2010125, JQ774019, KX096889, X, X, X; F. umbellata Vahl, FB/S2813 (BR), DQ455644, DQ455674, X, DQ455629, X; F. uncinulata Corner, unknown, AY063576, AY063537, EU087669, X, X; F. usambarensis Warb., unknown, DQ455653, DQ455677, X, X, X; F. vallis-choudae Delile, Rønsted 126 (K), AY063574, AY063535, EF092373, X, EU084373; F. variegata King, Jousselin, AY063578, AY063539, DQ367633, EU084323, DQ367653; F. variolosa Lindl. ex Benth., XL50, JN117651, JN117681, X, X, X; F. vasculosa Wall. ex. Miq., XL31, JN117652, JN117682, JN117715, X, X; F. vermifuga (Miq.) Miq., L.C. Pederneiras 751 (SP), KM186234, KM186233, KM186225, X, X; F. verruculosa Warb., Adjakidje 2779 (WAG), KJ845979, KJ845917, KJ846026, KJ846065, X; F. villosa Bl., Chase 19851 (K), AY730130, AY730217, EF092391, EU084340, EU084389; F. virens Aiton, unknown, JN117653, X, X, X, DQ367654; F. virgata Reinw. ex Blume, unknown, AF165417, AY730224, EF092404, EU084351, EU084393; F. vogeliana (Miq.) Miq., Rønsted 202 (K)/Rønsted 201 (K), EU091610, EU084440, EU087650, X, X; F. wassa Roxb., unknown, AF165418, EF092325, DQ367635, X, DQ367655; F. watkinsiana F.M.Bailey, Rønsted 83 (C)/Cook (NR256), EF545669, EF538784, EF538800, EU084310, EU084367; F. xylosycia Diels, unknown, AY063557, EF538785, EF538801, X, X; Fatoua villosa (Thunb.) Nakai, GLM-103136, KF137859, X, X, X, X; Helicostylis tomentosa (Poepp., Endl.) Rusby, unknown, FJ037846, X, X, X, X; Naucleopsis guianensis (Mildbr.) C.C.Berg, unknown, FJ037848, X, X, X, X; Poulsenia armata (Miq.) Standl., unknown, AY730144, AY730233, X, X, EU084353; Sparattosyce dioca Bur., Weiblen 1223 (MIN), AY730141, AY730231, EU087607, X, X. © 2018 The Linnean Society of London, Botanical Journal of the Linnean Society This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/open_access/funder_policies/chorus/standard_publication_model)

Hofmann, Christa-Ch

doi: 10.1093/botlinnean/boy035pmid: N/A

Abstract As scanning electron microscopy (SEM) has not been routinely used in palaeo-palynology, affiliations of fossil pollen to botanical equivalents, based only on light microscopy (LM), are always equivocal. Here 29 taxa of Ericales 55.8–41.2 My in age have been identified using LM and SEM, modifying the evolutionary and palaeo-biogeographical framework of some Ericales. Seven Ericaceae types, 17 Sapotaceae types, three Ebenaceae types, one Styracaceae type and one Theaceae type were identified to tribe or genus level. A 56-My-old Erica-type affiliated with Erica arborea might be the ancestor of African Erica taxa, whereas a younger type with affinities to palaeo-Arctic Erica might be the precursor of northern European heathers. The Kalmia-type suggests a European origin and a wider geographical distribution for this type during the mid-Eocene, and a Rhododendron-type of contemporaneous age corroborates the existence of numerous Rhododendron taxa in Europe. Both the 56-My-old Vaccinium-type, affiliated with temperate continental Asian taxa, and the Gaultheria-type, affiliated with several South American taxa, represent the oldest occurrences of these taxa and suggest a Eurasian origin. The Sapotaceae are represented by Sarcospermatoideae, with two Sarcosperma-types, Chrysophylloideae, with three Chrysophylleae-types affiliated with South American Pouteria/Elaeoluma and one affiliated with Asian Planchonella/Sersalisia, and Sapotoideae with three Mimusopeae/Isonandreae-types with mixed affinities, two Mimusopeae-types affiliated with African/Malegasy Mimusops/Vitellariopsis taxa, one Mimusopeae-type affiliated with South American Manilkara taxa, four Isonandreae-types affiliated with SE Asian Palaquium and one South American Sideroxylon-type. All Sapotaceae types are older than postulated in the literature by molecular dating analyses. Three Diospyros-types affiliated with African taxa corroborate a Gondwanan–African origin of the genus. A Styrax-type resembles extant South American and East Asian taxa and is interpreted as being an early European ancestor of Styrax. The contemporaneous Camellia-type represents the oldest record of Camellia. Camellia, Chrysophylloideae, Diospyros, Erica, Gaultheria, Kalmia, Rhododendron, Sarcospermatoideae, Sapotoideae, Styrax, Vaccinium INTRODUCTION The asterid ericalean clade (APG IV, 2016) is morphologically diverse with well-defined families (Schönenberger, Anderberg & Sytsma, 2005), of which several have been recently analysed using molecular phylogenetic techniques: Ericaceae by Kron et al. (1999), Kron, Powell & Luteyn (2002), Gillespie & Kron (2010), Kron & Luteyn (2005) and Kron & Powell (2009), Sapotaceae by Anderberg & Swenson (2003), Swenson, Bartish & Munzinger (2007), Swenson, Richardson & Bartish (2008), Bartish, Antonelli & Richardson (2011), Armstrong et al. (2014) and Richardson et al. (2014), Ebenaceae by Wallnhöfer (2001), Duangjai et al. (2006) and Geeraerts et al. (2009), Styracaceae by Fritsch (2001) and Fritsch et al. (2001, 2011), and Theaceae by Vijayan, Zhang & Tsou (2009). The evolution of Ericales and the asterids has been the subject of molecular phylogenetic investigations that used some fossils for age estimation of individual lineages within an order or clade (e.g. Magallón et al., 2015; Crepet, Nixon & Gandolfo, 2004; Bremer et al., 2002; Schönenberger et al., 2005; Martínez-Millán, 2010), but there are frequently discrepancies between the age of the fossils and molecular clock-estimated ages. Although there are many fossil representatives of ericalean families, several fossil occurrences have not been treated systematically and others are in need of thorough revision (Collinson, Boulter & Holmes, 1993; Martínez-Millán, 2010, p. 86; but see also Manchester et al., 2015). This is particularly the case with fossil pollen grains, which, in former times, were generally investigated only by light microscopy (LM) for stratigraphic purposes (e.g. Kruztsch & Vanhoorne, 1977; Gruas-Cavagnetto, 1976, p. 178; Kedves 1969) and only a few exceptions have demonstrated the need for scanning electron microscopy (SEM) in describing palaeo-palynomorphs (e.g. Doyle, Van Campo & Lugardon 1975; Walker & Walker, 1984). However, more recent studies have routinely used LM and SEM investigations of the same fossil grain (e.g. Zetter & Hesse, 1996; Zetter & Hofmann, 2001; Hofmann, Mohamed & Egger, 2011; Grímsson et al., 2015; Hofmann, Egger & King, 2015a, b; Bouchal et al., 2016, 2017; Schrank, 2017), and this paper follows this tradition. The routinely applied SEM investigation of fossil pollen is possible because of previous comprehensive LM and SEM studies on modern pollen. Particularly important for this study were the studies of Sarwar, Ito & Tahahashi (2006), Sarwar & Takahashi (2006, 2012, 2013, 2014) on Ericaceae, Harley (1991) on Sapotaceae, Geeraerts et al. (2009) on Ebenoideae, Morton & Dickison (1992) on Styracaceae and Zavada & Wei (1993) on Camellia L. The method applied in this paper, in which fossil pollen types are described, depicted and compared with the most similar looking pollen of modern species, represents a phenetic approach; whether the individual fossil pollen type is a member of a stem lineage, stem relatives or crown group cannot be resolved without a phylogenetic analysis. MATERIAL AND METHODS Sample preparation for palaeo-palynology followed standard procedures: samples were crushed by hand with a mortar and pestle; the resultant rock powder was treated with HCl to dissolve carbonate and washed three times and decanted. The remaining sample was treated with HF (cold process over 4 days) to dissolve silicates and decanted. The organic residue was washed three times, decanted and boiled in HCL for 5 min, decanted again and washed several times. The organic residue was not sieved and hence retained palynomorphs of < 10 µm. Further acetolysis reduced the amount of associated plant fragments. The remains were mixed with glycerine and stored in small, tightly closed glass vials. For LM investigation, the pollen of interest was transferred from a sample smeared on a glass slide into a clean drop of glycerine on a new slide, using a micro-manipulator, and then photographed with a Samsung digital camera. After being photographed, the pollen was transferred to an SEM stub with a micro-manipulator and carefully washed with 100% alcohol to remove the glycerine. It was then sputtered with gold and examined with SEM (FEI Inspect 500). Pollen terminology and descriptive terms generally follow Hesse et al. (2009). Stubs and photographs are stored in the Department of Palaeontology, University of Vienna, under inventory numbers IPUW 7836/BoBC_1-12 (1–10), IPUW7837a/Krappf_1-18 (1–10), IPUW7837b/Hain_1_1 (1–10), IPUW 7838/Brix_1-22 (1–10) and IPUW 7839/Cobham_1-12 (1–10). Geological background and stratigraphic position of the sample localities Samples from five localities have been studied, four localities from Europe and one from China. (1) The Brixton material used in this study came from three boreholes (samples B6-10.5, B10-11.9, B10-16.5, B13-12.3 and B13-12.3) drilled into the London Basin in south-west London (England, Fig. 1). The siltstone and mudstone sediments investigated come from the shallow marine Woolwich and terrestrial Reading Formations, both of which overlie the Upnor Formation, dated as NP9 (late Thanetian, Ellison & King, 2004). The carbon isotope excursion (CIE) marking the base of the Eocene and the Paleocene–Eocene Thermal Maximum (PETM) has been identified near the base of the Lower Mottled Clay of the Reading Formation (Thiry et al., 1998). (2) Sample material from the Cobham Lignite (Shorne Member) in Kent (England, Fig. 1) is also associated with the PETM. The Cobham Lignite overlies a 1-m-thick unnamed sandstone and mudstone unit which again overlies the Upnor Formation. The Apectodinium (L.I.Costa & C.Downie) J.K.Lentin & G.L.Williams acme, which in marine and marginal-marine environments is co-extensive with the PETM, occurs at the base of the Lower Shelly Beds of the Woolwich Formation in the Cobham area. The Cobham Lignite is composed of laminated and blocky lignite (Collinson et al., 2009) and the sample material of this study came from the laminated lignite (sample Cobham 48) donated by M. E. Collinson. (3) In the Krappfeld area (Carinthia, Austria), the terrestrial Holzer Formation of Ypresian age unconformably overlies marine upper Campanian marlstones (Upper Cretaceous) of the Pemberger Formation. The Holzer Formation comprises 8-m-thick soft kaolinitic green and red clays incorporating small coal lenses and the sample material (samples PQ5b, PQ8 and PQ9) is characterized by a relatively rich terrestrial palynoflora. The top of the clay can be assigned to the lower part of shallow benthic zone SBZ10, which has been correlated with calcareous nanoplankon zone NP12 and suggests that the claystone was deposited during the Early Eocene Climate Optimum (EECO; Hofmann et al., 2012). (4) The Borken sample material (Hesse, Germany, Fig. 1), which comes from a former underground mine in the ‘Borken brown coal field’ of the Borken Formation at Stolzenbach, south of Kassel (Oschkinis & Gregor, 1992, 2005), is of lower Bartonian age (Ritzkowski, 2005). A coal split of brownish clays and marls, which divides the upper part of the Main Seam into two, contains several vertebrate and plant macrofossils (summarized by Gregor, 2005). The material investigated (one sample), which was derived from a small lens of coal within the Main Seam at the seam split, incorporated several diaspores of Saururus L. (material from H.-J. Gregor). (5) The Changchang Formation of the Changchang Basin in north-western Hainan (southern China, Fig. 1) comprises lacustrine and fluvial mudsones, siltstones and sandstones with regularly intercalated coaly horizons. The coals in the lower part of the section are interpreted to have grown in a paludal to lacustrine setting. The sediments display a high diversity of megafossils (leaves and diaspores, see references in Spicer et al., 2014) and palynomorphs (Yao et al., 2009) and is assumed to be of Lutetian to Bartonian age. The sample material analysed here (samples I/14-1 and I/14-3) came from a coaly mudstone layer between the two thickest sandstone bodies in the middle part of the Changchang Formation and were donated by T. M. Kodrul and J. Jin. Further detailed work on these samples is in progress. Figure 1. View largeDownload slide Sketch map of the five localities: A, all four localities in Europe; B, Brixton (near London) and Cobham in England; C, Borken (near Kassel) in Germany; D, Krappfeld (near Klagenfurt) in Austria; and E, Changchang (Hainan). Figure 1. View largeDownload slide Sketch map of the five localities: A, all four localities in Europe; B, Brixton (near London) and Cobham in England; C, Borken (near Kassel) in Germany; D, Krappfeld (near Klagenfurt) in Austria; and E, Changchang (Hainan). Previous pollen data from the localities The pollen and spore composition of individual samples and the botanical affiliation of many taxa from the EECO section at Krappfeld, based on both LM and SEM, are given in Zetter & Hofmann (2001) and Hofmann & Zetter (2001). Hofmann et al. (2012) added new taxa and also summarized the global distribution and zonobiomes, climatic requirements and the seasonality of the extant relatives of the fossil taxa (c. 50% of the flora). The first results of pollen concerning early diverging angiosperm, monocot and seven eudicot families of the PETM from the Brixton drillcores and the EECO section at Krappfeld are given in Hofmann et al. (2015a, b; LM and SEM investigation). The pollen and spores of the Cobham Lignite section have been investigated with LM by Collinson et al. (2009) and overviews of the palynology of the coal-bearing strata of Stolzenbach from the Borken coal area are given in Hottenrott et al. (2010) and Hofmann & Gregor (2018; LM and SEM investigation). The pollen and spores of the Changchang Formation on Hainan have been investigated with LM by Yao et al. (2009). The pollen taxa described below are generally accessory elements that do not occur frequently, and if they do, then rarely in high numbers. RESULTS Pollen morphology is generally conservative, and thus, considering that the fossil pollen discussed in this paper are between 55.8 and c. 41.2 My old (time scale after Gradstein et al., 2012), the affiliation with a genus or tribe is here taken to indicate that the fossil pollen taxon could represent a precursor or a member of a lineage leading to that genus or tribe and sometimes might represent the genus itself; therefore, the epithet ‘ …...-type’ is used. The ericalean palynomorphs from the lower Bartonian Borken locality have been described by Hofmann & Gregor (2018), but, for comparative purposes and completeness, these taxa are discussed also in this paper. Here, 29 taxa of Ericales are described, imaged, and identified to family, tribe or genus level, with the help of LM and SEM (Table 1, Figs 2–9): seven Ericaceae (two Erica L.-, Kalmia L.-, Vaccinium L.-, Gaultheria Kalm ex L.- and one gen. indet.-type), 17 Sapotaceae (two Sarcosperma Hook.f.-, four Chrysophylleae-, three Mimusopeae/Isonandreae-, three Mimusopeae, four Isonandreae- and one Sideroxylon-types), three Ebenaceae (two Diospyros L.-type, one Diospyros/Euclea L.-type), one Styracaceae (Styrax L.-type) and one Theaceae (Camellia-type). Table 1. Summary of all pollen types described, their stratigraphic age, local occurrences and measurements Family Locality Number Age Nearest botanical affiliation Home of nearest botanical affiliation Climate Ma×. size of pollen (p×e) or tetrad (µm) Max wall thickness (µm) Size of endoapertures (h×w) (µm) Length of colpi (µm) Tectum Ornamentation Micro-ornamentation Er Erica-type Cobham 1 l. Ypr E. arborea Medit., Near East medit. 19.3 × 25.2 1.2 tec per fos rug gem Er Erica-type Borken 1 l. Bart Erica Europe temp. 21 × 19.6 1.3–1.5 tec fos rug stri Er Kalmia-type Borken 5 l. Bart Kalmia N. America wa tem 28.2 × 39.3 1.5–1.8 tec fos rug smo Er Rhododendron-type Borken 1 l. Bart Rhododendron Eurasia wa tem 24.1 × 23 1.5–1.8 tec per fos rug smo Er Vaccinium-type Cobham 1 l. Ypr Vaccinium Asia temp. 30 × 32 1.2–1.5 tec fos mrug smo Er Gaultheria-type Brixton 1 l. Ypr Gaultheria S America temp. 26.3 × 25.2 1.1 tec fos mrug smo Er gen indet-type Cobham 1 l. Ypr ? ? ? 25.1 × 20.2 1.1 tec fos rug inden S Sarcosperma-type Brixton 1 l. Ypr Sarcosperma Asia trop 18.6 × 14.8 1.2 2.8 × 3 11 tec per fos smo S Sarcosperma-type Krappfeld 1 Ypr Sarcosperma Asia trop 23.6 × 17.2 1.1 2.5 × 3 14 tec per fos smo S Chrysophylleae-type Cobham 2 l. Ypr Pouteria/Elaeol S. America trop 26.6 × 30.1 1.2–3 3 × 4.5 10–15 tec per fos mrug smo S Chrysophylleae-type Brixton 1 l. Ypr Planch./Sersalisia New Caled., Austral trop 31.6 × 30 1.7–1.9 4.5 × 8 15 tec per fos mrug stri S Chrysophylleae-type 1 Krappfeld 1 Ypr Pouteria/Elaeol S. America trop 21.5 × 21.1 1.2–1.8 1.2 × 3 8 tec fos per rug smo S Chrysophylleae-type 2 Krappfeld 1 Ypr Pouteria S. America trop 24.3 × 20.9 1.1–2.2 1.2 × 3.4 9 tec fos per smo S Mimusopeae/Isonandreae-type Brixton 1 l. Ypr Mimusop/Isonand Africa – Asia trop 30.2 × 25.1 1.5–1.7 1.7 × 4 11 tec fos mrug stri S Mimusopeae/Isonandreae-type 1 Borken 1 l. Bart Mimusop/Isonand Africa – Asia trop 29 × 21.9 1.4 1.4 × 3.5 22 tec per mver smo S Mimusopeae/Isonandreae-type 2 Borken 1 l. Bart Mimusop/Isonand Africa – Asia trop 29.1 × 22.7 1.6 1.5 × 5 17 tec per mver smo S Mimusopeae-type 1 Krappfeld 1 Ypr Mimusops/Vitellari. Africa, Madagasc trop 21.8 × 14.1 1 2 × 4.5 11.7 rec per fov mver smo S Mimusopeae-type 2 Krappfeld 2 Ypr Mimusops/Vitellari. Africa, Madagasc trop 28 × 22.1 2–2.4 2.1 × 2.4 13–14.8 tec per fos mver smo S Mimusopeae-type Borken 1 l. Bart Manilkara S. America trop 33.8 × 22.5 2–2.4 2 × 4.5 28–29.9 tec per fos mver ver S Isonandreae-type Brixton 1 P-E b Palaquium/Madhuca SE Asia trop 24.3 × 20.8 1.2 3.5 × 5 15.4 tec per mrug smo S Isonandreae-type Krappfeld 3 Ypr Palaquium SE Asia trop 30.5 × 22.1 1.3–1.8 4.2 × 5 25.5–27 tec per mrug smo S Isonandreae-type 1 Borken 1 l. Bart Palaquium SE Asia trop 26.2 × 13 1.2–1.4 2.8 × 2.5 20 tec per mrug smo S Isonandreae-type 2 Borken 1 l. Bart Palaquium SE Asia trop 26.6 × 20.9 2 3 × 6 19.4 tec per mrug smo S Sideroxylon-type Brixton 3 l. Ypr Sideroxylon S America, China trop 25.4-18.1 0.9–1.4 5.5 × 4.5 12.3–14 tec per fos rug smo Eb Diospyros/Euclea-type Brixton 1 l. Ypr Diospyros/Euclea Africa trop 38.5 × 20.9 0.9–1.4 3 × 3.5 24 tec fos per rug ech Eb Diospyros-type 1 Hainan 1 l. Bart Diospyros Africa trop 50 × 30.1 1.8–2.5 1.8 × 3.8 40 tec fos per rug rug Eb Diospyros-type 2 Hainan 4 l. Bart Diospyros Africa trop 36.9 × 26.4 1.2–1.4 1.3 × 2.5 24–31.5 tec fos mrug ann St Styrax-type Krappfeld 1 Ypr Styrax America, Asia wa tem 18.1 × 15.5 1 1 × 1.5 11.8 tec per crot smo T Camellia-type Krappfeld 1 Ypr Camellia China wa tem 18.2–195 1.2 w 3.5 tec fos mrug ann Family Locality Number Age Nearest botanical affiliation Home of nearest botanical affiliation Climate Ma×. size of pollen (p×e) or tetrad (µm) Max wall thickness (µm) Size of endoapertures (h×w) (µm) Length of colpi (µm) Tectum Ornamentation Micro-ornamentation Er Erica-type Cobham 1 l. Ypr E. arborea Medit., Near East medit. 19.3 × 25.2 1.2 tec per fos rug gem Er Erica-type Borken 1 l. Bart Erica Europe temp. 21 × 19.6 1.3–1.5 tec fos rug stri Er Kalmia-type Borken 5 l. Bart Kalmia N. America wa tem 28.2 × 39.3 1.5–1.8 tec fos rug smo Er Rhododendron-type Borken 1 l. Bart Rhododendron Eurasia wa tem 24.1 × 23 1.5–1.8 tec per fos rug smo Er Vaccinium-type Cobham 1 l. Ypr Vaccinium Asia temp. 30 × 32 1.2–1.5 tec fos mrug smo Er Gaultheria-type Brixton 1 l. Ypr Gaultheria S America temp. 26.3 × 25.2 1.1 tec fos mrug smo Er gen indet-type Cobham 1 l. Ypr ? ? ? 25.1 × 20.2 1.1 tec fos rug inden S Sarcosperma-type Brixton 1 l. Ypr Sarcosperma Asia trop 18.6 × 14.8 1.2 2.8 × 3 11 tec per fos smo S Sarcosperma-type Krappfeld 1 Ypr Sarcosperma Asia trop 23.6 × 17.2 1.1 2.5 × 3 14 tec per fos smo S Chrysophylleae-type Cobham 2 l. Ypr Pouteria/Elaeol S. America trop 26.6 × 30.1 1.2–3 3 × 4.5 10–15 tec per fos mrug smo S Chrysophylleae-type Brixton 1 l. Ypr Planch./Sersalisia New Caled., Austral trop 31.6 × 30 1.7–1.9 4.5 × 8 15 tec per fos mrug stri S Chrysophylleae-type 1 Krappfeld 1 Ypr Pouteria/Elaeol S. America trop 21.5 × 21.1 1.2–1.8 1.2 × 3 8 tec fos per rug smo S Chrysophylleae-type 2 Krappfeld 1 Ypr Pouteria S. America trop 24.3 × 20.9 1.1–2.2 1.2 × 3.4 9 tec fos per smo S Mimusopeae/Isonandreae-type Brixton 1 l. Ypr Mimusop/Isonand Africa – Asia trop 30.2 × 25.1 1.5–1.7 1.7 × 4 11 tec fos mrug stri S Mimusopeae/Isonandreae-type 1 Borken 1 l. Bart Mimusop/Isonand Africa – Asia trop 29 × 21.9 1.4 1.4 × 3.5 22 tec per mver smo S Mimusopeae/Isonandreae-type 2 Borken 1 l. Bart Mimusop/Isonand Africa – Asia trop 29.1 × 22.7 1.6 1.5 × 5 17 tec per mver smo S Mimusopeae-type 1 Krappfeld 1 Ypr Mimusops/Vitellari. Africa, Madagasc trop 21.8 × 14.1 1 2 × 4.5 11.7 rec per fov mver smo S Mimusopeae-type 2 Krappfeld 2 Ypr Mimusops/Vitellari. Africa, Madagasc trop 28 × 22.1 2–2.4 2.1 × 2.4 13–14.8 tec per fos mver smo S Mimusopeae-type Borken 1 l. Bart Manilkara S. America trop 33.8 × 22.5 2–2.4 2 × 4.5 28–29.9 tec per fos mver ver S Isonandreae-type Brixton 1 P-E b Palaquium/Madhuca SE Asia trop 24.3 × 20.8 1.2 3.5 × 5 15.4 tec per mrug smo S Isonandreae-type Krappfeld 3 Ypr Palaquium SE Asia trop 30.5 × 22.1 1.3–1.8 4.2 × 5 25.5–27 tec per mrug smo S Isonandreae-type 1 Borken 1 l. Bart Palaquium SE Asia trop 26.2 × 13 1.2–1.4 2.8 × 2.5 20 tec per mrug smo S Isonandreae-type 2 Borken 1 l. Bart Palaquium SE Asia trop 26.6 × 20.9 2 3 × 6 19.4 tec per mrug smo S Sideroxylon-type Brixton 3 l. Ypr Sideroxylon S America, China trop 25.4-18.1 0.9–1.4 5.5 × 4.5 12.3–14 tec per fos rug smo Eb Diospyros/Euclea-type Brixton 1 l. Ypr Diospyros/Euclea Africa trop 38.5 × 20.9 0.9–1.4 3 × 3.5 24 tec fos per rug ech Eb Diospyros-type 1 Hainan 1 l. Bart Diospyros Africa trop 50 × 30.1 1.8–2.5 1.8 × 3.8 40 tec fos per rug rug Eb Diospyros-type 2 Hainan 4 l. Bart Diospyros Africa trop 36.9 × 26.4 1.2–1.4 1.3 × 2.5 24–31.5 tec fos mrug ann St Styrax-type Krappfeld 1 Ypr Styrax America, Asia wa tem 18.1 × 15.5 1 1 × 1.5 11.8 tec per crot smo T Camellia-type Krappfeld 1 Ypr Camellia China wa tem 18.2–195 1.2 w 3.5 tec fos mrug ann Er, Ericaceae; S, Sapotaceae. Eb, Ebenaceae; St, Styracaceae; T, Theaceae; l. Ypr, lower Ypresian; l. Bart, lower Bartonian; P-E b, Palaeocene/Eocene boundary; medit., mediterranean; temp., temperate; wa tem, warm temperate; trop, tropical; tec, tectate; fos, fossulate; per, perforate; fov, foveolate; rug, rugulate; mrug, micro-rugulate; mver, micro-verrucate; ver, verrucate; crot, crotonoid pattern; gem, gemmate; stri, striate; smo, smooth; ech, echinate; ann, annelid pattern. View Large Table 1. Summary of all pollen types described, their stratigraphic age, local occurrences and measurements Family Locality Number Age Nearest botanical affiliation Home of nearest botanical affiliation Climate Ma×. size of pollen (p×e) or tetrad (µm) Max wall thickness (µm) Size of endoapertures (h×w) (µm) Length of colpi (µm) Tectum Ornamentation Micro-ornamentation Er Erica-type Cobham 1 l. Ypr E. arborea Medit., Near East medit. 19.3 × 25.2 1.2 tec per fos rug gem Er Erica-type Borken 1 l. Bart Erica Europe temp. 21 × 19.6 1.3–1.5 tec fos rug stri Er Kalmia-type Borken 5 l. Bart Kalmia N. America wa tem 28.2 × 39.3 1.5–1.8 tec fos rug smo Er Rhododendron-type Borken 1 l. Bart Rhododendron Eurasia wa tem 24.1 × 23 1.5–1.8 tec per fos rug smo Er Vaccinium-type Cobham 1 l. Ypr Vaccinium Asia temp. 30 × 32 1.2–1.5 tec fos mrug smo Er Gaultheria-type Brixton 1 l. Ypr Gaultheria S America temp. 26.3 × 25.2 1.1 tec fos mrug smo Er gen indet-type Cobham 1 l. Ypr ? ? ? 25.1 × 20.2 1.1 tec fos rug inden S Sarcosperma-type Brixton 1 l. Ypr Sarcosperma Asia trop 18.6 × 14.8 1.2 2.8 × 3 11 tec per fos smo S Sarcosperma-type Krappfeld 1 Ypr Sarcosperma Asia trop 23.6 × 17.2 1.1 2.5 × 3 14 tec per fos smo S Chrysophylleae-type Cobham 2 l. Ypr Pouteria/Elaeol S. America trop 26.6 × 30.1 1.2–3 3 × 4.5 10–15 tec per fos mrug smo S Chrysophylleae-type Brixton 1 l. Ypr Planch./Sersalisia New Caled., Austral trop 31.6 × 30 1.7–1.9 4.5 × 8 15 tec per fos mrug stri S Chrysophylleae-type 1 Krappfeld 1 Ypr Pouteria/Elaeol S. America trop 21.5 × 21.1 1.2–1.8 1.2 × 3 8 tec fos per rug smo S Chrysophylleae-type 2 Krappfeld 1 Ypr Pouteria S. America trop 24.3 × 20.9 1.1–2.2 1.2 × 3.4 9 tec fos per smo S Mimusopeae/Isonandreae-type Brixton 1 l. Ypr Mimusop/Isonand Africa – Asia trop 30.2 × 25.1 1.5–1.7 1.7 × 4 11 tec fos mrug stri S Mimusopeae/Isonandreae-type 1 Borken 1 l. Bart Mimusop/Isonand Africa – Asia trop 29 × 21.9 1.4 1.4 × 3.5 22 tec per mver smo S Mimusopeae/Isonandreae-type 2 Borken 1 l. Bart Mimusop/Isonand Africa – Asia trop 29.1 × 22.7 1.6 1.5 × 5 17 tec per mver smo S Mimusopeae-type 1 Krappfeld 1 Ypr Mimusops/Vitellari. Africa, Madagasc trop 21.8 × 14.1 1 2 × 4.5 11.7 rec per fov mver smo S Mimusopeae-type 2 Krappfeld 2 Ypr Mimusops/Vitellari. Africa, Madagasc trop 28 × 22.1 2–2.4 2.1 × 2.4 13–14.8 tec per fos mver smo S Mimusopeae-type Borken 1 l. Bart Manilkara S. America trop 33.8 × 22.5 2–2.4 2 × 4.5 28–29.9 tec per fos mver ver S Isonandreae-type Brixton 1 P-E b Palaquium/Madhuca SE Asia trop 24.3 × 20.8 1.2 3.5 × 5 15.4 tec per mrug smo S Isonandreae-type Krappfeld 3 Ypr Palaquium SE Asia trop 30.5 × 22.1 1.3–1.8 4.2 × 5 25.5–27 tec per mrug smo S Isonandreae-type 1 Borken 1 l. Bart Palaquium SE Asia trop 26.2 × 13 1.2–1.4 2.8 × 2.5 20 tec per mrug smo S Isonandreae-type 2 Borken 1 l. Bart Palaquium SE Asia trop 26.6 × 20.9 2 3 × 6 19.4 tec per mrug smo S Sideroxylon-type Brixton 3 l. Ypr Sideroxylon S America, China trop 25.4-18.1 0.9–1.4 5.5 × 4.5 12.3–14 tec per fos rug smo Eb Diospyros/Euclea-type Brixton 1 l. Ypr Diospyros/Euclea Africa trop 38.5 × 20.9 0.9–1.4 3 × 3.5 24 tec fos per rug ech Eb Diospyros-type 1 Hainan 1 l. Bart Diospyros Africa trop 50 × 30.1 1.8–2.5 1.8 × 3.8 40 tec fos per rug rug Eb Diospyros-type 2 Hainan 4 l. Bart Diospyros Africa trop 36.9 × 26.4 1.2–1.4 1.3 × 2.5 24–31.5 tec fos mrug ann St Styrax-type Krappfeld 1 Ypr Styrax America, Asia wa tem 18.1 × 15.5 1 1 × 1.5 11.8 tec per crot smo T Camellia-type Krappfeld 1 Ypr Camellia China wa tem 18.2–195 1.2 w 3.5 tec fos mrug ann Family Locality Number Age Nearest botanical affiliation Home of nearest botanical affiliation Climate Ma×. size of pollen (p×e) or tetrad (µm) Max wall thickness (µm) Size of endoapertures (h×w) (µm) Length of colpi (µm) Tectum Ornamentation Micro-ornamentation Er Erica-type Cobham 1 l. Ypr E. arborea Medit., Near East medit. 19.3 × 25.2 1.2 tec per fos rug gem Er Erica-type Borken 1 l. Bart Erica Europe temp. 21 × 19.6 1.3–1.5 tec fos rug stri Er Kalmia-type Borken 5 l. Bart Kalmia N. America wa tem 28.2 × 39.3 1.5–1.8 tec fos rug smo Er Rhododendron-type Borken 1 l. Bart Rhododendron Eurasia wa tem 24.1 × 23 1.5–1.8 tec per fos rug smo Er Vaccinium-type Cobham 1 l. Ypr Vaccinium Asia temp. 30 × 32 1.2–1.5 tec fos mrug smo Er Gaultheria-type Brixton 1 l. Ypr Gaultheria S America temp. 26.3 × 25.2 1.1 tec fos mrug smo Er gen indet-type Cobham 1 l. Ypr ? ? ? 25.1 × 20.2 1.1 tec fos rug inden S Sarcosperma-type Brixton 1 l. Ypr Sarcosperma Asia trop 18.6 × 14.8 1.2 2.8 × 3 11 tec per fos smo S Sarcosperma-type Krappfeld 1 Ypr Sarcosperma Asia trop 23.6 × 17.2 1.1 2.5 × 3 14 tec per fos smo S Chrysophylleae-type Cobham 2 l. Ypr Pouteria/Elaeol S. America trop 26.6 × 30.1 1.2–3 3 × 4.5 10–15 tec per fos mrug smo S Chrysophylleae-type Brixton 1 l. Ypr Planch./Sersalisia New Caled., Austral trop 31.6 × 30 1.7–1.9 4.5 × 8 15 tec per fos mrug stri S Chrysophylleae-type 1 Krappfeld 1 Ypr Pouteria/Elaeol S. America trop 21.5 × 21.1 1.2–1.8 1.2 × 3 8 tec fos per rug smo S Chrysophylleae-type 2 Krappfeld 1 Ypr Pouteria S. America trop 24.3 × 20.9 1.1–2.2 1.2 × 3.4 9 tec fos per smo S Mimusopeae/Isonandreae-type Brixton 1 l. Ypr Mimusop/Isonand Africa – Asia trop 30.2 × 25.1 1.5–1.7 1.7 × 4 11 tec fos mrug stri S Mimusopeae/Isonandreae-type 1 Borken 1 l. Bart Mimusop/Isonand Africa – Asia trop 29 × 21.9 1.4 1.4 × 3.5 22 tec per mver smo S Mimusopeae/Isonandreae-type 2 Borken 1 l. Bart Mimusop/Isonand Africa – Asia trop 29.1 × 22.7 1.6 1.5 × 5 17 tec per mver smo S Mimusopeae-type 1 Krappfeld 1 Ypr Mimusops/Vitellari. Africa, Madagasc trop 21.8 × 14.1 1 2 × 4.5 11.7 rec per fov mver smo S Mimusopeae-type 2 Krappfeld 2 Ypr Mimusops/Vitellari. Africa, Madagasc trop 28 × 22.1 2–2.4 2.1 × 2.4 13–14.8 tec per fos mver smo S Mimusopeae-type Borken 1 l. Bart Manilkara S. America trop 33.8 × 22.5 2–2.4 2 × 4.5 28–29.9 tec per fos mver ver S Isonandreae-type Brixton 1 P-E b Palaquium/Madhuca SE Asia trop 24.3 × 20.8 1.2 3.5 × 5 15.4 tec per mrug smo S Isonandreae-type Krappfeld 3 Ypr Palaquium SE Asia trop 30.5 × 22.1 1.3–1.8 4.2 × 5 25.5–27 tec per mrug smo S Isonandreae-type 1 Borken 1 l. Bart Palaquium SE Asia trop 26.2 × 13 1.2–1.4 2.8 × 2.5 20 tec per mrug smo S Isonandreae-type 2 Borken 1 l. Bart Palaquium SE Asia trop 26.6 × 20.9 2 3 × 6 19.4 tec per mrug smo S Sideroxylon-type Brixton 3 l. Ypr Sideroxylon S America, China trop 25.4-18.1 0.9–1.4 5.5 × 4.5 12.3–14 tec per fos rug smo Eb Diospyros/Euclea-type Brixton 1 l. Ypr Diospyros/Euclea Africa trop 38.5 × 20.9 0.9–1.4 3 × 3.5 24 tec fos per rug ech Eb Diospyros-type 1 Hainan 1 l. Bart Diospyros Africa trop 50 × 30.1 1.8–2.5 1.8 × 3.8 40 tec fos per rug rug Eb Diospyros-type 2 Hainan 4 l. Bart Diospyros Africa trop 36.9 × 26.4 1.2–1.4 1.3 × 2.5 24–31.5 tec fos mrug ann St Styrax-type Krappfeld 1 Ypr Styrax America, Asia wa tem 18.1 × 15.5 1 1 × 1.5 11.8 tec per crot smo T Camellia-type Krappfeld 1 Ypr Camellia China wa tem 18.2–195 1.2 w 3.5 tec fos mrug ann Er, Ericaceae; S, Sapotaceae. Eb, Ebenaceae; St, Styracaceae; T, Theaceae; l. Ypr, lower Ypresian; l. Bart, lower Bartonian; P-E b, Palaeocene/Eocene boundary; medit., mediterranean; temp., temperate; wa tem, warm temperate; trop, tropical; tec, tectate; fos, fossulate; per, perforate; fov, foveolate; rug, rugulate; mrug, micro-rugulate; mver, micro-verrucate; ver, verrucate; crot, crotonoid pattern; gem, gemmate; stri, striate; smo, smooth; ech, echinate; ann, annelid pattern. View Large Figure 2. View largeDownload slide Ericaceae. A–C, Erica-type from the Cobham Lignite: A, LM overview; B, SEM overview of the same grain as in A; C, SEM detail of the same grain as in A. D–F, Erica-type from Borken coalfield: D, LM overview; E, SEM overview of the same grain as in D; F, SEM detail of the same grain as in D. G–L, Kalmia-type from Borken coalfield: G, and J, LM overviews; H, SEM overview of the same grain as in G; K, SEM overview of the same grain as in J; I, SEM detail of the same grain as in G; L, SEM detail of the same grain as in J. LM photographs ×1000, bar in SEM overview 10 µm, bar in SEM detail 2 µm. Figure 2. View largeDownload slide Ericaceae. A–C, Erica-type from the Cobham Lignite: A, LM overview; B, SEM overview of the same grain as in A; C, SEM detail of the same grain as in A. D–F, Erica-type from Borken coalfield: D, LM overview; E, SEM overview of the same grain as in D; F, SEM detail of the same grain as in D. G–L, Kalmia-type from Borken coalfield: G, and J, LM overviews; H, SEM overview of the same grain as in G; K, SEM overview of the same grain as in J; I, SEM detail of the same grain as in G; L, SEM detail of the same grain as in J. LM photographs ×1000, bar in SEM overview 10 µm, bar in SEM detail 2 µm. Figure 3. View largeDownload slide Ericaceae. A–C, Rhododendron-type from Borken coal field: A, LM overview; B, SEM overview of the same grain as in A; C, SEM detail of the same grain as in A. D–F, Vaccinium-type from the Cobham Lignite: D, LM overview; E, SEM overview of the same grain as in D; F, SEM detail of the same grain as in D. G–I, Gaultheria-type from Brixton drill core 6: G, LM overview; H, SEM overview of the same grain as in G; I, SEM detail of the same grain as in G. J–L, Ericaceae gen. indet.-type from the Cobham Lignite: J, LM overview; K, SEM overview of the same grain as in J; L, SEM detail of the same grain as in J. LM photographs ×1000, bar in SEM overview 10 µm, bar in SEM detail 2 µm. Figure 3. View largeDownload slide Ericaceae. A–C, Rhododendron-type from Borken coal field: A, LM overview; B, SEM overview of the same grain as in A; C, SEM detail of the same grain as in A. D–F, Vaccinium-type from the Cobham Lignite: D, LM overview; E, SEM overview of the same grain as in D; F, SEM detail of the same grain as in D. G–I, Gaultheria-type from Brixton drill core 6: G, LM overview; H, SEM overview of the same grain as in G; I, SEM detail of the same grain as in G. J–L, Ericaceae gen. indet.-type from the Cobham Lignite: J, LM overview; K, SEM overview of the same grain as in J; L, SEM detail of the same grain as in J. LM photographs ×1000, bar in SEM overview 10 µm, bar in SEM detail 2 µm. Figure 4. View largeDownload slide Sapotaceae. A–C, Sarcosperma-type from Brixton drill core 6: A, LM overview; B, SEM overview of the same grain as in A; C, SEM detail of the same grain as in A. D–F, Sarcosperma-type from Krappfeld: D, LM overview; E, SEM overview of the same grain as in D; F, SEM detail of the same grain as in D. G–I, Chrysophylleae-type from the Cobham Lignite: G, LM overview; H, SEM overview of the same grain as in G; I, SEM detail of the same grain as in G. J–L, Chrysophylleae-type from Brixton drill core 10: J, LM overview; K, SEM overview of the same grain as in J; L, SEM detail of the same grain as in J. LM photographs ×1000, bar in SEM overview 10 µm, bar in SEM detail 2 µm. Figure 4. View largeDownload slide Sapotaceae. A–C, Sarcosperma-type from Brixton drill core 6: A, LM overview; B, SEM overview of the same grain as in A; C, SEM detail of the same grain as in A. D–F, Sarcosperma-type from Krappfeld: D, LM overview; E, SEM overview of the same grain as in D; F, SEM detail of the same grain as in D. G–I, Chrysophylleae-type from the Cobham Lignite: G, LM overview; H, SEM overview of the same grain as in G; I, SEM detail of the same grain as in G. J–L, Chrysophylleae-type from Brixton drill core 10: J, LM overview; K, SEM overview of the same grain as in J; L, SEM detail of the same grain as in J. LM photographs ×1000, bar in SEM overview 10 µm, bar in SEM detail 2 µm. Figure 5. View largeDownload slide Sapotaceae. A–C, Chrysophylleae-type sp. 1 from Krappfeld: A, LM overview; B, SEM overview of the same grain as in A; C, SEM detail of the same grain as in A. D–F, Chrysophylleae-type sp. 2 from Krappfeld: D, LM overview; E, SEM overview of the same grain as in D; F, SEM detail of the same grain as in D. G–I, Mimusopeae/Isonandreae-type from Brixton drill core 13: G, LM overview; H, SEM overview of the same grain as in G; I, SEM detail of the same grain as in G. J–L, Mimusopeae/Isonandreae-type sp. 1 from Borken coalfield: J, LM overview; K, SEM overview of the same grain as in J; L, SEM detail of the same grain as in J. LM photographs ×1000, bar in SEM overview 10 µm, bar in SEM detail 2 µm. Figure 5. View largeDownload slide Sapotaceae. A–C, Chrysophylleae-type sp. 1 from Krappfeld: A, LM overview; B, SEM overview of the same grain as in A; C, SEM detail of the same grain as in A. D–F, Chrysophylleae-type sp. 2 from Krappfeld: D, LM overview; E, SEM overview of the same grain as in D; F, SEM detail of the same grain as in D. G–I, Mimusopeae/Isonandreae-type from Brixton drill core 13: G, LM overview; H, SEM overview of the same grain as in G; I, SEM detail of the same grain as in G. J–L, Mimusopeae/Isonandreae-type sp. 1 from Borken coalfield: J, LM overview; K, SEM overview of the same grain as in J; L, SEM detail of the same grain as in J. LM photographs ×1000, bar in SEM overview 10 µm, bar in SEM detail 2 µm. Figure 6. View largeDownload slide Sapotaceae. A–C, Mimusopeae/Isonandreae-type sp. 2 from Borken coalfield: A, LM overview; B, SEM overview of the same grain as in A; C, SEM detail of the same grain as in A. D–F, Mimusopeae-type sp. 1 from Krappfeld: D, LM overview; E, SEM overview of the same grain as in D; F, SEM detail of the same grain as in D. G–I, Mimusopeae-type sp. 2 from Krappfeld: G, LM overview; H, SEM overview of the same grain as in G; I, SEM detail of the same grain as in G. J–L, Mimusopeae-type from Borken coalfield: J, LM overview; K, SEM overview of the same grain as in J; L, SEM detail of the same grain as in J. LM photographs ×1000, bar in SEM overview 10 µm, bar in SEM detail 2 µm. Figure 6. View largeDownload slide Sapotaceae. A–C, Mimusopeae/Isonandreae-type sp. 2 from Borken coalfield: A, LM overview; B, SEM overview of the same grain as in A; C, SEM detail of the same grain as in A. D–F, Mimusopeae-type sp. 1 from Krappfeld: D, LM overview; E, SEM overview of the same grain as in D; F, SEM detail of the same grain as in D. G–I, Mimusopeae-type sp. 2 from Krappfeld: G, LM overview; H, SEM overview of the same grain as in G; I, SEM detail of the same grain as in G. J–L, Mimusopeae-type from Borken coalfield: J, LM overview; K, SEM overview of the same grain as in J; L, SEM detail of the same grain as in J. LM photographs ×1000, bar in SEM overview 10 µm, bar in SEM detail 2 µm. Figure 7. View largeDownload slide Sapotaceae. A–C, Isonandreae-type from Brixton drill core 13: A, LM overview; B, SEM overview of the same grain as in A; C, SEM detail of the same grain as in A. D–F, Isonandreae-type from Krappfeld: D, LM overview; E, SEM overview of the same grain as in D; F, SEM detail of the same grain as in D. G–I, Isonandreae-type sp. 1 from Borken coalfield: G, LM overview; H, SEM overview of the same grain as in G; I, SEM detail of the same grain as in G. J–L, Isonandreae-type sp. 2 from Borken coalfield: J, LM overview; K, SEM overview of the same grain as in J; L, SEM detail of the same grain as in J. LM photographs ×1000, bar in SEM overview 10 µm, bar in SEM detail 2 µm. Figure 7. View largeDownload slide Sapotaceae. A–C, Isonandreae-type from Brixton drill core 13: A, LM overview; B, SEM overview of the same grain as in A; C, SEM detail of the same grain as in A. D–F, Isonandreae-type from Krappfeld: D, LM overview; E, SEM overview of the same grain as in D; F, SEM detail of the same grain as in D. G–I, Isonandreae-type sp. 1 from Borken coalfield: G, LM overview; H, SEM overview of the same grain as in G; I, SEM detail of the same grain as in G. J–L, Isonandreae-type sp. 2 from Borken coalfield: J, LM overview; K, SEM overview of the same grain as in J; L, SEM detail of the same grain as in J. LM photographs ×1000, bar in SEM overview 10 µm, bar in SEM detail 2 µm. Figure 8. View largeDownload slide Sapotaceae and Ebenaceae. A–F, Sideroxylon-type from Brixton drill core 6: A and D, LM overviews; B and E, SEM overviews of the same grains as in A and D, respectively; C and F, SEM details of the same grain as in A and D, respectively. G–I, Diospyros/Euclea-type from Brixton drill core 10: G, LM overview; H, SEM overview of the same grain as in G; I, SEM detail of the same grain as in G. J–L, Diospyros-type sp. 1 from Changchang, Hainan: J, LM overview; K, SEM overview of the same grain as in J; L, SEM detail of the same grain as in J. LM photographs ×1000, bar in SEM overview 10 µm, bar in SEM detail 2 µm. Figure 8. View largeDownload slide Sapotaceae and Ebenaceae. A–F, Sideroxylon-type from Brixton drill core 6: A and D, LM overviews; B and E, SEM overviews of the same grains as in A and D, respectively; C and F, SEM details of the same grain as in A and D, respectively. G–I, Diospyros/Euclea-type from Brixton drill core 10: G, LM overview; H, SEM overview of the same grain as in G; I, SEM detail of the same grain as in G. J–L, Diospyros-type sp. 1 from Changchang, Hainan: J, LM overview; K, SEM overview of the same grain as in J; L, SEM detail of the same grain as in J. LM photographs ×1000, bar in SEM overview 10 µm, bar in SEM detail 2 µm. Figure 9. View largeDownload slide Ebenaceae, Styracaceae and Theaceae. A–C, Diospyros-type sp. 2 from Changchang, Hainan: A, LM overview; B, SEM overview of the same grain as in A; C, SEM detail of the same grain as in A. D–F, Styrax-type from Krappfeld: D, LM overview; E, SEM overview of the same grain as in D; F, SEM detail of the same grain as in D. G–I, Camellia-type from Krappfeld: G, LM overview; H, SEM overview of the same grain as in G; I, SEM detail of the same grain as in G. LM photographs ×1000, bar in SEM overview 10 µm, bar in SEM detail 2 µm. Figure 9. View largeDownload slide Ebenaceae, Styracaceae and Theaceae. A–C, Diospyros-type sp. 2 from Changchang, Hainan: A, LM overview; B, SEM overview of the same grain as in A; C, SEM detail of the same grain as in A. D–F, Styrax-type from Krappfeld: D, LM overview; E, SEM overview of the same grain as in D; F, SEM detail of the same grain as in D. G–I, Camellia-type from Krappfeld: G, LM overview; H, SEM overview of the same grain as in G; I, SEM detail of the same grain as in G. LM photographs ×1000, bar in SEM overview 10 µm, bar in SEM detail 2 µm. DESCRIPTIONS OF FOSSIL POLLEN TAXA AND COMPARISON WITH EXTANT TAXA Ericales Ericaceae Ericoideae Ericeae Erica L. Erica-type from Cobham (Fig. 2A–C) LM Tetrahedral tetrad, elongated elliptical outline in compressed fossilized state in basal view (Fig. 2A); tetrad size 19.3 × 25.2 µm in diameter; pollen tricolporoidate/tricolporate; wall thickness c. 1.2 µm with sexine as thick as to slightly thicker than nexine; scabrate. SEM: Micro-rugulate, fossulate to sometimes perforate (Fig. 2B), micro-rugulae smallest in the mesocolpium areas (e.g. c. 0.2 × 0.4 µm) to medium sized (e.g. c. 0.5–1.2 µm) towards the margo and the areas connected to the neighbouring pollen of the tetrad, whereas in the margo and connecting areas the micro-rugulae are fused to a tectate belt up to 2 µm wide, all micro-rugulae and tectate areas are densely covered with regularly spaced micro-gemmae (Fig. 2C). Locality Sample 48 from the laminated lignite of Cobham Lignite (Schorne Member) in Kent. Comparative remarks Several extant Erica spp. display this distinctive micro-rugulate pattern with the regularly spaced micro-gemmae under SEM (see Erica spp. described on paldat.org and in Foss & Doyle, 1988, figs 8, 9, 11–13). However, their tetrads are mostly double the size of the fossil described here, so the best fit in both size and sculpture is E. arborea L. in Sarwar (2014; SEM Fig. 2A, E) and in Halbritter (2016), a species that is widely distributed in the Mediterranean, North Africa, the Near East and in the laurel forest of the Canary Islands. Erica-type from Borken (Fig. 2D–F) LM Tetrahedral tetrad, circular to slightly triangular in outline in basal view and triangular to circular in apical view (2D); tetrad size 21.0 × 19.6 µm in diameter; pollen tricolporoidate/tricolporate; wall thickness c. 1.3–1.5 µm with sexine thinner than nexine; scabrate. SEM: Tectum in the mesocolpium areas conspicuously fossulate and rugulate (fossulae border more or less angular, irregularly shaped rugulae); in the polar areas and around the colpi the tectum is faintly fossulate and perforate (Fig. 2E), the tectum in between is covered with parallel thin striae 0.5 µm long (Fig. 2F). Locality Coal lens in the upper part of the Main Seam of the Borken Formation, Stolzenbach underground mine. Comparative remarks Finely striated angular, irregularly shaped rugulae can be found in E. carnea L. and E. herbacea L. (E. herbacea is a synonym of E. carnea), in Halbritter (2016a, b), but the striae in the modern taxa are also covered by fine micro-echini. Some Vaccinium spp. also display fine striation (Sarwar, 2006, 2007; see below), but these species display striae on less pronounced rugulae (bordered by indistinct fossulae) and have tetrads as twice as large. Phyllodoceae Kalmia L. Kalmia-type from Borken (Fig. 2G-2L) LM Tetrahedral tetrads, circular to quadrangular in outline in lateral view (Fig. 2G, J); tetrad size ranges between 28.2 × 25.8 and 38.6 × 39.3 µm in diameter (N = 5); pollen tricolporoidate/tricolporate, wall thickness 1.5–1.8 µm with the sexine thinner than nexine; scabrate. SEM: The mesocolpium areas are conspicuously rugulate and fossulate (fossulae bordering the rugulae; Fig. 2H, K), the more or less angular (triangular to trapezoidal) rugulae are ≥ c. 1 µm long and ≥ c. 0.3 µm wide (Fig. 2I, L), the rugulate-fossulate pattern is more pronounced (deeper fossulae) in the mesocolpium areas and shallower around the colpi and at the poles, rugulae are smooth (Fig. 2I, L). Locality Coal lens in the upper part of the Main Seam of the Borken Formation, Stolzenbach underground mine. Comparative remarks LM and SEM images of this pollen type are similar to Kalmia spp. described and depicted in Sarwar & Takahashi (2014; fig. 1A, C, F, G, J, K; LM and SEM images). Rhodoreae Rhododendron L. Rhododendron-type from Borken (Fig. 3A–C) LM Tetrahedral tetrads, circular to triangular in outline in apical and basal view (Fig. 3A); tetrad size c. 24.1 × 23 µm; pollen tricolporoidate/tricolporate; wall thickness c. 1.5–1.8 µm with the sexine as thick as to slightly thinner than nexine; scabrate sculpturing. SEM: The mesocolpium areas are rugulate and fossulate to perforate (fossulae border more or less angular irregularly shaped rugulae). The fossulate-rugulate pattern is much shallower and smoother in the polar areas and around the colpi (Fig. 3B; faintly recognizable fossulae and perforations), the rugulae display faint parallel grooves (Fig. 3C); to the top right a viscin thread is visible (Fig. 3B). Locality Coal lens in the upper part of the Main Seam of the Borken Formation, Stolzenbach underground mine. Comparative remarks This fossil tetrad resembles in shape, size and faintly sculptured tectum surface Rhododendrum hirsutum L. in Halbritter (2016d) or the description and SEM images of the slightly larger tetrads of the Chinese and Japanese R. tsusiophyllum Sugim. in Sarwar & Takahashi (2013; fig. 1C overview; fig. 3F detail). Vaccinioideae Vaccinieae Vaccinium L. (Fig. 3D–F) Vaccinium-type from Cobham LM Tetrahedral tetrad, circular to triangular outline in slightly compressed basal view and circular to quadrangular in lateral view (Fig. 3D); tetrad size 30 × 32 µm; pollen tricolporoidate/tricolporate; wall thickness between 1.2 µm at the poles and 1.5 µm at the mesocolpium areas with the sexine as thick as to thicker than nexine; scabrate. SEM: The whole tectum surface is composed of tightly packed evenly shaped micro-rugulae (c. 0.13 µm wide and c. 0.6–0.8 µm long; Fig. 3E, F) that are partly parallel and partly irregularly arranged, with few intermingled micro-verrucae (c. 0.1 µm in diameter); this micro-rugulate pattern is more uneven and undulating and accompanied by fossulae and perforations in the mesocolpium areas (Fig. 3E). Locality Sample 48 from the laminated lignite of the Cobham Lignite (Schorne Member) in Kent. Comparative remarks Few Vaccinium spp. of Vaccinium section Conchophyllum Sleumer (V. emarginatum Hayata) and to a lesser extent of sections Bracteata Nakai (V. wrightii A.Gray) and Ciliata (V. oldhamii Miq.) display this characteristic micro-rugulate pattern (Sarwar, 2006; fig. 1F–J; Sawar, 2007, figs 3-42L–3-42O, 3-43K, 3-44H); all are continental Asian species. Gaultherieae Gaultheria Kalm ex. L. Gaultheria-type from Brixton (Fig. 3G–I) LM Tetrahedral tetrad, quadrangular to circular outline in compressed state lateral view (Fig. 3G); tetrad size 26.3 × 25.2 µm; pollen tricolporoidate; wall thickness c. 1.1 µm; scabrate. SEM: Micro-rugulate and fossulate, with the micro-rugulae tightly packed, shallow and more angular and unevenly sized (c. 0.2–0.6 µm wide and 0.5–1.5 µm long) at the polar and the colpi areas (Fig. 3H), and considerably smaller (c. 0.2–0.3 µm wide and 0.5 µm long), less angular in shape and loosely packed (with pronounced fossulae between micro-rugulae) in the mesocolpium areas; all micro-rugulae are smooth (Fig. 3I). Locality Brixton drill hole 6 at 10.5 m depth, Reading Formation, Upper Mottled Beds. Comparative remarks This tetrad type resembles in many aspects pollen tetrads from various Gaultheria taxa, such as G. prostrata W.W.Sm., G. bracteata (Cav.) G.Don and G. itatiaeae Wawra, in Sarwar (2006; figs 3G, H, 4E, respectively). All three are South American mountain taxa. Ericaceae indet.-type from Cobham (Fig. 3J–L) LM Tetrahedral tetrad, triangular to trilobate outline in apical and basal view (Fig. 3J); tetrad size c. 25.1 × 20.2 µm; pollen tricolporate/tricolporoidate; wall thickness c. 1 µm; scabrate. SEM: Rugulate and fossulate, rugulae are shallow, poly-angular with different sizes (0.5–2.5 µm in diameter) and bordered by fossulae at the polar and margo areas, whereas in the mesocolpium areas (Fig. 3K, L) the considerably smaller rugulae and fossulae are more pronounced, and rugulae display slightly uneven surfaces with faint indentions (Fig. 3L). Locality Sample 48 from the laminated lignite of the Cobham Lignite (Schorne Member) in Kent. Comparative remarks No further botanical assignment is possible. Sapotaceae Sarcospermatoideae Sarcosperma Hook.f. Sarcosperma-type from Brixton (Fig. 4A–C) LM Tetracolporate, prolate pollen grain, elliptical outline in equatorial view and slightly protruding apertural areas at the equator (Fig. 4A); polar axis 18.6 µm and equatorial axis 14.8 µm; wall thickness at the poles c. 1.2 µm with sexine as thick as to thicker than nexine, in the apertural areas only slightly thicker; colpi are c. 11 µm long, endopori circular to slightly lalongate (Fig. 4A), c. 2.8 µm high and c. 3 µm wide, costae not conspicuous; scabrate. SEM: Regularly spaced perforations on the poles and around the colpi, fossulate to perforate in the mesocolpium areas (Fig. 4B), fossulae orientation towards the mesocolpium in the equatorial areas (Fig. 4C), tectum surface slightly undulating without further ornamentation (Fig. 4B, 4C). Locality Brixton drill hole 6 at 10.5 m depth, Reading Formation, Upper Mottled Beds. Comparative remarks This pollen type resembles SEM images of pollen of Sarcosperma laurinum (Benth.) Hook.f. and S. griffithii Hook.f. ex C.B.Clark and LM images of S. arboreum Hook.f. of Harley’s Sarcosperma subtype IVA (Harley, 1991; fig. 17A, 1B, 1C, 1K, 1L, respectively); all three are Indian and Chinese taxa. Subtype IVA of Harley (1991) includes also two species of Chrysophylleae: Pradosia brevipes (Pierre) T.D.Penn. and Pouteria rigida Radlk. (fig. 17D–G, M, N), although both produce larger pollen grains and have circular endoapertures. Additionally, P. bevipes does not show costae under LM, the perforate fossulate tectum pattern under SEM is smaller-scaled than in Sarcosperma, and Pouteria rigida has much thicker pollen walls than Sarcosperma. Sarcosperma-type from Krappfeld (Fig. 4D–F) LM Tetracolporate, prolate pollen grain, elliptical outline in equatorial view and slightly protruding apertural areas at the equator (Fig. 4D); polar axis 23.6 µm and equatorial axis 17.2 µm; wall thickness at the poles c. 1.1 µm with the sexine as thick as the nexine; colpi are c. 14 µm long, endopori not easily discernible (Fig. 4D): circular to lalongate c. 2.5 µm high and c. 3 µm wide; scabrate. SEM: Regularly perforate fossulate, fossulae are often radially arranged (two to four) and start from perforations (Fig. 4E, F), tectum surface slightly undulating without further ornamentation (Fig. 4F). Locality Holzer Formation sample PQ9 in the Krappfeld (Carinthia). Comparative remarks This pollen type also resembles LM and SEM images of Sarcosperma taxa in Harley (1991, fig. 17A–C, H–L; see above). Remarks to the two Sarcosperma-types: These two pollen types could be interpreted to belong either to two quite closely related taxa or one variable taxon, but there is too little material to reach a conclusion. Chrysophylloideae Chrysophylleae Chrysophylleae-type from Cobham (Fig. 4G–I) LM Tetra- to pentacolporate subprolate to spheroidal pollen grain, circular to elliptical outline in equatorial view (Fig. 4G); polar axis ranges from 26.6 to 30.1 µm and equatorial axis from 26 to 27 µm (N = 2); wall thickness at the poles c. 1.2–1.5 µm with the sexine thicker than the nexine, in the apertural areas c. 2–3 µm thick with the nexine slightly thickening towards the equatorial area; endopori lalongate to circular 2.5–3.0 µm high and 3.3–4.5 µm wide (Fig. 4G); scabrate. SEM: Irregularly micro-rugulate, fossulate, perforate (Fig. 4H, I), fossulae are more pronounced in the mesocolpium areas, visible colpus membrane micro-granular (Fig. 4I); colpi are short, c. 10–15 µm long. Locality Sample 48 from the laminated lignite of the Cobham Lignite (Schorne Member) in Kent. Comparative remarks This pollen type resembles SEM and LM images of Pouteria robusta (Mart. & Eichler) Eyma [= P. cuspidata (A.DC.) Baehni subsp. robusta (Mart. & Eichler) T.D.Penn.] and SEM images of Elaeoluma nuda (Baehni) Aubrév. coll. Philcox 3257 of the Pichonia subtype VC of Harley (1991; fig. 22G–J, E, F, respectively). The related subtype VA includes, in addition to Chrysophylleae species, also the Mimusopae Autranella congolensis (De Wild.) Chev. ex Aubrév. & Pellegr. (Harley, 1991, fig. 20G, 2H), which is not comparable to the Chrysophylleae-type from Cobham, because it is twice as large and has conspicuous lalongate endopori. Another Elaeoluma nuda coll. Maquire 24596 in Harley (1991; fig. 38N) looks different: prolate and conspicuously perforate to foveolate in the polar areas. Both taxa occur in north-eastern South America. Chrysophylleae-type from Brixton (Fig. 4J–L) LM Pentacolporate subprolate to spheroidal pollen grain, elliptical to circular outline in equatorial view (Fig. 4J); polar axis 31.6 µm and equatorial axis 30 µm; wall thickness c. 1.5 µm with the sexine thicker than the nexine; endopori lalongate elliptical c. 4.5 µm high and c. 8 µm wide, colpi are short c. 15 µm; scabrate. SEM: micro-rugulate, fossulate with the fossulae bordering the micro-rugulae (Fig. 4K, L), poles are more perforate and smoother than the mesocolpium areas (Fig. 4K) where the fossulae are more pronounced and the more or less angular micro-rugulae (c. 0.3–0.5 µm wide and c. 0.5–1.5 µm long) display a perpendicular faint striate ornamentation (Fig. 4L); visible colpus membrane micro-granular to micro-rugulate (Fig. 4L). Locality Brixton drill hole 10 at 11.9 m depth, ?Reading Formation, channel sands above Lower Mottled Beds. Comparative remarks The tectum of this pollen type resembles to a certain degree SEM images of the New Caledonian ‘Pouteria’/Planchonella baillonii (Zahlbr.) Dubard, the North Australian ‘Pouteria’ sericea (Ait.) Baehni [synonym Sersalisia sericea (Ait.) R.Br.] in Harley (1991; figs 31M, 35N, respectively) and the LM image of P. linggensis (Burck) Baehni [synonym Planchonella linggensis (Burck) Pierre] in Harley (1991; fig. 33L), which is also from New Caledonia. Chrysophylleae-type sp. 1 from Krappfeld (Fig. 5A–C) LM: Tetracolporrate spheroidal pollen grain, circular outline in equatorial view (Fig. 5A); polar axis 21.5 µm and equatorial axis 21.1 µm; wall thickness at the poles c. 1.2 µm with the sexine thicker than the nexine, in the apertural areas c. 1.8 µm thick with the nexine slightly thickening towards the equatorial area (Fig. 5A); colpi are short, 8 µm long, endopori lalongate 1.2 µm high and 3 µm wide (Fig. 5C); scabrate. SEM: Irregularly fossulate and micro-rugulate, perforate to sparsely foveolate (Fig. 5B, C). Locality Holzer Formation sample PQ8 in the Krappfeld (Carinthia). Comparative remarks This pollen type closely resembles Pouteria robusta and Elaeoluma nuda coll. Philcox 3257, both in Harley (1991; fig. 22G and 22J, E, F, respectively); extant occurrences, see above. Chrysophylleae-type sp. 2 from Krappfeld (Fig. 5D–F) LM Tetracolporate subprolate pollen grain, elliptical to circular outline in equatorial view and slightly protruding apertural areas at the equator (Fig. 5D); polar axis 24.3 µm and equatorial axis 20.9 µm; wall thickness at the poles c. 1.1 µm with the sexine thicker than the nexine, in the apertural areas c. 2.2 µm thick; colpi are short, 9 µm long, endopori lalongate to circular 1.2 µm high and 3.6 µm wide (Fig. 5D); scabrate. SEM: irregularly sparsely fossulate, perforate, tectum surface slightly undulating without further ornamentation (Fig. 5E, F). Locality Holzer Formation sample PQ8 in the Krappfeld (Carinthia). Comparative remarks This pollen type resembles South American Pouteria robusta in Harley (1991; fig. 22G–J). Sapotoideae Mimusopeae/Isonandreae Mimusopeae/Isonandreae-type from Brixton (Fig. 5G–I) LM Pentacolporate, subprolate pollen grain, with an elliptical outline in equatorial view (Fig. 5G); polar axis 30.2 µm, equatorial axis 25.1 µm; wall thickness c. 1.5 µm with the sexine thinner than nexine; endopori lalongate c. 1.7 µm high and c. 4 µm wide, colpi are short, c. 11 µm long (Fig. 5G); scabrate. SEM: The tectum is unevenly criss-cross striate to micro-rugulate in the mesocolpium areas (Fig. 5H, I), the micro-rugulae are composed of sometimes parallel arranged and sometimes criss-crossing micro-striae c. 0.2–0.3 µm wide and c. 0.5–1.5 µm long (Fig. 5I); colpus membrane micro-granular (Fig. 5H). Locality Brixton drill hole 13 at 14.9 m depth, Reading Formation, Upper Mottled Beds. Comparative remarks To some extent in the apertures, wall and tectum this pollen resembles pollen of various Sapoteae, such as Labourdonnaisia revoluta Bojer in Harley (1991, fig. 11G, SEM) from Mauritius and Indian taxa, for example Palaquium obovatum (Griff.) Engl. (LM image, fig. 9H), Madhuca neriifolia (Moon) H.J.Lam (SEM overview, fig. 9J). Madhuca J.F.Gmel. sp. and M. tubulosa H.J.Lam (SEM and LM images, fig.11B, C–D), all in Harley (1991), but none is identical to the fossil taxon. Mimusopeae/Isonandreae-type sp. 1 from Borken (Fig. 5J–L) LM Tetracolporate, prolate pollen grain, elliptical to slightly rhomboidal outline in equatorial view (Fig. 5J); polar axis 29 µm, equatorial axis 21.9 µm; wall thickness c. 1.4 µm with the sexine as thick as nexine; endopori lalongate c. 1.4 µm high and c. 3.5 µm wide, colpi are c. 22 µm long (Fig. 5J); scabrate. SEM: The tectum is irregularly perforate in the mesocolpium areas (Fig. 5K), more regularly perforate in the polar areas and densely covered by minute, regularly arranged circular micro-verrucae c. 0.1 µm in diameter (Fig. 5L). Locality Coal lens in the upper part of the Main Seam of the Borken Formation, Stolzenbach underground mine. Comparative remarks This specimen strongly resembles LM and SEM images of the African Mimusops angel Chiov. and Malaysian‘Payena’ = Madhuca lanceolata (Merr.) Merr. in Harley (1991; figs 5G, 6D–F, respectively). Mimusopeae/Isonandreae-type sp. 2 from Borken (Fig. 6A–C) LM Tetracolporate, prolate pollen grain, elliptical outline in equatorial view (Fig. 6A); polar axis 29.1 µm, equatorial axis 22.7 µm; wall thickness c. 1.6 µm with the sexine as thick as nexine in the polar areas and slightly thicker than the nexine in the equatorial areas; endopori lalongate c. 1.5 µm high and c. 5 µm wide (Fig. 6A), colpi do not reach the polar areas (17 µm long); scabrate. SEM: The tectum is sparsely and irregularly perforate (Fig. 6B) and densely covered by shallow, minute, regularly arranged, slightly elongated micro-verrucae c. > 0.1 µm in diameter (Fig. 6C). Locality Coal lens in the upper part of the Main Seam of the Borken Formation, Stolzenbach underground mine. Comparative remarks This specimen is generally similar in size, colpus length, pollen wall and apertural measurements to the south-east African Mimusops obovata (Sond.) Nees ex Harv. and M. caffra E.Mey. ex A.DC. in Harley (1991; SEM and LM images fig. 5A–C, LM images fig. 6H, J), but the tectum surface under SEM is similar to Palaquium ottolanderi (Koord. & Valeton) Baehni in Harley (1991; fig. 7G) from Java. Mimusopeae Mimusopeae-type sp. 1 from Krappfeld (Fig. 6D–F) LM Tetracolporate, prolate pollen grain, elongated elliptical outline in equatorial view (Fig. 6D); polar axis 21.8 µm, equatorial axis 14.1 µm; wall thickness c. 1 µm with the sexine thicker than the nexine; endopori lalongate c. 2 µm high and 4.5 µm wide, colpi c. 11.7 µm long (Fig. 6D); scabrate. SEM: The poles are densely perforate to foveolate (abraded status of fossilized pollen) and the mesocolpium areas are sparsely perforate (Fig. 6E); the tectum is composed of evenly and densely arranged minute micro-verrucae (Fig. 6F; abraded tectum surface). Locality Holzer Formation sample PQ8 in the Krappfeld (Carinthia). Comparative remarks The conspicuously perforate to foveolate poles together with the evenly and densely packed micro-verrucae are known from Mimusops antongilensis Aubrév. (Madagascar) and M. marginata N.E.Br. [South Africa = Vitellariopsis marginata (N.E.Br.) Aubrév. in Harley (1991; LM and SEM images, fig. 10B–F)]. Mimusopeae-type sp. 2 from Krappfeld (Fig. 6G–I) LM Tri- to tetracolporate, prolate pollen grain, elongated elliptical outline in equatorial view and protruding apertures at the equator (Fig. 6G); polar axis c. 25–28 µm, equatorial axis c. 21.5–22.1 µm (N = 2); wall thickness c. 1.4–2.0 µm at the poles and c. 1.8–2.4 at the equator with the sexine thicker than the nexine; endopori lalongate c. 2.1–2.4 µm high and 3.1–3.5 µm wide, costae visible, colpi are c. 13.0–14.8 µm long (Fig. 6G); scabrate. SEM: Perforate to faintly fossulate, more pronounced at the poles (Fig. 6H), the tectum surface is slightly undulating and composed of evenly and densely arranged minute micro-verrucae < 0.1 µm in diameter (Fig. 6I). Locality Holzer Formation sample PQ5b in the Krappfeld (Carinthia). Comparative remarks The perforations and the evenly and densely packed micro-verrucae are similar to Mimusops antongilensis, but even more so to Mimusops marginata (=Vitellariopsis marginata) and Vitellariops kirkii (Barker) Dubard in Harley (1991, fig. 10B, 10C–F, SEM images fig. 25L, M, respectively); all three are African taxa. Mimusopeae-type from Borken (Fig. 6J–L) LM Tetracolporate, prolate pollen grain, elongated elliptical outline in equatorial view (Fig. 6J); polar axis 33.8 µm, equatorial axis 20.5 µm; wall thickness c. 2 µm with the sexine thinner than nexine at the poles and as thick as the nexine in the equatorial area; endopori lalongate c. 2 µm high and 4.5 µm wide, colpi reach the polar areas (Fig. 6J); scabrate sculpturing. SEM: The poles are irregularly perforate (Fig. 6K) and the tectum is densely covered by minute, shallow micro-verrucae and few rugulae (Fig. 6K, L), whereas in the mesocolpium areas the tectum is irregularly fossulate to perforate and rugulate, the fossulae border the irregularly shaped rugulae (c. 1.0–1.5 µm wide and 2–3 µm long), the rugulae are covered with minute, shallow micro-verrucae and sometimes micro-rugulae (sizes < 0.1 µm; Fig. 6L). Locality Coal lens in the upper part of the Main Seam of the Borken Formation, Stolzenbach underground mine. Comparative remarks The LM images resemble pollen of Manilkara pubicarpa Monach. in Harley (1991, fig. 9F, G), despite the higher endoapertures. However, the tectum pattern as seen with SEM is similar to that of Manilkara triflora (Allemão) Monach. in Harley (1991; fig. 9D as ‘Mimusops’ triflora Allemão); both taxa are from South America. Isonandreae Isonandreae-type from Brixton (Fig. 7A-7C) LM Tetracolporate, prolate to subprolate pollen grain, elliptical outline in equatorial view in a slightly oblique compressed state (Fig. 7A); polar axis 24.3 µm, equatorial axis 20.8 µm; wall thickness c. 1.2 µm with the sexine as thick as nexine; endopori lalongate c. 3.5 µm high and c. 5 µm wide, colpi c. 15.4 µm long (Fig. 7A); scabrate. SEM: Perforate and micro-striate, perforations and micro-striation more pronounced in the mesocolpium areas (Fig. 7B); micro-striae are c. 0.25 µm wide and c. 0.4–2.0 µm long and arranged in a regular criss-cross pattern around the perforations (Fig. 7C). Occurrence Brixton drill hole 13 at 12.3 m depth, Woolwich Formation, Upper Shelly Beds. Remarks LM images and images of the tectum under SEM resemble the Philippine Palaquium philippense (Perr.) C.B.Rob. and Malaysian Madhuca sp. and M. tubulosa in Harley (1991; figs 10H, L, 11B–D, respectively). Isonandreae-type from Krappfeld (Fig. 7D–F) LM Tetracolporate, prolate pollen grain, elliptical outline in equatorial view and slightly protruding apertural areas at the equator (Fig. 7D); polar axis ranges from 25.5 to 30.5 µm and equatorial axis from 17.5 to 22.1 µm (N = 3); wall thickness c. 1.3–1.5 µm at the poles and 1.5–1.8 µm at the equatorial region with the sexine as thick as to slightly thinner than nexine; endopori lalongate to elliptical c. 4.0–4.2 µm high and 4–5 µm wide (Fig. 7D), colpi reach the polar areas and are 25.5–27.0 µm long; scabrate. SEM: Regularly perforate and evenly micro-rugulate (Fig. 7E), micro-rugulae and micro-striae (sizes < 0.1 µm wide and 0.2–1.0 µm long) are partly arranged parallel or perpendicular to each other, the micro-striae tend to frame sets of parallel arranged shorter micro-rugulae (Fig. 7E,7F). Locality Holzer Formation sample PQ8 in the Krappfeld (Carinthia). Comparative remarks The SEM images of the tectum resemble Malaysian Palaquium leiocarpum Boerl. in Harley (1991; fig. 11H). The other characteristics (size, colpus length, pollen wall and apertural measurements) are similar to other taxa of Palaquium Blanco in Harley (1991; see subtype IIA). Isonandreae-type sp. 1 from Borken (Fig. 7G–I) LM Tetracolporate, prolate pollen grain, elongated elliptical outline in equatorial view (compressed state; Fig. 7G); polar axis 26.2 µm, equatorial axis 13 µm; wall thickness c. 1.2–1.4 µm with the sexine as thick as to thinner than nexine; endopori lalongate to elliptical c. 2.5–2.8 µm high and 4–5 µm wide (Fig. 7H), colpi reach the polar areas; scabrate sculpturing. SEM: The poles are sparsely but regularly perforate and the tectum is covered by micro-rugulae (Fig. 7H, 7I), whereas in the mesocolpium areas the tectum is regularly perforate and covered by micro-rugulae (Fig. 7I; sizes < 0.1 µm wide and 0.3 µm long). Locality Coal lens in the upper part of the Main Seam of the Borken Formation, Stolzenbach underground mine. Comparative remarks The SEM images also resemble the Malaysian Palaquium leiocarpum in Harley (1991; fig. 11H). The other characteristics (size, colpus length, pollen wall and apertural measurements) are comparable to other Palaquium taxa in Harley (1991; see subtype IIA). Isonandreae-type sp. 2 from Borken (Fig. 7J–L) LM Tetracolporate, prolate pollen grain, elliptical outline in equatorial view (Fig. 7J); polar axis 26.6 µm, equatorial axis 20.9 µm; wall thickness c. 2 µm with the sexine thinner than nexine in the polar areas and slightly thicker than to as thick as nexine in the equatorial area; endopori lalongate c. 2.8–3.0 µm high and c. 6 µm wide (Fig. 7J), colpi reach the polar areas; scabrate sculpturing. SEM: The tectum is faintly perforate (Fig. 7K) and covered by sets of parallel arranged micro-rugulae 0.1 µm wide and 0.3–1.0 µm long (Fig. 7L). Locality Coal lens in the upper part of the Main Seam of the Borken Formation, Stolzenbach underground mine. Comparative remarks The specimen is strongly reminiscent of Palaquium philippense pollen (Harley, 1991, fig 10G–L, LM and SEM images) and resembles the Isonandreae-type from Krappfeld. Remarks on all Isonandreae pollen types with Palaquium affinities: All fossil pollen taxa here described show strong affinities to Palaquium, but do not really resemble fossil pollen affiliated to this genus depicted as LM images in Gruas-Cavagnetto (1976: plate 4, figs 13, 12, 16; 27–28, for affiliations see also Armstrong, 2010: chapter 6). Sideroxyleae Sideroxylon L. Sideroxylon-type from Brixton (Fig. 8A–F) LM Tetracolporate, prolate pollen grain, elliptical outline in equatorial view (Fig. 8A, D); polar axis 20.7–25.4 µm, equatorial axis 15.6–18.1 µm (N = 3); wall thickness c. 0.9–1.4 µm with the sexine as thick as to slightly thinner than nexine; endopori more or less equi-dimensional c. 2.9–5.5 µm high and c. 3.0–4.5 µm wide Fig. 8D, F), costae visible, colpi are c. 12.3–14.3 µm long; scabrate. SEM: Perforate to foveolate, fossulate, micro-rugulate (Fig. 8B, E), with perforations, fossulae and micro-rugulae arranged in a criss-cross pattern more pronounced in the polar areas (Fig. 8F); the mesocolpium and colpi areas are more fossulate and perforate (Fig. 8B, E), fossulae can join perforations (Fig.8F) and tend to run more or less parallel to the polar axis in the mesocolpium areas and perpendicular to the colpi (Fig. 8C). Locality Brixton drill hole 6 at 10.5 m depth, Woolwich Formation, Upper Shelly Beds. Comparative remarks The tectum under the SEM strongly resembles LM and SEM images of the American S. reclinatum Michx. whereas the tectum under the SEM somewhat resembles the Chinese S. wightianum Hook. & Arn. in Harley (1991; figs 36A–D and 13G, respectively). Our pollen taxon is not comparable to Tetracolporopollenites halimbaense Kedves, T. megadolium (Potonié) Frederiksen or T. obscurus Pflug & Thomson depicted as LM images by Gruas-Cavagnetto (1976; plate 4 figs 5, 10–11, 14–15, respectively) from the lower Eocene of England, which have been interpreted as belonging to the ‘Sideroxylon pollen type’ of Harley (1991; see Armstrong, 2010: chapter 6). Ebenaceae Ebenoideae Diospyros L./Euclea L. Diospyros/Euclea-type from Brixton (Fig. 8G–I) LM Tricolporate, prolate pollen grain, elliptical outline in equatorial view and slightly protruding apertures at the equator (Fig. 8G); polar axis c. 38.5 µm, equatorial axis c. 20.9 µm; wall thickness c. 0.9–1.0 µm; colpi reach the polar areas, endopori rectangular to circular with thickened margins, c. 1.8–3.0 µm high and 3.5 µm wide (Fig. 8G); scabrate. SEM: Evenly micro-rugulate, fossulate to perforate (Fig. 8H), micro-rugulae are of irregularly rectangular shape 0.5 µm wide and c. 0.8–2.0 µm long that are regularly covered with rows of minute micro-echini that run perpendicular to the length of the micro-rugulae (Fig. 8I). Locality Brixton drill hole 10 at 16.5 m depth, ?Reading Formation, channel sands above Lower Mottled Beds. Comparative remarks SEM image of the tectum with supratectal minute micro-echini resembles somewhat the much smaller Central African pollen of Diospyros iturensis (Gürke) Letouzey & F.White, the much smaller Euclea tomentosa E.Mey. ex A.DC. pollen, a taxon that occurs today in the Cape Province, and slightly resembles D. viridicans Hiern, all in Geeraerts et al. (2009; figs 2N, 3H, 3J, respectively) from Africa. However, none of the extant taxa is especially similar to the fossil pollen type. Diospyros-type sp. 1 from Hainan (Fig. 8J–L) LM Tricolporate, prolate pollen grain, elongated elliptical outline in equatorial view and slightly protruding apertures at the equator (Fig. 8J); polar axis c. 50 µm, equatorial axis c. 30.1 µm; wall thickness c. 1.8–2.5 µm with the sexine as thick as nexine; colpi reach the polar areas and are c. 40 µm long, endopori rectangular to circular, c. 1.8 µm high and 3.75 µm wide with thickened margins (Fig. 8J); scabrate. SEM: Evenly micro-rugulate, fossulate and perforate (Fig. 8K, L), micro-rugulae are of irregularly rectangular to polygonal shapes with sizes ranging from 0.4 to 0.7 µm wide and c. 1 to 2 µm long that are ornamented by regular, mostly parallel arranged striae that run perpendicular to the length of the micro-rugulae (Fig. 8L), colpi have rounded ends (Fig. 8K). Locality Hainan, Changchang Formation, sample I/14-3 Comparative remarks The LM image is comparable with the LM images of the African Dipospyros mespiliformis Hochst. ex A.DC in size, apertures and pollen wall thickness (Gosling, Miller & Livingstone, 2013: plate XLIII, figs 1–3). The tectum under SEM resembles somewhat Diospyros longiflora Letouzey & F.White, which is endemic to Cameroon and Gabon, and D. monbuttensis Gürke in Geeraerts et al. (2009: figs 1O, 2F, respectively) from tropical west Africa, but the latter two are smaller than the fossil material. Diospyros-type sp. 2 from Hainan (Fig. 9A–C) LM Tricolporate, prolate pollen grain, elliptical to rhomboidal outline in equatorial view and slightly protruding apertures at the equator (Fig. 9A); polar axis 29.7–36.9 µm, equatorial axis 23.3–26.4 µm (N = 4); wall thickness c. 1.2–1.4 µm with the sexine as thick as nexine; colpi reach the polar areas and are c. 24.0–31.5 µm long, endopori lalongate c. 1.3–1.8 µm high and c. 2.5–4.8µm wide (Fig. 9A); faint costae just visible; scabrate. SEM: Evenly finely micro-striate (Fig. 9B, C), the striae are c. 0.14 µm wide and c. 0.4–1.5 µm long and display a faint ‘annelid pattern’ (Fig. 9C), the striae are arranged partly parallel to each other, are sometimes fused and sets of striae are arranged at an angle to each other (Fig. 9C); colpus ends are rounded (Fig. 9B). Locality Hainan, Changchang Formation sample I/14-1. Comparative remarks The LM image is comparable with the LM images of the African Dipospyros abyssinica Hiern in size, apertures and pollen wall thickness (Gosling et al., 2013: plate XCIII figs 1–4). SEM image of the tectum resembles three central African taxa of Diospyros: D. rotundifolia Hiern, D. batocana Hiern and D. usambarensis Hiern pollen (SEM images in Geeraerts et al., 2009: figs 1M, 2G, 2K, respectively). Styracaceae Styrax L. Styrax-type from Krappfeld (Fig. 9D-9F) LM Tricolporate, prolate pollen grain, rhomboidal to elliptical outline in equatorial view (Fig. 9D); polar axis 18.1 µm, equatorial axis 15.5 µm; wall thickness c. 1 µm with the sexine thicker than nexine in the equatorial area and as thick as to slightly thicker than nexine in the polar areas; colpi reach the polar areas, endopori lalongate to rectangular c. 1 µm high and 2.5 µm wide (Fig. 9D); psilate. SEM: Regularly perforate and shallowly micro-rugulate (Fig. 9E), the micro-rugulae produce a micro-crotonoid pattern around the perforations (Fig. 9F), sets of micro-crotonoid-like arranged micro-rugulae 0.5–0.8 µm; margo tectate (Fig. 9E) and up to 1.7 µm wide at the equator; ectexine produces a bridge-like structure c. 1.5 µm wide at the equator (Fig. 9D). Locality Holzer Formation sample PQ8 in the Krappfeld (Carinthia). Comparative remarks This pollen type resembles in its small size and aperture configuration the LM images of Huodendron Rehder sp. in Morton & Dickison (1992; fig. 2A, B, E, F) but not the SEM images, in which it is more reminiscent of the overview and detailed SEM images of Styrax japonica L. and, to a lesser extent, S. obassia Sieb. & Zucc. in Miyoshi, Fujiki & Kimura (2011: plate 208 figs 1–4, 5–8, respectively), further Styrax taxa in Morton & Dickison (1992; S. pallidus A.DC.: fig. 5C, D LM images, 5K, SEM image; S. officinalis L.: fig. 6A, B SEM images; S. japonica Sieb. & Zucc.: fig. 6D SEM image; S ovatus A.DC. fig. 7B SEM detail) and to an even lesser extent Rhederodendron macrocarpum Hu and Sinojackia xylocarpa Hu in Morton & Dickison (1992: figs 2K, 4N, both SEM images of tectum, respectively). All of the above mentioned taxa, except S. pallidus and S. ovatus (both South America), are deciduous Asian taxa (Morton & Dickison 1992) and all extant pollen grains are considerably larger than the fossil pollen except the pollen of Huodendron. Theaceae Theeae Camellia L. Camellia Krappfeld-type (Fig. 9G–I) LM Tricolporate pollen grain, oblate in compressed fossilized state, rounded triangular to circular outline in polar view (Fig. 9G); equatorial diameter 18.2–19.5 µm; wall thickness c. 1.2 µm with the sexine thinner than nexine, but as thick as nexine around the apertures; colpi reach the polar areas, endopori wide (c. 3.5 µm) but difficult to discern (Fig. 9G); scabrate. SEM: Densely covered by irregularly arranged micro-rugulae, perforate to fossulate (Fig. 9H), micro-rugulae c. 0.25 µm wide and up to 0.8 µm long with an ‘annelid pattern’ and partly covered by minute micro-gemmae (Fig. 9I). Locality Holzer Formation sample PQ9 in the Krappfeld (Carinthia). Comparative remarks Reminiscent of many Camellia spp. (see Zavada & Wei, 1993: all plates, Miyoshi et al., 2011: plate 69, figs 1–4), in particular C. sinensis (L.) Kuntze and C. yunnensis Cohen-Stuart (Fendt, 1996: plate 7, figs 1–7 and plate 8, figs 8–14, respectively), despite the fact that many extant Camellia have larger pollen grains. DISCUSSION The earliest records of mesofossils with ericalean affinities span the time between the late Coniacian and early Campanian, in which charred or coalified flowers have been found in the USA and Sweden (Friis, 1985; Schönberger & Friis, 2001; Martínez-Milán, Crepet & Nixon, 2009; Schönberger et al., 2012; Crepet, Nixon & Daghlian, 2013). No pollen with an ericalean affiliation has been reported from the associated sediments. Ericaceae The earliest occurrence of an ericaceous fossil in the fossil record is a fossilized flower Paleoenkianthus sayrevillensis Nixon & Crepet from North America (Nixon & Crepet, 1993, Crepet et al., 2013) probably of late Coniacian to early Turonian age (see age discussion in Massoni, Doyle & Sauquet, 2015). This is then followed by a Santonian fossil flower sharing characters of Actinidiaceae and Clethraceae, also from North America (Schönenberger et al., 2012). The first ‘Ericaceae s.l. pollen tetrads’ are known from the Maastrichtian of Europe (Krutzsch, 1970). These Ericaceae-like pollen tetrads were named in the literature as species or form genera names such as Ericipites Wodehouse, Tetradopollenites Thiele Pfeiffer and Ericaceoipollenites Potonié; the latter two have been claimed to be synomyms of Ericipites in a review paper on pollen tetrads by Krutzsch (1970). Krutzsch (1970) pointed out that these tetrads could also be related to Empetraceae, Pyrolaceae or Epacridaceae [groups assigned to Ericaceae sensuAPG IV (2016)]. There are also images of Campanian/Maastrichtian aged ‘Ericipites’ from Australia (Carpenter et al., 2015; fig. 7J), but if the picture is studied carefully one can see that it is not a compact tetrad of four pollen grains but three pollen grains in a tetrahedral arrangement with one pollen grain lost. It is possible that this ‘Ericipites’ is not a member of Ericaceae. After a time gap, there are Ericipites records from the Palaeocene of Argentina (Archangelsky & Zamaloa, 2003: Epacridaceae?) and Seymour Island in Antarctica (Askin, 1990: Epacridaceae?). From the Eocene onwards, Ericipites has been frequently reported from various European localities [e.g. Kedves, 1969; Kruztsch & Vanhoorne, 1977; Gruas-Cavagnetto, 1978, Tetradopollenites ericius (Potonié) Thomson & Pflug and T. callidus (Potonié) Thomson & Pflug; Zetter & Hesse, 1996, Rhododendron-like since the mid-Eocene] and also from localities in the Southern Hemisphere (Harris, 1972, mid-Eocene), followed later by a few fossilized fruit, seed and leaf findings (see below). Phylogenetic studies demonstrated the monophyly of Ericales (Bremer et al., 2002) and that of Ericaceae (Kron et al., 2002) and additionally revealed that the family is of Laurasian origin, either from North America or North America plus Eurasia (Kron & Luteyn, 2005). Erica is a huge genus with > 800 species in Europe, the Middle East and Africa; it has its highest diversity in the Cape Flora (McGuire & Kron, 2005). Erica arborea has the widest distribution of all Erica spp. extending from the Mediterranean biome in Europe and North Africa to the Ethiopian highlands. Phylogenetic studies (McGuire & Kron, 2005) have suggested that E. arborea is sister to all the African Erica taxa. They also assumed that the European ancestor of the African lineages must have been as widespread as the extant E. arborea and that migration and radiation through Africa to South Africa could have taken place since the mid-Miocene (c. 15 Mya), a result that was corroborated by Pirie et al. (2016). Of the two fossil Erica-type taxa described here, the older one from Cobham (PETM) closely resembles the extant E. arborea and might be related to the lineage that later led to the evolution and migration of the African taxa. Fossil pollen also fitting this type has been reported from the mid-Miocene in western Anatolia (Bouchal et al., 2016) and would corroborate the phylogenetic results of McGuire & Kron (2005) and Pirie et al. (2016). A slightly younger member related to this ancestor lineage has been encountered as fossil diaspores in coal-bearing upper Miocene sediments of the lower Rhenish basin (Germany) and named Erica palaeoarborea Van der Burgh (Van der Burgh, 1987). The second Erica-type, from the lower Bartonian Borken coal, resembles typical European Palaearctic ‘heather species’, such as E. carnea and E. herbacea, and is interpreted as representing a precursor of the Erica spp. living today in northern Europe. Extant Kalmia of tribe Phyllodoceae is a morphologically diverse genus comprising small sub-shrubs to small trees and has a generally North American distribution. The genus includes ten species, including one in Cuba and one inhabiting the northernmost parts of Eurasia; all thrive in swamps, peat bogs and on poor soils (Gillespie & Kron, 2010). The distribution of extant Phyllodoceae from the Pacific Northwest to eastern North America via northern Europe (plus Caucasus) to eastern Asia is interpreted to represent a biogeographical pattern consistent with the Arcto-Tertiary flora (Gillespie & Kron, 2010). The fossil record of Kalmia is also meagre and up to now only known from central Europe: fossil leaves of Kalmia saxonica Litke (Litke, 1966) and Kalmiophyllum Kr. & Weyl. of early to mid-Miocene age were reported from Lusatian coal seams in Germany (see Mai, 1995: 182). The fossil leaves, together with the mid-Eocene Kalmia-type tetrad described here, prove that the genus both had a much wider geographical distribution in the geological past (c. 40–15 Mya) than today and probably migrated between Europe and North America via North Atlantic land bridges (re-evaluated by Brikiatis, 2014). Rhododendron is an even larger genus than Erica with > 1000 species and centres of diversity in Asia, such as the Himalaya and Malaysia (Gao et al., 2002). Molecular phylogenetic studies, using various methods, revealed that Rhododendron is monophyletic (Gao et al., 2002; Goetsch, Eckert & Hall, 2005). Fossil seeds described as Rhododendron newburyanum Collinson & Crane (Collinson & Crane, 1978) were described from the lower Eocene Reading Formation (England). The Rhododendron-type pollen from Borken described here presumably represented an Asian Rhododendron lineage and together with the various other Rhododendron pollen depicted in Zetter & Hesse (1996) demonstrate that rhododendrons in general were well established in Europe during the mid-Eocene. Vaccinieae are a large and morphologically diverse tribe of woody plants with members widely distributed mostly in tropical zones but also in temperate montane areas of South America and East Malaysia (Kron et al., 1999, 2002). Kron et al. (1999) reported that the tribe is monophyletic, Vaccinium being the largest genus with > 400 species. These are generally distributed in the Northern Hemisphere, but there are also many species in the Old and New World tropics that are assumed to have diversified in several independent clades (Kron et al., 2002). Until now, the only fossil evidence for this genus has been diaspores described as Vaccinium miocenicum Van der Burgh (Van der Burgh, 1987), found in the upper Miocene sediments of Germany. The Vaccinium-type pollen described here from Cobham (PETM) is much older than Vaccinium miocenicum and resembles Vaccinium taxa from sections Conchophyllum and Bracteata and V. oldhamii (section Ciliata), which all live today in temperate continental Asia and are members of the closely related Agapetus and Bracteata-Oarianthe clades erected by Kron et al. (2002). The Vaccinium-type from Cobham is therefore interpreted to represent an ancient lineage of Vaccinieae that, if one considers the North American plus Eurasian origin of Ericaceae (Kron & Luteyn, 2005), was probably a Eurasian lineage that either migrated towards the east (temperate continental Asia) or came to Europe before the Eocene from the east. A literature search did not yield any pre-Eocene Ericaceae fossils from the Russian Federation; the earliest occurrence of Ericaceae is a mid-Eocene leaf fossil assigned to Leucothoë D.Don from the southern Ural mountains (Akhmetiev & Beniamovski, 2006; Akhmetiev, 2007; and the present author’s unpublished data from various Eurasian localities). Gaulterieae are a tribe of three genera, Gaulteria spp. being low growing woody half-shrubs distributed from temperate to tropical regions in the Americas, East to South-East Asia and Australasia (Fritsch et al., 2011). Molecular phylogenetic studies indicated that Gaulterieae are monophyletic (Bush et al., 2009) and frequently underwent reticulate evolution (Fritsch et al., 2011). The Gaultheria-type from Brixton (PETM) shows affinities with several Gautheria spp. that are today distributed in South America. In contrast to the fossil Vaccinium-type from Cobham, the Gaultheria-type from Brixton represents an ancient lineage that, if one considers the North American plus Eurasian origin of Ericaceae (Kron & Luteyn, 2005), either later migrated to the west (North America) or came from there before the Eocene to Europe. In either case, migration via the North Atlantic Land Bridges, as re-defined by Brikiatis (2014), would have been possible. Pre-Eocene fossil pollen tetrads of Ericaceae (or Gaultherieae) from North America are not known and thus a Eurasian origin is most likely. Sapotaceae Sapotaceae are a pantropical to pan-subtropical family of five morphologically defined tribes comprising trees and shrubs, with centres of diversity in Central and South America and southern Asia (Pennington, 1991). However, molecular phylogenetic investigations of Sapotaceae changed some of the characteristics used to define the tribes and their genera (Anderberg & Swenson, 2003; Swenson & Anderberg, 2005; Smedmark, Swenson & Anderberg, 2006; Armstrong et al., 2014; Richardson et al., 2014). The oldest Sapotaceae fossils are pollen from Upper Cretaceous sediments of Australia (Sapotaceoidaepollenites rotundus Harris; Harris, 1972 LM image; Stoian, 2002; the latter without images) that possibly could belong to Mimusopeae (according to Harley in Armstrong, 2010: chapter 6, appendix 6.1) and a Santonian/Campanian Sarcosperma-like Sapotaceae at Ajka (western Hungary, unpublished LM and SEM images by R. Zetter). These early finds are then followed by occurrences of various fossil Sapotaceae pollen grains from around the Palaeocene–Eocene boundary and during Eocene times worldwide (e.g. China, India, New Zealand and Australia, North Atlantic sediments and the USA; see compilation in Armstrong, 2010); unfortunately, these are rarely accompanied by reliable botanical affiliations because the identifications were made based on LM only, and therefore should be treated with caution. Reliable identifications of fossil Sapotaceae pollen are represented by the mid-Eocene Tieghemella heckelii Pierre ex A.Chev.-like pollen (African Mimusopeae) from the Isle of Wight (England) by Harley (1991; SEM investigation) and mid-Eocene Mimusops- and Madhuca?-like pollen (Taylor, 1989; SEM images) from the USA. Subfamily Sarcospermatoideae sensuSwenson & Anderberg (2005) comprise one genus, Sarcosperma, a genus of relatively small deciduous trees. This genus occurs from the eastern Himalaya through Malaysia to Sumatra (Pennington, 1991) and is potentially one of the earliest diverging Sapotaceae, because phylogenetic studies (Anderberg & Swenson, 2003; Swenson & Anderberg, 2005) indicate that Sarcosperma is sister to all other Sapotaceae and had split from the rest of the Sapotaceae > 98.8 Mya (Smedmark & Anderberg, 2007: table 2). The Sarcosperma-type pollen from Ajka (R. Zetter, unpublished LM and SEM images) proves that Sarcosperma existed during the late Cretaceous in Europe, and the Sarcosperma-type from Brixton significantly post-dates this old age. The two Sarcosperma-types, from Brixton and Krappfeld, described here resemble extant Asian mainland taxa and, together with the similar-looking Late Cretaceous Sarcosperma-type from Ajka, suggest a boreotropical origin of Sarcospermatoideae. The Chrysophylleae-type pollen grains documented in this paper partly corroborate the historical biogeographical reconstruction of the pantropical distribution of Chrysophylloideae, based on molecular phylogenetic studies, by Bartish et al. (2011). They postulated that the Chrysophylloideae evolved in Africa (c. 73–82 Mya) and split into South American and African lineages c. 62–72 Mya. Although both lineages dispersed afterwards, the African lineage further divided into Australasian and African lineages. However, with the lack of reliable pre-Eocene Sapotaceae fossils from Africa, the probable origin of Chrysophylleae in Africa remains a hypothesis. The different origins of the extant Chrysophylleae taxa that the fossil taxa of this paper resemble are remarkable: of the four fossil Chrysophylleae-types (two from the PETM and two from the EECO), three show clear affinities to South American species, and one (Chrysophylleae-Brixton type) resembles several extant species from Australasia and New Caledonia; the latter group forms a clade (Swenson et al., 2008: fig. 3). This means that during the PETM in England, Chrysophylleae had already split into a South American ‘Pouteria/Elaeoluma’ lineage and a ‘Sersalia/Planchonella’ lineage, which is now recognized as a clade in Australasia (Swenson et al., 2007). However, the presence of the four Chrysophylleae-type pollen during the lower Eocene in Europe neither refutes nor supports the hypothesis of Bartish et al. (2011), in which the South American Chrysophylloideae migrated from Africa via long-distance dispersal over the proto-South Atlantic across island chains (e.g. the Rio Grande Rise-Walvis Ridge) and not across the North Atlantic Land Bridge. The fossil data presented here show only that the members of Chrysophylleae were present in southern England c. 56 Mya and presumably went extinct during the following cooling events. No fossil proof for further migration towards the north-west, towards North America, has been found, although it would have been possible in that time frame, via the Thulean Route (Brikiatis, 2014). The three fossil pollen taxa that cannot be clearly separated into either Mimusopeae or Isonandreae, the four Isonandreae and two of the three Mimusopeae have all their ‘living equivalents’ in the Old World Tropics (e.g. India, Malaysia, Mauritius, Africa). Isonandreae today are an Old World tropical tribe, as are most extant Mimusopeae (inhabiting Africa, Madagascar the Mascarenes and the Seychelles), except Manilkara, which is also present in South America (see Armstrong, 2010). Molecular phylogenetic studies (Anderberg & Swenson, 2003; Swenson & Anderberg, 2005; Smedmark et al., 2007) on Sapotaceae demonstrated that Sideroxyleae, Mimusopeae and Isonandreae form a large clade. Furthermore, these studies suggested that after Sideroxyleae split off, the age of the Isonandreae/Mimusopeae node might be mid-Eocene. It is also suggested that this large clade originated in Africa (Anderberg & Swenson, 2003; Swenson & Anderberg, 2005; Smedmark et al., 2007); comparable ages (58–48 Mya) for the Mimusopeae/Isonandreae clade (including Inhambanella Dubard) were also proposed by Armstrong et al. (2014). Therefore, the various fossil pollen types described here might represent members of ancestral lineages that were already diversified: some of them (early diverging members of one or the other tribe) still had characteristics of both tribes during early Bartonian times, whereas others were already clearly separated into either Isonandreae or Mimusopeae during the Ypresian (see discussions below). The two Mimusopeae-types of early Ypresian age (Mimusops/Vitellariopsis affiliation) look similar to African and Madagascan taxa, and their presence in Austria certainly pre-dates the much younger node ages for Mimusops (28–17 Mya), and, at the moment, does not corroborate the probable origin of the genus in Africa, as postulated by Armstrong et al. (2014; table 1: node ages and ancestral areas of Sapotaceae). The same is shown for the fossil Mimusopeae-type from Borken of early Bartonian age (c. 40 Mya) that shows affinities to South American Manilkara: It also pre-dates the postulated node age (32–28 Mya) and does not necessarily support an African origin of Manikara, proposed by Armstrong et al. (2014: table 1). The authors do show that these calculated ages should be treated cautiously as they represent approximations that might change with time: if one compares proposed ages from Armstrong et al. (2014) and Richardson et al. (2014, partly the same authors as in Armstrong et al., 2014) the dates differ considerably because different data sets, methods and the quality and weighting of fossils for calibration were used. These uncertainties in the reconstruction of patterns of diversification and migration can only be overcome by more reliable fossil data, particularly the routine application of combined LM and SEM in palaeo-palynology. Because unequivocal pre-Eocene fossil Sapotaceae from Africa, Madagascar, India and South-East Asia are lacking, reconstructing their origins and migration paths remains a treacherous process. Today, Isonandreae have a geographical range from India towards the east and south-east (including the Pacific islands) to Australia, and Palaquium has a Malesian distribution (Pennington, 1991). The fossil occurrence of Isonandreae pollen (affinities to Palaquium leiocarpum, P. philippense and Madhuca sp.) from the PETM to the mid-Eocene in Europe does, to a certain extent, corroborate the stem and crown ages of Isonandreae (40.5 and 36.5 Mya, respectively) suggested by Richardson et al. (2014: table 1). However, these authors also postulated Africa as a most probable area of origin for stem Isonandreae and Sundania (Thailand, Indonesia, Borneo, Philippines, Java) for crown Isonandreae (e.g. Palaquium spp.). In contrast, the fossil data presented here show that pollen similar to Isonandreae crown taxa were already present in Europe 56 Mya and that they closely resemble East Asian Palaquium and Madhuca taxa from Borneo and the Philippines. There are no reliable pre-Eocene and Eocene Isonandreae fossils from Africa and South-East Asia (see discussion above, Harley 1991, and the author’s material from Hainan) that could indicate where stem Isonandreae came from. Sideroxyleae comprise trees and shrubs, with three genera at present that display a wide ecological range: their members inhabit arid areas and rainforests and have a pantropical distribution (Smedmark & Anderberg, 2007). Smedmark & Anderberg (2007) suggested that the extant groups of Sideroxyleae rapidly diversified between 65 and 34 Mya into different lineages, when a boreotropical flora was maintained by temperature maxima, climate optima and existing high-latitude land bridges between Eurasia and North America. The authors also postulated that Sideroxyleae entered the Americas via ‘North Atlantic routes’, as pointed out by Morley (2003) for the entire family Sapotaceae. On the other hand, the opposite scenario with America as the origin for the Sideroxyleae, but similar ages for diversification, was proposed by Stride, Nylinder & Swenson (2014). The 56-My-old Sideroxylon-type from Brixton closely resembles the south-eastern North American Sideroxylon reclinatum, but this does not resolve the question of where the Sideroxyleae lineage originated: Did the Sideroxyleae migrate from the Old World boreotropics via the re-evaluated Thulean Route of Brikiatis (2014) westwards to North America and diversify there later, between 53.2 and 51.4 Mya, as suggested by Smedmark & Anderberg (2007: table 3), or was it the other way round, from Central America to Africa with long-distance dispersal, as suggested by Stride et al. (2014)? The age of the Sideroxylon-type from Brixton would seem to corroborate the former hypothesis, but without reliable fossil data of pre-Eocene age from the Americas and Africa there is no proof for either model. Ebenaceae Ebenaceae are pantropical and comprise four genera. Molecular phylogenetic investigations demonstrate the monophyly of the family and the presence of two subfamilies: Lissocarpoideae (only Lissocarpa Benth., solely South American) and Ebenoideae (Diospyros, Euclea and Royena L.) which are generally African (Duangjai et al., 2006). Ebenaceae are mostly trees, sometimes shrubs, with highest diversity in Asia and the Indo-Pacific area, but morphological diversity is highest in Madagascar and Africa (Heywood, 1993; Wallnhöfer, 2001). The largest genus is Diospyros, a warm temperate to pantropical/subtropical genus comprising > 500 species (> 700 in The Plant List) with a wide ecological amplitude (Duangjai et al., 2006). The fossil record is so far restricted to pollen (Tricolporopollenites milonii Ollivier-Pierre) affiliated to Diospyros from the Amorican Massif (France) of lower Eocene age (cited in Muller, 1981), much younger Diospyros pollen from the late Eocene of the Florissant Formation in Colorado (Bouchal et al., 2016: figs 14F, 15I, J), diaspores of Diospyrocarpum senescens (Crié) Vaudois from the mid-Eocene of France (Vaudois-Miéja, 1980) and to numerous mummified leaves and flowers of Austrodiospyros cryptostoma Basinger & Christophel with associated ‘cf. Diospyros’ pollen from late mid-Eocene sediments of southern Australia (Basinger & Christophel, 1985; Christophel, Harris & Syber, 1987). The Diospyros from the Florissant Formation looks quite similar to the Diospyros-type sp. 2 from Hainan. The three new fossil pollen taxa described here resemble extant Diospyros spp. (and one to some extent Euclea) from subtropical and tropical areas of Africa, which is remarkable because two of the fossil pollen taxa are from the Changchang Formation of Hainan (southern China, presumably Lutetian to early Bartonian age). The oldest Diospyros/Euclea-type pollen (PETM, Brixton) is relatively large (polar axis 38 µm). Pollen grains of Euclea resemble those of Diospyros, but are generally much smaller than the normal size range for Diospyros (Morton & Kincaid, 1995: fig. 7; Geeraerts et al., 2009: table 1). Consequently, it can be suggested that 56 Mya, the lineage consisting of Diospyros and Euclea (and Royena) had already started to diversify. Euclea and Royenia remained in Africa, but members of Diospyros migrated to Asia, Australia and the Americas, with the result that around the mid-Eocene Diospyros was present in Europe (Vaudois-Miéja 1980), southern Australia (c. 39 Mya; Christophel et al., 1987: fig. 5) and with at least two types on Hainan, and at the end of the Eocene in the USA, but these latter three pollen types still resembled extant African Diospyros taxa. The new fossil data on Diospyros only slightly pre-date the age estimates for the split between Diospyros and Euclea plus Royena given by Turner et al. (2013) and probably corroborate a Gondwana–Africa origin. Styracaceae Styracaceae are a family of woody plants distributed across the Northern Hemisphere inhabiting warm temperate to tropical (only Styrax) climates and are assumed to be monophyletic (Fritsch, 2001; Fritsch et al., 2001). The largest genus, Styrax, is disjunctly distributed in southern parts of the USA, South America, eastern Asia and the Mediterranean region (Fritsch, 2001, Fritsch et al. 2001) and, with the Asian genus Huodendron, forms a clade (Fritsch et al., 2001). The earliest fossil occurrences of Styracaceae are seeds: Rehderodendron stonei Reid & Chandler has been found in mid-Eocene deposits at Grès Sabals, in France (Vaudois-Miéja, 1983) and Styrax spp. has been found from the upper Eocene of England (Chandler, 1925–26). Therefore, the occurrence of an older Ypresian Styrax-type pollen from Krappfeld is not completely anomalous. The sculpture and ornamentation of the tectum of the fossil Styrax-type pollen show many resemblances to the ornamentation of the considerably larger pollen of extant South American and eastern Asian Styrax species (and to a much lesser degree to Rehderodendron and Sinojackia Hu, both Asian taxa) and presumably represent an early or primitive characteristic of Styrax pollen. A similar observation was made by Fritsch (2001), who re-evaluated the above-mentioned fossil seeds, and suggested that the biogeographical origin of Styrax might be European/Eurasian or at least from the Northern Hemisphere and that the genus might represent a boreotropical element that migrated during the Eocene to North America via North Atlantic land bridges. Therefore, it can be suggested that the Styrax-type from Krappfeld might be close to the European ancestor of the Styrax/Huodendron clade sensuFritsch et al. (2001). Theaceae The monophyletic Camellia L. is the largest genus of Theaceae with 200–300 species distributed in warm temperate to tropical China and eastern to south-eastern Asia (Tsou, 1997; Vijayan et al., 2009). The fossil record of Theaceae starts at the end of the Cretaceous: Kvaček and Walther (1984) described Eurya crassitesta Knobloch, a fossil diaspore from the Maastrichtian of Austria. Numerous early to late Eocene diaspores from England and Germany were re-investigated and affiliated to Eurya by Mai (1971), followed by an investigation and description of numerous fossil diaspores from the mid-Eocene Claiborne Formation (USA) by Grote & Dilcher (1989, 1992) and their affiliation with the extant US genus Gordonia J.Ellis and Theaceae gen. indet. Fossil leaves and wood affiliated with Camellia are generally known since the upper Eocene to the Miocene from Japan, Taiwan, Korea and China (summarized in Huang et al., 2016) and therefore the Ypresian Camellia-type from Krappfeld represents the oldest record of a Camellia-lineage to date. CONCLUSIONS Twenty-nine pollen types of Ericales from five localities 55.8 to c. 41.2 My old have been described and identified down to tribe or genus level, using LM and SEM: seven members of Ericaceae, 17 Sapotaceae, three Ebenaceae, one Styracaceae and one Theaceae. Of the two fossil Erica-type taxa present, the older one, from the Cobham Lignite (PETM), resembles extant E. arborea and might be close to the ancestral lineage that later led to the evolution and migration of the African taxa. The younger Erica-type, from the lower Bartonian Borken coal, is reminiscent of European Palaearctic ‘heather species’ and might represent one of the precursors of Erica taxa living today in northern Europe. The earliest occurrence so far of Kalmia-type tetrads (early Bartonian) from Borken demonstrates a probable origin and much wider geographical distribution of this genus in the geological past, whereas Rhododendron-type taxa of the same age witness, with other fossil specimens, the establishment of various Rhododendron taxa during the mid-Eocene in Europe. The Vaccinium-type taxon from Cobham (PETM), which is the oldest occurrence of the genus Vaccinium, displays similarities to Vaccinium species of temperate continental Asia today and is interpreted to represent an ancient Eurasian lineage of Vaccinieae. The Gaultheria-type from the PETM at Brixton shows affinities to several Gautheria spp. that are now distributed in South America, but a Eurasian origin is most likely. This represents the oldest occurrence of the genus Gaultheria to date. Since the beginning of the Eocene to the mid-Eocene, all three subfamilies of Sapotaceae (Sarcospermatoideae, Chrysophylloideae Sapotoideae) and their tribes were already present in Europe (England, Austria and Germany): Sarcospermatoideae with two Sarcosperma-type taxa; Chrysophylloideae with three Chrysophylleae-type taxa allied with South American Elaeoluma/Pouteria, and one Asian Chrysophylleae-type taxon allied with Planchonella/Sersalisia; and Sapotoideae with three Mimusopeae/Isonandreae-type taxa with affinities to species of Mimusops, Madhuca, Palaquium and Labourdonnaisia Bojer, two Mimusopeae-type taxa showing affinities with African/Malegasy Mimusops/Vitellariopsis taxa, one Mimusopeae-type taxon showing affinity with South American Manilkara, three Isonandreae–type taxa showing affinities with South east Asian Palaquium and one Isonandreae-type taxaon showing affinities with Palaquium and Madhuca spp.; and one Sideroxylon-type showing affinities with South American species. However, reliable pre-Eocene Sapotaceae fossils from Africa, the Americas and Asia are not known to date, and because the oldest sapotaceous pollen fossils generally come from Europe, the origin of Sapotaceae could be assumed to lie in the boreotropics rather than in Africa/Gonwana; this is likely to be revealed in future palynological work on these under-investigated areas. All three fossil Ebenaceae pollen taxa (England and Hainan) resemble extant subtropical and tropical African Diospyros spp. (and one also to a certain extent Euclea) and probably corroborate a Gondwana–Africa origin for the genus (but see above). The Diospyros-type from Brixton (PETM) represents the oldest occurrence of Diospyros so far recognized. The Styrax-type taxon from Krappfeld shows many resemblances to extant South American and east Asian Styrax taxa and presumably represents an early type of Styrax pollen. It can be assumed that the Styrax-type might be close to the European ancestor of the Styrax/Huodendron clade sensuFritsch et al. (2001). The Ypresian Camellia-type taxon from Krappfeld represents the oldest record of the Camellia-lineage so far recognized. ACKNOWLEDGEMENTS The author thanks M. E. Collinson, H.-J. Gregor, Tania M. Kodrul and Jianhua Jin for donating samples, Fridgeir Grimsson and Reinhard Zetter for discussions, Kai-Uwe Hochhauser for computer support and A. Hugh N. Rice for correcting the English text. The reviews by James E. Doyle and Johannes M. Bouchal improved the manuscript greatly and are warmly acknowledged. REFERENCES Akhmetiev MA . 2007 . Paleocene and Eocene floras of Russia and adjacent regions: climatic conditions of their development . Paleontological Journal 41 : 1032 – 1039 . Google Scholar CrossRef Search ADS Akhmetiev MA , Beniamovski VN . 2006 . The Paleogene and Eocene in the Russian part of west Eurasia . Stratigraphy and Geological Correlation 14 : 49 – 72 . Google Scholar CrossRef Search ADS Anderberg A , Swenson U . 2003 . Evolutionary lineages in Sapotaceae (Ericales): a cladistic analysis based on ndhF sequence data . International Journal of Plant Sciences 164 : 763 – 773 . Google Scholar CrossRef Search ADS APG IV . 2016 . An update of the Angiosperm Phylogeny Group classification for the orders and families of flowering plant: APG IV . Botanical Journal of the Linnean Society 181 : 1 – 20 . CrossRef Search ADS Archangelsky A , Zamaloa María del C . 2003 . Primeros resultados palinológicos del Paleógeno del sector oriental de la Sierra La Colonia, provincia del Chubut, Argentina . Revista Museo Argentino Ciencas Naturale 5 : 119 – 123 . Google Scholar CrossRef Search ADS Armstrong K . 2010 . Systematics and biogeography of the pantropical genus Manilkara Adans. (Sapotaceae) . PhD Thesis, Institute of Evolutionary Biology, School of Biological Sciences of the University Edinburgh and Royal Botanic Garden Edinburgh . Armstrong KE , Stone GN , Nicholls JA , Valderrama E , Anderberg AA , Smedmark J , Gautier L , Naciri Y , Milne R , Richardson JE . 2014 . Patterns of diversification amongst tropical regions compared: a case study in Sapotaceae . Frontiers in Genetics 5 : 362 . Google Scholar CrossRef Search ADS PubMed Askin RA . 1990 . Campanian to Paleocene spore and pollen assemblages of Seymour Island, Antarctica . Review of Palaeobotany and Palynology 65 : 105 – 113 . Google Scholar CrossRef Search ADS Bartish IV , Antonelli A , Richardson JE . 2011 . Vicariance or long distance dispersal: historical biogeography of the pantropical subfamily Chrysophylloideae (Sapotaceae) . Journal of Biogeography 38 : 177 – 190 . Google Scholar CrossRef Search ADS Basinger JF , Christophel DC . 1985 . Fossil flowers of the Ebenaceae from the Eocene of southern Australia . Canadian Journal of Botany 63 : 1825 – 1843 . Google Scholar CrossRef Search ADS Bouchal JM , Mayda S , Zetter R , Grimsson F , Akgün F , Denk T . 2017 . Miocene palynofloras of the Tinaz lignite mine, Mugla, southwest Anatolia: taxonomy, palaeoecology and local vegetation change . Review of Palaeobotany and Palynology 243 : 1 – 36 . Google Scholar CrossRef Search ADS Bouchal JM , Zetter R , Grimsson F , Denk T . 2016 . The middle Miocene palynoflora and palaeoenvironments of Eskihisar (Yatağan basin, south-western Anatolia): a combined LM and SEM investigation . Botanical Journal of the Linnean Society 182 : 14 – 79 . Google Scholar CrossRef Search ADS Bush CM , Lu L , Fritsch PW , Li D-Z , Kron KA . 2009 . Phylogeny of Gaultherieae (Ericaceae; Vaccinioideae) based on DNY sequence data from matK, ndhF, and nrITS . International Journal of Plant Sciences 170 : 355 – 364 . Google Scholar CrossRef Search ADS Brikiatis L . 2014 . The De Geer, Thulean and Beringia routes: key concepts for understanding early Cenozoic biogeography . Journal of Biogeography 41 : 1036 – 1054 . Google Scholar CrossRef Search ADS Bremer B , Bremer K , Heidari N , Erixon P , Olmstead RG , Anderberg AA , Källersjö M , Barkhordarian E . 2002 . Phylogenetics of asterids based on 3 coding and 3 non-coding chloroplast DNA markers and the utility of non-coding DNA at higher taxonomic levels . Molecular Phylogenetics and Evolution 24 : 274 – 301 . Google Scholar CrossRef Search ADS PubMed Carpenter RJ , McPhail MK , Jordan GJ , Hill RS . 2015 . Fossil evidence for open, Proteaceae-dominated heathland and fire in the Late Cretaceous . American Journal of Botany 102 : 1 – 16 . Google Scholar CrossRef Search ADS Chandler ME. 1925 – 26 . The Upper Eocene Flora of Hordle, Hants . London : Palaeontographical Society Monographs Part 1 & Part 2. Christophel DC , Harris WK , Syber AK . 1987 . The Eocene flora of the Anglesea locality, Victoria . Alcheringia 11 : 303 – 323 . Google Scholar CrossRef Search ADS Collinson ME , Boulter MC , Holmes PL . 1993 . Magnoliophyta (Angiospermae) . In Benton MJ , ed. The fossil record . London : Chapman & Hall , 809 – 841 . Collinson ME , Crane PR . 1978 . Rhododendron seeds from the Palaeocene of southern England . Botanical Journal of the Linnean Society 76 : 195 – 205 . Google Scholar CrossRef Search ADS Collinson ME , Steart DC , Harrington GJ , Hooker JJ , Scott AC , Allen LO , Glasspool IJ , Gibbons SJ . 2009 . Palynological evidence of vegetation dynamics in response to palaeoenvironmental change across the onset of the Palaeogene-Eocene Thermal Maximum at Cobham, Southern England . Grana 48 : 38 – 66 . Google Scholar CrossRef Search ADS Crepet WL , Nixon KC , Gandolfo MA . 2004 . Fossil evidence and phylogeny: the age of major angiosperm clades based on mesofossil evidence from Cretaceous deposits . American Journal of Botany 91 : 1666 – 1682 . Google Scholar CrossRef Search ADS PubMed Crepet WL , Nixon KC , Daghlian CP . 2013 . Fossil Ericales from the upper Cretaceous of New Jersey . International Journal of Plant Sciences 174 : 572 – 584 . Google Scholar CrossRef Search ADS Doyle JA , Van Campo M , Lugardon B . 1975 . Observations on exine structure of Eucommiidites and Lower Cretaceous angiosperm pollen . Pollen et Spores 17 : 429 – 486 Duangjai S , Wallnöfer B , Samuel R , Munzinger J , Chase MW . 2006 . Generic delimitation and relationships in Ebenaceae sensu lato: evidence from six plastid DNA regions . American Journal of Botany 93 : 1808 – 1827 . Google Scholar CrossRef Search ADS PubMed Ellison RA , King C . 2004 . Chapter 4. Palaeogene: Paleocene . In: Ellison RA , Woods A , Allen DJ , Forster A , Pharoah TC , King C , eds. Geology of London . Keyworth : British Geological Survey , 22 – 43 . Fendt A . 1996 . Beiträge zur Palynologie rezenter und fossiler Theaceae . Unpublished Master’s Thesis, University Vienna . Foss PJ , Doyle GJ . 1988 . A palynological study of Irish Ericaceae . Pollen et Spores 30 : 151 – 177 . Friis EM . 1985 . Actinocalyx gen. nov., sympetalous angiosperm flowers from the Upper Cretaceous of southern Sweden . Review of Palaeobotany and Palynology 45 : 171 – 183 . Google Scholar CrossRef Search ADS Fritsch PW . 2001 . Phylogeny and biogeography of the flowering plant genus Styrax (Styracaceae) based on chloroplast DNA restriction sites and DNA sequences of the internal transcribed spacer region . Molecular Phylogenetics and Evolution 19 : 387 – 408 . Google Scholar CrossRef Search ADS PubMed Fritsch PW , Lu L , Bush CM , Cruz BC , Kron KA , Li D-Z . 2011 . Phylogenetic analysis of the wintergreen group (Ericaceae) based on six genic regions . Systematic Botany 36 : 990 – 1003 . Google Scholar CrossRef Search ADS Fritsch PW , Morton CM , Chen T , Meldrum C . 2001 . Phylogeny and biogeography of the Styraceae . International Journal of Plant Sciences 162 : S65 – S116 . Google Scholar CrossRef Search ADS Gao L , Li D-Z , Zhang C-Q , Yang J-B . 2002 . Infrageneric and sectional relationships in the genus Rhododendron (Ericaceae) inferred from ITS sequence data . Acta Botanica Sinica 44 : 1351 – 1356 . Geeraerts A , Raeymaekers JAM , Vickier S , Pletsers A , Smets E , Huysmans S . 2009 . Systematic palynology in Ebenaceae with focus on Ebenoideae: morphological diversity and character evolution . Review of Palaeobotany and Palynology 153 : 336 – 353 . Google Scholar CrossRef Search ADS Gillespie E , Kron K . 2010 . Molecular phylogenetic relationships and a revised classification of the subfamily Ericoideae (Ericaceae) . Molecular Phylogenetics and Evolution 56 : 343 – 354 . Google Scholar CrossRef Search ADS PubMed Goetsch L , Eckert AJ , Hall BD . 2005 . The molecular systematics of Rhododendron (Ericaceae): a phylogeny based upon RPB2 gene sequences . Systematic Botany 30 : 616 – 626 . Google Scholar CrossRef Search ADS Gosling WD , Miller CS , Livingstone DA . 2013 . Atlas of the tropical West African flora . Review of Palaeobotany and Palynology 199 : 1 – 135 . Google Scholar CrossRef Search ADS Gradstein FM , Ogg JG , Schmitz M , Ogg G . 2012 . A geologic time scale . Amsterdam : Elsevier . Gregor H-J . 2005 . Pflanzen und Tiere aus den eozänen Braunkohlen des Untertagebaues Stolzenbach bei Kassel . Philippia 12 : 147 – 181 . Grímsson F , Zetter R , Grimm G , Pedersen GK , Pedersen A , Denk T . 2015 . Fagaceae pollen from the early Cenozoic of West Greenland: revisiting Engler’s and Chaney’s Arcto-Tertiary hypothesis . Plant Systematics and Evolution 301 : 809 – 832 . Google Scholar CrossRef Search ADS PubMed Grote PJ , Dilcher DL . 1989 . Investigations of angiosperms from the Eocene of North America: a new genus of Theaceae based on fruit and seed remains . Botanical Gazette 150 : 190 – 206 . Google Scholar CrossRef Search ADS Grote PJ , Dilcher DL . 1992 . Fruits and seeds of tribe Gordonieae (Theaceae) from the Eocene of North America . American Journal of Botany 79 : 744 – 753 . Google Scholar CrossRef Search ADS Gruas-Cavagnetto C . 1976 . Etude palynologique du Paleogène du Sud Angleterre . Cahiers Micropaléontologie 1 : 1 – 49 . Gruas-Cavagnetto C . 1978 . Etude palynologique de l Èocène du Bassin Anglo-Parisien . Memoir Societé Géologique France, Nouveau Séries 56 Memoire 131 : 1 – 64 . Harris WK . 1972 . New form species from southern Australia early Tertiary sediments . Transactions of the Royal Society of Australia 96 : 56 – 61 . Halbritter H . 2016a . Erica arborea . In Paldat (2016-11-11) – a palynological database. Available at: https://www.paldat.org/pub/Erica_arborea/302157 Halbritter H . 2016b . Erica carnea . In Paldat (2016-11-25) – a palynological database. Available at: https://www.paldat.org/pub/Erica_carnea/302417 Halbritter H . 2016c . Erica herbacea . In Paldat (2016-11-25) – a palynological database. Available at: https://www.paldat.org/pub/Erica_carnea/302418 Halbritter H . 2016d . Rhododendron hirsutum . In Paldat (2016-12-04) – a palynological database. Available at: https://www.paldat.org/pub/Rhododendron_hirsutum/302695 Harley MM . 1991 . The pollen morphology of the Sapotaceae . Kew Bulletin 46 : 379 – 491 . Google Scholar CrossRef Search ADS Hesse M , Halbritter H , Zetter R , Weber M , Buchner R , Frosch-Radivo A , Ulrich S . 2009 . Pollen terminology – an illustrated handbook . Vienna : Springer . Heywood VH . 1993 . Flowering plants of the world . London : BT Batsford . Hofmann CC , Egger H , King C . 2015a . LM and SEM investigations of pollen from the PETM and EECO localities of Austria and Great Britain: new findings of Atherospermataceae, Annonaceae, Araceae, and Arecaceae from the lower Eocene . Plant Systematics and Evolution 301 : 773 – 793 . Google Scholar CrossRef Search ADS Hofmann CC , Egger H , King C . 2015b . SEM investigation of pollen from the lower Eocene (Carinthia and Salzburg in Austria and Brixton, London area, in England): new findings of Vitaceae, Euphorbiaceae, Phyllanthaceae, Fabaceae, Anacardiaceae, Araliaceae and Apiaceae . Plant Systematics and Evolution 301 : 2291 – 2312 . Google Scholar CrossRef Search ADS Hofmann CC , Gregor HJ . 2018 . Scanning electron microscope investigations of pollen from an atypical mid-Eocene coal facies in Stolzenbach mine (PreußenElektra) near Borken (Kassel, Lower Hesse, Germany) . Review of Palaeobotany and Palynology 252 : 41 – 63 . Google Scholar CrossRef Search ADS Hofmann CC , Mohamed O , Egger H . 2011 . A new terrestrial palynoflora from the Palaeocene/Eocene boundary in the northwestern Tethyan realm (St. Pankraz, Austria) . Review of Palaeobotany and Palynology 166 : 295 – 310 . Google Scholar CrossRef Search ADS Hofmann CC , Pancost R , Ottner F , Egger H , Taylor K , Zetter R . 2012 . Palynology, biomarker and clay mineralogy of the Early Eocene Climate Optimum (EECO) in the transgressive Krappfeld succession (eastern Alps, Austria) . Austrian Journal of Earth Sciences 105 : 224 – 239 . Hofmann CC , Zetter R . 2001 . Palynological investigations of the Krappfeld area, Palaeocene/Eocene, Carinthia (Austria) . Palaeontographica B 259 : 47 – 64 . Hottenrott M , Gregor H-J , Oschkinis V . 2010 . Die eozänen Braunkohleschichten aus dem Untertagebau Stolzenbach bei Kassel (PreußenElektra, Niederhessen) VII Die Mikroflora . Documenta naturae 181 : 29 – 43 . Huang LL , Jin JH , Quan C , Oskolski AA . 2016 . Camellia nanningensis sp. nov.: the earliest fossil wood record of the genus Camellia (Theaceae) from East Asia . Journal of Plant Research 129 : 823 – 831 . Google Scholar CrossRef Search ADS PubMed Kedves M . 1969 . Palynological studies on Hungarian early Tertiary deposits . Budapest : Publishing House of the Hungarian Academy of Sciences . Kron KA , Fuller R , Crayn DM , Gadek PA , Quinn CJ . 1999 . Phylogenetic relationships of epacrids and vaccinioids (Ericaceae s.l.) based on matK sequence data . Plant Systematics and Evolution 218 : 55 – 65 . Google Scholar CrossRef Search ADS Kron KA , Luteyn JE . 2005 . Origins and biogeographic patterns in Ericaceae: new insights from recent phylogenetic analyses . Kongelige Danske Videnskabernes Selskab Biologiske Skrifter 55 : 479 – 500 . Kron KA , Powell EA . 2009 . Molecular systematics of Rhododendron subgenus Tsutsusi (Rhodoreae, Ericoideae, Ericaceae) . Edinbourgh Journal of Botany 66 : 81 – 95 . Google Scholar CrossRef Search ADS Kron KA , Powell EA , Luteyn JL . 2002 . Phylogenetic relationships within the blueberry tribe (Vaccinieae, Ericaceae) based on sequence data from matK and nuclear ribosomal ITS regions, with comments on the placement of Satyria . American Journal of Botany 89 : 327 – 336 . Google Scholar CrossRef Search ADS PubMed Krutzsch W . 1970 . Zur Kenntnis fossiler disperser Tetradenpollen . Paläontologische Abhandlungen Abteilung B Paläobotanik 3 : 399 – 433 . Krutzsch W , Vanhoorne R . 1977 . Die Pollenflora von Epinois und Loksbergen in Belgien . Palaeontographica Abteilung B 163 : 1 – 110 . Kvaček Z , Walther H . 1984 . Nachweis Tertiärer Theaceen Mitteleuropas nach blatt-epidermalen Untersuchungen. II Bestimmung fossiler Theaceen Sippen . Feddes Repertorium 95 : 331 – 346 . Litke R . 1966 . Kutikularanalytische Untersuchungen im Niederlausitzer Unterflöz . Paläontologische Abhandlungen B II : 327 – 426 . Magallón S , Gómez-Acevedo S , Sánchez-Reyes LL , Hernandéz-Hernandéz T . 2015 . A metacalibrated time-tree documents the early rise of flowering plant phylogenetic diversity . New Phytologist . 207 : 437 – 453 . Google Scholar CrossRef Search ADS PubMed Mai DH . 1971 . Über fossile Lauraceae und Theaceae in Mitteleuropa . Feddes Repertorium 82 : 313 – 341 . Google Scholar CrossRef Search ADS Mai DH . 1995 . Tertiäre Vegetationsgeschichte Europas . Jena: Gustav Fischer Verlag . Manchester SR , Grímsson F , Zetter R . 2015 . Assessing the fossil record of asterids in the context of our current phylogenetic framework . Annals of the Missouri Botanical Gardens 100 : 329 – 363 . Google Scholar CrossRef Search ADS Massoni J , Doyle J , Sauquet H . 2015 . Fossil calibration of Magnoliideae an ancient lineage of angiosperms . Palaeontologica Electronica 17 : 1 – 25 . Martínez-Millán M . 2010 . Fossil record and age of the Asteridae . Botanical Review 76 : 83 – 135 . Google Scholar CrossRef Search ADS Martínez-Millán M , Crepet WL , Nixon KC . 2009 . Pentapetalum trifasciculandricus gen. et sp. nov., a thealean fossil flower from the Raritan Formation, New Jersey, USA (Turonian, Late Cretaceous) . American Journal of Botany 96 : 933 – 949 . Google Scholar CrossRef Search ADS PubMed McGuire AF , Kron KA . 2005 . Phylogenetic relationships of European and African ericas . International Journal of Plant Sciences 166 : 311 – 318 . Google Scholar CrossRef Search ADS Miyoshi N , Fujiki T , Kimura H . 2011 . Pollen flora of Japan . Japan : Hokaido University Press . Morley RJ . 2003 . Interplate dispersal paths for megathermal angiosperms . Perspectives of Plant Evolution and Systematics 6 : 5 – 20 . Google Scholar CrossRef Search ADS Morton CM , Dickison WC . 1992 . Comparative pollen morphology of the Styracaceae . Grana 31 : 1 – 15 . Google Scholar CrossRef Search ADS Morton CM , Kincaid DT . 1995 . A model for coding pollen size in reference to phylogeny using examples from the Ebenaceae . American Journal of Botany 82 : 1173 – 1178 . Google Scholar CrossRef Search ADS Muller J . 1981 . Fossil pollen records of the extant angiosperms . Botanical Review 47 : 1 – 142 . Google Scholar CrossRef Search ADS Nixon KC , Crepet WL . 1993 . Late Cretaceous fossil flowers of ericalean affinity . American Journal of Botany 80 : 616 – 623 . Google Scholar CrossRef Search ADS Oschkinis V , Gregor HJ . 1992 . Paläontologische Funde aus der eozänen Braunkohle des Untertagebaus Stolzenbach (PreußenElektra) in Niederhessen. I. Die Flora . Documenta naturae 72 : 1 – 31 . Oschkinis V , Gregor H-J . 2005 . Die eozänen Braunkohleschichten aus dem Untertagebau Stolzenbach bei Kassel (PreußenElektra, Niederhessen). I. Geologie und Fundumstände . Documenta naturae 156 : 1 – 25 . Pennington TD . 1991 . The genera of Sapotaceae . Kew : Royal Botanic Gardens . Pirie MD , Oliver EG , Mugrabi de Kuppler A , Gehrke B , Le Maitre NC , Kandziora M , Bellstedt DU . 2016 . The biodiversity hotspot as evolutionary hot-bed: spectacular radiation of Erica in the Cape Floristic region . BMC Evolutionary Biology 16 : 190 . Google Scholar CrossRef Search ADS PubMed Palazzesi L , Barreda V . 2004 . Primer registro palinológico de la Formación Puerto Madryn (Chubut, Argentinia) . Ameghiniana 41 : 355 – 362 . Richardson JE , Bakar AM , Tosh J , Armstrong K , Smedmark J , Anderberg AA , Slik F , Wilkie P . 2014 . The influence of tectonics, sea-level changes and dispersal on migration and diversification of Isonandrae (Sapotaceae) . Botanical Journal of the Linnean Society 174 : 130 – 140 . Google Scholar CrossRef Search ADS Ritzkowski S . 2005 . Das Tertiär der Hessischen Senke in der stratigraphischen Tabelle von Deutschland 2002 . Newsletter of Stratigraphy 41 : 339 – 346 . Google Scholar CrossRef Search ADS Sarwar AKMG . 2007 . Pollen morphology and its systematic significance in Ericaceae . PhD thesis, Hokkaido University . Sarwar AKMG , Ito T , Tahahashi H . 2006 . An overview of pollen morphology and its relevance to the sectional classification of Vaccinium L. (Ericaceae) . Japanese Journal of Palynology 152 : 15 – 34 . Sarwar GAKM , Takahashi H . 2012 . Pollen morphology of Kalmia L. (Phyllodoceae, Ericaceae) and its taxonomic significance . Bangladesh Journal of Plant Taxonomy 19 : 123 – 133 . Google Scholar CrossRef Search ADS Sarwar AKMG , Takahashi H . 2013 . Pollen morphology of Rhododendron and related genera and it taxonomic significance . Bangladesh Journal of Plant Taxonomy 20 : 185 – 199 . Google Scholar CrossRef Search ADS Sarwar AKMG , Takahashi H . 2014 . Pollen morphology of Erica L. and related genera and its taxonomic significance . Grana 53 : 221 – 231 . Google Scholar CrossRef Search ADS Schönenberger J , Anderberg AA , Sytsma KJ . 2005 . Molecular phylogenetics and patterns of floral evolution in the Ericales . International Journal of Plant Sciences 166 : 265 – 288 . Google Scholar CrossRef Search ADS Schönenberger J , Friis EM . 2001 . Fossil flowers of ericalean affinity from the Late Cretaceous of southern Sweden . American Journal of Botany 88 : 467 – 480 . Google Scholar CrossRef Search ADS PubMed Schönenberger J , von Balthazar M , Takahashi M , Xiao X , Crane PR , Herendeen PS . 2012 . Glandulocalyx upatoiensis, a fossil flower of Ericales (Actinidiaceae/Clethraceae) from the Late Cretaceous (Santonian) of Georgia, USA . Annals of Botany 109 : 921 – 936 . Google Scholar CrossRef Search ADS PubMed Schrank E . 2017 . Palynology of the Albian Makhtesch Quatan site, northern Negev (Israel) . Review of Palaeobotany and Palynology 246 : 185 – 215 Google Scholar CrossRef Search ADS Smedmark JE , Anderberg AA . 2007 . Boreotropical migration explains hybridization between geographically distant lineages in the pantropical clade Sideroxyleae (Sapotaceae) . American Journal of Botany 94 : 1491 – 1505 . Google Scholar CrossRef Search ADS PubMed Smedmark JE , Swenson U , Anderberg AA . 2006 . Accounting for variation of substitution rates through time in Bayesian phylogeny reconstruction of Sapotoideae (Sapotaceae) . Molecular Phylogenetics and Evolution 39 : 706 – 721 . Google Scholar CrossRef Search ADS PubMed Spicer RA , Herman AB , Liao W , Spicer TEV , Kodrul TM . 2014 . Cool tropics in the Middle Eocene: evidence from the Changchang flora, Hainan, Island, China . Palaeogeography Palaeoclimatology Palaeoecology 412 : 1 – 16 . Google Scholar CrossRef Search ADS Stoian LM . 2002 . Late Cretaceous - Late Eocene palynofloras from drillhole Troas 1, offshore Otway basin, South Australia . Australia : Department of Primary Industries and Resources . Report book 2002/010, Gov. Stride G , Nylinder S , Swenson U . 2014 . Revisiting the biogeography of Sideroxylon (Sapotaceae) and an evaluation of the taxonomic status of Argania and Spiniluma . Australian Systematic Botany 27 : 104 – 118 . Google Scholar CrossRef Search ADS Swenson U , Anderberg AA . 2005 . Phylogeny, character evolution, and classification of Sapotaceae (Ericales) . Cladistics 21 : 101 – 130 . Google Scholar CrossRef Search ADS Swenson U , Bartish IV , Munzinger J . 2007 . Phylogeny, diagnostic characters and generic limitation of Australasian Chrysophylloideae (Sapotaceae, Ericales). Evidence from ITS sequence data and morphology . Cladistics 23 : 201 – 228 Google Scholar CrossRef Search ADS Swenson U , Richardson JE , Bartish IV . 2008 . Multi-gene phylogeny of the pantropical subfamily Chrysophylloideae (Sapotaceae): evidence of generic polyphyly and extensive morphological homoplasy . Cladistics 24 : 1006 – 1031 . Google Scholar CrossRef Search ADS Taylor DW . 1989 . Select palynomorphs from the middle Eocene Claiborne Formation, Tenn. (U.S.A.) . Review of Palaeobotany and Palynology 58 : 111 – 128 . Google Scholar CrossRef Search ADS Thiry M , Dupuis C , Aubry M-P , Berggren WA , Ellison RA , Knox RWO’B , Sinha A , Stott L . 1998 . Tentative correlations between continental deposits of the argiles plastiques (Paris Basin) and Reading Beds (London Basin), based on chemostratigraphy . Strata 9 : 125 – 129 . Tsou C-H . 1997 . Embryology of the Theaceae – anther and ovule development of Camellia, Franklinia, and Schima . American Journal of Botany 84 : 369 – 381 . Google Scholar CrossRef Search ADS PubMed Turner B , Munzinger J , Duangjai S , Temsch EM , Stockenhuber R , Barfuss MH , Chase MW , Samuel R . 2013 . Molecular phylogenetics of New Caledonian Diospyros (Ebenaceae) using plastid and nuclear markers . Molecular Phylogenetics and Evolution 69 : 740 – 763 . Google Scholar CrossRef Search ADS PubMed Van der Burgh J . 1987 . Miocene floras in the lower Rhenish basin and their ecological interpretation . Review of Palaeobotany and Palynology 52 : 299 – 366 . Google Scholar CrossRef Search ADS Vaudois-Miéja N . 1980 . Paléocarpologie des Ebénacées de la flore des Grès à Sabals (Anjou, France) . Review of Palaeobotany and Palynology 29 : 49 – 64 . Google Scholar CrossRef Search ADS Vaudois-Miéja N . 1983 . Extension paléogéographique en Europe de Ĺactuel genre asiatique Rehderodendron Hu (Styracacées) . Compte-Rendus des Seances Académie des Sciences 2 , 296 : 125 – 130 . Vijayan K , Zhang WJ , Tsou C-H . 2009 . Molecular taxonomy of Camellia (Theaceae) inferred from nrITS sequences . American Journal of Botany 96 : 1348 – 1360 . Google Scholar CrossRef Search ADS PubMed Wallnhöfer B . 2001 . The biology and systematics of Ebenaceae: a review . Annalen des Naturhistorischen Museums Wien B 103 : 485 – 512 . Walker JW , Walker AG . 1984 . Ultrastructure of Lower Cretaceous angiosperm pollen and the origin and evolution of flowering plants . Annals of the Missouri Botanical Garden 71 : 464 – 521 . Google Scholar CrossRef Search ADS Yao YF , Bera S , Ferguson DK , Mosbrugger V , Paudayal KN , Jin J . 2009 . Reconstruction of paleovegetation and paleoclimate in the Early and Middle Eocene, Hainan Island, China . Climate Change 92 : 169 – 189 . Google Scholar CrossRef Search ADS Zetter R , Hesse M . 1996 . The morphology of pollen tetrads and viscin threads in some Tertiary Rhododendron-like Ericaceae . Grana 35 : 285 – 294 . Google Scholar CrossRef Search ADS Zetter R, Hofmann CC. 2001. New aspects on the palynoflora of the lowermost Eocene (Krappfeld, Carinthia). In: WE Piller, MW Rasser (eds.) Paleogene of the Eastern Alps. Verlag der Österreichischen Akademie der Wissenschaften 12: 473–507. Zavada MS , Wei Z-X . 1993 . A contribution to pollen morphology of Camellia (Theaceae) . Grana 32 : 233 – 242 . Google Scholar CrossRef Search ADS © 2018 The Linnean Society of London, Botanical Journal of the Linnean Society This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/open_access/funder_policies/chorus/standard_publication_model)

Migliore, Jérémy;Baumel, Alex;Leriche, Agathe;Juin, Marianick;Médail, Frédéric

doi: 10.1093/botlinnean/boy032pmid: N/A

Abstract The simultaneous application of species distribution modelling (SDM) and study of genetic imprints left by range dynamics is appropriate when examining the biogeographical processes that have favoured the survival of plants through past climate changes. Nevertheless, such an approach is rarely performed on the scale of the entire Mediterranean and almost never concerns widespread thermophilous plants. Here, we examine the biogeographical responses of an important Mediterranean shrub, Myrtus communis (Myrtaceae), to severe Quaternary climate conditions. Our analysis combines SDM and phylogeography based on plastid/nuclear DNA sequences and AFLP data. Palaeoclimatic models using MaxEnt and levels of genetic diversity in M. communis are used to infer drastic changes in areas of climatic suitability during the last 130000 years, with a southward range contraction during the Last Glacial Maximum. Modelling of past areas of suitability for M. communis identifies a few relatively small long-term refugia, suggesting that it survived in temporary refugia during glacial periods. Myrtus communis is characterized by a higher genetic diversity and distinctiveness in the southern part of its range, where it was less impacted by glaciations. The structure of genetic diversity reveals stronger range expansions in the western part of the range, whereas migration processes remained much more restricted in the eastern Mediterranean. Mediterranean, Myrtus communis (myrtle, palaeoecology, phylogeography, Quaternary glaciations, range shift, refugia, species distribution modelling INTRODUCTION Mediterranean climate ecosystems harbour the richest extra-tropical floras, but contemporary ecological factors and environmental heterogeneity do not entirely explain their remarkable biodiversity and endemism or their spatial patterns (Rundel et al., 2016). It has been suggested for the Cape Region and south-western Australia that long-term environmental stability allowed ancient lineages to continue their diversification (Cowling et al., 2015). In contrast, clades are generally younger in the Mediterranean region, leading to the hypothesis that environmental instability has prevented the long-term accumulation of diversity here (Valente & Vargas, 2013; Cowling et al., 2015). Indeed, several old lineages, i.e. those of early to mid-Cenozoic origins, were pruned by extinction c. 3.2 Mya, during the transition from a tropical to a Mediterranean climate (Postigo Mijarra et al., 2009; Rodríguez-Sánchez et al., 2010). The Quaternary glaciations starting at c. 2.7 Mya subsequently led to increased climatic severity characterized by more arid conditions and a mean temperature of the coldest month c. 12 °C lower than at present (Wu et al., 2007). These drastic climatic conditions favoured open landscapes of grasslands and steppe vegetation to the detriment of more thermophilous and woody vegetation (Quézel & Médail, 2003; Thompson, 2005). Nevertheless, recent phylogenetic evidence suggests that most of plant clades of the Mediterranean region diversified between the Miocene and Pliocene (from 23 Mya), i.e. before the onset of the Mediterranean climate (Vargas, Fernández-Mazuecos & Heleno, 2018), so that extinction of older lineages was counteracted by diversifying processes in younger lineages. Species distribution modelling (SDM) with palaeoclimatic explanatory variables can provide insights into the processes shaping modern genetic and species diversity at the same time as tracking the contraction/expansion of populations in response to successive Quaternary glacial/interglacial oscillations (Waltari et al., 2007; Carnaval et al., 2009). The hypothesis of long-term refugia proposes the stability of suitable areas that allows viable populations to persist through climatic oscillations (Ashcroft, 2010; Stewart et al., 2010). In the framework of refuge theory, we should expect to find a positive correlation between the stability of suitable areas and genetic singularity, because long-term populations tend to harbour more diversity than recently expanding ones (Svenning et al., 2011; Hampe et al., 2013; Gavin et al., 2014). These long-term refugia have been defined as ‘phylogeographical hotspots’, i.e. significant reservoirs of unique genetic diversity of Mediterranean plant species that probably played a key role in speciation processes (Médail & Diadema, 2009). The complex spatio-temporal trajectories of species and populations in and out of refugia need to be better integrated (Gavin et al., 2014). Combining palaeoclimatic SDM and phylogeography allows us to infer the existence of putative Quaternary long-term refugia. This was the case for Olea europaea L., for which a large part of its populations located in the south-western and the coastal eastern Mediterranean persisted under continuously suitable areas during the Last Inter-Glacial (LIG, c. 116–130 kya BP), the Last Glacial Maximum (LGM, c. 19.0–26.5 kya) and current climatic conditions (Besnard et al., 2013). The simultaneous analysis of both SDM and genetic imprints left by range dynamics has been rarely performed on the scale of the entire Mediterranean and even less so with regard to widespread thermophilous plants. In this context, we chose to focus on the myrtle, Myrtus communis L. (Myrtaceae), because it is one of the few Mediterranean thermophilous evergreen shrubs occurring in coastal and lowland environments, below an elevation of 500 m. A few common woody plants of the thermo-Mediterranean shrublands have been studied using a phylogeographical approach (Pinus pinea L.: Vendramin et al., 2008; Cistus monspeliensis L.: Fernández-Mazuecos & Vargas, 2010; Chamaerops humilis L.: García-Castaño et al., 2014; Guzmán et al., 2017). Myrtle is characterized by a long biogeographical history, having persisted in the Mediterranean region through the Cenozoic since at least the Neogene period (Migliore et al., 2012). Found primarily in low-elevation habitats, this shrub has probably been grately affected by both climatic and habitat changes through latitudinal shifts of climatic conditions and eustatic oscillations of the Mediterranean Sea (Pirazzoli, 2005). Thus, the framework of the expansion/contraction model (Bennett & Provan, 2008; Gavin et al., 2014) could be relevant for past Myrtus range dynamics, especially during the climatic oscillations of the Quaternary. To examine the biogeographical responses of this Mediterranean shrub to severe Quaternary climate conditions, we used MaxEnt species distribution models, and detected the current and past suitable areas for M. communis on the scale of its entire Mediterranean and Macaronesian distribution range. From this we inferred potential range shifts and identified refugia on the basis of bioclimatic suitability during the LIG, LGM and the Mid Holocene (c. 6 kya). We compared the implications of the SDM study with the phylogeography of M. communis based on plastid and nuclear DNA sequences and multilocus AFLP markers. MATERIAL AND METHODS Modelling of suitable areas for Myrtus communis Occurrence data The study area was divided into a grid of 3019 1 × 1-km cells using ArcGis 10.2 (ESRI, Redlands, WA, USA). All the cells under consideration comprised at least 75% land area. An occurrence dataset, consisting of 4168 observations of M. communis (Fig. 1A), was collected from the following sources: the Global Biodiversity Information Facility data portal (2011 occurrences; http://www.gbif.org/), our own field GPS data (179 Mediterranean occurrences), GPS data from the French CBNMED database (758 occurrences; http://flore.silene.eu), the Atlas of the Aegean Flora (462 occurrences; A. Strid, unpubl. data), Flora Croatica (660 occurrences; Nikolić, 2015), and scientific literature on the flora of northern Africa and Italy (98 occurrences). For modelling purposes, the dataset was reduced to one occurrence per unit cell (n = 3019). Figure 1. View largeDownload slide (A) Known occurrences of Myrtus communis in the Mediterranean Basin, gathered from botanical databases and literature in addition to field collected data (n = 4168). (B) Current suitable areas for the species, applying species distribution modelling (SDM) from actual occurrences using MaxEnt with six bioclimatic variables (see Methods). Figure 1. View largeDownload slide (A) Known occurrences of Myrtus communis in the Mediterranean Basin, gathered from botanical databases and literature in addition to field collected data (n = 4168). (B) Current suitable areas for the species, applying species distribution modelling (SDM) from actual occurrences using MaxEnt with six bioclimatic variables (see Methods). Bioclimatic variables Six variables related to the ecological requirements of the species were chosen and used to fit the bioclimatic distribution model of M. communis. Isothermality (BIO3) reflects the influence of larger or smaller temperature fluctuations within a month, relative to the year (generally useful for studies of insular and maritime environments). Minimal temperature of the coldest month (BIO6) quantifies potentially lethal frost events and generalized stress due to low temperatures. Annual temperature range (BIO7) is useful for examining the effect of the ranges of annual temperature extremes. Annual precipitation (BIO12) approximates average water availability, and precipitation of the driest month (BIO14) describes the extremes associated with potentially lethal drought events and quantifies stress due to low water availability. Finally, because M. communis requires higher temperatures during the wettest period of year for its germination, mean temperature in the wettest quarter (BIO8) was also considered. These bioclimatic variables were extracted from the Bioclim dataset, provided by WorldClim 1.4 in a GIS-based raster format (1 × 1-km resolution; Hijmans et al., 2005; O’Donnell & Ignizio, 2012). As a safeguard against redundancy among our variables, we checked that all pairwise Pearson correlation coefficients between any two of the six selected variables were < 0.71. Modelling method The maximum entropy algorithm as implemented in MaxEnt 3.3.1 was used to define suitable areas for M. communis (Phillips, Anderson & Schapire, 2006; Phillips & Dudík, 2008). MaxEnt estimates suitable areas for species by finding the distribution of maximum entropy (i.e. closest to uniform) subject to the constraint that the expected value of each environmental variable (or its transform and/or interactions) under the estimated distribution matches its empirical average. The method has been shown to be well adapted to presence-only data and has consistently demonstrated performance that is competitive with other methods (Elith et al., 2006; but see Yackulic et al., 2013; Phillips et al., 2017). Several models were fitted by investigating all feature types (linear, quadratic and product) and all combinations of features, except for the threshold feature, which is known to fit overly conservative models. Ten replicates were performed for each model (cross-validation method). The machine learning fitting process was set to end when training gain fell below the threshold value of 0.0001, or after 3000 iterations. The user-defined number of background points was set to 50000, and individual points were chosen to correct for latitudinal and sampling biases (i.e. an orientated random selection of 50000 points across the entire study area, paying attention to cell area, which is a function of latitude, and to sampling bias). The extrapolated function was set to ‘no’. Model selection The Akaike information criterion (AIC) was calculated for all models and averaged by combination to define the most parsimonious combination of features. Area under the curve (AUC) and standard deviation metrics were checked for congruence with AIC results. The final model of the current suitable areas for M. communis was fitted using the selected combination of features, with 50 replicates (identical parameters to the fitting procedure). A jackknife analysis was performed to evaluate the contributions of variables to the model, and response curves were created to assess how each variable influences the level of suitability. Model projections and persistence assessment The final model was projected using previous climate history to determine the suitable areas for M. communis during the Mid Holocene, LGM and LIG. It was assumed that the ecological requirements of M. communis have remained similar over the last Pleistocene climatic cycles (Valiente-Banuet et al., 2006; Nogués-Bravo, 2009). Reconstructed bioclimatic variables were provided by: the recently updated Coupled Model Intercomparison Project Phase 5 (WorldClim 1.4); testing data from Global Climate Models (GCM) CCSM4 (Community Climate System Model); MIROC-ESM (Model for Interdisciplinary Research on Climate); and the MPI-ESM-P (Max Planck Institute Earth system model) for the Mid Holocene (30-s resolution, i.e. c. 1 × 1 km) and LGM (2.5-min resolution, i.e. c. 5 × 5 km; this layer was resampled to a 1 × 1 km resolution raster but not interpolated, meaning the resolution of interpretation for the LGM model and derived inferences must be 5 × 5 km). For the LIG chronozone, data from Otto-Bliesner et al. (2006) were used (30-s resolution). Suitability values as modelled by MaxEnt were converted into presence/absence, using the 10-percentile threshold of 0.305571, the minimum probability for presence cells after discarding the lowest 10% of the suitability distribution (Pearson et al., 2007). Long-term areas of persistence were inferred as continuously suitable areas (> 10-percentile value) over the LIG, LGM and Mid Holocene. Phylogeographical reconstruction of Myrtus communis Genetic sampling strategy and molecular methods We collected 118 samples from the Mediterranean and Macaronesian distribution area of M. communis. One randomly sampled individual per site was newly genotyped using AFLP (Table S1; Vos et al., 1995). The AFLP reaction was performed on 300–500 ng DNA, as described in Migliore et al. (2011). After screening selective primers, three primer combinations giving clearly visible band profiles were chosen (EcoRI-AAGG/MseI-CCAG, EcoRI-AAC/MseI-CAC, EcoRI-AAC/MseI-CAA). Polyacrylamide electrophoreses (0.4%) were performed using a 96-capillary automated sequencer (Megabace 1000, Amersham Bioscience), with manual scoring used to detect error-prone markers with unreliable peaks. The reproducibility of the AFLP markers was checked by repeating the complete analysis on 20 samples for each pair of primers (Bonin et al., 2004). A genotyping error rate of 5.7% was obtained, due to the unreliability of some markers, which were subsequently removed from the dataset. We also included sequence data for the plastid DNA trnL-trnF and rpl2-trnH intergenic regions and the nuclear DNA external transcribed spacer (ETS) region and the internal transcribed spacer (ITS1–5.8S–ITS2) region (n = 309 and 176, respectively; Supporting Information Table S1). The methods used to extract DNA and to sequence DNA markers are described by Migliore et al. (2012), because these sequences were previously generated for phylogenetic reconstruction and molecular dating. Genetic analyses AFLP genotypes were assigned to genetically homogeneous clusters via the model-based clustering algorithm provided by STRUCTURE 2.3.3 (Pritchard, Stephens & Donnelly, 2000; Falush, Stephens & Pritchard, 2007). Bayesian analysis was run for 1000000 generations (burn-in of 100000), and for K = 2–21 with ten iterations for each K-value. The admixture and recessive allele models were then chosen. The most likely number of clusters was determined using the criteria lnPr(X/K) and ΔK (Evanno, Regnaut & Goudet, 2005). After independently concatenating the two plastid and the two nuclear regions, plastid haplotype and nuclear ribotype median-joining networks were constructed, using NETWORK 4.613 (Bandelt, Forster & Röhl, 1999). To check the congruence of the genetic patterns from the same samples studied with AFLP and DNA sequences, a Mantel test was undertaken, comparing inter-AFLP genotype distances (Jaccard index) to nucleotide divergence of plastid DNA haplotypes and nuclear DNA ribotypes (p distance). The Mantel test was based on 10000 permutations (GenAlEx 6.5; Peakall & Smouse, 2012). Genetic diversity and singularity were summarized by several indices calculated using GenAlEx 6.5: the number of plastid DNA haplotypes (NH) and nuclear DNA ribotypes (NR) including private haplotypes and ribotypes; the unbiased diversity (uh) for all the markers; the mean genetic distance within each group of samples sequenced (GD); the number of AFLP bands (NB); the number of private AFLP bands (PB); and the percentage of AFLP polymorphic loci (%P). All these indices were computed according to (1) the main AFLP genetic clusters detected, and (2) the main geographical trends of climatic suitability found according to species distribution models, excluding the Macaronesian genetic cluster whose sampling was too low. RESULTS Modelled current and past suitable areas for Myrtus communis The most parsimonious fitted model (based on AIC, Table S2, Fig. 1B) was a combination of linear and quadratic features, with a True Skill Statistics (TSS) value of 0.73 (calculated using the 10-percentile threshold for a binary transformation; and an AUC value for training data of 0.88 ± 0.0068 (gain threshold reached after 1040 iterations). Jackknife and response curves indicated that higher suitability values were inferred with high values of ‘minimal temperature of coldest month’ with an optimum at 6 °C [BIO6; test gain when used in isolation (Tg) of 0.88]. Suitability also increased with moderate values of ‘mean temperature in the wettest quarter’ with an optimum at 18 °C (BIO8; Tg = 0.49) and low values of ‘annual temperature range’ (optimum = 8 °C for BIO7; Tg = 0.31). Annual precipitation values also positively influenced the level of suitability (Tg = 0.25), with an inferred maximum suitability of 825 mm annual precipitation. The variables ‘precipitation in the driest month’ (BIO14) and ‘isothermality’ (BIO3) had a slight influence on the current model (Tg = 0.025 and 0.023, respectively). In summary, M. communis appeared to be particularly sensitive to cold stress, with suitable areas characterized by mean minimal temperature above frost threshold and mild winter temperatures. Areas with low levels of precipitation (from 220 mm) are suitable for M. communis, but its optimum is around the maximum average annual precipitation value that defines the Mediterranean climate. The low number of available Myrtle fossils from the LGM and Mid Holocene (Migliore et al., 2012) makes it difficult to choose between one of the three past-climate models (CCSM4, MIROC-ESM and MPI-ESM-P) for projecting SDM. The range dynamics of M. communis over time using the MIROC-ESM climate model is the only one presented here (Fig. 2; results with the two other past-climate models are presented as in the Supporting Information). However, the SDM inference of long-term areas of persistence is presented using all three past-climate model projections (Fig. 3). Figure 2. View largeDownload slide Suitable areas for Myrtus communis (projection of the most parsimonious MaxEnt model fitted to current climatic conditions from actual occurrences) during the Last Inter-Glacial (LIG) period (A), the Last Glacial Maximum (LGM) using the MIROC-ESM climate model (B), the Mid Holocene using the MIROC-ESM climate model (C) and the Present (D). Continuous suitability values modelled with MaxEnt were converted into presence/absence, using the 10-percentile threshold ≥ 0.3 (see Methods and Figs S1–4). Shading indicates palaeodistributions modelled just before the chronozone represented. Figure 2. View largeDownload slide Suitable areas for Myrtus communis (projection of the most parsimonious MaxEnt model fitted to current climatic conditions from actual occurrences) during the Last Inter-Glacial (LIG) period (A), the Last Glacial Maximum (LGM) using the MIROC-ESM climate model (B), the Mid Holocene using the MIROC-ESM climate model (C) and the Present (D). Continuous suitability values modelled with MaxEnt were converted into presence/absence, using the 10-percentile threshold ≥ 0.3 (see Methods and Figs S1–4). Shading indicates palaeodistributions modelled just before the chronozone represented. Figure 3. View largeDownload slide Areas of long-term persistence inferred as continuously suitable areas over the Last Inter-Glacial (LIG), the Last Glacial Maximum (LGM) and Mid Holocene using the CCSM4 (A), MIROC-ESM (B) and MPI-ESM-P (C) climate models (10-percentile threshold ≥ 0.3; 5 × 5-km resolution). Figure 3. View largeDownload slide Areas of long-term persistence inferred as continuously suitable areas over the Last Inter-Glacial (LIG), the Last Glacial Maximum (LGM) and Mid Holocene using the CCSM4 (A), MIROC-ESM (B) and MPI-ESM-P (C) climate models (10-percentile threshold ≥ 0.3; 5 × 5-km resolution). During the LIG, M. communis was inferred to be potentially widespread throughout the entire Mediterranean Basin, albeit with a suitable area along the European and northern African coasts that was narrower than the current one (Fig. 2A). The total inferred size of suitable areas approached 4468025 km2 during the LIG versus 13482343 km2 at present. A high level of suitability was suggested along the Atlantic coast (Portugal, Gibraltar, south-western Iberian Peninsula and western Morocco). In the eastern Mediterranean, the spatial range was similar to the current one, with a slight southward shift of suitable areas. During the LGM, the whole study area was much less suitable for M. communis than it is under current climate conditions (Fig. 2B). A drastic restriction of its potential suitable range is suggested: on average, the climate models showed 4343671 km2 of suitable areas lost between the LIG and the LGM (Figs S1, 2). The physiography of the Mediterranean Basin was quite different during these periods from how it is at present: in particular, the Adriatic Gulf had seen a sea-level drop of 100–120 m relative to the present. Using the LIG map as a baseline for analysing the changes occurring during the LGM, we can infer that the suitable areas for M. communis could have moved down towards lower altitudes and southward towards lower latitudes. This was apparent on western and central Mediterranean islands, along the Atlantic coastline, and, to a lesser degree, in Greek and Turkish areas (Fig. 2B). Among the climate models used for the LGM, the CCSM4 model was the most restrictive with only 91052 km2 of suitable area, in contrast to 120428 km2 for the MIROC-ESM model (with greater areas of suitability in the Iberian Peninsula and Gibraltar Strait) and 161585 km2 for the MPI-ESM-P model (with greater areas of suitability in western Mediterranean islands) (Figs S1, 2). During the Mid Holocene, climatic conditions were inferred to have been more suitable than during the LGM, thus allowing for a range expansion of M. communis (Figs 2C, S3, S4). Averaging over the three climate models, this gain in area comes to 196412 km2 in total. An important but unexpected result was the low number and small size of inferred long-term refugia (i.e. continuously suitable areas over the LIG, LGM and Mid Holocene), whatever past-climate model is used (Fig. 3). We infer drastic changes in the spatial range of the species between these periods together with a quasi-absence of overlap between suitable areas during the three contrasted climatic periods, apart from parts of the Macaronesian islands (Azores and Madeira), around the Gibraltar Strait, on the island of Crete and in the Levant (Fig. 3). Above the 41°N parallel, no long-term persistence areas over the LIG, LGM and Mid Holocene were detected in the northern part of the Mediterranean Basin (Fig. 3). Multi-markers genetic diversity of Myrtus communis The AFLP analysis on 118 individuals of M. communis resulted in 199 scored fragments, 191 of which were found to be polymorphic, with a length varying from 66 to 460 bp (dataset available on request). The most probable partitioning of the genetic variation of M. communis according to STRUCTURE was in K = 2 and K = 4 genetic clusters (Figs 4A, B, S5). Figure 4. View largeDownload slide Spatial genetic structure of Myrtus communis through its Mediterranean distribution range (A), based on the AFLP data summarized from STRUCTURE, based on K = 4 genetic clusters (B), and the median-joining networks from concatenated plastid DNA sequences trnL-trnF and rpl2-trnH intergenic regions (C) and concatenated nuclear DNA sequences external transcribed spacer (ETS) region and internal transcribed spacer (ITS1–5.8S–ITS2) (D). Colour chart is based on the four main genetic clusters identified by AFLP: EM (eastern Mediterranean in orange), WM (western Mediterranean in green), CM (circum-Mediterranean in blue) and Macaronesia (in purple). Numbers in A refer to spatially closed populations for which samples have the same genetic clusters identified. Haplotype and ribotype identifiers are detailed in Figure S6; circle sizes in networks are proportional to haplotype/ribotype frequencies. Figure 4. View largeDownload slide Spatial genetic structure of Myrtus communis through its Mediterranean distribution range (A), based on the AFLP data summarized from STRUCTURE, based on K = 4 genetic clusters (B), and the median-joining networks from concatenated plastid DNA sequences trnL-trnF and rpl2-trnH intergenic regions (C) and concatenated nuclear DNA sequences external transcribed spacer (ETS) region and internal transcribed spacer (ITS1–5.8S–ITS2) (D). Colour chart is based on the four main genetic clusters identified by AFLP: EM (eastern Mediterranean in orange), WM (western Mediterranean in green), CM (circum-Mediterranean in blue) and Macaronesia (in purple). Numbers in A refer to spatially closed populations for which samples have the same genetic clusters identified. Haplotype and ribotype identifiers are detailed in Figure S6; circle sizes in networks are proportional to haplotype/ribotype frequencies. The genetic distances calculated from AFLP, plastid and nuclear DNA datasets were positively correlated and the Mantel tests were significant for the Mediterranean samples (n = 110: plastid DNA vs. AFLP: R2 = 0.20, P = 0.0001; plastid DNA vs. nuclear DNA: R2 = 0.19, P = 0.0001; nuclear DNA vs. AFLP: R2 = 0.20, P = 0.0001). Genetic diversity indices were computed for each AFLP genetic cluster: for the three molecular markers, the western Mediterranean (WM) and eastern Mediterranean (EM) clusters had similar levels of diversity when accounting for the different levels of sampling (uh; Table 1). The circum-Mediterranean (CM) cluster had the lowest diversity of plastid DNA markers, but had the highest diversity of nuclear DNA, and it also had the highest percentage of polymorphic markers for AFLP (%P; Table 1). This contrast could be due to the intermediate geographical position of the CM cluster providing a higher admixture of the WM or EM clusters, i.e. contact zones (see below). Table 1. Genetic (plastid DNA sequences, nuclear DNA sequences and AFLP data) features for Mediterranean samples of Myrtus communis in relation to the three detected genetic clusters (eastern, circum- and western Mediterranean) and with the distribution of samples to the north or south of the 41°N parallel: sample size (N), number of plastid DNA haplotypes (NH with number of private haplotypes in parentheses), unbiased diversity (uh with standard errors in parentheses), mean genetic distance (GD), number of nuclear DNA ribotypes (NR with number of private nrDNA ribotypes in parentheses), number of AFLP bands (NB), number of private AFLP bands (PB) and percentage of polymorphic loci (%P) Plastid DNA sequences Nuclear DNA sequences AFLP data N NH uh GD n NR uh GD n NB PB uh %P Cluster EM, eastern Mediterranean 60 2 0.051 (0.051) 0.463 26 6 0.042 (0.021) 6.108 10 110 12 0.145 (0.013) 43.72 Cluster CM, circum-Mediterranean 86 3 0.037 (0.028) 0.333 52 16 0.091 (0.030) 7.177 48 151 23 0.124 (0.011) 67.34 Cluster WM, western Mediterranean 155 4 0.055 (0.032) 0.491 94 15 0.041 (0.018) 3.831 55 132 19 0.130 (0.013) 54.27 All data 301 9 0.212 (0.057) / 172 37 0.294 (0.035) / 113 199 / 0.278 (0.018) / NM, northern Mediterranean (41°N) 133 4 (0) 0.085 (0.048) 0.768 86 19 (11) 0.186 (0.032) 8.332 43 135 10 0.161 (0.014) 56.78 SM, southern Mediterranean (41°N) 168 9 (5) 0.267 (0.061) 2.407 86 26 (17) 0.340 (0.036) 11.844 70 175 50 0.173 (0.013) 82.41 Plastid DNA sequences Nuclear DNA sequences AFLP data N NH uh GD n NR uh GD n NB PB uh %P Cluster EM, eastern Mediterranean 60 2 0.051 (0.051) 0.463 26 6 0.042 (0.021) 6.108 10 110 12 0.145 (0.013) 43.72 Cluster CM, circum-Mediterranean 86 3 0.037 (0.028) 0.333 52 16 0.091 (0.030) 7.177 48 151 23 0.124 (0.011) 67.34 Cluster WM, western Mediterranean 155 4 0.055 (0.032) 0.491 94 15 0.041 (0.018) 3.831 55 132 19 0.130 (0.013) 54.27 All data 301 9 0.212 (0.057) / 172 37 0.294 (0.035) / 113 199 / 0.278 (0.018) / NM, northern Mediterranean (41°N) 133 4 (0) 0.085 (0.048) 0.768 86 19 (11) 0.186 (0.032) 8.332 43 135 10 0.161 (0.014) 56.78 SM, southern Mediterranean (41°N) 168 9 (5) 0.267 (0.061) 2.407 86 26 (17) 0.340 (0.036) 11.844 70 175 50 0.173 (0.013) 82.41 View Large Table 1. Genetic (plastid DNA sequences, nuclear DNA sequences and AFLP data) features for Mediterranean samples of Myrtus communis in relation to the three detected genetic clusters (eastern, circum- and western Mediterranean) and with the distribution of samples to the north or south of the 41°N parallel: sample size (N), number of plastid DNA haplotypes (NH with number of private haplotypes in parentheses), unbiased diversity (uh with standard errors in parentheses), mean genetic distance (GD), number of nuclear DNA ribotypes (NR with number of private nrDNA ribotypes in parentheses), number of AFLP bands (NB), number of private AFLP bands (PB) and percentage of polymorphic loci (%P) Plastid DNA sequences Nuclear DNA sequences AFLP data N NH uh GD n NR uh GD n NB PB uh %P Cluster EM, eastern Mediterranean 60 2 0.051 (0.051) 0.463 26 6 0.042 (0.021) 6.108 10 110 12 0.145 (0.013) 43.72 Cluster CM, circum-Mediterranean 86 3 0.037 (0.028) 0.333 52 16 0.091 (0.030) 7.177 48 151 23 0.124 (0.011) 67.34 Cluster WM, western Mediterranean 155 4 0.055 (0.032) 0.491 94 15 0.041 (0.018) 3.831 55 132 19 0.130 (0.013) 54.27 All data 301 9 0.212 (0.057) / 172 37 0.294 (0.035) / 113 199 / 0.278 (0.018) / NM, northern Mediterranean (41°N) 133 4 (0) 0.085 (0.048) 0.768 86 19 (11) 0.186 (0.032) 8.332 43 135 10 0.161 (0.014) 56.78 SM, southern Mediterranean (41°N) 168 9 (5) 0.267 (0.061) 2.407 86 26 (17) 0.340 (0.036) 11.844 70 175 50 0.173 (0.013) 82.41 Plastid DNA sequences Nuclear DNA sequences AFLP data N NH uh GD n NR uh GD n NB PB uh %P Cluster EM, eastern Mediterranean 60 2 0.051 (0.051) 0.463 26 6 0.042 (0.021) 6.108 10 110 12 0.145 (0.013) 43.72 Cluster CM, circum-Mediterranean 86 3 0.037 (0.028) 0.333 52 16 0.091 (0.030) 7.177 48 151 23 0.124 (0.011) 67.34 Cluster WM, western Mediterranean 155 4 0.055 (0.032) 0.491 94 15 0.041 (0.018) 3.831 55 132 19 0.130 (0.013) 54.27 All data 301 9 0.212 (0.057) / 172 37 0.294 (0.035) / 113 199 / 0.278 (0.018) / NM, northern Mediterranean (41°N) 133 4 (0) 0.085 (0.048) 0.768 86 19 (11) 0.186 (0.032) 8.332 43 135 10 0.161 (0.014) 56.78 SM, southern Mediterranean (41°N) 168 9 (5) 0.267 (0.061) 2.407 86 26 (17) 0.340 (0.036) 11.844 70 175 50 0.173 (0.013) 82.41 View Large The plastid DNA network was characterized by 12 haplotypes (Figs 4A–C, S6A–C) geographically structured in the Mediterranean Basin and between the Macaronesian and Mediterranean regions. The simultaneous analysis of ETS and ITS nuclear DNA data revealed a similar genetic pattern to that obtained with plastid DNA data; 41 ribotypes were distributed through the three main genetic clusters (Figs 4A–D, S6B–D). Higher values of genetic singularity and diversity were detected south of the 41°N parallel (Fig. 3), as follows: nine plastid DNA haplotypes (vs. four to the north) of which five (vs. zero) are private; 26 ribotypes (vs. 19) of which 17 (vs. 11) are private; 175 AFLP bands (vs. 135) of which 50 (vs. ten) are private (Table 1). Higher genetic diversity in the southern Mediterranean was detected similarly for plastid DNA, nuclear DNA and AFLP molecular markers (see uh and %P in Table 1). Southern France and Corsica, which were the most well-sampled areas in the northern Mediterranean, appeared to have a low genetic originality (no private markers), even though France was a contact zone between the WM and CM genetic clusters (Figs 4, S6). Considering each of these clusters independently, we found that the areas harbouring the highest number of plastid DNA haplotypes were situated in North Africa (WM, Algeria and Tunisia), in Sicily (CM), and in the Levant and Cyprus (EM). The number of ribotypes was also higher in North Africa (WM), the Levant and Cyprus (EM) (Fig. S6). At the scale of the Mediterranean Basin, the contact zones between genetic clusters occurred longitudinally, and originated from WM and CM clusters, whereas plastid haplotypes and ribotypes of the EM cluster remained isolated (Fig. 4). The major plastid DNA contact zone was detected between EM and CM clusters in Turkey and the Levant, implying widespread haplotypes 4 and 12 (Fig. S6). The same pattern was observed for nuclear DNA markers (AFLP and sequences), except that the WM cluster was detected in Turkey and the Levant. In summary, the three molecular markers are consistent in indicating that the western and central parts of the Mediterranean could have been stronger sources of range expansions than the eastern parts. DISCUSSION Palaeoclimatic SDM and levels of genetic diversity suggest drastic changes in suitable areas for M. communis during the last 130000 years. In contrast to the phylogeography of Olea europaea (Besnard et al., 2013), past suitability modelling hindcasts the persistence of populations of M. communis in a few small long-term refugia. Projected past suitable areas for M. communis indicate a strong signature of the last glaciation on the potential geographical range of the species. On average, 97% of suitable areas were inferred to be lost between the LIG and the LGM (Fig. 2). The decrease in genetic diversity and singularity revealed by molecular markers (Table 1) is consistent with the idea that northern populations along the Mediterranean coastline were the most affected by LGM climate change, as also evidenced for Laurus nobilis (Rodríguez-Sánchez & Arroyo, 2008), Erica scoparia (Désamoré et al., 2012) and Chamaerops humilis (García-Castaño et al., 2014). Furthermore, dramatic consequences of glaciations with major bottlenecks, or even complete extinction in the Mediterranean, were suggested for some thermophilous plants, as Nerium oleander L. (Mateu-Andrés et al., 2015) and Ceratonia siliqua L. (Ramón-Laca & Mabberley, 2004). The strong range contraction inferred for M. communis during the LGM raises the issue of its long-term survival through the Pleistocene. Because fossil data are often too scarce for formally validating SDM, the identification of refugia using SDM approaches remains tentative (Ashcroft, 2010). According to SDM, the size of the long-term refugia for M. communis is highly variable with (1) two main poles of persistence in the western Mediterranean (Atlantic coast and Baetic–Rifan complex) and in the east (Levantine coast), and (2) small pockets along the mainland and island coastlines (Fig. 3). The reduced overlap of suitable areas among the three palaeomaps (LIG, LGM and Mid Holocene) leads to a few reduced long-term refugia compared to the size of the glacial refugia (Figs 2, 3). These results, rarely found in the literature, question the long-term refugia hypothesis and suggest the possibility of a regional persistence of M. communis by range shifts toward temporary refugia, where conditions became suitable. It must be noted that the SDM projections at the LIG, LGM and Mid Holocene do not aim to infer M. communis distribution at those times, but rather to identify potential suitable areas considering the environmental variables selected. Distribution and suitable areas would be consistent only under the assumption of unlimited dispersal of the species. At this juncture, we may examine our capacity to reject the possibility of long-term and in situ refugia for M. communis that could be cryptic for our methods. The role of local microrefugia may have played a more important role than expected, as the heterogeneity of micro-habitats may have allowed species to cope with environmental stress at finer spatial scales (Stewart & Lister, 2001; Serra-Diaz et al., 2015; Meineri & Hylander, 2017). However, it is worth considering studies conducted on M. communis at small scales. Studies in southern Spain are consistent with a low likelihood of long-term persistence of M. communis in small remnant patches of woodlands (González-Varo et al., 2015). The regeneration and genetic diversity of M. communis are generally lower in small populations and are highly sensitive to fragmentation (e.g. González-Varo, Arroyo & Aparicio, 2009; Nora, Albaladejo & Aparicio, 2015). Although places such as Corsica can harbour conditions for local microrefugia (Médail & Diadema, 2009; Gavin et al., 2014), local persistence is unlikely, perhaps due to the severe effects of glaciations (Conchon, 1986). Despite its abundance in the lowlands of Corsica up to 500 m elevation, and irrespective of which molecular marker was used, we found a lack of genetic differentiation, with for example only two plastid DNA haplotypes (8, 9) detected after an extensive sampling. These haplotypes are not private to Corsica and belong to the WM genetic cluster, occurring in Sardinia, south-east France, Italy and eastern Algeria (Figs 4, S6). Moreover, all 12 Corsican palaeoecological records for Myrtus correspond to the late Holocene (Reille, 1977, 1984, 1992; Vella, 2010). Reille (1992) suggested that M. communis is ‘the latest arrival in the thermophilous vegetation directly related to human action’. There are caveats to our questioning of long-term in situ refugia for M. communis. Specifically, reports that fleshy myrtle berries are mainly dispersed by frugivorous birds (Herrera, 1984; Traveset, Riera & Mas, 2001; González-Varo, 2010) are consistent with an active role for migration. Moreover, widespread plastid DNA haplotypes and ribotypes (Fig. S6) suggest episodes of frequent and/or long-distance migrations of M. communis in the western and central parts of the Mediterranean Basin. Our analyses revealed that migrations have been less pronounced in the eastern Mediterranean. Genetic data also suggest that migrations occurred rapidly, recently and/or over long distances to prevent molecular divergence between remote sites (e.g. the CM genetic cluster). The phylogeography of Smilax aspera L., a climbing plant with sclerophyllous leaves and red berries that are dispersed by birds, is also congruent with the pattern described for M. communis with a longitudinal rather than a latitudinal structure of genetic diversity (Chen et al., 2014). Such a longitudinal organization of phylogroups being more or less isolated, but rarely totally disjunct, thus constitutes one major phylogeographical trend of the circum-Mediterranean woody species (see the review by Nieto Feliner, 2014). As for M. communis, the phylogeography of Erica arborea L. (Désamoré et al., 2011), Arbutus unedo L. (Santiso et al., 2016) and Laurus nobilis L. (Rodríguez-Sánchez et al., 2009) support a stronger diversification within the western Mediterranean, and inferred eastward migrations for M. communis and A. unedo (see Santiso et al., 2016). CONCLUSIONS Documentation of past range dynamics of Mediterranean woody taxa supports the idea that glaciations led to significant changes in Mediterranean vegetation, such as in the abundance of most thermophilous plants. Our results are in accordance with recent findings indicating that the Mediterranean Basin was less stable from a palaeoenvironmental point of view, having experienced higher extinction rates of species and lineages compared to other Mediterranean-type ecosystems (Valente & Vargas, 2013; Cowling et al., 2015). Myrtus communis is representative of a Tertiary cold-sensitive lineage, and we propose here that the severe effects of Pleistocene glaciations were faced despite a relatively reduced role of long-term refugia. Myrtus communis seems to have survived the Quaternary glaciations by regional range shifts towards temporary refugia, a response which has, to our knowledge, not been considered in previous Mediterranean phylogeographical or SDM studies. This hypothesis represents an important issue regarding conservation or restoration of connectivity between populations of M. communis and, more generally, between fragments of woodlands to ensure the future resilience of Mediterranean thermophilous shrublands. SUPPORTING INFORMATION Additional Supporting Information may be found in the online version of this article at the publisher’s web-site. Figure S1. Palaeodistribution of Myrtus communis during the LIG and LGM. Figure S2. Presence/absence of Myrtus communis during the LIG and LGM. Figure S3. Palaeodistribution of Myrtus communis during the Mid Holocene. Figure S4. Presence/absence of Myrtus communis during the Mid Holocene. Figure S5. Analysis of STRUCTURE results for AFLP data. Figure S6. Spatial genetic structure with haplotype identifiers of Myrtus communis, based on plastid and nuclear DNA. Table S1. List of samples of Myrtus communis used for genetic analyses (plastid and nuclear sequencing, and AFLP genotyping). Table S2. AIC scores and number of parameters for fitted MaxEnt models. ACKNOWLEDGEMENTS This work was supported by the Provence-Alpes-Côte d’Azur Region, the Conservatoire Botanique National de Corse, and the Conservatoire Botanique National Méditerranéen de Porquerolles. Molecular biology analyses were carried out at the Molecular Biology Common Service of IMBE (Arbois) and the Molecular Biology Platform of INRA URFM (Avignon), with particular thanks to Bruno Fady and Anne Roig. We are grateful to Michelle Barbier-Leydet for her help with the European Pollen Database. The English text was edited by Kolo Wamba (Proediting, Redwood City, CA, USA). We thank the numerous contributors to the sampling of M. communis, cited in Migliore et al. (2012). Finally, we thank the Associate Editor Božo Frajman, Gonzalo Nieto Feliner and another reviewer for their very constructive comments, which helped to improve the manuscript. REFERENCES Ashcroft MB . 2010 . Identifying refugia from climate change . Journal of Biogeography 37 : 1407 – 1413 . Bandelt HJ , Forster P , Röhl A . 1999 . Median-joining networks for inferring intraspecific phylogenies . Molecular Biology and Evolution 16 : 37 – 48 . Google Scholar CrossRef Search ADS PubMed Bennett KD , Provan J . 2008 . What do we mean by ‘refugia’ ? Quaternary Science Reviews 27 : 2449 – 2455 . Google Scholar CrossRef Search ADS Besnard G , Khadari B , Navascués M , Fernández-Mazuecos M , El Bakkali A , Arrigo N , Baali-Cherif D , Brunini-Bronzini de Caraffa V , Santoni S , Vargas P , Savolainen V . 2013 . The complex history of the olive tree: from Late Quaternary diversification of Mediterranean lineages to primary domestication in the northern Levant . Proceedings of the Royal Society B: Biological Sciences 280 : 20122833 . Google Scholar CrossRef Search ADS Bonin A , Bellemain E , Bronken Eidesen P , Pompanon F , Brochmann C , Taberlet P . 2004 . How to track and assess genotyping errors in population genetics studies . Molecular Ecology 13 : 3261 – 3273 . Google Scholar CrossRef Search ADS PubMed Carnaval AC , Hickerson MJ , Haddad CF , Rodrigues MT , Moritz C . 2009 . Stability predicts genetic diversity in the Brazilian Atlantic forest hotspot . Science (New York, N.Y.) 323 : 785 – 789 . Google Scholar CrossRef Search ADS PubMed Chen C , Qi ZC , Xu XH , Koch MA , Jin X‐J , Fu C‐X , Qiu Y-X . 2014 . Understanding the formation of Mediterranean–African–Asian disjunctions: evidence for Miocene climate-driven vicariance and recent long-distance dispersal in the Tertiary relict Smilax aspera (Smilacaceae) . New Phytologist 204 : 43 – 255 . Google Scholar CrossRef Search ADS Conchon O . 1986 . Quaternary glaciations in Corsica . Quaternary Science Reviews 5 : 429 – 432 . Google Scholar CrossRef Search ADS Cowling RM , Potts AJ , Bradshaw PL , Colville J , Arianoutsou M , Ferrier S , Forest F , Fyllas NM , Hopper SD , Ojeda F , Proches S , Smith RJ , Rundel PW , Vassilakis E , Zutta BR . 2015 . Variation in plant diversity in Mediterranean-climate ecosystems: the role of climatic and topographical stability . Journal of Biogeography 42 : 552 – 564 . Google Scholar CrossRef Search ADS Désamoré A , Laenen B , Devos N , Popp M , González-Mancebo JM , Carine MA , Vanderpoorten A . 2011 . Out of Africa: north-westwards Pleistocene expansions of the heather Erica arborea . Journal of Biogeography 38 : 164 – 176 . Google Scholar CrossRef Search ADS Désamoré A , Laenen B , Gonzalez-Mancebo JM , Jaen Molina R , Bystriakova N , Martinez-Klimova E , Carine MA , Vanderpoorten A . 2012 . Inverted patterns of genetic diversity in continental and island populations of the heather Erica scoparia s.l . Journal of Biogeography 3 : 574 – 584 . Google Scholar CrossRef Search ADS Elith J , Graham CH , Anderson RP , Dudík M , Ferrier S , Guisan A , Hijmans RJ , Huettmann F , Leathwick JR , Lehmann A , Li J , Lohmann LG , Loiselle BA , Manion G , Moritz C , Nakamura M , Nakazawa Y , Overton JMcCM , Townsend Peterson A , Phillips SJ , Richardson K , Scachetti-Pereira R , Schapire RE , Soberón J , Williams S , Wisz MS , Zimmermann NE 2006. Novel methods improve prediction of species’ distributions from occurrence data . Ecography 29 : 129 – 151 . CrossRef Search ADS Evanno G , Regnaut S , Goudet J . 2005 . Detecting the number of clusters of individuals using the software STRUCTURE: a simulation study . Molecular Ecology 14 : 2611 – 2620 . Google Scholar CrossRef Search ADS PubMed Falush D , Stephens M , Pritchard JK . 2007 . Inference of population structure using multilocus genotype data: dominant markers and null alleles . Molecular Ecology Notes 7 : 574 – 578 . Google Scholar CrossRef Search ADS PubMed Fernández-Mazuecos M , Vargas P . 2010 . Ecological rather than geographical isolation dominates Quaternary formation of Mediterranean Cistus species . Molecular Ecology 19 : 1381 – 1395 . Google Scholar CrossRef Search ADS PubMed García-Castaño JL , Terrab A , Ortiz MÁ , Stuessy TF , Talavera S . 2014 . Patterns of phylogeography and vicariance of Chamaerops humilis L. (Palmae) . Turkish Journal of Botany 38 : 1132 – 1146 . Google Scholar CrossRef Search ADS Gavin DG , Fitzpatrick MC , Gugger PF , Heath KD , Rodríguez-Sánchez F , Dobrowski SZ , Hampe A , Hu FS , Ashcroft MB , Bartlein PJ , Blois JL , Carstens BC , Davis EB , de Lafontaine G , Edwards ME , Fernandez M , Henne PD , Herring EM , Holden ZA , Kong WS , Liu J , Magri D , Matzke NJ , McGlone MS , Saltré F , Stigall AL , Tsai YH , Williams JW . 2014 . Climate refugia: joint inference from fossil records, species distribution models and phylogeography . The New Phytologist 204 : 37 – 54 . Google Scholar CrossRef Search ADS PubMed Guzmán B , Fedriani JM , Delibes M , Vargas P . 2017 . The colonization history of the Mediterranean dwarf palm (Chamaerops humilis L., Palmae) . Tree Genetics & Genomes 13 : 24 . Google Scholar CrossRef Search ADS González-Varo JP . 2010 . Fragmentation, habitat composition and the dispersal/predation balance in interactions between the Mediterranean myrtle and avian frugivores . Ecography 33 : 185 – 197 . Google Scholar CrossRef Search ADS González-Varo JP , Albaladejo RG , Aizen MA , Arroyo J , Aparicio A . 2015 . Extinction debt of a common shrub in a fragmented landscape . Journal of Applied Ecology 52 : 580 – 589 . Google Scholar CrossRef Search ADS González-Varo JP , Arroyo J , Aparicio A . 2009 . Effects of fragmentation on pollinator assemblage, pollen limitation and seed production of Mediterranean myrtle (Myrtus communis) . Biological Conservation 142 : 1058 – 1065 . Google Scholar CrossRef Search ADS Hampe A , Rodríguez-Sánchez F , Dobrowski S , Hu FS , Gavin DG . 2013 . Climate refugia: from the Last Glacial Maximum to the twenty-first century . The New Phytologist 197 : 16 – 18 . Google Scholar CrossRef Search ADS PubMed Herrera CM . 1984 . A study of avian frugivores, bird-dispersed plants, and their interaction in Mediterranean scrublands . Ecological Monographs 54 : 1 – 23 . Google Scholar CrossRef Search ADS Hijmans RJ , Cameron SE , Parra JL , Jones PG , Jarvis A . 2005 . Very high resolution interpolated climate surfaces for global land areas . International Journal of Climatology 25 : 1965 – 1978 . Google Scholar CrossRef Search ADS Mateu-Andrés I , Ciurana MJ , Aguilella A , Boisset F , Guara M , Laguna E , Currás R , Ferrer P , Vela E , Puche MF , Pedrola-Monfort J . 2015 . Plastid DNA homogeneity in Celtis australis L. (Cannabaceae) and Nerium oleander L. (Apocynaceae) throughout the Mediterranean Basin . International Journal of Plant Sciences 176 : 421 – 432 . Google Scholar CrossRef Search ADS Médail F , Diadema K . 2009 . Glacial refugia influence plant diversity patterns in the Mediterranean Basin . Journal of Biogeography 36 : 1333 – 1345 . Google Scholar CrossRef Search ADS Meineri E , Hylander K . 2017 . Fine-grain, large-domain climate models based on climate station and comprehensive topographic information improve microrefugia detection . Ecography 40 : 1003 – 1013 . Google Scholar CrossRef Search ADS Migliore J , Baumel A , Juin M , et al. 2011 . Genetic diversity and structure of a Mediterranean endemic plant in Corsica (Mercurialis corsica, Euphorbiaceae) . Population Ecology 53 : 573 – 586 . Google Scholar CrossRef Search ADS Migliore J , Baumel A , Juin M , Médail F . 2012 . From Mediterranean shores to central Saharan mountains: key phylogeographical insights from the genus Myrtus . Journal of Biogeography 39 : 942 – 956 . Google Scholar CrossRef Search ADS Nieto Feliner G . 2014 . Patterns and processes in plant phylogeography in the Mediterranean Basin. A review . Perspectives in Plant Ecology, Evolution and Systematics 16 : 265 – 278 . Google Scholar CrossRef Search ADS Nikolić T . 2015 . Myrtus communis distribution in Croatia. Flora Croatica Database . Faculty of Science, University of Zagreb . Available at: http://hirc.botanic.hr/fcd (accessed January 2015 ). Nogués-Bravo D . 2009 . Predicting the past distribution of species climatic niches . Global Ecology and Biogeography 18 : 521 – 531 . Google Scholar CrossRef Search ADS Nora S , Albaladejo RG , Aparicio A . 2015 . Genetic variation and structure in the Mediterranean shrubs Myrtus communis and Pistacia lentiscus in different landscape contexts . Plant Biology (Stuttgart, Germany) 17 : 311 – 319 . Google Scholar CrossRef Search ADS PubMed O’Donnell MS , Ignizio DA . 2012 . Bioclimatic predictors for supporting ecological applications in the conterminous United States . U.S. Geological Survey Data Series 691 : 10 . Otto-Bliesner BL , Marshall SJ , Overpeck JT , Miller GH , Hu A . 2006 . Simulating Arctic climate warmth and icefield retreat in the last interglaciation . Science (New York, N.Y.) 311 : 1751 – 1753 . Google Scholar CrossRef Search ADS PubMed Peakall R , Smouse PE . 2012 . GenAlEx 6.5: genetic analysis in Excel. Population genetic software for teaching and research–an update . Bioinformatics 28 : 2537 – 2539 . Google Scholar CrossRef Search ADS PubMed Pearson RG , Raxworthy CJ , Nakamura M , Townsend Peterson A . 2007 . Predicting species distributions from small numbers of occurrence records: a test case using cryptic geckos in Madagascar . Journal of Biogeography 34 : 102 – 117 . Google Scholar CrossRef Search ADS Phillips SJ , Anderson RP , Schapire RE . 2006 . Maximum entropy modeling of species geographic distributions . Ecological Modelling 190 : 231 – 259 . Google Scholar CrossRef Search ADS Phillips SJ , Dudík M . 2008 . Modeling of species distributions with MaxEnt: new extensions and a comprehensive evaluation . Ecography 31 : 161 – 175 . Google Scholar CrossRef Search ADS Phillips SJ , Anderson RP , Dudík M , Schapire RE , Blair M . 2017 . Opening the black box: an open-source release of MaxEnt . Ecography 40 : 887 – 893 . Google Scholar CrossRef Search ADS Pirazzoli PA . 2005 . A review of possible eustatic, isostatic and tectonic contributions in eight late-Holocene relative sea-level histories from the Mediterranean area . Quaternary Science Reviews 24 : 1989 – 2001 . Google Scholar CrossRef Search ADS Postigo Mijarra JM , Barron E , Gomez Manzaneque F , Morla C . 2009 . Floristic changes in the Iberian Peninsula and Balearic Islands (south-west Europe) during the Cenozoic . Journal of Biogeography 36 : 2025 – 2043 . Google Scholar CrossRef Search ADS Pritchard JK , Stephens M , Donnelly P . 2000 . Inference of population structure using multilocus genotype data . Genetics 155 : 945 – 950 Google Scholar PubMed Quézel P , Médail F . 2003 . Ecologie et biogéographie des forêts du bassin méditerranéen . Amsterdam : Elsevier . Ramón-Laca L , Mabberley DJ . 2004 . The ecological status of the carob-tree (Ceratonia siliqua, Leguminosae) in the Mediterranean . Botanical Journal of the Linnean Society 144 : 431 – 436 . Google Scholar CrossRef Search ADS Reille M . 1977 . Quelques aspects de l’activité humaine en Corse durant le Subatlantique et ses conséquences sur la végétation. ‘Approche écologique de l’homme fossile’ . Supplément au Bulletin de l’AFEQ 47 : 329 – 342 . Reille M . 1984 . Origine de la végétation actuelle de la Corse sud-orientale; analyse pollinique de cinq marais côtiers . Pollen et Spores 26 : 43 – 60 . Reille M . 1992 . New pollen- analytical researches in Corsica: the problem of Quercus ilex L. and Erica arborea L., the origin of Pinus halepensis Miller forests . New Phytologist 122 : 359 – 378 . Google Scholar CrossRef Search ADS Rodríguez-Sánchez F , Arroyo J . 2008 . Reconstructing the demise of Tethyan plants: climate-driven range dynamics of Laurus since the Pliocene . Global Ecology and Biogeography 17 : 685 – 695 . Google Scholar CrossRef Search ADS Rodríguez-Sánchez F , Guzman B , Valido A , Vargas P , Arroyo J . 2009 . Late Neogene history of the laurel tree (Laurus L., Lauraceae) based on phylogeographical analyses of Mediterranean and Macaronesian populations . Journal of Biogeography 36 : 1270 – 1281 . Google Scholar CrossRef Search ADS Rodríguez-Sánchez F , Hampe A , Jordano P , Arroyo J . 2010 . Past tree range dynamics in the Iberian Peninsula inferred through phylogeography and palaeodistribution modelling: a review . Review of Palaeobotany and Palynology 162 : 507 – 521 . Google Scholar CrossRef Search ADS Rundel PW , Arroyo MTK , Cowling RM , Keeley JE , Lamont BB , Vargas P . 2016 . Mediterranean biomes: evolution of their vegetation, floras, and climate . Annual Review of Ecology, Evolution, and Systematics 47 : 383 – 407 . Google Scholar CrossRef Search ADS Santiso X , Lopez L , Retuerto R , Barreiro R . 2016 . Phylogeography of a widespread species: pre-glacial vicariance, refugia, occasional blocking straits and long-distance migrations . AoB PLANTS 8 : plw003. Serra-Diaz JM , Scheller RM , Syphard AD , Franklin J . 2015 . Disturbance and climate microrefugia mediate tree range shifts during climate change . Landscape Ecology 30 : 1039 – 1053 . Google Scholar CrossRef Search ADS Stewart JR , Lister AM . 2001 . Cryptic northern refugia and the origins of the modern biota . Trends in Ecology and Evolution 16 : 608 – 613 . Google Scholar CrossRef Search ADS Stewart JR , Lister AM , Barnes I , Dalén L . 2010 . Refugia revisited: individualistic responses of species in space and time . Proceedings of the Royal Society B Biological Sciences 277 : 661 – 671 . Google Scholar CrossRef Search ADS PubMed Svenning JC , Fløjgaard C , Marske KA , Nógues-Bravo D , Normand S . 2011 . Applications of species distribution modeling to paleobiology . Quaternary Science Reviews 30 : 2930 – 2947 . Google Scholar CrossRef Search ADS Thompson JD . 2005 . Plant evolution in the Mediterranean . New York : Oxford University Press . Google Scholar CrossRef Search ADS Traveset A , Riera N , Mas RE . 2001 . Ecology of fruit-colour polymorphism in Myrtus communis and differential effects of birds and mammals on seed germination and seedling growth . Journal of Ecology 89 : 749 – 760 . Google Scholar CrossRef Search ADS Valente LM , Vargas P . 2013 . Contrasting evolutionary hypotheses between two Mediterranean-climate floristic hotspots: the Cape of southern Africa and the Mediterranean Basin . Journal of Biogeography 40 : 2032 – 2046 . Google Scholar CrossRef Search ADS Valiente-Banuet A , Rumebe AV , Verdú M , Callaway RM . 2006 . Modern Quaternary plant lineages promote diversity through facilitation of ancient Tertiary lineages . Proceedings of the National Academy of Sciences of the United States of America 103 : 16812 – 16817 . Google Scholar CrossRef Search ADS PubMed Vargas P , Fernández-Mazuecos M , Heleno R . 2018 . Phylogenetic evidence for a Miocene origin of Mediterranean lineages: species diversity, reproductive traits and geographical isolation . Plant Biology (Stuttgart, Germany) 20 : 157 – 165 . Google Scholar CrossRef Search ADS PubMed Vella M-A . 2010 . Approches géomorphologique et géophysique des interactions sociétés/milieux en Corse au cours de l’Holocène . PhD Thesis, Université de Corse . Vendramin GG , Fady B , González-Martínez SC , Hu FS , Scotti I , Sebastiani F , Soto A , Petit RJ . 2008 . Genetically depauperate but widespread: the case of an emblematic Mediterranean pine . Evolution 62 : 680 – 688 . Google Scholar CrossRef Search ADS PubMed Vos P , Hogers R , Bleeker M , Reijans M , van de Lee T , Hornes M , Friters A , Pot J , Paleman J , Kuiper M , Zabeau M . 1995 . AFLP: a new technique for DNA fingerprinting . Nucleic Acids Research 23 : 4407 – 4414 . Google Scholar CrossRef Search ADS PubMed Waltari E , Hijmans RJ , Peterson AT , Nyári ÁS , Perkins SL , Guralnick RP . 2007 . Locating Pleistocene refugia: comparing phylogeographic and ecological niche model predictions . PLoS One 2 : e563 . Google Scholar CrossRef Search ADS PubMed Wu H , Guiot J , Brewer S , et al. 2007 . Climatic changes in Eurasia and Africa at the last glacial maximum and mid-Holocene: reconstruction from pollen data using inverse vegetation modelling . Climate Dynamics 29 : 211 – 229 . Google Scholar CrossRef Search ADS Yackulic CB , Chandler R , Zipkin EF , Zhengtang G . 2013 . Presence-only modelling using MaxEnt: when can we trust the inferences ? Methods in Ecology and Evolution 4 : 236 – 243 . Google Scholar CrossRef Search ADS © 2018 The Linnean Society of London, Botanical Journal of the Linnean Society This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/open_access/funder_policies/chorus/standard_publication_model)

Maia, Fabiano Rodrigo;Telles, Francismeire Jane;Goldenberg, Renato

doi: 10.1093/botlinnean/boy039pmid: N/A

Abstract Studies on the reproduction of plant species occurring in sites with different climatic conditions and landscapes can give insights into the dynamics and environmental pressures that might contribute to adaptive differentiation of individuals across the distribution range. To investigate any possible variation across sites, produced by biotic and/or abiotic factors, we selected individuals of T. hatschbachii, an endemic to grasslands on rocky outcrops with distinct climate and geological features [subtropical humid granitic outcrops (GO) vs. seasonally dry sandstone outcrops (SO)] and evaluated the phenology and reproductive biology across sites. We found asynchronous flowering periods between sites, but not asynchronous fruiting seasons. Microclimate variation (local temperature and precipitation) between sites (GO and SO) better explained divergence in phenology. Differences in abundance and frequency of floral visitors influenced intrinsic individual characteristics (pollen quality/quantity). Overall, individuals of T. hatschbachii at both sites are self–compatible, with differences in fruit set, seed set, germination rates and abundance of floral visitors, suggesting an effect from other factors, not just the reproductive system. The reproductive strategy of T. hatschbachii is comparable to that revealed in previous studies of plants in subtropical grassland areas, indicating a common pattern of adaptive reactions modulated by the intrinsic characteristics of individuals, the behaviour of floral visitors and landscape features where the individuals are located (spatial structure and vegetation age differences). abiotic drivers, biotic drivers, flowering asynchrony, reproductive strategies INTRODUCTION Biotic and abiotic factors influence the reproductive biology of plants (Arroyo, Armesto & Primack, 1985; Brito & Sazima, 2012; Maia, Malucelli & Varassin, 2013; Rech, 2014). Nevertheless, these factors are usually considered in isolation or studied under limited sampling circumstances in a single area (Goldenberg & Shepherd, 1998; Maia, Varassin & Goldenberg, 2016). Possible variation through space (different areas) and important information regarding plant occurrence, such as local conditions and geological history, are usually not taken into account (Thompson, 2005; Dart et al., 2012; Rech, 2014). Subtropical southern Brazil is characterized by a wide range of habitats with distinct climate and geomorphological conditions, resulting in a vegetation mosaic (grasslands/forests) formed during Palaeoclimate events in the Quaternary (Maack, 1981; Labiak, 2014). Sometimes the variation is greater even over short distances, e.g. between the moist and warm Atlantic forest along the coast and upper montane areas and the seasonally dry areas covered with Araucaria Juss. forests, grasslands and the southern limits of the Brazilian savannas (Maack, 1981; Labiak, 2014). Some areas on the top of mountain ranges with rocky outcrops are covered with isolated patches of relictual grassland vegetation. These grasslands are rather peculiar in the Atlantic forest, where they occur on granitic outcrops (GO; Fig. 1A) or in savannas, where they occur on sandstone outcrops (SO; Fig. 1B). The vegetation dynamics of these areas reflect distinct temporal processes (Behling and Negrelle, 2001; Behling, 2002). The expansion of the forest vegetation promoted the reduction and isolation of dry areas on GO (c. 11 000 years ago; Behling & Negrelle, 2001) with a more recent isolation of SO (c. 5000 years ago; Behling, 2002; Maia et al., 2017a). Figure 1. View largeDownload slide Habitats and flowers of Tibouchina hatschbachii: (AB) granitic outcrops (GO); (C–D) sandstone outcrops (SO); (E) dimorphic stamens, (I) antesepalous (larger) and (II) antepetalous (smaller); (F) flower being visited by Bombus morio, one of its effective pollinators. Figure 1. View largeDownload slide Habitats and flowers of Tibouchina hatschbachii: (AB) granitic outcrops (GO); (C–D) sandstone outcrops (SO); (E) dimorphic stamens, (I) antesepalous (larger) and (II) antepetalous (smaller); (F) flower being visited by Bombus morio, one of its effective pollinators. The existence of spatial and temporal differences on GO and SO could have promoted differential selective effects on the establishment and persistence of plants occurring in the two areas. One of these possible effects could be the emergence of differences in reproductive strategies, as demonstrated for other plant species from the Atlantic Forest (Brito & Sazima, 2012). If the former applies to species found in both environments, we would expect plants from GO and SO to show differences in their phenology, pollen dynamics and reproductive biology. Such variations would arise from differential pressures imposed by microclimate (local temperature and precipitation) and landscape conditions, besides the different ages of the environments (Behling & Negrelle, 2001; Behling, 2002; Labiak, 2014; Maia, 2017). One such species that is found in both areas is Tibouchina hatschbachii Wurdack (Melastomataceae: Melastomeae), an endemic shrub restricted to the subtropical grasslands of southern Brazil. Effects of spatial and temporal variations tend to be strong for species with specialized pollination systems (Renner, 1993; Goldenberg & Shepherd, 1998; Price et al., 2005). Most species of Melastomataceae only offer pollen as a reward to visitors (for exceptions, see Varassin, Penneys & Michelangeli, 2008; Dellinger et al., 2014; Maia et al., 2016) and the pollen grains in this family are kept in poricidal anthers, restricting the effective pollinators to those capable of removing the reward through anther sonication (also known as buzz pollination, Buchmann, 1983). Floral visitation rates and insect abundance can be affected by local conditions across areas. Both areas are at high elevations (800–1930 m a.s.l.), where activity of floral visitors is believed to be scarce due to low temperatures and strong winds (Arroyo et al., 1985; Freitas & Sazima, 2006). Under these conditions, we expect a reduction in the number of available pollinators and, consequently, in pollen grains transfer between conspecific flowers (Arroyo et al., 1985; Brito & Sazima, 2012). In addition, GO are more isolated than SO, which could constrain genetic flow between sites (Maia, 2017). The aims of the present study were to investigate the existence of divergence in reproductive strategies (phenology and reproductive biology) and in pollinator species richness and abundance of T. hatschbachii, due to possible differences in environmental conditions across the occurrence area of the species. We selected individuals from sites on GO and SO to compare attributes related to pre- and post-fertilization events, such as reproductive phenology, pollen dynamics, floral visitors, fruit and seed set and viability of the progeny. Specifically, we aimed to answer the following questions. (1) Are there differences regarding the reproductive phenology of individuals across sites? (2) Are there differences in pollen dynamics (pollen viability and availability and pollen load deposited on stigmas) between the two sites? (3) Is there any variation in the reproductive biology between sites? (4) Are there differences in fruit set, seed set and germination of seeds produced by manual reproductive treatments between sites? (5) How abundant and diverse are flower visitors at these sites? These questions will allow us to test the hypothesis that individuals in GO and SO sites show differences in reproductive strategies (phenology and reproductive biology), species richness and abundance of pollinators, due to climate differences, structure of the subtropical landscape and different ages between these environments. If our hypothesis holds, we expect: (1) differences in phenological patterns between sites, generated by the transition from a subtropical humid (GO) to a seasonally dry (SO) climate; (2) pollen limitation in GO, due to the higher environmental isolation when compared to SO and (3) reduced progeny fitness on SO, as a result of inbreeding depression, due to the more recent occupation of the area. MATERIAL AND METHODS Model species Individuals of T. hatschbachii from GO and SO sites occur in vegetation patches with different degrees of isolation (Fig. 1A–D). Plants offer solely pollen, concealed in poricidal anthers, as a reward for their pollinators that must be capable of removing it through sonication (Fig. 1F). The flowers are pentamerous, hermaphroditic, herkogamous, with ten dimorphic stamens in two sets differing in size. The five antesepalous stamens are larger than the five antepetalous ones (Fig. 1E). This dimorphism is sometimes followed by differences in the quantity and quality of pollen grains (functional pollen dimorphism), a strategy known as division of labour (Darwin, 1862) and that has been functionally demonstrated for some species of Melastomataceae (Luo, Zhang & Renner, 2008, Luo, Gu & Zhang, 2009). The anthesis lasts approximately one day and plants in GO and SO produce on average 21 and 27 flowers/day, respectively. Fruits are capsular and, when mature, they release many minute, autochorous seeds (seeds self-dispersed by gravity; Meyer, Guimarães & Goldenberg, 2009). Vouchers (T. hatschbachii in GO – UEC 188776 and SO – UPCB 75240) were deposited in the herbarium of the Universidade Estadual de Campinas (UEC). Study sites, climate and vegetation We studied T. hatschbachii in two sites 172.65 km apart from November 2013 to November 2014. The sites were selected in order to account for the extreme environmental variation, considering the strong differences in climate between areas (Maia et al., 2017a). The first site is located at 1100 m a.s.l, on the eastern slope of the Serra do Mar, in the Environmental Protection Area (Portuguese acronym: APA) Pico Paraná (25° 15’ 27.95’’S, 48° 45’ 45.54’’W), in the Antonina municipality, Paraná state, Brazil. The area is mainly covered with dense Atlantic rain forest, but the individuals of T. hatschbachii were sampled on isolated GO (Fig. 1A, B), which are covered with grasslands and scattered shrubs. The climate in this GO region is subtropical humid, without a dry season or frosts (Alvares et al., 2014). During the study, the average annual temperature was 21.9 °C and the monthly precipitation ranged between 60.6 mm and 458.0 mm, with greater intensity in February (Fig. S1; SIMEPAR – Meteorological System of Paraná – Antonina). The second site is located at 878 m a.s.l., in grasslands on SO (Fig. 1C, D), interspersed with patches of Araucaria Forest (Maack, 1981). This area belongs to the Guartelá State Park (24° 33’ 16.79’’S, 50° 13’ 58.26’’W), which lies in the APA Escarpa Devoniana, Tibagi municipality, Paraná state, Brazil. The climate in the area is seasonally dry, with a well-defined dry season and frequent frosts (Alvares et al., 2014). During the study, the average annual temperature was 19.6 °C and the monthly precipitation ranged between 29.8 mm and 253.0 mm, with greater intensity in January (Fig. S1; SIMEPAR–Telemaco Borba meteorological station). Reproductive phenology To assess the flowering and fruiting phenophases, we regularly visited the study sites at 30-day intervals. At each site, we randomly chose 50 individuals and monitored the phenology by recording their flowering period (the presence or absence of flowers) and fruiting period (the presence or absence of fruits), without discriminating between developmental stages (Tannus, Assis & Morellato, 2006; Morellato, Camargo & Gressler, 2013). Pollen dynamics Due to the presence of stamen dimorphism, the analyses of pollen dynamics of T. hatschbachii were performed considering both anther types. To evaluate pollen viability within and between sites, we selected 20 individuals at each site and isolated one flower at pre-anthesis phase per individual. From each previous isolated flower, we randomly selected two anthers during anthesis, one of each size (small and large), totalling 40 anthers per area. Anthers were fixed in 70% formalin–acetic acid–alcohol (FAA) solution. Pollen grains were stained with acetic carmine and counted under a light microscope (Kearns & Inouye, 1993). Acetic carmine evaluates pollen malformation and we used this criterion as an indicator of pollen viability (Hoffman & Varassin 2011). At least 200 pollen grains were counted per sample, following Maia et al. (2016). To evaluate pollen removal rates during anthesis, we collected open flowers at both sites at different times of the day (07:00 h, 09:00 h, 11:00 h, 13:00 h and 15:00 h), following Dafni, Kevan & Husband (2005). Every two hours, we randomly selected ten flowers at each site (one per individual) and removed one large and one small anther from each flower (50 flowers and 100 anthers per site). These anthers were individually stored in Eppendorf tubes in 1.5 ml 70 % ethanol. Pollen grains were released inside the Eppendorf tube by macerating the anther. After that, we stirred the solution in a vortex for 30 seconds and transferred 100 µl to a glass slide. We counted all pollen grains in the slide under a light microscope and calculated the ratio to the initial volume (Kearns & Inouye, 1993). Once the number of grains per anther was estimated, this value was multiplied by five (number of anthers of each type in one flower) to estimate the number of grains per anther type (large/small) in each flower. We calculated the amount of pollen for both anthers at both sites for each observed time. To estimate the pollen load on the stigmatic surface, we used the same flowers and individuals from GO and SO as described above, at the same time intervals. We collected and accommodated each sampled stigma (50 stigmas per site; ten per time interval) in Eppendorf tubes with 1.5 ml 70% ethanol. Pollen load was estimated under a light microscope using the pollen load index, which considers five categories from 0 to 4 (0%; 1–10%; 11–25%; 26–50% and 51–100%), according to the percentage of the stigmatic area covered with pollen grains (see Brito & Sazima, 2012). For each interval of time, we calculated the average index value for each site. Reproductive system, production and germination of seeds To determine the reproductive system at each site, we performed the following controlled pollination experiments using isolated pre-anthetic flowers (Radford et al., 1974): openpollination (OP); hand-self–pollination (HSP); cross-pollination between close (1–2 m away; hereafter referred to as CCP) and distant individuals (100–200 m away; hereafter referred to as DCP); spontaneous self-pollination (SSP) and apomixis (AP). We used 50 individuals per site, each of them receiving all treatments. Once the treatments were finished, the marked flowers were bagged (except for the OP treatment) and observed until fruit ripening, to then establish the fruit set (calculated as the ratio between flowers resulting in fully developed fruits in each treatment). To assess the pollen tube growth on the stigmatic surface, we performed self- and cross-pollination treatments using ten flowers (five per treatment) from five individuals at each site, following a modified version of Martin’s protocol (1959; modification described in Maia et al., 2016). Flowers were collected and fixed in 70% FAA, 24, 48 and 72 h after treatments. To determine the seed set, we randomly collected between seven to ten almost ripe, but still closed fruits (depending on availability in the field), resulting from the OP, HSP, CCP and DCP treatments on GO and SO. The total number of well-formed seeds from sampled fruits (used as the measure of the seed set) was established considering the presence of embryos. To facilitate visualization and quantification of embryos, we treated seeds with 5% sodium hypochlorite solution for two hours. Subsequent to that, we washed and immersed the seeds in distilled water for 40 minutes (modified from Sofiatti et al., 2008), we then placed the seeds in a Petri dish and counted the visible embryos. The second step in the analysis of seed set was the germination ratio. From those seeds presenting embryo, we selected 50 from each of the four reproductive treatments (OP, HSP, CCP and DCP), totalling 200 seeds, and placed them on moistened filter paper in germination boxes. After 30 days, we estimated the percentage of germinated seeds for each reproductive treatment. We used radicle protrusion through the seed coat as our criterion for germination. A cross-pollination treatment between GO and SO (CPBS) was carried out to investigate the existence of possible reproductive barriers between individuals. Twenty-five individuals from GO were the pollen donors and 25 individuals from SO were the recipients. Due to logistical limitations, we did not perform cross-pollination treatments inverting donors (SO) and recipients (GO). The anthers in pre-anthetic flowers used during treatments were stored in Eppendorf tubes in a freezer at an average temperature of –10 ± 1 °C to avoid pollen viability loss (Brito et al., 2010). Floral visitors We determined the species richness and abundance through focal observations between 06:00 h and 16:00 h (five consecutive days for each site). According to their behaviour, visitors were considered to be pollinators when they legitimately visited the flowers by vibrating the full set of anthers (Dafni et al., 2005), to be robbers when they chewed or cut the anthers causing damage or even to be thieves when they removed pollen by vibrating individual anthers or passively collected it from petals, without damaging anthers (Inouye, 1980). Visitors were collected with an entomological net, sacrificed in ethyl acetate and sent to specialists for taxonomic identification. The insects are deposited at the entomological collection ‘Pe. Jesus S. Moure’ Entomology Museum, (DZUP–UFPR), Paraná state. Statistical analysis To assess variation in the reproductive phenophases (measured as the period of occurrence, peak and duration) between GO and SO, we performed circular statistical analysis for phenological events using the software ORIANA 4.0 (Kovach, 2016). For the analyses, months were converted into angles at 30° intervals, where January corresponds to 15°, February to 45°, March to 75° and so forth (Morellato, Alberti & Hudson, 2010). We calculated the average angle (u), which corresponds to the mean date of the phenophase occurrence, the length of the average vector (r), a measure of the concentration around the average angle that indicates whether the phenophase is concentrated around a peak and whether there is synchronism among individuals. The vector r ranges from 0 (no seasonality) to 1 (all individuals reproduce synchronically; Morellato et al., 2000). In addition, we performed the Rayleigh test (Z; Zar, 2010), which determines the significance of the average angle for all unimodal distributions; when the average angle is significant, the pattern is considered seasonal and it corresponds to the average date of the year concentrating the phenological events (Morellato et al., 2000, Morellato et al., 2010). To test any possible effect of abiotic factors on the phenophases of GO and SO, we evaluated the relationship between phenophases and the average temperature and precipitation during the period of the study. For the analysis, we used the number of individuals in each phenophase/month and the available climate data (local temperature and precipitation) associated with the month to set generalized linear models (GLM) with a Poisson distribution. When overdispersion was detected, we corrected the standard error using a quasi-generalized linear model (Zuur, Leno & Elphick, 2009). Finally, we evaluated phenophase overlapping among sites using the Spearman rank correlation analysis (rs; Zar, 2010). To determine the reproductive system of the individuals at each site, we used the self-incompatibility index (ISI sensuZapata & Arroyo, 1978), calculated as the ratio between the fruit set after self and cross-pollination treatments (Lloyd & Schoen, 1992). According to this index, values below 0.2 indicate a self-incompatible system (Zapata & Arroyo, 1978). To explore the results of pollen dynamics (pollen viability, pollen removal rates and pollen load deposited on stigma) and reproductive system experiments (pollination treatments, fruit set, seed set and seed germination), we fitted our data to different models according to the distribution of the response variables. The analyses were divided in two sets. In the first set, we explored any possible differences on pollen viability and availability, considering the anther type (large and small), sites (GO and SO) and the number of pollen grains remaining after interval periods of two hours (from 7:00 h to 15:00 h). Pollen viability differences between anther types and sites were analysed by means of GLM with binomial distribution and logit link function. To determine differences on pollen removal rates after visitation in each sampled interval (from 07:00 h to 15:00 h) and sites, we used a GLM with a Poisson distribution. In the second set of analyses, we explored differences on fruit set, seed set and germination rate, according to the pollination treatments and study sites. For fruit set, we used a generalized linear mixed model (GLMM), assuming a binomial distribution (with N = number of flowers and P = probability of a flower to produce fruit), logit link function and the plant individual as the random term. For the analyses on the seed set (number of seeds with embryo) and seed germination rate, we used GLM with Poisson and binomial distributions, respectively. When necessary, post hoc tests were performed using the glht function (multicomp package, Hothorn et al., 2016) with a Tukey contrast for comparisons. All statistical analyses were performed using R software, version 3.2.5 (RDCT, 2016) and vegan (Oksanen et al., 2013), car (Fox and Weisberg, 2011), lme4 (Bates et al., 2015) and nlme (Pinheiro et al., 2016) packages. RESULTS Reproductive phenology Phenophases were seasonal and asynchronous between sites (Fig. 2; Table 1). The flowering season period differed between sites, being a little longer in SO than in GO (Fig. 2A, B; rs = 0.51, P < 0.05). However, the fruiting seasons were extensive and similar across sites (Fig. 2C, D; rs = 0.95, P < 0.001). Average dates for flowering peaks varied between sites (15th January for GO and 5th March for SO, Fig. 2A, B), but were similar for fruiting peaks (30th March, two months after the flowering peak, in GO and 8th April, one month after the flowering peak, in SO; Fig. 2C, D). Within sites, individuals showed high mean r values in both phenophases, which reflects a high synchrony between individuals at the same site, although those differences were not so pronounced for the fruiting season (Fig. 2; Table 1). Figure 2. View largeDownload slide Phenological events of individuals (N = 50 in each area). Individuals flowering on (A) granitic outcrops (GO) and (B) sandstone outcrops (SO); individuals fruiting on (C) GO and (D) SO. Concentric circles indicate the number of individuals in different phenophases, ranging from 0 (none) to 100 (all individuals). Outermost circle (continuous line) represents the year and months are indicated at 30° intervals. Arrows indicate the mean date of the phenological event considering the number of individuals. The length of the arrows represents the value of vector r, ranging from 0 to 1, which is a measure of the concentration of the phenophase around the mean date (seasonality degree; P < 0.0001; see Method section for details). Figure 2. View largeDownload slide Phenological events of individuals (N = 50 in each area). Individuals flowering on (A) granitic outcrops (GO) and (B) sandstone outcrops (SO); individuals fruiting on (C) GO and (D) SO. Concentric circles indicate the number of individuals in different phenophases, ranging from 0 (none) to 100 (all individuals). Outermost circle (continuous line) represents the year and months are indicated at 30° intervals. Arrows indicate the mean date of the phenological event considering the number of individuals. The length of the arrows represents the value of vector r, ranging from 0 to 1, which is a measure of the concentration of the phenophase around the mean date (seasonality degree; P < 0.0001; see Method section for details). Table 1. Results of the phenological analysis of Tibouchina hatschbachii (50 individuals in each site). (N) = the total number of observations during the year; (u) = mean angles, all significant according to the Rayleigh test (P < 0.0001). GO = site on granitic outcrops; SO = site on sandstone outcrops. Phenophase Circular analysis Site GO SO Flowering Observations (N) 78 126 Mean angle (u) ± SD 15.571° ± 11.644° 64.273° ± 24.267° Mean date 15th January 5th March Length of mean vector (r) 0.953 0.914 Rayleigh test (Z) 70.903 105.309 Rayleigh test (P) < 0.0001 < 0.0001 Fruiting Observations (N) 197 226 Mean angle (u) ± SD 89.264° ± 39.905° 98.46° ± 42.033° Mean date 30th March 8th April Length of mean vector (r) 0.785 0.764 Rayleigh test (Z) 121.283 131.939 Rayleigh Test (P) < 0.0001 < 0.0001 Phenophase Circular analysis Site GO SO Flowering Observations (N) 78 126 Mean angle (u) ± SD 15.571° ± 11.644° 64.273° ± 24.267° Mean date 15th January 5th March Length of mean vector (r) 0.953 0.914 Rayleigh test (Z) 70.903 105.309 Rayleigh test (P) < 0.0001 < 0.0001 Fruiting Observations (N) 197 226 Mean angle (u) ± SD 89.264° ± 39.905° 98.46° ± 42.033° Mean date 30th March 8th April Length of mean vector (r) 0.785 0.764 Rayleigh test (Z) 121.283 131.939 Rayleigh Test (P) < 0.0001 < 0.0001 View Large Table 1. Results of the phenological analysis of Tibouchina hatschbachii (50 individuals in each site). (N) = the total number of observations during the year; (u) = mean angles, all significant according to the Rayleigh test (P < 0.0001). GO = site on granitic outcrops; SO = site on sandstone outcrops. Phenophase Circular analysis Site GO SO Flowering Observations (N) 78 126 Mean angle (u) ± SD 15.571° ± 11.644° 64.273° ± 24.267° Mean date 15th January 5th March Length of mean vector (r) 0.953 0.914 Rayleigh test (Z) 70.903 105.309 Rayleigh test (P) < 0.0001 < 0.0001 Fruiting Observations (N) 197 226 Mean angle (u) ± SD 89.264° ± 39.905° 98.46° ± 42.033° Mean date 30th March 8th April Length of mean vector (r) 0.785 0.764 Rayleigh test (Z) 121.283 131.939 Rayleigh Test (P) < 0.0001 < 0.0001 Phenophase Circular analysis Site GO SO Flowering Observations (N) 78 126 Mean angle (u) ± SD 15.571° ± 11.644° 64.273° ± 24.267° Mean date 15th January 5th March Length of mean vector (r) 0.953 0.914 Rayleigh test (Z) 70.903 105.309 Rayleigh test (P) < 0.0001 < 0.0001 Fruiting Observations (N) 197 226 Mean angle (u) ± SD 89.264° ± 39.905° 98.46° ± 42.033° Mean date 30th March 8th April Length of mean vector (r) 0.785 0.764 Rayleigh test (Z) 121.283 131.939 Rayleigh Test (P) < 0.0001 < 0.0001 View Large Flowering and fruiting were influenced by climate conditions at both sites. Flowering response was influenced by temperature in SO (χ2 = 4.74; d.f. = 1; P = 0.03) and temperature and precipitation increased the flowering response of individuals in GO (χ2 temperature = 129.27; χ2 precipitation = 15.90; d.f. = 1; P < 0.001). At both sites, the number of fruiting individuals was only influenced by precipitation (χ2 > 13.94; d.f. = 1; P < 0.001) and a subsequent increase in precipitation favoured the number of individuals producing fruits, although fruit maturation follows the reduction of precipitation in the following months. Pollen dynamics Pollen viability differed between sites (χ2 = 512.78; d.f. = 1; P < 0.001), with individuals from SO presenting lower viability in both large and small anthers (Fig. 3; large anthers difference between sites: χ2 = 206.64; d.f. = 1; P < 0.001; small anthers difference between sites: χ2 = 11.23; d.f. = 1; P < 0.001). Pollen viability differed between anther types (large and small) at SO, being higher in large anthers (Fig. 3; χ2 = 5.13; d.f. = 1; P = 0.02). There was no difference in pollen viability between anther types at GO (Fig. 3; χ2 = 1.16; d.f. = 1; P = 0.28). Figure 3. View largeDownload slide Pollen viability of large (LA) and small (SA) anthers in the two sites (N = 25 individuals for site) of Tibouchina hatschbachii. Each point indicates the mean (± SE) proportion of pollen viability. Figure 3. View largeDownload slide Pollen viability of large (LA) and small (SA) anthers in the two sites (N = 25 individuals for site) of Tibouchina hatschbachii. Each point indicates the mean (± SE) proportion of pollen viability. Overall, sites differed in the amount of pollen available during each interval, reflecting the activity of floral visitors at each site. We found more pollen available in plants from GO than from SO (χ2 = 366; d.f. = 1; P < 0.0001; Fig. 4). On average, larger anthers from both sites had more pollen than smaller anthers (χ2 = 652.92; d.f. = 1; P < 0.001; Fig. 4). There was a difference in pollen removal rates on the first day of the anthesis, which was more intense in SO (χ2 = 2.733.367; d.f. = 4; P < 0.001; Fig. 4). Pollen removal by visitors was more intense in the first morning hours (GO: χ2 = 1.239.014, d.f. = 4, P < 0.001, Fig. 4A; SO: χ2 = 1.618.294; d.f. = 4; P < 0.001, Fig. 4B) and the removal differed between anther types within sites (GO: χ2 = 408.313, d.f. = 1, P < 0.001, Fig. 4A; SO: χ2 = 253.716; d.f. = 1; P < 0.001, Fig. 4B), with visitors mainly removing pollen from small anthers. The pollen load on the stigmatic surface in individuals from GO reached stigmatic saturation later (13:00 h; Fig. 5) than individuals from SO (09:00 h; Fig. 5). Figure 4. View largeDownload slide Pollen removal rates of large and small anthers on (A) granitic outcrops (GO) and (B) sandstone outcrops (SO), respectively (N = 10 individuals for site) of Tibouchina hatschbachii. Each point indicates the mean (± SE) quantity of pollen remaining from large and small anthers at different times of the day. Figure 4. View largeDownload slide Pollen removal rates of large and small anthers on (A) granitic outcrops (GO) and (B) sandstone outcrops (SO), respectively (N = 10 individuals for site) of Tibouchina hatschbachii. Each point indicates the mean (± SE) quantity of pollen remaining from large and small anthers at different times of the day. Figure 5. View largeDownload slide Mean pollen load index (± SE) on the stigmas during the first day of anthesis in the two sites (N = 10 individuals for site) of Tibouchina hatschbachii. Figure 5. View largeDownload slide Mean pollen load index (± SE) on the stigmas during the first day of anthesis in the two sites (N = 10 individuals for site) of Tibouchina hatschbachii. Reproductive system, production and germination of seeds Individuals from the two sites are self-compatible and do not produce fruits by autonomous apomixis [ISI in GO (1.53) and SO (0.92); Fig. 6, AP treatment]. Although the plants at SO tend to produce more fruits than at GO (coefficient of the model = 0.187; Fig. 6), the total fruit set was not significantly different between sites (χ2 = 0.93; d.f. = 1; P > 0.001; Fig. 6). Nonetheless, there were differences in the fruit set among pollination treatments (χ2 = 47.29; d.f. = 5; P < 0.001; Fig. 6). For both GO and SO, the treatments open pollination (OP), hand self-pollination (HSP) and cross-pollination between close (CCP) and distant individuals (DCP) equally contributed to fruit set and differ from spontaneous self-pollination (SSP) treatment (Fig. 6). The SSP treatment produced few fruits and none completed development at either site (Fig. 6). Pollen tubes developed similarly on the stigmatic surface of samples from both sites, irrespective of the treatment, indicating absence of incompatibility reaction sites (Fig. S2). The fruit set from inter-site crosses (SO × GO) was low [cross-pollination treatment between sites (CPBS) = 0.16 fruits; N = 25 flowers], with all fruits aborting before full development. Figure 6. View largeDownload slide Fruit set (%± SE) in the two sites (N = 50 individuals for site) of Tibouchina hatschbachii after treatments: open-pollination (OP); hand self-pollination (HSP); close cross-pollination (CCP); distant cross-pollination (DCP); spontaneous self-pollination (SSP) and apomixis (AP). Treatments followed by the same letter do not statistically differ (P > 0.001). HSP and AP treatments present the same values (52% and 0%, respectively) between areas, superimposing the data on the figure. Figure 6. View largeDownload slide Fruit set (%± SE) in the two sites (N = 50 individuals for site) of Tibouchina hatschbachii after treatments: open-pollination (OP); hand self-pollination (HSP); close cross-pollination (CCP); distant cross-pollination (DCP); spontaneous self-pollination (SSP) and apomixis (AP). Treatments followed by the same letter do not statistically differ (P > 0.001). HSP and AP treatments present the same values (52% and 0%, respectively) between areas, superimposing the data on the figure. The total number of well-formed seeds (seed set) was lower at SO than at GO (χ2 = 7346.1; d.f. = 1; P < 0.001; Fig. 7). Seed set differed between pollination treatments across and within sites (χ2 = 1242.5; d.f. = 3; P < 0.001; Fig. 7). At SO, seed production was lower for the HSP treatment, whereas the CCP treatment at GO resulted in a higher number of well-formed seeds. Since no fruits resulting from the SSP treatment developed, seed set could not be evaluated (Fig. 7). Figure 7. View largeDownload slide Number of seeds with embryo (mean ± SE) for each site after treatments: open-pollination (OP); hand self-pollination (HSP); close cross-pollination (CCP) and distant cross-pollination (DCP). Same letter indicates the means are not statistically significant (P > 0.001). Figure 7. View largeDownload slide Number of seeds with embryo (mean ± SE) for each site after treatments: open-pollination (OP); hand self-pollination (HSP); close cross-pollination (CCP) and distant cross-pollination (DCP). Same letter indicates the means are not statistically significant (P > 0.001). Seed germination differed between sites, always being lower at SO (χ2 = 34.47; d.f. = 1; P < 0.001; Fig. 8) and between treatments (χ2 = 23.47; d.f. = 3; P < 0.001; Fig. 8). Germinated seed ratio among treatments varied between GO and SO (GO: χ2 = 19.42; d.f. = 3; P < 0.001; SO: χ2 = 22.791; d.f. = 3; P < 0.001; Fig. 7). Within sites, the hand self-pollinated (HSP) treatment presented the lowest germinated seed ratio in SO, significantly differing from the other treatments (Fig. 7). For GO, seed germination ratio was smaller in the close cross pollination (CCP) treatment, statistically differing from HSP and distant cross pollination (DCP) treatments and marginally from open pollination (OP) treatment (Fig. 7). Figure 8. View largeDownload slide Seed germination (mean ± SE) for each site after treatments: open-pollination (OP); hand self-pollination (HSP); close cross-pollination (CCP) and distant cross-pollination (DCP). Same letter indicates the means are not statistically significant (P > 0.001). Figure 8. View largeDownload slide Seed germination (mean ± SE) for each site after treatments: open-pollination (OP); hand self-pollination (HSP); close cross-pollination (CCP) and distant cross-pollination (DCP). Same letter indicates the means are not statistically significant (P > 0.001). Floral visitors Large bees are the pollinators of T. hatschbachii, whereas small bees act as robbers, accessing pollen grains by cutting the anthers or thieves by collecting pollen grains that fall on the petals after a legitimate visit or by indiscriminately buzzing single anthers (Table S1). There was no difference in the richness of floral visitors between sites (15 species for each site; Table 2). However, the abundance was higher in SO (82 visits at vs 48 at GO; Table 2). Overall, most visitors were illegitimate, acting as pollen thieves or robbers (Table 2), with Trigona spinipes being the most frequent visitor from this category at both sites. The most frequent pollinators were Bombus morio at GO (N = 16; 33% of visits) and Bombus pauloensis at SO (N = 32; 67% of visits; Table S1; Fig. 1D). The remaining pollinators, despite their contribution to pollination success, were less frequent (one to three visits; Table S1). The peak of visitor activity was between 07:00 and 10:00 h at both sites, with pollen thieves and robbers being frequent throughout anthesis at both sites. Table 2. Richness (number of species) and abundance (number of visits) of flower visitors of Tibouchina hatschbachii from January to March/2014, on granitic outcrops (GO) and sandstone outcrops (SO). 50 hours of observation per site. Flower visitors Richness (frequency) Abundance (frequency) GO SO GO SO Pollinator 6 (0.40) 4 (0.27) 16 (0.33) 19 (0.23) Thief/Robber 9 (0.60) 11 (0.73) 32 (0.67) 63 (0.77) TOTAL 15 15 48 82 Flower visitors Richness (frequency) Abundance (frequency) GO SO GO SO Pollinator 6 (0.40) 4 (0.27) 16 (0.33) 19 (0.23) Thief/Robber 9 (0.60) 11 (0.73) 32 (0.67) 63 (0.77) TOTAL 15 15 48 82 View Large Table 2. Richness (number of species) and abundance (number of visits) of flower visitors of Tibouchina hatschbachii from January to March/2014, on granitic outcrops (GO) and sandstone outcrops (SO). 50 hours of observation per site. Flower visitors Richness (frequency) Abundance (frequency) GO SO GO SO Pollinator 6 (0.40) 4 (0.27) 16 (0.33) 19 (0.23) Thief/Robber 9 (0.60) 11 (0.73) 32 (0.67) 63 (0.77) TOTAL 15 15 48 82 Flower visitors Richness (frequency) Abundance (frequency) GO SO GO SO Pollinator 6 (0.40) 4 (0.27) 16 (0.33) 19 (0.23) Thief/Robber 9 (0.60) 11 (0.73) 32 (0.67) 63 (0.77) TOTAL 15 15 48 82 View Large DISCUSSION Despite the absence of variation in the reproductive system of T. hatschbachii among GO and SO areas, we noticed relevant differences in the reproductive strategies, related to the climate and dynamic of floral visitors between areas. This indicates that other factors tightly connected to fitness, such as plant adaptation strategies, with individuals responding to local environmental conditions, are possibly affecting the occurrence of T. hatschbachii in subtropical grasslands in South America. Variation in reproductive phenology Flowering and fruiting events were well-defined at GO and SO, but flowering phenology differed between them. Variations in microclimate (local precipitation and temperature) between environments seem to influence the phenology of individuals in the studied sites, determining the flowering and fruiting periods and peaks. The macroclimate in both areas is clearly seasonal, but the intensity of the microclimate variation during the year was higher at GO, resulting in a marked seasonality during the flowering period at this site. Fruiting season can be considerably extended when compared to the flowering season, overlapping between sites. Tibouchina hatschbachii seems to follow a pattern that is common among plants occurring in different vegetation types in South America, with well-delimited flowering periods, but a fruiting season that is more extended along the time (Morellato et al., 2013). Despite being asynchronous between sites, the flowering season at both sites coincided with the summer, which indirectly favours cross-pollination, since the availability and richness of visitors, especially bees, tend to be higher during this period (Gonçalves & Melo, 2005; Gonçalvez, Melo & Aguiar, 2009). However, the flowering asynchrony between sites may have a negative impact on pollen flow, promoting reproductive isolation between sites and consequently negatively affecting individual reproductive fitness (Ollerton & Lack, 1998; Tarayre et al., 2007). Local environmental conditions seem also to contribute to the evolution of reproductive barriers between plants from the same species at both GO and SO (possible populations; Palma-Silva et al., 2011). Phylogeographic studies previously carried on T. hatschbachii, including the sites studied here, demonstrated that the distribution of the species was influenced by the recent movements of a geographical barrier in the region (a deep geological deformation known as the Ribeira do Iguape valley; Maia et al., 2017a). This barrier appears to have separated the ancestral population in two geographically structured genetic lineages, c. 1 Mya (Maia et al., 2017a). The byproduct of such adaptation may include a post-zygotic barrier between populations. An indirect evidence of such divergence was the result of the inter-site cross experiment: all fruits resulting from the treatment aborted before ripening. This abortion could indicate an accumulation of allelic differences between populations (Richards, 1997; Palma-Silva et al., 2011; Pinheiro et al., 2013), as verified in contemporary genetic structuration of T. hatschbachii (Maia et al., 2017b). Although fruit development began in the rainy season, fruit ripening and dehiscence tend to occur in the following months, coinciding with the dry season. This pattern is usually associated with autochoric species (self-dispersed or dispersed by the wind; Rathcke & Lacey, 1985), such as T. hatschbachii. We are aware that our interpretations of phenological variations are based only in a unique reproductive event and thus phenology may be triggered by other factors apart from climate (e.g. photoperiod, inter-annual variations in climate). To evaluate better whether these differences are environmentally triggered or genetically fixed, we suggest complementary phenological studies considering extended periods (several years) or plant transplant experiments between sites (GO and SO) to evaluate directly the soil and climate effects of environments on the reproductive phenological events. Variation in pollen dynamics and floral visitors We did not find evidence of limitation in pollen deposition or removal at either site, despite the differences in quality and quantity between them. This result contradicts the expectations for plant species in high elevation areas, such as Andean, alpine and upper montane environments, in which pollen limitation increases proportionally with elevation, often associated with a reduction in density and activity of some pollinators, specially bees (Arroyo et al., 1985; Totland, 1993; García-Camacho & Totland, 2009; Brito & Sazima, 2012). Despite the tendency for a lower frequency in GO (located at a higher elevation than SO), the number of pollinator visits in both areas was enough to prevent pollen limitation. In these environments, the conditions may not be so extreme as to limit the activity of bees. The relation between viability (an indicator of quality) and availability of pollen grains (an indicator of quantity) at both sites for T. hatschbachii may reflect a functional strategy related to its reproductive biology. This strategy allows plants to deal with the trade-off between the abundance of visitors acting as pollinators and as thieves and robbers, the dilemma of pollen grain loss to bees (feeding) and the assurance of reproduction (Luo et al., 2008; Ferreira & Araújo, 2016). At both sites, pollen removal was more intense in small anthers and the most frequent visitors were small bees acting as thieves and robbers. Illegitimate visitors (classified here as thieves and robbers) could exert biotic pressures on the pattern of pollen production and removal dynamics in these environments. At GO, we found fewer illegitimate visitors and, consequently, more pollen available in both anthers. On the other hand, we found less pollen available in both anther types at SO, as a consequence of the abundance of pollen thieves and robbers. At SO, individuals invested more in pollen quality in large anthers, although the overall viability was lower in comparison with individuals at GO. As a strategy to ensure the dispersion of male gametes, both sites have more pollen available in larger anthers, indicating a functional division-of-labour among stamens (Darwin, 1862; Luo et al., 2008, Luo et al., 2009). These anthers are morphologically adapted to place the pollen on the body of the pollinator (Fig. 1F), out of the grooming reach of the bees (Darwin, 1862; Luo et al., 2008, Luo et al., 2009). Our results based on the seed set resulting from natural pollination treatment across sites, suggest that both sites are responding to pressures imposed by biotic (even with slight differences in the number of visitors) and abiotic drivers. Recent studies indicate a relation between plants with high pollen viability and an increase in the visits by bees (Maia et al., 2016). High pollen viability and availability found at GO seem to explain the high abundance of pollinators, by increasing the frequency of pollinators that do not compete against thieves and robbers for the resource and assuring the destination of the pollen carried by these agents. Variation in reproductive biology Tibouchina hatschbachii is self-compatible at both sites. This agrees with previous studies on congeneric species (Goldenberg & Shepherd, 1998; dos Santos et al., 2012; Brito & Sazima, 2012; Maia et al., 2016) and indicates that this reproductive system tends to be conserved in the genus. However, despite the self-compatibility, fruit production differed between treatments and sites and this may be influenced by factors other than solely the reproductive strategy adopted by the species. Aspects related to the fragmented landscape and phenology may explain the overall differences in fruit set between sites. On SO, fruit production through cross-pollination between distant (DCP) and close (CCP) individuals was higher when compared to the same treatment on GO, indicating that cross-pollination was more efficient on SO. The spatial arrangement of the plants on SO (with individuals relatively close to each other) may be responsible for a higher proportion of fruits resulting from DCP and CCP treatments, since T. hatschbachii is pollinated by large bees known to fly for long periods and distances, such as Bombus, Centris, Euglossa and Xylocopa (Zurbuchen et al., 2010; Hagen, Wikelski & Kissling, 2011), consequently visiting more individuals along the flight route (Zimmerman, 1982; Zurbuchen et al., 2010). Furthermore, and despite the difference in viability, flowering on SO tends to be longer, which makes the resource (pollen) more predictable over time, contributing to the maintenance of pollinator visits and pollen flow (Brito & Sazima, 2012). Indeed, previous studies evaluating the contemporary genetic structure of the populations of this species in the same regions show that the contribution of pollen to the gene flow is almost four times greater than the contribution by seeds in populations occurring on SO than on GO, corroborating the efficiency of pollen transport on SO (Maia, 2017). Seed abortion rates were pronounced on SO. Seed abortion was always higher in the hand self-pollination (HSP) treatment and may be due to maternal resources restrictions, which may limit the transference of these resources to a self-fertilized progeny, and by post-zygotic mechanisms that may act as barriers to self-fertilization (Stephenson, 1981; Lloyd & Schoen, 1992; Harder & Routley, 2006). This, in turn, may also favour the reproductive success of individuals originated from cross-pollination (Lloyd & Schoen, 1992). Evidence for this is that 98.6% of the self-pollinated seeds did not complete their development, suggesting a decrease in the viability of the autogamous offspring and thus a possible selection against homozygotes (Gibbs & Sassaki, 1998). These results show that even considering the self-compatibility, the reproductive success is higher when cross-pollination occurs, since seeds resulting from the DCP treatment were more viable and presented higher germination rates. Indeed, as evidenced by previous studies evaluating the contemporary genetic structure of the populations, there is a high rate of apparent crossing between individuals in this area, corroborating the allogamous origin of the new responses on SO (Maia et al., 2017b; Maia, 2017). We conclude that spatiotemporal factors affect the reproductive biology of T. hatschbachii in subtropical areas, since fruit set and seed set resulting from different pollination treatments were different in each site. These factors include (1) climate variations between sites that may have shaped the reproductive phenology, which also seems to have been the result of adaptation to historical events; (2) differences in the landscape isolation between the two areas and also in the abundance and frequency of floral visitors and, more important, thieves and robbers, and in intrinsic individual characteristics (pollen quality/quantity) influencing the dynamics of pollen in these areas. These factors may also affect fruit set and the fitness of the individuals, shaping seed set and germination rates. Given that this is a likely scenario shared by many taxa from the Brazilian subtropical grasslands, other reproductive biology studies evaluating the individual and differential response of plants in this region may confirm these patterns. SUPPORTING INFORMATION Additional Supporting Information may be found in the online version of this article at the publisher’s web-site: Table S1. Floral visitors of Tibouchina hatschbachii on granitic outcrops (GO) and sandstone outcrops (SO). Figure S1. Precipitation and temperature during the period of study. (A) granitic outcrops (GO); (B) sandstone outcrops (SO). Source of climate data: SIMEPAR. Figure S2. Pollen tubes in Tibouchina hatschbachii in the two studied sites (N = 5 individuals for each site). (A–C) Granitic outcrops (GO): (A) pollen tubes reaching the embryonic egg sac, between 24–48 hours (flowers from the HSP treatment); (B) pollen tubes reaching the ovary, between 24–48h (flowers from CCP treatment); (C) pollen tubes reaching the ovary, between 48–72h (flowers from DCP treatment). Sandstone outcrops (SO): (D–E) pollen tube reaches the embryo from the egg sac between 24–48 hours in the treatment HSP and CCP, respectively; (F) pollen tubes germinating along the style between 24–48 hours in DCP treatment. Pollination treatments: hand self-pollination (HSP), close cross-pollination (CCP) and distant cross-pollination (DCP). ACKNOWLEDGEMENTS This work was supported by the Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq; project financing proc. 457510–2014–5) and we thank the CNPq for the doctorate grant (to FRM) and productivity grant (to RG). We thank IAP and COPEL for permissions and logistical support; P.A.P. Franzoi, C. Ribeiro and V.R.C. Maia for valuable help in the field; V.L.G. Brito, S. Koehler and A.R. Barbosa for helpful comments on a previous version of the manuscript. We also thank anonymous reviewers for substantial contributions to the final text. FJT thanks CAPES for her postdoc grants at UFPR and UFU. REFERENCES Alvares CA , Stape JL , Sentelhas PC , Golçalves JLM , Sparovek G . 2014 . Köppen’s climate classification map for Brazil . Meteorologische Zeitschrift 22 : 711 – 728 . Google Scholar CrossRef Search ADS Arroyo MTK , Armesto JJ , Primack RB . 1985 . Community studies in pollination ecology in the high temperate Andes of central Chile. II. Effect of temperature on visitation rates and pollination possibilities . Plant Systematics and Evolution 149 : 187 – 203 . Google Scholar CrossRef Search ADS Bates D , Maechler M , Bolker B , Walker S . 2015 . Fitting linear mixed-effects models using lme4 . Journal of Statistical Software 67 : 1 – 48 . Google Scholar CrossRef Search ADS Behling H , Negrelle RRB . 2001 . Tropical rain forest and climate dynamics of the Atlantic lowland, southern Brazil, during the late Quartenary . Quaternary Research 56 : 383 – 389 . Google Scholar CrossRef Search ADS Behling H . 2002 . South and southeast Brazilian grasslands during late Quaternary times: a synthesis . Palaeogeography, Palaeoclimatology and Palaeoecology 177 : 19 – 27 . Google Scholar CrossRef Search ADS Brito AC , Souza JD , Rebouças TNH , Amaral CLF . 2010 . Propriedades do pólen e do estigma de Ocimum basilicum L. (cultivar Maria Bonita) para aumentar a eficiência de cruzamentos em programas de melhoramento . Revista Brasileira de Plantas Medicinais 12 : 208 – 214 . Google Scholar CrossRef Search ADS Brito VLG , Sazima M . 2012 . Tibouchina pulchra (Melastomataceae): reproductive biology of a tree species at two sites of an elevational gradient in the Atlantic rainforest in Brazil . Plant Systematics and Evolution 298 : 1271 – 1279 . Google Scholar CrossRef Search ADS Buchmann SL . 1983 . Buzz pollination in angiosperms . In: Jones CE , Little RJ , eds. Handbook of experimental pollination biology . New York : Van Nostrand Reinhold , 73 – 113 . Dafni A , Kevan PG , Husband BC . 2005 . Practical pollination biology . Cambridge : Envirosquest . Dart SR , Samis KE , Austen E , Eckert CG . 2012 . Broad geographic covariation between floral traits and the mating system in Camissoniopsis cheiranthifolia (Onagraceae): multiple stable mixed mating systems across the species’ range ? Annals of Botany 109 : 599 – 611 . Google Scholar CrossRef Search ADS PubMed Darwin C . 1862 . Letter to Asa Gray, 22 January . Available at: http://www.darwinproject.ac.uk/con-tent/view/28/38/ (accessed 09.07.16 ). Dellinger AS , Penneys DS , Stedler YM , Fragner L , Weckwerth W , Schönenberger J . 2014 . A specialized bird pollination system with a bellows mechanism for pollen transfer and staminal food body rewards . Current Biology 14 : 1 – 5 . Ferreira QIX , Araújo FP . 2016 . Economia de pólen favorecida pela heteranteria em Desmocelis villosa (Melastomataceae) . Rodriguésia 67 : 347 – 355 . Google Scholar CrossRef Search ADS Freitas L , Sazima M . 2006 . Pollination biology in a tropical high altitude grassland in Brazil: interaction at the community level . Annals of the Missouri Botanical Garden 93 : 465 – 516 . Google Scholar CrossRef Search ADS Fox J , Weisberg S . 2011 . Functions and datasets to accompany. An R companion to applied regression , 2nd edn. Thousand Oaks : Sage . García-Camacho R , Totland O . 2009 . Pollen limitation in the Alpine: a meta-analysis . Arctic, Antarctic and Alpine Research 41 : 103 – 111 . Google Scholar CrossRef Search ADS Gibbs PE , Sassaki R . 1998 . Reproductive biology of Dalbergia miscolobium Benth. (Leguminosae – Papilionoidae) in SE Brazil: the effect of pistillate sorting on fruit-set . Annals of Botany 81 : 735 – 740 . Google Scholar CrossRef Search ADS Goldenberg R , Shepherd GJ . 1998 . Studies on the reproductive biology of Melastomataceae in ‘cerrado’ vegetation . Plant Systematics and Evolution 211 : 13 – 29 . Google Scholar CrossRef Search ADS Gonçalves RB , Melo GAR . 2005 . A comunidade de abelhas (Hymenoptera, Apidae s. l.) em uma área restrita de campo natural no Parque Estadual de Vila Velha, Paraná: diversidade, fenologia e fontes florais de alimento . Revista Brasileira de Entomologia 49 : 557 – 571 . Google Scholar CrossRef Search ADS Gonçalvez RB , Melo GAR , Aguiar AJC . 2009 . A assembleia de abelhas (Hymenoptera, Apidae) de uma área restrita de campos naturais do Parque Estadual de Vila Velha, Paraná e comparações com áreas de campos e cerrado . Papeis Avulsos de Zoologia 49 : 163 – 181 . Hagen M , Wikelski M , Kissling WD . 2011 . Space use of bumblebees (Bombus spp.) revealed by radio-tracking . PLoS One 6 : e19997 . Google Scholar CrossRef Search ADS PubMed Harder LD , Routley MB . 2006 . Pollen and ovule fates and reproductive performance by flowering plants . In: Harder LD , Barrett SCH , eds. Ecology and evolution of flowers . Oxford : Oxford University Press , 1 – 80 . Hoffman GM , Varassin IG . 2011 . Variação da viabilidade polínica em Tibouchina (Melastomataceae) . Rodriguésia 62 : 223 – 228 . Google Scholar CrossRef Search ADS Hothorn T , Bretz F , Westfall P , Heiberger RM , Schuetzenmeister A , Scheibe S . 2016 . Package multcomp: simultaneous inference in general parametric models. R package version 3.2.5 . Available at: https://cran.r–project.org/web/packages/multcomp/multcomp.pdf (accessed 21.03.17 ). Inouye DW . 1980 . The terminology of floral larceny . Ecology 61 : 1251 – 1252 . Google Scholar CrossRef Search ADS Kearns CA , Inouye D . 1993 . Techniques for pollination biologists . Niwot : University Press of Colorado . Kovach WL . 2016 . Oriana – circular statistics for Windows, version 4 . Pentraeth : Kovach Computing Services . Labiak PHE . 2014 . Plantas vasculares do Paraná. In: Kaehler M, Goldenberg R, Evangelista PHL, Ribas OS, Vieira AOS, Hatschbach GG, eds . Aspectos fitogeográficos do Paraná. Curitiba: Departamento de Botânica/UFPR, Paraná, Brasil, 7–22. Lloyd DG , Schoen DJ . 1992 . Self- and cross-fertilization in plants. II. The selection of self-fertilization . International Journal of Plant Sciences 153 : 370 – 380 . Google Scholar CrossRef Search ADS Luo Z , Zhang D , Renner SS . 2008 . Why two kinds of stamens in buzz-pollinated flowers? Experimental support for Darwin’s division-of-labour hypothesis . Functional Ecology 22 : 794 – 800 . Google Scholar CrossRef Search ADS Luo Z , Gu L , Zhang D . 2009 . Intrafloral differentiation of stamens in heterantherous flowers . Journal of Systematics and Evolution 47 : 43 – 56 . Google Scholar CrossRef Search ADS Maack R . 1981 . Geografia física do estado do Paraná . Curitiba : J. Olympio . Maia FR , Malucelli TS , Varassin IG . 2013 . Ecological factors affecting the fruiting success of a Tibouchina trichopoda (DC.) Baill. (Melastomataceae) flower . Acta Botanica Brasilica 27 : 142 – 146 . Google Scholar CrossRef Search ADS Maia FR , Varassin IG , Goldenberg R . 2016 . Apomixis does not affect visitation to flowers of Melastomataceae, but pollen sterility does . Plant Biology (Stuttgart, Germany) 18 : 132 – 138 . Google Scholar CrossRef Search ADS PubMed Maia FR , Zwiener VP , Morokawa R , Silva–Pereira V , Goldenberg R . 2017a . Phylogeography and ecological niche modeling uncover the evolutionary history of Tibouchina hatschbachii (Melastomataceae), a taxon restricted to the subtropical grasslands of South America . Botanical Journal of The Linnean Society 183 : 616 – 632 . Google Scholar CrossRef Search ADS Maia FR , Sujii PS , Silva-Pereira V , Goldenberg R . 2017b . Naturally fragmented and isolated distribution in subtropical grassland patches affects genetic diversity and structure at different spatial scales: the case of Tibouchina hatschbachii, an endemic shrub from Brazil . American Journal of Botany 104 : 1 – 11 . Google Scholar CrossRef Search ADS Maia FR . 2017 . The effect of time and space in the evolution of a taxon restricted to subtropical grasslands of South America . Unpublished Thesis, Universidade Estadual de Campinas , Campinas, Brazil . Martin FN . 1959 . Staining and observation of pollen tubes in the style by means of fluorescence . Stain Technology 34 : 125 – 128 . Google Scholar CrossRef Search ADS PubMed Meyer FS , Guimarães PJF , Goldenberg R . 2009 . Uma nova espécie de Tibouchina Aubl. (Melastomataceae) e notas taxonômicas sobre o gênero no Estado do Paraná, Brasil . Hoehnea 36 : 139 – 147 . Google Scholar CrossRef Search ADS Morellato LPC , Talora DC , Takahasi A , Bencke CC , Romera EC , Zipparro VB . 2000 . Phenology of Atlantic Rain Forest trees: a comparative study . Biotropica 32 : 811 – 823 . Google Scholar CrossRef Search ADS Morellato LPC , Alberti LF , Hudson IL . 2010 . Applications of circular statistics in plant phenology: a case studies approach . In: Hudson IL , Keatley M , eds. Phenological research: methods for environmental and climate change analysis . Dordrecht , Springer , 357 – 371 . Google Scholar CrossRef Search ADS Morellato LPC , Camargo MGG , Gressler E . 2013 . A review of plant phenology in South and Central America . In: Schwartz MD , ed. Phenology: an integrative environmental science . Dordrecht , Springer , 91 – 113 . Google Scholar CrossRef Search ADS Oksanen J , Blanchet FG , Kindt R , Legendre P , Minchin PR , O’Hara RB , Simpson GL, Solymos P, Henry M, Stevens H, Szoecs E, Wagner H. 2013 . vegan: community ecology package. R package version 3.2.5 . Available at: http://CRAN.Rproject.org/package=vegan (accessed 07.09.16 ). Ollerton J , Lack A . 1998 . Relationships between flowering phenology, plant size and reproductive success in Lotus corniculatus (Fabaceae) . Plant Ecology 139 : 35 – 47 . Google Scholar CrossRef Search ADS Palma-Silva C , Wendt T , Pinheiro F , Barbará T , Fay MF , Cozzolino S , Lexer C . 2011 . Sympatric bromeliad species (Pitcairnia spp.) facilitate tests of mechanisms involved in species cohesion and reproductive isolation in Neotropical inselbergs . Molecular Ecology 20 : 3185 – 3201 . Google Scholar CrossRef Search ADS PubMed Pinheiro F , Cozzolino S , de Barros F , Gouveia TM , Suzuki RM , Fay MF , Palma-Silva C . 2013 . Phylogeographic structure and outbreeding depression reveal early stages of reproductive isolation in the Neotropical orchid Epidendrum denticulatum . Evolution 67 : 2024 – 2039 . Google Scholar CrossRef Search ADS PubMed Pinheiro J , Bates D , DebRoy S , Sarkar D . 2016 . nlme: linear and nonlinear mixed effects models. R package version 3 , 1 – 128 . Price MV , Waser NM , Irwin RE , Campbell DR , Brody AK . 2005 . Temporal and spatial variation in pollination of a montane herb: a seven-year study . Ecological Society of America 86 : 2106 – 2116 RDCT - R Development Core Team . 2016 . R: a language and environment for statistical computing . Vienna, Austria : R Foundation for Statistical Computing . Radford AE . 1974 . Vascular plant systematics . New York : Harper and Row . Rech A . 2014 . Walking through the flower fields: the role of time and space on the evolution of pollination strategies . Unpublished Thesis, Universidade Estadual de Campinas , Campinas, Brazil . Rathcke B , Lacey EP . 1985 . Phenological patterns of terrestrial plants . Annual Review of Ecology and Systematics 16 : 179 – 214 . Google Scholar CrossRef Search ADS Renner SS . 1993 . Phylogeny and classification of the Melastomataceae and Memecylaceae . Nordic Journal of Botany 13 : 519 – 540 . Google Scholar CrossRef Search ADS Richards AJ . 1997 . Plant breeding systems . London : Chapman & Hall . Google Scholar CrossRef Search ADS dos Santos AP , Fracasso CM , Luciene dos Santos M , Romero R , Sazima M , Oliveira PE . 2012 . Reproductive biology and species geographical distribution in the Melastomataceae: a survey based on New World taxa . Annals of Botany 110 : 667 – 679 . Google Scholar CrossRef Search ADS PubMed SIMEPAR – Sistema meteorológico do Paraná [online] . Observatório agrometereológico de Telêmaco Borba e Antonina , Available at: http://www.simepar.br/ (accessed 27.03.15 ). Sofiatti V , Araújo EF , Araújo RF , Reis MS , Silva LVBD , Cargnin A . 2008 . Uso do hipoclorito de sódio para degradação do endocarpo de sementes de cafeeiro com diferentes graus de umidade . Revista Braileira de Sementes 30 : 150 – 160 . Google Scholar CrossRef Search ADS Stephenson AG . 1981 . Flower and fruit abortion: proximate causes and ultimate functions . Annual Review of Ecology and Systematics 12 : 253 – 279 . Google Scholar CrossRef Search ADS Tannus JLS , Assis MA , Morellato LPC . 2006 . Fenologia reprodutiva em campo sujo e campo úmido numa área de Cerrado no sudeste do Brasil, Itirapina – SP . Biota Neotropica 6 : 1 – 23 . Google Scholar CrossRef Search ADS Tarayre M , Bowman G , Schermann-Legionnet A , Barat M , Atlan A . 2007 . Flowering phenology of Ulex europaeus: ecological consequences of variation within and among populations . Evolutionary Ecology Research 21 : 395 – 409 . Google Scholar CrossRef Search ADS Thompson JN . 2005 . The geographic mosaic of coevolution . Chicago : University of Chicago Press . Totland Ø . 1993 . Pollination in alpine Norway: flowering phenology, insect visitors, and visitation rates in two plant communities . Canadian Journal of Botany 71 : 1072 – 1079 . Google Scholar CrossRef Search ADS Varassin IG , Penneys DS , Michelangeli FA . 2008 . Comparative anatomy and morphology of nectar-producing Melastomataceae . Annals of Botany 102 : 899 – 909 . Google Scholar CrossRef Search ADS PubMed Zapata TR , Arroyo MTK . 1978 . Plant reproductive ecology of a secondary deciduous tropical forest in Venezuela . Biotropica 10 : 221 – 230 . Google Scholar CrossRef Search ADS Zar JH . 2010 . Biostatistical analysis . Upper Saddle River : Prince-Hall International . Zimmerman M . 1982 . Optimal foraging: random movement by pollen collecting bumblebees . Oecologia 53 : 394 – 398 . Google Scholar CrossRef Search ADS PubMed Zurbuchen A , Landert L , Klaiber J , Muller A , Hein S , Dorn S . 2010 . Maximum foraging ranges in solitary bees: only few individuals have the capability to cover long foraging . Biological Conservation 143 : 669 – 676 . Google Scholar CrossRef Search ADS Zuur AF , Leno EN , Elphick CS . 2010 . A protocol for data exploration to avoid common statistical problems . Methods in Ecology and Evolution 1 : 3 – 14 . Google Scholar CrossRef Search ADS © 2018 The Linnean Society of London, Botanical Journal of the Linnean Society This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/open_access/funder_policies/chorus/standard_publication_model)

Phuphumirat, Wongkot;Leeratiwong, Charan;Malaikanok, Paranchai;Zetter, Reinhard

doi: 10.1093/botlinnean/boy033pmid: N/A

Abstract The morphological plasticity that can occur in pollen of Gnetum due to harmomegathy demands a comparative overview of pollen characters in Gnetum when studying systematics, particularly when fossil pollen grains are taken into account. Using LM and SEM, we studied acetolyzed and dry and rehydrated non-acetolyzed pollen grains of eight Gnetum spp. in Thailand. They are subspheroidal, inaperturate, microechinate monads with blunt-ended microechini. In acetolyzed and non-acetolyzed pollen grains, the raised exine regions and the ridge–like folds appear to be characteristic of the dehydrated state, whereas a psilate surface between microechini is representative of the hydrated state. Although we were able to measure and record some variation in average values of certain features of all the pollen studied and in some characters of the dehydrated specimens, the general morphological similarity of all the examined Gnetum pollen makes it difficult to establish any systematic significance. Nonetheless, their continuous tectum supports the molecular phylogenetic information that suggests Asian Gnetum spp. are monophyletic. This study presents the first report of the pollen morphology of Gnetum microcarpum Blume and G. tenuifolium Ridl. © 2018 The Linnean Society of London, Biological Journal of the Linnean Society, 2018, XX, 000–000. exine, Gnetum, harmomegathy, palynology INTRODUCTION Gnetales are a monophyletic order, comprising three families of gymnosperms, Ephedraceae, Gnetaceae and Welwitschiaceae. Gnetum L. belongs to Gnetaceae and comprises evergreen dioecious (or rarely monoecious) trees, shrubs and woody climbers, growing in moist tropical and subtropical regions of Asia, Africa and South America (Markgraf, 1929; Crane, 1985; Kubitzki, 1990; Price, 1996; Friis, Crane & Pedersen, 2011). There are c. 43 species: four in Africa, seven in South America and c. 32 in Asia (Biye, Balkwill & Cron, 2014; Hou et al., 2015). The systematic relationships among Gnetum spp. have been well studied (Griffith, 1859; Markgraf, 1929; Carlquist, 1996; Won & Renner, 2005, 2006; Hou et al., 2015). Based on morphological data, two sections have been established, including Gnetum (synonyms: sections Erecta Griff. and Gnemonomorphi Markgr.) and Scandentia Griff. (synonym: section Cylindrostachys Markgr.). Gnetum section Gnetum includes three subsections, Araeognemones Markgr. (American taxa), Micrognemones Markgr. (African taxa) and Eugnemones Markgr. (arborescent Asian taxa). Gnetum section Scandentia (lianoid Asian taxa) is further divided into two subsections, Stipitati Markgr. and Sessiles Markgr. (Griffith, 1859; Markgraf, 1929; Price, 1996). Markgraf (1929) proposed that arborescent Gnetum taxa are ancestral and had diverged earlier than other species in the genus. The molecular analyses of Hou et al. (2015) suggested that section Gnetum is paraphyletic and section Scandentia is monophyletic. However, the subsections of section Gnetum were monophyletic, whereas those of section Scandentia were not. Their results also provided strong support for the existence of three geographically defined major clades of Gnetum: the American, African and Asian clades. The pollen morphology of Gnetum has been studied for more than 80 years (Wodehouse, 1935; Erdtman, 1965; Gullvåg, 1966; Hesse, 1980; Orel, Kuprianova & Golubeva, 1986; Gillespie & Nowicke, 1994; Osborn, 2000; Yao et al., 2004; Ickert-Bond & Renner, 2016; Tekleva, 2016). The pollen has been generally described as small, spheroidal or subspheroidal, inaperturate and microechinate (Erdtman, 1965; Gillespie & Nowicke, 1994; Osborn, 2000; Yao et al., 2004; Tekleva, 2016). However, a pore-like region was mentioned in one previous study (Erdtman, 1965). The pollen wall ultrastructure showed a thin tectum, granular infratectum, lamellate endexine and an indistinct or absent foot layer (Osborn, 2000; Yao et al., 2004; Tekleva, 2016). Based on differences observed in the tectal ornamentations by light microscopy (LM) alone, Erdtman (1954, 1965) divided Gnetum pollen into three types: spinulose (Asian taxa), pilate (Neotropical species) and insulous (African species). Different results were obtained from scanning electron microscopy (SEM) analyses as only two pollen types were recognized. These are the Asian type, represented by a continuous tectum with microechini, and the American and African type, characterized by a discontinuous tectum forming plate-like regions or islands with more numerous microechini (Gillespie & Nowicke, 1994; Tekleva, 2016). The raised exine region and ridge (or fold) around the echinus or echini, forming a network-like pattern, are interesting characters of pollen from Asian Gnetum (Gillespie & Nowicke, 1994; Yao et al., 2004; Tekleva, 2016). Gillespie & Nowicke (1994) observed the differences in the tectum surfaces of pollen samples from G. latifolium Blume and stated that the variation of folding found was influenced by sample preparation procedures. However, in a natural environment, pollen morphology can change to accommodate the pressure changes within pollen grains in response to wet or dry conditions (Wodehouse, 1935; Hesse et al., 2009) and numerous studies have demonstrated the plasticity of pollen morphology due to harmomegathy (Payne, 1972; Pacini, 1990; Scotland, Barnes & Blackmore, 1990; Gillespie & Nowicke, 1994; Hesse, 2000; Hesse et al., 2001; Halbritter & Hesse, 2004; Hesse et al., 2009; Halbritter, Weber & Hesse 2010; Katifori et al., 2010; Volkova, Severova & Polevova, 2013). This causes remarkable differences in the appearance of pollen even from the same plant species, leading to problems when making comparative interpretations, particularly in the fossil record. In Thailand, eight Gnetum taxa have been observed: G. cuspidatum Blume; G. latifolium var. funiculare (Blume) Markgr.; G. leptostachyum Blume; G. gnemon L. var. tenerum Markgr.; G. macrostachyum Hook.f.; G. microcarpum Blume; G. montanum Markgr. and G. tenuifolium Ridl. (Phengklai, 1973). Besides the pollen morphology of G. microcarpum and G. tenuifolium, other pollen morphologies have been reported (Gillespie & Nowicke, 1994; Osborn, 2000; Yao et al., 2004; Tekleva, 2016). However, these morphological characterizations were based on examinations of acetolyzed or dry pollen grains from herbaria. No hydrated or rehydrated pollen grains treated with different methods have been previously examined. Since Gnetum pollen exhibits morphological flexibility depending on the environmental conditions or even the preparation technique used, palynological data, particularly from fossil Gnetum pollen, can be misinterpreted. Therefore, this present study aimed to document the pollen morphology of Gnetum pollen in Thailand and harmomegathic characters in order to provide a proper perspective on the comparative morphology of these pollen to establish their systematic significance. MATERIAL AND METHODS Pollen from all the Gnetum spp. recorded in Thailand (Phengklai, 1973) were studied. Except for pollen from two specimens of G. microcarpum and one specimen of G. macrostachyum, which were collected from Kho Hong Hill, Songkhla province, southern Thailand, the pollen samples of all species were taken from the Herbarium, Biology Department, Faculty of Science, Prince of Songkla University (PSU), the Herbarium of the Department of Biology, Faculty of Science, Chang-Mai University (CMU) and the Queen Sirikit Botanic Garden Herbarium (QBG), Botanical Garden Organization, Thailand (see Appendix). To ensure the correct identification of the pollen from G. microcarpum, collected plant samples were compared with specimens deposited in the PSU herbarium. For LM study, pollen grains were acetolyzed and mixed with glycerin before mounting on slides (Erdtman, 1969). The pollen size, outline, exine thickness and apertural condition were investigated. The pollen was then photographed under light microscopy, BX43, Olympus. For SEM analysis, two types of pollen samples were examined: acetolyzed and non-acetolyzed. All the non-acetolyzed pollen samples were studied in dry and rehydrated conditions. The acetolyzed pollen samples were placed on stubs using a nose hair-tipped needle (Zetter, 1989). The dry pollen samples were further dehydrated and cleaned in acidified 2, 2–dimethoxypropane (DMP) for 30 min, followed by critical point drying (Halbritter, 1998). This preparation technique was also performed for the study of the pollen morphology of rehydrated pollen grains, but before treatment with acidified DMP, the pollen samples were rehydrated with Agepon solution in distilled water (1:200) for 30 min. Other dry-untreated pollen samples from all Gnetum spp. were also investigated under SEM in order to ensure that their obtained characters of dehydration were not caused by sample preparation techniques. All pollen samples on SEM stubs were sputter-coated with gold using an SPI-module sputter coater (with SPI supplies) and then examined and photographed using an FEI Quanta 400 SEM. The characters studied were the pollen size, the type of microechinus tip, the echinus height and base width, the distance between microechini, the number of microechini per 25 µm2 surface area and the number of microechini in each inflated region. The FEI XT microscope control software was used to measure the parameters mentioned. Twenty pollen grains were examined for each species in each treatment. However, measurements were not made in the study of the dry-untreated pollen samples. Pollen morphology was described using the terminology published in Punt et al. (2007) and Hesse et al. (2009). RESULTS The results are shown in evolutionary order, referring to the phylogenetic information studies of Hou et al. (2015) and Won & Renner (2006). In the present study, we examined pollen from eight Gnetum taxa found in Thailand: one arborescent Asian taxa (Gnetum gnemon var. tenerum from clade G) and seven lianoid Asian taxa: G. montanum from clade O, or the hainanense clade; G. latifolium var. funiculare and G. leptostachyum from clade M, or the latifolium clade; and G. macrostachyum, G. tenuifolium, G. cuspidatum and G. microcarpum from clade N, or the cuspidatum clade. The morphologies of their pollen are described as follows. Gnetum gnemon var. Tenerum Pollen are monads, subspheroidal, elliptic or ovate in the long axis view, circular or elliptic in the short axis view, long axis diameter 14–20 µm in LM, 11.13–16.78 µm in SEM (acetolyzed grains), 8.71–16.69 µm in SEM (non-acetolyzed grains), short axis diameter 12.0–17.5 µm in LM, 10.01–16.18 µm in SEM (acetolyzed grains), 8.27–13.11 µm in SEM (non-acetolyzed grains), inaperturate, exine 0.5–0.8 µm thick (LM), tectate, continuous tectum, microechinate in LM (showing inapparent microechini), microechinate with psilate or ridged surface between microechini in SEM, with a bluntly round tip. On some grains, there are remarkably raised exine regions surrounded by a ridge, with one to five microechini on each region. Microechini are small, 0.18–0.83 µm high and 0.28–0.64 µm in base diameter (acetolyzed grains), 0.16–0.44 µm high and 0.26–0.89 µm in base diameter (non-acetolyzed grains). There are 10–19 microechini per surface area of 25 µm2 on acetolyzed grains and 11–22 microechini per surface area of 25 µm2 on non-acetolyzed grains. The distances between microechini on acetolyzed grains range from 0.95 to 2.31 µm, whereas those on non-acetolyzed grains vary from 0.23 to 2.62 µm (Tables 1–3; Figs 1A, 2A–C, 4A–D). Figure 1. View largeDownload slide Light micrographs (LM) of Gnetum pollen species in Thailand (A-H). A, Gnetum gnemon. B, Gnetum montanum. C, Gnetum latifolium. D, Gnetum leptostachyum. E, Gnetum macrostachyum. F, Gnetum tenuifolium. G, Gnetum cuspidatum. H, Gnetum microcarpum. Scale bars: 10 μm (A–H). Figure 1. View largeDownload slide Light micrographs (LM) of Gnetum pollen species in Thailand (A-H). A, Gnetum gnemon. B, Gnetum montanum. C, Gnetum latifolium. D, Gnetum leptostachyum. E, Gnetum macrostachyum. F, Gnetum tenuifolium. G, Gnetum cuspidatum. H, Gnetum microcarpum. Scale bars: 10 μm (A–H). Figure 2. View largeDownload slide Scanning electron micrographs (SEM) of acetolyzed pollen grains from Gnetum species in Thailand showing the variation in sculptures (A–L). A–C, Gnetum gnemon. D–F, Gnetum montanum. G–I, Gnetum latifolium. J–L, Gnetum leptostachyum (A, D, G and J, general view; B, E, H and K, echinate with apparent ridge and raised exine region; C, F, I and L, echinate with no, or inapparent, ridge and raised exine region). Arrows indicate a raised exine region surrounded by a ridge. An asterisk indicates a dome-shaped microechinus. Scale bars: 10 μm (A, D, G and J); 1 μm (B, C, E, F, H, I, K and L). Figure 2. View largeDownload slide Scanning electron micrographs (SEM) of acetolyzed pollen grains from Gnetum species in Thailand showing the variation in sculptures (A–L). A–C, Gnetum gnemon. D–F, Gnetum montanum. G–I, Gnetum latifolium. J–L, Gnetum leptostachyum (A, D, G and J, general view; B, E, H and K, echinate with apparent ridge and raised exine region; C, F, I and L, echinate with no, or inapparent, ridge and raised exine region). Arrows indicate a raised exine region surrounded by a ridge. An asterisk indicates a dome-shaped microechinus. Scale bars: 10 μm (A, D, G and J); 1 μm (B, C, E, F, H, I, K and L). Figure 3. View largeDownload slide Scanning electron micrographs (SEM) of acetolyzed pollen grains from Gnetum species in Thailand showing the variation in sculptures (A–L). A–C, Gnetum macrostachyum. D–F, Gnetum tenuifolium. G–I, Gnetum cuspidatum. J–L, Gnetum microcapum (A, D, G and J, general view; B, E, H and K, echinate with apparent ridge and raised exine region; C, crumpled pollen grain; F, I and L, echinate with no, or inapparent, ridge and raised exine region). Arrows indicate a raised exine region surrounded by a ridge. An asterisk indicates a dome-shaped microechinus. Scale bars: 10 μm (A, D, G and J); 1 μm (B, C, E, F, H, I, K and L). Figure 3. View largeDownload slide Scanning electron micrographs (SEM) of acetolyzed pollen grains from Gnetum species in Thailand showing the variation in sculptures (A–L). A–C, Gnetum macrostachyum. D–F, Gnetum tenuifolium. G–I, Gnetum cuspidatum. J–L, Gnetum microcapum (A, D, G and J, general view; B, E, H and K, echinate with apparent ridge and raised exine region; C, crumpled pollen grain; F, I and L, echinate with no, or inapparent, ridge and raised exine region). Arrows indicate a raised exine region surrounded by a ridge. An asterisk indicates a dome-shaped microechinus. Scale bars: 10 μm (A, D, G and J); 1 μm (B, C, E, F, H, I, K and L). Figure 4. View largeDownload slide Scanning electron micrographs (SEM) of non-acetolyzed pollen grains from Gnetum species in Thailand showing the variation in sculptures (A–P). A–D, Gnetum gnemon. E–H, Gnetum montanum. I–L, Gnetum latifolium. M–P, Gnetum leptostachyum (A, E, I and M, dry-treated samples, general view; B, F, J and N, rehydrated samples, general view; C, G and O, dry-treated samples, echinate with apparent ridge and raised exine region; K, dry-treated sample, crumpled pollen grain; D, H, L and P, rehydrated samples, echinate with no, or inapparent, ridge and raised exine region). Arrows indicate a raised exine region surrounded by a ridge. An asterisk indicates a dome–shaped microechinus. Scale bars: 10 μm (A, B, E, F, I, J, M and N); 1 μm (C, D, G, H, K, L, O and P). Figure 4. View largeDownload slide Scanning electron micrographs (SEM) of non-acetolyzed pollen grains from Gnetum species in Thailand showing the variation in sculptures (A–P). A–D, Gnetum gnemon. E–H, Gnetum montanum. I–L, Gnetum latifolium. M–P, Gnetum leptostachyum (A, E, I and M, dry-treated samples, general view; B, F, J and N, rehydrated samples, general view; C, G and O, dry-treated samples, echinate with apparent ridge and raised exine region; K, dry-treated sample, crumpled pollen grain; D, H, L and P, rehydrated samples, echinate with no, or inapparent, ridge and raised exine region). Arrows indicate a raised exine region surrounded by a ridge. An asterisk indicates a dome–shaped microechinus. Scale bars: 10 μm (A, B, E, F, I, J, M and N); 1 μm (C, D, G, H, K, L, O and P). Table 1. Morphology of acetolyzed pollen grains from studied Gnetum species under LM. The average value and standard deviation are shown in parentheses. Species G. gnemon G. montanum G. latifolium G. leptostachyum G. macrostachyum G. tenuifolium G. cuspidatum G. microcarpum Pollen size (μm) 12.0–17.5 × 14.0–20.0 (14.6 ± 1.6 × 16.5 ± 1.5) 11.0–15.0 × 12.0–17.5 (13.5 ± 1.4 × 15.3 ± 1.7) 10.0–15.0 × 13.0–16.0 (13.3 ± 1.7 × 14.6 ± 0.7) 10.0–14.0 × 12.0–16.0 (12.2 ± 1.0 × 13.7 ± 0.9) 12.0–14.0 × 12.5–16.5 (12.6 ± 0.7 × 14.8 ± 1.2) 11.0–15.0 × 12.0–17.5 (12.9 ± 1.4 × 14.5 ± 1.3) 11.0–14.0 × 12.0–16.0 (12.8 ± 0.9 × 14.3 ± 0.9) 11.0–15.4 × 12.4–19.2 (12.5 ± 1.1 × 14.5 ± 1.8) Exine thicknessa (μm) 0.5–0.8 (0.6 ± 0.1) 0.6–0.9 (0.7 ± 0.1) 0.4–0.7 (0.6 ± 0.1) 0.4–0.8 (0.6 ± 0.1) 0.5–0.8 (0.6 ± 0.1) 0.4–0.8 (0.6 ± 0.1) 0.5–0.8 (0.7 ± 0.1) 0.5–0.6 (0.5 ± 0.1) Species G. gnemon G. montanum G. latifolium G. leptostachyum G. macrostachyum G. tenuifolium G. cuspidatum G. microcarpum Pollen size (μm) 12.0–17.5 × 14.0–20.0 (14.6 ± 1.6 × 16.5 ± 1.5) 11.0–15.0 × 12.0–17.5 (13.5 ± 1.4 × 15.3 ± 1.7) 10.0–15.0 × 13.0–16.0 (13.3 ± 1.7 × 14.6 ± 0.7) 10.0–14.0 × 12.0–16.0 (12.2 ± 1.0 × 13.7 ± 0.9) 12.0–14.0 × 12.5–16.5 (12.6 ± 0.7 × 14.8 ± 1.2) 11.0–15.0 × 12.0–17.5 (12.9 ± 1.4 × 14.5 ± 1.3) 11.0–14.0 × 12.0–16.0 (12.8 ± 0.9 × 14.3 ± 0.9) 11.0–15.4 × 12.4–19.2 (12.5 ± 1.1 × 14.5 ± 1.8) Exine thicknessa (μm) 0.5–0.8 (0.6 ± 0.1) 0.6–0.9 (0.7 ± 0.1) 0.4–0.7 (0.6 ± 0.1) 0.4–0.8 (0.6 ± 0.1) 0.5–0.8 (0.6 ± 0.1) 0.4–0.8 (0.6 ± 0.1) 0.5–0.8 (0.7 ± 0.1) 0.5–0.6 (0.5 ± 0.1) aExine thickness excluding the echinus height View Large Table 1. Morphology of acetolyzed pollen grains from studied Gnetum species under LM. The average value and standard deviation are shown in parentheses. Species G. gnemon G. montanum G. latifolium G. leptostachyum G. macrostachyum G. tenuifolium G. cuspidatum G. microcarpum Pollen size (μm) 12.0–17.5 × 14.0–20.0 (14.6 ± 1.6 × 16.5 ± 1.5) 11.0–15.0 × 12.0–17.5 (13.5 ± 1.4 × 15.3 ± 1.7) 10.0–15.0 × 13.0–16.0 (13.3 ± 1.7 × 14.6 ± 0.7) 10.0–14.0 × 12.0–16.0 (12.2 ± 1.0 × 13.7 ± 0.9) 12.0–14.0 × 12.5–16.5 (12.6 ± 0.7 × 14.8 ± 1.2) 11.0–15.0 × 12.0–17.5 (12.9 ± 1.4 × 14.5 ± 1.3) 11.0–14.0 × 12.0–16.0 (12.8 ± 0.9 × 14.3 ± 0.9) 11.0–15.4 × 12.4–19.2 (12.5 ± 1.1 × 14.5 ± 1.8) Exine thicknessa (μm) 0.5–0.8 (0.6 ± 0.1) 0.6–0.9 (0.7 ± 0.1) 0.4–0.7 (0.6 ± 0.1) 0.4–0.8 (0.6 ± 0.1) 0.5–0.8 (0.6 ± 0.1) 0.4–0.8 (0.6 ± 0.1) 0.5–0.8 (0.7 ± 0.1) 0.5–0.6 (0.5 ± 0.1) Species G. gnemon G. montanum G. latifolium G. leptostachyum G. macrostachyum G. tenuifolium G. cuspidatum G. microcarpum Pollen size (μm) 12.0–17.5 × 14.0–20.0 (14.6 ± 1.6 × 16.5 ± 1.5) 11.0–15.0 × 12.0–17.5 (13.5 ± 1.4 × 15.3 ± 1.7) 10.0–15.0 × 13.0–16.0 (13.3 ± 1.7 × 14.6 ± 0.7) 10.0–14.0 × 12.0–16.0 (12.2 ± 1.0 × 13.7 ± 0.9) 12.0–14.0 × 12.5–16.5 (12.6 ± 0.7 × 14.8 ± 1.2) 11.0–15.0 × 12.0–17.5 (12.9 ± 1.4 × 14.5 ± 1.3) 11.0–14.0 × 12.0–16.0 (12.8 ± 0.9 × 14.3 ± 0.9) 11.0–15.4 × 12.4–19.2 (12.5 ± 1.1 × 14.5 ± 1.8) Exine thicknessa (μm) 0.5–0.8 (0.6 ± 0.1) 0.6–0.9 (0.7 ± 0.1) 0.4–0.7 (0.6 ± 0.1) 0.4–0.8 (0.6 ± 0.1) 0.5–0.8 (0.6 ± 0.1) 0.4–0.8 (0.6 ± 0.1) 0.5–0.8 (0.7 ± 0.1) 0.5–0.6 (0.5 ± 0.1) aExine thickness excluding the echinus height View Large Table 2. Morphology of acetolyzed pollen grains from studied Gnetum species under SEM (an = 11; bn = 7). The average value and standard deviation are shown in parentheses. Species G. gnemon G. montanum G. latifolium G. leptostachyum G. macrostachyum G. tenuifolium G. cuspidatum G. microcarpum Acetolyzed pollen grains showing apparent ridge and raised exine region (SEM) Pollen size (µm) 10.01–12.53 × 11.13–13.74 (11.73 ± 0.84 × 12.61 ± 0.70) 8.58–15–85 × 10.15–16.48 (12.18 ± 2.79 × 13.57 ± 2.06) 8.03–15.14 × 9.84–15.89 (11.22 ± 1.98 × 12.42 ± 1.60) 10.98–13.81 × 11.90–14.83 (11.84 ± 077 × 13.00 ± 0.79) 11.23–16.77 × 11.87–17.76 (13.31 ± 1.47 × 13.96 ± 1.53) 11.33–15.11 × 11.82–16.42 (12.85 ± 1.28 × 13.99 ± 1.95) 11.20–15.62 × 11.62–16.03 (13.38 ± 1.82 × 13.69 ± 1.94) 11.20–13.62 × 12.34–14.47 (11.81 ± 0.87 × 13.35 ± 0.87b) Microechinus height (µm) 0.34–0.83 (0.59 ± 0.13) 0.33–0.69 (0.47 ± 0.09) 0.39–0.61 (0.47 ± .05) 0.16–0.36 (0.26 ± 0.06) 0.60–0.89 (0.72 ± 0.08) 0.22–0.78 (0.45 ± 0.16) 0.38–1.09 (0.71 ± 0.25) 0.23–0.36 (0.31 ± 0.04b) Microechinus base diameter (µm) 0.28–0.59 (0.47 ± 0.9) 0.29–0.75 (0.51 ± 0.09) 0.36–0.79 (0.48 ± 0.09) 0.21–0.67 (0.40 ± 0.13) 0.78–1.22 (0.93 ± 0.13) 0.38–1.34 (0.82 ± 0.28) 0.47–1.72 (0.99 ± 0.41) 0.34–0.58 (0.45 ± 0.08b) Distance between microechini (µm) 1.03–1.81 (1.40 ± 0.25) 0.93–2.11 (1.42 ± 0.29) 0.55–2.32 (1.42 ± 0.47) 0.84–2.15 (1.41 ± 0.24) 1.15–1.93 (1.47 ± 0.26) 0.78–2.33 (1.42 ± 0.40) 1.04–2.30 (1.45 ± 0.34) 1.15–2.18 (1.62 ± 0.39b) Number of microechini per 25 µm2 (microechini/25 µm2) 14–19 (17.05 ± 1.50) 13–18 (15.55 ± 1.43) 10–16 (12.15 ± 1.46) 10–15 (12.75 ± 1.52) 8–11 (8.85 ± 0.81) 14–19 (16.00 ± 1.30) 9–14 (11.75 ± 1.25) 10–13 (11.14 ± 1.07) Number of microechini in each raised exine region 1–5 1–8 1–3 1–6 1–3 2–11 1–3 1–3b Acetolyzed pollen grains showing no (or inapparent) ridge and raised exine region (SEM) Pollen size (µm) 11.80–16.18 × 13.51–16.78 (13.94 ± 1.54 × 15.34 ± 0.95) 8.34–14.65 × 10.49–16.68 (11.66 ± 2.5 × 13.46 ± 2.10) 9.33–15.56 × 11.52–16.36 (12.11 ± 1.77 × 13.88 ± 1.67) 10.03–13.50 × 11.95–14.92 (11.82 ± 1.10 × 13.22 ± 0.82a) - 10.56–15.80 × 11.51–17.88 (12.91 ± 1.90 × 14.63 ± 2.53) 11.15–16.56 × 11.85–17.00 (13.30 ± 1.73 × 13.98 ± 1.80) 10.27–15.61 × 11.99–21.70 (12.60 ± 1.66 × 14.38 ± 3.41) Microechinus height (µm) 0.18–0.42 (0.28 ± 0.07) 0.37–0.69 (0.54 ± 0.07) 0.28–0.71 (0.48 ± 0.09) 0.30–0.41 (0.35 ± 0.04a) - 0.15–0.78 (0.42 ± 0.19) 0.22–0.75 (0.44 ± 0.13) 0.16–0.49 (0.32 ± 0.09) Microechinus base diameter (µm) 0.32–0.64 (0.45 ± 0.09) 0.46–0.75 (0.59 ± 0.07) 0.29–0.67 (0.51 ± 0.80) 0.38–0.67 (0.55 ± 0.08a) - 0.21–0.75 (0.45 ± 0.13) 0.38–0.76 (0.53 ± 0.12) 0.26–0.51 (0.40 ± 0.07) Distance between microechini (µm) 0.95–2.31 (1.63 ± 0.36) 1.00–2.89 (1.71 ± 0.40) 1.12–2.58 (1.71 ± 0.38) 1.22–2.29 (1.73 ± 0.29a) - 1.29–2.68 (1.86 ± 0.40) 1.32–2.93 (1.92 ± 0.49) 1.28–2.10 (1.64 ± 0.19) Number of microechini per 25 µm2 (microechini/25 µm2) 10 -14 (12.45 ± 1.23) 10–15 (11.40 ± 1.27) 9–13 (10.35 ± 1.35) 10–13 (11.82 ± 0.75 a) - 6–12 (8.95 ± 1.76) 6–10 (8.20 ± 1.15) 7–13 (9.60 ± 1.67) Species G. gnemon G. montanum G. latifolium G. leptostachyum G. macrostachyum G. tenuifolium G. cuspidatum G. microcarpum Acetolyzed pollen grains showing apparent ridge and raised exine region (SEM) Pollen size (µm) 10.01–12.53 × 11.13–13.74 (11.73 ± 0.84 × 12.61 ± 0.70) 8.58–15–85 × 10.15–16.48 (12.18 ± 2.79 × 13.57 ± 2.06) 8.03–15.14 × 9.84–15.89 (11.22 ± 1.98 × 12.42 ± 1.60) 10.98–13.81 × 11.90–14.83 (11.84 ± 077 × 13.00 ± 0.79) 11.23–16.77 × 11.87–17.76 (13.31 ± 1.47 × 13.96 ± 1.53) 11.33–15.11 × 11.82–16.42 (12.85 ± 1.28 × 13.99 ± 1.95) 11.20–15.62 × 11.62–16.03 (13.38 ± 1.82 × 13.69 ± 1.94) 11.20–13.62 × 12.34–14.47 (11.81 ± 0.87 × 13.35 ± 0.87b) Microechinus height (µm) 0.34–0.83 (0.59 ± 0.13) 0.33–0.69 (0.47 ± 0.09) 0.39–0.61 (0.47 ± .05) 0.16–0.36 (0.26 ± 0.06) 0.60–0.89 (0.72 ± 0.08) 0.22–0.78 (0.45 ± 0.16) 0.38–1.09 (0.71 ± 0.25) 0.23–0.36 (0.31 ± 0.04b) Microechinus base diameter (µm) 0.28–0.59 (0.47 ± 0.9) 0.29–0.75 (0.51 ± 0.09) 0.36–0.79 (0.48 ± 0.09) 0.21–0.67 (0.40 ± 0.13) 0.78–1.22 (0.93 ± 0.13) 0.38–1.34 (0.82 ± 0.28) 0.47–1.72 (0.99 ± 0.41) 0.34–0.58 (0.45 ± 0.08b) Distance between microechini (µm) 1.03–1.81 (1.40 ± 0.25) 0.93–2.11 (1.42 ± 0.29) 0.55–2.32 (1.42 ± 0.47) 0.84–2.15 (1.41 ± 0.24) 1.15–1.93 (1.47 ± 0.26) 0.78–2.33 (1.42 ± 0.40) 1.04–2.30 (1.45 ± 0.34) 1.15–2.18 (1.62 ± 0.39b) Number of microechini per 25 µm2 (microechini/25 µm2) 14–19 (17.05 ± 1.50) 13–18 (15.55 ± 1.43) 10–16 (12.15 ± 1.46) 10–15 (12.75 ± 1.52) 8–11 (8.85 ± 0.81) 14–19 (16.00 ± 1.30) 9–14 (11.75 ± 1.25) 10–13 (11.14 ± 1.07) Number of microechini in each raised exine region 1–5 1–8 1–3 1–6 1–3 2–11 1–3 1–3b Acetolyzed pollen grains showing no (or inapparent) ridge and raised exine region (SEM) Pollen size (µm) 11.80–16.18 × 13.51–16.78 (13.94 ± 1.54 × 15.34 ± 0.95) 8.34–14.65 × 10.49–16.68 (11.66 ± 2.5 × 13.46 ± 2.10) 9.33–15.56 × 11.52–16.36 (12.11 ± 1.77 × 13.88 ± 1.67) 10.03–13.50 × 11.95–14.92 (11.82 ± 1.10 × 13.22 ± 0.82a) - 10.56–15.80 × 11.51–17.88 (12.91 ± 1.90 × 14.63 ± 2.53) 11.15–16.56 × 11.85–17.00 (13.30 ± 1.73 × 13.98 ± 1.80) 10.27–15.61 × 11.99–21.70 (12.60 ± 1.66 × 14.38 ± 3.41) Microechinus height (µm) 0.18–0.42 (0.28 ± 0.07) 0.37–0.69 (0.54 ± 0.07) 0.28–0.71 (0.48 ± 0.09) 0.30–0.41 (0.35 ± 0.04a) - 0.15–0.78 (0.42 ± 0.19) 0.22–0.75 (0.44 ± 0.13) 0.16–0.49 (0.32 ± 0.09) Microechinus base diameter (µm) 0.32–0.64 (0.45 ± 0.09) 0.46–0.75 (0.59 ± 0.07) 0.29–0.67 (0.51 ± 0.80) 0.38–0.67 (0.55 ± 0.08a) - 0.21–0.75 (0.45 ± 0.13) 0.38–0.76 (0.53 ± 0.12) 0.26–0.51 (0.40 ± 0.07) Distance between microechini (µm) 0.95–2.31 (1.63 ± 0.36) 1.00–2.89 (1.71 ± 0.40) 1.12–2.58 (1.71 ± 0.38) 1.22–2.29 (1.73 ± 0.29a) - 1.29–2.68 (1.86 ± 0.40) 1.32–2.93 (1.92 ± 0.49) 1.28–2.10 (1.64 ± 0.19) Number of microechini per 25 µm2 (microechini/25 µm2) 10 -14 (12.45 ± 1.23) 10–15 (11.40 ± 1.27) 9–13 (10.35 ± 1.35) 10–13 (11.82 ± 0.75 a) - 6–12 (8.95 ± 1.76) 6–10 (8.20 ± 1.15) 7–13 (9.60 ± 1.67) View Large Table 2. Morphology of acetolyzed pollen grains from studied Gnetum species under SEM (an = 11; bn = 7). The average value and standard deviation are shown in parentheses. Species G. gnemon G. montanum G. latifolium G. leptostachyum G. macrostachyum G. tenuifolium G. cuspidatum G. microcarpum Acetolyzed pollen grains showing apparent ridge and raised exine region (SEM) Pollen size (µm) 10.01–12.53 × 11.13–13.74 (11.73 ± 0.84 × 12.61 ± 0.70) 8.58–15–85 × 10.15–16.48 (12.18 ± 2.79 × 13.57 ± 2.06) 8.03–15.14 × 9.84–15.89 (11.22 ± 1.98 × 12.42 ± 1.60) 10.98–13.81 × 11.90–14.83 (11.84 ± 077 × 13.00 ± 0.79) 11.23–16.77 × 11.87–17.76 (13.31 ± 1.47 × 13.96 ± 1.53) 11.33–15.11 × 11.82–16.42 (12.85 ± 1.28 × 13.99 ± 1.95) 11.20–15.62 × 11.62–16.03 (13.38 ± 1.82 × 13.69 ± 1.94) 11.20–13.62 × 12.34–14.47 (11.81 ± 0.87 × 13.35 ± 0.87b) Microechinus height (µm) 0.34–0.83 (0.59 ± 0.13) 0.33–0.69 (0.47 ± 0.09) 0.39–0.61 (0.47 ± .05) 0.16–0.36 (0.26 ± 0.06) 0.60–0.89 (0.72 ± 0.08) 0.22–0.78 (0.45 ± 0.16) 0.38–1.09 (0.71 ± 0.25) 0.23–0.36 (0.31 ± 0.04b) Microechinus base diameter (µm) 0.28–0.59 (0.47 ± 0.9) 0.29–0.75 (0.51 ± 0.09) 0.36–0.79 (0.48 ± 0.09) 0.21–0.67 (0.40 ± 0.13) 0.78–1.22 (0.93 ± 0.13) 0.38–1.34 (0.82 ± 0.28) 0.47–1.72 (0.99 ± 0.41) 0.34–0.58 (0.45 ± 0.08b) Distance between microechini (µm) 1.03–1.81 (1.40 ± 0.25) 0.93–2.11 (1.42 ± 0.29) 0.55–2.32 (1.42 ± 0.47) 0.84–2.15 (1.41 ± 0.24) 1.15–1.93 (1.47 ± 0.26) 0.78–2.33 (1.42 ± 0.40) 1.04–2.30 (1.45 ± 0.34) 1.15–2.18 (1.62 ± 0.39b) Number of microechini per 25 µm2 (microechini/25 µm2) 14–19 (17.05 ± 1.50) 13–18 (15.55 ± 1.43) 10–16 (12.15 ± 1.46) 10–15 (12.75 ± 1.52) 8–11 (8.85 ± 0.81) 14–19 (16.00 ± 1.30) 9–14 (11.75 ± 1.25) 10–13 (11.14 ± 1.07) Number of microechini in each raised exine region 1–5 1–8 1–3 1–6 1–3 2–11 1–3 1–3b Acetolyzed pollen grains showing no (or inapparent) ridge and raised exine region (SEM) Pollen size (µm) 11.80–16.18 × 13.51–16.78 (13.94 ± 1.54 × 15.34 ± 0.95) 8.34–14.65 × 10.49–16.68 (11.66 ± 2.5 × 13.46 ± 2.10) 9.33–15.56 × 11.52–16.36 (12.11 ± 1.77 × 13.88 ± 1.67) 10.03–13.50 × 11.95–14.92 (11.82 ± 1.10 × 13.22 ± 0.82a) - 10.56–15.80 × 11.51–17.88 (12.91 ± 1.90 × 14.63 ± 2.53) 11.15–16.56 × 11.85–17.00 (13.30 ± 1.73 × 13.98 ± 1.80) 10.27–15.61 × 11.99–21.70 (12.60 ± 1.66 × 14.38 ± 3.41) Microechinus height (µm) 0.18–0.42 (0.28 ± 0.07) 0.37–0.69 (0.54 ± 0.07) 0.28–0.71 (0.48 ± 0.09) 0.30–0.41 (0.35 ± 0.04a) - 0.15–0.78 (0.42 ± 0.19) 0.22–0.75 (0.44 ± 0.13) 0.16–0.49 (0.32 ± 0.09) Microechinus base diameter (µm) 0.32–0.64 (0.45 ± 0.09) 0.46–0.75 (0.59 ± 0.07) 0.29–0.67 (0.51 ± 0.80) 0.38–0.67 (0.55 ± 0.08a) - 0.21–0.75 (0.45 ± 0.13) 0.38–0.76 (0.53 ± 0.12) 0.26–0.51 (0.40 ± 0.07) Distance between microechini (µm) 0.95–2.31 (1.63 ± 0.36) 1.00–2.89 (1.71 ± 0.40) 1.12–2.58 (1.71 ± 0.38) 1.22–2.29 (1.73 ± 0.29a) - 1.29–2.68 (1.86 ± 0.40) 1.32–2.93 (1.92 ± 0.49) 1.28–2.10 (1.64 ± 0.19) Number of microechini per 25 µm2 (microechini/25 µm2) 10 -14 (12.45 ± 1.23) 10–15 (11.40 ± 1.27) 9–13 (10.35 ± 1.35) 10–13 (11.82 ± 0.75 a) - 6–12 (8.95 ± 1.76) 6–10 (8.20 ± 1.15) 7–13 (9.60 ± 1.67) Species G. gnemon G. montanum G. latifolium G. leptostachyum G. macrostachyum G. tenuifolium G. cuspidatum G. microcarpum Acetolyzed pollen grains showing apparent ridge and raised exine region (SEM) Pollen size (µm) 10.01–12.53 × 11.13–13.74 (11.73 ± 0.84 × 12.61 ± 0.70) 8.58–15–85 × 10.15–16.48 (12.18 ± 2.79 × 13.57 ± 2.06) 8.03–15.14 × 9.84–15.89 (11.22 ± 1.98 × 12.42 ± 1.60) 10.98–13.81 × 11.90–14.83 (11.84 ± 077 × 13.00 ± 0.79) 11.23–16.77 × 11.87–17.76 (13.31 ± 1.47 × 13.96 ± 1.53) 11.33–15.11 × 11.82–16.42 (12.85 ± 1.28 × 13.99 ± 1.95) 11.20–15.62 × 11.62–16.03 (13.38 ± 1.82 × 13.69 ± 1.94) 11.20–13.62 × 12.34–14.47 (11.81 ± 0.87 × 13.35 ± 0.87b) Microechinus height (µm) 0.34–0.83 (0.59 ± 0.13) 0.33–0.69 (0.47 ± 0.09) 0.39–0.61 (0.47 ± .05) 0.16–0.36 (0.26 ± 0.06) 0.60–0.89 (0.72 ± 0.08) 0.22–0.78 (0.45 ± 0.16) 0.38–1.09 (0.71 ± 0.25) 0.23–0.36 (0.31 ± 0.04b) Microechinus base diameter (µm) 0.28–0.59 (0.47 ± 0.9) 0.29–0.75 (0.51 ± 0.09) 0.36–0.79 (0.48 ± 0.09) 0.21–0.67 (0.40 ± 0.13) 0.78–1.22 (0.93 ± 0.13) 0.38–1.34 (0.82 ± 0.28) 0.47–1.72 (0.99 ± 0.41) 0.34–0.58 (0.45 ± 0.08b) Distance between microechini (µm) 1.03–1.81 (1.40 ± 0.25) 0.93–2.11 (1.42 ± 0.29) 0.55–2.32 (1.42 ± 0.47) 0.84–2.15 (1.41 ± 0.24) 1.15–1.93 (1.47 ± 0.26) 0.78–2.33 (1.42 ± 0.40) 1.04–2.30 (1.45 ± 0.34) 1.15–2.18 (1.62 ± 0.39b) Number of microechini per 25 µm2 (microechini/25 µm2) 14–19 (17.05 ± 1.50) 13–18 (15.55 ± 1.43) 10–16 (12.15 ± 1.46) 10–15 (12.75 ± 1.52) 8–11 (8.85 ± 0.81) 14–19 (16.00 ± 1.30) 9–14 (11.75 ± 1.25) 10–13 (11.14 ± 1.07) Number of microechini in each raised exine region 1–5 1–8 1–3 1–6 1–3 2–11 1–3 1–3b Acetolyzed pollen grains showing no (or inapparent) ridge and raised exine region (SEM) Pollen size (µm) 11.80–16.18 × 13.51–16.78 (13.94 ± 1.54 × 15.34 ± 0.95) 8.34–14.65 × 10.49–16.68 (11.66 ± 2.5 × 13.46 ± 2.10) 9.33–15.56 × 11.52–16.36 (12.11 ± 1.77 × 13.88 ± 1.67) 10.03–13.50 × 11.95–14.92 (11.82 ± 1.10 × 13.22 ± 0.82a) - 10.56–15.80 × 11.51–17.88 (12.91 ± 1.90 × 14.63 ± 2.53) 11.15–16.56 × 11.85–17.00 (13.30 ± 1.73 × 13.98 ± 1.80) 10.27–15.61 × 11.99–21.70 (12.60 ± 1.66 × 14.38 ± 3.41) Microechinus height (µm) 0.18–0.42 (0.28 ± 0.07) 0.37–0.69 (0.54 ± 0.07) 0.28–0.71 (0.48 ± 0.09) 0.30–0.41 (0.35 ± 0.04a) - 0.15–0.78 (0.42 ± 0.19) 0.22–0.75 (0.44 ± 0.13) 0.16–0.49 (0.32 ± 0.09) Microechinus base diameter (µm) 0.32–0.64 (0.45 ± 0.09) 0.46–0.75 (0.59 ± 0.07) 0.29–0.67 (0.51 ± 0.80) 0.38–0.67 (0.55 ± 0.08a) - 0.21–0.75 (0.45 ± 0.13) 0.38–0.76 (0.53 ± 0.12) 0.26–0.51 (0.40 ± 0.07) Distance between microechini (µm) 0.95–2.31 (1.63 ± 0.36) 1.00–2.89 (1.71 ± 0.40) 1.12–2.58 (1.71 ± 0.38) 1.22–2.29 (1.73 ± 0.29a) - 1.29–2.68 (1.86 ± 0.40) 1.32–2.93 (1.92 ± 0.49) 1.28–2.10 (1.64 ± 0.19) Number of microechini per 25 µm2 (microechini/25 µm2) 10 -14 (12.45 ± 1.23) 10–15 (11.40 ± 1.27) 9–13 (10.35 ± 1.35) 10–13 (11.82 ± 0.75 a) - 6–12 (8.95 ± 1.76) 6–10 (8.20 ± 1.15) 7–13 (9.60 ± 1.67) View Large Table 3. Morphology of non-acetolyzed pollen grains from studied Gnetum species under SEM (an = 2; bn = 8). The average value and standard deviation are shown in parentheses. Species G. gnemon G. montanum G. latifolium G. leptostachyum G. macrostachyum G. tenuifolium G. cuspidatum G. microcarpum Dry-treated pollen grains showing apparent ridge and raised exine region (SEM) Pollen size (µm) 8.27–12.17 × 8.71–12.48 (9.91 ± 1.23 × 10.84 ± 1.18) 9.05–15.55 × 10.15–16.23 (12.27 ± 2.70 × 13.33 ± 2.07) 8.54–14.2 × 10.31–16.46 (11.07 ± 1.74 × 12.61 ± 1.91) 10.54–12.90 × 11.46–14.10 (11.74 ± 0.56 × 12.70 ± 0.60) 11.04–13.43 × 12.06–14.30 (11.90 ± 0.70 × 12.75 ± 0.61) 9.72–16.29 × 10.45–17.06 (13.06 ± 2.46 × 13.86 ± 2.55) 11.01–15.62 × 12.61–16.07 (13.34 ± 1.94 × 14.36 ± 1.39 9.70–12.51 × 10.88–12.68 (10.61 ± 0.85 × 11.68 ± 0.37) Microechinus height (µm) 0.16–0.44 (0.31 ± 0.07) 0.57–0.81 (0.70 ± 0.08) 0.36–0.61 (0.52 ± 0.08) 0.25–0.75 (0.46 ± 0.11) 0.44–0.83 (0.63 ± 0.11) 0.17–0.75 (0.42 ± 0.17) 0.51–1.05 (0.77 ± 0.16) 0.47–0.89 (0.61 ± 0.13) Microechinus base diameter (µm) 0.26–0.89 (0.51 ± 0.14) 0.44–0.82 (0.62 ± 0.11) 0.28–0.60 (0.46 ± 0.07) 0.47–1.34 (0.79 ± 0.20) 0.70–1.26 (0.98 ± 0.16) 0.34–1.20 (0.71 ± 0.32) 0.39–1.71 (1.02 ± 0.35) 0.71–1.31 (0.88 ± 0.14) Distance between microechini (µm) 0.23–1.04 (0.70 ± 0.24) 0.73–2.10 (1.26 ± 0.33) 0.40–2.04 (1.14 ± 0.61) 0.73–1.80 (1.26 ± 0.25) 0.94–1.81 (1.18 ± 0.23) 0.64–1.60 (1.28 ± 0.26) 1.06–1.88 (1.44 ± 0.23) 0.78–1.56 (1.09 ± 0.23) Number of microechini per 25 µm2 (microechini/25 µm2) 15–22 (19.45 ± 2.16) 12–16 (13.75 ± 1.25) 10–32 (13.65 ± 5.46) 12–17 (15.00 ± 1.26) 10–14 (12.25 ± 1.07) 12–20 (16.25 ± 2.27) 8–14 (11.50 ± 1.50) 13–21 (17.35 ± 1.78) Number of microechini in each raised exine region 1–4 1–5 1–5 1–7 1–2 1–4 1–4 1–3 Rehydrated pollen grains showing no (or inapparent) ridge and raised exine region (SEM) Pollen size (µm) 10.36–13.11 × 13.14–16.69 (12.17 ± 0.69 × 14.08 ± 0.89) 9.44–15.21 × 10.98–16.71 (12.75 ± 2.00 × 14.32 ± 1.94) 9.15-12-93 × 11.51–16.58 (11.13 ± 1.12 × 12.88 ± 0.96) 10.36–12.37 × 11.60–13.61 (11.43 ± 0.57 × 12.93 ± 0.65) 12.22–13.03 × 12.72–13.27 (12.63 ± 0.56 × 13.00 ± 0.39a) 10.35–16.20 × 11.71–17.85 (12.79 ± 1.98 × 14.59 ± 2.51) 11.96–16.56 × 12.31–16.88 (14.24 ± 1.98 × 14.68 ± 1.97b) 11.51–18.85 × 12.01–21.90 (13.99 ± 2.59 × 15.85 ± 3.84) Microechinus height (µm) 0.19–0.42 (0.32 ± 0.06) 0.21–0.48 (0.35 ± 0.08) 0.32–0.64 (0.52 ± 0.09) 0.22–0.45 (0.35 ± 0.06) 0.47–0.52 (0.50 ± 0.04a) 0.14–0.41 (0.26 ± 0.07) 0.33–0.72 (0.50 ± 0.13b) 0.15–0.49 (0.34 ± 0.08) Microechinus base diameter (µm) 0.29–0.63 (0.45 ± 0.09) 0.35–0.62 (0.47 ± 0.08) 0.38–0.77 (0.55 ± 0.11) 0.43–0.83 (0.67 ± 0.11) 0.88–0.91 (0.90 ± 0.02a) 0.2–0.7 (0.47 ± 0.13) 0.48–1.11 (0.75 ± 0.21b) 0.19–0.56 (0.40 ± 0.09) Distance between microechini (µm) 0.85–2.62 (1.59 ± 0.53) 1.09–2.78 (1.87 ± 0.43) 1.11–2.56 (1.78 ± 0.43) 1.07–2.31 (1.61 ± 0.28) 2.07–2.69 (2.38 ± 0.44a) 1.38–2.50 (1.84 ± 0.32) 1.40–2.29 (1.90 ± 0.33b) 1.19–2.11 (1.62 ± 0.25) Number of microechini per 25 µm2 (microechini/25 µm2) 11–16 (13.40 ± 1.14) 11–14 (12.70 ± 0.98) 10 -14 (11.50 ± 1.00) 9–13 (10.85 ± 1.09) 10–11 (10.50 ± 0.71a) 7–12 (8.60 ± 1.19) 8–12 (9.87 ± 1.55) 11–15 (12.85 ± 1.04) Species G. gnemon G. montanum G. latifolium G. leptostachyum G. macrostachyum G. tenuifolium G. cuspidatum G. microcarpum Dry-treated pollen grains showing apparent ridge and raised exine region (SEM) Pollen size (µm) 8.27–12.17 × 8.71–12.48 (9.91 ± 1.23 × 10.84 ± 1.18) 9.05–15.55 × 10.15–16.23 (12.27 ± 2.70 × 13.33 ± 2.07) 8.54–14.2 × 10.31–16.46 (11.07 ± 1.74 × 12.61 ± 1.91) 10.54–12.90 × 11.46–14.10 (11.74 ± 0.56 × 12.70 ± 0.60) 11.04–13.43 × 12.06–14.30 (11.90 ± 0.70 × 12.75 ± 0.61) 9.72–16.29 × 10.45–17.06 (13.06 ± 2.46 × 13.86 ± 2.55) 11.01–15.62 × 12.61–16.07 (13.34 ± 1.94 × 14.36 ± 1.39 9.70–12.51 × 10.88–12.68 (10.61 ± 0.85 × 11.68 ± 0.37) Microechinus height (µm) 0.16–0.44 (0.31 ± 0.07) 0.57–0.81 (0.70 ± 0.08) 0.36–0.61 (0.52 ± 0.08) 0.25–0.75 (0.46 ± 0.11) 0.44–0.83 (0.63 ± 0.11) 0.17–0.75 (0.42 ± 0.17) 0.51–1.05 (0.77 ± 0.16) 0.47–0.89 (0.61 ± 0.13) Microechinus base diameter (µm) 0.26–0.89 (0.51 ± 0.14) 0.44–0.82 (0.62 ± 0.11) 0.28–0.60 (0.46 ± 0.07) 0.47–1.34 (0.79 ± 0.20) 0.70–1.26 (0.98 ± 0.16) 0.34–1.20 (0.71 ± 0.32) 0.39–1.71 (1.02 ± 0.35) 0.71–1.31 (0.88 ± 0.14) Distance between microechini (µm) 0.23–1.04 (0.70 ± 0.24) 0.73–2.10 (1.26 ± 0.33) 0.40–2.04 (1.14 ± 0.61) 0.73–1.80 (1.26 ± 0.25) 0.94–1.81 (1.18 ± 0.23) 0.64–1.60 (1.28 ± 0.26) 1.06–1.88 (1.44 ± 0.23) 0.78–1.56 (1.09 ± 0.23) Number of microechini per 25 µm2 (microechini/25 µm2) 15–22 (19.45 ± 2.16) 12–16 (13.75 ± 1.25) 10–32 (13.65 ± 5.46) 12–17 (15.00 ± 1.26) 10–14 (12.25 ± 1.07) 12–20 (16.25 ± 2.27) 8–14 (11.50 ± 1.50) 13–21 (17.35 ± 1.78) Number of microechini in each raised exine region 1–4 1–5 1–5 1–7 1–2 1–4 1–4 1–3 Rehydrated pollen grains showing no (or inapparent) ridge and raised exine region (SEM) Pollen size (µm) 10.36–13.11 × 13.14–16.69 (12.17 ± 0.69 × 14.08 ± 0.89) 9.44–15.21 × 10.98–16.71 (12.75 ± 2.00 × 14.32 ± 1.94) 9.15-12-93 × 11.51–16.58 (11.13 ± 1.12 × 12.88 ± 0.96) 10.36–12.37 × 11.60–13.61 (11.43 ± 0.57 × 12.93 ± 0.65) 12.22–13.03 × 12.72–13.27 (12.63 ± 0.56 × 13.00 ± 0.39a) 10.35–16.20 × 11.71–17.85 (12.79 ± 1.98 × 14.59 ± 2.51) 11.96–16.56 × 12.31–16.88 (14.24 ± 1.98 × 14.68 ± 1.97b) 11.51–18.85 × 12.01–21.90 (13.99 ± 2.59 × 15.85 ± 3.84) Microechinus height (µm) 0.19–0.42 (0.32 ± 0.06) 0.21–0.48 (0.35 ± 0.08) 0.32–0.64 (0.52 ± 0.09) 0.22–0.45 (0.35 ± 0.06) 0.47–0.52 (0.50 ± 0.04a) 0.14–0.41 (0.26 ± 0.07) 0.33–0.72 (0.50 ± 0.13b) 0.15–0.49 (0.34 ± 0.08) Microechinus base diameter (µm) 0.29–0.63 (0.45 ± 0.09) 0.35–0.62 (0.47 ± 0.08) 0.38–0.77 (0.55 ± 0.11) 0.43–0.83 (0.67 ± 0.11) 0.88–0.91 (0.90 ± 0.02a) 0.2–0.7 (0.47 ± 0.13) 0.48–1.11 (0.75 ± 0.21b) 0.19–0.56 (0.40 ± 0.09) Distance between microechini (µm) 0.85–2.62 (1.59 ± 0.53) 1.09–2.78 (1.87 ± 0.43) 1.11–2.56 (1.78 ± 0.43) 1.07–2.31 (1.61 ± 0.28) 2.07–2.69 (2.38 ± 0.44a) 1.38–2.50 (1.84 ± 0.32) 1.40–2.29 (1.90 ± 0.33b) 1.19–2.11 (1.62 ± 0.25) Number of microechini per 25 µm2 (microechini/25 µm2) 11–16 (13.40 ± 1.14) 11–14 (12.70 ± 0.98) 10 -14 (11.50 ± 1.00) 9–13 (10.85 ± 1.09) 10–11 (10.50 ± 0.71a) 7–12 (8.60 ± 1.19) 8–12 (9.87 ± 1.55) 11–15 (12.85 ± 1.04) View Large Table 3. Morphology of non-acetolyzed pollen grains from studied Gnetum species under SEM (an = 2; bn = 8). The average value and standard deviation are shown in parentheses. Species G. gnemon G. montanum G. latifolium G. leptostachyum G. macrostachyum G. tenuifolium G. cuspidatum G. microcarpum Dry-treated pollen grains showing apparent ridge and raised exine region (SEM) Pollen size (µm) 8.27–12.17 × 8.71–12.48 (9.91 ± 1.23 × 10.84 ± 1.18) 9.05–15.55 × 10.15–16.23 (12.27 ± 2.70 × 13.33 ± 2.07) 8.54–14.2 × 10.31–16.46 (11.07 ± 1.74 × 12.61 ± 1.91) 10.54–12.90 × 11.46–14.10 (11.74 ± 0.56 × 12.70 ± 0.60) 11.04–13.43 × 12.06–14.30 (11.90 ± 0.70 × 12.75 ± 0.61) 9.72–16.29 × 10.45–17.06 (13.06 ± 2.46 × 13.86 ± 2.55) 11.01–15.62 × 12.61–16.07 (13.34 ± 1.94 × 14.36 ± 1.39 9.70–12.51 × 10.88–12.68 (10.61 ± 0.85 × 11.68 ± 0.37) Microechinus height (µm) 0.16–0.44 (0.31 ± 0.07) 0.57–0.81 (0.70 ± 0.08) 0.36–0.61 (0.52 ± 0.08) 0.25–0.75 (0.46 ± 0.11) 0.44–0.83 (0.63 ± 0.11) 0.17–0.75 (0.42 ± 0.17) 0.51–1.05 (0.77 ± 0.16) 0.47–0.89 (0.61 ± 0.13) Microechinus base diameter (µm) 0.26–0.89 (0.51 ± 0.14) 0.44–0.82 (0.62 ± 0.11) 0.28–0.60 (0.46 ± 0.07) 0.47–1.34 (0.79 ± 0.20) 0.70–1.26 (0.98 ± 0.16) 0.34–1.20 (0.71 ± 0.32) 0.39–1.71 (1.02 ± 0.35) 0.71–1.31 (0.88 ± 0.14) Distance between microechini (µm) 0.23–1.04 (0.70 ± 0.24) 0.73–2.10 (1.26 ± 0.33) 0.40–2.04 (1.14 ± 0.61) 0.73–1.80 (1.26 ± 0.25) 0.94–1.81 (1.18 ± 0.23) 0.64–1.60 (1.28 ± 0.26) 1.06–1.88 (1.44 ± 0.23) 0.78–1.56 (1.09 ± 0.23) Number of microechini per 25 µm2 (microechini/25 µm2) 15–22 (19.45 ± 2.16) 12–16 (13.75 ± 1.25) 10–32 (13.65 ± 5.46) 12–17 (15.00 ± 1.26) 10–14 (12.25 ± 1.07) 12–20 (16.25 ± 2.27) 8–14 (11.50 ± 1.50) 13–21 (17.35 ± 1.78) Number of microechini in each raised exine region 1–4 1–5 1–5 1–7 1–2 1–4 1–4 1–3 Rehydrated pollen grains showing no (or inapparent) ridge and raised exine region (SEM) Pollen size (µm) 10.36–13.11 × 13.14–16.69 (12.17 ± 0.69 × 14.08 ± 0.89) 9.44–15.21 × 10.98–16.71 (12.75 ± 2.00 × 14.32 ± 1.94) 9.15-12-93 × 11.51–16.58 (11.13 ± 1.12 × 12.88 ± 0.96) 10.36–12.37 × 11.60–13.61 (11.43 ± 0.57 × 12.93 ± 0.65) 12.22–13.03 × 12.72–13.27 (12.63 ± 0.56 × 13.00 ± 0.39a) 10.35–16.20 × 11.71–17.85 (12.79 ± 1.98 × 14.59 ± 2.51) 11.96–16.56 × 12.31–16.88 (14.24 ± 1.98 × 14.68 ± 1.97b) 11.51–18.85 × 12.01–21.90 (13.99 ± 2.59 × 15.85 ± 3.84) Microechinus height (µm) 0.19–0.42 (0.32 ± 0.06) 0.21–0.48 (0.35 ± 0.08) 0.32–0.64 (0.52 ± 0.09) 0.22–0.45 (0.35 ± 0.06) 0.47–0.52 (0.50 ± 0.04a) 0.14–0.41 (0.26 ± 0.07) 0.33–0.72 (0.50 ± 0.13b) 0.15–0.49 (0.34 ± 0.08) Microechinus base diameter (µm) 0.29–0.63 (0.45 ± 0.09) 0.35–0.62 (0.47 ± 0.08) 0.38–0.77 (0.55 ± 0.11) 0.43–0.83 (0.67 ± 0.11) 0.88–0.91 (0.90 ± 0.02a) 0.2–0.7 (0.47 ± 0.13) 0.48–1.11 (0.75 ± 0.21b) 0.19–0.56 (0.40 ± 0.09) Distance between microechini (µm) 0.85–2.62 (1.59 ± 0.53) 1.09–2.78 (1.87 ± 0.43) 1.11–2.56 (1.78 ± 0.43) 1.07–2.31 (1.61 ± 0.28) 2.07–2.69 (2.38 ± 0.44a) 1.38–2.50 (1.84 ± 0.32) 1.40–2.29 (1.90 ± 0.33b) 1.19–2.11 (1.62 ± 0.25) Number of microechini per 25 µm2 (microechini/25 µm2) 11–16 (13.40 ± 1.14) 11–14 (12.70 ± 0.98) 10 -14 (11.50 ± 1.00) 9–13 (10.85 ± 1.09) 10–11 (10.50 ± 0.71a) 7–12 (8.60 ± 1.19) 8–12 (9.87 ± 1.55) 11–15 (12.85 ± 1.04) Species G. gnemon G. montanum G. latifolium G. leptostachyum G. macrostachyum G. tenuifolium G. cuspidatum G. microcarpum Dry-treated pollen grains showing apparent ridge and raised exine region (SEM) Pollen size (µm) 8.27–12.17 × 8.71–12.48 (9.91 ± 1.23 × 10.84 ± 1.18) 9.05–15.55 × 10.15–16.23 (12.27 ± 2.70 × 13.33 ± 2.07) 8.54–14.2 × 10.31–16.46 (11.07 ± 1.74 × 12.61 ± 1.91) 10.54–12.90 × 11.46–14.10 (11.74 ± 0.56 × 12.70 ± 0.60) 11.04–13.43 × 12.06–14.30 (11.90 ± 0.70 × 12.75 ± 0.61) 9.72–16.29 × 10.45–17.06 (13.06 ± 2.46 × 13.86 ± 2.55) 11.01–15.62 × 12.61–16.07 (13.34 ± 1.94 × 14.36 ± 1.39 9.70–12.51 × 10.88–12.68 (10.61 ± 0.85 × 11.68 ± 0.37) Microechinus height (µm) 0.16–0.44 (0.31 ± 0.07) 0.57–0.81 (0.70 ± 0.08) 0.36–0.61 (0.52 ± 0.08) 0.25–0.75 (0.46 ± 0.11) 0.44–0.83 (0.63 ± 0.11) 0.17–0.75 (0.42 ± 0.17) 0.51–1.05 (0.77 ± 0.16) 0.47–0.89 (0.61 ± 0.13) Microechinus base diameter (µm) 0.26–0.89 (0.51 ± 0.14) 0.44–0.82 (0.62 ± 0.11) 0.28–0.60 (0.46 ± 0.07) 0.47–1.34 (0.79 ± 0.20) 0.70–1.26 (0.98 ± 0.16) 0.34–1.20 (0.71 ± 0.32) 0.39–1.71 (1.02 ± 0.35) 0.71–1.31 (0.88 ± 0.14) Distance between microechini (µm) 0.23–1.04 (0.70 ± 0.24) 0.73–2.10 (1.26 ± 0.33) 0.40–2.04 (1.14 ± 0.61) 0.73–1.80 (1.26 ± 0.25) 0.94–1.81 (1.18 ± 0.23) 0.64–1.60 (1.28 ± 0.26) 1.06–1.88 (1.44 ± 0.23) 0.78–1.56 (1.09 ± 0.23) Number of microechini per 25 µm2 (microechini/25 µm2) 15–22 (19.45 ± 2.16) 12–16 (13.75 ± 1.25) 10–32 (13.65 ± 5.46) 12–17 (15.00 ± 1.26) 10–14 (12.25 ± 1.07) 12–20 (16.25 ± 2.27) 8–14 (11.50 ± 1.50) 13–21 (17.35 ± 1.78) Number of microechini in each raised exine region 1–4 1–5 1–5 1–7 1–2 1–4 1–4 1–3 Rehydrated pollen grains showing no (or inapparent) ridge and raised exine region (SEM) Pollen size (µm) 10.36–13.11 × 13.14–16.69 (12.17 ± 0.69 × 14.08 ± 0.89) 9.44–15.21 × 10.98–16.71 (12.75 ± 2.00 × 14.32 ± 1.94) 9.15-12-93 × 11.51–16.58 (11.13 ± 1.12 × 12.88 ± 0.96) 10.36–12.37 × 11.60–13.61 (11.43 ± 0.57 × 12.93 ± 0.65) 12.22–13.03 × 12.72–13.27 (12.63 ± 0.56 × 13.00 ± 0.39a) 10.35–16.20 × 11.71–17.85 (12.79 ± 1.98 × 14.59 ± 2.51) 11.96–16.56 × 12.31–16.88 (14.24 ± 1.98 × 14.68 ± 1.97b) 11.51–18.85 × 12.01–21.90 (13.99 ± 2.59 × 15.85 ± 3.84) Microechinus height (µm) 0.19–0.42 (0.32 ± 0.06) 0.21–0.48 (0.35 ± 0.08) 0.32–0.64 (0.52 ± 0.09) 0.22–0.45 (0.35 ± 0.06) 0.47–0.52 (0.50 ± 0.04a) 0.14–0.41 (0.26 ± 0.07) 0.33–0.72 (0.50 ± 0.13b) 0.15–0.49 (0.34 ± 0.08) Microechinus base diameter (µm) 0.29–0.63 (0.45 ± 0.09) 0.35–0.62 (0.47 ± 0.08) 0.38–0.77 (0.55 ± 0.11) 0.43–0.83 (0.67 ± 0.11) 0.88–0.91 (0.90 ± 0.02a) 0.2–0.7 (0.47 ± 0.13) 0.48–1.11 (0.75 ± 0.21b) 0.19–0.56 (0.40 ± 0.09) Distance between microechini (µm) 0.85–2.62 (1.59 ± 0.53) 1.09–2.78 (1.87 ± 0.43) 1.11–2.56 (1.78 ± 0.43) 1.07–2.31 (1.61 ± 0.28) 2.07–2.69 (2.38 ± 0.44a) 1.38–2.50 (1.84 ± 0.32) 1.40–2.29 (1.90 ± 0.33b) 1.19–2.11 (1.62 ± 0.25) Number of microechini per 25 µm2 (microechini/25 µm2) 11–16 (13.40 ± 1.14) 11–14 (12.70 ± 0.98) 10 -14 (11.50 ± 1.00) 9–13 (10.85 ± 1.09) 10–11 (10.50 ± 0.71a) 7–12 (8.60 ± 1.19) 8–12 (9.87 ± 1.55) 11–15 (12.85 ± 1.04) View Large Gnetum montanum Pollen are monads, subspheroidal, elliptic or ovate in the long axis view, circular or elliptic in the short axis view, long axis diameter 12.0–17.5 µm in LM, 10.15–16.68 µm in SEM (acetolyzed grains), 10.15–16.71 µm in SEM (non-acetolyzed grains), short axis diameter 11.0–15.0 µm in LM, 8.34–15.85 µm in SEM (acetolyzed grains), 9.05–15.55 µm in SEM (non-acetolyzed grains), inaperturate, exine 0.6–0.9 µm thick (LM), tectate, continuous tectum, microechinate in LM, microechinate with psilate or ridged surface between microechini in SEM, with blunt-ended microechini. Some grains show slightly or considerably raised exine areas enclosed by a ridge. Numbers of microechini on each raised area vary from one to eight. Microechini are small, 0.33–0.69 µm high and 0.29–0.75 µm in base diameter (acetolyzed grains) and 0.21–0.81 µm high and 0.35–0.82 µm in base diameter (non-acetolyzed grains). In an area of 25 µm2 on the tectum surface, ten to 18 and 11–16 microechini were found on acetolyzed grains and non-acetolyzed grains, respectively. The distances between microechini vary from 0.93 to 2.89 µm on acetolyzed grains and from 0.73 to 2.78 µm on non-acetolyzed grains (Tables 1–3; Figs 1B, 2D–F, 4E–H). Gnetum latifolium var. Funiculare Pollen are monads, subspheroidal, elliptic or ovate in the long axis view, circular or elliptic in the short axis view, long axis diameter 13–16 µm in LM, 9.84–16.36 µm in SEM (acetolyzed grains), 10.31–16.58 µm in SEM (non-acetolyzed grains), short axis diameter 10.0–15.0 µm in LM, 8.03–15.56 µm in SEM (acetolyzed grains), 8.54–14.20 µm in SEM (non-acetolyzed grains), inaperturate, exine 0.4–0.7 µm thick (LM), tectate, continuous tectum, microechinate in LM, microechinate with psilate or ridged surface between microechini in SEM, with a bluntly round tip. Some grains display distinct raised exine regions bordered by a ridge, with one to five microechini on each region. Microechini are 0.28–0.71 µm high and 0.29–0.79 µm in base diameter (acetolyzed grains), 0.32–0.64 µm high and 0.28–0.77 µm in base diameter (non-acetolyzed grains). There are nine to 16 microechini per surface area of 25 µm2 on acetolyzed grains and ten to 32 on non-acetolyzed grains. The distances between microechini on acetolyzed grains are 0.55–2.58 µm, whereas those on non-acetolyzed grains are 0.40–2.56 µm (Tables 1–3; Figs 1C, 2G–I, 4I–L). Gnetum leptostachyum Pollen are monads, subspheroidal, elliptic or ovate in the long axis view, circular or elliptic in the short axis view, long axis diameter 12–16 µm in LM, 11.90–14.92 µm in SEM (acetolyzed grains), 11.46–14.10 µm in SEM (non-acetolyzed grains), short axis diameter 10–14 µm in LM, 10.03–13.81 µm in SEM (acetolyzed grains), 10.36–12.90 µm in SEM (non-acetolyzed grains), inaperturate, exine 0.4–0.8 µm thick (LM), tectate, continuous tectum, microechinate in LM, microechinate with psilate or ridged surface between microechini in SEM, with blunt-ended microechini. Some grains demonstrate considerably raised exine areas surrounded by a ridge, with one to seven microechini on each area. Sometimes, the base of a microechinus is swollen forming a dome-shaped microechinus. Microechini are small, 0.16–0.41 µm high and 0.21–0.67 µm in base diameter (acetolyzed grains), 0.22–0.75 µm high and 0.43–1.34 µm in base diameter (non-acetolyzed grains). In an area of 25 µm2 on the tectum surface, ten to 15 and nine to 17 microechini were observed on acetolyzed grains and non-acetolyzed grains, respectively. The distances between microechini vary from 0.84 to 2.29 µm on acetolyzed grains and from 0.73 to 2.31 µm on non-acetolyzed grains (Tables 1–3; Figs 1D, 2J–L, 4M–P). Gnetum macrostachyum Pollen are monads, subspheroidal, elliptic or ovate in the long axis view, circular or elliptic in the short axis view, long axis diameter 12.5–16.5 µm in LM, 11.87–17.76 µm in SEM (acetolyzed grains), 12.06–14.30 µm in SEM (non-acetolyzed grains), short axis diameter 12.0–14.0 µm in LM, 11.23–16.77 µm in SEM (acetolyzed grains), 11.04–13.43 µm in SEM (non-acetolyzed grains), inaperturate, exine 0.5–0.8 µm thick (LM), tectate, continuous tectum, microechinate in LM, microechinate with psilate (not present in the acetolyzed sample) or ridged surface between microechini in SEM, with a bluntly round tip. On both acetolyzed and non-acetolyzed grains, dome-shaped microechini with blunt-ended tips were frequently observed adjacent to each other, a ridge surrounding one or a few of them, forming a network-like pattern on the pollen wall. Pollen grains with a crumpled tectum surface could be observed in the acetolyzed pollen samples. Microechini are small, 0.60–0.89 µm high and 0.78–1.22 µm in base diameter (acetolyzed grains), 0.44–0.83 µm high and 0.70–1.26 µm in base diameter (non-acetolyzed grains). There are eight to 11 microechini per surface area of 25 µm2 on acetolyzed grains and ten to 14 on non-acetolyzed grains. The distances between microechini on acetolyzed grains are 1.15–1.93 µm, whereas those on non-acetolyzed grains are 0.94–2.69 µm (Tables 1–3; Figs 1E, 3A–C, 5A–D). Gnetum tenuifolium Pollen are monads, subspheroidal, elliptic or ovate in the long axis view, circular or elliptic in the short axis view, long axis diameter 12.0–17.5 µm in LM, 11.51–17.88 µm in SEM (acetolyzed grains), 10.45–17.85 µm in SEM (non-acetolyzed grains), short axis diameter 11–15 µm in LM, 10.56–15.80 µm in SEM (acetolyzed grains), 9.72–16.29 µm in SEM (non-acetolyzed grains), inaperturate, exine 0.4–0.8 µm thick (LM), tectate, continuous tectum, microechinate in LM, microechinate with psilate (rare) or ridged surface between microechini in SEM, with blunt-ended microechini. Pollen grains with distinct and large raised exine regions surrounded by a ridge, with two to 11 microechini on each region, were frequently recognized. A dome-shaped microechinus with blunt apex was also observed. Microechini are 0.15–0.78 µm high and 0.21–1.34 µm in base diameter (acetolyzed grains), 0.14–0.75 µm high and 0.20–1.20 µm in base diameter (non-acetolyzed grains). In an area of 25 µm2 on the tectum surface, six to 19 and seven to 20 microechini were recognized on acetolyzed grains and non-acetolyzed grains, respectively. The distances between microechini vary from 0.78 to 2.68 µm on acetolyzed grains and from 0.64 to 2.50 µm on non-acetolyzed grains (Tables 1–3; Figs 1F, 3D–F, 5E–H). Figure 5. View largeDownload slide Scanning electron micrographs (SEM) of non-acetolyzed pollen grains from Gnetum species in Thailand showing the variation in sculptures (A–P). A–D, Gnetum macrostachyum. E–H, Gnetum tenuifolium. I–L, Gnetum cuspidatum. M–P, Gnetum microcarpum (A, E, I and M, dry-treated samples, general view; B, F, J and N, rehydrated samples, general view; C, G, K and O, dry-treated samples, echinate with apparent ridge and raised exine region; D, H, L and P, rehydrated samples, echinate with no, or inapparent, ridge and raised exine region). Arrows indicate a raised exine region surrounded by a ridge. An asterisk indicates a dome-shaped microechinus. Scale bars: 10 μm (A, B, E, F, I, J, M and N); 1 μm (C, D, G, H, K, L, O and P). Figure 5. View largeDownload slide Scanning electron micrographs (SEM) of non-acetolyzed pollen grains from Gnetum species in Thailand showing the variation in sculptures (A–P). A–D, Gnetum macrostachyum. E–H, Gnetum tenuifolium. I–L, Gnetum cuspidatum. M–P, Gnetum microcarpum (A, E, I and M, dry-treated samples, general view; B, F, J and N, rehydrated samples, general view; C, G, K and O, dry-treated samples, echinate with apparent ridge and raised exine region; D, H, L and P, rehydrated samples, echinate with no, or inapparent, ridge and raised exine region). Arrows indicate a raised exine region surrounded by a ridge. An asterisk indicates a dome-shaped microechinus. Scale bars: 10 μm (A, B, E, F, I, J, M and N); 1 μm (C, D, G, H, K, L, O and P). Gnetum cuspidatum Pollen are monads, subspheroidal, elliptic or ovate in the long axis view, circular or elliptic in the short axis view, long axis diameter 12–16 µm in LM, 11.62–17.00 µm in SEM (acetolyzed grains), 12.31–16.88 µm in SEM (non-acetolyzed grains), short axis diameter 11–14 µm in LM, 11.15–16.62 µm in SEM (acetolyzed grains), 11.01–16.56 µm in SEM (non-acetolyzed grains), inaperturate, exine 0.5–0.8 µm thick (LM), tectate, continuous tectum, microechinate in LM, microechinate with psilate or ridged surface between microechini in SEM, with a bluntly round tip. Some grains demonstrate slightly or considerably raised exine areas surrounded by the ridge, with one to four microechini on each area. On dry pollen samples, dome-shaped microechini appear to be more adjacent to each other than those on acetolyzed samples. Microechini are small, 0.22–1.09 µm high and 0.38–1.72 µm in base diameter (acetolyzed grains), 0.33–1.05 µm high and 0.39–1.71 µm in base diameter (non-acetolyzed grains). There are six to 14 microechini per surface area of 25 µm2 on acetolyzed grains and eight to 14 microechini per surface area of 25 µm2 on non-acetolyzed grains. The distances between microechini vary from 1.04 to 2.93 µm on acetolyzed grains and from 1.06 to 2.29 on non-acetolyzed grains (Tables 1–3; Figs 1G, 3G–I, 5I–L). Gnetum microcarpum Pollen are monads, subspheroidal, elliptic or ovate in the long axis view, circular or elliptic in the short axis view, long axis diameter 12.4–19.2 µm in LM, 11.99–21.70 µm in SEM (acetolyzed grains), 10.88–21.90 µm in SEM (non-acetolyzed grains), short axis diameter 11.0–15.4 µm in LM, 10.27–15.61 µm in SEM (acetolyzed grains), 9.70–18.85 µm in SEM (non-acetolyzed grains), inaperturate, exine 0.5–0.6 µm thick (LM), tectate, microechinate in LM, microechinate with a psilate or (rarely) ridged surface between microechini in SEM, with a bluntly round tip. Some grains demonstrate slightly raised exine areas enclosed by a ridge, with one to three microechini on each area. Microechini are small, 0.16–0.49 µm high and 0.26–0.58 µm in base diameter (acetolyzed grains), 0.15–0.89 µm high and 0.19–1.31 µm in base diameter (non-acetolyzed grains). In an area of 25 µm2 on the tectum surface, seven to 13 and 11–21 microechini were observed on acetolyzed grains and non-acetolyzed grains, respectively. The distances between microechini vary from 1.15 to 2.18 µm on acetolyzed grains and from 0.78 to 2.11 on non-acetolyzed grains (Tables 1–3; Figs 1H, 3J–L, 5M–P). Raised exine regions enclosed by ridges were found in both dry-treated and dry-untreated pollen grains of all Gnetum spp. (Figs 4A, C, E, G, I, K, M and O, 5A, C, E, G, I, K, M and O and 6A–P). Figure 6. View largeDownload slide Scanning electron micrographs (SEM) of dry-untreated pollen grains from Gnetum species in Thailand showing the typical characters of the dehydrated state (A–P). A–B, Gnetum gnemon. C–D, Gnetum montanum. E–F, Gnetum latifolium. G–H, Gnetum leptostachyum. I–J, Gnetum macrostachyum. K–L, Gnetum tenuifolium. M–N, Gnetum cuspidatum. O–P, Gnetum microcarpum (A, C, E, G, I, K, M and O, general view; B, D, F, H, J, L, N and P, echinate with apparent ridge and raised exine region. Arrows indicate a raised exine region surrounded by a ridge. An asterisk indicates a dome-shaped microechinus. Scale bars: 10 μm (A, C, E, G, I, K, M and O); 1 μm (B, D, F, H, J, L, N and P). Figure 6. View largeDownload slide Scanning electron micrographs (SEM) of dry-untreated pollen grains from Gnetum species in Thailand showing the typical characters of the dehydrated state (A–P). A–B, Gnetum gnemon. C–D, Gnetum montanum. E–F, Gnetum latifolium. G–H, Gnetum leptostachyum. I–J, Gnetum macrostachyum. K–L, Gnetum tenuifolium. M–N, Gnetum cuspidatum. O–P, Gnetum microcarpum (A, C, E, G, I, K, M and O, general view; B, D, F, H, J, L, N and P, echinate with apparent ridge and raised exine region. Arrows indicate a raised exine region surrounded by a ridge. An asterisk indicates a dome-shaped microechinus. Scale bars: 10 μm (A, C, E, G, I, K, M and O); 1 μm (B, D, F, H, J, L, N and P). DISCUSSION Pollen morphology of Gnetum species Pollen from all Gnetum spp. studied were monads, subspheroidal, elliptic or ovate in the long axis view, circular or elliptic in the short axis view, inaperturate, tectate with a continuous tectum, and microechinate with psilate or ridged surface between microechini. Their morphology is generally similar to those of Asian Gnetum spp. described in previous studies, including G. cleistostachyum C.Y.Cheng, G. costatum K.Schum., G. cuspidatum, G. gnemon, G. hainanense C.Y.Cheng, G. latifolium var. funiculare (as G. funiculare Blume), G. leptostachyum, G. luofuense C.Y.Cheng, G. macrostachyum, G. montanum, G. neglectum Blume, G. parvifolium (Warb.) Cheng (=G. indicum Merr.) and G. pendulum C.Y.Cheng (Erdtman 1954, 1965; Hesse, 1980; Gillespie & Nowicke, 1994; Osborn, 2000; Yao et al., 2004; Tekleva, 2016). Even though, based on plant morphology, G. gnemon and G. costatum are placed in a different section to other Asian taxa (Griffith, 1859; Markgraf, 1929; Price, 1996), their pollen present a continuous tectum with microechini, which is similar to those of other Asian Gnetum spp. (Figs 1–6; Gillespie & Nowicke, 1994, plate I, fig. 1). Thus, this character supports the molecular phylogenetic information from studies by Won & Renner (2006), Hou et al. (2015) and Hou et al. (2016), revealing Asian Gnetum taxa to be monophyletic. In other words, it indicates that Gnetum section Gnetum subsection Eugnemones is more closely related to Gnetum section Scandentia than the other two groups in its own section [subsections Araeognemones (American taxa) and Micrognemones (African taxa)], which show a discontinuous tectum forming islands with numerous microechini (Tekleva, 2016, fig. 1a, b; Gillespie & Nowicke, 1994, plate II, figs 7–12). From the different tectal ornamentations observed under LM, Erdtman (1954, 1965) proposed three different groups of Gnetum pollen with non-overlapping geographical distributions, Asian, American and African. This proposal provides support for the three major phylogenetic clusters put forward by Hou et al. (2015). However, Gillespie & Nowicke (1994) showed that, in SEM, the tectal ornamentation of Gnetum pollen from the examined American species was not pilate as shown in those previous LM studies, being more similar to those of African taxa, plate II, figs 7–12). According to Hou et al. (2015), Gnetum diverged in the Late Cretaceous when the breakup of Gondwana occurred and the Asian crown group was dated to the Cretaceous-Palaeogene (K-Pg) boundary. These findings disagree with the hypothesis proposed by Markgraf (1929) that arborescent Gnetum taxa appeared to be an earlier diverged ancestor of other species in the genus. The discovery of the first fossil male strobiles with an affinity to Gnetum, Khitania columnispicata Guo, Sha, Bian & Qiu, from the Early Cretaceous in the Yixian Formation, Liaoning, China also confirms the results from the molecular study (Guo et al., 2009). However, fossil palynomorphs comparable to pollen of extant Gnetum taxa have not yet been reported (Friis et al., 2011). Such disappearances from the geological record are possibly associated with their thin ectexine layer (Tekleva, 2016). Pollen of all examined Gnetum spp. displayed morphological plasticity. Some grains exhibited a psilate surface between microechini, whereas others showed slightly or considerably raised exine regions surrounded by various patterns of ridge-like folds, with one or more microechini on each region. Other studies on Asian Gnetum pollen also found similar raised regions and folding (Gillespie & Nowicke, 1994; Osborn, 2000; Yao et al., 2004; Tekleva, 2016) but only Gillespie & Nowicke (1994; plate I, figs 2–3) reported the variation in tectum ornamentation of Gnetum pollen and also pointed out that the different sculptures observed could be caused by the sample preparation. However, two months after the first examination was carried out, we re-examined the acetolyzed grains with psilate surfaces between the microechini and found that the exines were slightly raised and ridged, and that the same characters that indicate the dehydrated state of Gnetum pollen grains were present in the dry-treated samples of all species as well as in the dry-untreated pollen grains. Thus, it can be concluded that, rather than sample preparation techniques, such morphological plasticity could be associated with harmomegathy, which we will discuss in the next section. Even though the exine ultrastructures of pollen from Ephedra L., Gnetum and Welwitschia Hook.f. are similar, being described as granular infratectum, lamellate endexine, tectum protruding outwards, their exine ornamentations are different, being plicate in Ephedra and Welwitschia and microechinate in Gnetum (Gullvåg, 1966; Hesse, 1980, 1984; Osborn, Taylor & de Lima, 1993; Gillespie & Nowicke, 1994; Ickert-Bond, Skvarla & Chissoe, 2003; Yao et al., 2004; Bolinder, Niklas & Rydin, 2015). Since this ornamentation of Gnetum pollen is unique among gymnosperms (Gillespie & Nowicke, 1994), it appears to be a potential synapomorphy for this plant group. Tekleva (2016) stated that in Asian Gnetum spp., raised exine areas with microechini are equivalent to the plate-like exine wall of G. africanum Welw. and the plicae of Ephedra and Welwitschia. Gnetum have been considered entomophilous, although various anemophilous syndromes have been recognized (Kato, Inoue & Nagamitsu, 1995; Kato, 1996; Momose, et al. 1998; Kato et al., 2008; Gong et al., 2015). It is possible that the adaptation of the Gnetum lineage reached an entomophilous pollination mode along a different line than Ephedra and Welwitschia (Tekleva, 2016). Jörgensen & Rydin (2015) proposed that the loss of sterile ovules in African Gnetum is probably linked to wind pollination. However, the correlation remains unclear between wind pollination and a discontinuous tectum as the pollination biology of American species that also produce pollen with a discontinuous tectum has not yet been investigated. A cup-like depression was found in some grains under LM and SEM in our study but without any indication of its being an exit for a pollen tube. This depression was interpreted as a leptoma (pore-like area) by Erdtman (1965) and Kuprianova (1983). Gillespie & Nowicke (1994), Yao et al., (2004) and Tekleva (2016) also found, under SEM and TEM, no evidence of thinning on the exine wall to indicate the existence of an apertural region in Gnetum pollen. In this study, a considerable overlap in qualitative and quantitative morphological characters exists among species and also among samples of the same species prepared by different techniques. Therefore, it is not possible to classify Thai Gnetum pollen into species, particularly because the pollen has a psilate surface between microechini. However, some differences could be noticed. Pollen from G. gnemon var. tenerum in this study shows the highest average values of grain size (in LM), and number of microechini within a tectum surface area of 25 µm2 and the lowest average values of distance between microechini. These results imply that pollen of this taxon are large with numerous microechini. Moreover, under LM, their microechini are less prominent than those of others. The study by Osborn (2000), based on SEM, described G. gnemon pollen as spheroidal in shape, whereas they appeared subspheroidal in our LM investigation. Since a coverslip was not used in our study, the pollen could be turned over with a hair-tipped needle. As a result, we could obtain the correct position for measuring and photographing the pollen under LM. Due to the impossibility of moving pollen grains during SEM analysis, the dimensions of pollen measured under SEM differ from their dimensions measured under LM. Yao et al. (2004) separated Gnetum pollen into two types according to their spinule (echinus) characters: one with inflated bases in G. parvifolium and G. cleistostachyum, and one without inflated bases in G. montanum, G. luofuense, G. pendulum and G. hainanens. However, in our analysis, both types were found in all examined pollen species due to their harmomegathic mechanism, although some different patterns of inflated areas could be recognized. In this study, the inflated bases of the microechini of pollen from G. gnemon var. tenerum, G. latifolium var. funiculare, G. leptostachyum, G. montanum and G. tenuifolium tended to merge with the bases of adjacent microechini, forming a raised exine region enclosed by a ridge with more than two microechini on each area (Figs 2B, E, H and K, 3E, 4C, G, K and O, 5G, 6A–H, K and L). This morphology was also found in pollen from G. gnemon, G. latifolium var. funiculare, G. leptostachyum and G. montanum studied by Gillespie & Nowicke (1994, plate I, figs 4–5), Halbritter (2015) and Tekleva (2016, fig. 1c–d). In the present study, numerous microechini on a large, merged, inflated base comprised a morphological character found only in pollen grains from G. tenuifolium (Figs 3D and E, 5E and G, 6K and L). The same morphological feature was also observed in pollen samples from different plants of this taxon. However, a single dome-shaped microechinus with an isolated, inflated base on a raised area were also found in all four species, more especially in dry pollen samples (both treated and untreated) from G. leptostachyum and G. tenuifolium. The dome-shaped microechinus in our study, was more frequently observed in pollen from G. macrostachyum, G. cuspidatum and G. microcarpum (Figs 3B, H and K, 5C, K and O, 6I–J and M–P). One or a few dome-shaped microechini of G. macrostachyum pollen were bordered by a prominent ridge, resulting in the formation of a network-like sculpture on the pollen wall. These characters are similar to those of G. macrostachyum shown in Tekleva (2016, fig. 1 h–i). However, similar network-like patterns of ridges were also found in pollen from G. indicum (Tekleva, 2016, fig. 1e–f) and G. latifolium (Gillespie & Nowicke, 1994, plate I, fig. 3) but without the inflated base of the microechinus. Likewise, in this present study, the base of microechini of acetolyzed pollen of G. cuspidatum and G. microcarpum, and of the dry pollen samples of the latter species, appear slightly inflated and ridged. This was also recognized in G. cuspidatum pollen examined by Gillespie & Nowicke (1994, plate I, fig. 6). Even though the morphologies of both dry-treated and dry-untreated pollen samples appeared to be similar, differences in some specific characters were observed. In G. tenuifolium, for example, the islands on the dry-treated pollen tended to be more distant from each other than they were on the dry-untreated pollen, even though both samples were taken from the same plant material (Figs 5E, G, 6K, L). This suggested that the variation in the characters of the dehydrated pollen could have been produced by the sample preparation methods and by the age of the specimens. Therefore, it should be ensured that the same technique is employed when a comparative study is made. However, different characters of dry pollen from G. latifolium were recognized even when the same preparation procedure was used (Fig. 6E, F; Tekleva, 2016, fig. 1c–d). Thus, it should be emphasized that intraspecific variation should also be taken into account when investigating pollen of this plant group. THE HARMOMEGATHIC CHARACTER In nature, pollen pass through many developmental stages that differ in relation to their hydrated and dehydrated states. These stages include the microsporogenesis phase (microspore or pollen immersed in locular fluid), the anther dehydration phase (the locular fluid disappears), the presentation phase (the anther opens), the dispersal phase (pollen are exposed to the environment) and the pollen-stigma interaction phase (pollen rehydrate on the stigma) (Firon, Nepi & Pacini, 2012). During these phases, the exine elasticity of pollen adapts to facilitate harmomegathy (Mohl, 1834). In this study, the average values of pollen size and distance between microechini of non-acetolyzed grains with psilate surfaces between microechini (rehydrated pollen) tended to be higher than those of non-acetolyzed grains with apparent ridges and raised exine areas (dry pollen), whereas the latter grains had higher average values for the microechinus height and base diameter and for the number of microechini in the defined area. The same tendency also existed in the acetolyzed pollen samples. In other words, pollen grains with apparent ridges and raised exine regions usually had smaller volumes, suggesting that these two features are possibly indicative of the dehydrated state of Asian Gnetum pollen, whereas a psilate surface between microechini indicates its hydrated state. Previous studies pointed out that Gnetum pollen grains contract when dried leading to folding and crumpling of their tectum surface, providing a potential way to protect themselves from damage due to changes in their volume (Wodehouse, 1935; Tekleva, 2016). Numerous examples of ‘dry Gnetum pollen samples’ with apparent ridges and/or inflated exine areas have been shown in many studies (Yao et al., 2004; Halbritter, 2015; Tekleva, 2016). In two other gnetalean plants, Ephedra and Welwitschia, even though their pollen is polyplicate, which already provides a harmomegathic function (Traverse, 2007), ridge–like folds similar to those of Gnetum were also found on their plicae (Ickert-Bond, Skvarla & Chissoe, 2003, plate I, fig. 5, plate II, fig. 12, 13; Doores, Osborn & El-Ghazaly, 2007, fig. 8C; Bolinder et al., 2015, figs 4K, 5A, C, F, 6A; Halbritter, 2016). Moreover, two types of ridge, straight and sinuous, were discovered in Ephedra pollen (Ickert-Bond, Skvarla & Chissoe, 2003, plate II, figs 12, 13), whereas in Welwitschia pollen, a hinge-like structure was observed directly over the furrow (Osborn, 2000, fig. 13). Such folding structures were also recognized in fossil ephedroid pollen (Osborn et al., 1993, plate II, fig. 4–6). Theoretically, pollen exhibit morphological alterations, including exine expansion and contraction, folding of the exine and modification of the lacuna volume to accommodate differences in internal osmotic pressure during wet and dry cycles (Wodehouse, 1935; Blackmore & Barnes, 1986; Hesse et al., 2009). The first two processes are likely to contribute to the harmomegathy of Gnetum pollen in our study. In this present study, the harmomegathic characters of Gnetum pollen in their dehydrated and hydrated states were found in both acetolyzed and non-acetolyzed grains. This indicated that the exine elasticity is still effective even in acetolyzed pollen from old herbarium material. Moreover, besides the crumpled appearance exhibited by some dry pollen, the morphology of acetolyzed and non-acetolyzed pollen grains in the same harmomegathic state was generally similar. This is not in agreement with earlier reports that acetolyzed pollen grains of other plant species normally do not show the same shape as those of the hydrated pollen and that the typical infolding present in fresh samples of some pollen species was not found in those from old herbarium specimens (Halbritter & Hesse, 2004). The different appearances between acetolyzed and hydrated pollen from Acanthaceae and Araceae were shown by Scotland et al. (1990) (SEM) and Weber, Halbritter & Hesse (1999) (LM). The harmomegathic mechanism is found in a wide range of fossil gymnosperm pollen from the Upper Palaeozoic and Early Mesozoic eras. The corpus shows strips of exine (taeniae) separated by grooves (striae) allowing the pollen to fold onto itself when drying out during pollination. This feature can still be found in the extant dicot Nilgirianthus warrensis (Dalzell) Bremek. (Acanthaceae). In angiosperms, such foldings of pollen exine appear in the membranous colpus area, allowing the pollen to become sealed (Traverse, 2007). Since there is no colpus in the ulcerate (mono-porate) and inaperturate grains, they exhibit different approaches to regulating pollen volume. During dehydration, their pollen walls fold into a mirror-buckled shape (Katifori et al., 2010). This study and previous studies, observed a cup-like feature in dry Gnetum pollen (Yao et al., 2004, figs 11, 12; Tekleva, 2016, fig. 1d–f). Therefore, in addition to the raised exine area and the ridge, this mirror-buckling character would be another indicator of dehydration in this pollen group. Halbritter & Hesse (2004) proposed that infolding patterns of pollen were controlled by many factors, such as the aperture condition, the thickness of the pollen wall, thickening and thinning of wall strata, pollen size, ornamentation and pollenkitt. Thus, it would be difficult to envision the shape of the dehydrated pollen grain. For example, even though pollen from Echinops L. (Asteraceae) and Gnetum show similar exine sculpture (echinate), their harmomegathic characters are different since the folding of Echinops pollen occurred in the apertural areas (Garnatje & Martin, 2007, fig. 1A and B). When water passes through the exine stratum in plant spores and pollen, it does so without the development of any structural change in the exine and it is then absorbed by the intine, stimulating the first state of increased spore and pollen volume. In the second state, the ability of sporomorphs to increase their volume to accommodate water uptake by the intine and cytoplasm is determined by exine elasticity (Pacini, 1990). According to ultrastructural studies, many structures prevent pollen destruction during hydration, e.g. the columellate infratectal layer in Boraginaceae, the spongy endexine in Adansonia L. or the fragmentation of the foot–layer in many plant groups (Hesse, 2000; Volkova et al., 2013; Rasoamanana et al., 2015). It is possible that the exine flexibility of Gnetum pollen found in this present study is facilitated not only by the concentration of infratectal granules between the lamellate endexine and the central part of the raised exine region shown in Hesse (1980), Yao et al. (2004) and Tekleva (2016), but also by the loose connection between the individual exine layers. In zona-aperturate pollen grains of Zamioculcas Schott and Gonatopus Engl. the lamellate endexine accommodates morphological plasticity in response to harmomegathic stress (Hesse et al., 2001). Hesse (2000) mentioned that not only the space within the infratectal stratum, but also the separation of exine units plays a part in pollen volume regulation. However, the ultrastructure of Gnetum pollen walls in different states of hydration and dehydration has not yet been examined. Therefore, further studies using TEM are necessary for a better understanding of Gnetum palynology and the harmomegathic mechanisms of pollen of the species. CONCLUSION The pollen morphology of all the Gnetum spp. is generally similar. Microechini on a psilate surface appears to be the character most representative of the hydrated state, whereas the dehydrated state is characterized by raised exine regions and ridge-like folds. Since harmomegathy can cause morphological plasticity, some overlapping of qualitative and quantitative morphological characters occurs between different species and among samples of the same species prepared by different techniques. Therefore, it was not possible to establish any systematic significance in these eight studied pollen species. However, the suggestion that Asian Gnetum taxa are monophyletic, which is based on the molecular phylogenetic data, is supported by the existence of the continuous tectum observed in all the Gnetum pollen in the present study. Moreover, the authors conclude that comparative palynological study of the variation in the pollen characters of this plant group must take into account the influence of preparation techniques as well as the numbers and conditions of the plants studied. ACKNOWLEDGEMENTS The authors are grateful to the Herbarium of Department Biology, Faculty of Science, Prince of Songkla University, the Herbarium of Department of Biology, Faculty of Science, Chang-Mai University and the Queen Sirikit Botanic Garden Herbarium, Botanical Garden Organization, Thailand for providing Gnetum specimens, making this project possible. Mr Thomas Duncan Coyne is also thanked for improvements to the English. This project was funded by the Plant Genetic Conservation Project under the Royal Initiative of Her Royal Highness Princess Maha Chakri Sirindhorn, project number SCI590816S. APPENDIX Examined herbarium specimens: 1. G. cuspidatum Blume, Sirirugsa, P., Ban Tung Lung, Songkhla, Thailand, Specimen number 184064, PSU herbarium. 2. G. latifolium Blume var. funiculare (Blume) Markgr., Maxwell JF., Khao Luang National Park, Nakhon Si Thammarat, Specimen number 184038, PSU herbarium; Sirirugsa, P., Doi Phu Kha National Park, Pua, Nan, Specimen number 17664, QBG herbarium. 3. G. leptostachyum Blume, Maxwell JF., Chiang Mai, Thailand, Specimen number 01953, CMU herbarium. 4. G. gnemon L. var. tenerum Markgr., Muangyen N., Chiang Mai, Specimen number 71989, gifted from QBG herbarium; Newman M., Khao Luang National Park, Nakhon Si Thammarat, Specimen number 184034, PSU herbarium; Romkloa Botanical Garden, Specimen number 62505, QBG herbarium. 5. G. macrostachyum Hook.f., Maxwell JF., Nakhonnayok, Thailand, Specimen number 20117, CMU herbarium. 6. G. montanum Markgr., Maknoi, C., Si satchanalai, Sukhothai, Specimen number 82654, gifted from QBG herbarium, PSU herbarium; La-ongsri W, Tatiya P, Satatha S., Doi Pha Chang, Nan, Specimen number 0015257, gifted from QBG herbarium, PSU herbarium. 7. G. tenuifolium Ridl., Sirirugsa, P., Kho Hong hill, Songkhla, Specimen number 184048; 184049, PSU herbarium. REFERENCES Biye EH , Balkwill K , Cron GV . 2014 . A clarification of Gnetum L. (Gnetaceae) in Africa and the description of two new species . Plant Systematics and Evolution 300 : 263 – 272 . Google Scholar CrossRef Search ADS Blackmore S , Barnes SH . 1986 . Harmomegathic mechanisms in pollen grains . In: Blackmore S , Ferguson IK , eds. Pollen and spores: form and function . London : Academic Press , 137 – 149 . Bolinder K , Niklas KJ , Rydin C . 2015 . Aerodynamics and pollen ultrastructure in Ephedra . American Journal of Botany 102 : 457 – 470 . Google Scholar CrossRef Search ADS PubMed Carlquist S . 1996 . Wood, bark, and stem anatomy of Gnetales: a summary . International Journal of Plant Sciences 157 : 58 – 76 . Google Scholar CrossRef Search ADS Crane PR . 1985 . Phylogenetic analysis of seed plants and the origin of angiosperms . Annals of the Missouri Botanical Garden 72 : 716 – 793 . Google Scholar CrossRef Search ADS Doores AS , Osborn JM , El-Ghazaly G . 2007 . Pollen ontogeny in Ephedra americana (Gnetales) . International Journal of Plant Sciences 168 : 985 – 997 . Google Scholar CrossRef Search ADS Erdtman G . 1954 . Pollen morphology and plant taxonomy . Botaniska Notiser 2 : 65 – 81 . Erdtman G . 1965 . Pollen and spore morphology/Plant taxonomy. Gymnospermae, Pteridophyta, Bryophyta . Stockholm : Almquist & Wiksell /New York: Ronald Press. Erdtman G . 1969 . Handbook of palynology, an introduction to the study of pollen grains and spores . New York : Hafner Publishing Co . Firon N , Nepi M , Pacini E . 2012 . Water status and associated processes mark critical stages in pollen development and functioning . Annals of Botany 109 : 1201 – 1214 . Google Scholar CrossRef Search ADS PubMed Friis EM , Crane PR , Pedersen KR . 2011 . Early flowers and angiosperm evolution . New York : Cambridge University Press . Google Scholar CrossRef Search ADS Garnatje T , Martin J . 2007 . Pollen studies in the genus Echinops L. and Xeranthemum group (Asteraceae) . Botanical Journal of the Linnean Society 154 : 549 – 557 . Google Scholar CrossRef Search ADS Gillespie LJ , Nowicke JW . 1994 . Systematic implications of pollen morphology in Gnetum . Acta Botanica Gallica 141 : 131 – 139 . Google Scholar CrossRef Search ADS Gong YB , Yang M , Vamosi JC , Yang HM , Mu WX , Liu JK , Wan T . 2015 . Wind or insect pollination? Ambophily in a subtropical gymnosperm Gnetum parvifolium (Gnetales) . Plant Species Biology 31 : 272 – 279 . Google Scholar CrossRef Search ADS Griffith W . 1859 . Remarks on Gnetum . Transactions of the Linnean Society of London 26 : 299 – 312 . Google Scholar CrossRef Search ADS Gullvåg BM . 1966 . The fine structure of some gymnosperm pollen walls . Grana Palynologica 6 : 435 – 475 . Google Scholar CrossRef Search ADS Guo SX , Sha JG , Bian LZ , Qiu YL . 2009 . Male spike strobiles with Gnetum affinity from the early Cretaceous in western Liaoning, Northeast China . Journal of Systematics and Evolution 47 : 93 – 102 . Google Scholar CrossRef Search ADS Halbritter H . 1998 . Preparing living pollen material for scanning electron microscopy using 2,2–dimethoxypropane (DMP) and critical point drying . Biotechnic and Histochemistry 73 : 137 – 143 Google Scholar CrossRef Search ADS PubMed Halbritter H . 2015 . Gnetum gnemon . In: PalDat—a palynological database. Published on the Internet. Available at: https://www.paldat.org/pub/Gnetum_gnemon/300040 (accessed 3 May 2017). Halbritter H . 2016 . Welwitschia mirabilis . In: PalDat—a palynological database. Published on the Internet. Available at: https://www.paldat.org/pub/Welwitschia_mirabilis/302371 (accessed 3 May 2017). Halbritter H , Hesse M . 2004 . Principal modes of infoldings in tricolp(or)ate angiosperm pollen . Grana 43 : 1 – 14 . Google Scholar CrossRef Search ADS Halbritter H , Weber M , Hesse M . 2010 . Unique aperture stratification in Carex (Cyperaceae) pollen . Grana 49 : 1 – 11 . Google Scholar CrossRef Search ADS Hesse M . 1980 . Pollenkitt is lacking in Gnetum gnemon (Gnetaceae) . Plant Systematics and Evolution 136 : 41 – 46 . Google Scholar CrossRef Search ADS Hesse M . 1984 . Pollenkitt is lacking in Gnetatae: Ephedra and Welwitschia: further proof for its restriction to the angiosperms . Plant Systematics and Evolution 144 : 9 – 16 . Google Scholar CrossRef Search ADS Hesse M . 2000 . Pollen wall stratification and pollination . Plant Systematics and Evolution 222 : 1 – 17 . Google Scholar CrossRef Search ADS Hesse M , Bogner J , Halbritter H , Weber M . 2001 . Palynology of the perigoniate Aroideae: Zamioculcas, Gonatopus and Stylochaeton (Araceae) . Grana 40 : 26 – 34 . Google Scholar CrossRef Search ADS Hesse M , Halbritter H , Zetter R , Weber M , Buchner R , Frosch-Radivo A , Ulrich S . 2009 . Pollen terminology. An illustrated handbook . Vienna, New York : Springer . Hou C , Humphreys AM , Thureborn O , Rydin C . 2015 . New insights into the evolutionary history of Gnetum (Gnetales) . Taxon 64 : 239 – 253 . Google Scholar CrossRef Search ADS Hou C , Wikström N , Strijk JS , Rydin C . 2016 . Resolving phylogenetic relationships and species delimitations in closely related gymnosperms using high-throughput NGS, Sanger sequencing and morphology . Plant Systematics and Evolution 302 : 1345 – 1365 . Google Scholar CrossRef Search ADS Ickert-Bond SM , Renner SS . 2016 . The Gnetales: recent insights on their morphology, reproductive biology, chromosome numbers, biogeography, and divergence times . Journal of Systematics and Evolution 54 : 1 – 16 . Google Scholar CrossRef Search ADS Ickert-Bond SM , Skvarla JJ , Chissoe WF . 2003 . Pollen dimorphism in Ephedra L. (Ephedraceae) . Review of Palaeobotany and Palynology 124 : 325 – 334 . Google Scholar CrossRef Search ADS Jörgensen A , Rydin C . 2015 . Reproductive morphology in the Gnetum cuspidatum group (Gnetales) and its implications for pollination biology in the Gnetales . Plant Ecology and Evolution 148 : 387 – 396 . Google Scholar CrossRef Search ADS Kato M . 1996 . Plant–pollinator interactions in the understory of a lowland mixed dipterocarp forest in Sarawak . American Journal of Botany 83 : 732 – 743 . Google Scholar CrossRef Search ADS Kato M , Inoue T , Nagamitsu T . 1995 . Pollination biology of Gnetum (Gnetaceae) in a lowland mixed dipterocarp forest in Sarawak . American Journal of Botany 82 : 862 – 868 . Google Scholar CrossRef Search ADS Kato M , Kosaka Y , Kawakita A , Okuyama Y , Kobayashi C , Phimminith T , Thongphan D . 2008 . Plant–pollinator interactions in tropical monsoon forests in Southeast Asia . American Journal of Botany 95 : 1375 – 1394 . Google Scholar CrossRef Search ADS PubMed Katifori E , Alben S , Cerda E , Nelson DR , Dumais J . 2010 . Foldable structures and the natural design of pollen grains . Proceedings of the National Academy of Sciences USA 10 : 7635 – 7639 . Google Scholar CrossRef Search ADS Kubitzki K . 1990 . Gnetatae . In: Kramer KU , Green PS , eds. The families and genera of vascular plants, vol. 1, pteridophytes and gymnosperms . Berlin, Heidelberg : Springer Verlag , 378 – 391 . Kuprianova LA . 1983 . Struktura pyl’cy gnetuma‒skhodstvo ili rodstvo ? Priroda 12 : 70 – 71 . Markgraf F . 1929 . Monographie der Gattung Gnetum . Bulletin du Jardin botanique de Buitenzorg 10 : 407 – 511 . Mohl H . 1834 . Über den Bau und die Formen der Pollenkörner. Beiträge zur Anatomie und Physiologie der Gewächse . Erstes Heft . Bern : Christain Fischer und Comp . Momose K , Yumoto T , Nagamitsu T , Kato M , Nagamasu H , Sakai S , Harrison RD , Itioka T , Hamid AA , Inoue T . 1998 . Pollination biology in a lowland dipterocarp forest in Sarawak, Malaysia. I. Characteristics of the plant-pollinator community in a lowland dipterocarp forest . American Journal of Botany 85 : 1477 – 1501 . Google Scholar CrossRef Search ADS PubMed Orel LI , Kuprianova LA , Golubeva EA . 1986 . Ul’trastruktura acetolicirovannyh obolochek tapetal’nyh kletok i pyl’cevyh zjoren u Gnetum africanum (Gnetaceae) . Botanicheskii Zhurnal (Moscow & Leningrad) 71 : 750 – 754 . Osborn JM . 2000 . Pollen morphology and ultrastructure of gymnospermous anthophytes . In: Harley MM , Morton CM , Blackmore S , eds. Pollen and spores: morphology and biology . London : Royal Botanical Gardens, Kew , 163 – 185 . Osborn JM , Taylor TN , de Lima MR . 1993 . The ultrastructure of fossil ephedroid pollen with gnetalean affinities from the Lower Cretaceous of Brazil . Review of Palaeobotany and Palynology 77 : 171 – 184 . Google Scholar CrossRef Search ADS Pacini E . 1990 . Harmomegathic characters of Pteridophyta spores and Spermatophyta pollen . Plant Systematics and Evolution, Supplement 5 : 53 – 69 . Google Scholar CrossRef Search ADS Payne WW . 1972 . Observations of harmomegathy in pollen of Anthophyta . Grana 12 : 93 – 98 . Google Scholar CrossRef Search ADS Phengklai C . 1973 . Studies in flora of Thailand: Gnetaceae . Thai Forest Bulletin (Botany) 7 : 21 – 31 . Price RA . 1996 . Systematics of the Gnetales: a review of morphological and molecular evidence . International Journal of Plant Sciences (Supplement) 157 : S40 – S49 . Google Scholar CrossRef Search ADS Punt W , Hoen PP , Blackmore S , Nilsson S , Le Thomas A . 2007 . Glossary of pollen and spore terminology . Review of Palaeobotany and Palynology 143 : 1 – 81 . Google Scholar CrossRef Search ADS Rasoamanana EN , Razanamaro O , Ramavovololona P , Ramamonjisoa RZ , Verdeil JL , Danthu P , Suárez-Cervera M . 2015 . Pollen wall ultrastructure of the genus Adansonia L. species . Plant Systematics and Evolution 301 : 541 – 554 . Google Scholar CrossRef Search ADS Scotland RW , Barnes HB , Blackmore S . 1990 . Harmomegathy in the Acanthaceae . Grana 29 : 37 – 45 . Google Scholar CrossRef Search ADS Tekleva M . 2016 . Pollen morphology and ultrastructure of several Gnetum species: an electron microscopic study . Plant Systematics and Evolution 302 : 291 – 303 . Google Scholar CrossRef Search ADS Traverse A . 2007 . Paleopalynology . Dordrecht : Springer . Volkova OA , Severova EE , Polevova SV , 2013 . Structural basis of harmomegathy: evidence from Boraginaceae pollen . Plant Systematics and Evolution 299 : 1769 – 1779 . Google Scholar CrossRef Search ADS Weber M , Halbritter H , Hesse M . 1999 . The basic pollen wall types in Araceae . International Journal of Plant Sciences 160 : 415 – 423 . Google Scholar CrossRef Search ADS Wodehouse RP . 1935 . Pollen grains–their structure, identification and significance in science and medicine . New York : McGraw-Hill . Won H , Renner SS . 2005 . The internal transcribed spacer of nuclear ribosomal DNA in the gymnosperm Gnetum . Molecular Phylogenetics and Evolution 36 : 581 – 597 . Google Scholar CrossRef Search ADS PubMed Won H , Renner SS . 2006 . Dating dispersal and radiation in the gymnosperm Gnetum (Gnetales)—clock calibration when outgroup relationships are uncertain . Systematic Biology 55 : 610 – 622 . Google Scholar CrossRef Search ADS PubMed Yao Y-F , Xi Y-Z , Geng B-Y , Li C-S . 2004 . The exine ultrastructure of pollen grains in Gnetum (Gnetaceae) from China and its bearing on the relationship with the ANITA Group . Botanical Journal of the Linnean Society 146 : 415 – 425 . Google Scholar CrossRef Search ADS Zetter R . 1989 . Methodik und Bedeutung einer routinemäßig kombinierten lichtmikroskopischen und rasterelektonenmikroskopischen Untersuchung fossiler Mikrofloren . Courier Forschungsinstitut Senckenberg 109 : 41 – 50 . © 2018 The Linnean Society of London, Botanical Journal of the Linnean Society This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/open_access/funder_policies/chorus/standard_publication_model)

Feio, Ana Carla;Meira, Renata M S A;Riina, Ricarda

doi: 10.1093/botlinnean/boy038pmid: N/A

Abstract Establishing species limits in clades in Croton using characters from external morphology and common molecular markers has proved cumbersome, especially in Croton section Cyclostigma, a group of 50 Neotropical arborescent species commonly known as dragon’s blood. Given this, we explored leaf and shoot apex anatomical characters for their utility in distinguishing species or groups of species in this section. We analysed 90 specimens belonging to section Cyclostigma and 14 specimens from related groups (sections Adenophylli, Cupreati and Sampatik), recording and describing 45 qualitative leaf characters. These characters were assembled into a matrix and analysed using statistical clustering methods based on similarity. Our results show that trichomes are one of the most diverse and variable anatomical features among the studied specimens. Novel anatomical features for Croton include the presence of a hypodermis and two previously unreported types of non-glandular trichomes. Although we did not detect a single anatomical character uniting section Cyclostigma, combinations of anatomical characters were useful to establish species limits and taxonomic identities within this section. colleters, hypodermis, idioblasts, nectaries, non-articulated laticifers, non-secretory trichomes INTRODUCTION Croton L. is a megadiverse genus with >1200 species (Christenhusz, Fay & Chase, 2017; Haber et al., 2017). The medicinally important dragon’s blood trees belong exclusively to Croton section Cyclostigma Griseb. (Riina, Berry & van Ee, 2009), one of the main Neotropical clades of the genus. About 750 species in the genus occur in the New World, and these are organized into 31 sections in the most recent classification (van Ee, Riina & Berry, 2011). According to ongoing taxonomic studies, section Cyclostigma consists of c. 50 species distributed from Mexico to northern Argentina (Riina et al., pers. comm.). All species in this section produce abundant reddish to orange/yellowish latex, which is widely used in many Latin American countries, mainly as a treatment for skin wounds and burns (Meza, 1999; Borges & King, 2000; Jones, 2003; Salatino, Salatino & Negri, 2007). Dragon′s blood trees grow predominantly in moist forests, including riverine, lowland and montane forests, but a few species occur in dry forests (van Ee et al., 2011). Like the other 30 sections of New World Croton, section Cyclostigma is monophyletic. However, the resolution of the relationships between and within most sections is still a work in progress. For instance, a monophyletic section Cyclostigma was established by excluding several putative species, which were retrieved in seven other clades of Croton. The Cyclostigma clade is sister to the monotypic section Cupreati Riina in the phylogenetic analysis based on trnL-F or sister to section Adenophylli Griseb. (section Cascarilla in Riina et al., 2009) in the ITS tree. In addition, resolution at the species level in the Cyclostigma clade was low and many dragon’s blood species remained to be sampled (Riina et al., 2009). The only potential synapomorphy for section Cyclostigma is the presence of non-articulated laticifers in wood rays (Wiedenhoeft, Riina & Berry, 2009). On the other hand, all species in the section share a suite of morphological characteristics including arborescent habit, reddish latex, stellate indumentum, conspicuous stipules, acropetiolar glands, long thyrsoid inflorescences with bisexual cymules predominantly at the base and staminate flowers with >16 stamens (Riina et al., 2009). However, most of the individual characters from the suite described above are also present in other clades of Croton (van Ee et al., 2011). Although recent taxonomic studies of the dragon’s blood group have identified new and morphologically distinct species (Riina, Berry & Cornejo, 2007; Riina, Vigo & Cerón, 2014; Riina et al., 2015; Feio et al., in press), establishing species limits in section Cyclostigma has proved difficult due to the prevalent homoplasy in most of the traditional morphological characters and the lack of resolution at the species level with the molecular markers used so far (Riina et al., 2009; van Ee et al., 2011). Given this scenario, we conducted a comparative analysis of leaf anatomical features using a broad sampling of section Cyclostigma, a selection of species from its sister clades from nuclear and plastid phylogenetic analyses (sections Adenophylli and Cupreati) and a representative of a more distant clade [section Sampatik (G.L.Webster) Riina]. Our goals were to identify additional anatomical features that could help to tease apart species or groups in section Cyclostigma and to expand the database of characters available for studies of morphological evolution in Euphorbiaceae. MATERIAL AND METHODS Taxon sampling We sampled 104 specimens, of which 90 belong to Croton section Cyclostigma or have morphological affinities with that section according to ongoing taxonomic studies (R. Riina, pers. comm.), 11 specimens representing six species were from Croton section Adenophylli, one specimen from the only known species (C. cupreatus Croizat) of section Cupreati and two specimens of C. piptocalyx Müll.Arg. from section Sampatik. Specimens were identified to section based on their morphological affinities according to van Ee et al. (2011). Plant material for anatomical studies was obtained from herbarium specimens ( Appendix 1). Herbarium acronyms follow those of Thiers (2016, continuously updated). Light microscopy Herbarium samples of fully expanded leaves and shoot apex were rehydrated (Smith & Smith, 1942) and stored in 70% ethanol. Subsequently, fragments from the middle and basal portions of the leaf blade (midrib and margin) and the middle and distal portions of the petiole were dehydrated in decreasing ethanol series and were embedded in 2-hydroxyethyl methacrylate resin for sectioning (Historesin Leica®, solutions were prepared according to manufacturer’s instructions) following Meira & Martins (2003). All samples were transversally and longitudinally sectioned with a thickness of 3–7 µm, using an automatic rotary microtome (model RM 2265, Leica® Biosystems, Nussloch GmbH, Heidelberg, Germany) using disposable glass knives (Leica®, Biosystems, Nussloch GmbH). Sections were stained with toluidine blue at pH 4.6 (O’Brien, Feder & McCully, 1965) and slides were mounted in resin (Permount®, Fisher Scientific, NJ, USA). Some of the samples were also cleared with 5% sodium hydroxide and 20% hypochlorite solutions, stained with 50% ethanol-diluted fuchsin (Shobe & Lersten, 1967) and mounted in glycerinated gelatin (Kaiser, 1880). Observations and photographic documentation were performed with a light microscope (Model AX70TRF, Olympus Optical, Tokyo, Japan) equipped with a U-Photo system, polarizing filter and digital camera (AxioCam HRc; ©Carl Zeiss, Jena, Germany). Macro images were obtained using a stereomicroscope (Stemi 2000-C®, ©Carl Zeiss Microscopy GmbH, Jena, Germany) with a coupled digital camera (AxioCam ERc5s®, ©Carl Zeiss Microscopy GmbH, Jena, Germany). Scanning electron microscopy Some of samples stored in 70% ethanol were dehydrated in ethanol and critical point dried with CO2 (Bozzola & Russel, 1992) in a 020 CPD dryer (Bal-Tec; Balzers, Liechtenstein). They were mounted on aluminium stubs and coated with gold using a FDU 010 sputter coater (Bal-Tec). Examination of specimens and image capture were conducted using a scanning electron microscope Leo 1430VP (Zeiss, Cambridge, UK) at the Centro de Microscopia e Microanálises, Universidade Federal de Viçosa. Description and analysis of anatomical characters Anatomical descriptions were based on terms from Metcalfe & Chalk (1979, 1983) and Evert (2006). Trichome terminology followed the classification of Webster, Del-Arco Aguilar & Smith (1996). Forty-five qualitative characters from leaf and shoot apex were recorded (Table 1). A matrix of specimens by characters was built, in which characters were coded as binary states ( Appendix 2). A distance matrix was calculated using the Dice-Sorensen and Jaccard coefficients and similarity dendrograms were generated using the software PAST3 vs. 3.06 (Hammer, Harper & Ryan, 2001). The resulting dendrograms were further edited in FigTree vs.1.4 (Rambaut, 2012) and CorelDRAW X3®. Table 1. Qualitative anatomical characters used in cluster analysis coded as present (1) or absent (0) in the matrix Number Character and states 1 Secretory idioblasts 2 Hypodermis 3 Continuous palisade parenchyma in the midrib 4 Bundle-sheath extension 5 Margin with continuous palisade parenchyma 6 Colleters on the leaf margin 7 Basilaminar/acropetiolar colleters 8 EFNs basilaminar 9 EFNs acropetiolar 10 EFNs dispersed on the blade 11 EFNs marginal 12 EFNs sessile 13 EFNs stipitate 14 EFNs with concave surface 15 EFNs with convex surface 16 EFNs with flat surface 17 Lepidote trichomes 18 Rosulate trichomes 19 Dendritic trichomes 20 Fasciculate cushion-shaped trichomes 21 Stipitate-stellate trichomes 22 Stipitate-stellate porrect trichomes 23 Appressed-stellate porrect trichomes 24 Appressed-stellate trichomes 25 Multiradiate trichomes 26 Fasciculate trichomes 27 Simple trichomes 28 Hypostomatic leaf 29 Amphistomatic leaf 30 Paracytic stomata 31 Secretory idioblasts only in the abaxial epidermis 32 Secretory idioblasts only in the adaxial epidermis 33 Secretory idioblasts in both abaxial epidermis and palisade parenchyma 34 Secretory idioblasts in the palisade parenchyma 35 Secretory idioblasts dispersed in the mesophyll 36 Non-articulated branched laticifers 37 Articulated laticifers 38 Unistratified epidermis 39 Dorsiventral mesophyll 40 Biconvex midrib 41 Midrib uniconvex, with flat to slightly concave adaxial face 42 Midrib with colateral bundles arranged in an open arch-like with one to five dorsal bundles 43 Midrib with colateral bundles arranged in an open arch-like without dorsal bundles 44 Revolute margin 45 Involute margin Number Character and states 1 Secretory idioblasts 2 Hypodermis 3 Continuous palisade parenchyma in the midrib 4 Bundle-sheath extension 5 Margin with continuous palisade parenchyma 6 Colleters on the leaf margin 7 Basilaminar/acropetiolar colleters 8 EFNs basilaminar 9 EFNs acropetiolar 10 EFNs dispersed on the blade 11 EFNs marginal 12 EFNs sessile 13 EFNs stipitate 14 EFNs with concave surface 15 EFNs with convex surface 16 EFNs with flat surface 17 Lepidote trichomes 18 Rosulate trichomes 19 Dendritic trichomes 20 Fasciculate cushion-shaped trichomes 21 Stipitate-stellate trichomes 22 Stipitate-stellate porrect trichomes 23 Appressed-stellate porrect trichomes 24 Appressed-stellate trichomes 25 Multiradiate trichomes 26 Fasciculate trichomes 27 Simple trichomes 28 Hypostomatic leaf 29 Amphistomatic leaf 30 Paracytic stomata 31 Secretory idioblasts only in the abaxial epidermis 32 Secretory idioblasts only in the adaxial epidermis 33 Secretory idioblasts in both abaxial epidermis and palisade parenchyma 34 Secretory idioblasts in the palisade parenchyma 35 Secretory idioblasts dispersed in the mesophyll 36 Non-articulated branched laticifers 37 Articulated laticifers 38 Unistratified epidermis 39 Dorsiventral mesophyll 40 Biconvex midrib 41 Midrib uniconvex, with flat to slightly concave adaxial face 42 Midrib with colateral bundles arranged in an open arch-like with one to five dorsal bundles 43 Midrib with colateral bundles arranged in an open arch-like without dorsal bundles 44 Revolute margin 45 Involute margin View Large Table 1. Qualitative anatomical characters used in cluster analysis coded as present (1) or absent (0) in the matrix Number Character and states 1 Secretory idioblasts 2 Hypodermis 3 Continuous palisade parenchyma in the midrib 4 Bundle-sheath extension 5 Margin with continuous palisade parenchyma 6 Colleters on the leaf margin 7 Basilaminar/acropetiolar colleters 8 EFNs basilaminar 9 EFNs acropetiolar 10 EFNs dispersed on the blade 11 EFNs marginal 12 EFNs sessile 13 EFNs stipitate 14 EFNs with concave surface 15 EFNs with convex surface 16 EFNs with flat surface 17 Lepidote trichomes 18 Rosulate trichomes 19 Dendritic trichomes 20 Fasciculate cushion-shaped trichomes 21 Stipitate-stellate trichomes 22 Stipitate-stellate porrect trichomes 23 Appressed-stellate porrect trichomes 24 Appressed-stellate trichomes 25 Multiradiate trichomes 26 Fasciculate trichomes 27 Simple trichomes 28 Hypostomatic leaf 29 Amphistomatic leaf 30 Paracytic stomata 31 Secretory idioblasts only in the abaxial epidermis 32 Secretory idioblasts only in the adaxial epidermis 33 Secretory idioblasts in both abaxial epidermis and palisade parenchyma 34 Secretory idioblasts in the palisade parenchyma 35 Secretory idioblasts dispersed in the mesophyll 36 Non-articulated branched laticifers 37 Articulated laticifers 38 Unistratified epidermis 39 Dorsiventral mesophyll 40 Biconvex midrib 41 Midrib uniconvex, with flat to slightly concave adaxial face 42 Midrib with colateral bundles arranged in an open arch-like with one to five dorsal bundles 43 Midrib with colateral bundles arranged in an open arch-like without dorsal bundles 44 Revolute margin 45 Involute margin Number Character and states 1 Secretory idioblasts 2 Hypodermis 3 Continuous palisade parenchyma in the midrib 4 Bundle-sheath extension 5 Margin with continuous palisade parenchyma 6 Colleters on the leaf margin 7 Basilaminar/acropetiolar colleters 8 EFNs basilaminar 9 EFNs acropetiolar 10 EFNs dispersed on the blade 11 EFNs marginal 12 EFNs sessile 13 EFNs stipitate 14 EFNs with concave surface 15 EFNs with convex surface 16 EFNs with flat surface 17 Lepidote trichomes 18 Rosulate trichomes 19 Dendritic trichomes 20 Fasciculate cushion-shaped trichomes 21 Stipitate-stellate trichomes 22 Stipitate-stellate porrect trichomes 23 Appressed-stellate porrect trichomes 24 Appressed-stellate trichomes 25 Multiradiate trichomes 26 Fasciculate trichomes 27 Simple trichomes 28 Hypostomatic leaf 29 Amphistomatic leaf 30 Paracytic stomata 31 Secretory idioblasts only in the abaxial epidermis 32 Secretory idioblasts only in the adaxial epidermis 33 Secretory idioblasts in both abaxial epidermis and palisade parenchyma 34 Secretory idioblasts in the palisade parenchyma 35 Secretory idioblasts dispersed in the mesophyll 36 Non-articulated branched laticifers 37 Articulated laticifers 38 Unistratified epidermis 39 Dorsiventral mesophyll 40 Biconvex midrib 41 Midrib uniconvex, with flat to slightly concave adaxial face 42 Midrib with colateral bundles arranged in an open arch-like with one to five dorsal bundles 43 Midrib with colateral bundles arranged in an open arch-like without dorsal bundles 44 Revolute margin 45 Involute margin View Large The taxonomic identity of the specimens was confirmed or modified by integrating information from the resulting clustering pattern, morphology from taxonomic descriptions, field observations, type specimens, habitat and geographical distribution. RESULTS Leaf anatomy The matrix of anatomical characters has 1.3% of missing data coded as ‘?’ (Table 2). There were 17 specimens (Table 2), for which we could not evaluate some characters (1–11) because the original herbarium material was too old or poorly preserved. In section Cyclostigma the most critical specimens regarding missing data were C. bogotanus Cuatrec. (Riina-1591) and C. mutisianus Kunth (Barkley-3768), both with the maximum percentage of missing characters (21%) (Table 2). Table 2. Specimens with missing data in the leaf character matrix and their respective numbers and percentage. Total number of characters coded is 45, and total number of cells in the matrix is 5408 (45 characters × 104 specimens) Specimens (Voucher) # missing characters % missing characters C. aequatoris (Riina-1434) 6 12 C. bogotanus (Riina-1591) 11 21 C. bonplandianus (Riina-1517) 10 19 C. conduplicatus (Riina-1296) 6 12 C. hibiscifolius (Breteler-3446) 4 8 C. hibiscifolius (Contreras-042) 2 4 C. mutisianus-2 (Barkley-3768) 11 21 C. piptotocalyx (Bortoluzzi-379) 2 4 C. pseudopopulus (Mota-2276) 1 2 C. pseudopopulus (Mota-2284) 1 2 C. pseudopopulus (Mota-2291) 1 2 C. rimbachii (Riina-1402) 3 6 C. rusbyi (Riina-1479) 1 2 C. rusbyi (Riina-1481) 1 2 C. urucurana-1 (Leitão-Filho-1603) 3 6 C. urucurana-1 (Pollito-VA 001) 5 10 C. vulnerarius-1 (Forero-8148) 2 4 Total (entire matrix) 70 1.3 Specimens (Voucher) # missing characters % missing characters C. aequatoris (Riina-1434) 6 12 C. bogotanus (Riina-1591) 11 21 C. bonplandianus (Riina-1517) 10 19 C. conduplicatus (Riina-1296) 6 12 C. hibiscifolius (Breteler-3446) 4 8 C. hibiscifolius (Contreras-042) 2 4 C. mutisianus-2 (Barkley-3768) 11 21 C. piptotocalyx (Bortoluzzi-379) 2 4 C. pseudopopulus (Mota-2276) 1 2 C. pseudopopulus (Mota-2284) 1 2 C. pseudopopulus (Mota-2291) 1 2 C. rimbachii (Riina-1402) 3 6 C. rusbyi (Riina-1479) 1 2 C. rusbyi (Riina-1481) 1 2 C. urucurana-1 (Leitão-Filho-1603) 3 6 C. urucurana-1 (Pollito-VA 001) 5 10 C. vulnerarius-1 (Forero-8148) 2 4 Total (entire matrix) 70 1.3 View Large Table 2. Specimens with missing data in the leaf character matrix and their respective numbers and percentage. Total number of characters coded is 45, and total number of cells in the matrix is 5408 (45 characters × 104 specimens) Specimens (Voucher) # missing characters % missing characters C. aequatoris (Riina-1434) 6 12 C. bogotanus (Riina-1591) 11 21 C. bonplandianus (Riina-1517) 10 19 C. conduplicatus (Riina-1296) 6 12 C. hibiscifolius (Breteler-3446) 4 8 C. hibiscifolius (Contreras-042) 2 4 C. mutisianus-2 (Barkley-3768) 11 21 C. piptotocalyx (Bortoluzzi-379) 2 4 C. pseudopopulus (Mota-2276) 1 2 C. pseudopopulus (Mota-2284) 1 2 C. pseudopopulus (Mota-2291) 1 2 C. rimbachii (Riina-1402) 3 6 C. rusbyi (Riina-1479) 1 2 C. rusbyi (Riina-1481) 1 2 C. urucurana-1 (Leitão-Filho-1603) 3 6 C. urucurana-1 (Pollito-VA 001) 5 10 C. vulnerarius-1 (Forero-8148) 2 4 Total (entire matrix) 70 1.3 Specimens (Voucher) # missing characters % missing characters C. aequatoris (Riina-1434) 6 12 C. bogotanus (Riina-1591) 11 21 C. bonplandianus (Riina-1517) 10 19 C. conduplicatus (Riina-1296) 6 12 C. hibiscifolius (Breteler-3446) 4 8 C. hibiscifolius (Contreras-042) 2 4 C. mutisianus-2 (Barkley-3768) 11 21 C. piptotocalyx (Bortoluzzi-379) 2 4 C. pseudopopulus (Mota-2276) 1 2 C. pseudopopulus (Mota-2284) 1 2 C. pseudopopulus (Mota-2291) 1 2 C. rimbachii (Riina-1402) 3 6 C. rusbyi (Riina-1479) 1 2 C. rusbyi (Riina-1481) 1 2 C. urucurana-1 (Leitão-Filho-1603) 3 6 C. urucurana-1 (Pollito-VA 001) 5 10 C. vulnerarius-1 (Forero-8148) 2 4 Total (entire matrix) 70 1.3 View Large Of all the surveyed characters, trichomes were the most diverse and variable and may differ even between closely related species (Fig. 1A–O), but no type was exclusive to section Cyclostigma or any of the other outgroup species. Non-glandular trichomes were lepidote (Fig. 1A, B), dendritic (Fig. 1C, D), rosulate (Fig. 1E, F), fasciculate cushion-shaped (Fig. 1G, H), stipitate-stellate (Fig. 1I), stipitate-stellate porrect (Fig. 1J), appressed-stellate porrect (Fig. 1K), appressed-stellate (Fig. 1L), multiradiate (Fig. 1M, N) or simple (Fig. 1O). Out of these tn types, only the rosulate, stipitate-stellate porrect, appressed-stellate porrect and simple types are more typical of section Cyclostigma. Figure 1. View largeDownload slide Diversity of non-glandular trichomes in Croton. A, B. Lepidote trichomes; A. C. cupreatus; B. C. urucurana-1; C, D. Dendritic trichomes in C. vulnerarius; E, F. Rosulate trichomes in C. floccosus; G, H. Fasciculate cushion-shaped in C. speciosus; I. Stipitate-stellate trichomes in C. alchorneicarpus; J. Stipitate-stellate porrect trichomes in C. bonplandianus; K. Appressed-stellate porrect trichomes in C. redolens; K. L. Appressed-stellate trichomes in C. piptocalyx; M, N. Multiradiate trichomes in C. hibiscifolius; O. Simple trichomes in C. coriaceus. Bars: 60 µm (L); 100 µm (E, G, I, M, O); 150 µm (F, H); 200 µm (J, K); 300 µm (A–C, N); 500 µm (D). Figure 1. View largeDownload slide Diversity of non-glandular trichomes in Croton. A, B. Lepidote trichomes; A. C. cupreatus; B. C. urucurana-1; C, D. Dendritic trichomes in C. vulnerarius; E, F. Rosulate trichomes in C. floccosus; G, H. Fasciculate cushion-shaped in C. speciosus; I. Stipitate-stellate trichomes in C. alchorneicarpus; J. Stipitate-stellate porrect trichomes in C. bonplandianus; K. Appressed-stellate porrect trichomes in C. redolens; K. L. Appressed-stellate trichomes in C. piptocalyx; M, N. Multiradiate trichomes in C. hibiscifolius; O. Simple trichomes in C. coriaceus. Bars: 60 µm (L); 100 µm (E, G, I, M, O); 150 µm (F, H); 200 µm (J, K); 300 µm (A–C, N); 500 µm (D). Lepidote trichomes [present in two specimens (C. urucurana Baill. -1) of section Cyclostigma and in sections Adenophylli (C. ruizianus Müll.Arg.) and Cupreati (C. cupreatus)], dendritic trichomes [present in C. vulnerarius Baill. -2, C. draco Schltdl. -1, and C. celtidifolius Baill. -1] and fasciculate cushion-shaped trichomes [present in C. huberi Steyerm., C. speciosus Müll.Arg., C. medusae Müll.Arg. -1, C. medusae-4 and C. coriaceus Kunth -3] were less common. All specimens present hypostomatic leaves with parasitic stomata (Fig. 2A–C), epidermis with sinuous anticlinal walls on the abaxial side (Fig. 2A, B), straight to slightly sinuous walls on the adaxial side (Fig. 2C) and a slightly striated cuticle covering the epidermis on both sides (Fig. 2B). In cross section, they present a single-layered epidermis, thin cuticle, and stomata placed at the same level of the ordinary epidermal cells (Fig. 2D, E). In some specimens of section Cyclostigma (C. amentiformis Riina, C. bogotanus, C. coriaceus-2, C. coriaceus-3, C. floccosus B.A.Sm., C. rimbachii Croizat) and in section Cupreati, we observed a single-layered hypodermis underlying the adaxial epidermis (Fig. 2D). Figure 2. View largeDownload slide General anatomical features. A–C. Frontal view; D–P. Cross-sections; A. Abaxial epidermis with sinuous anticlinal walls in C. celtidifolius; B. Detail of abaxial epidermis, showing paracitic stomata and cuticle with attenuated stretch in C. macrobothrys; C. Adaxial epidermis with straight to slightly sinuous walls in C. macrobothrys; D. Dorsiventral mesophyll with unistratified epidermis in C. amentiformis, note: hypodermis (arrowhead), bundle-shealth extension (arrow); E. Dorsiventral mesophyll with collateral bundles (arrow) in C. ruizianus; F. Druse crystals under polarized light in the longitudinal section of petiole in C. alchorneicarpus; G, H, I. Leaf margin, G. Revolute in C. conduplicatus, note inset in C. amentiformis; H. Involute in C. bogotanus, Arrowhead: continuous palisade parenchyma; I. Discontinuous palisade parenchyma in C. lechleri; J. Biconvex midrib in C. lechleri; K. Midrib with flat to slightly concave adaxial side in C. conduplicatus; L. Vascular system of midrib with dorsal bundles in C. vulnerarius; M. Vascular system of midrib without dorsal bundles in C. aequatoris; N, O, P. Petiole of C. charaguesis; N. General view of petiole with round shape, horseshoe-shaped vascular system and two accessory bundles (arrows); O. Detail of petiole showing epidermis, cortical region and part of vascular system; P. Detail of adaxial side of petiole with two accessory bundles. Pp: palisade parenchyma, Sp: spongy parenchyma, Bars: 100 µm (B); 150 µm (D, E, F, I); 200 µm (A, C); 250 µm (M); 300 µm (G, H, K, L, O); 600 µm (insets Fig. G and H, P); 700 µm (J); 800 µm (N). Figure 2. View largeDownload slide General anatomical features. A–C. Frontal view; D–P. Cross-sections; A. Abaxial epidermis with sinuous anticlinal walls in C. celtidifolius; B. Detail of abaxial epidermis, showing paracitic stomata and cuticle with attenuated stretch in C. macrobothrys; C. Adaxial epidermis with straight to slightly sinuous walls in C. macrobothrys; D. Dorsiventral mesophyll with unistratified epidermis in C. amentiformis, note: hypodermis (arrowhead), bundle-shealth extension (arrow); E. Dorsiventral mesophyll with collateral bundles (arrow) in C. ruizianus; F. Druse crystals under polarized light in the longitudinal section of petiole in C. alchorneicarpus; G, H, I. Leaf margin, G. Revolute in C. conduplicatus, note inset in C. amentiformis; H. Involute in C. bogotanus, Arrowhead: continuous palisade parenchyma; I. Discontinuous palisade parenchyma in C. lechleri; J. Biconvex midrib in C. lechleri; K. Midrib with flat to slightly concave adaxial side in C. conduplicatus; L. Vascular system of midrib with dorsal bundles in C. vulnerarius; M. Vascular system of midrib without dorsal bundles in C. aequatoris; N, O, P. Petiole of C. charaguesis; N. General view of petiole with round shape, horseshoe-shaped vascular system and two accessory bundles (arrows); O. Detail of petiole showing epidermis, cortical region and part of vascular system; P. Detail of adaxial side of petiole with two accessory bundles. Pp: palisade parenchyma, Sp: spongy parenchyma, Bars: 100 µm (B); 150 µm (D, E, F, I); 200 µm (A, C); 250 µm (M); 300 µm (G, H, K, L, O); 600 µm (insets Fig. G and H, P); 700 µm (J); 800 µm (N). The mesophyll is dorsiventral in all specimens, with one or two layers of palisade parenchyma and four or five layers of spongy parenchyma (Fig. 2D, E). The vascular system is formed by collateral bundles (Fig. 2E), which possess bundle-sheath extensions only in sections Cyclostigma and Cupreati (Fig. 2D). Under polarized light, we detected abundant druse crystals in all the studied specimens, but they are found mostly in the ground parenchyma of different regions of the blade (Fig. 2F). Leaf margins are predominantly revolute (Fig. 2G, inset); however, in section Cyclostigma there are specimens with a slightly involute margin (Fig. 2H, inset) as in C. bogotanus, C. draco-2, C. echinocarpus Müll.Arg., C. erythrochilum Müll.Arg., C. erythrochyloides Croizat, C. floccosus and Riina-1416. Leaf margins can also present either continuous (Fig. 2G, H) or discontinuous (Fig. 2I) palisade parenchyma. The midrib is biconvex (Fig. 2J) in sections Cyclostigma, Cupreati and Sampatik and C. abutilifolius Croizat and C. gracilipes Baill. (section Adenophylli) and uniconvex with the adaxial side flat to slightly concave in C. conduplicatus Kunth and C. ruizianus (section Adenophylli) (Fig. 2K). Sections Cyclostigma and Adenophylli possess a midrib with five to ten layers of angular-annular collenchyma (Fig. 2J, L) and a continuous palisade parenchyma (Fig. 2K, L), but the palisade parenchyma is not continuous in species of sections Cupreati and Sampatik. In all the studied species, except C. aequatoris Croizat and C. conduplicatus in section Adenophylli, the vascular system consists of collateral bundles arranged in a main open arch and one to five smaller dorsal bundles and all bundles are surrounded by fibres (Fig. 2J, L, M). In contrast, C. aequatoris and C. conduplicatus are the only species with a single main vascular bundle (Fig. 2K). In all specimens, the petiole in cross section is rounded with a slight depression in the central region of the adaxial side (Fig. 2N). The epidermis is single-layered with tiny cells (Fig. 2O) and covered by a thin cuticle. In the cortical region, nine or ten layers of angular collenchyma are present underlying the epidermis, followed by seven to ten layers of parenchyma. The collateral vascular system presents horseshoe-shaped and two accessory bundles, with fibres in the perivascular region (Fig. 2O). In the adaxial face, adjacent to the main vascular bundle, there are two accessory bundles surrounded by fibres (Fig. 2P). Secretory structures We identified four types of secretory structures with different topology and histological organization distributed in all studied specimens. These were idioblasts, laticifers, colleters and extrafloral nectaries (EFNs). Secretory idioblasts were common in all specimens, and consist of large cells, sometimes strongly stained due to the secretion accumulated inside the vacuole. They were located in the abaxial epidermis (Fig. 3A) and scattered in the mesophyll (Fig. 3B) in most specimens of section Cyclostigma and in all specimens of sections Adenophylli and Sampatik. In section Cupreati, idioblasts were present both in the abaxial epidermis and palisade parenchyma (Fig. 3C), and present only in the palisade parenchyma (Fig. 3D) in C. pseudopopulus Baill., C. floccosus, Riina-1416 and C. urucurana-1. Some idioblasts were situated on the stipe of trichomes (Fig. 3E) in some specimens from all sampled sections, but the presence of these structures in specimens with stipitate trichomes did not show a consistent pattern within species. Figure 3. View largeDownload slide Diversity of secretory structures in Croton. A–E. Secretory idioblasts (black arrows); F–H. Laticifers (white arrows); I. Colleter; J–T. Extrafloral nectaries (EFN); A. Idioblasts in the abaxial epidermis in C. alchorneicarpus; B. Idioblasts scattered in the mesophyll in C. rimbachii; C. Idioblasts in the epidermis and palisade parenchyma in C. cupreatus; D. Idioblasts in the palisade parenchyma in C. pseudopopulus; E. idioblasts situated at the base of the trichomes in C. lechleri; F–H. Non-articulated branched laticifers in the shoot apical meristem (white arrows); F. C. amentiformis; G, H. C. celtidifolius; I. Non-vascularized colleter of the standard type in C. bogotanus; J. Stipitate EFN basilaminar in C. macrobothrys; K, L. EFN acropetiolar; K. Stipitate EFN in C. celtidifolius; L. Sessile EFN in Riina-1520; M. Dispersed EFN on the blade in C. piptocalyx; N. Sessile EFN on the leaf margin in C. piptocalyx; O, P. EFN in both acropetiolar and basilaminar; O. Sessile EFN in C. redolens; P. Stipitate EFN in C. pseudopopulus; Q, R. EFN with concave surface in Riina-1592 and C. aequatoris, respectively; S. EFN with convex surface in C. gossypifolius; T. EFN with flat surface in C. cupreatus. Bars: 50 µm (G, H); 150 µm (A–E); 250 µm (F, I); 300 µm (R); 800 µm (Q, S, T); 0.5 mm (M–O); 1 mm (J–L, P). Figure 3. View largeDownload slide Diversity of secretory structures in Croton. A–E. Secretory idioblasts (black arrows); F–H. Laticifers (white arrows); I. Colleter; J–T. Extrafloral nectaries (EFN); A. Idioblasts in the abaxial epidermis in C. alchorneicarpus; B. Idioblasts scattered in the mesophyll in C. rimbachii; C. Idioblasts in the epidermis and palisade parenchyma in C. cupreatus; D. Idioblasts in the palisade parenchyma in C. pseudopopulus; E. idioblasts situated at the base of the trichomes in C. lechleri; F–H. Non-articulated branched laticifers in the shoot apical meristem (white arrows); F. C. amentiformis; G, H. C. celtidifolius; I. Non-vascularized colleter of the standard type in C. bogotanus; J. Stipitate EFN basilaminar in C. macrobothrys; K, L. EFN acropetiolar; K. Stipitate EFN in C. celtidifolius; L. Sessile EFN in Riina-1520; M. Dispersed EFN on the blade in C. piptocalyx; N. Sessile EFN on the leaf margin in C. piptocalyx; O, P. EFN in both acropetiolar and basilaminar; O. Sessile EFN in C. redolens; P. Stipitate EFN in C. pseudopopulus; Q, R. EFN with concave surface in Riina-1592 and C. aequatoris, respectively; S. EFN with convex surface in C. gossypifolius; T. EFN with flat surface in C. cupreatus. Bars: 50 µm (G, H); 150 µm (A–E); 250 µm (F, I); 300 µm (R); 800 µm (Q, S, T); 0.5 mm (M–O); 1 mm (J–L, P). The shoot meristems of all specimens examined presented laticifers dispersed in ground tissues of leaves, which exhibited a ‘Y’ branching pattern (Fig. 3F–H) and evident secretory activity. These laticifers appear to be non-articulated, although the anatomical preparations from the herbarium material used did not allow a clear enough visualization of the pattern of cellular division to be totally confident about the laticifer type. Colleters occurred exclusively along the leaf margin of all analysed specimens, and they were of the standard type (Fig. 3I), non-vascularized, with a secretory epidermis with a single-layered, radially arranged palisade and a thin cuticle. They have a slightly constricted base and a central axis composed by fundamental parenchyma where secretory idioblasts, druse crystals and laticifers can be present. In fully expanded and mature leaves, colleters eventually fall off with age and they are regarded as deciduous. Extrafloral nectaries were also present in all specimens. They were located either at the base of the leaf blade (basilaminar) (Fig. 3J) or at the distal portion of the petiole (acropetiolar) (Fig. 3K, L). Acropetiolar or basilaminar EFNs were usually in one pair. There were EFNs dispersed on the blade (Fig. 3M) and along the leaf margin (Fig. 3N, inset) only in C. piptocalyx (section Sampatik). The only specimens in section Cyclostigma with more than two acropetiolar and basilaminar EFNs were C. redolens Pittier, C. pseudopopulus and medusae-1 (Carvalho-3789) (Fig. 3O, P). EFNs varied among and within species from sessile (Fig. 3L–O) to stipitate (Fig. 3J, K, P), and their secretory surface can be concave (Fig. 3Q, R), convex (Fig. 3S) or flat (Fig. 3T). Croton piptocalyx was the only species with EFNs scattered over the abaxial side of the blade near the margins (Fig. 3M), the apex of the petiole and alternating with colleters along the margins (Fig. 3N). Basilaminar and acropetiolar EFNs were similar in their structure and were present in all species sampled (Fig. 3Q–T). They consist of a single-layered palisade secretory epidermis, a dense protoplast and a smooth and thick cuticle. Underlying the secretory epidermis, there were ten to 12 layers of nectariferous parenchyma of isodiametric cells with dense cytoplasm, containing druse crystals, secretory idioblasts and laticifers. EFNs were vascularized by bundles originating from the blade and/or the petiole and consisting of both xylem and phloem, but with the latter being more abundant (Fig. 3Q, T). Cluster analysis of anatomical characters The topologies of dendrograms obtained using the Dice-Sorensen or Jaccard similarity indices were identical in their branching and clustering pattern, and we will use the dendrogram based on the widely used Jaccard’s similarity index (Fig. 4) for describing and discussing our main results below. The cluster analysis performed on the 45 characters and 104 specimens examined yielded 44 groups (Fig. 4). These groups were established using similarity values greater or equal to 0.8 (J ≥ 0.8) based on their anatomical similarities (a value of 1 indicates 100% similarity among specimens). We assigned the most appropriate specific taxonomic identity integrating anatomical similarity, shown in the cluster analysis, with information on external morphology, ecological knowledge and geographical data. Thus, of the 44 determined and/or confirmed species, 36 in fact belong to section Cyclostigma, and only six to section Adenophylli, one to section Sampatik and one to section Cupreati (Fig. 4). Some specimens belonging to the same species of Cyclostigma did not group together (e.g. Croton draco, C. coriaceus, C. vulnerarius), so we assigned them a serial number (e.g. draco-1, draco-2). To better explain the overall similarity pattern among specimens, we divided the dendrogram from its base to the tips according to the beginning of the main clusters at different levels. This resulted in three major levels (L1, L2, L3) indicated by dotted vertical lines (Fig. 4). These levels correspond to 0.57, 0.60 and 0.66 similarity values, respectively. Figure 4. View largeDownload slide Dendrogram of the 104 specimens analysed using the Jaccard’s similarity coefficient. Taxonomic identities assigned to each group are next to specimens. Note: specimens without species name belong to Croton section Cyclostigma (star). Labels L1–3 and g1–14 correspond to levels of similarity and groups (see text). Figure 4. View largeDownload slide Dendrogram of the 104 specimens analysed using the Jaccard’s similarity coefficient. Taxonomic identities assigned to each group are next to specimens. Note: specimens without species name belong to Croton section Cyclostigma (star). Labels L1–3 and g1–14 correspond to levels of similarity and groups (see text). Specimens of sections Sampatik, Cupreati and a group of Adenophylli (C. abutilifolius, C. gracilipes) appear more similar anatomically to section Cyclostigma than specimens of Adenophylli g1 and g2. However, in contrast with the position of the Sampatik and Cupreati clusters, the two species of Adenophylli are nested well in the Cyclostigma group (Fig. 4). Thirteen specimens, marked with a star (Fig. 4), could not be assigned to any known species, but they all have morphological affinities to members of section Cyclostigma. They represent about 11 undescribed species of section Cyclostigma. Some of these undetermined specimens, which appear in four different separate clusters in the dendrogram, were assigned to an informal taxonomic group (the ‘medusae’ group) based on leaf anatomy and external morphology (Fig. 4). Subgroups g1 and g2 in Adenophylli were further divided in four subgroups (g6, g7, g8 and g9) based on characteristics, such as acropetiolar EFNs (g6: C. conduplicatus), basilaminar EFNs and appressed-stellate trichomes (g7: C. bonplandianus Baill.), appressed-stellate porrect trichomes and absence of associated bundles in the midrib (g8: C. aequatoris), and presence of one to five associated bundles in the midrib (g9: C. ruizianus). The other group of section Adenophylli (C. abutilifolius, C. gracilipes) and most species of section Cyclostigma were clustered together essentially because the presence of secretory idioblasts on the abaxial epidermis and stipitate-stellate porrect trichomes (Fig. 4). DISCUSSION Significance of leaf anatomical features for Croton systematics Anatomical characters were useful for the taxonomic identification of 36 species of section Cyclostigma and for the detection of either intraspecific variation or wrongly identified specimens. Among the character states exclusive to section Cyclostigma and useful to separate species, were the presence of hypodermis, fasciculate cushion-shaped trichomes (both present in five species), dendritic trichomes (restricted to two species) and the absence of nectaries from leaf margin and blade (in all species of section Cyclostigma). On the other hand, some characters states (presence of hypostomatic leaves, paracytic stomata, dorsiventral mesophyll, collateral bundles, druse crystals, basilaminar/acropetiolar extrafloral nectaries, colleters, non-articulated laticifers and stellate trichomes) were common to all Croton spp. sampled. Several studies, focused on one or a couple of species of section Cyclostigma, have also reported these characters (Rudall, 1994; Webster et al., 1996; Sá-Haiad et al., 2009; Vitarelli et al., 2015; Feio, Riina & Meira, 2016). We identified ten types of non-glandular trichomes and confirmed that section Cyclostigma presented all types described by Webster et al. (1996) plus two additional types (fasciculate cushion-shaped and appressed-stellate porrect trichomes), which are here described for the first time in Croton. Previous research about comparative anatomy of Croton (Webster et al., 1996; Senakun & Chantaranothai, 2010; Liu, Deng & Liao, 2013) or Crotoneae (Sá-Haiad et al., 2009; Vitarelli et al., 2015) showed that trichome types are useful for taxonomy mostly at the species level. On the other hand, we noted trichome morphology seems to be plastic among the studied species, sometimes even within the same individual. Moreover, we demonstrated that simple trichomes may sometimes correspond to immature stages of a more complex type. Thus, trichome characters should be used with caution for taxonomic purposes in Cyclostigma, and potentially in all other sections of Croton. In addition, the absence of lepidote trichomes has been used as a character state to distinguish section Cyclostigma from other clades of Croton such as sections Cupreati, Lasiogyne (Klotzsch) Baill. and Lamprocroton (Müll.Arg.) Pax in Engl. & Prantl (van Ee et al., 2011). However, such trichomes were recently reported in C. macrobothrys Baill. and C. urucurana (Sá-Haiad et al., 2009; Caruzo et al., 2016; Feio, 2016). Here we confirmed that, indeed, all species of section Cyclostigma studied here typically lack these trichomes, except for two specimens of C. urucurana (Pollito-VA-001 and Leitão-Filho-1603), and their presence may vary between populations. Thus, we recommend that these two species be confirmed with additional sampling, given that both taxa have rather large geographical ranges in the Brazilian Atlantic Rain Forest (Caruzo et al., 2016; Santos, Riina & Caruzo, 2017). The presence of a hypodermis in six of the 36 species of section Cyclostigma examined and in C. cupreatus is a new report for Croton. Species of section Cyclostigma occur along mountain creeks, edges and gaps of forests, river sides and along roads (Riina et al., 2009; van Ee et al., 2011). Despite the high humidity present in some of these habitats (e.g. cloud forest, river sides), these Croton spp. are exposed to high luminosity and the presence of a hypodermis is clearly advantageous. In fact, all species with a hypodermis occur in montane forest at high elevation in the Andes, but they are not the only species in section Cyclostigma occupying these habitat types. Systematic implications of secretory structures Previous studies of Crotoneae have reported idioblasts producing only lipophilic secretion (e.g. Sá-Haiad et al., 2009; Vitarelli et al., 2015). However, in a detailed study of two species of section Cyclostigma, Feio et al. (2016) also detected polysaccharides, phenolic compounds, alkaloids and proteins, suggesting a richer and more complex nature of the secretion of idioblasts in members of that section. Vitarelli et al. (2015) described articulated laticifers in leaves and shoots apices in most species of Crotoneae they analysed [i.e. Croton sections Cleodora (Klotzsch) Baill. (12 species) and Lamprocroton (Müll.Arg.) Pax (11 species) and the related genera Astraea Klotzsch (one species) and Brasiliocroton P.E.Berry & Cordeiro (one species)] and suggested that the articulated type might be widespread in subfamily Crotonoideae. In contrast, non-articulated laticifers were reported by Rudall (1994) for seven Croton spp. (C. antisyphiliticus Mart. ex Müll.Arg., C. conduplicatus, C. heteropleurus Urb., C. megalobotrys Müll.Arg., C. occidentalis Müll.Arg., C. sagraeanus Müll.Arg., C. sylvaticus Hochst. and four undetermined species) belonging to three different clades of Croton [sections Geiseleria (A.Gray) Baill. and Adenophylli and the Old World Croton clade]. Non-articulated laticifers were also found in wood rays of section Cyclostigma (Wiedenhoeft et al., 2009) and additionally in leaves and flowers of two representatives of this section, C. echinocarpus and C. urucurana (Feio et al., 2016). This evidence and our own results suggest that both types of laticifers are probably present in the subfamily, although additional studies including a comprehensive taxon sampling is needed to determine which type is more common in Crotonoideae. Colleters have been previously reported in vegetative (leaves) and reproductive organs in three species of section Cyclostigma (Riina et al., 2015; Feio et al., 2016) and representatives of four other Croton sections, namely C. glandulosus L. (section Geiseleria) (Machado et al., 2015) and sections Alabamenses B.W.van Ee, Lamprocroton, and Cleodora (Vitarelli et al., 2015). This last study also reported colleters in two other genera, Brasiliocroton and Astraea, which are closely related to Croton. There is no consensus about the importance of vascularization in colleters. Carlquist (1969) pointed out that the presence of vascularization is directly related to the size of a structure and not necessarily to its state of development. On the other hand, Thomas (1991) suggested that the vascularization of colleters was more related to the derivation or not of vascular bundles from the organs to which colleters are attached, regardless the size of the structure (Arekal & Ramakrishna, 1980; Appezzato-da-Glória & Estelita, 2000). In all studied species, colleters on leaves were non-vascularized and sometimes deciduous (on mature leaves). Deciduous, non-vascularized colleters have been previously reported on the leaves of other Croton spp. (Riina et al., 2015; Vitarelli et al., 2015; Feio et al., 2016). Colleter-deciduousness on mature leaves could be related to the lack of a continuous nourishment in these secretory structures, but this hypothesis needs further investigation. Structural variation and location of extrafloral nectaries were useful features distinguishing group of species at different levels in the cluster analysis. For example, species of section Adenophylli were separated in three subgroups: one group with sessile EFNs and convex surface (g1); one with stipitate EFNs and concave surface (g2) and one (C. abutilifolius, C. gracilipes) clustering with members of section Cyclostigma. The presence of nectaries scattered on the blade and along the leaf margin (alternating with colleters) confirmed the classification of Riina et al. (2009), which removed C. piptocalyx from section Cyclostigma sensuWebster (1993). Delimitation of groups and taxonomic implications Specimens of sections Sampatik, Cupreati and a group of Adenophylli (C. abutilifolius, C. gracilipes) appear more similar anatomically to section Cyclostigma than specimens of Adenophylli g1 and g2. This gives an indication of the homoplasious nature of the anatomical characters used in this study, which is similar to the pattern found with external morphological characters (Riina et al., 2009). The morphoanatomical characters alone will not be enough to separate groups at the sectional level in the genus Croton. However, our results show that the use of anatomical characters could facilitate taxonomic delimitation at the species level in sections of Croton, and it is possible that using anatomical and morphological characters in combination could be even more useful for establishing species limits. There is a recent study establishing species limits in Croton occurring in a region of the northern Andes (Luján, León & Riina, 2015), but these authors only analysed vegetative morphological characters. Specimens Soto-414 and Soto-430 are from the same locality in Bolivia. They are morphologically identical and were initially identified based on external morphology as C. aff. urucurana. These two specimens clustered with Riina-1498, which was initially identified as C. tyndaridum Croizat. It is possible that these three specimens belong to the same species given their high similarity (100%) in anatomical characters. Riina-1498 is problematic because it comes from a cultivated plant in the city of Lima (Peru) and there were no records of its provenance, and its identification as C. tyndaridum was always questionable, since C. tyndaridum is only known from the type locality (Rio Perené, Junín, Peru). Because Soto-414 and Soto-430 did not group with any of the C. urucurana groups, and the uncertainty regarding the identity and origin of Riina-1498, we did not assign any name to this group and consider it as an unknown and probably undescribed species. Croton draco-1 (Berry-7595), despite being originally identified as C. draco based on morphology, clustered with C. rusbyi Britton ex Rusby and not with C. draco-2 (Marquez-558 and Martínez-6054). Croton draco-1 has a bundle-sheath extension, dendritic trichomes and revolute margin, which are features absent in C. draco-2. Croton draco is another widespread and variable species occurring from Mexico to Panama (Webster & Burch, 1968; Webster & Huft, 1988) and, given its wide geographical range, it would have been desirable to have a better sampling of it. Our results show that there might be at least two morphotypes that could be easily separated by anatomical characters. This group obviously merits further investigation using anatomy, morphology and molecular data and a better sampling across its entire range. Croton mutisianus-1 and C. mutisianus-2 also did not cluster together as expected. Instead, C. mutisianus-2 clustered with C. quadrisetosus Lam. It is possible that this grouping is an artefact since C. mutisianus-2 has the highest percentage of missing data (21%) in our matrix. In fact, C. mutisianus-2 (Barkley-3768) was a poor sample and it was not possible to obtain information about EFN and stem apex to characterize the type of laticifers. External morphology does not support this grouping either, since C. mutisianus-2 and C. quadrisetosus have many differences in floral morphology (e.g. flowers with long pedicels, and stigmas bifid and terete in C. mutisianus vs. flowers sessile to subsessile and stigmas multifid with expanded and flattened tips in C. quadrisetosus). In addition, C. mutisianus is widespread in montane forests of Colombia, and C. quadrisetosus is only known from montane forests of Peru (Berry et al., 2015). Unexpected grouping patterns, similar to those discussed above, were found for specimens previously identified using external morphology as C. celtidifolius, C. vulnerarius and C. coriaceus. Specimens from these three species show relatively low levels of similarity in anatomical characters appearing in separate groups of the dendrogram. It is possible that in some cases the species are indeed more polymorphic in relation to anatomical characters than to characters from external morphology. However, the limited sampling prevents us from using anatomical data alone to further subdivide such species, especially because the external morphology does not support the observed clustering pattern based on anatomy. All these taxa need further investigation using a broader sampling across their geographical range and integration of anatomy and morphology in the evaluation of species limits. Our study provides additional characters that can contribute to clarify the taxonomic affinities of species in section Cyclostigma, and it gives support to the taxonomic position of the section adopted by Riina et al. (2009) and van Ee et al. (2011). As with external morphology (Riina et al., 2009), section Cyclostigma can also be characterized by a suite of anatomical characters and so far only one anatomical wood character (secondary xylem rays containing non-articulated laticifers) has been identified as a possible synapomorphy for the section (Wiedenhoeft et al., 2009). CONCLUSIONS AND FUTURE WORK Although we did not find an anatomical character present in all members of section Cyclostigma, the use of a suite of characters was useful to group specimens based on their similarities and to confirm and/or to determine taxonomic identities at the species level in most cases. However, we also recognize the limitation of these characters in several taxa for which taxonomic resolution was not achieved. The presence of different types of stellate trichomes, non-articulated laticifers, absence of marginal EFNs and the lack of lepidote trichomes can be used in combination to define section Cyclostigma anatomically. Three new anatomical features were described for Croton: hypodermis, fasciculate cushion-shaped trichomes and appressed-stellate porrect trichomes. Our results do not fully support the hypothesis of a widespread occurrence of articulated laticifers in Crotonoideae, but future studies for re-evaluating laticifers in this subfamily are necessary and should include analyses of their ontogeny and the nature of their secretions. Finally, we highlight the partial utility of anatomical characters at the species level and the need of using them along with other characters in an integrative taxonomic framework. ACKNOWLEDGEMENTS ACF thanks FAPEMIG for the scholarship in Brazil during her doctoral degree, the CAPES foundation through PDSE (99999.009518/2014-02) supporting her stay at the Real Jardín Botánico-CSIC (Madrid) and the Programa de Capacitação Institucional (MPEG/MCTI) for the current CNPq research grant. RMSAM was supported by CNPq (477867/2013–8). RR was supported by a Visiting Scholar Fellowship to Brazil funded by FAPEMIG (CRA-BPV-00043-14) and CNPq research grant (477867/2013–8). Additional support was provided by Centro de Ciências Ambientais Floresta Escola (SECTES/UNESCO/HidroEX/FAPEMIG). We are grateful to two reviewers and the associate editor for their valuable suggestions and critical comments to earlier versions of this manuscript. REFERENCES Appezzato-da-Glória B , Estelita MEM . 2000 . Development, structure and distribution of colleters in Mandevilla illustris and M. velutina (Apocynaceae) . Revista brasileira de Botânica 23 : 113 – 120 . Arekal GD , Ramakrishna TM . 1980 . Extrafloral nectaries of Calotropis gigantea and Wattakaka volubilis . Phytomorphology 30 : 303 – 306 . Berry P , Riina R , van Ee B . 2015 . Croton mutisianus Kunth . In: Bernal R , Gradstein SR , Celis M , eds. Catálogo de plantas y líquenes de Colombia. Instituto de Ciencias Naturales, Universidad Nacional de Colombia, Bogotá . Available at: http://catalogoplantasdecolombia.unal.edu.co. Borges JR , King SR . 2000 . Croton lechleri, sustainable utilization of an Amazonian pioneer species . Medicinal Plant Conservation 6 : 24 – 26 . Bozzola JJ , Russel LD . 1992 . Electron microscopy . Boston : Jones and Bartlett Publishers . Carlquist S . 1969 . Toward acceptable evolutionary interpretations of floral anatomy . Phytomorphology 19 : 332 – 362 . Caruzo RFS , Santos RF , Feio AC , Meira RMSA , Riina R . 2016 . Redefinition of Croton macrobothrys (Euphorbiaceae), a tree species from the Brazilian Atlantic Forest, with the description of a new subspecies . Phytotaxa 273 : 51 – 58 . Christenhusz MJ , Fay MF , Chase MW . 2017 . Plants of the world: an illustrated encyclopedia of vascular plants . Kew and Chicago : Kew Publishing/University of Chicago Press . Evert RF . 2006 . Esau’s plant anatomy. Meristems, cells, and tissues of the plant body – their structure, function, and development , 3th edn . Hoboken : Wiley . Feio AC . 2016 . Croton sect. Cyclostigma (Euphorbiaceae): novidades anatômicas e taxonômicas . PhD Thesis, Universidade Federal de Viçosa , Minas Gerais . Feio AC , Riina R , Meira RMSA . 2016 . Secretory structures in leaves and flowers of two dragon’s blood Croton (Euphorbiaceae): new evidence and interpretations . International Journal of Plant Sciences 177 : 511 – 522 . Feio AC , Ore-Rengifo MI , Berry PE , Riina R . 2018 . Four new species of dragon’s blood Croton (Euphorbiaceae) from South America . Systematic Botany ( in press ). Haber EA , Kainulainen K , Van Ee BW , Oyserman BO , Berry PE . 2017 . Phylogenetic relationships of a major diversification of Croton (Euphorbiaceae) in the western Indian Ocean region . Botanical Journal of the Linnean Society 183 : 532 – 544 . Hammer Ø , Harper DAT , Ryan PD . 2001 . PAST: paleontological statistics software package for education and data analysis . Palaeontologia Electronica 4 : 1 – 9 . Jones K . 2003 . Review of sangre de drago (Croton lechleri) – a South American tree sap in the treatment of diarrhea, inflammation, insect bites, viral infections, and wounds: traditional uses to clinical research . The Journal of Alternative and Complementary Medicine 9 : 877 – 896 . Kaiser E . 1880 . Verfahren zur Herstellung einer tadellosen Glycerin-gelatine . Botanisch Zentralblatt, Stuttgart 180 : 25 – 26 . Liu HF , Deng YF , Liao JP . 2013 . Foliar trichomes of Croton L. (Euphorbiaceae: Crotonoideae) from China and its taxonomic implications . Bangladesh Journal of Plant Taxonomy 20 : 85 – 94 . Luján M , León Y , Riina R . 2015 . Synopsis of Croton (Euphobiaceae) in the Andes of Mérida, Venezuela . Caldasia 37 : 73 – 90 . Machado SR , Paleari LM , Paiva EAS , Rodrigues TM . 2015 . Colleters on the inflorescence axis of Croton glandulosus (Euphorbiaceae): structural and functional characterization . International Journal of Plant Sciences 176 : 86 – 93 . Meira RMSA , Martins FM . 2003 . Inclusão de material herborizado em metacrilato para estudos de anatomia vegetal . Revista Árvore 27 : 109 – 112 . Metcalfe CR , Chalk L . 1979 . Anatomy of the dicotyledons: systematic anatomy of the leaf and stem , Vol. I , 2nd edn . Oxford : Oxford University Press . Metcalfe CR , Chalk L . 1983 . Anatomy of the dicotyledons , Vol. 2 , 2nd edn . Oxford : Clarendon Press . Meza EN . 1999 . Nombres aborıgenes peruanos de las especies de Croton que producen el latex denominado ‘sangre de grado’ . In: Mesa EN , ed. Desarrollando nuestra diversidad cultural: ‘sangre de grado’ y el reto de su producción en el Peru . San Marcos : Universidad Nacional Mayor de San Marcos, Fondo Editorial , 25 – 44 . O’Brien TP , Feder N , McCully ME . 1965 . Polychromatic staining of plant cells walls by toluidine blue O . Protoplasma 59 : 368 – 373 . Riina R , Berry PE , Cornejo X . 2007 . A new species of ‘sangre de drago’ (Croton section Cyclostigma, Euphorbiaceae) from coastal Ecuador . Brittonia 59 : 97 – 101 . Riina R , Berry PE , van Ee BW . 2009 . Molecular phylogenetics of the dragon’s blood Croton section Cyclostigma (Euphorbiaceae): a polyphyletic assemblage unraveled . Systematic Botany 34 : 360 – 374 . Riina R , Vigo MA , Cerón CE . 2014 . Croton condorensis: an enigmatic new species of Euphorbiaceae from southern Ecuador . Phytotaxa 164 : 154 – 158 . Riina R , Cumbicus N , Feio AC , Cerón CE , Meira RMSA , Berry PE . 2015 . A new species of dragon’s blood Croton (Euphorbiaceae) from South America with singular inflorescences . Webbia 70 : 187 – 192 . Rudall PJ . 1994 . Laticifers in Crotonoideae (Euphorbiaceae): homology and evolution . Annals of the Missouri Botanical Garden 81 : 270 – 282 . Sá-Haiad B , Serpa-Ribeiro ACC , Barbosa CN , Pizzini D , Leal DO , Senna-Valle L , Santiago-Fernandes LDR . 2009 . Leaf structure of species from three closely related genera from tribe Crotoneae Dumort. (Euphorbiaceae s.s., Malpighiales) . Plant Systematics and Evolution 283 : 179 – 202 . Salatino A , Salatino MLF , Negri G . 2007 . Traditional uses, chemistry and pharmacology of Croton species (Euphorbiaceae) . Journal of the Brazilian Chemical Society 18 : 11 – 33 . Santos RF , Riina R , Caruzo MBR . 2017 . Diversity of arborescent lineages of Crotoneae (Euphorbiaceae) in the Brazilian Atlantic Rain Forest . Plant Systematics and Evolution 303 : 1467 – 1497 Senakun C , Chantaranothai P . 2010 . A morphological survey of foliar trichomes of Croton L. (Euphorbiaceae) in Thailand . Thai Forest Bulletin, Botany 38 : 167 – 172 . Shobe WR , Lersten NR . 1967 . A technique for clearing and staining gymnosperm leaves . Botanical Gazette 128 : 150 – 152 . Smith FH , Smith EC . 1942 . Anatomy of the inferior ovary of Darbya . American Journal of Botany 29 : 464 – 471 . Thiers B . 2017 . [continuously updated]. Index Herbariorum: a global directory of public herbaria and associated staff . New York : New York Botanical Garden’s Virtual Herbarium . Available at: http://sweetgum.nybg.org/science/ih/. Thomas V . 1991 . Structural, functional and phylogenetic aspects of the colleter . Annals of Botany 68 : 287 – 305 . van Ee BW , Riina R , Berry PE . 2011 . A revised infrageneric classification and molecular phylogeny of new world Croton (Euphorbiaceae) . Taxon 60 : 791 – 823 . Vitarelli NC , Riina R , Caruzo MBR , Cordeiro I , Fuertes-Aguilar J , Meira RMSA . 2015 . Foliar secretory structures in Crotoneae (Euphorbiaceae): diversity, anatomy, and evolutionary significance . American Journal of Botany 12 : 833 – 847 . Webster GL . 1993 . A provisional synopsis of the sections of the genus Croton (Euphorbiaceae) . Taxon 42 : 793 – 823 . Webster GL , Burch D . 1968 . Euphorbiaceae. In: Flora of Panama . Annals of the Missouri Botanical Garden 54 : 211 – 350 . Webster GL , Huft MJ . 1988 . Revised synopsis of Panamanian Euphorbiaceae . Annals of the Missouri Botanical Garden 75 : 1087 – 1144 . Webster GL , Del-Arco Aguilar MJ , Smith BA . 1996 . Systematic distribution of foliar trichome types in Croton (Euphorbiaceae) . Botanical Journal of the Linnean Society 121 : 41 – 57 . Wiedenhoeft AC , Riina R , Berry PE . 2009 . ‘Ray-intrusive’ laticifers in species of Croton section Cyclostigma (Euphorbiaceae) . IAWA Journal 30 : 135 – 148 . Appendix 1. List of species with their most recent taxonomic identification (previous to this study) and voucher information. Herbarium acronyms in parentheses according to Thiers (2016). Species Specimen number, collector and collection number (herbarium) Croton section Cyclostigma C. alchorneicarpus Croizat 1: Riina 1526 (MICH); 2: Barreto 206 (MG); 3: Riina 1531 (MICH); 4: Occhioni 8208 (MBM) C. amentiformis Riina 5: Jorgensen 97 (MO) C. bogotanus Cuatrec. 6: Riina 1591 (MICH) C. aff. celtidifolius/pseudopopulus 7: Cavalcante 1166 (MG) C. celtidifolius Baill. 8: Catharino 640 (MG); 9: Riina 1520 (MICH); 10: Riina 1530 (MICH); 11: Shepherd 12205 (MG) C. charaguensis Standl. 12: Riina 1513 (MICH) C. churutensis Riina & Cornejo 13: Cornejo 7590 (MICH) C. coriaceus Kunth 14: Berry 7603 (MICH); 15: Riina 1403 (MICH); 16: Riina 1417 (MICH) C. aff. coriaceus Kunth 17: Berry 7620 (MICH) C. draco Schltdl. 18: Berry 7595 (MICH); 19: Marquez 558 (MBM); 20: Martínez 6054 (MBM) C. echinocarpus Müll.Arg. 21: Riina 1316 (MICH); 22: Riina 1519 (MICH); 23: Batista 01 (VIC) C. erythrochyloides Croizat 24: Riina 1503 (MICH) C. fastuosus Baill. 25: Costa s/n (BHCB) C. floccosus B.A.Sm. 26: Riina 1405 (MICH); 27: Riina 1406 (MICH); 28: Riina 1407 (MICH) C. glaziovii Müll.Arg. 29: Riina 1521 (MICH) C. gossypiifolius Vahl 30: Riina 1261 (MICH); 31: Riina 1303 (MICH) C. hibiscifolius Kunth ex Spreng. 32: Breteler 3446 (MG); 33: Contreras 042 (MICH); 34: Riina 1413 (MICH); 35: Riina 1414 (MICH) C. aff. hibiscifolius 36: Riina 1592 (MA); 37: Orsini 2013-13 (MYF) C. huberi Steyerm. 38: Riina 1276 (MICH); 39: Steyermark s.n. (MG) C. lechleri Müll.Arg. 40: Riina 1443 (MICH); 41: Riina 1449 (MICH); 42: Riina 1496 (MICH); 43: Riina 1497 (MICH) C. aff. lechleri Müll.Arg. 44: Riina 1453 (MICH) C. macrobothrys Baill. 45: Barros 2014 (MG); 46: Custodio Filho 1905 (MG); 47: Mori s.n. (MG); 48: Riina 1522 (MICH) C. aff. medusae Müll.Arg. 49: Pirani 4982 (SPF) C. aff. mutisianus Kunth 50: Riina 1590 (UEN, MA) C. mutisianus Kunth 51: Barkley 3768 (MBM) C. perspeciosus Croizat 52: Riina 1435 (MICH); 53: Riina 1441 (MICH); 54: Riina 1444 (MICH) C. pilulifer Rusby 55: Riina 1500 (MICH); 56: Riina 1508 (MICH) C. aff. pilulifer Rusby 57: Riina 1461 (MICH); 58: Riina 1462 (MICH); 59: Riina 1483 (MICH) C. plagiograptus Müll.Arg. 60: Carvalho 3789 (MBM) C. pseudopopulus Baill. 61: Mota 2276 (VIC); 62: Mota 2284 (VIC); 63: Mota 2291 (VIC) C. quadrisetosus Lam. 64: Daza 4051 (E) C. redolens Pittier 65: Riina 1848 (MICH); 66: Riina 1850 (MICH); 67: Webster 23688 (MICH) C. aff. redolens Pittier 68: Webster 23620 (MICH) C. rimbachii Croizat 69: Riina 1402 (MICH); 70: Riina 1422 (MICH); 71: Riina 1440 (MICH) C. rusbyi Britton ex Rusby 72: Riina 1479 (MICH); 73: Riina 1481 (MICH) C. sp1 74: Daza 4572 (E) C. sp2 75: Riina 1426 (MICH) C. sp3 76: Riina 1416 (MICH) C. speciosus Müll.Arg. 77: Berry 7590 (MICH); 78: Riina 1262 (MICH); 79: Riina 1278 (MICH) C. tyndaridum Croizat 80: Riina 1498 (MICH) C. urucurana Baill. 81: Heringer 18473 (MG); 82: Leitão-Filho-1603 (MG); 83: Pollito-VA-001 (MG) C. aff. urucurana Baill. 84: Cordeiro 3325 (SP); 85: Soto 430 (USZ) C. aff. urucurana/lechleri 86: Soto 414 (USZ) C. vulnerarius Baill. 87: Cordeiro 345 (MG); 88: Forero 8148 (MG); 89: Saran 07 (MG); 90: Silva 1242 (VIC) Croton section Adenophylli C. abutilifolius Croizat 91: Riina 1505 (MA) C. aequatoris Croizat 92: Riina 1434 (MA) C. bonplandianus Baill. 93: Riina 1517 (LPB) C. conduplicatus Kunth 94: Riina 1266 (VEN); 95: Riina 1302 (VEN); 96: Riina 1296 (MICH); 97: Riina 1837 (MA) C. gracilipes Baill. 98: Riina 1501 (MA) C. ruizianus Müll.Arg. 99: Riina 1386 (MICH); 100: Riina 1486 (MICH); 101: Riina 1487 (MICH) Croton section Sampatik C. piptocalyx Müll. Arg. 102: Bortoluzzi 379 (VIC); 103: Bortoluzzi 381 (VIC) Croton section Cupreati C. cupreatus Croizat 104: Riina 1408 (MA) Species Specimen number, collector and collection number (herbarium) Croton section Cyclostigma C. alchorneicarpus Croizat 1: Riina 1526 (MICH); 2: Barreto 206 (MG); 3: Riina 1531 (MICH); 4: Occhioni 8208 (MBM) C. amentiformis Riina 5: Jorgensen 97 (MO) C. bogotanus Cuatrec. 6: Riina 1591 (MICH) C. aff. celtidifolius/pseudopopulus 7: Cavalcante 1166 (MG) C. celtidifolius Baill. 8: Catharino 640 (MG); 9: Riina 1520 (MICH); 10: Riina 1530 (MICH); 11: Shepherd 12205 (MG) C. charaguensis Standl. 12: Riina 1513 (MICH) C. churutensis Riina & Cornejo 13: Cornejo 7590 (MICH) C. coriaceus Kunth 14: Berry 7603 (MICH); 15: Riina 1403 (MICH); 16: Riina 1417 (MICH) C. aff. coriaceus Kunth 17: Berry 7620 (MICH) C. draco Schltdl. 18: Berry 7595 (MICH); 19: Marquez 558 (MBM); 20: Martínez 6054 (MBM) C. echinocarpus Müll.Arg. 21: Riina 1316 (MICH); 22: Riina 1519 (MICH); 23: Batista 01 (VIC) C. erythrochyloides Croizat 24: Riina 1503 (MICH) C. fastuosus Baill. 25: Costa s/n (BHCB) C. floccosus B.A.Sm. 26: Riina 1405 (MICH); 27: Riina 1406 (MICH); 28: Riina 1407 (MICH) C. glaziovii Müll.Arg. 29: Riina 1521 (MICH) C. gossypiifolius Vahl 30: Riina 1261 (MICH); 31: Riina 1303 (MICH) C. hibiscifolius Kunth ex Spreng. 32: Breteler 3446 (MG); 33: Contreras 042 (MICH); 34: Riina 1413 (MICH); 35: Riina 1414 (MICH) C. aff. hibiscifolius 36: Riina 1592 (MA); 37: Orsini 2013-13 (MYF) C. huberi Steyerm. 38: Riina 1276 (MICH); 39: Steyermark s.n. (MG) C. lechleri Müll.Arg. 40: Riina 1443 (MICH); 41: Riina 1449 (MICH); 42: Riina 1496 (MICH); 43: Riina 1497 (MICH) C. aff. lechleri Müll.Arg. 44: Riina 1453 (MICH) C. macrobothrys Baill. 45: Barros 2014 (MG); 46: Custodio Filho 1905 (MG); 47: Mori s.n. (MG); 48: Riina 1522 (MICH) C. aff. medusae Müll.Arg. 49: Pirani 4982 (SPF) C. aff. mutisianus Kunth 50: Riina 1590 (UEN, MA) C. mutisianus Kunth 51: Barkley 3768 (MBM) C. perspeciosus Croizat 52: Riina 1435 (MICH); 53: Riina 1441 (MICH); 54: Riina 1444 (MICH) C. pilulifer Rusby 55: Riina 1500 (MICH); 56: Riina 1508 (MICH) C. aff. pilulifer Rusby 57: Riina 1461 (MICH); 58: Riina 1462 (MICH); 59: Riina 1483 (MICH) C. plagiograptus Müll.Arg. 60: Carvalho 3789 (MBM) C. pseudopopulus Baill. 61: Mota 2276 (VIC); 62: Mota 2284 (VIC); 63: Mota 2291 (VIC) C. quadrisetosus Lam. 64: Daza 4051 (E) C. redolens Pittier 65: Riina 1848 (MICH); 66: Riina 1850 (MICH); 67: Webster 23688 (MICH) C. aff. redolens Pittier 68: Webster 23620 (MICH) C. rimbachii Croizat 69: Riina 1402 (MICH); 70: Riina 1422 (MICH); 71: Riina 1440 (MICH) C. rusbyi Britton ex Rusby 72: Riina 1479 (MICH); 73: Riina 1481 (MICH) C. sp1 74: Daza 4572 (E) C. sp2 75: Riina 1426 (MICH) C. sp3 76: Riina 1416 (MICH) C. speciosus Müll.Arg. 77: Berry 7590 (MICH); 78: Riina 1262 (MICH); 79: Riina 1278 (MICH) C. tyndaridum Croizat 80: Riina 1498 (MICH) C. urucurana Baill. 81: Heringer 18473 (MG); 82: Leitão-Filho-1603 (MG); 83: Pollito-VA-001 (MG) C. aff. urucurana Baill. 84: Cordeiro 3325 (SP); 85: Soto 430 (USZ) C. aff. urucurana/lechleri 86: Soto 414 (USZ) C. vulnerarius Baill. 87: Cordeiro 345 (MG); 88: Forero 8148 (MG); 89: Saran 07 (MG); 90: Silva 1242 (VIC) Croton section Adenophylli C. abutilifolius Croizat 91: Riina 1505 (MA) C. aequatoris Croizat 92: Riina 1434 (MA) C. bonplandianus Baill. 93: Riina 1517 (LPB) C. conduplicatus Kunth 94: Riina 1266 (VEN); 95: Riina 1302 (VEN); 96: Riina 1296 (MICH); 97: Riina 1837 (MA) C. gracilipes Baill. 98: Riina 1501 (MA) C. ruizianus Müll.Arg. 99: Riina 1386 (MICH); 100: Riina 1486 (MICH); 101: Riina 1487 (MICH) Croton section Sampatik C. piptocalyx Müll. Arg. 102: Bortoluzzi 379 (VIC); 103: Bortoluzzi 381 (VIC) Croton section Cupreati C. cupreatus Croizat 104: Riina 1408 (MA) Appendix 1. List of species with their most recent taxonomic identification (previous to this study) and voucher information. Herbarium acronyms in parentheses according to Thiers (2016). Species Specimen number, collector and collection number (herbarium) Croton section Cyclostigma C. alchorneicarpus Croizat 1: Riina 1526 (MICH); 2: Barreto 206 (MG); 3: Riina 1531 (MICH); 4: Occhioni 8208 (MBM) C. amentiformis Riina 5: Jorgensen 97 (MO) C. bogotanus Cuatrec. 6: Riina 1591 (MICH) C. aff. celtidifolius/pseudopopulus 7: Cavalcante 1166 (MG) C. celtidifolius Baill. 8: Catharino 640 (MG); 9: Riina 1520 (MICH); 10: Riina 1530 (MICH); 11: Shepherd 12205 (MG) C. charaguensis Standl. 12: Riina 1513 (MICH) C. churutensis Riina & Cornejo 13: Cornejo 7590 (MICH) C. coriaceus Kunth 14: Berry 7603 (MICH); 15: Riina 1403 (MICH); 16: Riina 1417 (MICH) C. aff. coriaceus Kunth 17: Berry 7620 (MICH) C. draco Schltdl. 18: Berry 7595 (MICH); 19: Marquez 558 (MBM); 20: Martínez 6054 (MBM) C. echinocarpus Müll.Arg. 21: Riina 1316 (MICH); 22: Riina 1519 (MICH); 23: Batista 01 (VIC) C. erythrochyloides Croizat 24: Riina 1503 (MICH) C. fastuosus Baill. 25: Costa s/n (BHCB) C. floccosus B.A.Sm. 26: Riina 1405 (MICH); 27: Riina 1406 (MICH); 28: Riina 1407 (MICH) C. glaziovii Müll.Arg. 29: Riina 1521 (MICH) C. gossypiifolius Vahl 30: Riina 1261 (MICH); 31: Riina 1303 (MICH) C. hibiscifolius Kunth ex Spreng. 32: Breteler 3446 (MG); 33: Contreras 042 (MICH); 34: Riina 1413 (MICH); 35: Riina 1414 (MICH) C. aff. hibiscifolius 36: Riina 1592 (MA); 37: Orsini 2013-13 (MYF) C. huberi Steyerm. 38: Riina 1276 (MICH); 39: Steyermark s.n. (MG) C. lechleri Müll.Arg. 40: Riina 1443 (MICH); 41: Riina 1449 (MICH); 42: Riina 1496 (MICH); 43: Riina 1497 (MICH) C. aff. lechleri Müll.Arg. 44: Riina 1453 (MICH) C. macrobothrys Baill. 45: Barros 2014 (MG); 46: Custodio Filho 1905 (MG); 47: Mori s.n. (MG); 48: Riina 1522 (MICH) C. aff. medusae Müll.Arg. 49: Pirani 4982 (SPF) C. aff. mutisianus Kunth 50: Riina 1590 (UEN, MA) C. mutisianus Kunth 51: Barkley 3768 (MBM) C. perspeciosus Croizat 52: Riina 1435 (MICH); 53: Riina 1441 (MICH); 54: Riina 1444 (MICH) C. pilulifer Rusby 55: Riina 1500 (MICH); 56: Riina 1508 (MICH) C. aff. pilulifer Rusby 57: Riina 1461 (MICH); 58: Riina 1462 (MICH); 59: Riina 1483 (MICH) C. plagiograptus Müll.Arg. 60: Carvalho 3789 (MBM) C. pseudopopulus Baill. 61: Mota 2276 (VIC); 62: Mota 2284 (VIC); 63: Mota 2291 (VIC) C. quadrisetosus Lam. 64: Daza 4051 (E) C. redolens Pittier 65: Riina 1848 (MICH); 66: Riina 1850 (MICH); 67: Webster 23688 (MICH) C. aff. redolens Pittier 68: Webster 23620 (MICH) C. rimbachii Croizat 69: Riina 1402 (MICH); 70: Riina 1422 (MICH); 71: Riina 1440 (MICH) C. rusbyi Britton ex Rusby 72: Riina 1479 (MICH); 73: Riina 1481 (MICH) C. sp1 74: Daza 4572 (E) C. sp2 75: Riina 1426 (MICH) C. sp3 76: Riina 1416 (MICH) C. speciosus Müll.Arg. 77: Berry 7590 (MICH); 78: Riina 1262 (MICH); 79: Riina 1278 (MICH) C. tyndaridum Croizat 80: Riina 1498 (MICH) C. urucurana Baill. 81: Heringer 18473 (MG); 82: Leitão-Filho-1603 (MG); 83: Pollito-VA-001 (MG) C. aff. urucurana Baill. 84: Cordeiro 3325 (SP); 85: Soto 430 (USZ) C. aff. urucurana/lechleri 86: Soto 414 (USZ) C. vulnerarius Baill. 87: Cordeiro 345 (MG); 88: Forero 8148 (MG); 89: Saran 07 (MG); 90: Silva 1242 (VIC) Croton section Adenophylli C. abutilifolius Croizat 91: Riina 1505 (MA) C. aequatoris Croizat 92: Riina 1434 (MA) C. bonplandianus Baill. 93: Riina 1517 (LPB) C. conduplicatus Kunth 94: Riina 1266 (VEN); 95: Riina 1302 (VEN); 96: Riina 1296 (MICH); 97: Riina 1837 (MA) C. gracilipes Baill. 98: Riina 1501 (MA) C. ruizianus Müll.Arg. 99: Riina 1386 (MICH); 100: Riina 1486 (MICH); 101: Riina 1487 (MICH) Croton section Sampatik C. piptocalyx Müll. Arg. 102: Bortoluzzi 379 (VIC); 103: Bortoluzzi 381 (VIC) Croton section Cupreati C. cupreatus Croizat 104: Riina 1408 (MA) Species Specimen number, collector and collection number (herbarium) Croton section Cyclostigma C. alchorneicarpus Croizat 1: Riina 1526 (MICH); 2: Barreto 206 (MG); 3: Riina 1531 (MICH); 4: Occhioni 8208 (MBM) C. amentiformis Riina 5: Jorgensen 97 (MO) C. bogotanus Cuatrec. 6: Riina 1591 (MICH) C. aff. celtidifolius/pseudopopulus 7: Cavalcante 1166 (MG) C. celtidifolius Baill. 8: Catharino 640 (MG); 9: Riina 1520 (MICH); 10: Riina 1530 (MICH); 11: Shepherd 12205 (MG) C. charaguensis Standl. 12: Riina 1513 (MICH) C. churutensis Riina & Cornejo 13: Cornejo 7590 (MICH) C. coriaceus Kunth 14: Berry 7603 (MICH); 15: Riina 1403 (MICH); 16: Riina 1417 (MICH) C. aff. coriaceus Kunth 17: Berry 7620 (MICH) C. draco Schltdl. 18: Berry 7595 (MICH); 19: Marquez 558 (MBM); 20: Martínez 6054 (MBM) C. echinocarpus Müll.Arg. 21: Riina 1316 (MICH); 22: Riina 1519 (MICH); 23: Batista 01 (VIC) C. erythrochyloides Croizat 24: Riina 1503 (MICH) C. fastuosus Baill. 25: Costa s/n (BHCB) C. floccosus B.A.Sm. 26: Riina 1405 (MICH); 27: Riina 1406 (MICH); 28: Riina 1407 (MICH) C. glaziovii Müll.Arg. 29: Riina 1521 (MICH) C. gossypiifolius Vahl 30: Riina 1261 (MICH); 31: Riina 1303 (MICH) C. hibiscifolius Kunth ex Spreng. 32: Breteler 3446 (MG); 33: Contreras 042 (MICH); 34: Riina 1413 (MICH); 35: Riina 1414 (MICH) C. aff. hibiscifolius 36: Riina 1592 (MA); 37: Orsini 2013-13 (MYF) C. huberi Steyerm. 38: Riina 1276 (MICH); 39: Steyermark s.n. (MG) C. lechleri Müll.Arg. 40: Riina 1443 (MICH); 41: Riina 1449 (MICH); 42: Riina 1496 (MICH); 43: Riina 1497 (MICH) C. aff. lechleri Müll.Arg. 44: Riina 1453 (MICH) C. macrobothrys Baill. 45: Barros 2014 (MG); 46: Custodio Filho 1905 (MG); 47: Mori s.n. (MG); 48: Riina 1522 (MICH) C. aff. medusae Müll.Arg. 49: Pirani 4982 (SPF) C. aff. mutisianus Kunth 50: Riina 1590 (UEN, MA) C. mutisianus Kunth 51: Barkley 3768 (MBM) C. perspeciosus Croizat 52: Riina 1435 (MICH); 53: Riina 1441 (MICH); 54: Riina 1444 (MICH) C. pilulifer Rusby 55: Riina 1500 (MICH); 56: Riina 1508 (MICH) C. aff. pilulifer Rusby 57: Riina 1461 (MICH); 58: Riina 1462 (MICH); 59: Riina 1483 (MICH) C. plagiograptus Müll.Arg. 60: Carvalho 3789 (MBM) C. pseudopopulus Baill. 61: Mota 2276 (VIC); 62: Mota 2284 (VIC); 63: Mota 2291 (VIC) C. quadrisetosus Lam. 64: Daza 4051 (E) C. redolens Pittier 65: Riina 1848 (MICH); 66: Riina 1850 (MICH); 67: Webster 23688 (MICH) C. aff. redolens Pittier 68: Webster 23620 (MICH) C. rimbachii Croizat 69: Riina 1402 (MICH); 70: Riina 1422 (MICH); 71: Riina 1440 (MICH) C. rusbyi Britton ex Rusby 72: Riina 1479 (MICH); 73: Riina 1481 (MICH) C. sp1 74: Daza 4572 (E) C. sp2 75: Riina 1426 (MICH) C. sp3 76: Riina 1416 (MICH) C. speciosus Müll.Arg. 77: Berry 7590 (MICH); 78: Riina 1262 (MICH); 79: Riina 1278 (MICH) C. tyndaridum Croizat 80: Riina 1498 (MICH) C. urucurana Baill. 81: Heringer 18473 (MG); 82: Leitão-Filho-1603 (MG); 83: Pollito-VA-001 (MG) C. aff. urucurana Baill. 84: Cordeiro 3325 (SP); 85: Soto 430 (USZ) C. aff. urucurana/lechleri 86: Soto 414 (USZ) C. vulnerarius Baill. 87: Cordeiro 345 (MG); 88: Forero 8148 (MG); 89: Saran 07 (MG); 90: Silva 1242 (VIC) Croton section Adenophylli C. abutilifolius Croizat 91: Riina 1505 (MA) C. aequatoris Croizat 92: Riina 1434 (MA) C. bonplandianus Baill. 93: Riina 1517 (LPB) C. conduplicatus Kunth 94: Riina 1266 (VEN); 95: Riina 1302 (VEN); 96: Riina 1296 (MICH); 97: Riina 1837 (MA) C. gracilipes Baill. 98: Riina 1501 (MA) C. ruizianus Müll.Arg. 99: Riina 1386 (MICH); 100: Riina 1486 (MICH); 101: Riina 1487 (MICH) Croton section Sampatik C. piptocalyx Müll. Arg. 102: Bortoluzzi 379 (VIC); 103: Bortoluzzi 381 (VIC) Croton section Cupreati C. cupreatus Croizat 104: Riina 1408 (MA) Appendix 2. Binary matrix with 104 specimens and 45 characters. (0) absent; (1) present; (?) unknown; (-) inapplicable. Specimen Anatomical characters 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 1 1 0 0 1 0 1 0 0 1 0 0 0 1 1 0 0 0 0 0 0 1 0 0 0 1 0 1 0 1 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 2 1 0 0 1 0 1 0 0 1 0 0 0 1 0 1 0 0 0 0 0 1 0 0 0 1 0 1 0 1 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 3 1 0 0 1 0 1 0 0 1 0 0 0 1 1 0 0 0 0 0 0 1 0 0 0 1 0 1 0 1 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 4 1 0 0 1 0 1 0 0 1 0 0 0 1 1 0 0 0 0 0 0 1 0 0 0 1 0 1 0 1 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 5 1 1 1 1 0 1 0 1 1 0 0 0 1 0 1 0 0 0 0 0 0 1 0 0 0 0 1 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 6 1 0 0 1 1 1 0 1 0 0 0 0 1 1 0 0 ? ? ? ? ? ? ? ? ? ? ? 0 1 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 7 1 0 1 1 1 1 0 0 1 0 0 0 1 0 0 1 0 0 1 0 1 0 0 0 0 0 0 0 1 1 1 0 0 0 1 1 0 1 1 1 0 1 0 0 0 8 1 0 0 1 0 1 0 0 1 0 0 1 0 0 1 0 0 0 0 0 1 1 0 0 0 0 0 0 1 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 9 1 0 0 1 1 1 0 0 1 0 0 0 1 0 1 0 0 0 0 0 0 1 0 0 0 0 0 0 1 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 10 1 0 0 1 1 1 0 0 1 0 0 1 0 0 1 0 0 0 0 0 1 1 0 0 0 0 0 0 1 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 11 1 0 0 1 1 1 0 0 1 0 0 1 0 0 1 0 0 0 0 0 0 0 0 1 0 0 0 0 1 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 12 1 0 0 0 1 1 0 0 1 0 0 0 1 0 1 0 0 0 0 0 0 0 1 0 0 0 0 0 1 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 13 1 0 0 1 1 1 0 0 1 0 0 1 0 0 1 0 0 0 0 0 1 0 1 0 0 0 1 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 14 1 0 0 1 1 1 0 1 0 0 0 0 1 1 0 0 0 0 0 0 0 1 1 0 0 0 1 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 15 1 1 0 1 1 1 0 1 0 0 0 0 1 0 0 1 0 0 0 1 1 1 0 0 0 0 1 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 16 1 1 0 1 1 1 0 1 0 0 0 0 1 1 0 0 0 0 0 0 0 1 1 0 0 0 1 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 17 1 0 0 1 1 1 0 1 0 0 0 0 1 1 0 0 0 0 0 0 0 1 0 0 0 0 0 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 18 1 0 0 1 1 1 0 0 1 0 0 0 1 0 1 0 0 1 1 0 0 0 0 0 1 0 0 0 1 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 19 1 0 0 0 1 1 0 0 1 0 0 0 1 0 1 0 0 1 0 0 0 0 0 0 1 0 0 0 1 1 1 0 0 0 1 1 0 1 1 1 0 1 0 0 1 20 1 0 0 0 1 1 0 0 1 0 0 0 1 0 1 0 0 1 0 0 0 0 0 1 1 0 0 0 1 1 1 0 0 0 1 1 0 1 1 1 0 1 0 0 1 21 1 0 0 1 1 1 0 1 0 0 0 1 0 0 1 0 0 0 0 0 1 1 0 0 0 0 1 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 0 1 22 1 0 0 1 1 1 0 1 0 0 0 1 0 0 1 0 0 0 0 0 1 1 0 0 0 0 0 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 0 1 23 1 0 0 1 1 1 0 1 0 0 0 1 0 0 1 0 0 0 0 0 1 1 0 0 0 0 0 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 0 1 24 1 0 0 1 1 1 0 0 1 0 0 1 0 0 1 0 0 0 0 0 0 1 0 0 0 0 0 0 1 1 1 0 0 0 1 1 0 1 1 1 0 1 0 0 1 25 1 0 0 1 1 1 0 1 0 0 0 0 1 0 1 0 0 0 0 0 0 1 0 0 0 0 0 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 26 1 1 0 1 1 1 0 1 0 0 0 0 1 0 1 0 0 1 0 0 0 0 1 0 1 0 0 1 0 1 0 1 0 0 1 1 0 1 1 1 0 1 0 0 1 27 1 1 0 1 1 1 0 0 1 0 0 0 1 0 1 0 0 1 0 0 0 1 0 0 1 0 0 1 0 1 0 1 0 0 1 1 0 1 1 1 0 1 0 0 1 28 1 1 0 1 1 1 0 1 0 0 0 0 1 0 1 0 0 1 0 0 0 1 0 0 1 0 0 1 0 1 0 1 0 0 1 1 0 1 1 1 0 1 0 0 1 29 1 0 0 1 1 1 0 0 1 0 0 0 1 0 1 0 0 0 0 1 0 1 0 0 0 0 0 0 1 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 30 1 0 0 1 0 1 0 0 1 0 0 1 0 0 1 0 0 0 0 0 0 1 0 0 1 0 1 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 31 1 0 0 1 0 1 0 0 1 0 0 1 0 0 1 0 0 0 0 0 0 1 0 0 1 0 0 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 32 1 0 0 ? ? 1 0 0 1 0 0 1 0 0 1 0 0 1 0 0 0 1 0 0 0 0 0 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 ? ? 33 1 0 0 1 1 1 0 0 1 0 0 1 0 0 1 0 0 1 0 0 0 1 0 0 1 0 0 1 0 1 1 0 0 0 1 ? ? 1 1 1 0 1 0 0 0 34 1 0 0 1 1 1 0 0 1 0 0 1 0 0 1 0 0 1 0 0 0 0 0 0 1 0 0 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 0 0 35 1 0 0 1 1 1 0 0 1 0 0 1 0 0 1 0 0 1 0 0 0 0 0 0 1 0 0 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 0 0 36 1 0 0 1 0 1 0 1 0 0 0 1 0 0 1 0 0 0 0 0 0 0 1 0 0 0 0 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 0 0 37 1 0 1 1 0 1 0 1 0 0 0 1 0 0 1 0 0 0 0 0 0 1 0 0 0 0 0 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 38 1 0 0 1 1 1 0 1 0 0 0 0 1 0 1 0 0 1 0 1 0 1 1 0 0 0 0 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 0 0 39 1 0 0 1 1 1 0 1 0 0 0 0 1 0 1 0 0 0 0 1 0 1 1 0 0 0 0 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 0 0 40 1 0 0 1 1 1 0 1 0 0 0 1 0 0 1 0 0 1 0 0 0 1 0 0 0 0 0 0 1 1 1 0 0 0 1 1 0 1 1 1 0 1 0 0 1 41 1 0 0 1 0 1 0 1 0 0 0 1 0 0 1 0 0 1 0 0 0 1 0 0 0 0 0 0 1 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 42 1 0 0 1 0 1 0 1 0 0 0 1 0 0 1 0 0 1 0 0 0 1 0 0 0 0 0 0 1 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 43 1 0 0 1 0 1 0 1 0 0 0 1 0 0 1 0 0 1 0 0 0 1 0 0 0 0 0 0 1 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 44 1 0 0 1 0 1 0 1 0 0 0 1 0 0 1 0 0 1 0 0 0 1 0 0 0 0 0 0 1 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 45 1 0 0 1 1 1 0 0 1 0 0 0 1 1 0 0 0 0 0 0 0 1 1 1 0 0 0 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 46 1 0 0 1 1 1 0 0 1 0 0 0 1 1 0 0 0 0 0 0 0 1 1 1 0 0 0 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 47 1 0 0 1 1 1 0 0 1 0 0 0 1 0 1 0 0 0 0 0 0 0 1 1 0 0 0 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 48 1 0 0 1 1 1 0 0 1 0 0 0 1 1 0 0 0 0 0 0 0 0 0 1 0 0 0 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 49 1 0 0 1 1 1 0 0 1 0 0 0 1 0 1 0 0 0 0 0 0 1 0 0 0 0 1 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 50 1 1 0 1 1 1 0 1 0 0 0 0 1 0 1 0 0 0 0 0 0 1 0 1 0 0 1 0 1 1 1 0 0 0 1 1 0 1 1 1 0 1 0 0 1 51 1 0 0 ? ? 1 0 1 0 0 0 ? ? ? ? ? 0 1 0 0 0 1 1 0 0 0 0 1 0 1 1 0 0 0 1 ? ? 1 1 1 0 1 0 ? ? 52 1 0 0 1 0 1 0 0 1 0 0 0 1 1 0 0 0 0 0 0 0 1 0 0 0 0 1 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 53 1 0 0 1 0 1 0 0 1 0 0 0 1 1 0 0 0 0 0 0 0 1 0 0 0 0 1 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 54 1 0 0 1 0 1 0 0 1 0 0 0 1 1 0 0 0 0 0 0 0 1 0 0 0 0 1 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 55 1 0 0 1 1 1 0 1 0 0 0 0 1 0 1 0 0 0 0 0 0 1 1 0 0 0 0 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 56 1 0 0 1 1 1 0 1 0 0 0 0 1 0 1 0 0 0 0 0 0 1 1 0 0 0 0 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 57 1 0 0 1 1 1 0 1 0 0 0 0 1 0 1 0 0 0 0 0 0 1 1 0 0 0 0 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 58 1 0 0 1 1 1 0 1 0 0 0 0 1 0 1 0 0 0 0 0 0 1 1 0 0 0 0 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 59 1 0 0 1 1 1 0 1 0 0 0 0 1 0 1 0 0 0 0 0 0 1 1 0 0 0 0 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 60 1 0 0 1 1 1 0 1 1 0 0 1 1 1 1 0 0 0 0 1 0 1 0 0 0 0 1 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 61 1 0 0 ? 0 1 0 0 1 0 0 0 1 1 0 0 0 1 0 0 0 1 0 0 0 0 0 1 0 1 1 0 0 1 0 1 0 1 1 1 0 1 0 0 0 62 1 0 0 ? 0 1 0 0 1 0 0 0 1 1 0 0 0 1 0 0 0 1 0 0 0 0 0 1 0 1 1 0 0 1 0 1 0 1 1 1 0 1 0 0 0 63 1 0 0 ? 0 1 0 0 1 0 0 0 1 1 0 0 0 1 0 0 0 1 0 0 0 0 0 1 0 1 1 0 0 1 0 1 0 1 1 1 0 1 0 0 0 64 1 0 0 1 1 1 0 1 0 0 0 0 1 1 0 0 0 1 0 0 0 1 1 0 0 0 0 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 65 1 0 1 1 1 1 0 0 1 0 0 0 1 1 1 0 0 0 0 0 0 1 1 0 0 0 1 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 0 0 66 1 0 1 1 1 1 0 0 1 0 0 0 1 1 1 0 0 0 0 0 0 1 1 0 0 0 1 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 0 0 67 1 0 1 1 1 1 0 0 1 0 0 0 1 1 1 0 0 0 0 0 0 1 1 0 0 0 1 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 0 0 68 1 0 1 1 1 1 0 0 1 0 0 0 1 0 1 0 0 0 0 0 0 1 1 0 0 0 0 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 0 0 69 1 1 0 1 1 1 0 0 1 0 0 0 1 ? ? ? 0 0 0 0 0 1 1 0 0 0 0 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 70 1 1 0 1 1 1 0 0 1 0 0 0 1 0 1 0 0 0 0 0 0 1 1 0 0 0 0 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 71 1 1 0 1 1 1 0 0 1 0 0 0 1 0 1 0 0 0 0 0 0 1 1 0 0 0 0 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 72 1 0 0 1 1 1 0 0 1 0 0 0 1 0 1 0 0 ? 0 0 0 0 0 0 1 0 0 0 1 1 1 0 0 0 1 1 0 1 1 1 0 1 0 0 0 73 1 0 0 1 1 1 0 0 1 0 0 0 1 0 1 0 0 ? 0 0 0 0 0 0 1 0 0 0 1 1 1 0 0 0 1 1 0 1 1 1 0 1 0 0 0 74 1 0 0 1 1 1 0 0 1 0 0 1 0 0 1 0 0 0 0 0 0 1 1 0 0 0 0 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 75 1 0 1 1 0 1 0 1 0 0 0 0 1 0 1 0 0 1 0 0 0 1 1 0 0 0 0 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 76 1 1 0 1 0 1 0 1 0 0 0 0 1 0 1 1 0 0 0 0 0 1 0 0 0 0 0 1 0 1 0 0 0 1 0 1 0 1 1 1 0 1 0 0 1 77 1 0 0 1 0 1 0 0 1 0 0 0 1 0 1 0 0 0 0 1 0 1 0 0 0 0 1 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 78 1 0 0 1 0 1 0 0 1 0 0 0 1 0 1 0 0 0 0 1 0 1 0 0 0 0 1 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 79 1 0 0 1 0 1 0 0 1 0 0 0 1 0 1 0 0 0 0 1 0 1 0 0 0 0 1 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 80 1 0 1 1 1 1 0 1 0 0 0 1 0 0 1 0 0 0 0 0 0 1 0 0 0 0 0 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 0 0 81 1 0 1 1 1 1 0 0 1 0 0 1 0 0 1 0 0 0 0 0 0 1 1 0 0 0 0 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 0 0 82 1 0 0 1 1 ? 0 0 1 0 0 ? ? 0 1 0 1 0 0 0 0 0 0 0 0 0 0 1 0 1 1 0 0 1 0 1 0 1 1 1 0 1 0 1 0 83 1 0 0 1 1 ? 0 0 1 0 0 ? ? 0 1 0 1 0 0 0 0 0 0 0 0 0 0 1 0 1 1 0 0 1 0 ? ? 1 1 1 0 1 0 0 0 84 1 0 1 1 0 1 0 0 1 0 0 1 0 0 1 0 0 0 0 0 0 1 0 0 0 0 1 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 0 0 85 1 0 1 1 1 1 0 1 0 0 0 1 0 0 1 0 0 0 0 0 0 1 0 0 0 0 0 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 0 0 86 1 0 1 1 1 1 0 1 0 0 0 1 0 0 1 0 0 0 0 0 0 1 0 0 0 0 0 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 0 0 87 1 0 1 1 1 1 0 1 0 0 0 0 1 1 0 0 0 0 1 0 0 1 0 0 0 0 1 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 0 0 88 1 0 1 1 ? 1 0 0 1 0 0 0 1 1 0 0 0 0 0 0 0 1 0 0 0 0 1 1 ? 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 89 1 0 1 1 1 1 0 1 0 0 0 0 1 1 0 0 0 0 1 0 0 1 0 0 0 0 0 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 90 1 0 1 1 1 1 0 1 0 0 0 0 1 1 0 0 0 0 1 0 0 1 0 0 0 0 0 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 91 1 0 1 0 1 1 0 1 0 0 0 1 0 0 1 0 0 0 0 0 0 1 0 0 0 0 0 0 1 1 1 0 0 0 1 1 0 1 1 1 0 1 0 0 0 92 ? 0 1 0 0 1 0 1 0 0 0 0 1 1 0 0 0 0 0 0 0 0 1 0 0 0 0 1 0 1 ? ? ? ? ? 1 0 1 1 1 0 0 1 1 0 93 0 0 ? 0 1 1 0 1 0 0 0 1 0 0 1 0 0 0 0 0 0 0 0 1 0 0 0 0 1 1 ? ? ? ? ? 1 0 1 1 ? ? ? ? 0 0 94 1 0 1 0 1 1 0 0 1 0 0 1 0 0 1 0 0 0 0 0 0 1 0 0 0 0 0 1 0 1 1 0 0 0 1 1 0 1 1 0 1 0 1 1 0 95 1 0 1 0 1 1 0 0 1 0 0 1 0 0 1 0 0 0 0 0 0 1 0 0 0 0 0 1 0 1 1 0 0 0 1 1 0 1 1 0 1 0 1 1 0 96 ? 0 1 0 1 1 0 0 1 0 0 1 0 0 1 0 0 0 0 0 0 1 0 0 0 0 0 1 0 1 ? ? ? ? ? 1 0 1 1 0 1 0 1 1 0 97 1 0 1 0 1 1 0 0 1 0 0 1 0 0 1 0 0 0 0 0 0 1 0 0 0 0 0 1 0 1 1 0 0 0 1 1 0 1 1 0 1 0 1 1 0 98 1 0 1 0 1 1 0 0 1 0 0 1 0 0 1 0 0 0 0 0 0 1 0 0 0 0 0 0 1 1 1 0 0 0 1 1 0 1 1 1 0 1 0 0 0 99 1 0 0 0 1 1 0 1 0 0 0 0 1 1 0 0 1 0 0 0 0 1 0 0 0 0 0 1 0 1 0 0 0 0 1 1 0 1 1 1 0 1 0 0 0 100 1 0 1 0 0 1 0 1 0 0 0 0 1 0 0 1 0 0 0 0 0 1 0 0 0 0 0 1 0 1 0 0 0 0 1 1 0 1 1 0 1 1 0 0 0 101 1 0 1 0 1 1 0 1 0 0 0 0 1 1 0 0 1 0 0 0 0 1 0 0 0 0 0 1 0 1 0 0 0 0 1 1 0 1 1 0 1 1 0 0 0 102 1 0 0 0 0 1 0 0 1 1 1 1 0 1 1 0 0 0 0 0 0 0 1 1 0 0 0 1 0 1 1 0 0 0 1 ? ? 1 1 1 0 1 0 1 0 103 1 0 0 0 0 1 0 0 1 1 1 1 0 1 1 0 0 0 0 0 0 0 1 1 1 0 0 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 104 1 1 0 1 1 1 0 0 1 0 0 0 1 0 0 1 1 0 0 0 0 0 0 0 0 0 0 1 0 1 0 0 1 0 1 1 0 1 1 1 0 1 0 1 0 Specimen Anatomical characters 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 1 1 0 0 1 0 1 0 0 1 0 0 0 1 1 0 0 0 0 0 0 1 0 0 0 1 0 1 0 1 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 2 1 0 0 1 0 1 0 0 1 0 0 0 1 0 1 0 0 0 0 0 1 0 0 0 1 0 1 0 1 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 3 1 0 0 1 0 1 0 0 1 0 0 0 1 1 0 0 0 0 0 0 1 0 0 0 1 0 1 0 1 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 4 1 0 0 1 0 1 0 0 1 0 0 0 1 1 0 0 0 0 0 0 1 0 0 0 1 0 1 0 1 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 5 1 1 1 1 0 1 0 1 1 0 0 0 1 0 1 0 0 0 0 0 0 1 0 0 0 0 1 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 6 1 0 0 1 1 1 0 1 0 0 0 0 1 1 0 0 ? ? ? ? ? ? ? ? ? ? ? 0 1 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 7 1 0 1 1 1 1 0 0 1 0 0 0 1 0 0 1 0 0 1 0 1 0 0 0 0 0 0 0 1 1 1 0 0 0 1 1 0 1 1 1 0 1 0 0 0 8 1 0 0 1 0 1 0 0 1 0 0 1 0 0 1 0 0 0 0 0 1 1 0 0 0 0 0 0 1 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 9 1 0 0 1 1 1 0 0 1 0 0 0 1 0 1 0 0 0 0 0 0 1 0 0 0 0 0 0 1 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 10 1 0 0 1 1 1 0 0 1 0 0 1 0 0 1 0 0 0 0 0 1 1 0 0 0 0 0 0 1 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 11 1 0 0 1 1 1 0 0 1 0 0 1 0 0 1 0 0 0 0 0 0 0 0 1 0 0 0 0 1 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 12 1 0 0 0 1 1 0 0 1 0 0 0 1 0 1 0 0 0 0 0 0 0 1 0 0 0 0 0 1 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 13 1 0 0 1 1 1 0 0 1 0 0 1 0 0 1 0 0 0 0 0 1 0 1 0 0 0 1 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 14 1 0 0 1 1 1 0 1 0 0 0 0 1 1 0 0 0 0 0 0 0 1 1 0 0 0 1 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 15 1 1 0 1 1 1 0 1 0 0 0 0 1 0 0 1 0 0 0 1 1 1 0 0 0 0 1 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 16 1 1 0 1 1 1 0 1 0 0 0 0 1 1 0 0 0 0 0 0 0 1 1 0 0 0 1 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 17 1 0 0 1 1 1 0 1 0 0 0 0 1 1 0 0 0 0 0 0 0 1 0 0 0 0 0 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 18 1 0 0 1 1 1 0 0 1 0 0 0 1 0 1 0 0 1 1 0 0 0 0 0 1 0 0 0 1 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 19 1 0 0 0 1 1 0 0 1 0 0 0 1 0 1 0 0 1 0 0 0 0 0 0 1 0 0 0 1 1 1 0 0 0 1 1 0 1 1 1 0 1 0 0 1 20 1 0 0 0 1 1 0 0 1 0 0 0 1 0 1 0 0 1 0 0 0 0 0 1 1 0 0 0 1 1 1 0 0 0 1 1 0 1 1 1 0 1 0 0 1 21 1 0 0 1 1 1 0 1 0 0 0 1 0 0 1 0 0 0 0 0 1 1 0 0 0 0 1 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 0 1 22 1 0 0 1 1 1 0 1 0 0 0 1 0 0 1 0 0 0 0 0 1 1 0 0 0 0 0 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 0 1 23 1 0 0 1 1 1 0 1 0 0 0 1 0 0 1 0 0 0 0 0 1 1 0 0 0 0 0 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 0 1 24 1 0 0 1 1 1 0 0 1 0 0 1 0 0 1 0 0 0 0 0 0 1 0 0 0 0 0 0 1 1 1 0 0 0 1 1 0 1 1 1 0 1 0 0 1 25 1 0 0 1 1 1 0 1 0 0 0 0 1 0 1 0 0 0 0 0 0 1 0 0 0 0 0 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 26 1 1 0 1 1 1 0 1 0 0 0 0 1 0 1 0 0 1 0 0 0 0 1 0 1 0 0 1 0 1 0 1 0 0 1 1 0 1 1 1 0 1 0 0 1 27 1 1 0 1 1 1 0 0 1 0 0 0 1 0 1 0 0 1 0 0 0 1 0 0 1 0 0 1 0 1 0 1 0 0 1 1 0 1 1 1 0 1 0 0 1 28 1 1 0 1 1 1 0 1 0 0 0 0 1 0 1 0 0 1 0 0 0 1 0 0 1 0 0 1 0 1 0 1 0 0 1 1 0 1 1 1 0 1 0 0 1 29 1 0 0 1 1 1 0 0 1 0 0 0 1 0 1 0 0 0 0 1 0 1 0 0 0 0 0 0 1 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 30 1 0 0 1 0 1 0 0 1 0 0 1 0 0 1 0 0 0 0 0 0 1 0 0 1 0 1 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 31 1 0 0 1 0 1 0 0 1 0 0 1 0 0 1 0 0 0 0 0 0 1 0 0 1 0 0 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 32 1 0 0 ? ? 1 0 0 1 0 0 1 0 0 1 0 0 1 0 0 0 1 0 0 0 0 0 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 ? ? 33 1 0 0 1 1 1 0 0 1 0 0 1 0 0 1 0 0 1 0 0 0 1 0 0 1 0 0 1 0 1 1 0 0 0 1 ? ? 1 1 1 0 1 0 0 0 34 1 0 0 1 1 1 0 0 1 0 0 1 0 0 1 0 0 1 0 0 0 0 0 0 1 0 0 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 0 0 35 1 0 0 1 1 1 0 0 1 0 0 1 0 0 1 0 0 1 0 0 0 0 0 0 1 0 0 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 0 0 36 1 0 0 1 0 1 0 1 0 0 0 1 0 0 1 0 0 0 0 0 0 0 1 0 0 0 0 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 0 0 37 1 0 1 1 0 1 0 1 0 0 0 1 0 0 1 0 0 0 0 0 0 1 0 0 0 0 0 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 38 1 0 0 1 1 1 0 1 0 0 0 0 1 0 1 0 0 1 0 1 0 1 1 0 0 0 0 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 0 0 39 1 0 0 1 1 1 0 1 0 0 0 0 1 0 1 0 0 0 0 1 0 1 1 0 0 0 0 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 0 0 40 1 0 0 1 1 1 0 1 0 0 0 1 0 0 1 0 0 1 0 0 0 1 0 0 0 0 0 0 1 1 1 0 0 0 1 1 0 1 1 1 0 1 0 0 1 41 1 0 0 1 0 1 0 1 0 0 0 1 0 0 1 0 0 1 0 0 0 1 0 0 0 0 0 0 1 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 42 1 0 0 1 0 1 0 1 0 0 0 1 0 0 1 0 0 1 0 0 0 1 0 0 0 0 0 0 1 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 43 1 0 0 1 0 1 0 1 0 0 0 1 0 0 1 0 0 1 0 0 0 1 0 0 0 0 0 0 1 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 44 1 0 0 1 0 1 0 1 0 0 0 1 0 0 1 0 0 1 0 0 0 1 0 0 0 0 0 0 1 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 45 1 0 0 1 1 1 0 0 1 0 0 0 1 1 0 0 0 0 0 0 0 1 1 1 0 0 0 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 46 1 0 0 1 1 1 0 0 1 0 0 0 1 1 0 0 0 0 0 0 0 1 1 1 0 0 0 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 47 1 0 0 1 1 1 0 0 1 0 0 0 1 0 1 0 0 0 0 0 0 0 1 1 0 0 0 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 48 1 0 0 1 1 1 0 0 1 0 0 0 1 1 0 0 0 0 0 0 0 0 0 1 0 0 0 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 49 1 0 0 1 1 1 0 0 1 0 0 0 1 0 1 0 0 0 0 0 0 1 0 0 0 0 1 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 50 1 1 0 1 1 1 0 1 0 0 0 0 1 0 1 0 0 0 0 0 0 1 0 1 0 0 1 0 1 1 1 0 0 0 1 1 0 1 1 1 0 1 0 0 1 51 1 0 0 ? ? 1 0 1 0 0 0 ? ? ? ? ? 0 1 0 0 0 1 1 0 0 0 0 1 0 1 1 0 0 0 1 ? ? 1 1 1 0 1 0 ? ? 52 1 0 0 1 0 1 0 0 1 0 0 0 1 1 0 0 0 0 0 0 0 1 0 0 0 0 1 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 53 1 0 0 1 0 1 0 0 1 0 0 0 1 1 0 0 0 0 0 0 0 1 0 0 0 0 1 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 54 1 0 0 1 0 1 0 0 1 0 0 0 1 1 0 0 0 0 0 0 0 1 0 0 0 0 1 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 55 1 0 0 1 1 1 0 1 0 0 0 0 1 0 1 0 0 0 0 0 0 1 1 0 0 0 0 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 56 1 0 0 1 1 1 0 1 0 0 0 0 1 0 1 0 0 0 0 0 0 1 1 0 0 0 0 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 57 1 0 0 1 1 1 0 1 0 0 0 0 1 0 1 0 0 0 0 0 0 1 1 0 0 0 0 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 58 1 0 0 1 1 1 0 1 0 0 0 0 1 0 1 0 0 0 0 0 0 1 1 0 0 0 0 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 59 1 0 0 1 1 1 0 1 0 0 0 0 1 0 1 0 0 0 0 0 0 1 1 0 0 0 0 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 60 1 0 0 1 1 1 0 1 1 0 0 1 1 1 1 0 0 0 0 1 0 1 0 0 0 0 1 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 61 1 0 0 ? 0 1 0 0 1 0 0 0 1 1 0 0 0 1 0 0 0 1 0 0 0 0 0 1 0 1 1 0 0 1 0 1 0 1 1 1 0 1 0 0 0 62 1 0 0 ? 0 1 0 0 1 0 0 0 1 1 0 0 0 1 0 0 0 1 0 0 0 0 0 1 0 1 1 0 0 1 0 1 0 1 1 1 0 1 0 0 0 63 1 0 0 ? 0 1 0 0 1 0 0 0 1 1 0 0 0 1 0 0 0 1 0 0 0 0 0 1 0 1 1 0 0 1 0 1 0 1 1 1 0 1 0 0 0 64 1 0 0 1 1 1 0 1 0 0 0 0 1 1 0 0 0 1 0 0 0 1 1 0 0 0 0 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 65 1 0 1 1 1 1 0 0 1 0 0 0 1 1 1 0 0 0 0 0 0 1 1 0 0 0 1 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 0 0 66 1 0 1 1 1 1 0 0 1 0 0 0 1 1 1 0 0 0 0 0 0 1 1 0 0 0 1 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 0 0 67 1 0 1 1 1 1 0 0 1 0 0 0 1 1 1 0 0 0 0 0 0 1 1 0 0 0 1 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 0 0 68 1 0 1 1 1 1 0 0 1 0 0 0 1 0 1 0 0 0 0 0 0 1 1 0 0 0 0 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 0 0 69 1 1 0 1 1 1 0 0 1 0 0 0 1 ? ? ? 0 0 0 0 0 1 1 0 0 0 0 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 70 1 1 0 1 1 1 0 0 1 0 0 0 1 0 1 0 0 0 0 0 0 1 1 0 0 0 0 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 71 1 1 0 1 1 1 0 0 1 0 0 0 1 0 1 0 0 0 0 0 0 1 1 0 0 0 0 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 72 1 0 0 1 1 1 0 0 1 0 0 0 1 0 1 0 0 ? 0 0 0 0 0 0 1 0 0 0 1 1 1 0 0 0 1 1 0 1 1 1 0 1 0 0 0 73 1 0 0 1 1 1 0 0 1 0 0 0 1 0 1 0 0 ? 0 0 0 0 0 0 1 0 0 0 1 1 1 0 0 0 1 1 0 1 1 1 0 1 0 0 0 74 1 0 0 1 1 1 0 0 1 0 0 1 0 0 1 0 0 0 0 0 0 1 1 0 0 0 0 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 75 1 0 1 1 0 1 0 1 0 0 0 0 1 0 1 0 0 1 0 0 0 1 1 0 0 0 0 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 76 1 1 0 1 0 1 0 1 0 0 0 0 1 0 1 1 0 0 0 0 0 1 0 0 0 0 0 1 0 1 0 0 0 1 0 1 0 1 1 1 0 1 0 0 1 77 1 0 0 1 0 1 0 0 1 0 0 0 1 0 1 0 0 0 0 1 0 1 0 0 0 0 1 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 78 1 0 0 1 0 1 0 0 1 0 0 0 1 0 1 0 0 0 0 1 0 1 0 0 0 0 1 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 79 1 0 0 1 0 1 0 0 1 0 0 0 1 0 1 0 0 0 0 1 0 1 0 0 0 0 1 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 80 1 0 1 1 1 1 0 1 0 0 0 1 0 0 1 0 0 0 0 0 0 1 0 0 0 0 0 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 0 0 81 1 0 1 1 1 1 0 0 1 0 0 1 0 0 1 0 0 0 0 0 0 1 1 0 0 0 0 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 0 0 82 1 0 0 1 1 ? 0 0 1 0 0 ? ? 0 1 0 1 0 0 0 0 0 0 0 0 0 0 1 0 1 1 0 0 1 0 1 0 1 1 1 0 1 0 1 0 83 1 0 0 1 1 ? 0 0 1 0 0 ? ? 0 1 0 1 0 0 0 0 0 0 0 0 0 0 1 0 1 1 0 0 1 0 ? ? 1 1 1 0 1 0 0 0 84 1 0 1 1 0 1 0 0 1 0 0 1 0 0 1 0 0 0 0 0 0 1 0 0 0 0 1 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 0 0 85 1 0 1 1 1 1 0 1 0 0 0 1 0 0 1 0 0 0 0 0 0 1 0 0 0 0 0 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 0 0 86 1 0 1 1 1 1 0 1 0 0 0 1 0 0 1 0 0 0 0 0 0 1 0 0 0 0 0 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 0 0 87 1 0 1 1 1 1 0 1 0 0 0 0 1 1 0 0 0 0 1 0 0 1 0 0 0 0 1 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 0 0 88 1 0 1 1 ? 1 0 0 1 0 0 0 1 1 0 0 0 0 0 0 0 1 0 0 0 0 1 1 ? 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 89 1 0 1 1 1 1 0 1 0 0 0 0 1 1 0 0 0 0 1 0 0 1 0 0 0 0 0 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 90 1 0 1 1 1 1 0 1 0 0 0 0 1 1 0 0 0 0 1 0 0 1 0 0 0 0 0 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 91 1 0 1 0 1 1 0 1 0 0 0 1 0 0 1 0 0 0 0 0 0 1 0 0 0 0 0 0 1 1 1 0 0 0 1 1 0 1 1 1 0 1 0 0 0 92 ? 0 1 0 0 1 0 1 0 0 0 0 1 1 0 0 0 0 0 0 0 0 1 0 0 0 0 1 0 1 ? ? ? ? ? 1 0 1 1 1 0 0 1 1 0 93 0 0 ? 0 1 1 0 1 0 0 0 1 0 0 1 0 0 0 0 0 0 0 0 1 0 0 0 0 1 1 ? ? ? ? ? 1 0 1 1 ? ? ? ? 0 0 94 1 0 1 0 1 1 0 0 1 0 0 1 0 0 1 0 0 0 0 0 0 1 0 0 0 0 0 1 0 1 1 0 0 0 1 1 0 1 1 0 1 0 1 1 0 95 1 0 1 0 1 1 0 0 1 0 0 1 0 0 1 0 0 0 0 0 0 1 0 0 0 0 0 1 0 1 1 0 0 0 1 1 0 1 1 0 1 0 1 1 0 96 ? 0 1 0 1 1 0 0 1 0 0 1 0 0 1 0 0 0 0 0 0 1 0 0 0 0 0 1 0 1 ? ? ? ? ? 1 0 1 1 0 1 0 1 1 0 97 1 0 1 0 1 1 0 0 1 0 0 1 0 0 1 0 0 0 0 0 0 1 0 0 0 0 0 1 0 1 1 0 0 0 1 1 0 1 1 0 1 0 1 1 0 98 1 0 1 0 1 1 0 0 1 0 0 1 0 0 1 0 0 0 0 0 0 1 0 0 0 0 0 0 1 1 1 0 0 0 1 1 0 1 1 1 0 1 0 0 0 99 1 0 0 0 1 1 0 1 0 0 0 0 1 1 0 0 1 0 0 0 0 1 0 0 0 0 0 1 0 1 0 0 0 0 1 1 0 1 1 1 0 1 0 0 0 100 1 0 1 0 0 1 0 1 0 0 0 0 1 0 0 1 0 0 0 0 0 1 0 0 0 0 0 1 0 1 0 0 0 0 1 1 0 1 1 0 1 1 0 0 0 101 1 0 1 0 1 1 0 1 0 0 0 0 1 1 0 0 1 0 0 0 0 1 0 0 0 0 0 1 0 1 0 0 0 0 1 1 0 1 1 0 1 1 0 0 0 102 1 0 0 0 0 1 0 0 1 1 1 1 0 1 1 0 0 0 0 0 0 0 1 1 0 0 0 1 0 1 1 0 0 0 1 ? ? 1 1 1 0 1 0 1 0 103 1 0 0 0 0 1 0 0 1 1 1 1 0 1 1 0 0 0 0 0 0 0 1 1 1 0 0 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 104 1 1 0 1 1 1 0 0 1 0 0 0 1 0 0 1 1 0 0 0 0 0 0 0 0 0 0 1 0 1 0 0 1 0 1 1 0 1 1 1 0 1 0 1 0 Appendix 2. Binary matrix with 104 specimens and 45 characters. (0) absent; (1) present; (?) unknown; (-) inapplicable. Specimen Anatomical characters 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 1 1 0 0 1 0 1 0 0 1 0 0 0 1 1 0 0 0 0 0 0 1 0 0 0 1 0 1 0 1 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 2 1 0 0 1 0 1 0 0 1 0 0 0 1 0 1 0 0 0 0 0 1 0 0 0 1 0 1 0 1 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 3 1 0 0 1 0 1 0 0 1 0 0 0 1 1 0 0 0 0 0 0 1 0 0 0 1 0 1 0 1 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 4 1 0 0 1 0 1 0 0 1 0 0 0 1 1 0 0 0 0 0 0 1 0 0 0 1 0 1 0 1 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 5 1 1 1 1 0 1 0 1 1 0 0 0 1 0 1 0 0 0 0 0 0 1 0 0 0 0 1 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 6 1 0 0 1 1 1 0 1 0 0 0 0 1 1 0 0 ? ? ? ? ? ? ? ? ? ? ? 0 1 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 7 1 0 1 1 1 1 0 0 1 0 0 0 1 0 0 1 0 0 1 0 1 0 0 0 0 0 0 0 1 1 1 0 0 0 1 1 0 1 1 1 0 1 0 0 0 8 1 0 0 1 0 1 0 0 1 0 0 1 0 0 1 0 0 0 0 0 1 1 0 0 0 0 0 0 1 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 9 1 0 0 1 1 1 0 0 1 0 0 0 1 0 1 0 0 0 0 0 0 1 0 0 0 0 0 0 1 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 10 1 0 0 1 1 1 0 0 1 0 0 1 0 0 1 0 0 0 0 0 1 1 0 0 0 0 0 0 1 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 11 1 0 0 1 1 1 0 0 1 0 0 1 0 0 1 0 0 0 0 0 0 0 0 1 0 0 0 0 1 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 12 1 0 0 0 1 1 0 0 1 0 0 0 1 0 1 0 0 0 0 0 0 0 1 0 0 0 0 0 1 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 13 1 0 0 1 1 1 0 0 1 0 0 1 0 0 1 0 0 0 0 0 1 0 1 0 0 0 1 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 14 1 0 0 1 1 1 0 1 0 0 0 0 1 1 0 0 0 0 0 0 0 1 1 0 0 0 1 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 15 1 1 0 1 1 1 0 1 0 0 0 0 1 0 0 1 0 0 0 1 1 1 0 0 0 0 1 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 16 1 1 0 1 1 1 0 1 0 0 0 0 1 1 0 0 0 0 0 0 0 1 1 0 0 0 1 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 17 1 0 0 1 1 1 0 1 0 0 0 0 1 1 0 0 0 0 0 0 0 1 0 0 0 0 0 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 18 1 0 0 1 1 1 0 0 1 0 0 0 1 0 1 0 0 1 1 0 0 0 0 0 1 0 0 0 1 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 19 1 0 0 0 1 1 0 0 1 0 0 0 1 0 1 0 0 1 0 0 0 0 0 0 1 0 0 0 1 1 1 0 0 0 1 1 0 1 1 1 0 1 0 0 1 20 1 0 0 0 1 1 0 0 1 0 0 0 1 0 1 0 0 1 0 0 0 0 0 1 1 0 0 0 1 1 1 0 0 0 1 1 0 1 1 1 0 1 0 0 1 21 1 0 0 1 1 1 0 1 0 0 0 1 0 0 1 0 0 0 0 0 1 1 0 0 0 0 1 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 0 1 22 1 0 0 1 1 1 0 1 0 0 0 1 0 0 1 0 0 0 0 0 1 1 0 0 0 0 0 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 0 1 23 1 0 0 1 1 1 0 1 0 0 0 1 0 0 1 0 0 0 0 0 1 1 0 0 0 0 0 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 0 1 24 1 0 0 1 1 1 0 0 1 0 0 1 0 0 1 0 0 0 0 0 0 1 0 0 0 0 0 0 1 1 1 0 0 0 1 1 0 1 1 1 0 1 0 0 1 25 1 0 0 1 1 1 0 1 0 0 0 0 1 0 1 0 0 0 0 0 0 1 0 0 0 0 0 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 26 1 1 0 1 1 1 0 1 0 0 0 0 1 0 1 0 0 1 0 0 0 0 1 0 1 0 0 1 0 1 0 1 0 0 1 1 0 1 1 1 0 1 0 0 1 27 1 1 0 1 1 1 0 0 1 0 0 0 1 0 1 0 0 1 0 0 0 1 0 0 1 0 0 1 0 1 0 1 0 0 1 1 0 1 1 1 0 1 0 0 1 28 1 1 0 1 1 1 0 1 0 0 0 0 1 0 1 0 0 1 0 0 0 1 0 0 1 0 0 1 0 1 0 1 0 0 1 1 0 1 1 1 0 1 0 0 1 29 1 0 0 1 1 1 0 0 1 0 0 0 1 0 1 0 0 0 0 1 0 1 0 0 0 0 0 0 1 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 30 1 0 0 1 0 1 0 0 1 0 0 1 0 0 1 0 0 0 0 0 0 1 0 0 1 0 1 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 31 1 0 0 1 0 1 0 0 1 0 0 1 0 0 1 0 0 0 0 0 0 1 0 0 1 0 0 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 32 1 0 0 ? ? 1 0 0 1 0 0 1 0 0 1 0 0 1 0 0 0 1 0 0 0 0 0 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 ? ? 33 1 0 0 1 1 1 0 0 1 0 0 1 0 0 1 0 0 1 0 0 0 1 0 0 1 0 0 1 0 1 1 0 0 0 1 ? ? 1 1 1 0 1 0 0 0 34 1 0 0 1 1 1 0 0 1 0 0 1 0 0 1 0 0 1 0 0 0 0 0 0 1 0 0 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 0 0 35 1 0 0 1 1 1 0 0 1 0 0 1 0 0 1 0 0 1 0 0 0 0 0 0 1 0 0 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 0 0 36 1 0 0 1 0 1 0 1 0 0 0 1 0 0 1 0 0 0 0 0 0 0 1 0 0 0 0 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 0 0 37 1 0 1 1 0 1 0 1 0 0 0 1 0 0 1 0 0 0 0 0 0 1 0 0 0 0 0 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 38 1 0 0 1 1 1 0 1 0 0 0 0 1 0 1 0 0 1 0 1 0 1 1 0 0 0 0 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 0 0 39 1 0 0 1 1 1 0 1 0 0 0 0 1 0 1 0 0 0 0 1 0 1 1 0 0 0 0 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 0 0 40 1 0 0 1 1 1 0 1 0 0 0 1 0 0 1 0 0 1 0 0 0 1 0 0 0 0 0 0 1 1 1 0 0 0 1 1 0 1 1 1 0 1 0 0 1 41 1 0 0 1 0 1 0 1 0 0 0 1 0 0 1 0 0 1 0 0 0 1 0 0 0 0 0 0 1 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 42 1 0 0 1 0 1 0 1 0 0 0 1 0 0 1 0 0 1 0 0 0 1 0 0 0 0 0 0 1 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 43 1 0 0 1 0 1 0 1 0 0 0 1 0 0 1 0 0 1 0 0 0 1 0 0 0 0 0 0 1 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 44 1 0 0 1 0 1 0 1 0 0 0 1 0 0 1 0 0 1 0 0 0 1 0 0 0 0 0 0 1 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 45 1 0 0 1 1 1 0 0 1 0 0 0 1 1 0 0 0 0 0 0 0 1 1 1 0 0 0 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 46 1 0 0 1 1 1 0 0 1 0 0 0 1 1 0 0 0 0 0 0 0 1 1 1 0 0 0 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 47 1 0 0 1 1 1 0 0 1 0 0 0 1 0 1 0 0 0 0 0 0 0 1 1 0 0 0 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 48 1 0 0 1 1 1 0 0 1 0 0 0 1 1 0 0 0 0 0 0 0 0 0 1 0 0 0 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 49 1 0 0 1 1 1 0 0 1 0 0 0 1 0 1 0 0 0 0 0 0 1 0 0 0 0 1 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 50 1 1 0 1 1 1 0 1 0 0 0 0 1 0 1 0 0 0 0 0 0 1 0 1 0 0 1 0 1 1 1 0 0 0 1 1 0 1 1 1 0 1 0 0 1 51 1 0 0 ? ? 1 0 1 0 0 0 ? ? ? ? ? 0 1 0 0 0 1 1 0 0 0 0 1 0 1 1 0 0 0 1 ? ? 1 1 1 0 1 0 ? ? 52 1 0 0 1 0 1 0 0 1 0 0 0 1 1 0 0 0 0 0 0 0 1 0 0 0 0 1 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 53 1 0 0 1 0 1 0 0 1 0 0 0 1 1 0 0 0 0 0 0 0 1 0 0 0 0 1 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 54 1 0 0 1 0 1 0 0 1 0 0 0 1 1 0 0 0 0 0 0 0 1 0 0 0 0 1 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 55 1 0 0 1 1 1 0 1 0 0 0 0 1 0 1 0 0 0 0 0 0 1 1 0 0 0 0 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 56 1 0 0 1 1 1 0 1 0 0 0 0 1 0 1 0 0 0 0 0 0 1 1 0 0 0 0 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 57 1 0 0 1 1 1 0 1 0 0 0 0 1 0 1 0 0 0 0 0 0 1 1 0 0 0 0 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 58 1 0 0 1 1 1 0 1 0 0 0 0 1 0 1 0 0 0 0 0 0 1 1 0 0 0 0 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 59 1 0 0 1 1 1 0 1 0 0 0 0 1 0 1 0 0 0 0 0 0 1 1 0 0 0 0 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 60 1 0 0 1 1 1 0 1 1 0 0 1 1 1 1 0 0 0 0 1 0 1 0 0 0 0 1 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 61 1 0 0 ? 0 1 0 0 1 0 0 0 1 1 0 0 0 1 0 0 0 1 0 0 0 0 0 1 0 1 1 0 0 1 0 1 0 1 1 1 0 1 0 0 0 62 1 0 0 ? 0 1 0 0 1 0 0 0 1 1 0 0 0 1 0 0 0 1 0 0 0 0 0 1 0 1 1 0 0 1 0 1 0 1 1 1 0 1 0 0 0 63 1 0 0 ? 0 1 0 0 1 0 0 0 1 1 0 0 0 1 0 0 0 1 0 0 0 0 0 1 0 1 1 0 0 1 0 1 0 1 1 1 0 1 0 0 0 64 1 0 0 1 1 1 0 1 0 0 0 0 1 1 0 0 0 1 0 0 0 1 1 0 0 0 0 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 65 1 0 1 1 1 1 0 0 1 0 0 0 1 1 1 0 0 0 0 0 0 1 1 0 0 0 1 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 0 0 66 1 0 1 1 1 1 0 0 1 0 0 0 1 1 1 0 0 0 0 0 0 1 1 0 0 0 1 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 0 0 67 1 0 1 1 1 1 0 0 1 0 0 0 1 1 1 0 0 0 0 0 0 1 1 0 0 0 1 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 0 0 68 1 0 1 1 1 1 0 0 1 0 0 0 1 0 1 0 0 0 0 0 0 1 1 0 0 0 0 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 0 0 69 1 1 0 1 1 1 0 0 1 0 0 0 1 ? ? ? 0 0 0 0 0 1 1 0 0 0 0 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 70 1 1 0 1 1 1 0 0 1 0 0 0 1 0 1 0 0 0 0 0 0 1 1 0 0 0 0 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 71 1 1 0 1 1 1 0 0 1 0 0 0 1 0 1 0 0 0 0 0 0 1 1 0 0 0 0 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 72 1 0 0 1 1 1 0 0 1 0 0 0 1 0 1 0 0 ? 0 0 0 0 0 0 1 0 0 0 1 1 1 0 0 0 1 1 0 1 1 1 0 1 0 0 0 73 1 0 0 1 1 1 0 0 1 0 0 0 1 0 1 0 0 ? 0 0 0 0 0 0 1 0 0 0 1 1 1 0 0 0 1 1 0 1 1 1 0 1 0 0 0 74 1 0 0 1 1 1 0 0 1 0 0 1 0 0 1 0 0 0 0 0 0 1 1 0 0 0 0 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 75 1 0 1 1 0 1 0 1 0 0 0 0 1 0 1 0 0 1 0 0 0 1 1 0 0 0 0 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 76 1 1 0 1 0 1 0 1 0 0 0 0 1 0 1 1 0 0 0 0 0 1 0 0 0 0 0 1 0 1 0 0 0 1 0 1 0 1 1 1 0 1 0 0 1 77 1 0 0 1 0 1 0 0 1 0 0 0 1 0 1 0 0 0 0 1 0 1 0 0 0 0 1 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 78 1 0 0 1 0 1 0 0 1 0 0 0 1 0 1 0 0 0 0 1 0 1 0 0 0 0 1 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 79 1 0 0 1 0 1 0 0 1 0 0 0 1 0 1 0 0 0 0 1 0 1 0 0 0 0 1 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 80 1 0 1 1 1 1 0 1 0 0 0 1 0 0 1 0 0 0 0 0 0 1 0 0 0 0 0 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 0 0 81 1 0 1 1 1 1 0 0 1 0 0 1 0 0 1 0 0 0 0 0 0 1 1 0 0 0 0 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 0 0 82 1 0 0 1 1 ? 0 0 1 0 0 ? ? 0 1 0 1 0 0 0 0 0 0 0 0 0 0 1 0 1 1 0 0 1 0 1 0 1 1 1 0 1 0 1 0 83 1 0 0 1 1 ? 0 0 1 0 0 ? ? 0 1 0 1 0 0 0 0 0 0 0 0 0 0 1 0 1 1 0 0 1 0 ? ? 1 1 1 0 1 0 0 0 84 1 0 1 1 0 1 0 0 1 0 0 1 0 0 1 0 0 0 0 0 0 1 0 0 0 0 1 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 0 0 85 1 0 1 1 1 1 0 1 0 0 0 1 0 0 1 0 0 0 0 0 0 1 0 0 0 0 0 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 0 0 86 1 0 1 1 1 1 0 1 0 0 0 1 0 0 1 0 0 0 0 0 0 1 0 0 0 0 0 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 0 0 87 1 0 1 1 1 1 0 1 0 0 0 0 1 1 0 0 0 0 1 0 0 1 0 0 0 0 1 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 0 0 88 1 0 1 1 ? 1 0 0 1 0 0 0 1 1 0 0 0 0 0 0 0 1 0 0 0 0 1 1 ? 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 89 1 0 1 1 1 1 0 1 0 0 0 0 1 1 0 0 0 0 1 0 0 1 0 0 0 0 0 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 90 1 0 1 1 1 1 0 1 0 0 0 0 1 1 0 0 0 0 1 0 0 1 0 0 0 0 0 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 91 1 0 1 0 1 1 0 1 0 0 0 1 0 0 1 0 0 0 0 0 0 1 0 0 0 0 0 0 1 1 1 0 0 0 1 1 0 1 1 1 0 1 0 0 0 92 ? 0 1 0 0 1 0 1 0 0 0 0 1 1 0 0 0 0 0 0 0 0 1 0 0 0 0 1 0 1 ? ? ? ? ? 1 0 1 1 1 0 0 1 1 0 93 0 0 ? 0 1 1 0 1 0 0 0 1 0 0 1 0 0 0 0 0 0 0 0 1 0 0 0 0 1 1 ? ? ? ? ? 1 0 1 1 ? ? ? ? 0 0 94 1 0 1 0 1 1 0 0 1 0 0 1 0 0 1 0 0 0 0 0 0 1 0 0 0 0 0 1 0 1 1 0 0 0 1 1 0 1 1 0 1 0 1 1 0 95 1 0 1 0 1 1 0 0 1 0 0 1 0 0 1 0 0 0 0 0 0 1 0 0 0 0 0 1 0 1 1 0 0 0 1 1 0 1 1 0 1 0 1 1 0 96 ? 0 1 0 1 1 0 0 1 0 0 1 0 0 1 0 0 0 0 0 0 1 0 0 0 0 0 1 0 1 ? ? ? ? ? 1 0 1 1 0 1 0 1 1 0 97 1 0 1 0 1 1 0 0 1 0 0 1 0 0 1 0 0 0 0 0 0 1 0 0 0 0 0 1 0 1 1 0 0 0 1 1 0 1 1 0 1 0 1 1 0 98 1 0 1 0 1 1 0 0 1 0 0 1 0 0 1 0 0 0 0 0 0 1 0 0 0 0 0 0 1 1 1 0 0 0 1 1 0 1 1 1 0 1 0 0 0 99 1 0 0 0 1 1 0 1 0 0 0 0 1 1 0 0 1 0 0 0 0 1 0 0 0 0 0 1 0 1 0 0 0 0 1 1 0 1 1 1 0 1 0 0 0 100 1 0 1 0 0 1 0 1 0 0 0 0 1 0 0 1 0 0 0 0 0 1 0 0 0 0 0 1 0 1 0 0 0 0 1 1 0 1 1 0 1 1 0 0 0 101 1 0 1 0 1 1 0 1 0 0 0 0 1 1 0 0 1 0 0 0 0 1 0 0 0 0 0 1 0 1 0 0 0 0 1 1 0 1 1 0 1 1 0 0 0 102 1 0 0 0 0 1 0 0 1 1 1 1 0 1 1 0 0 0 0 0 0 0 1 1 0 0 0 1 0 1 1 0 0 0 1 ? ? 1 1 1 0 1 0 1 0 103 1 0 0 0 0 1 0 0 1 1 1 1 0 1 1 0 0 0 0 0 0 0 1 1 1 0 0 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 104 1 1 0 1 1 1 0 0 1 0 0 0 1 0 0 1 1 0 0 0 0 0 0 0 0 0 0 1 0 1 0 0 1 0 1 1 0 1 1 1 0 1 0 1 0 Specimen Anatomical characters 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 1 1 0 0 1 0 1 0 0 1 0 0 0 1 1 0 0 0 0 0 0 1 0 0 0 1 0 1 0 1 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 2 1 0 0 1 0 1 0 0 1 0 0 0 1 0 1 0 0 0 0 0 1 0 0 0 1 0 1 0 1 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 3 1 0 0 1 0 1 0 0 1 0 0 0 1 1 0 0 0 0 0 0 1 0 0 0 1 0 1 0 1 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 4 1 0 0 1 0 1 0 0 1 0 0 0 1 1 0 0 0 0 0 0 1 0 0 0 1 0 1 0 1 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 5 1 1 1 1 0 1 0 1 1 0 0 0 1 0 1 0 0 0 0 0 0 1 0 0 0 0 1 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 6 1 0 0 1 1 1 0 1 0 0 0 0 1 1 0 0 ? ? ? ? ? ? ? ? ? ? ? 0 1 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 7 1 0 1 1 1 1 0 0 1 0 0 0 1 0 0 1 0 0 1 0 1 0 0 0 0 0 0 0 1 1 1 0 0 0 1 1 0 1 1 1 0 1 0 0 0 8 1 0 0 1 0 1 0 0 1 0 0 1 0 0 1 0 0 0 0 0 1 1 0 0 0 0 0 0 1 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 9 1 0 0 1 1 1 0 0 1 0 0 0 1 0 1 0 0 0 0 0 0 1 0 0 0 0 0 0 1 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 10 1 0 0 1 1 1 0 0 1 0 0 1 0 0 1 0 0 0 0 0 1 1 0 0 0 0 0 0 1 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 11 1 0 0 1 1 1 0 0 1 0 0 1 0 0 1 0 0 0 0 0 0 0 0 1 0 0 0 0 1 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 12 1 0 0 0 1 1 0 0 1 0 0 0 1 0 1 0 0 0 0 0 0 0 1 0 0 0 0 0 1 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 13 1 0 0 1 1 1 0 0 1 0 0 1 0 0 1 0 0 0 0 0 1 0 1 0 0 0 1 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 14 1 0 0 1 1 1 0 1 0 0 0 0 1 1 0 0 0 0 0 0 0 1 1 0 0 0 1 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 15 1 1 0 1 1 1 0 1 0 0 0 0 1 0 0 1 0 0 0 1 1 1 0 0 0 0 1 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 16 1 1 0 1 1 1 0 1 0 0 0 0 1 1 0 0 0 0 0 0 0 1 1 0 0 0 1 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 17 1 0 0 1 1 1 0 1 0 0 0 0 1 1 0 0 0 0 0 0 0 1 0 0 0 0 0 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 18 1 0 0 1 1 1 0 0 1 0 0 0 1 0 1 0 0 1 1 0 0 0 0 0 1 0 0 0 1 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 19 1 0 0 0 1 1 0 0 1 0 0 0 1 0 1 0 0 1 0 0 0 0 0 0 1 0 0 0 1 1 1 0 0 0 1 1 0 1 1 1 0 1 0 0 1 20 1 0 0 0 1 1 0 0 1 0 0 0 1 0 1 0 0 1 0 0 0 0 0 1 1 0 0 0 1 1 1 0 0 0 1 1 0 1 1 1 0 1 0 0 1 21 1 0 0 1 1 1 0 1 0 0 0 1 0 0 1 0 0 0 0 0 1 1 0 0 0 0 1 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 0 1 22 1 0 0 1 1 1 0 1 0 0 0 1 0 0 1 0 0 0 0 0 1 1 0 0 0 0 0 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 0 1 23 1 0 0 1 1 1 0 1 0 0 0 1 0 0 1 0 0 0 0 0 1 1 0 0 0 0 0 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 0 1 24 1 0 0 1 1 1 0 0 1 0 0 1 0 0 1 0 0 0 0 0 0 1 0 0 0 0 0 0 1 1 1 0 0 0 1 1 0 1 1 1 0 1 0 0 1 25 1 0 0 1 1 1 0 1 0 0 0 0 1 0 1 0 0 0 0 0 0 1 0 0 0 0 0 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 26 1 1 0 1 1 1 0 1 0 0 0 0 1 0 1 0 0 1 0 0 0 0 1 0 1 0 0 1 0 1 0 1 0 0 1 1 0 1 1 1 0 1 0 0 1 27 1 1 0 1 1 1 0 0 1 0 0 0 1 0 1 0 0 1 0 0 0 1 0 0 1 0 0 1 0 1 0 1 0 0 1 1 0 1 1 1 0 1 0 0 1 28 1 1 0 1 1 1 0 1 0 0 0 0 1 0 1 0 0 1 0 0 0 1 0 0 1 0 0 1 0 1 0 1 0 0 1 1 0 1 1 1 0 1 0 0 1 29 1 0 0 1 1 1 0 0 1 0 0 0 1 0 1 0 0 0 0 1 0 1 0 0 0 0 0 0 1 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 30 1 0 0 1 0 1 0 0 1 0 0 1 0 0 1 0 0 0 0 0 0 1 0 0 1 0 1 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 31 1 0 0 1 0 1 0 0 1 0 0 1 0 0 1 0 0 0 0 0 0 1 0 0 1 0 0 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 32 1 0 0 ? ? 1 0 0 1 0 0 1 0 0 1 0 0 1 0 0 0 1 0 0 0 0 0 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 ? ? 33 1 0 0 1 1 1 0 0 1 0 0 1 0 0 1 0 0 1 0 0 0 1 0 0 1 0 0 1 0 1 1 0 0 0 1 ? ? 1 1 1 0 1 0 0 0 34 1 0 0 1 1 1 0 0 1 0 0 1 0 0 1 0 0 1 0 0 0 0 0 0 1 0 0 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 0 0 35 1 0 0 1 1 1 0 0 1 0 0 1 0 0 1 0 0 1 0 0 0 0 0 0 1 0 0 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 0 0 36 1 0 0 1 0 1 0 1 0 0 0 1 0 0 1 0 0 0 0 0 0 0 1 0 0 0 0 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 0 0 37 1 0 1 1 0 1 0 1 0 0 0 1 0 0 1 0 0 0 0 0 0 1 0 0 0 0 0 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 38 1 0 0 1 1 1 0 1 0 0 0 0 1 0 1 0 0 1 0 1 0 1 1 0 0 0 0 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 0 0 39 1 0 0 1 1 1 0 1 0 0 0 0 1 0 1 0 0 0 0 1 0 1 1 0 0 0 0 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 0 0 40 1 0 0 1 1 1 0 1 0 0 0 1 0 0 1 0 0 1 0 0 0 1 0 0 0 0 0 0 1 1 1 0 0 0 1 1 0 1 1 1 0 1 0 0 1 41 1 0 0 1 0 1 0 1 0 0 0 1 0 0 1 0 0 1 0 0 0 1 0 0 0 0 0 0 1 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 42 1 0 0 1 0 1 0 1 0 0 0 1 0 0 1 0 0 1 0 0 0 1 0 0 0 0 0 0 1 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 43 1 0 0 1 0 1 0 1 0 0 0 1 0 0 1 0 0 1 0 0 0 1 0 0 0 0 0 0 1 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 44 1 0 0 1 0 1 0 1 0 0 0 1 0 0 1 0 0 1 0 0 0 1 0 0 0 0 0 0 1 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 45 1 0 0 1 1 1 0 0 1 0 0 0 1 1 0 0 0 0 0 0 0 1 1 1 0 0 0 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 46 1 0 0 1 1 1 0 0 1 0 0 0 1 1 0 0 0 0 0 0 0 1 1 1 0 0 0 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 47 1 0 0 1 1 1 0 0 1 0 0 0 1 0 1 0 0 0 0 0 0 0 1 1 0 0 0 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 48 1 0 0 1 1 1 0 0 1 0 0 0 1 1 0 0 0 0 0 0 0 0 0 1 0 0 0 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 49 1 0 0 1 1 1 0 0 1 0 0 0 1 0 1 0 0 0 0 0 0 1 0 0 0 0 1 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 50 1 1 0 1 1 1 0 1 0 0 0 0 1 0 1 0 0 0 0 0 0 1 0 1 0 0 1 0 1 1 1 0 0 0 1 1 0 1 1 1 0 1 0 0 1 51 1 0 0 ? ? 1 0 1 0 0 0 ? ? ? ? ? 0 1 0 0 0 1 1 0 0 0 0 1 0 1 1 0 0 0 1 ? ? 1 1 1 0 1 0 ? ? 52 1 0 0 1 0 1 0 0 1 0 0 0 1 1 0 0 0 0 0 0 0 1 0 0 0 0 1 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 53 1 0 0 1 0 1 0 0 1 0 0 0 1 1 0 0 0 0 0 0 0 1 0 0 0 0 1 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 54 1 0 0 1 0 1 0 0 1 0 0 0 1 1 0 0 0 0 0 0 0 1 0 0 0 0 1 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 55 1 0 0 1 1 1 0 1 0 0 0 0 1 0 1 0 0 0 0 0 0 1 1 0 0 0 0 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 56 1 0 0 1 1 1 0 1 0 0 0 0 1 0 1 0 0 0 0 0 0 1 1 0 0 0 0 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 57 1 0 0 1 1 1 0 1 0 0 0 0 1 0 1 0 0 0 0 0 0 1 1 0 0 0 0 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 58 1 0 0 1 1 1 0 1 0 0 0 0 1 0 1 0 0 0 0 0 0 1 1 0 0 0 0 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 59 1 0 0 1 1 1 0 1 0 0 0 0 1 0 1 0 0 0 0 0 0 1 1 0 0 0 0 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 60 1 0 0 1 1 1 0 1 1 0 0 1 1 1 1 0 0 0 0 1 0 1 0 0 0 0 1 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 61 1 0 0 ? 0 1 0 0 1 0 0 0 1 1 0 0 0 1 0 0 0 1 0 0 0 0 0 1 0 1 1 0 0 1 0 1 0 1 1 1 0 1 0 0 0 62 1 0 0 ? 0 1 0 0 1 0 0 0 1 1 0 0 0 1 0 0 0 1 0 0 0 0 0 1 0 1 1 0 0 1 0 1 0 1 1 1 0 1 0 0 0 63 1 0 0 ? 0 1 0 0 1 0 0 0 1 1 0 0 0 1 0 0 0 1 0 0 0 0 0 1 0 1 1 0 0 1 0 1 0 1 1 1 0 1 0 0 0 64 1 0 0 1 1 1 0 1 0 0 0 0 1 1 0 0 0 1 0 0 0 1 1 0 0 0 0 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 65 1 0 1 1 1 1 0 0 1 0 0 0 1 1 1 0 0 0 0 0 0 1 1 0 0 0 1 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 0 0 66 1 0 1 1 1 1 0 0 1 0 0 0 1 1 1 0 0 0 0 0 0 1 1 0 0 0 1 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 0 0 67 1 0 1 1 1 1 0 0 1 0 0 0 1 1 1 0 0 0 0 0 0 1 1 0 0 0 1 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 0 0 68 1 0 1 1 1 1 0 0 1 0 0 0 1 0 1 0 0 0 0 0 0 1 1 0 0 0 0 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 0 0 69 1 1 0 1 1 1 0 0 1 0 0 0 1 ? ? ? 0 0 0 0 0 1 1 0 0 0 0 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 70 1 1 0 1 1 1 0 0 1 0 0 0 1 0 1 0 0 0 0 0 0 1 1 0 0 0 0 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 71 1 1 0 1 1 1 0 0 1 0 0 0 1 0 1 0 0 0 0 0 0 1 1 0 0 0 0 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 72 1 0 0 1 1 1 0 0 1 0 0 0 1 0 1 0 0 ? 0 0 0 0 0 0 1 0 0 0 1 1 1 0 0 0 1 1 0 1 1 1 0 1 0 0 0 73 1 0 0 1 1 1 0 0 1 0 0 0 1 0 1 0 0 ? 0 0 0 0 0 0 1 0 0 0 1 1 1 0 0 0 1 1 0 1 1 1 0 1 0 0 0 74 1 0 0 1 1 1 0 0 1 0 0 1 0 0 1 0 0 0 0 0 0 1 1 0 0 0 0 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 75 1 0 1 1 0 1 0 1 0 0 0 0 1 0 1 0 0 1 0 0 0 1 1 0 0 0 0 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 76 1 1 0 1 0 1 0 1 0 0 0 0 1 0 1 1 0 0 0 0 0 1 0 0 0 0 0 1 0 1 0 0 0 1 0 1 0 1 1 1 0 1 0 0 1 77 1 0 0 1 0 1 0 0 1 0 0 0 1 0 1 0 0 0 0 1 0 1 0 0 0 0 1 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 78 1 0 0 1 0 1 0 0 1 0 0 0 1 0 1 0 0 0 0 1 0 1 0 0 0 0 1 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 79 1 0 0 1 0 1 0 0 1 0 0 0 1 0 1 0 0 0 0 1 0 1 0 0 0 0 1 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 80 1 0 1 1 1 1 0 1 0 0 0 1 0 0 1 0 0 0 0 0 0 1 0 0 0 0 0 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 0 0 81 1 0 1 1 1 1 0 0 1 0 0 1 0 0 1 0 0 0 0 0 0 1 1 0 0 0 0 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 0 0 82 1 0 0 1 1 ? 0 0 1 0 0 ? ? 0 1 0 1 0 0 0 0 0 0 0 0 0 0 1 0 1 1 0 0 1 0 1 0 1 1 1 0 1 0 1 0 83 1 0 0 1 1 ? 0 0 1 0 0 ? ? 0 1 0 1 0 0 0 0 0 0 0 0 0 0 1 0 1 1 0 0 1 0 ? ? 1 1 1 0 1 0 0 0 84 1 0 1 1 0 1 0 0 1 0 0 1 0 0 1 0 0 0 0 0 0 1 0 0 0 0 1 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 0 0 85 1 0 1 1 1 1 0 1 0 0 0 1 0 0 1 0 0 0 0 0 0 1 0 0 0 0 0 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 0 0 86 1 0 1 1 1 1 0 1 0 0 0 1 0 0 1 0 0 0 0 0 0 1 0 0 0 0 0 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 0 0 87 1 0 1 1 1 1 0 1 0 0 0 0 1 1 0 0 0 0 1 0 0 1 0 0 0 0 1 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 0 0 88 1 0 1 1 ? 1 0 0 1 0 0 0 1 1 0 0 0 0 0 0 0 1 0 0 0 0 1 1 ? 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 89 1 0 1 1 1 1 0 1 0 0 0 0 1 1 0 0 0 0 1 0 0 1 0 0 0 0 0 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 90 1 0 1 1 1 1 0 1 0 0 0 0 1 1 0 0 0 0 1 0 0 1 0 0 0 0 0 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 91 1 0 1 0 1 1 0 1 0 0 0 1 0 0 1 0 0 0 0 0 0 1 0 0 0 0 0 0 1 1 1 0 0 0 1 1 0 1 1 1 0 1 0 0 0 92 ? 0 1 0 0 1 0 1 0 0 0 0 1 1 0 0 0 0 0 0 0 0 1 0 0 0 0 1 0 1 ? ? ? ? ? 1 0 1 1 1 0 0 1 1 0 93 0 0 ? 0 1 1 0 1 0 0 0 1 0 0 1 0 0 0 0 0 0 0 0 1 0 0 0 0 1 1 ? ? ? ? ? 1 0 1 1 ? ? ? ? 0 0 94 1 0 1 0 1 1 0 0 1 0 0 1 0 0 1 0 0 0 0 0 0 1 0 0 0 0 0 1 0 1 1 0 0 0 1 1 0 1 1 0 1 0 1 1 0 95 1 0 1 0 1 1 0 0 1 0 0 1 0 0 1 0 0 0 0 0 0 1 0 0 0 0 0 1 0 1 1 0 0 0 1 1 0 1 1 0 1 0 1 1 0 96 ? 0 1 0 1 1 0 0 1 0 0 1 0 0 1 0 0 0 0 0 0 1 0 0 0 0 0 1 0 1 ? ? ? ? ? 1 0 1 1 0 1 0 1 1 0 97 1 0 1 0 1 1 0 0 1 0 0 1 0 0 1 0 0 0 0 0 0 1 0 0 0 0 0 1 0 1 1 0 0 0 1 1 0 1 1 0 1 0 1 1 0 98 1 0 1 0 1 1 0 0 1 0 0 1 0 0 1 0 0 0 0 0 0 1 0 0 0 0 0 0 1 1 1 0 0 0 1 1 0 1 1 1 0 1 0 0 0 99 1 0 0 0 1 1 0 1 0 0 0 0 1 1 0 0 1 0 0 0 0 1 0 0 0 0 0 1 0 1 0 0 0 0 1 1 0 1 1 1 0 1 0 0 0 100 1 0 1 0 0 1 0 1 0 0 0 0 1 0 0 1 0 0 0 0 0 1 0 0 0 0 0 1 0 1 0 0 0 0 1 1 0 1 1 0 1 1 0 0 0 101 1 0 1 0 1 1 0 1 0 0 0 0 1 1 0 0 1 0 0 0 0 1 0 0 0 0 0 1 0 1 0 0 0 0 1 1 0 1 1 0 1 1 0 0 0 102 1 0 0 0 0 1 0 0 1 1 1 1 0 1 1 0 0 0 0 0 0 0 1 1 0 0 0 1 0 1 1 0 0 0 1 ? ? 1 1 1 0 1 0 1 0 103 1 0 0 0 0 1 0 0 1 1 1 1 0 1 1 0 0 0 0 0 0 0 1 1 1 0 0 1 0 1 1 0 0 0 1 1 0 1 1 1 0 1 0 1 0 104 1 1 0 1 1 1 0 0 1 0 0 0 1 0 0 1 1 0 0 0 0 0 0 0 0 0 0 1 0 1 0 0 1 0 1 1 0 1 1 1 0 1 0 1 0 © 2018 The Linnean Society of London, Botanical Journal of the Linnean Society This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/open_access/funder_policies/chorus/standard_publication_model)

Sánchez-Tapia, Andrea;Garbin, Mário L;Siqueira, Marinez F;Guidoni-Martins, Karlo G;FLS, Fabio R Scarano,;Carrijo, Tatiana T

doi: 10.1093/botlinnean/boy034pmid: N/A

Abstract Studying species with different distributions and ecological requirements in the same taxonomic group allows an understanding of the specialization strategies across the environmental conditions of core vs. peripheral habitats. Here, we studied patterns of niche and geographical partitioning among Myrsine spp. and mapped areas of high alpha and beta diversity in the Brazilian Atlantic Forest. Thirteen species were classified a priori according to their ecological preferences and geographical distributions as (1) widely distributed, (2) generalists, (3) core formation specialists and (4) peripheral formation specialists. Their niche breadths were reconstructed using PCA. Ecological niche models were built using an ensemble modelling approach and retaining a majority consensus between them, aiming to build their potential distribution. Environmental and geographical overlaps were calculated, and the potential richness and species turnover were analysed. Generalists had wider niches than specialists, but peripheral species did not have narrower niches than core species. The geographical segregation pattern was correlated with niche differences. Even though the Central Corridor and the Serra do Mar sub-regions are richest in species, satellite areas were most important in protecting specialist species. Given the high species turnover between the areas with high and intermediate richness, conservation must target both types of areas. geographical distribution, Myrsinaceae, plant conservation, species diversity INTRODUCTION Shifts in environmental conditions between core and peripheral habitats impose limits to plant distribution. For example, transitions between vegetation types, biome boundaries and coastal regions constrain the limits of physiological plasticity for plant survival and reproduction (Crawford, 2008). The Brazilian Atlantic Forest (BAF) is a vegetational complex (sensu Scarano, 2002) that encompasses several core formations, including evergreen rain forest (≤ 1000 m of the eastern slopes of the coastline mountain chain), semi-deciduous forest (the plateau in the centre and south-eastern interior of the country), the subtropical Araucaria Juss. forest in the south, the north-eastern brejo forests and several peripheral neighbouring open vegetation types, including restingas, rocky outcrops and swampy areas (see Morellato & Haddad, 2000; Oliveira-Filho & Fontes, 2000; Scarano, 2002). Peripheral communities are subjected to environmental conditions, such as drought, flooding, salinity and temperature extremes, notably different from those of the core formations (Scarano, 2002, 2009). The successful establishment of plant species from core formations of the Atlantic Forest into peripheral areas requires a certain degree of ecological plasticity (Scarano, 2009), and specialization in these peripheral areas is expected to restrict distributions (Scarano et al., 2005). Many species in the Atlantic Forest have restricted distribution ranges (Rezende et al., 2014) and are thought to be more prone to threat or extinction (Pimm et al., 2014). The potential of species in the same taxonomic group to occupy a habitat or to become specialists in different environments can provide information about the levels of specialization and species turnover (a component of beta diversity, Baselga, 2010) among environments. Database collections (REFLORA, 2017; speciesLink, 2017) are the main sources of information for geographical distribution of species in Brazil, providing data on preferential habitats and conservation status. However, herbarium records can have biases and uncertainties, such as the ‘museum effect’, collection gaps or misidentified collections (Ponder et al., 2001; Murray-Smith et al., 2009; Lacerda & Nimmo, 2010; Werneck et al., 2011; Dutra et al., 2015) that compromise the understanding of species distribution patterns (Meyer, Weigelt, Kreft, 2015). Moreover, our limited knowledge about the distribution ranges of many plant species underestimates our prediction about species richness and beta diversity (e.g. how these species share the geographical space). This lack of information has theoretical and practical implications. It restricts (1) our ability to reveal ecological and evolutionary processes behind the origin and maintenance of species richness at large scales (McKnight et al., 2007) and (2) the development of effective conservation strategies (Socolar et al., 2016). The number and quality of occurrence records influence the accuracy and quality of ecological niche models and therefore a thorough data-step cleaning is essential for improving the results, even if the resulting number of records is low. Ecological niche occurs in the environmental space, i.e. the variables or dimensions of the niche, whereas the geographical space refers to the geographical configuration of such conditions and their effects on the species distribution. Ecological niche modeling (ENM) performs an estimation of the ecological niche (a set of environmental conditions that meet the conditions for a species survival) and project these conditions back to the geographical space (areas that offer appropriate abiotic conditions) (Peterson et al., 2011). ENM allows correcting biases by transforming the point-based patterns of collection records into continuous surfaces, reflecting the probability of species occurrence (Graham et al., 2004). ENM has become a common procedure for determining the geographical distribution of species in conservation research (De Marco Jr & Siqueira, 2009). Studies applying this approach are mainly focused on species lists compiled for specific regions (de Siqueira and Durigan, 2007; Rezende et al., 2014), for single-species analysis of invasives (Dunlop et al., 2006; Taylor et al., 2012; Taylor & Kumar, 2016) or for finding endemic and rare species (de Siqueira et al., 2009; Kamino et al., 2012). Studies on potential patterns of species distribution help to generate hypotheses about their conservation status (Sousa-Baena, Garcia, Peterson, 2014) and their partitioning of the ecological and geographical spaces (Casazza et al., 2013). Despite such importance, few contributions apply this approach to taxonomic groups, such as species within a genus. For example, Lirio, Peixoto & de Siqueira (2015) found a relationship between the potential geographical distribution of species of Macrotorus Perkins in different habitats of the Atlantic Forest, which was mainly related to differences in precipitation indexes. Based on species distribution models, Vargas et al., (2004) found that conservation units in Ecuador protected only c. 3% of the endemic species of Anthurium Schott. Such refinement is only possible by processing species distribution within specific taxonomic groups. Genera and families for which recent taxonomic revisions are available are especially interesting in this regard because they provide updated information on the geographical distribution of species. In this context, we chose Myrsine L. for understanding patterns of niche partitioning in the environmental space and of potential distribution in the Brazilian Atlantic Forest. Myrsine is a pantropical genus of Primulaceae. Of the 25 species listed in Brazil, 23 occur in the Brazilian Atlantic Forest, from north-eastern to southern Brazil (Flora do Brasil [Brazilian Flora], 2020). Three of these species are considered ‘endangered’ in the Brazilian Red List of Threatened Species: M. congesta (Sw.) Pipoly, M. glazioviana Warm. and M. villosissima Mart. (Freitas et al., 2013). Myrsine spp. disperse and establish in different habitats in the Atlantic Forest and some seem to specialize in narrow niche breadths. Given these ecological traits, three Myrsine spp. have been used to understand patterns of ecophysiological performance between the core and marginal areas of the Atlantic Forest (Scarano et al., 2001; Scarano, 2002, 2009; Duarte et al., 2005). The quantification of distributional patterns of Myrsine spp. and their niche breadth will allow a better understanding of the variation in species richness and its causes in the Brazilian Atlantic Forest. The goal of this study was to compare the patterns of spatial distribution of Myrsine spp. in the Brazilian Atlantic Forest, relating them to their niche breadth and habitat preferences in this biome. We expect that species present in both core and peripheral formations (generalist species) have wider niches than species restricted to core or peripheral formations (specialist species). We expect specialists restricted to peripheral areas to show the narrowest niches due to the filtering effect of peripheral ecosystems in comparison to core conditions. We also hypothesize that the potential area of the species will be directly related to their niche breadth. Specifically, our objectives were: (1) to reconstruct the niche occupied by a species in the environmental space and model their potential distribution in the geographical space; (2) to assess niche differences between specialist and generalist species and their geographical and ecological segregation patterns; and (3) to map the gradient of species richness and turnover. MATERIAL AND METHODS Species selection, occurrence data and criteria adopted Thirteen Myrsine spp. were selected for this study [M. balansae (Mez) Otegui, M. coriacea (Sw.) R.Br. ex Roem. & Schult., M. gardneriana A.DC., Myrsine guianensis (Aubl.) Kuntze, M. hermogenesii (Jung-Mend. & Bernacci) M.F.Freitas & Kin.-Gouv., M. lancifolia Mart., M. leuconeura Mart., M. parvifolia A.DC., M. parvula (Mez) Otegui, M. rubra M.F.Freitas & Kin.-Gouv., M. umbellata Mart., M. venosa A.DC. and M. villosissima Mart.]. The criterion for the species selection was the quality of datasets, which were recently updated under a floristic-taxonomic study for Primulaceae (Carrijo et al., in press). The occurrence of Myrsine spp. in different biomes of Brazil and vegetation types of Atlantic Forest (Table 1) was consulted in specialized bibliographies and databases (Jung, 1981; Siqueira, 1987, 1993; Jung-Mendaçolli and Bernacci, 1997, 2001; Jung-Mendaçolli, Bernacci, Freitas 2005; Freitas & Carrijo, 2008; BFG, 2015; Brazilian Flora 2020; Carrijo et al., in press). Table 1. Occurrences in Brazilian biomes and vegetation types for 13 Myrsine spp. in the Brazilian Atlantic Forest Species/Author Occurrence in biomes of Brazil Vegetation types M. balansae (Mez) Otegui Endemic to the Atlantic Forest Rainforest M. coriacea (Sw.) R.Br. ex Roem. & Schult. Cerrado and Atlantic Forest High elevation campos, rocky outcrops, coastal ecosystems (restinga vegetation) and open areas and below canopy gaps, forest edges in rain and semideciduous forests M. gardneriana A.DC. Cerrado and Atlantic Forest High elevation campos, rocky outcrops, rainforest M. guianensis (Aubl.) Kuntze Amazon, caatinga, cerrado, Atlantic Forest Rainforest, rocky outcrops and coastal ecosystems (restinga vegetation) M. hermogenesii (Jung-Mend. & Bernacci) M.F.Freitas & Kin. -Gouv. Endemic to the Atlantic Forest Rainforest M. lancifolia Mart. Cerrado and Atlantic Forest Rainforest M. leuconeura Mart. Cerrado and Atlantic Forest Coastal ecosystems (riparian vegetation) M. parvifolia A.DC. Cerrado, Atlantic Forest Coastal ecosystems (restinga vegetation) M. parvula (Mez) Otegui Cerrado, Atlantic Forest Rainforest, rocky outcrops M. rubra M.F.Freitas & Kin.-Gouv. Endemic to the Atlantic Forest Coastal ecosystems (swamp forest) M. umbellata Mart. Amazon, caatinga, cerrado, Atlantic Forest High elevation campos, rocky outcrops, rainforest, coastal ecosystems (restinga vegetation) M. venosa A.DC. Cerrado, Atlantic Forest Rainforest, coastal ecosystems (restinga vegetation) M. villosissima Mart. Endemic to the Atlantic Forest Rainforest, rocky outcrops Species/Author Occurrence in biomes of Brazil Vegetation types M. balansae (Mez) Otegui Endemic to the Atlantic Forest Rainforest M. coriacea (Sw.) R.Br. ex Roem. & Schult. Cerrado and Atlantic Forest High elevation campos, rocky outcrops, coastal ecosystems (restinga vegetation) and open areas and below canopy gaps, forest edges in rain and semideciduous forests M. gardneriana A.DC. Cerrado and Atlantic Forest High elevation campos, rocky outcrops, rainforest M. guianensis (Aubl.) Kuntze Amazon, caatinga, cerrado, Atlantic Forest Rainforest, rocky outcrops and coastal ecosystems (restinga vegetation) M. hermogenesii (Jung-Mend. & Bernacci) M.F.Freitas & Kin. -Gouv. Endemic to the Atlantic Forest Rainforest M. lancifolia Mart. Cerrado and Atlantic Forest Rainforest M. leuconeura Mart. Cerrado and Atlantic Forest Coastal ecosystems (riparian vegetation) M. parvifolia A.DC. Cerrado, Atlantic Forest Coastal ecosystems (restinga vegetation) M. parvula (Mez) Otegui Cerrado, Atlantic Forest Rainforest, rocky outcrops M. rubra M.F.Freitas & Kin.-Gouv. Endemic to the Atlantic Forest Coastal ecosystems (swamp forest) M. umbellata Mart. Amazon, caatinga, cerrado, Atlantic Forest High elevation campos, rocky outcrops, rainforest, coastal ecosystems (restinga vegetation) M. venosa A.DC. Cerrado, Atlantic Forest Rainforest, coastal ecosystems (restinga vegetation) M. villosissima Mart. Endemic to the Atlantic Forest Rainforest, rocky outcrops View Large Table 1. Occurrences in Brazilian biomes and vegetation types for 13 Myrsine spp. in the Brazilian Atlantic Forest Species/Author Occurrence in biomes of Brazil Vegetation types M. balansae (Mez) Otegui Endemic to the Atlantic Forest Rainforest M. coriacea (Sw.) R.Br. ex Roem. & Schult. Cerrado and Atlantic Forest High elevation campos, rocky outcrops, coastal ecosystems (restinga vegetation) and open areas and below canopy gaps, forest edges in rain and semideciduous forests M. gardneriana A.DC. Cerrado and Atlantic Forest High elevation campos, rocky outcrops, rainforest M. guianensis (Aubl.) Kuntze Amazon, caatinga, cerrado, Atlantic Forest Rainforest, rocky outcrops and coastal ecosystems (restinga vegetation) M. hermogenesii (Jung-Mend. & Bernacci) M.F.Freitas & Kin. -Gouv. Endemic to the Atlantic Forest Rainforest M. lancifolia Mart. Cerrado and Atlantic Forest Rainforest M. leuconeura Mart. Cerrado and Atlantic Forest Coastal ecosystems (riparian vegetation) M. parvifolia A.DC. Cerrado, Atlantic Forest Coastal ecosystems (restinga vegetation) M. parvula (Mez) Otegui Cerrado, Atlantic Forest Rainforest, rocky outcrops M. rubra M.F.Freitas & Kin.-Gouv. Endemic to the Atlantic Forest Coastal ecosystems (swamp forest) M. umbellata Mart. Amazon, caatinga, cerrado, Atlantic Forest High elevation campos, rocky outcrops, rainforest, coastal ecosystems (restinga vegetation) M. venosa A.DC. Cerrado, Atlantic Forest Rainforest, coastal ecosystems (restinga vegetation) M. villosissima Mart. Endemic to the Atlantic Forest Rainforest, rocky outcrops Species/Author Occurrence in biomes of Brazil Vegetation types M. balansae (Mez) Otegui Endemic to the Atlantic Forest Rainforest M. coriacea (Sw.) R.Br. ex Roem. & Schult. Cerrado and Atlantic Forest High elevation campos, rocky outcrops, coastal ecosystems (restinga vegetation) and open areas and below canopy gaps, forest edges in rain and semideciduous forests M. gardneriana A.DC. Cerrado and Atlantic Forest High elevation campos, rocky outcrops, rainforest M. guianensis (Aubl.) Kuntze Amazon, caatinga, cerrado, Atlantic Forest Rainforest, rocky outcrops and coastal ecosystems (restinga vegetation) M. hermogenesii (Jung-Mend. & Bernacci) M.F.Freitas & Kin. -Gouv. Endemic to the Atlantic Forest Rainforest M. lancifolia Mart. Cerrado and Atlantic Forest Rainforest M. leuconeura Mart. Cerrado and Atlantic Forest Coastal ecosystems (riparian vegetation) M. parvifolia A.DC. Cerrado, Atlantic Forest Coastal ecosystems (restinga vegetation) M. parvula (Mez) Otegui Cerrado, Atlantic Forest Rainforest, rocky outcrops M. rubra M.F.Freitas & Kin.-Gouv. Endemic to the Atlantic Forest Coastal ecosystems (swamp forest) M. umbellata Mart. Amazon, caatinga, cerrado, Atlantic Forest High elevation campos, rocky outcrops, rainforest, coastal ecosystems (restinga vegetation) M. venosa A.DC. Cerrado, Atlantic Forest Rainforest, coastal ecosystems (restinga vegetation) M. villosissima Mart. Endemic to the Atlantic Forest Rainforest, rocky outcrops View Large These species were classified into groups, considering independently: (1) the types of vegetation in which they occur in the Atlantic Forest (core and/or peripheral, in ecological terms) and (2) the amplitude of their known geographical distribution (in geographical terms). In ecological terms, species occurring in both core and peripheral ecosystems were classified as generalists and species that occur only in core formations or peripheral formations were considered as specialists. We kept the core/peripheral division in specialist species because of the fundamental ecological differences between these formations and the limits to survival in peripheral areas compared with core areas. In geographical terms, we made a distinction between generalists that occur mainly inside the Atlantic Forest and generalists that occupy large tracts of South and Central America and classified the former as regional and the latter as widely distributed. The geographical distribution of specialist species was either regional or local and was not used to subclassify specialists (Table 2). We defined thus four groups of species: widely distributed generalists; regionally distributed generalists; core formation specialists and peripheral formation specialists. Hereafter, these groups will be referred to as (1) widely distributed, (2) generalist, (3) core and (4) peripheral, respectively. Table 2. Myrsine spp., range of habitats where they are most commonly found, the group classification used throughout the analyses, number of occurrences and relative niche breadth (for M. coriacea NB = 1). Species are sorted by the number of occurrences Species/Author Range of habitats in Atlantic Forest Ecological classification Group Number of occurrences Niche breadth M. coriacea (Sw.) R.Br. ex Roem. & Schult. Core formation / coastal ecosystems / rocky outcrops / high elevation campos Generalist Widely distributed 963 1 M. umbellata Mart. Core formation / coastal ecosystems / rocky outcrops Generalist Widely distributed 603 0.632 M. guianensis (Aubl.) Kuntze Core formation / coastal ecosystems / rocky outcrops Generalist Widely distributed 544 0.78 M. gardneriana A.DC. Core / rocky outcrops / high elevation campos Generalist Generalist 285 0.484 M. venosa A.DC. Core formation / coastal areas Generalist Generalist 137 0.311 M. parvula (Mez) Otegui Core formation / rocky outcrops Generalist Generalist 114 0.302 M. parvifolia A.DC. Coastal ecosystems Specialist Peripheral 110 0.234 M. balansae (Mez) Otegui Core formation Specialist Core 57 0.137 M. hermogenesii (Jung- Mend. & Bernacci) M.F.Freitas & Kin.-Gouv. Core formation Specialist Core 40 0.192 M. lancifolia Mart. Core formation Specialist Core 33 0.137 M. leuconeura Mart. Coastal ecosystems Specialist Peripheral 25 0.085 M. villosissima Mart. Core / rocky outcrops / high elevation campos Generalist Generalist 22 0.085 M. rubra M.F.Freitas & Kin.-Gouv. Coastal ecosystems Specialist Peripheral 7 0.009 Species/Author Range of habitats in Atlantic Forest Ecological classification Group Number of occurrences Niche breadth M. coriacea (Sw.) R.Br. ex Roem. & Schult. Core formation / coastal ecosystems / rocky outcrops / high elevation campos Generalist Widely distributed 963 1 M. umbellata Mart. Core formation / coastal ecosystems / rocky outcrops Generalist Widely distributed 603 0.632 M. guianensis (Aubl.) Kuntze Core formation / coastal ecosystems / rocky outcrops Generalist Widely distributed 544 0.78 M. gardneriana A.DC. Core / rocky outcrops / high elevation campos Generalist Generalist 285 0.484 M. venosa A.DC. Core formation / coastal areas Generalist Generalist 137 0.311 M. parvula (Mez) Otegui Core formation / rocky outcrops Generalist Generalist 114 0.302 M. parvifolia A.DC. Coastal ecosystems Specialist Peripheral 110 0.234 M. balansae (Mez) Otegui Core formation Specialist Core 57 0.137 M. hermogenesii (Jung- Mend. & Bernacci) M.F.Freitas & Kin.-Gouv. Core formation Specialist Core 40 0.192 M. lancifolia Mart. Core formation Specialist Core 33 0.137 M. leuconeura Mart. Coastal ecosystems Specialist Peripheral 25 0.085 M. villosissima Mart. Core / rocky outcrops / high elevation campos Generalist Generalist 22 0.085 M. rubra M.F.Freitas & Kin.-Gouv. Coastal ecosystems Specialist Peripheral 7 0.009 View Large Table 2. Myrsine spp., range of habitats where they are most commonly found, the group classification used throughout the analyses, number of occurrences and relative niche breadth (for M. coriacea NB = 1). Species are sorted by the number of occurrences Species/Author Range of habitats in Atlantic Forest Ecological classification Group Number of occurrences Niche breadth M. coriacea (Sw.) R.Br. ex Roem. & Schult. Core formation / coastal ecosystems / rocky outcrops / high elevation campos Generalist Widely distributed 963 1 M. umbellata Mart. Core formation / coastal ecosystems / rocky outcrops Generalist Widely distributed 603 0.632 M. guianensis (Aubl.) Kuntze Core formation / coastal ecosystems / rocky outcrops Generalist Widely distributed 544 0.78 M. gardneriana A.DC. Core / rocky outcrops / high elevation campos Generalist Generalist 285 0.484 M. venosa A.DC. Core formation / coastal areas Generalist Generalist 137 0.311 M. parvula (Mez) Otegui Core formation / rocky outcrops Generalist Generalist 114 0.302 M. parvifolia A.DC. Coastal ecosystems Specialist Peripheral 110 0.234 M. balansae (Mez) Otegui Core formation Specialist Core 57 0.137 M. hermogenesii (Jung- Mend. & Bernacci) M.F.Freitas & Kin.-Gouv. Core formation Specialist Core 40 0.192 M. lancifolia Mart. Core formation Specialist Core 33 0.137 M. leuconeura Mart. Coastal ecosystems Specialist Peripheral 25 0.085 M. villosissima Mart. Core / rocky outcrops / high elevation campos Generalist Generalist 22 0.085 M. rubra M.F.Freitas & Kin.-Gouv. Coastal ecosystems Specialist Peripheral 7 0.009 Species/Author Range of habitats in Atlantic Forest Ecological classification Group Number of occurrences Niche breadth M. coriacea (Sw.) R.Br. ex Roem. & Schult. Core formation / coastal ecosystems / rocky outcrops / high elevation campos Generalist Widely distributed 963 1 M. umbellata Mart. Core formation / coastal ecosystems / rocky outcrops Generalist Widely distributed 603 0.632 M. guianensis (Aubl.) Kuntze Core formation / coastal ecosystems / rocky outcrops Generalist Widely distributed 544 0.78 M. gardneriana A.DC. Core / rocky outcrops / high elevation campos Generalist Generalist 285 0.484 M. venosa A.DC. Core formation / coastal areas Generalist Generalist 137 0.311 M. parvula (Mez) Otegui Core formation / rocky outcrops Generalist Generalist 114 0.302 M. parvifolia A.DC. Coastal ecosystems Specialist Peripheral 110 0.234 M. balansae (Mez) Otegui Core formation Specialist Core 57 0.137 M. hermogenesii (Jung- Mend. & Bernacci) M.F.Freitas & Kin.-Gouv. Core formation Specialist Core 40 0.192 M. lancifolia Mart. Core formation Specialist Core 33 0.137 M. leuconeura Mart. Coastal ecosystems Specialist Peripheral 25 0.085 M. villosissima Mart. Core / rocky outcrops / high elevation campos Generalist Generalist 22 0.085 M. rubra M.F.Freitas & Kin.-Gouv. Coastal ecosystems Specialist Peripheral 7 0.009 View Large Species occurrences were downloaded from GBIF (www.gbif.org) and the speciesLink databases (speciesLink, 2017, http://www.splink.cria.com.br). The search in GBIF was performed using the rgbif R package (Chamberlain et al., 2016) selecting records with the following filters: ‘with coordinates’; ‘based on preserved specimens’ and ‘without geospatial issues’. The search in speciesLink was performed to include: (1) synonyms in the List of Species of the Brazilian Flora; (2) only records with original coordinates and (3) non-suspicious coordinates. The resulting occurrences were complemented with a personal database with herbarium specimens revised by T.T. Carrijo (N = 300). GBIF includes the species records deposited in RB (Rio de Janeiro Botanical Garden) and speciesLink, which includes all other Brazilian herbaria. This double query may appear redundant; however, speciesLink included recent changes in the taxonomic determination of some records. Therefore, we initially kept both data sources and excluded these duplicate records later in the data cleaning process (Supporting Information, Appendix S1). The resulting occurrences were checked for taxonomic consistency by looking for collector numbers and checking whether the species was identified by a specialist. Whenever a record (e.g. same collector and species number) was present in different collections, we gave priority to the names given by taxonomic specialists and excluded records with more than one name. We excluded some suspicious coordinates that entered the list despite the applied filters, such as the centroid of countries and states, common in collections made before 1980. We also excluded suspicious coordinates, such as new occurrences in states of Brazil with no known identifier, and misidentifications, which were checked in the collection, for all species. Finally, we removed duplicates and occurrences in the same pixel of the environmental layers (5 arcmin resolution). Ecological niche breadth and environmental space partitioning We used Worldclim’s bioclimatic variables at 5 arcmin resolution (Hijmans et al., 2005), cropped for South and Central America as environmental variables. This spatial resolution is in agreement with the quality of the herbarium records (Giannini et al., 2012), once a part of the collection is georeferenced at the same resolution as the municipality. For ecological niche reconstruction, the environmental conditions at the occurrence points were summarized by principal components analysis (PCA; Legendre & Legendre, 2012), using a subset of the variables with absolute pairwise correlations below 0.7. Elevation (h_dem variable in HYDRO1k, available from the USGS) was correlated to several bioclimatic variables (e.g. Bio1, annual mean temperature, r = −0.76, Bio5, max temperature of warmest month, r = −0.80, Bio 10, mean temperature of the warmest quarter, r = −0.87), and we did not select it when controlling for collinearity. The PCA allowed us to visualize the environmental space occupied by each species. Niche breadth was estimated as the area of the convex hull formed by the species occurrences along the two principal components of the PCA, taking the widest niche as reference (M. coriacea equal to 1). We also calculated the distance between each occurrence point and the centroid of the convex hull for the species and calculated their mean and standard deviation, which can be interpreted as a multivariate measure of the variance of the niche extent (the variability of habitat conditions used by the species) (Anderson, 2006; Oksanen et al., 2016). Niche areas and variances were calculated using the community ecology package vegan v. 2.3-0 (Oksanen et al., 2015) in R (R Core Team, 2015). Ecological niche modelling We transformed the 19 bioclimatic variables into a set of six orthogonal variables resulting from a PCA performed for the totality of the environmental space. We did not select variables according to the particular environmental requirements of each species but created these variables to represent the variety of sites where the studied taxa occur. This procedure allows for discrimination and comparison of niches and geographical distributions within the common environmental space of the BAF (Broennimann et al., 2012). We used a k-fold cross-validation approach for each species with ten or more records, with k = 3. The occurrence dataset was split in three, with three modelling rounds performed on each one, using two-thirds of the data for training the models and the remaining third for testing them. For M. rubra, which had N = 7 occurrence records, a jackknife (leave-one-out) approach was performed, i.e. setting k = N. We randomly sampled 100 pseudoabsences per occurrence point (Lobo & Tognelli, 2011) with an upper limit of 10000. Pseudoabsences were randomly selected within a maximum distance buffer, i.e. the width of the buffer corresponding to the maximum geographical distance between occurrence points. This procedure assured that pseudoabsences were sampled in areas where species could disperse, at least theoretically (Barve et al., 2011; Stokland et al., 2011; Barbet-Massin et al., 2012), while controlling the low prevalence associated with generating pseudoabsences inside large ranges. In addition, we sampled pseudoabsences outside the pixels where occurrence points were recorded, excluding exact environmental information from the background (Hirzel et al., 2001, Lütolf et al., 2006). The modelling algorithms were MaxEnt, Bioclim, minimum environmental distance, Random Forests, General Linear Models (GLM) and Support Vector Machines (SVM). MaxEnt and Bioclim are implemented in the R package dismo (Hijmans et al., 2015), SVM in the package kernlab (Karatzoglou et al., 2004), Random Forests in the package randomForest (Liaw & Wiener, 2002) and GLM in base R (R Core Team, 2015). Minimum environmental distance algorithm was calculated as the minimum value of the Euclidean Distance between the environmental conditions at each pixel and the values for the set of occurrence points. To assess the performance of each model, we calculated the area under the ROC curve (AUC), which varies between 0 and 1, and the maximum true skill statistic (TSS; Allouche et al., 2006), which varies between −1 and 1. To generate the final model per algorithm, we retained the models with AUC > 0.7 and TSS > 0.5. The final models were generated by cutting each continuous model by the threshold that maximizes its TSS. To build the ensemble model we calculated the mean of the resulting binary models. To generate a robust estimation of the potential distribution area, we retained the areas predicted by at least 50% of the selected models (majority consensus; Araújo & New, 2007). All analyses were carried out in the R statistical environment (R Core Team, 2015). Relationship between niche breadth and potential geographical areas The modelled potential area for each species and their niche breadth were compared through regression for all species and between groups (widely distributed generalists, regional generalists, core specialists, peripheral specialists) using permutation-based ANOVAs with pairwise post hoc tests (Basso, 2009). Partition of the environmental and geographical space To assess the partitioning of the environmental and geographical spaces by the studied species, we applied Broennimann et al.’s (2012) framework on the PCA environmental space and the resulting ENM. This framework builds on the calculation the D and I overlap statistics (Warren, Glor & Turelli 2008), but corrects for record density and differences in resolution by applying a Kernel correction on a gridded space. We used the I statistic and a further correction for differences in the species prevalence. The significance of pairwise niche differences and the potential distributions was assessed through permutation, using the ecospat 2.2.0 R package (Broennimann, Cola & Guisan, 2016). At last, a Mantel’s test was performed to compare the relationship between the partition of the environmental and the geographical spaces, using 1-D as a distance measure. Species richness and relative importance of generalist and specialist species A potential richness map was built by adding the final models for each species in the BAF. Since the richest areas include all species, widely distributed species and generalist species accounted for almost half of the potential richness. We analysed the contribution of each group to species richness per group to highlight the contribution of each group to total richness. This was based on choosing areas where widely distributed species account for more than half of the richness (i.e. areas with four species, two of which are widely distributed); (1) areas where generalists account for > 30% of the richness (four out of 13 species: 30%) and (2) areas where core and peripheral species account for > 20% of the richness (three out of 13 species: 23%). Species turnover We estimated the species turnover among pixels of the predicted binary models, using Jaccard-based pairwise dissimilarity proposed by Baselga (2010). This pairwise dissimilarity decomposes the total beta diversity between species turnover and nestedness. Whereas species turnover reflects the replacement of some species by others, species nestedness reflects only differences in species richness between two communities (Baselga, 2010) or pixels, as in our study. Total beta diversity was estimated based on the mean beta diversity values between a focal pixel and 16 adjacent pixels (second order adjacencies) (Melo, Rangel, Diniz-Filho, 2009). This analysis was performed after reducing the resolution of the models by a factor of five (i.e. joining pixels in 5 × 5 groups) and retaining the potential presence of each species. The resulting maps depict the species turnover values. This analysis was performed in the R package ‘betapart’ (Baselga et al., 2017). RESULTS We found 11 748 records for the 13 Myrsine spp. analysed: 7875 records from GBIF, 3573 from speciesLink, and 300 from the personal database of T. Carrijo. Following data cleaning, we retained 2940 unique environmental occurrences (Supporting Information, Appendix S1). The final number of occurrences per species is shown in Table 2. Based on their presence in core and peripheral vegetation types and their geographical distributions, the classification into groups resulted as follows: (1) widely distributed generalist species: M. coriacea, M. umbellata and M. guianensis; (2) generalist species: M. gardneriana, M. parvula, M. venosa and M. villosissima; (3) core formation specialists: M. balansae, M. hermogenesii and M. lancifolia and (4) peripheral formation specialists: M. leuconeura, M. parvifolia and M. rubra. The seven variables selected for summarizing the environmental space in the PCA were Bio_1 (annual mean temperature), Bio_2 (mean temperature diurnal range), Bio_3 (isothermality), Bio_7 (temperature annual range), Bio_13 (precipitation of wettest month), Bio_15 (precipitation seasonality) and Bio_18 (mean precipitation of warmest quarter) (Fig. 1). These variables explained 54.09% of the variation in the two first axes, and divided the environmental space along the first axis from homogeneously warm (high temperature and isothermality) to seasonal climates (high annual and diurnal temperature range) (Fig. 1). Figure 1. View largeDownload slide Principal components analysis of the environmental space occupied by Myrsine spp. The grey dots represent the whole set of occurrences (N = 2940; top left) and the polygons highlight the occurrences for each species superimposed on the occurrences of all other species. Bio_1 (annual mean temperature), Bio_2 (mean temperature diurnal range), Bio_3 (isothermality), Bio_7 (temperature annual range), Bio_13 (precipitation of wettest month), Bio_15 (precipitation seasonality) and Bio_18 (mean precipitation of warmest quarter). Niche area corresponds to the area of the hull relative to the largest polygon (Myrsine coriacea = 1). Species are depicted in order of decreasing niche breadth, and are classified as widely distributed (red), generalist (green), core (yellow) or peripheral (blue). Figure 1. View largeDownload slide Principal components analysis of the environmental space occupied by Myrsine spp. The grey dots represent the whole set of occurrences (N = 2940; top left) and the polygons highlight the occurrences for each species superimposed on the occurrences of all other species. Bio_1 (annual mean temperature), Bio_2 (mean temperature diurnal range), Bio_3 (isothermality), Bio_7 (temperature annual range), Bio_13 (precipitation of wettest month), Bio_15 (precipitation seasonality) and Bio_18 (mean precipitation of warmest quarter). Niche area corresponds to the area of the hull relative to the largest polygon (Myrsine coriacea = 1). Species are depicted in order of decreasing niche breadth, and are classified as widely distributed (red), generalist (green), core (yellow) or peripheral (blue). PCA reflects the differences in niche breadth between the studied Myrsine spp. Widely distributed species had the widest niches, as expected [M. coriacea niche breadth (hereafter NB) = 1, M. guianensis NB = 0.78, M. umbellata NB = 0.63]. However, we found a mix of niche breadth in generalist species, especially regarding M. villosissima (M. gardneriana, NB = 0.48, M. venosa, NB = 0.31, M. parvula, NB = 0.3, M. villosissima NB = 0.08). The niche breadth of core species (M. balansae NB = 0.14, M. hermogenesii NB = 0.19, M. lancifolia NB = 0.14), was relatively constant, but peripheral species (M. leuconeura NB = 0.09, M. parvifolia NB = 0.23, M. rubra NB = 0.01) had large differences between them, especially due to the large niche in M. parvifolia. Overall, only niche area for widely distributed species was significantly different from the other groups (global P value = 0.002), and the niche of peripheral species was not narrower than that of core species (P = 0.866) (Table 2; Fig. 1). Pairwise overlap in the environmental space revealed a similar use of the environmental space of widely distributed and generalist species, that occupy most of the environmental space and species-specific patterns, independent of the ecological group (Supporting Information, Appendix S2). The number of models retained based on TSS and AUC values varied from 13 to 18 out of the models fitted for each species (Supporting Information, Appendix S3). All the models were retained for M. coriacea, M. gardneriana, M. parvula and M. umbellata. The highest percentage of models was retained for widely distributed species, but in other cases the number of retained models was unrelated to the number of original occurrences. The mean TSS of the retained models showed no relationship with groups or number of occurrences. Regarding the algorithms applied, BioClim, SVM and GLM under-performed when compared with Euclidean Distance, Maxent and Random Forests. No significant relationship between the original number of occurrences and the final potential richness was found (Supporting Information, Appendix S4), suggesting that our procedure was efficient in avoiding sampling biases inherent to biodiversity datasets. The ecological niche models (Figs 2–5) revealed the use of geographical space in Myrsine spp. As expected, the widely distributed species (M. coriacea, M. guianensis, M. umbellata) occupy large geographical extensions in the BAF. Ecological generalists (M. gardneriana, M. parvula, M. venosa and M. villosissima) had intermediate potential areas and niches, except for M. villosissima, that had both a small potential area and relatively narrow niche (Fig. 6). Generalists had a small geographical overlap in the Serra do Mar, but otherwise occupy different geographical areas in the BAF (Supporting Information, Appendix 2). Areas for core specialists ranged from 1600 to 3900 km2. Myrsine balansae, M. hermogenesii and M. lancifolia occupy different portions of the BAF, with M. balansae in the south-west, M. hermogenesii in the Serra do Mar and the south of Bahia and M. lancifolia restricted to the Serra do Mar. Peripheral species had different potential areas, ranging from 600 to 4500 km2. Myrsine parvifolia and M. rubra were restricted to coastal restinga areas, following the distribution of these ecosystems, whereas M. leuconeura had a larger potential distribution than expected given its relatively narrow niche, probably because it is also present in riparian areas in cerrado and occupies mainly swampy areas in the BAF. Potential area differed significantly only between the widely distributed species and each of the other groups (core, peripheral and generalists; P = 0.008). Overall, species with wider niches tended to show larger potential geographical areas (R2 = 0.66, P < 0.001, Fig. 6) and their overlap in the environmental space was correlated with their overlap in the geographical space (Mantel statistic r: 0.7526, significance: 0.001 with 999 permutations, Supporting Information, Appendix 2). Figure 2. View largeDownload slide Potential geographical distribution (green areas) of species with wide distributions in the Brazilian Atlantic Forest: Myrsine coriacea, M. guianensis and M. umbellata (occurrence in red dots). Figure 2. View largeDownload slide Potential geographical distribution (green areas) of species with wide distributions in the Brazilian Atlantic Forest: Myrsine coriacea, M. guianensis and M. umbellata (occurrence in red dots). Figure 3. View largeDownload slide Potential geographical distribution (green areas) of generalist species in the Brazilian Atlantic Forest: Myrsine gardneriana, M. parvula, M. venosa and M. villosissima (occurrence in red dots). Figure 3. View largeDownload slide Potential geographical distribution (green areas) of generalist species in the Brazilian Atlantic Forest: Myrsine gardneriana, M. parvula, M. venosa and M. villosissima (occurrence in red dots). Figure 4. View largeDownload slide Potential geographical distribution (green areas) of core species in the Brazilian Atlantic Forest: Myrsine balansae, M. hermogenesii and M. lancifolia (occurrence in red dots). Figure 4. View largeDownload slide Potential geographical distribution (green areas) of core species in the Brazilian Atlantic Forest: Myrsine balansae, M. hermogenesii and M. lancifolia (occurrence in red dots). Figure 5. View largeDownload slide Potential geographical distribution (green areas) of the peripheral species in the Brazilian Atlantic Forest: Myrsine leuconeura, M. parvifolia and M. rubra (occurrence in red dots). Figure 5. View largeDownload slide Potential geographical distribution (green areas) of the peripheral species in the Brazilian Atlantic Forest: Myrsine leuconeura, M. parvifolia and M. rubra (occurrence in red dots). Figure 6. View largeDownload slide Relationship between niche breadth and their potential distribution area (km2) of Myrsine spp. (grey circles) in the Brazilian Atlantic Forest. rub, Myrsine rubra; leu, M. leuconeura; bal, M. balansae; leu, M. leuconeura; parvu, M. parvula; lan, M. lancifolia; parvi, M. parvifolia; her, M. hermogenesii; vil, M. vilosissima; gui, M. guianensis; ven, M. venosa; gar, M. gardneriana; cor, M. coriacea; umb, M. umbellata. There was a general linear relationship between niche breadth and potential area. Figure 6. View largeDownload slide Relationship between niche breadth and their potential distribution area (km2) of Myrsine spp. (grey circles) in the Brazilian Atlantic Forest. rub, Myrsine rubra; leu, M. leuconeura; bal, M. balansae; leu, M. leuconeura; parvu, M. parvula; lan, M. lancifolia; parvi, M. parvifolia; her, M. hermogenesii; vil, M. vilosissima; gui, M. guianensis; ven, M. venosa; gar, M. gardneriana; cor, M. coriacea; umb, M. umbellata. There was a general linear relationship between niche breadth and potential area. The potential species richness of Myrsine in the Brazilian Atlantic Forest is shown in Figure 7. The most species-rich quadrats (nine to 12 species) are located along the montane areas of Espírito Santo (Atlantic Forest Central Corridor) and of Rio de Janeiro and São Paulo (Serra do Mar), the central areas of BAF. These areas shelter most of Myrsine spp. that occur preferentially in the core formation, the high elevation campos and rocky outcrops (Table 2). Intermediate richness areas (five to eight species) could be found around this central portion, in southern Bahia, southern Minas Gerais and the inner parts of São Paulo, Paraná and Rio Grande do Sul areas (satellite regions). When the proportion of widely distributed, generalist, core and peripheral species was analysed independently (Fig. 8), the regions that appeared as species-rich for each of these groups differed from the richest areas. Core species were concentrated in southern and inland western areas, and peripheral species were most important towards the north in areas of transition with Caatinga and Cerrado and along the coast. Generalists were important throughout the area and widely distributed species were important in some low diversity areas (Fig. 8). Turnover between species was concentrated in areas with intermediate richness, away from the Central Corridor and the Serra do Mar, which concentrates the highest diversity (Fig. 9). These areas comprised Bahia and Espírito Santo coastal areas, where peripheral species reached their highest potential richness, and southern areas where core species attained their highest importance. Figure 7. View largeDownload slide Potential richness of Myrsine spp. in the Atlantic Forest, obtained by adding the final model for each species. The richest areas correspond to mountain regions in the Espírito Santo, Rio de Janeiro and São Paulo states, in the so-called Central corridor of the BAF. Figure 7. View largeDownload slide Potential richness of Myrsine spp. in the Atlantic Forest, obtained by adding the final model for each species. The richest areas correspond to mountain regions in the Espírito Santo, Rio de Janeiro and São Paulo states, in the so-called Central corridor of the BAF. Figure 8. View largeDownload slide Areas with a high proportion of widely distributed, generalist, core and peripheral Myrsine spp. in the Brazilian Atlantic Forest. Widely distributed species dominated (i.e. accounted for more than half the richness) the less rich northern portion and the central area that corresponds to rainforest to deciduous forest and forest to Cerrado transitions. Generalist species proportion was higher (i.e. > 30% of the richness) in the southern states. Core and peripheral species proportion was higher in the central portion and the coastal areas, respectively. Core and peripheral importance overlapped in the Bahia hotspot, between the São Francisco and Rio Doce rivers. Figure 8. View largeDownload slide Areas with a high proportion of widely distributed, generalist, core and peripheral Myrsine spp. in the Brazilian Atlantic Forest. Widely distributed species dominated (i.e. accounted for more than half the richness) the less rich northern portion and the central area that corresponds to rainforest to deciduous forest and forest to Cerrado transitions. Generalist species proportion was higher (i.e. > 30% of the richness) in the southern states. Core and peripheral species proportion was higher in the central portion and the coastal areas, respectively. Core and peripheral importance overlapped in the Bahia hotspot, between the São Francisco and Rio Doce rivers. Figure 9. View largeDownload slide The species turnover component of beta diversity for 13 Myrsine spp. in the Brazilian Atlantic Forest. The highest species turnover was found in the northern Bahia and Espírito Santo hotspots, and in the southern portions, away from the central corridor that has the highest species richness. Figure 9. View largeDownload slide The species turnover component of beta diversity for 13 Myrsine spp. in the Brazilian Atlantic Forest. The highest species turnover was found in the northern Bahia and Espírito Santo hotspots, and in the southern portions, away from the central corridor that has the highest species richness. DISCUSSION Our results confirmed the hypothesis of a general relationship between niche breadth and potential distribution area for Myrsine spp. of the Brazilian Atlantic Forest. Widely distributed and generalist species had wider niches than specialists, but peripheral species did not have smaller niches or more restricted potential distributions than core species. The potential species richness was highest in the Serra do Mar, but some areas with intermediate richness had a high proportion of core and peripheral species, highlighting the need for conservation strategies that include these areas. The positive relationship between niche breadth and geographical range is a well-known pattern (Slatyer, Hirst & Sexton, 2013), but the mechanisms behind it are not always clearly understood. Here, we hypothesized that the ability to colonize different environments would lead to both larger niches and potential geographical areas. We also expected peripheral species to be more restricted than core species, due to the filtering effects of peripheral environments. However, core species behaved as a homogeneous group in their niche breadths and potential areas, whereas peripheral species behaved differently depending on the vegetation type they occupy. Scarano et al. (2005) suggested that two types of peripheral species exist in the BAF, those that colonize recent environments such as quaternary coastal plains (restingas), and those that are highly adapted to relictual environments. However, their potential area may depend on the prevalence of the preferred environmental conditions independently from the niche breadth of the species. For instance, M. parvifolia has the largest niche breadth among the specialist species, but it is restricted to the marginal coastal (restinga) areas suggesting that the niche breadth is related to the latitudinal gradient. In the case of M. leuconeura, restricted to riparian forests, the resulting geographical distribution spans a large area despite a narrow niche and, in the case of M. rubra, its restriction to swampy vegetation determines both a small niche and a restricted distribution. Niche segregation patterns among Myrsine spp. were highly species-specific and mostly independent of the ecological classification, with widely distributed and generalist species occupying large areas of the environmental and geographical spaces, and species-specific patterns among the rest of species. The partition of the environmental space was related to spatial segregation, as indicated by the Mantel test. However, not only climate constraints are expected to determine such patterns. Positive interactions have been shown to be an important structuring force in the peripheral regions of the Atlantic Forest (Dias et al., 2005; Dias & Scarano, 2007; Garbin et al., 2012) and biotic interactions can affect species distributions at large macroecological scales (Araújo & Luoto, 2007). Moreover, maintenance of species diversity in peripheral habitats, with unfavourable climatic conditions, may depend on such positive interactions (Valiente-Banuet et al., 2006). Thus, the restriction of Myrsine spp. to some of the peripheral habitats may depend on the facilitative effect of nurse plants. Potential richness patterns revealed that high species richness was concentrated in the Serra do Mar, which is a well-known region of endemism in the BAF (Carnaval & Moritz, 2008; Ribeiro et al., 2009; da Silva & Casteleti, 2009; Werneck et al., 2011). Conservation strategies based on species richness would overestimate this region as a conservation priority. However, species richness is maximized with the joint presence of the widely distributed and generalist species. The presence of these species masks the importance of intermediate-richness, satellite areas, in conserving Myrsine, such as the Bahia hotspot and the Espirito Santo coast, which form the Central Corridor of the BAF, and some inland areas corresponding to semideciduous forest and Araucaria forest. The analysis of richness patterns per group of species allowed us to uncover areas with differences in the importance of core, peripheral, generalist and widely distributed species. Sites that preserve peripheral species (Fig. 8) had mostly intermediate species richness and were outside the central high diversity regions. Likewise, core specialists had a high importance in the Araucaria and the interior regions in the south-west (sensu da Silva & Casteleti, 2003). Overall, satellite areas were as important as central corridors because they have intermediate richness in areas normally neglected for habitat protection and include species with relative importance suggesting that they are of great interest for biodiversity conservation. Moreover, conserving such areas would also protect marginal populations of Myrsine spp. and the genetic diversity and evolutionary adaptations involved in the plasticity of these species that allow colonization of such habitats. This has serious implications for the development of conservation strategies in an era of climatic changes, when predicting and understanding how plants acclimate and adapt to extreme conditions (Reyer et al., 2013), commonly found in the marginal habitats, are paramount to effective conservation (Scarano & Ceotto, 2015). Our results provide evidence to Scarano’s (2002, 2009) argument that conservation strategies should treat all of the Brazilian Atlantic Forest as a complex entity rather than focus on core formations alone. Species turnover, an important component of species richness in hotspots such as the Atlantic Forest, was concentrated in regions with intermediate species richness, which highlights our previous result that satellite regions should be taken into account for conservation planning. Turnover pattern relies on the existence of several restricted species across geographical space (Socolar et al., 2016). The use of a single taxonomic group such as Myrsine can be informative for setting conservation priorities and strategies. If from an alpha-diversity perspective, priority would be set at the central high-diversity regions (Serra do Mar and the Central Corridor), from a beta-diversity perspective, Myrsine shows a turnover pattern, whereby species present at one site are replaced by other species absent from the first. After examining plant communities, Socolar et al., (2016) argue that turnover across natural sites implies that conservation must target multiple sites. We argue that a similar logic can be applied to biodiversity conservation at the genus level. Some models for conservation priority emphasize a restricted distribution as a measure of the degree of threat and probability of extinction (Myers et al., 2000; Pimm et al., 2014). This implies that restricted species are expected to be more fragile than common species. For instance, M. villosissima showed a narrow niche breadth and a small potential area and is the only species of this study included in the Brazilian Red List of Threatened Species (Freitas et al., 2013). However, this may not always be the case, as pointed out by Scarano (2009). This indicates that being a narrow-range species does not necessarily mean biological fragility and that being a widely distributed species (e.g. M. balansae, M. lancifolia) does not necessarily imply a wide niche. At the genus level, Myrsine can be used as an indicator of beta-diversity at these larger spatial scales, instead of focusing on local communities. For future work, it might be important also include other genera to allow more generalized conclusions in this regard. CONCLUSION Our study revealed that Myrsine spp. partition their geographical distribution, and that both peripheral areas and core formations are important for species conservation in the Atlantic Forest. Even though satellite areas showed lower species richness when compared with the central high diversity regions, they maintain species geographically restricted to these areas, as well as species that share the core and peripheral areas. We showed that at larger geographical scales, the potential distribution of genera such as Myrsine may be used as an indicator of beta-diversity. Thus, conservation priorities should encompass both high and intermediate species richness areas. Further research is required to understand the effect of environmental factors and biotic interactions at finer scales on the final distribution of Myrsine spp. SUPPORTING INFORMATION Additional Supporting Information may be found in the online version of this article at the publisher’s web-site. Appendix S1. Data cleaning process. Three data sources were used, speciesLink (N = 3573), GBIF (N = 7875) and Tatiana Carrijo’s personal database for these species (N = 300). After suspicious coordinates, dubious taxonomic determination and duplicates were removed, the database had 5455 records, of which 2940 are unique environmental records (one record per species per pixel). Appendix S2. Pairwise species overlap in the environmental and the geographic spaces, following Broenimann et al. (2012). Convex hulls and rasters are shown for simplicity, but the calculations were made based on Broenimann’s gridded space and kernel correction, setting the resolution to 100. Numbers represent the D statistic and bold characters indicate significance at 0.05 with 99 permutations. A Mantel test revealed a significant relationship between both overlap matrices (Mantel statistic 0.65, P < 0.001). Appendix S3. Number of partitions retained during the species distribution modeling processes. In parentheses, the number of selected partitions (AUC > 0.7, TSS > 0.5). GLM – generalized linear model; maxent – maximum entropy, mindist – minimum distance; SVM – support vector machines; TSS – maximum true skill statistic. Appendix S4. Relationship between the original number of records and the final potential richness per pixel (based on the ENM). High potential richness did not arise from a sampling effect and higher record number did not imply in higher potential richness (adjusted R2 = 0.3923). The jitter reveals the high number of pixels with low record numbers that have potential richness from zero to 11 species. ACKNOWLEDGEMENTS This study was supported by Fundação de Amparo à Pesquisa do Espírito Santo – FAPES (grant: 61860808/2013). We also thank the Coordenação de Aperfeiçoamento de Pessoal de Nível Superior – CAPES – and for Fundação de Amparo à Pesquisa do Espírito Santo (FAPES) for the scholarship granted to the first author and fourth authors, respectively; and the Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq) for the grants to M.F. Siqueira (309296/2014-5, 461572/2014-1 and 441929/2016-8) and T.T. Carrijo (305821/2016-4). Finally, we gratefully thank the curators of the RB, MBML, CVRD and VIES herbaria for sending biological collections as loans or donations, making it possible to improve the quality of databases. REFERENCES Allouche O , Tsoar A , Kadmon R . 2006 . Assessing the accuracy of species distribution models: prevalence, kappa and the true skill statistic (TSS) . Journal of Applied Ecology 43 : 1223 – 1232 . Google Scholar CrossRef Search ADS Anderson MJ . 2006 . Distance-based tests for homogeneity of multivariate dispersions . Biometrics 62 : 245 – 253 . Google Scholar CrossRef Search ADS PubMed Araújo MB , Luoto M . 2007 . The importance of biotic interactions for modelling species distributions under climate change . Global Ecology and Biogeography 16 : 743 – 753 . Google Scholar CrossRef Search ADS Araújo M , New M . 2007 . Ensemble forecasting of species distributions . Trends in Ecology & Evolution 22 : 42 – 47 . Google Scholar CrossRef Search ADS PubMed Barbet-Massin M , Jiguet F , Albert CH , Thuiller W . 2012 . Selecting pseudo-absences for species distribution models: how, where and how many ? Methods in Ecology and Evolution 3 : 327 – 338 . Google Scholar CrossRef Search ADS Barve N , Barve V , Jiménez-Valverde A , Lira-Noriega A , Maher SP , Peterson AT , Soberón J . 2011 . The crucial role of the accessible area in ecological niche modeling and species distribution modeling . Ecological Modelling 222 : 1810 – 1819 . Google Scholar CrossRef Search ADS Baselga A . 2010 . Partitioning the turnover and nestedness components of beta diversity . Global Ecology and Biogeography 19 : 134 – 143 . Google Scholar CrossRef Search ADS Baselga A , Orme D , Villeger S , De Bortoli J , Leprieur F . 2017 . betapart: partitioning beta diversity into turnover and nestedness components . Available at: https://cran.r-project.org/web/packages/betapart/betapart.pdf Basso D , Pesarin F , Salmaso L , Solari A . 2009 . Permutation tests for stochastic ordering and ANOVA: theory and applications with R . London NY : Springer Verlag . BFG . 2015 . Growing knowledge: an overview of seed plant diversity in Brazil . Rodriguesia 66 : 1085 – 1113 . CrossRef Search ADS Broennimann O , Cola VD , Guisan A . 2016 . ecospat: spatial ecology miscellaneous methods . R package version 2.2.0. Available at: https://CRAN.R-project.org/package=ecospat. Broennimann O , Fitzpatrick MC , Pearman PB , Petitpierre B , Pellissier L , Yoccoz NG , Thuiller W , Fortin M‐J , Randin C , Zimmermann NE , Graham CH , Guisan A . 2012 . Measuring ecological niche overlap from occurrence and spatial environmental data . Global Ecology and Biogeography 21 : 481 – 497 . Google Scholar CrossRef Search ADS Carnaval AC , Moritz C . 2008 . Historical climate modelling predicts patterns of current biodiversity in the Brazilian Atlantic forest . Journal of Biogeography 35 : 1187 – 1201 . Google Scholar CrossRef Search ADS Casazza G , Marsili S , Granato L , Conti E , Noble V , Selvaggi A , Minuto L . 2013 . Distribution range and ecological niche of Primula marginata Curtis (Primulaceae) . Plant Biosystems – An International Journal Dealing with all Aspects of Plant Biology 147 : 593 – 600 . Google Scholar CrossRef Search ADS Carrijo TT , Tuler AC , Luber J , Costa RG da , Santos MC dos , Paschoa RP da , Freitas M de F , Carrijo TT , Tuler AC , Luber J , Costa RG da , Santos MC dos , Paschoa RP da , Freitas M de F . 2017 . Flora of Espírito Santo: Primulaceae . Rodriguésia 68 : 1829 – 1856 . Google Scholar CrossRef Search ADS Chamberlain S , Ram K , Barve V , Mcglinn D . 2016 . rgbif: interface to the Global Biodiversity Information Facility API . R package version 0.9.4. Available at: https://CRAN.R-project.org/package=rgbif. Crawford RM . 2008 . Plants at the margin: ecological limits and climate change. Cambridge, New York : Cambridge University Press . Google Scholar CrossRef Search ADS De Marco Jr P , de Siqueira MF . 2009 . Como determinar a distribuição potencial de espécies sob uma abordagem conservacionista . Megadiversidade 5 : 65 – 76 . Dias ATC , Scarano FR . 2007 . Clusia as nurse plant . In: Clusia: a woody neotropical genus of remarkable plasticity and diversity . Berlin, Heidelberg : Springer Science & Business Media , 55 – 71 . Google Scholar CrossRef Search ADS Dias ATC , Teixeira Zaluar HL , Ganade G , et al. 2005 . Canopy composition influencing plant patch dynamics in a Brazilian sandy coastal plain . Journal of Tropical Ecology 21 : 343 – 347 . Google Scholar CrossRef Search ADS Duarte HM , Geßler A , Scarano FR , Franco AC , de Mattos EA , Nahm M , Rennenberg H , Rodrigues PJFP , Zaluar HLT , Lüttge U . 2005 . Ecophysiology of six selected shrub species in different plant communities at the periphery of the Atlantic Forest of SE-Brazil . Flora—Morphology, Distribution, Functional Ecology of Plants 200 : 456 – 476 . Google Scholar CrossRef Search ADS Dunlop EA , Wilson JC , Mackey AP . 2006 . The potential geographic distribution of the invasive weed Senna obtusifolia in Australia . Weed Research 46 : 404 – 413 . Google Scholar CrossRef Search ADS Dutra VF , Alves-Araújo A , Carrijo TT . 2015 . Angiosperm checklist of Espírito Santo: using electronic tools to improve the knowledge of an Atlantic Forest biodiversity hotspot . Rodriguésia 66 : 1145 – 1152 . Google Scholar CrossRef Search ADS Flora do Brasil . 2020 under construction. Available at: http://floradobrasil.jbrj.gov.br. Accessed: 22 Aug. 2017 . Freitas M de F , Carrijo TT . 2008 . A família Myrsinaceae nos contrafortes do maciço da Tijuca e entorno do Jardim Botânico do Rio de Janeiro, Brasil . Rodriguésia 59: 813 – 828 . Freitas M de F , Judice DM , Reis Júnior JS , Santos Filho LAF dos , Abreu MB de , Borges RAX . 2013 . Myrsinaceae . In: Martinelli G , Moraes MA , eds. Livro vermelho da flora do Brasil . Rio de Janeiro : CNC Flora , 711 – 713 . Garbin ML , Carrijo TT , Sansevero JBB , Sánchez-Tapia A , Scarano FR . 2012 . Subordinate, not dominant, woody species promote the diversity of climbing plants . Perspectives in Plant Ecology, Evolution and Systematics 14 : 257 – 265 . Google Scholar CrossRef Search ADS Giannini TC , Siqueira MF , Acosta AL , Barreto FC , Saraiva AM , Alves-dos-Santos I . 2012 . Current challenges of species distribution predictive modelling . Rodriguésia 63 : 733 – 749 . Google Scholar CrossRef Search ADS Graham C , Ferrier S , Huettman F , Moritz C , Peterson A . 2004 . New developments in museum-based informatics and applications in biodiversity analysis . Trends in Ecology & Evolution 19 : 497 – 503 . Google Scholar CrossRef Search ADS PubMed Hijmans RJ , Cameron SE , Parra JL , Jones PG , Jarvis A . 2005 . Very high resolution interpolated climate surfaces for global land areas . International Journal of Climatology 25 : 1965 – 1978 . Google Scholar CrossRef Search ADS Hijmans RJ , Phillips S , Leathwick J , Elith J . 2015 . dismo: species distribution modeling. R package version 1.0–12 . Vienna : The R Foundation for Statistical Computing . Available at: http://cran. r-project. org Hirzel AH , Helfer V , Metral F . 2001 . Assessing habitat-suitability models with a virtual species . Ecological Modelling 145 : 111 – 121 . Google Scholar CrossRef Search ADS HYDRO1K | The Long Term Archive . https://lta.cr.usgs.gov/HYDRO1K Jung SL . 1981 . Flora fanerogâmica da Reserva do Parque Estadual das Fontes do Ipiranga (São Paulo, Brasil): 74. Myrsinaceae . Hoehnea 9 : 88 – 91 . Jung-Mendaçolli SL , Bernacci LC . 1997 . Flora fanerogâmica da Ilha do Cardoso (SP, Brasil): Myrsinaceae . In: Melo MMRF , Barros F , Chiea SAC et al. , eds. Flora fanerogâmica da Ilha do Cardoso, São Paulo . São Paulo, SP, Brazil : Instituto de Botânica , 81 – 98 . Jung-Mendaçolli SL , Bernacci LC . 2001 . Myrsinaceae da APA de Cairuçu, Parati (Rio de Janeiro, Brasil) . Rodriguésia : 52: 49 – 64 . Jung-Mendaçolli SL , Bernacci LC , Freitas MF . 2005 . Myrsinaceae . In: Wanderley MGL , Shepherd GJ , Giulietti AM , eds. Flora fanerogâmica do Estado de São Paulo . Brazil: Instituto de Botânica, 279 – 300 . Kamino LHY , de Siqueira MF , Sánchez-Tapia A , et al. 2012 . Reassessment of the extinction risk of endemic species in the Neotropics: how can modelling tools help us ? Natureza & Conservação 10 : 191 – 198 . Google Scholar CrossRef Search ADS Karatzoglou A , Smola A , Hornik K , Zeileis A . 2004 . kernlab—an S4 package for kernel methods in R . Journal of Statistical Software 11 : 1 – 20 . Google Scholar CrossRef Search ADS de Lacerda AEB , Nimmo ER . 2010 . Can we really manage tropical forests without knowing the species within? Getting back to the basics of forest management through taxonomy . Forest Ecology and Management 259 : 995 – 1002 . Google Scholar CrossRef Search ADS Legendre P , Legendre L . 2012 . Numerical ecology . Amsterdam, Oxford : Elsevier . Liaw A , Wiener M . 2002 . Classification and regression by randomForest . R News 2 : 18 – 22 . Lirio EJD , Peixoto AL , De Siqueira MF . 2015 . Taxonomy, conservation, geographic and potential distribution of Macrotorus Perkins (Mollinedioideae, Monimiaceae), and a key to the Neotropical genera of Monimiaceae . Phytotaxa 234 : 201 . Google Scholar CrossRef Search ADS Lobo JM , Tognelli MF . 2011 . Exploring the effects of quantity and location of pseudo-absences and sampling biases on the performance of distribution models with limited point occurrence data . Journal for Nature Conservation 19 : 1 – 7 . Google Scholar CrossRef Search ADS Lütolf M , Kienast F , Guisan A . 2006 . The ghost of past species occurrence: improving species distribution models for presence-only data . Journal of Applied Ecology 43 : 802 – 815 . Google Scholar CrossRef Search ADS McKnight MW , White PS , McDonald RI , Lamoreux JF , Sechrest W , Ridgely RS , Stuart SN . 2007 . Putting beta-diversity on the map: broad-scale congruence and coincidence in the extremes . PLoS Biology 5 : e272 . Google Scholar CrossRef Search ADS PubMed Melo AS , Rangel TFLVB , Diniz-Filho JAF . 2009 . Environmental drivers of beta-diversity patterns in New-World birds and mammals . Ecography 32 : 226 – 236 . Google Scholar CrossRef Search ADS Meyer C , Weigelt P , Kreft H . 2015 . Multidimensional biases, gaps and uncertainties in global plant occurrence information . Ecology Letters 19 : 1 – 15 . Morellato LPC , Haddad CFB . 2000 . Introduction: the Brazilian Atlantic Forest . Biotropica 32 : 786 – 792 . Google Scholar CrossRef Search ADS Murray-Smith C , Brummitt NA , Oliveira-Filho AT , Bachman S , Moat J , Lughadha EMN , Lucas EJ . 2009 . Plant diversity hotspots in the Atlantic coastal forests of Brazil . Conservation Biology 23 : 151 – 163 . Google Scholar CrossRef Search ADS PubMed Myers N , Mittermeier RA , Mittermeier CG , da Fonseca GAB , Kent J . 2000 . Biodiversity hotspots for conservation priorities . Nature 403 : 853 – 858 . Google Scholar CrossRef Search ADS PubMed Oksanen J , Blanchet FG , Friendly M , Kindt R , Legendre P , McGlinn D , Minchin PR , O’Hara RB , Simpson GL , Solymos P , Stevens MHH , Szoecs E , Wagner H . 2016 . vegan: community ecology package R package version 2.4-1. Available at: https://CRAN.R-project.org/package=vegan. Oliveira-Filho AT , Fontes MAL . 2000 . Patterns of floristic differentiation among Atlantic Forests in southeastern Brazil and the influence of climate . Biotropica 32 : 793 – 810 . Google Scholar CrossRef Search ADS Otegui M . 1998 . Sinopsis del género Myrsine L. (Myrsinaceae) en el Cono Sur de América del Sur . Candollea 53 : 133 – 157 . Peterson AT , Soberón J , Pearson RG , Anderson RP , Martínez-Meyer E , Nakamura M . 2011 . Ecological niches and geographic distributions . Princeton : Princeton University Press . Pimm SL , Jenkins CN , Abell R , Brooks TM , Gittleman JL , Joppa LN , Raven PH , Roberts CM , Sexton JO . 2014 . The biodiversity of species and their rates of extinction, distribution, and protection . Science 344 : 1246752 – 1246752 . Google Scholar CrossRef Search ADS PubMed Ponder WF , Carter GA , Flemons P , et al. 2001 . Evaluation of museum collection data for use in biodiversity assessment . Conservation Biology 15 : 648 – 657 . Google Scholar CrossRef Search ADS R Core Team . 2015 . R: a language and environment for statistical computing . Vienna : R Foundation for Statistical Computing . Reflora VH . 2017 . Herbário Virtual REFLORA . Available at: http://reflora.jbrj.gov.br/reflora/herbarioVirtual/. Accessed: 22 Aug. 2017 . Reyer CP , Leuzinger S , Rammig A , Wolf A, Bartholomeus RP , Bonfante A , de Lorenzi F , Dury M , Gloning P , Abou Jaoudé R , Klein R , Kuster TM , Martins M , Niedrist G , Riccardi M , Wohlfahrt G , de Angelis P , de Dato G , François L , Menzel A , Pereira M . 2013 . A plant’s perspective of extremes: terrestrial plant responses to changing climatic variability . Global Change Biology 19 : 75 – 89 . Google Scholar CrossRef Search ADS PubMed Rezende VL , de Oliveira-Filho AT , Eisenlohr PV , Kamino LHY , Vibrans AC . 2014 . Restricted geographic distribution of tree species calls for urgent conservation efforts in the Subtropical Atlantic Forest . Biodiversity and Conservation 24 : 1057 – 1071 . Google Scholar CrossRef Search ADS Ribeiro MC , Metzger JP , Martensen AC , Ponzoni FJ , Hirota MM . 2009 . The Brazilian Atlantic Forest: how much is left, and how is the remaining forest distributed? Implications for conservation . Biological Conservation 142 : 1141 – 1153 . Google Scholar CrossRef Search ADS Scarano FR . 2002 . Structure, function and floristic relationships of plant communities in stressful habitats marginal to the Brazilian Atlantic Rainforest . Annals of Botany 90 : 517 – 524 . Google Scholar CrossRef Search ADS PubMed Scarano FR . 2009 . Plant communities at the periphery of the Atlantic rain forest: rare-species bias and its risks for conservation . Biological Conservation 142 : 1201 – 1208 . Google Scholar CrossRef Search ADS Scarano FR , Ceotto P . 2015 . Brazilian Atlantic forest: impact, vulnerability, and adaptation to climate change . Biodiversity and Conservation 24 : 2319 – 2331 . Google Scholar CrossRef Search ADS Scarano FR , Duarte HM , Franco AC , Geβler A , Mattos EA de , Rennenberg H , Lüttge U . 2005 . Physiological synecology of tree species in relation to geographic distribution and ecophysiological parameters at the Atlantic forest periphery in Brazil: an overview . Trees 19 : 493 – 496 . Google Scholar CrossRef Search ADS Scarano FR , Duarte HM , Ribeiro KT , Rodrigues PJFP , Barcellos EMB , Franco AC , Brulfert J , Deléens E , Lüttge U . 2001 . Four sites with contrasting environmental stress in southeastern Brazil: relations of species, life form diversity, and geographic distribution to ecophysiological parameters . Botanical Journal of the Linnean Society 136 : 345 – 364 . Google Scholar CrossRef Search ADS da Silva JMC , Casteleti CHM . 2003 . Status of the biodiversity of the Atlantic Forest of Brazil . In: Leal CG , Câmara I de G , eds. The Atlantic Forest of South America: biodiversity status, threats, and outlook . Washington : Island Press , 43 – 59 . Siqueira JC . 1987 . Considerações taxonômicas sobre as espécies do gênero Rapanea Aublet (Myrsinaceae) ocorrentes no Rio Grande do Sul . Pesquisas Botânica 38 : 147 – 156 . Siqueira JC . 1993 . O gênero Rapanea Aublet (Myrsinaceae) na região serrana do estado do Rio de Janeiro: aspectos taxonômicos e ecológicos das espécies . Pesquisas Botânicas 44: 41 – 52 . de Siqueira MF , Durigan G . 2007 . Modelagem da distribuição geográfica de espécies lenhosas de cerrado no Estado de São Paulo . Revista Brasileira de Botânica 30 : 233 – 243 . de Siqueira MF , Durigan G , De Marco Júnior P , et al. 2009 . Something from nothing: using landscape similarity and ecological niche modeling to find rare plant species . Journal for Nature Conservation 17 : 25 – 32 . Google Scholar CrossRef Search ADS Slatyer RA , Hirst M , Sexton JP . 2013 . Niche breadth predicts geographical range size: a general ecological pattern . Ecology Letters 16 : 1104 – 1114 . Google Scholar CrossRef Search ADS PubMed Socolar JB , Gilroy JJ , Kunin WE , et al. 2016 . How should beta-diversity inform biodiversity conservation ? Trends in Ecology & Evolution 31 : 67 – 80 . Google Scholar CrossRef Search ADS PubMed Sousa-Baena MS , Garcia LC , Peterson AT . 2014 . Completeness of digital accessible knowledge of the plants of Brazil and priorities for survey and inventory . Diversity and Distributions 20 : 369 – 381 . Google Scholar CrossRef Search ADS speciesLink . 2017 . speciesLink: sistema de informação distribuído para coleções biológicas . Available at: http://splink.cria.org.br/ Accessed: 27 Nov. 2017 . Stokland JN , Halvorsen R , Støa B . 2011 . Species distribution modelling – effect of design and sample size of pseudo-absence observations . Ecological Modelling 222 : 1800 – 1809 . Google Scholar CrossRef Search ADS Taylor S , Kumar L . 2016 . Will climate change impact the potential distribution of a native vine (Merremia peltata) which is behaving invasively in the Pacific region ? Ecology and Evolution 6 : 742 – 754 . Google Scholar CrossRef Search ADS PubMed Taylor S , Kumar L , Reid N , et al. 2012 . Climate change and the potential distribution of an invasive shrub, Lantana camara L . PLoS One 7 : e35565 . Google Scholar CrossRef Search ADS PubMed Valiente-Banuet A , Rumebe AV , Verdú M , et al. 2006 . Modern Quaternary plant lineages promote diversity through facilitation of ancient Tertiary lineages . Proceedings of the National Academy of Sciences 103 : 16812 – 16817 . Google Scholar CrossRef Search ADS Vargas JH , Consiglio T , Jørgensen PM , et al. 2004 . Modelling distribution patterns in a species-rich plant genus, Anthurium (Araceae), in Ecuador . Diversity and Distributions 10 : 211 – 216 . Google Scholar CrossRef Search ADS Warren DL , Glor RE , Turelli M . 2008 . Environmental niche equivalency versus conservatism: quantitative approaches to niche evolution . Evolution 62 : 2868 – 2883 . Google Scholar CrossRef Search ADS PubMed Werneck M de S , Sobral MEG , Rocha CTV , Landau EC , Stehmann JR . 2011 . Distribution and endemism of angiosperms in the Atlantic Forest . Natureza & Conservação 9 : 188 – 193 . Google Scholar CrossRef Search ADS © 2018 The Linnean Society of London, Botanical Journal of the Linnean Society This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/open_access/funder_policies/chorus/standard_publication_model)

Kim, Jung Sung;Kim, Joo-Hwan

doi: 10.1093/botlinnean/boy031pmid: N/A

Abstract Liliaceae sensu APG IV include c. 600 species, and have been circumscribed into three subfamilies, Lilioideae, Calochortoideae and Streptopoideae. Molecular phylogenetic approaches to this family have produced dynamic changes in the generic circumscriptions. We conducted molecular phylogenetic analyses, time estimations and biogeographical analyses to confirm generic relationships, discuss their circumscription for classification and clarify the evolutionary history of the family. A phylogenetic analysis is presented as a combined data set of c. 6.1 kb from four plastid coding gene sequences for 142 taxa representing all the genera of Liliaceae. Medeoloideae were newly defined as an independent subfamily from Lilioideae. Tricyrtis of Calochortoideae is embedded in Streptopoideae, despite its unique characteristics. The crown age of Liliaceae was calculated at c. 85 Mya, and Liliaceae are considered to have originated in temperate Asia in the late Cretaceous and to have expanded their distribution via dispersal with the occurrence of repetitive vicariance events during their evolution. distribution, Medeoloideae, molecular dating, molecular phylogeny, subfamilial circumscription, Tricyrtis INTRODUCTION Liliaceae are mainly distributed in the Northern Hemisphere in the temperate zones of eastern Asia and North America, and include many ornamental plants widely grown for their attractive flowers. In APG IV (2016), this family includes 15 genera and > 600 species. They belong to the petaloid monocots and are generally characterized as herbaceous with a bulb storage organ or rhizome and flowers with six petaloid tepals, six stamens and a superior ovary. In particular, their large and colourful flowers are of commercial, pharmaceutical and ethnobotanical interest. Taxonomically, Liliaceae have been used in a wider range of circumscriptions than any other family (Fay & Chase, 2000), and the classification of the family has changed dramatically based on modern molecular phylogenetic studies (Table 1). In the past, taxonomic confusion about the broad sense of the family severely hampered our understanding of the diversity of Liliaceae. Table 1. Comparison of the classfication systems for Liliaceae Bentham & Hooker (1883) Engler (1888) Hutchinson (1959) Dahlgren et al. (1985) Takhtajan (1997) Tamura (1998) Patterson & Givnish (2002) Liliiflorae Liliales Liliales Liliales Liliales Liliales Liliaceae Liliaceae Liliaceae Liliaceae Liliaceae Liliaceae Liliaceae Subfamily Lilioideae Subfamily Lilioideae Subfamily Lilioideae 1. Tribe Tulipeae 1. Tribe Tulipeae 1. Tribe Tulipeae 1. Tribe Lilieae 1. Tribe Lilieae Fritillaria Fritillaria Fritillaria Fritillaria Fritillaria Fritillaria Fritillaria Lilium Lilium Lilium Lilium Lilium Lilium Lilium Calochortus Cardiocrinum Cardiocrinum Cardiocrinum Cardiocrinum Cardiocrinum Notholirion Notholirion Notholirion Notholirion Notholirion Nomocharis Nomocharis Nomocharis Nomocharis Calochortus Calochortus 2. Tribe Tulipeae 2. Tribe Tulipeae Tulipa Tulipa Tulipa Tulipa Tulipa Tulipa (incl. Amana) Tulipa Erythronium Erythronium Erythronium Erythronium Erythronium Erythronium Erythronium Amana Amana 3. Tribe Lloydieae Amana Lloydia Lloydia Lloydia Lloydia Lloydia Lloydia Lloydia Subfamily Allioideae 1. Tribe Allieeae Gagea Gagea Gagea Gagea Gagea Gagea Gagea Subfamily Asparagoideae 2. Tribe Medeoleae 1. Tribe Parideae Uvulariaceae Medeolaceae Subfamily Medeoloideae Subfamily Medeoloideae Medeola Medeola Medeola Medeola Medeola Medeola Scoliopus Scoliopus Scoliopus Colchicales 2. Tribe Polygonateae 2. Tribe Polygonateae Uvulariaceae Clintonia Clintonia Clintonia Clintonia Clintonia Clintonia Clintonia Calochortaceae 1. Tribe Tricyrtideae Subfamily Streptopoideae 3. Tribe Polygonateae Prosartes Prosartes Prosartes Prosartes Prosartes Streptopus Streptopus Streptopus Streptopus Streptopus Streptopus Streptopus Scoliopaceae Scoliopus Scoliopus Scoliopus Subfamily Melanthioideae 4. Tribe Uvularieae 1. Tribe Uvularieae 3. Tribe Tricyrtideae Tricyrtidaceae Subfamily Calochortoideae Tricyrtis Tricyrtis Tricyrtis Tricyrtis Tricyrtis Tricyrtis Tricyrtis Trilliaceae Calochortaceae Calochortaceae 2. Tribe Calochorteae Scoliopus Calochortus Calochortus Calochortus Calochortus Bentham & Hooker (1883) Engler (1888) Hutchinson (1959) Dahlgren et al. (1985) Takhtajan (1997) Tamura (1998) Patterson & Givnish (2002) Liliiflorae Liliales Liliales Liliales Liliales Liliales Liliaceae Liliaceae Liliaceae Liliaceae Liliaceae Liliaceae Liliaceae Subfamily Lilioideae Subfamily Lilioideae Subfamily Lilioideae 1. Tribe Tulipeae 1. Tribe Tulipeae 1. Tribe Tulipeae 1. Tribe Lilieae 1. Tribe Lilieae Fritillaria Fritillaria Fritillaria Fritillaria Fritillaria Fritillaria Fritillaria Lilium Lilium Lilium Lilium Lilium Lilium Lilium Calochortus Cardiocrinum Cardiocrinum Cardiocrinum Cardiocrinum Cardiocrinum Notholirion Notholirion Notholirion Notholirion Notholirion Nomocharis Nomocharis Nomocharis Nomocharis Calochortus Calochortus 2. Tribe Tulipeae 2. Tribe Tulipeae Tulipa Tulipa Tulipa Tulipa Tulipa Tulipa (incl. Amana) Tulipa Erythronium Erythronium Erythronium Erythronium Erythronium Erythronium Erythronium Amana Amana 3. Tribe Lloydieae Amana Lloydia Lloydia Lloydia Lloydia Lloydia Lloydia Lloydia Subfamily Allioideae 1. Tribe Allieeae Gagea Gagea Gagea Gagea Gagea Gagea Gagea Subfamily Asparagoideae 2. Tribe Medeoleae 1. Tribe Parideae Uvulariaceae Medeolaceae Subfamily Medeoloideae Subfamily Medeoloideae Medeola Medeola Medeola Medeola Medeola Medeola Scoliopus Scoliopus Scoliopus Colchicales 2. Tribe Polygonateae 2. Tribe Polygonateae Uvulariaceae Clintonia Clintonia Clintonia Clintonia Clintonia Clintonia Clintonia Calochortaceae 1. Tribe Tricyrtideae Subfamily Streptopoideae 3. Tribe Polygonateae Prosartes Prosartes Prosartes Prosartes Prosartes Streptopus Streptopus Streptopus Streptopus Streptopus Streptopus Streptopus Scoliopaceae Scoliopus Scoliopus Scoliopus Subfamily Melanthioideae 4. Tribe Uvularieae 1. Tribe Uvularieae 3. Tribe Tricyrtideae Tricyrtidaceae Subfamily Calochortoideae Tricyrtis Tricyrtis Tricyrtis Tricyrtis Tricyrtis Tricyrtis Tricyrtis Trilliaceae Calochortaceae Calochortaceae 2. Tribe Calochorteae Scoliopus Calochortus Calochortus Calochortus Calochortus View Large Table 1. Comparison of the classfication systems for Liliaceae Bentham & Hooker (1883) Engler (1888) Hutchinson (1959) Dahlgren et al. (1985) Takhtajan (1997) Tamura (1998) Patterson & Givnish (2002) Liliiflorae Liliales Liliales Liliales Liliales Liliales Liliaceae Liliaceae Liliaceae Liliaceae Liliaceae Liliaceae Liliaceae Subfamily Lilioideae Subfamily Lilioideae Subfamily Lilioideae 1. Tribe Tulipeae 1. Tribe Tulipeae 1. Tribe Tulipeae 1. Tribe Lilieae 1. Tribe Lilieae Fritillaria Fritillaria Fritillaria Fritillaria Fritillaria Fritillaria Fritillaria Lilium Lilium Lilium Lilium Lilium Lilium Lilium Calochortus Cardiocrinum Cardiocrinum Cardiocrinum Cardiocrinum Cardiocrinum Notholirion Notholirion Notholirion Notholirion Notholirion Nomocharis Nomocharis Nomocharis Nomocharis Calochortus Calochortus 2. Tribe Tulipeae 2. Tribe Tulipeae Tulipa Tulipa Tulipa Tulipa Tulipa Tulipa (incl. Amana) Tulipa Erythronium Erythronium Erythronium Erythronium Erythronium Erythronium Erythronium Amana Amana 3. Tribe Lloydieae Amana Lloydia Lloydia Lloydia Lloydia Lloydia Lloydia Lloydia Subfamily Allioideae 1. Tribe Allieeae Gagea Gagea Gagea Gagea Gagea Gagea Gagea Subfamily Asparagoideae 2. Tribe Medeoleae 1. Tribe Parideae Uvulariaceae Medeolaceae Subfamily Medeoloideae Subfamily Medeoloideae Medeola Medeola Medeola Medeola Medeola Medeola Scoliopus Scoliopus Scoliopus Colchicales 2. Tribe Polygonateae 2. Tribe Polygonateae Uvulariaceae Clintonia Clintonia Clintonia Clintonia Clintonia Clintonia Clintonia Calochortaceae 1. Tribe Tricyrtideae Subfamily Streptopoideae 3. Tribe Polygonateae Prosartes Prosartes Prosartes Prosartes Prosartes Streptopus Streptopus Streptopus Streptopus Streptopus Streptopus Streptopus Scoliopaceae Scoliopus Scoliopus Scoliopus Subfamily Melanthioideae 4. Tribe Uvularieae 1. Tribe Uvularieae 3. Tribe Tricyrtideae Tricyrtidaceae Subfamily Calochortoideae Tricyrtis Tricyrtis Tricyrtis Tricyrtis Tricyrtis Tricyrtis Tricyrtis Trilliaceae Calochortaceae Calochortaceae 2. Tribe Calochorteae Scoliopus Calochortus Calochortus Calochortus Calochortus Bentham & Hooker (1883) Engler (1888) Hutchinson (1959) Dahlgren et al. (1985) Takhtajan (1997) Tamura (1998) Patterson & Givnish (2002) Liliiflorae Liliales Liliales Liliales Liliales Liliales Liliaceae Liliaceae Liliaceae Liliaceae Liliaceae Liliaceae Liliaceae Subfamily Lilioideae Subfamily Lilioideae Subfamily Lilioideae 1. Tribe Tulipeae 1. Tribe Tulipeae 1. Tribe Tulipeae 1. Tribe Lilieae 1. Tribe Lilieae Fritillaria Fritillaria Fritillaria Fritillaria Fritillaria Fritillaria Fritillaria Lilium Lilium Lilium Lilium Lilium Lilium Lilium Calochortus Cardiocrinum Cardiocrinum Cardiocrinum Cardiocrinum Cardiocrinum Notholirion Notholirion Notholirion Notholirion Notholirion Nomocharis Nomocharis Nomocharis Nomocharis Calochortus Calochortus 2. Tribe Tulipeae 2. Tribe Tulipeae Tulipa Tulipa Tulipa Tulipa Tulipa Tulipa (incl. Amana) Tulipa Erythronium Erythronium Erythronium Erythronium Erythronium Erythronium Erythronium Amana Amana 3. Tribe Lloydieae Amana Lloydia Lloydia Lloydia Lloydia Lloydia Lloydia Lloydia Subfamily Allioideae 1. Tribe Allieeae Gagea Gagea Gagea Gagea Gagea Gagea Gagea Subfamily Asparagoideae 2. Tribe Medeoleae 1. Tribe Parideae Uvulariaceae Medeolaceae Subfamily Medeoloideae Subfamily Medeoloideae Medeola Medeola Medeola Medeola Medeola Medeola Scoliopus Scoliopus Scoliopus Colchicales 2. Tribe Polygonateae 2. Tribe Polygonateae Uvulariaceae Clintonia Clintonia Clintonia Clintonia Clintonia Clintonia Clintonia Calochortaceae 1. Tribe Tricyrtideae Subfamily Streptopoideae 3. Tribe Polygonateae Prosartes Prosartes Prosartes Prosartes Prosartes Streptopus Streptopus Streptopus Streptopus Streptopus Streptopus Streptopus Scoliopaceae Scoliopus Scoliopus Scoliopus Subfamily Melanthioideae 4. Tribe Uvularieae 1. Tribe Uvularieae 3. Tribe Tricyrtideae Tricyrtidaceae Subfamily Calochortoideae Tricyrtis Tricyrtis Tricyrtis Tricyrtis Tricyrtis Tricyrtis Tricyrtis Trilliaceae Calochortaceae Calochortaceae 2. Tribe Calochorteae Scoliopus Calochortus Calochortus Calochortus Calochortus View Large From initially broad and diverse viewpoints (Bentham & Hooker, 1883; Engler, 1888; Hutchinson, 1959; Dahlgren, Clifford & Yeo, 1985), classification systems for the family have varied based on morphological, karyological, embryological and molecular phylogenetic studies. Beginning with the first molecular phylogenetic study on the superorder Lilianae (Chase et al., 1995), the contemporary classification of monocots has been rearranged on the basis of their phylogenetic relationships. The narrow concept of the order Liliales was also established, including the infrafamilial classification of larger families. However, in Liliaceae, the taxonomic rank of tribe Medeoleae in subfamily Lilioideae and the circumscription of Calochortoideae and Streptopoideae has been long debated due to disagreement about the phylogenetic position of Calochortus Pursh and Tricyrtis Wallich (Table 1). The monophyly of Calochortoideae has been a highly controversial problem in Liliaceae. Liliaceae were sometimes positioned in the same clade to support their monophyly (Patterson & Givnish, 2002; Petersen, Seberg & Davis, 2013) or divided into different clades with other subfamily groups. In other molecular phylogenetic studies, Calochortus was sister to the rest of the family (Vinnersten & Bremer, 2001) or showed an independent and closer relationship with Streptopoideae against the sister relationship of Tricyrtis and Lilioideae (Fay et al., 2006). However, Kim et al. (2013) proposed a different subfamilial circumscription based on a molecular phylogenetic study using a broad sampling of Liliales that expanded the Streptopoideae to include Tricyrtis, with Calochortoideae consisting of only Calochortus. In their study, the sister relationship between Lilioideae and Calochortoideae, excluding Tricyrtis, was moderately well supported. Lilioideae, the most widely studied and well-defined subfamily, include several genera with expanded circumscriptions that need to be revised according to their phylogenetic relationships. Although the monophyly of Fritillaria L. and Lilium L. (including Nomocharis Franchet) as questioned by Hayashi and Kawano (2000) has been resolved by numerous molecular phylogenetic approaches, conflict remains between the infrageneric circumscriptions and their phylogenetic relationships (Nishikawa et al., 1999, 2001; Ronsted et al., 2005; Gao et al., 2012, 2013; Day et al., 2014). Nomocharis (seven species) was long recognized as the intermediate genus between Lilium and Fritillaria, but is now nested in Lilium (e.g. Gao et al., 2012). Gagea Salisb. and Lloydia Salisb. ex Rchb., based on molecular phylogenetic analyses (e.g. Peterson et al., 2004), are now treated as a single genus, Gagea, composed of 14 sections including section Lloydia (Salisb. ex Rchb.) Peruzzi, J.-M.Tison, A.Peterson & J.Peterson (Peterson, Levichev & Peterson, 2008; Zarrei et al., 2009; Peruzzi, 2011), although the monophyly of each section and sectional relationships remain unclear. The phylogenetic relationships among Tulipa L., Amana Honda and Erythronium L. have also been controversial, due to a lack of strong support (e.g. Allen, Soltis & Soltis, 2003; Petersen et al., 2013). The phylogenetic position of Amana has been an issue for understanding the relationships among the three genera, although it was often regarded as a member of Tulipa (Tamura, 1998). In recent molecular phylogenetic studies, Amana was sister to the remaining two genera (Christenhusz et al., 2013) or was closer to Erythronium than to Tulipa (Clennett et al., 2012; Kim et al., 2013); Tulipa was not monophyletic in the latter case. The taxonomic treatment of Amana and its phylogenetic relationship remain to be clarified in the phylogenetics of Liliaceae. Molecular phylogenetic analysis is widely used for calculating divergence times and understanding the distributional diversities of plant groups in their evolutionary histories. The Cretaceous origin of the monocots has been clearly established from the results of molecular phylogenetics (Bremer, 2000; Janssen & Bremer, 2004; Hertweck et al., 2015) and from the fossil record (Daghlian, 1981). Vinnersten & Bremer (2001) estimated the divergence time of Liliales and recognized four main clades back to 65 Mya: Campynemataceae, Melanthiaceae, Smilacaceae+Liliaceae, and Alstroemeriaceae (including Luzuriagaceae)+Colchicaceae. However, their tree topology differed from recent molecular phylogenetic studies. Fossil records of Liliales have been reported from restricted taxa for Luzuriaga Ruiz & Pav. (Conran et al., 2014), Ripogonum J.R.Forst. & G.Forst. (Conran, Carpenter & Jordan, 2009; Carpenter et al., 2014) and Smilax L. (Logan, Smiley & Eglinton, 1995; Wilde & Frankenhauser, 1998; Wilf, 2000; Ding et al., 2011). Recently, Iles et al. (2016) proposed 34 fossils for calibrating the age of major monocot clades and suggested that the fossils of Luzuriaga peterbannisteri Conran, Bannister, Mildenh. & D.E.Lee and Ripogonum tasmanicum Conran, R.G.Carp. & G.J.Jord. are suitable for Liliales. Estimated divergence time and biogeography have been explored at the genus and family level for Liliales [Lilium (Gao et al., 2013), Smilax (Zhao et al., 2013; Chen et al., 2014), Alstroemeriaceae (Chacon et al., 2012), Colchicaceae (Chacon & Renner, 2014) and Corsiaceae (Mennes et al., 2015)]. Givnish et al. (2016) provided a phylogenomic and biogeographical history of Liliales. However, the weak and moderate support values in the phylogenetic trees raised led to uncertainty regarding relationships in Liliaceae and among those families including climbers. We rebuilt a molecular phylogenetic tree for Liliaceae using expanded samples including representatives of the major groups without taxonomic gaps based on four plastid DNA gene sequences. Our results provide new insights into the phylogenetic relationships of the family with strong support and the circumscription of the subfamilies and genera. For a comprehensive understanding of the evolutionary history of the family, we also conducted dating and biogeographical analysis. MATERIAL AND METHODS Plant materials We included 142 taxa, representing all three subfamilies and all genera of the family and 11 outgroup species (Supporting Information Table S1), obtained through different methods: (1) fresh leaves directly collected on field trips in South Korea, China and Taiwan; (2) DNA materials from the DNA Bank of the Royal Botanic Gardens, Kew; and (3) herbarium specimen samples collected from the Herbarium of the Kumming Institute of Botany (KIB), Harvard University Herbaria (HUH), Kyoto University Herbarium (KYO), Herbarium of the University of Tokyo (TI) and the National Museum of Natural Science Botanical Garden (Taichung, Taiwan). All plant materials used in this study were collected through the KNRRC (Medicinal Plants Resources Bank NRF-2010-0005790) supported by the Korea Research Foundation, where resources were provided by the Ministry of Education, Science and Technology in 2014. Species names followed the World Checklist of Selected Plant Families (http://apps.kew.org/wcsp/home.do) and the International Plant Name Index (http://www.ipni.org/ipni/plantnamesearchpage.do). DNA sequencing and phylogenetic analysis Total genomic DNA was extracted from 0.2–1.5 g of fresh or silica gel-dried leaves (Chase & Hills, 1991) using the CTAB (cetyltrimethylammonium bromide) buffer method (Doyle & Doyle, 1987). Lipids were removed with SEVAG (24:1 chloroform/isoamyl alcohol) and stored at −20 °C until use after dissolving in 1× TE buffer. The DNA concentration was determined with a UV-Vis Spectrophotometer (BioSpec-nano/0.7 mm, Shimadzu Corp.). For the molecular phylogenetic analysis, four plastid genes (rbcL, matK, ndhF and atpB) were amplified with the GenAmp PCR System 9700 (Life Technologies Co.) and Solg e-Taq DNA polymerase (SolGent Co., Ltd). Details of the primers used to amplify the four genes are listed in Table S2. For the matK gene region, PCR was profiled as a 35-cycle reaction with denaturation at 94 °C for 1 min, annealing at 53 °C for 1 min and extension at 72 °C for 1–2 min, in addition to initial denaturation at 94 °C for 3 min and final extension at 72 °C for 7 min. To amplify the rbcL, ndhF and atpB genes, we followed the reaction conditions of Shinwari et al. (1994a), Olmstead et al. (2000) and Hoot, Culham & Crane (1995), respectively. All PCR products were purified using the QIAquick PCR purification kit (Qiagen Inc.) and cycle sequencing reactions were performed using the BigDye Terminator v3.1 Cycle Sequencing Kit (Perkin Elmer Applied Biosystem) according to the manufacturer’s protocol. Sequence editing and assembly of contigs were performed using Sequencher (ver. 4.10). For sequence alignment, we used MUSCLE embedded in the Geneious program ver. 7.1.7 (Kearse et al., 2012) and made manual adjustments following the guidelines of Kelchner (2000). From the aligned sequences of 161 taxa, we obtained two phylogenetic trees through maximum likelihood (ML) and Bayesian analyses. For both analyses, nucleotide substitution model parameters were determined using jModelTest ver. 0.1.1 (Posada, 2008) and TVM+I+G was selected as the most suitable for the datasets (nst = 6, rates = gamma). We used the RAxML Blackbox program (Stamatakis, Hoover & Rougemont, 2008) through the relevant web-server (http://embnet.vital-it.ch/raxml-bb/) for making the ML tree searches with 1000 replicates of bootstrap analysis. Bayesian analysis was performed using MrBayes ver. 3.1.2 (Ronquist & Huelsenbeck, 2003). For the analyses, two simultaneous runs were performed starting from random trees for 5 × 106 generations, and Markov chains were sampled every 1000 generations. In total, 25% of the total trees were discarded as burn-in samples. The trees were illustrated using the FigTree v1.4.0 program (http://tree.bio.ed.ac.uk/software/figtree/) with supporting values. Divergence time estimation On the basis of the combined data set, we estimated the divergence time of Liliaceae using a relaxed clock log normal implemented in BEAST 2.1.3 (Drummond et al., 2006, 2012; Bouckaert et al., 2014). The substitution model TVM+I+G was determined as the best model under Akaike’s information criterion (AIC) in jModel TEST and GTR+I+G was used. A Yule process of speciation was specified as a tree prior in the analysis. Posterior distributions of internal nodes were calculated by Markov chain Monte Carlo (MCMC) analyses of 100 million generations with 10% burn-in. Sampling adequacy and convergence of the chains were checked using Tracer 1.6, and maximum clade credibility (MCC) chronograms were visualized using FigTree 1.4.0. Calibration point priors were modelled as log normal distributions with three dated fossils for Liliales used as offset values. The mean and standard deviation of these distributions were set to 1.0 and 1.25, respectively, and a 95% confidence interval spanning approximately 10 My. On the basis of leaf microfossils of R. tasmanicum (Conran et al., 2009; Carpenter et al., 2014), the stem node of the Ripogonaceae was set to 51 Mya. On the basis of fossil leaves of L. peterbannisteri (Conran et al., 2014), the stem node of Luzuriaga was set to 23.2 Mya. For the crown node of Smilax, different calibration records from this genus were reported as Late Palaeocene–Early Eocene (48.6–55.8 Mya; Wilf, 2000), and Mid-Eocene (37.2–48.6 Mya; Wilde & Frankenhauser, 1998); therefore, we set the mean age of the older records to 46.5 Mya, even though younger fossil records have also been reported from China (Ding et al., 2011). Despite debates about the crown age of Liliales [43–105 Mya: Mennes et al. (2013); 82 Mya: Vinnersten & Bremer (2001); 117 Mya: Janssen & Bremer (2004); 121 Mya: Hertweck et al. (2015)], the root age of the tree was constrained to 90 Mya on the basis of the most recent molecular clock test estimate (Mennes et al., 2015). Ancestral area reconstruction The biogeographical history of Liliaceae was inferred from the analysis of the ancestral area reconstruction using Bayesian binary MCMC (BBM) implemented in the Reconstruct Ancestral States in Phylogenies (RASP) program ver. 3.02 (Yu, Harris & He, 2010; Yu et al., 2015) using the trees retained from the BEAST analysis. Seven distributional areas were defined for Liliaceae and outgroups: Africa (A); Asia-Temperate (B); Asia-Tropical (C); Europe (D); South America (E); North America (F); and Australasia (G). We applied ten chains, optimized the fixed JC+G model for 500000 cycles, sampled the posterior substitution every 100 generations and allowed for a maximum of four areas for constraining the ancestral ranges. RESULTS Alignment and phylogenetic analysis The combined data matrix was composed of 6152 aligned sites, of which 4184 were constant and 1419 were potentially parsimony-informative. In contrast to rbcL and atpB, length variations among the taxa were detected in matK and ndhF for Liliaceae. These variable regions included five genus-specific indels: a 7-bp insertion (ATCTATT) in Clintonia Raf., a 6-bp insertion (ATATAG) in Notholirion Boiss. and a 6-bp shared deletion (TTCTTT) in Streptopus Michx., Scoliopus Torrey and Prosartes D.Don in matK; and a 3-bp shared insertion (ATA) in Tulipa and Erythronium and a 6-bp shared deletion (GATCAA) in Gagea (including Lloydia) in ndhF. The topology was congruent between the ML and Bayesian trees, which were generated using the combined data matrix (Fig. 1). The monophyly of Liliaceae was supported in both trees, and Liliaceae and Smilacaceae were sister groups, with moderate support. In Liliaceae, three major clades were recognized: (1) Streptopoideae + Tricyrtis [bootstrap value (BP) 93, posterior probability (PP) 0.99]; (2) Calochortus alone (BP 100, PP 1.0); and (3) Lilioideae (BP 100, PP 1.0). Members of Calochortoideae in the traditional sense were found in two clades with strong support: Calochortus was sister to Lilioideae but not to the Streptopoideae-Tricyrtis clade (BP 86, PP 0.98). In Lilioideae, Clintonia-Medeola were sister to the remaining genera, which were divided into three clades: (1) Tulipa-Erythronium-Amana; (2) Gagea (including Lloydia); and (3) Fritillaria-Lilium (including Nomocharis)-Cardiocrinum-Notholirion. All genera received strong support for their monophyly. Figure 1. View largeDownload slide View largeDownload slide Maximum likelihood tree for Liliaceae generated from four plastid gene sequences with infrafamilial classification. An asterisk on a branch indicates that it was supported by 100% bootstrap value (BP) in the RAxML trees and a Bayesian posterior probability of 1.00. Abbreviations of subfamilies: Str, Streptopoideae; Cal, Calochortoideae; Med, Medeoloideae; Lil, Lilioideae. Figure 1. View largeDownload slide View largeDownload slide Maximum likelihood tree for Liliaceae generated from four plastid gene sequences with infrafamilial classification. An asterisk on a branch indicates that it was supported by 100% bootstrap value (BP) in the RAxML trees and a Bayesian posterior probability of 1.00. Abbreviations of subfamilies: Str, Streptopoideae; Cal, Calochortoideae; Med, Medeoloideae; Lil, Lilioideae. Different settings of sister groups for Liliaceae gave slightly different support values for some clades in the same tree topology (Fig. S1). Because the relationship between Smilacaceae and Liliaceae was moderately well supported when three related families (Smilacaceae, Ripogonaceae and Philesiaceae) were included in the analysis as outgroups, we compared the support values when excluding each sister group member. The support values of the inner clades in Liliaceae were generally higher in the tree generated with Smilacaceae as the outgroup than that with Rhipogonaceae+Philesiaceae as the outgroup. From the results of this updated molecular phylogenetic analysis, we suggest that four subfamilies should be recognized: Lilioideae; Medeoloideae (composed of Medeola and Clintonia); Calochortoideae (composed of just Calochortus); and Streptopoideae (including Tricyrtis). We use this classification to interpret the results of the dating and biogeographical analyses. Molecular dating and biogeographical reconstruction The BEAST-derived chronogram for Liliaceae based on the combined sequence data is presented in Figure 2. We designated eight main nodes for the family (nodes A–H in Fig. 2 and Table 2) and a node for each genus (nodes 1–12). From this result, we estimated the age of the crown group of Liliaceae (node A) at c. 85.12 Mya [95% highest posterior density (HPD), 64.66–91.14 Mya, Table 2]. At the subfamilial level, we estimated divergence times of the crown groups of 54.78, 61.81, 46.15 and 23.03 Mya for Streptopoideae, Lilioideae, Medeoloideae and Calochortoideae, respectively (see Discussion for further details). The divergence time of each genus was also estimated (nodes 1–12 in Fig. 2 and Table 2). In Streptopoideae, the crown age of Tricyrtis was c. 14.95 Mya (node 12, 95% HPD, 5.71–23.31 Mya), and the others showed younger divergence ages. In Lilioideae, the crown age of Gagea was the oldest among the genera (node 5, 44.24 Mya, HPD 23.32–43.45 Mya). Fritillaria and Lilium were estimated to have originated at 28.12 and 27.96 Mya (nodes 1 and 2), respectively. Table 2. Summary of divergence time estimation results under a Bayeasian approach as implented in BEAST based on the combined plastid DNA sequence data Node Calibration point discription Mean age (Mya) 95% HPD A Family Liliaceae crown node 85.12 64.66–91.14 B Subfamily Streptopoideae crown node 54.78 36.62–83.53 C divergence of subfamilies Calochortoideae and Lilioideae 82.95 61.66–90.4 D Subfamily Lilioideae crown node 66.36 49.5–81.39 E Subfamily Medeoloideae crown node 46.15 22.52–58.94 F Subfamily Lilioideae crown node 61.81 44.8–74.6 G Tribe Tulipeae crown node 57.63 39.48–66.86 H Tribe Lilieae crown node 42.89 24.51–50.21 1 Genus Lilium (+Nomocharis) crown node 27.96 12.85–28.89 2 Genus Fritillaria crown node 28.12 15.58–32.42 3 Genus Cardiocrinum crown node 16.53 1.55–18.17 4 Genus Notholirion crown node 21.17 4.56–24.25 5 Genus Gagea (+Lloydia) crown node 44.24 23.32–43.45 6 Genus Tulipa crown node 20.74 10.2–23.99 7 Genus Erythronium crown node 24.38 13.7–29.5 8 Genus Clintonia crown node 12.8 4.76–20.25 9 Genus Calochortus crown node (subfamily Calochortoideae) 23.03 13.88–32.96 10 Genus Streptopus crown node 2.02 1.09–6.28 11 Genus Scoliopus crown node 0.91 0.01–3.63 12 Genus Tricyrtis crown node 14.95 5.71–23.31 Node Calibration point discription Mean age (Mya) 95% HPD A Family Liliaceae crown node 85.12 64.66–91.14 B Subfamily Streptopoideae crown node 54.78 36.62–83.53 C divergence of subfamilies Calochortoideae and Lilioideae 82.95 61.66–90.4 D Subfamily Lilioideae crown node 66.36 49.5–81.39 E Subfamily Medeoloideae crown node 46.15 22.52–58.94 F Subfamily Lilioideae crown node 61.81 44.8–74.6 G Tribe Tulipeae crown node 57.63 39.48–66.86 H Tribe Lilieae crown node 42.89 24.51–50.21 1 Genus Lilium (+Nomocharis) crown node 27.96 12.85–28.89 2 Genus Fritillaria crown node 28.12 15.58–32.42 3 Genus Cardiocrinum crown node 16.53 1.55–18.17 4 Genus Notholirion crown node 21.17 4.56–24.25 5 Genus Gagea (+Lloydia) crown node 44.24 23.32–43.45 6 Genus Tulipa crown node 20.74 10.2–23.99 7 Genus Erythronium crown node 24.38 13.7–29.5 8 Genus Clintonia crown node 12.8 4.76–20.25 9 Genus Calochortus crown node (subfamily Calochortoideae) 23.03 13.88–32.96 10 Genus Streptopus crown node 2.02 1.09–6.28 11 Genus Scoliopus crown node 0.91 0.01–3.63 12 Genus Tricyrtis crown node 14.95 5.71–23.31 View Large Table 2. Summary of divergence time estimation results under a Bayeasian approach as implented in BEAST based on the combined plastid DNA sequence data Node Calibration point discription Mean age (Mya) 95% HPD A Family Liliaceae crown node 85.12 64.66–91.14 B Subfamily Streptopoideae crown node 54.78 36.62–83.53 C divergence of subfamilies Calochortoideae and Lilioideae 82.95 61.66–90.4 D Subfamily Lilioideae crown node 66.36 49.5–81.39 E Subfamily Medeoloideae crown node 46.15 22.52–58.94 F Subfamily Lilioideae crown node 61.81 44.8–74.6 G Tribe Tulipeae crown node 57.63 39.48–66.86 H Tribe Lilieae crown node 42.89 24.51–50.21 1 Genus Lilium (+Nomocharis) crown node 27.96 12.85–28.89 2 Genus Fritillaria crown node 28.12 15.58–32.42 3 Genus Cardiocrinum crown node 16.53 1.55–18.17 4 Genus Notholirion crown node 21.17 4.56–24.25 5 Genus Gagea (+Lloydia) crown node 44.24 23.32–43.45 6 Genus Tulipa crown node 20.74 10.2–23.99 7 Genus Erythronium crown node 24.38 13.7–29.5 8 Genus Clintonia crown node 12.8 4.76–20.25 9 Genus Calochortus crown node (subfamily Calochortoideae) 23.03 13.88–32.96 10 Genus Streptopus crown node 2.02 1.09–6.28 11 Genus Scoliopus crown node 0.91 0.01–3.63 12 Genus Tricyrtis crown node 14.95 5.71–23.31 Node Calibration point discription Mean age (Mya) 95% HPD A Family Liliaceae crown node 85.12 64.66–91.14 B Subfamily Streptopoideae crown node 54.78 36.62–83.53 C divergence of subfamilies Calochortoideae and Lilioideae 82.95 61.66–90.4 D Subfamily Lilioideae crown node 66.36 49.5–81.39 E Subfamily Medeoloideae crown node 46.15 22.52–58.94 F Subfamily Lilioideae crown node 61.81 44.8–74.6 G Tribe Tulipeae crown node 57.63 39.48–66.86 H Tribe Lilieae crown node 42.89 24.51–50.21 1 Genus Lilium (+Nomocharis) crown node 27.96 12.85–28.89 2 Genus Fritillaria crown node 28.12 15.58–32.42 3 Genus Cardiocrinum crown node 16.53 1.55–18.17 4 Genus Notholirion crown node 21.17 4.56–24.25 5 Genus Gagea (+Lloydia) crown node 44.24 23.32–43.45 6 Genus Tulipa crown node 20.74 10.2–23.99 7 Genus Erythronium crown node 24.38 13.7–29.5 8 Genus Clintonia crown node 12.8 4.76–20.25 9 Genus Calochortus crown node (subfamily Calochortoideae) 23.03 13.88–32.96 10 Genus Streptopus crown node 2.02 1.09–6.28 11 Genus Scoliopus crown node 0.91 0.01–3.63 12 Genus Tricyrtis crown node 14.95 5.71–23.31 View Large Figure 2. View largeDownload slide Maximum clade credibility tree of Liliaceae inferred from the relaxed clock lognormal method, based on the combined data set of four plastid genes, three fossil calibration points marked with yellow stars and assuming at Yule process of speciation. Blue bars on nodes represent 95% confidence intervals for divergence time. Figure 2. View largeDownload slide Maximum clade credibility tree of Liliaceae inferred from the relaxed clock lognormal method, based on the combined data set of four plastid genes, three fossil calibration points marked with yellow stars and assuming at Yule process of speciation. Blue bars on nodes represent 95% confidence intervals for divergence time. The most recent common ancestor of Liliaceae occurred in temperate Asia in the late Cretaceous (ancestral area B, Fig. 3). After dispersal and vicariance events occurred in the family, the distribution expanded to North America (ancestral area F) and returned to Asia (ancestral areas B and C). The ancestral area of the Streptopoideae was temperate Asia (ancestral area B) and that for Calochortoideae was North America (ancestral area F). This distribution expanded again to Asia, including the tropical regions (ancestral area C), with geographical diversification events of dispersal and vicariance in the divergence of Medeoloideae and Lilioideae. Figure 3. View largeDownload slide Chronogram for Liliaceae based on the Baysian binary Markov chain Monte Carlo (BBM) method in RASP using the BEAST-derived chronogram (Fig. 2). Pie charts on the branches indicate relative probabilities for each ancestral area derived from BBM analysis based on a maximum area number of four. Inferred dispersal and vicariance events are indicated by pink and yellow stars, respectively, on internal branches. Figure 3. View largeDownload slide Chronogram for Liliaceae based on the Baysian binary Markov chain Monte Carlo (BBM) method in RASP using the BEAST-derived chronogram (Fig. 2). Pie charts on the branches indicate relative probabilities for each ancestral area derived from BBM analysis based on a maximum area number of four. Inferred dispersal and vicariance events are indicated by pink and yellow stars, respectively, on internal branches. DISCUSSION Phylogenetic relationship and updated classification of Liliaceae All currently accepted genera of Liliaceae sensu APG IV are included in the present study. The monophyly of Liliaceae is confirmed and is congruent with a previous study (Kim et al., 2013) with strong support values. Based on the updated molecular phylogenetic tree, we suggest four subfamilies: Streptopoideae, Calochortoideae, Medeoloideae and Lilioideae (Table 3). The major change is in the circumscription between Calochortoideae and Streptopoideae due to the position of Tricyrtis. Recently, Givnish et al. (2016) conducted phylogenomic analysis of all ten families in Liliales and showed that Tricyrtis was sister to Lilioideae, whereas Calochortus was closer to members of Streptopoideae. However, the bootstrap support values for these relationships were weak in the supermatrix analysis, which included only rbcL and matK. Although support values were moderate in the data set of 75 genes, these relationships were uncertain because of the deficiency in taxon sampling. Specifically, in their supermatrix analysis, there were only one, two and three samples of Calochortus, Streptopus and Tricyrtis, respectively. In our study, more representatives of Calochortus, Tricyrtis and Streptopus were used (Table S1). As a result, Tricyrtis was clearly embedded in Streptopoideae, and Calochortus was sister to Lilioideae with strong support values. This updated result allows us to recognize a monotypic Calochortoideae, excluding Tricyrtis. The second change was the elevation in rank of tribe Medeoleae of Lilioideae to an independent subfamily Medeoloideae based on the monophyly of this clade in the tree and its distinguishing morphological characteristics excluding the Fritillaria-type embryo sac formation (Table 3). We therefore recommend that this new subfamily classification system is more reasonable in understanding relationships in the family. Referring to our new concept, Liliaceae consist of four monophyletic subfamilies sharing morphological and cytological features (Table 3). Streptopoideae are recognized by rhizome type and reticulate leaf venation; they are sister to the rest of the family and consist of four genera (Tricyrtis, Scoliopus, Prosartes and Streptopus). Among the genera, Tricyrtis is distinguished from the others by having relatively larger and conspicuous flowers and a different basic chromosome number, x = 13. Prosartes, formerly in Disporum Salisb. (Shinwari et al., 1994a, 1994b) and reclassified within Liliaceae, shows similar characteristics to Streptopus, even though it is closer to Scoliopus in the tree. The monotypic Calochortoideae comprise c. 60 species, with large and conspicuous flowers, larger inner tepals and a basic chromosome number x = 9, and have a centre of diversity in California. Patterson & Givnish (2003) suggested that the speciation was caused by chromosomal evolution in the genus and in particular by limited dispersal that led to the narrow endemism of the species, the geographical cohesion of clades, parallel radiations in habitat preference, floral form and tolerance to serpentinite bedrock. Medeoloideae consist of two genera Medeola and Clintonia and have intermediate genome sizes between Lilioideae and the other subfamilies. We concluded that they should be recognized as an independent subfamily based on characters of the rhizome, reticulate venation, a small and inconspicuous flower, berry and a lower basic chromosome number (x = 7), even though it is clearly closer to Lilioideae in Fritillaria-type embryo sac formation. This closer relationship matched the results of Hayashi et al. (2001). Lilioideae are the largest in Liliaceae and include eight genera: Amana, Erythronium, Tulipa, Gagea (expanded to include Lloydia), Notholirion, Cardiocrinum, Lilium (expanded to include Nomocharis) and Fritillaria. The typical features of Liliaceae [bulb, parallel venation, large flower, loculicidal capsule and higher chromosome number (x = 12)] can be found in this subfamily. They were divided into two main clades corresponding to tribes Tulipeae and Lilieae with strong support values (Fig. 1), congruent with previous molecular phylogenetic studies (Patterson & Givnish, 2002; Fay et al., 2006; Gao et al., 2012; Kim et al., 2013). Table 3. Character comparison of newly defined subfamilies Subfamily Streptopoideae Calochortoideae Medeoloideae Lilioideae Genus Tricyrtis Scoliopus Prosartes Streptopus Calochortus Medeola Clintonia Amana Erythronium Tulipa Gagea (+Lloydia) Notholirion Cardiocrinum Lilium (+Nomocharis) Fritillaria Embryo sac formation (5) Polygonum type Fritillaria type Storage organ (6,7) rhizome bulb rhizome bulb Leaf venation (6) reticulate parallel reticulate parallel Tepal (6) large/ conspicuous small/inconspicuous large/ conspicuous small/ inconspicuous large/ conspicuous Size of outer(O) vs inner (I) tepal (5,7) O ≈ I O ≥ I O < I O ≈ I O ≈ I Fruit (5,7) capsule (septical) berry capsule (septical) berry capsule (loculicidal) Chromosome number (9) x = 13 x = 8 x = 8 (8–9) x = 8 x = 9 (6–10) x = 7 x = 12 Mean 1Cx (pg) (8,9) 4.5 9.2 3.4 6.6 5.4 14.2 18.9 21.5 32.8* 25.4* 14.2 35.3 38.6 39.6* 50.9* Distribution (10) AS NA NA AS, EU, NA NA NA AS, NA AS NA, EU, AS NA, EU, AS, AF NA, EU, AS, AF AS AS NA, AS, NA, EU, AS Subfamily Streptopoideae Calochortoideae Medeoloideae Lilioideae Genus Tricyrtis Scoliopus Prosartes Streptopus Calochortus Medeola Clintonia Amana Erythronium Tulipa Gagea (+Lloydia) Notholirion Cardiocrinum Lilium (+Nomocharis) Fritillaria Embryo sac formation (5) Polygonum type Fritillaria type Storage organ (6,7) rhizome bulb rhizome bulb Leaf venation (6) reticulate parallel reticulate parallel Tepal (6) large/ conspicuous small/inconspicuous large/ conspicuous small/ inconspicuous large/ conspicuous Size of outer(O) vs inner (I) tepal (5,7) O ≈ I O ≥ I O < I O ≈ I O ≈ I Fruit (5,7) capsule (septical) berry capsule (septical) berry capsule (loculicidal) Chromosome number (9) x = 13 x = 8 x = 8 (8–9) x = 8 x = 9 (6–10) x = 7 x = 12 Mean 1Cx (pg) (8,9) 4.5 9.2 3.4 6.6 5.4 14.2 18.9 21.5 32.8* 25.4* 14.2 35.3 38.6 39.6* 50.9* Distribution (10) AS NA NA AS, EU, NA NA NA AS, NA AS NA, EU, AS NA, EU, AS, AF NA, EU, AS, AF AS AS NA, AS, NA, EU, AS (1) Smith (1911), (2) McAllister (1914), (3) Berg (1962), (4) Björnstad (1970), (5) Dahlgren et al. (1985), (6) Tamura (1998), (7) Patterson & Givnish (2002), (8) Leitch et al. (2007), (9) Peruzzi et al. (2009), (10) e-monocot (http://www.e-monocot.org). View Large Table 3. Character comparison of newly defined subfamilies Subfamily Streptopoideae Calochortoideae Medeoloideae Lilioideae Genus Tricyrtis Scoliopus Prosartes Streptopus Calochortus Medeola Clintonia Amana Erythronium Tulipa Gagea (+Lloydia) Notholirion Cardiocrinum Lilium (+Nomocharis) Fritillaria Embryo sac formation (5) Polygonum type Fritillaria type Storage organ (6,7) rhizome bulb rhizome bulb Leaf venation (6) reticulate parallel reticulate parallel Tepal (6) large/ conspicuous small/inconspicuous large/ conspicuous small/ inconspicuous large/ conspicuous Size of outer(O) vs inner (I) tepal (5,7) O ≈ I O ≥ I O < I O ≈ I O ≈ I Fruit (5,7) capsule (septical) berry capsule (septical) berry capsule (loculicidal) Chromosome number (9) x = 13 x = 8 x = 8 (8–9) x = 8 x = 9 (6–10) x = 7 x = 12 Mean 1Cx (pg) (8,9) 4.5 9.2 3.4 6.6 5.4 14.2 18.9 21.5 32.8* 25.4* 14.2 35.3 38.6 39.6* 50.9* Distribution (10) AS NA NA AS, EU, NA NA NA AS, NA AS NA, EU, AS NA, EU, AS, AF NA, EU, AS, AF AS AS NA, AS, NA, EU, AS Subfamily Streptopoideae Calochortoideae Medeoloideae Lilioideae Genus Tricyrtis Scoliopus Prosartes Streptopus Calochortus Medeola Clintonia Amana Erythronium Tulipa Gagea (+Lloydia) Notholirion Cardiocrinum Lilium (+Nomocharis) Fritillaria Embryo sac formation (5) Polygonum type Fritillaria type Storage organ (6,7) rhizome bulb rhizome bulb Leaf venation (6) reticulate parallel reticulate parallel Tepal (6) large/ conspicuous small/inconspicuous large/ conspicuous small/ inconspicuous large/ conspicuous Size of outer(O) vs inner (I) tepal (5,7) O ≈ I O ≥ I O < I O ≈ I O ≈ I Fruit (5,7) capsule (septical) berry capsule (septical) berry capsule (loculicidal) Chromosome number (9) x = 13 x = 8 x = 8 (8–9) x = 8 x = 9 (6–10) x = 7 x = 12 Mean 1Cx (pg) (8,9) 4.5 9.2 3.4 6.6 5.4 14.2 18.9 21.5 32.8* 25.4* 14.2 35.3 38.6 39.6* 50.9* Distribution (10) AS NA NA AS, EU, NA NA NA AS, NA AS NA, EU, AS NA, EU, AS, AF NA, EU, AS, AF AS AS NA, AS, NA, EU, AS (1) Smith (1911), (2) McAllister (1914), (3) Berg (1962), (4) Björnstad (1970), (5) Dahlgren et al. (1985), (6) Tamura (1998), (7) Patterson & Givnish (2002), (8) Leitch et al. (2007), (9) Peruzzi et al. (2009), (10) e-monocot (http://www.e-monocot.org). View Large Circumscription of Lilioideae In Tulipeae, the monophyly of Gagea including Lloydia was confirmed, with all Lloydia spp. being embedded in the clade. This was congruent with previous molecular phylogenetic studies at the generic level (Peterson et al., 2008; Zarrei et al., 2009). On the other hand, Amana was revealed to be sister to Tulipa and Erythronium, as described in many molecular phylogenetic studies (Allen et al., 2003; Rønsted et al., 2005; Clennett et al., 2012; Christenhusz et al., 2013; Kim et al., 2013), even though the relationship among these three genera has been frequently confused. Therefore, Amana should be treated as a genus separate from Tulipa. The phylogenetic relationship within Lilieae indicated that Notholirion was the sister to the other three genera, and Fritillaria was closely related to Lilium (including Nomocharis) with strong support. The results also supported the monophyly of the expanded concept of Lilium to include Nomocharis, the taxonomic placement and relationship of which has been debated in recent decades (Liang, 1984; Nishikawa et al., 1999, 2001; Hayashi & Kawano, 2000; Liang & Tamura, 2000; Peruzzi et al., 2009; Lee et al., 2011). It was finally defined by Gao et al. (2012) based on reconstruction of its phylogenetic relationships. On the basis of this updated molecular phylogenetic analysis, we suggest the following new classification for Liliaceae composed of 15 genera: 1. Subfamily Lilioideae (eight genera) 1.1. Tribe Lilieae: Lilium (including Nomocharis), Fritillaria, Cardiocrinum, Notholirion 1.2. Tribe Tulipeae: Tulipa, Erythronium, Amana, Gagea (including Lloydia) 2. Subfamily Medeoloideae (two genera): Clintonia, Medeola 3. Subfamily Calochortoideae (one genus): Calochortus 4. Subfamily Streptopoideae (four genera): Scoliopus, Prosartes, Streptopus, Tricyrtis Asian origin and biogeographical history of Liliaceae Liliaceae have been regarded to have originated recently in Liliales. To calculate the age of Liliales, Vinnersten & Bremer (2001) suggested that Liliaceae originated c. 37–41 Mya, but historical study of the mycoheterotrophic family Corsiaceae in Liliales (Mennes et al., 2015) suggested that Liliaceae are older than the previous calculation of Vinnersten & Bremer (2001). Mennes et al. (2015) estimated the age of the family to 46 Mya (crown node) or 56 Mya (stem node) using nuclear 18S rDNA and mitochondrial atpA and matR gene sequence data. Additionally, Givnish et al. (2016) showed that Melanthiaceae and Liliaceae were the oldest families in Liliales, although the crown age of Liliaceae was 66.8 Mya (95% HPD, 48.3–92.3 Mya). However, in the present analyses, the crown age of Liliaceae was c. 85 Mya in the Late Cretaceous (Fig. 2 and Table 2), indicating that they are a relatively ancient clade of Liliales. Lilioideae are the oldest of the subfamilies (61.81 Mya) and Calochortoideae the youngest (23.03 Mya). These results revealed that Gagea is the oldest in the family, and numerous genera have estimated ages of 20–28 Mya. Gao et al. (2013) estimated the ages of Lilieae and Lilium at 16 and 13.6 Mya, respectively, but these divergence times are older in our estimation. The crown ages of Lilieae and Lilium were 42.89 and 27.96 Mya, respectively. Various different divergence times across Liliaceae have been reported in different studies, possibly caused by the different topologies of the phylogenetic trees and calibration points. A North American origin of Liliaceae was proposed by Vinnersten & Bremer (2001) and Patterson & Givnish (2002). It was also assumed that multiple intercontinental exchanges and expansions between North America and Eurasia occurred repeatedly in the family and within the genera via the Bering land bridge. Recently, Givnish et al. (2016) suggested that the origin of Liliaceae was possibly from North America or Asia based on a biogeographical analysis among families of Liliales. Our results showed that the ancestral area of the Liliaceae was temperate Asia (Fig. 3). This difference in origin may be due to variances in phylogenetic relationships among Tricyrtis, Calochortus and Streptopus in their study compared to the present analysis. Our results also suggested that dispersal and vicariance events occurred in the main lineages of the family at least 12 and five times, respectively. Vicariance was estimated to have occurred simultaneously with the dispersal event via Beringia between temperate Asia (B) and North America (F). Specifically, Streptopoideae originated in Asia and expanded to North America and Europe. On the other hand, ancestors of the remaining subfamilies arrived in North America and differentiated before expanding into Asia and Europe. CONCLUSIONS We provide an updated circumscription of Liliaceae composed of the four subfamilies Lilioideae, Medeol oideae, Calochortoideae and Streptopoideae. Dating analyses revealed that Liliaceae and Melathiaceae were the oldest families in Liliales. From an Asian origin, we assumed that the distribution of Liliaceae initially expanded to North America; then the family have again experienced complicated expansion between the two major distribution areas during its evolution. SUPPORTING INFORMATION Additional Supporting Information may be found in the online version of this article at the publisher’s web-site. Figure S1. Comparison of the ML tree topologies and support values with different outgroup settings. (A) Original analysis, which includes three most closely related families (Smilacaceae, Ripogonaceae and Philesiaceae) together. (B) Smilacaceae were excluded from the analysis. (C) Smilacaceae were included as the sole sister group in the analysis. Table S1. Plant materials used in this study and their accession numbers. Genes missing from the data matrix are labeled with a dash (-). Table S2. Primer sequences and references for PCR and sequencing reactions performed in this study. FUNDING This work was supported by the National Research Foundation of Korea (NRF) Grant Fund (MEST 2010-0029131) and Korea National Arboretum (KNA 1-2-13, 14-2). ACKNOWLEDGMENTS Thanks to Ms Hee-Woon Jang, Mr Linh Nguyen Vu and Dr Do Hoa Dong Khoa at Gachon University for their sincere contribution and support for this research. Authors thank to DNA bank of Royal Botanic Gardens, Kew (http://apps.kew.org/dnabank/homepage.html), Herbarium of New York Botanical Garden (NYBG) and the Herbarium of the Kunming Institute of Botany, CAS (KUN) for providing the DNAs and plant materials for this study. Joo-Hwan Kim would like to pay special thanks and appreciation to Prof Yoon Shik Kim who inspired me to pursue the path of science. His encouragment and support made it possible to continue my life as a scientist. I dedicate this article to him. REFERENCES Allen GA , Soltis DE , Soltis PS . 2003 . Phylogeny and biogeography of Erythronium (Liliaceae) inferred from chloroplast matK and nuclear rDNA ITS sequences . Systematic Botany 28 : 512 – 523 . APG IV . 2016 . An update of the Angiosperm Phylogeny Group classification for the orders and families of flowering plants: APG IV . Botanical Journal of the Linnean Society 181 : 1 – 20 . Bentham G , Hooker JD . 1883 . Genera plantarum ad exemplaria imprimis in herbariis kewensibus servata definita . London : Reeve & Co . Berg RY . 1962 . Contribution to the comparative embryology of the Liliaceae: Scoliopus, Trillium, Paris, and Medeola . Skrifter, Det Norske Videnskaps-Akademi, I. Mat.-Naturv. Klasse, Ny Serie 4 : 64 . Björnstad IN . 1970 . Comparative embryology of Asparagoideae-Polygonatae, Liliaceae . Nytt Magasin for Botanikk 17 : 169 – 207 . Bouckaert R , Heled J , Kuhnert D , Vaughan T , Wu CH , Xie D , Suchard MA , Rambaut A , Drummond AJ . 2014 . BEAST 2: a software platform for Bayesian evolutionary analysis . PLoS Computational Biology 10 : e1003537 . Bremer K . 2000 . Early Cretaceous lineages of monocot flowering plants . Proceedings of the National Academy of Sciences of the United States of America 97 : 4707 – 4711 . Carpenter RJ , Wilf P , Conran JG , Cuneo NR . 2014 . A Paleogene trans-Antarctic distribution for Ripogonum (Ripogonaceae: Liliales) . Palaeontologia Electronica 17 : 39A . Chacon J , de Assis MC , Meerow AW , Renner SS . 2012 . From East Gondwana to Central America: historical biogeography of the Alstroemeriaceae . Journal of Biogeography 39 : 1806 – 1818 . Chacon J , Renner SS . 2014 . Assessing model sensitivity in ancestral area reconstruction using Lagrange: a case study using the Colchicaceae family . Journal of Biogeography 41 : 1414 – 1427 . Chase MW , Duvall MR , Hills HG , Conran JG , Cox AV , Eguiarte LE , Hartwell J , Fay MF , Caddick LR , Cameron KM , Hoot SB . 1995 . Molecular phylogenetics of Lilianae . In: Rudall PJ , Cribb PJ , Cutler DF , Humphries CJ , eds. Monocotyledons: systematics and evolution . Kew : Royal Botanic Gardens, Kew , 109 – 137 . Chase MW , Hills HG . 1991 . Silica gel—an ideal material for field preservation of leaf samples for DNA studies . Taxon 40 : 215 – 220 . Chen C , Qi ZC , Xu XH , Comes HP , Koch MA , Jin XJ , Fu CX , Qiu YX . 2014 . Understanding the formation of Mediterranean–African–Asian disjunctions: evidence for Miocene climate-driven vicariance and recent long-distance dispersal in the Tertiary relict Smilax aspera (Smilacaceae) . New Phytologist 204 : 243 – 255 . Christenhusz MJM , Govaerts R , David JC , Hall T , Borland K , Roberts PS , Tuomisto A , Buerki S , Chase MW , Fay MF . 2013 . Tiptoe through the tulips—cultural history, molecular phylogenetics and classification of Tulipa (Liliaceae) . Botanical Journal of the Linnean Society 172 : 280 – 328 . Clennett JCB , Chase MW , Forest F , Maurin O , Wilkin P . 2012 . Phylogenetic systematics of Erythronium (Liliaceae): morphological and molecular analyses . Botanical Journal of the Linnean Society 170 : 504 – 528 . Conran JG , Bannister JM , Mildenhall DC , Lee DE , Chacon J , Renner SS . 2014 . Leaf fossils of Luzuriaga and a Monocot flower with in situ pollen of Liliacidites contortus Mildenh. & Bannister sp nov (Alstroemeriaceae) from the early Miocene . American Journal of Botany 101 : 141 – 155 . Conran JG , Carpenter RJ , Jordan GJ . 2009 . Early Eocene Ripogonum (Liliales: Ripogonaceae) leaf macrofossils from southern Australia . Australian Systematic Botany 22 : 219 – 228 . Daghlian CP . 1981 . A review of the fossil record of monocotyledons . Botanical Review 47 : 517 – 555 . Dahlgren RMT , Clifford HT , Yeo PF . 1985 . The families of the monocotyledons: structure, evolution, and taxonomy . Berlin : Springer . Day PD , Berger M , Hill L , Fay MF , Leitch AR , Leitch IJ , Kelly LJ . 2014 . Evolutionary relationships in the medicinally important genus Fritillaria L. (Liliaceae) . Molecular Phylogenetics and Evolution 80 : 11 – 19 . Ding ST , Sun BN , Wu JY , Li XC . 2011 . Miocene Smilax leaves and associated epiphyllous fungi from Zhejiang, East China and their paleoecological implications . Review of Palaeobotany and Palynology 165 : 209 – 223 . Doyle JJ , Doyle JL . 1987 . A rapid DNA isolation procedure for small quantities of fresh leaf tissue . Phytochemical Bulletin, Botanical Society of America 19 : 11 – 15 . Drummond AJ , Ho SYW , Phillips MJ , Rambaut A . 2006 . Relaxed phylogenetics and dating with confidence . PLoS Biology 4 : 699 – 710 . Drummond AJ , Suchard MA , Xie D , Rambaut A . 2012 . Bayesian phylogenetics with BEAUti and the BEAST 1.7 . Molecular Biology and Evolution 29 : 1969 – 1973 . Engler A . 1888 . Liliaceae . In: Engler A , Prantl K , eds. Die Natürlichen Pflanzenfamilien II (5) . Leipzig : W. Engelmann , 10 – 91 . Fay MF , Chase MW . 2000 . Modern concepts of Liliaceae with a focus on the relationships of Fritillaria . Curtis’s Botanical Magazine 17 : 146 – 149 . Fay MF , Chase MW , Rønsted NAH , Devey DS , Pillon Y , Pires JC , Petersen G , Seberg O , Davis JI . 2006 . Phylogenetics of Liliales: summarized evidence from combined analyses of five plastid and one mitochondrial loci . Aliso 22 : 559 – 565 . Gao YD , Harris AJ , Zhou SD , He XJ . 2013 . Evolutionary events in Lilium (including Nomocharis, Liliaceae) are temporally correlated with orogenies of the Q-T plateau and the Hengduan Mountains . Molecular Phylogenetics and Evolution 68 : 443 – 460 . Gao YD , Hohenegger M , Harris AJ , Zhou SD , He XJ , Wan J . 2012 . A new species in the genus Nomocharis Franchet (Liliaceae): evidence that brings the genus Nomocharis into Lilium . Plant Systematics and Evolution 298 : 69 – 85 . Givnish TJ , Zuluaga A , Marques I , Lam VKY , Gomez MS , Iles WJD , Ames M , Spalink D , Moeller JR , Briggs BG , Lyon SP , Stevenson DW , Zomlefer W , Graham SW . 2016 . Phylogenomics and historical biogeography of the monocot order Liliales: out of Australia and through Antarctica . Cladistics 32 : 581 – 605 . Hayashi K , Kawano S . 2000 . Molecular systematics of Lilium and allied genera (Liliaceae): phylogenetic relationships among Lilium and related genera based on the rbcL and matK gene sequence data . Plant Species Biology 15 : 73 – 93 . Hertweck KL , Kinney MS , Stuart SA , Maurin O , Mathews S , Chase MW , Gandolfo MA , Pires JC . 2015 . Phylogenetics, divergence times and diversification from three genomic partitions in monocots . Botanical Journal of the Linnean Society 178 : 375 – 393 . Hoot SB , Culham A , Crane PR . 1995 . The utility of atpB gene-sequences in resolving phylogenetic-relationships—comparison with rbcL and 18S ribosomal DNA-sequences in the Lardizabalaceae . Annals of the Missouri Botanical Garden 82 : 194 – 207 . Hutchinson J . 1959 . The families of flowering plants, arranged according to a new system based on their probable phylogeny . Oxford : Clarendon Press . Iles WJD , Smith SY , Gandolfo MA , Graham SW . 2016 . Monocot fossils suitable for molecular dating analyses . Botanical Journal of the Linnean Society 178 : 346 – 374 . Janssen T , Bremer K . 2004 . The age of major monocot groups inferred from 800+rbcL sequences . Botanical Journal of the Linnean Society 146 : 385 – 398 . Kearse M , Moir R , Wilson A , Stones-Havas S , Cheung M , Sturrock S , Buxton S , Cooper A , Markowitz S , Duran C , Thierer T , Ashton B , Meintjes P , Drummond A . 2012 . Geneious basic: an integrated and extendable desktop software platform for the organization and analysis of sequence data . Bioinformatics 28 : 1647 – 1649 . Kelchner SA . 2000 . The evolution of non-coding chloroplast DNA and its application in plant systematics . Annals of the Missouri Botanical Garden 87 : 482 – 498 . Kim JS , Hong JK , Chase MW , Fay MF , Kim JH . 2013 . Familial relationships of the monocot order Liliales based on a molecular phylogenetic analysis using four plastid loci: matK, rbcL, atpB and atpF-H . Botanical Journal of the Linnean Society 172 : 5 – 21 . Lee CS, Kim SC, Yeau SH, Lee NS. 2011. Major lineages of the genus Lilium (Liliaceae) based on nrDNA ITS sequences, with special emphasis on the Korean species. Journal of Plant Biolology 54: 159–171. Leitch IJ, Beaulieu JM, Cheung K, Hanson L, Lysak MA, Fay MF. 2007. Punctuated genome size evolution in Liliaceae. Journal of Evolutionary Biology 20: 2296–2308. Liang SY. 1984. Studies on the genus Nomocharis (Liliaceae). Bulletin of Botanical Research 4: 163–178. Liang SY, Tamura M. 2000. Lilium L.; Nomocahris Franchet. In: Wu ZY, Raven PH, eds. Flora of China, vol. 24. Beijing: Science Press; St. Louis: Missouri Botanical Garden Press, 135–159. Logan GA , Smiley CJ , Eglinton G . 1995 . Preservation of fossil leaf waxes in association with their source tissues, Clarkia, Northern Idaho, USA . Geochimica et Cosmochimica Acta 59 : 751 – 763 . McAllister F . 1914 . The development of the embryo sac in the Convallariaceae . Botanical Gazette 58 : 137 – 153 . Mennes CB , Lam VKY , Rudall PJ , Lyon SP , Graham SW , Smets EF , Merckx VSFT . 2015 . Ancient Gondwana break-up explains the distribution of the mycoheterotrophic family Corsiaceae (Liliales) . Journal of Biogeography 42 : 1123 – 1136 . Mennes CB , Smets EF , Moses SN , Merckx VSFT . 2013 . New insights in the long-debated evolutionary history of Triuridaceae (Pandanales) . Molecular Phylogenetics and Evolution 69 : 994 – 1004 . Nishikawa T , Okazaki K , Arakawa K , Nagamine T . 2001 . Phylogenetic analysis of section Sinomartagon in genus Lilium using sequences of the internal transcribed spacer region in nuclear ribosomal DNA . Breeding Science 51 : 39 – 46 . Nishikawa T , Okazaki K , Uchino T , Arakawa K , Nagamine T . 1999 . A molecular phylogeny of Lilium in the internal transcribed spacer region of nuclear ribosomal DNA . Journal of Molecular Evolution 49 : 238 – 249 . Olmstead RG , Kim KJ , Jansen RK , Wagstaff SJ . 2000 . The phylogeny of the Asteridae sensu lato based on chloroplast ndhF gene sequences . Molecular Phylogenetics and Evolution 16 : 96 – 112 . Patterson TB , Givnish TJ . 2002 . Phylogeny, concerted convergence, and phylogenetic niche conservatism in the core Liliales: insights from rbcL and ndhF sequence data . Evolution 56 : 233 – 252 . Patterson TB, Givnish TJ; Blackwell Publishing Ltd. 2003. Geographic cohesion, chromosomal evolution, parallel adaptive radiations, and consequent floral adaptations in Calochortus (Calochortaceae): evidence from a cpDNA phylogeny. New Phytologist 161: 253–264. Peruzzi L, Leitch IJ, Caparelli KF. 2009. Chromosome diversity and evolution in Liliaceae. Annals of Botany 103: 459–475. Peruzzi L . 2011 . Nomenclatural novelties at sectional level in Gagea (Liliaceae) . Atti della Società Toscana di Scienze Naturali, Serie B 118 : 23 – 24 . Petersen G , Seberg O , Davis JI . 2013 . Phylogeny of the Liliales (Monocotyledons) with special emphasis on data partition congruence and RNA editing . Cladistics 29 : 274 – 295 . Peterson A , John H , Koch E , Peterson J . 2004 . A molecular phylogeny of the genus Gagea (Liliaceae) in Germany inferred from non-coding chloroplast and nuclear DNA sequences . Plant Systematics and Evolution 245 : 145 – 162 . Peterson A , Levichev IG , Peterson J . 2008 . Systematics of Gagea and Lloydia (Liliaceae) and infrageneric classification of Gagea based on molecular and morphological data . Molecular Phylogenetics and Evolution 46 : 446 – 465 . Posada D . 2008 . jModelTest: phylogenetic model averaging . Molecular Biology and Evolution 25 : 1253 – 1256 . Ronquist F , Huelsenbeck JP . 2003 . MrBayes 3: Bayesian phylogenetic inference under mixed models . Bioinformatics 19 : 1572 – 1574 . Rønsted N , Law S , Thornton H , Fay MF , Chase MW . 2005 . Molecular phylogenetic evidence for the monophyly of Fritillaria and Lilium (Liliaceae; Liliales) and the infrageneric classification of Fritillaria . Molecular Phylogenetics and Evolution 35 : 509 – 527 . Shinwari ZK , Kato H , Terauchi R , Kawano S . 1994a . Phylogenetic-relationships among genera in the Liliaceae–Asparagoideae–Polygonatae s.l. inferred from rbcL gene sequence data . Plant Systematics and Evolution 192 : 263 – 277 . Shinwari ZK , Terauchi R , Utech FH , Kawano S . 1994b . Recognition of the new-world Disporum section Prosartes as Prosartes (Liliaceae) based on the sequence data of the rbcL gene . Taxon 43 : 353 – 366 . Smith R . 1911 . The tetranucleate embryo sac of Clintonia . Botanical Gazette 52 : 209 – 217 . Stamatakis A , Hoover P , Rougemont J . 2008 . A rapid bootstrap algorithm for the RAxML web servers . Systematic Biology 57 : 758 – 771 . Tamura MN . 1998 . Flowering plants. Monocotyledons. Lilianae (except Orchidaceae) . Berlin : Springer-Verlag . Vinnersten A , Bremer K . 2001 . Age and biogeography of major clades in Liliales . American Journal of Botany 88 : 1695 – 1703 . Wilde V , Frankenhauser H . 1998 . The middle Eocene plant taphocoenosis from Eckfeld (Eifel, Germany) . Review of Palaeobotany and Palynology 101 : 7 – 28 . Wilf P . 2000 . Late Paleocene-early Eocene climate changes in southwestern Wyoming: paleobotanical analysis . Geological Society of America Bulletin 112 : 292 – 307 . Yu Y , Harris AJ , Blair C , He XJ . 2015 . RASP (reconstruct ancestral state in phylogenies): a tool for historical biogeography . Molecular Phylogenetics and Evolution 87 : 46 – 49 . Yu Y , Harris AJ , He XJ . 2010 . S-DIVA (statistical dispersal-vicariance analysis): a tool for inferring biogeographic histories . Molecular Phylogenetics and Evolution 56 : 848 – 850 . Zarrei M , Wilkin P , Fay MF , Ingrouille MJ , Zarre S , Chase MW . 2009 . Molecular systematics of Gagea and Lloydia (Liliaceae; Liliales): implications of analyses of nuclear ribosomal and plastid DNA sequences for infrageneric classification . Annals of Botany 104 : 125 – 142 . Zhao YP , Qi ZC , Ma WW , Dai QY , Li P , Cameron KM , Lee J , Xiang QY , Fu CX . 2013 . Comparative phylogeography of the Smilax hispida group (Smilacaceae) in eastern Asia and North America—implications for allopatric speciation, causes of diversity disparity, and origins of temperate elements in Mexico . Molecular Phylogenetics and Evolution 68 : 300 – 311 . © 2018 The Linnean Society of London, Botanical Journal of the Linnean Society This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices)