Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Population density and size influence pollen dispersal pattern and mating system of the predominantly outcrossed Banksia nivea (Proteaceae) in a threatened ecological community

Population density and size influence pollen dispersal pattern and mating system of the... Abstract Gene flow is a critical component of plant mating systems and influences population fitness. However, pollen dispersal can be highly variable and influenced by natural and anthropogenic fragmentation. Gene flow through pollen dispersal was investigated in two populations of contrasting size and habitat context in Banksia nivea ssp. uliginosa, a rare species in the Busselton ironstone threatened ecological community of Western Australia with a naturally fragmented distribution. Paternity analysis was conducted with seven microsatellite loci to determine mating system parameters and patterns of pollen dispersal. Outcrossing was high in both populations with a similar level of selfing for both populations despite differences in population size, density and vegetation matrix. Most mating occurred within 10 m of a mother plant in the small, clumped population, while more dispersed mating, up to 50 m from a mother plant, was recorded in the large, less dense population. Our results show that population density and size are important influences on mating system parameters and level of pollen dispersal. gene flow, fragmentation, microsatellite, paternity analysis, pollinator movement INTRODUCTION Gene flow through pollen dispersal is necessary to maintain long-term population viability (Segelbacher et al., 2010), and is largely dependent on the mechanisms of dispersal or mode of reproduction and the degree of connectivity among populations (Ellstrand, 2014; Lowe, Harris & Ashton, 2004; Pannell & Fields, 2014). Fragmentation creates physical barriers by increasing isolation among patches and reducing total habitat area (Frankham, Ballou & Briscoe, 2002). Small, irregularly shaped fragments can exacerbate edge effects on populations within a fragment (Bennett, 2003), expose fragments to an increased risk of disease or exotic species invasion (Mendonça et al., 2015), affect pollen dispersal and the behaviour of pollinators (Byrne et al., 2007; Yates et al., 2007a; Llorens et al., 2012) and change community dynamics (Haddad et al., 2015). Roads and fences can alter the shape and distribution of small and degraded remnant populations to form linear strips with a high edge to area ratio. In addition, small population fragments may receive fewer pollinator visits because they do not provide enough flowering plants to attract or support pollinators (Yates et al., 2007a). Fragmentation impacts pollinators differently depending on several factors, such as number of fragments, size and spatial pattern of fragments, time since fragmentation, dispersal ability of the species, diversity of resources for pollinators, and the environment around the fragments and its impact on foraging range (Aguilar et al., 2006). These factors may lead to a reduction or loss of pollinators, and/or changed patterns of pollinator behaviour, causing reduced reproductive success of plant species that depend on these pollinators (Reed, 2005). Isolation and small population size also increase the chance of self-pollination, mating between related individuals and correlated mating as the result of fewer mates (Ouborg, Vergeer & Mix, 2006; Yates et al., 2007b), as well as reduction in pollen immigration (Byrne et al., 2007). Due to the large number of factors influencing both historical and contemporary gene flow in fragmented landscapes, it is not surprising that a range of responses of plant populations to fragmentation has been shown (Young, Boyle & Brown, 1996; Honnay & Jacquemyn, 2007; Aguilar et al., 2008; Eckert et al., 2010). Limited gene flow can result in genetic drift or differential selection in local environments leading to genetic differentiation between populations (Coates & Byrne, 2005). An increase in selfing in normally outcrossed species may lead to reduced genetic diversity and increased differentiation between populations (Coates, Sampson & Yates, 2007). Increases in inbreeding can lead to expression of lethal or harmful recessive mutations that reduce rates of growth and flowering, decrease pollen quantity, lower seed production, reduce fecundity through low germination success, and increase mortality (Burbidge & James, 1991; Dudash & Fenster, 2000; Richards, 2000; O’Connell, Mosseler & Rajora, 2006). However, some studies have shown no significant difference in estimates of outcrossing rate in large or small populations in fragmented landscapes (Coates et al., 2007; Yates et al., 2007a). One possible explanation for this is that bird pollinators transport pollen between populations and buffer small populations from reduced mate availability through increased effective population size (Yates et al., 2007a). Another explanation is the presence of post-zygotic seed abortion mechanisms involving deleterious recessive genes that selectively decrease progeny numbers but increase the multilocus outcrossing rate (Burbidge & James, 1991). In addition, species in naturally fragmented populations may be adapted to this landscape context and be less affected by anthropogenic fragmentation (Sampson et al., 2014, 2015; Nelson-Tunley, Morgan Richards & Trewick, 2016). The Busselton ironstone ecological community of Western Australia provides a suitable landscape context to investigate the effects of natural and anthropogenic fragmentation on gene flow through pollen dispersal. This community occurs on ironstone soils that are associated with shallow seasonal inundation as a consequence of the impermeable nature of subsurface layers of ironstone and associated heavy soil (English, 1999). The plant community on these ironstone soils is listed as a Critically Endangered Threatened Ecological Community due to its restricted occurrence (Fig. 1) and there are also several rare endemic species restricted to the community. The community has been affected by land clearing, dieback disease (particularly affecting species in the family Proteaceae), weed invasion, grazing, altered fire regimes, hydrological change and salinity (English, 1999). Figure 1. View largeDownload slide Vegetation layer and presence of ironstone in the Busselton area of Western Australia showing natural fragmentation of ironstone soil patches and anthropogenic fragmentation of vegetation through clearing. Remaining vegetation is shown in grey, cleared land is shown in white, and ironstone soils are outlined in red. Location of the Tutunup and Gale populations indicated by arrows. Note Banksia nivea ssp. uliginosa only occurs in 15 populations on some ironstone soils with remaining vegetation. Figure 1. View largeDownload slide Vegetation layer and presence of ironstone in the Busselton area of Western Australia showing natural fragmentation of ironstone soil patches and anthropogenic fragmentation of vegetation through clearing. Remaining vegetation is shown in grey, cleared land is shown in white, and ironstone soils are outlined in red. Location of the Tutunup and Gale populations indicated by arrows. Note Banksia nivea ssp. uliginosa only occurs in 15 populations on some ironstone soils with remaining vegetation. Banksia nivea ssp. uliginosa (A.S. George) A.R. Mast & K.R. Thiele (swamp honeypot) is one such species that is restricted to the ironstone habitat around Busselton (George, 1996; Mast & Thiele, 2007) and has a naturally fragmented population distribution (Fig. 1). It is isolated from the common subspecies by >100 km. While the ecological characteristics of B. nivea (formerly Dryandra nivea) are well described by George (1996) and Cavanagh & Pieroni (2006), little is known about its reproductive biology. A pollinator exclusion experiment has shown that mammals are the predominant pollinator with some bird pollination (R. Thavornkanlapachai, unpubl. data). Based on primarily mammal pollination with some bird pollination, we predict that patterns of pollen dispersal both within and between populations would be influenced by population size and isolation, with lower pollen dispersal distance in smaller road verge populations compared to larger populations in nature reserves. This study aimed to test the hypothesis that pollen dispersal in B. nivea ssp. uliginosa in fragmented populations is consistent with population characteristics of size and density, through examining the mating system and pollen dispersal in two populations of different size and density in a naturally fragmented landscape. MATERIAL AND METHODS Study species and study sites Banksia nivea ssp. uliginosa has a patchy distribution in two main areas, east of Busselton and on the Scott River Plain, approximately 220 and 260 km, respectively, south of Perth, Western Australia (George, 1996). It grows as a mounded shrub up to 1.5 m tall and 1.5 m across. Flowers are yellowish brown with maroon-coloured style and green pollen presenter, have a mousey odour, and are well hidden within the plant (Brown, Thomson-Dans & Marchant, 1998; Cavanagh & Pieroni, 2006). Flowering occurs between July and September. Each inflorescence has 86.7 ± 2.9 flowers on average (R. Thavornkanlapachai, unpubl. data). Although some follicles open to release seeds periodically, many remain on the plant for a long period time. If plants die after a disturbance, such as fire, mass recruitment occurs (Dixon, Dixon & Krauss, 2003). This study was undertaken in two populations of differing size and habitat context. The location of sampled plants in each population is shown in Figure 2. One population is located on the northern side of Tutunup Road, south-east of Busselton. Adjacent to this area, there is a larger number of plants on the southern side of the road. The size of the TutunupA population is 45 plants, and the majority occur in dense clumps scattered along the road verge with an average number of 21.7 ± 11.4 plants in a radius of 10 m. The overall density is 86 plants/ha, although this overall value is not representative of the plant distribution due to the presence of two more isolated plants, and the density in the main distribution is 144 plants/ha (Fig. 2A). The other population is located in a moderately sized nature reserve at Gale Road. This population is large and dispersed with an average number of 7.5 ± 3.1 plants in a radius of 10 m, and is isolated from other populations (Fig. 2B). The estimated population size is > 500 plants with a density of 134.2 plants/ha that is representative of the plant distribution. Figure 2. View largeDownload slide Location of plants sampled for paternity analysis in two populations of Banksia nivea ssp. uliginosa: (A) Gale population and (B) TutunupA population. ●, sampled plant;⋆, plant sampled for seed; number identifies mother plants; lines represent roads. Figure 2. View largeDownload slide Location of plants sampled for paternity analysis in two populations of Banksia nivea ssp. uliginosa: (A) Gale population and (B) TutunupA population. ●, sampled plant;⋆, plant sampled for seed; number identifies mother plants; lines represent roads. Plant material and genotyping Leaf samples were collected from all 45 plants in the TutunupA population, and 101 plants sampled from a 106 × 71-m patch in the centre of the reserve on Gale Road. Twenty infructescences produced in 2008 from each of ten randomly selected mature plants were collected in both populations. For each plant, 25 seeds were extracted, sterilized in 1% sodium hypochlorite solution for 10 min, and rinsed in sterile water four times. The seeds were placed on 7% Amy Media bacteriological agar in 9-cm Petri dishes, wrapped with Parafilm, and incubated at 15 °C with a 12-h dark/light cycle. Seedlings were grown for at least 5 weeks prior to DNA extraction. Both leaf and seedling samples were freeze dried for 2 days. DNA from leaf samples was extracted using the CTAB-PVP method (Byrne, Parrish & Moran, 1998) with sodium sulphite added to the extraction buffer (Byrne, Macdonald & Francki, 2001). DNA from seedlings was extracted using a small-scale version of the Doyle & Doyle (1990) method. DNA samples were normalized to 20 ng/µL and genotyped for seven microsatellite loci, DnA105, DnA129, DnA117, DnB008, DnB114, DnC010 and DnD007, following Millar & Byrne (2008). Modifications were made as follows: addition of another microsatellite locus, DnC010; forward primer sequence 5′-TGTTGCATGTTGGGATAAGG-3′ and reverse primer sequence 5′-TGCATCCAGCCTTGTCATAG-3′; and MgCl2 concentration was adjusted to 1.0 mM (DnA105, DnA117, DnC010, DnD007), 1.5 mM (DnB008, DnB114) or 2.0 mM (DnA129). Data analysis CERVUS 3.0 was used to obtain maximum likelihood-based paternity assignment and measure genetic diversity in the populations [number of alleles per locus, observed and expected heterozygosity, and polymorphic information content (PIC)] (Marshall et al., 1998). Paternity assignment used the following parameters: 10000 simulated offspring; number and proportion in blankets of candidate fathers of 45 (1.0) and 101 (0.1) for TutunupA and Gale, respectively; proportion of loci typed 0.98 and 0.99; inbreeding rates of 0.035 and 0.058; proportion of mistypes of 0.01; minimum number of loci of 4; and confidence level for assignment of 80%. Differences in measures of genetic diversity between populations, as well as between parents and progenies, were statistically tested using Friedman’s rank sum and Wilcoxon’s signed-rank tests implemented in the R version 3.4.3 statistical package (R Core Team, 2017). NEWPAT was also used in conjunction with CERVUS 3.0 as it can identify multiple-matching male parents in addition to having specific algorithms for dealing with mismatches potentially caused by null alleles (Amos, 2000). In cases where two possible fathers with the same likelihood were identified, the closest father by distance to the mother plant was selected as the most likely father. Cryptic pollen flow, or the probability that a seed was assigned to a pollen donor within a population when the true parent is located outside of the population, was calculated as 1 − (P2n) where P2 is the total exclusion probability for the second parent (Dow & Ashley, 1998). Pollen dispersal distances (between mother plant and pollen donor) were determined based on paternity assignment and the number of pollen donors for each 10-m distance class was plotted. Mating system parameters were estimated using MLTR v3.0 (Ritland, 2002) based on the expectation maximization (EM) method for determining maternal and pollen allele frequencies and the Newton Raphson method for joint maximum likelihood estimation of the multi-locus outcrossing rate (tm) and single-locus outcrossing estimate averages over all loci (ts). Estimates of bi-parental inbreeding (tmts), correlation of paternity (rp), correlation of t among loci (rs), correlation of selfing among loci within progeny arrays (rsloc) and correlation of selfing among families (rsfam) were also obtained. Standard errors were computed based on 500 bootstraps. The correlation of paternity was translated to provide estimates of the effective number of pollen donors using the equation Nep = 1/rp (Ritland, 2002). The standard error was used to determine whether mating parameters were significantly lower than one or greater than zero. Pollen and ovule frequencies were not pooled. The fixation index of seedlings (FO) was estimated, FO=1−(HO/HE), where HO a direct estimate of observed heterozygosity from the seed calculated in GENALEX v6.5 (Peakall & Smouse, 2006) and HE is expected heterozygosity, manually calculated using HE=1−∑​p2, where p is the frequency of each allele of a given locus in the pollen pool (Nei, 1977). Mean co-ancestry of adult plants was estimated as Loiselle’s kinship coefficient (Loiselle et al., 1995) using SPAGeDi (Hardy & Vekemans, 2002). Mean co-ancestry within families was manually calculated as Θ = 0.125(1+Fm)[4s + (tm2 +stmrs )(1+rpm)], where Fm is the fixation index of mother trees, s is the selfing rate, estimated as s =1−tm, rs is correlation of selfing among families, and rpm is multilocus correlation of paternity (Sebbenn, 2006). We inferred pollen-mediated gene dispersal patterns by modelling parentage probabilities of offspring using the neighbourhood model as implemented in NM+ v1.1 (Chybicki & Burczyk, 2010). NM+ uses a maximum likelihood approach to estimate the proportion of offspring resulting from selfing (s), pollen immigration from outside the sampling area (mp) and shape parameter of pollen (bp). Initial parameters were set as follows: neighbourhood of both populations to infinite (to sample all individuals in the sampling area), error rates to 0.01, pollen immigration rate (mp) and selfing rate (s) following the values in Table 3, and initial value for bp to one. Average pollen dispersal distance (dp) was calculated from the average distance of assigned father to mother plant. The level of pollen immigration, to, was estimated as to= 1 − (ti + s), where s is the proportion of selfed progeny and ti is the proportion of outcross events originating from pollen sources within the population. The heterogeneity of the pollen pools sampled from seedlings of each maternal tree was analysed using the TWOGENER method (Smouse et al., 2001) implemented in GENALEX v6.5 (Peakall & Smouse, 2006). This analysis is based on analysis of molecular variance (AMOVA) (Excoffier, Smouse & Quattro, 1992) and identifies the effects of fathers on pollen pool structure (Smouse et al., 2001). RESULTS Microsatellite variation The microsatellite loci were highly variable in the sampled populations of B. nivea ssp. uliginosa with a total of 55 and 53 alleles detected across the seven loci in the TutunupA and Gale populations, respectively (Table 1). The average numbers of alleles per locus (A) were similar in the TutunupA and Gale populations (7.9 ± 4.2 and 7.6 ± 5.0, respectively). Observed heterozygosity (HO) in the TutunupA population was high (0.581 ± 0.051) but not significantly different from that in the Gale population (0.477 ± 0.09; W = 33, P = 0.318). There were no significant differences in observed or expected heterozygosity for parent and progenies in either population (TutunupA: χ2 = 0.110, d.f. = 6, P > 0.05; Gale: χ2 = 0.094, d.f. = 6, P > 0.05). Estimates of null allele frequencies were lower than 0.1, except for DnA129 in both populations and DnA117 in the TutunupA population. As mismatches generated by null alleles are treated as typing errors in paternity assignment, null alleles with low frequencies do not have a significant effect. The polymorphic information content of the loci in both populations was moderate and not significantly different (W = 35, P = 0.209). Table 1. Allelic diversity values for microsatellite markers in two populations of Banksia nivea ssp. uliginosa Population  Locus  A  HO  HE  PIC  Excl  Null  TutunupA  DnA105  4  0.502  0.537  0.467  0.725  0.039  N = 245  DnA117  5  0.465  0.487  0.433  0.745  0.023    DnA129  9  0.498  0.763  0.730  0.441  0.214    DnB008  14  0.755  0.763  0.745  0.405  0.005    DnB114  13  0.769  0.807  0.786  0.358  0.019    DnC010  6  0.457  0.438  0.386  0.779  0.030    DnD007  4  0.624  0.718  0.663  0.541  0.073    Mean  7.9 (4.2)  0.581 (0.051)  0.645 (0.058)  0.601 (0.063)    0.058 (0.027)    Total  62.9        0.985    Gale  DnA105  5  0.616  0.626  0.575  0.619  0.003  N = 294  DnA117  4  0.158  0.242  0.229  0.874  0.200    DnA129  10  0.224  0.370  0.349  0.793  0.252    DnB008  5  0.399  0.432  0.406  0.751  0.046    DnB114  18  0.830  0.878  0.864  0.245  0.027    DnC010  6  0.678  0.680  0.622  0.582  0.004    DnD007  5  0.432  0.403  0.371  0.783  0.044    Mean  7.6 (5.0)  0.477 (0.092)  0.519 (0.083)  0.488 (0.081)    0.082 (0.038)    Total  186.3        0.964    Population  Locus  A  HO  HE  PIC  Excl  Null  TutunupA  DnA105  4  0.502  0.537  0.467  0.725  0.039  N = 245  DnA117  5  0.465  0.487  0.433  0.745  0.023    DnA129  9  0.498  0.763  0.730  0.441  0.214    DnB008  14  0.755  0.763  0.745  0.405  0.005    DnB114  13  0.769  0.807  0.786  0.358  0.019    DnC010  6  0.457  0.438  0.386  0.779  0.030    DnD007  4  0.624  0.718  0.663  0.541  0.073    Mean  7.9 (4.2)  0.581 (0.051)  0.645 (0.058)  0.601 (0.063)    0.058 (0.027)    Total  62.9        0.985    Gale  DnA105  5  0.616  0.626  0.575  0.619  0.003  N = 294  DnA117  4  0.158  0.242  0.229  0.874  0.200    DnA129  10  0.224  0.370  0.349  0.793  0.252    DnB008  5  0.399  0.432  0.406  0.751  0.046    DnB114  18  0.830  0.878  0.864  0.245  0.027    DnC010  6  0.678  0.680  0.622  0.582  0.004    DnD007  5  0.432  0.403  0.371  0.783  0.044    Mean  7.6 (5.0)  0.477 (0.092)  0.519 (0.083)  0.488 (0.081)    0.082 (0.038)    Total  186.3        0.964    A, number of alleles per locus; HO, observed heterozygosity; HE, expected heterozygosity; PIC, polymorphic information content; Excl, exclusion probability for the second parent; Null, estimated frequency of null alleles; N, total number of plants sampled, including adult plants and seedling progeny. Standard deviation is given after allele mean numbers. Standard error is in parentheses. View Large Table 1. Allelic diversity values for microsatellite markers in two populations of Banksia nivea ssp. uliginosa Population  Locus  A  HO  HE  PIC  Excl  Null  TutunupA  DnA105  4  0.502  0.537  0.467  0.725  0.039  N = 245  DnA117  5  0.465  0.487  0.433  0.745  0.023    DnA129  9  0.498  0.763  0.730  0.441  0.214    DnB008  14  0.755  0.763  0.745  0.405  0.005    DnB114  13  0.769  0.807  0.786  0.358  0.019    DnC010  6  0.457  0.438  0.386  0.779  0.030    DnD007  4  0.624  0.718  0.663  0.541  0.073    Mean  7.9 (4.2)  0.581 (0.051)  0.645 (0.058)  0.601 (0.063)    0.058 (0.027)    Total  62.9        0.985    Gale  DnA105  5  0.616  0.626  0.575  0.619  0.003  N = 294  DnA117  4  0.158  0.242  0.229  0.874  0.200    DnA129  10  0.224  0.370  0.349  0.793  0.252    DnB008  5  0.399  0.432  0.406  0.751  0.046    DnB114  18  0.830  0.878  0.864  0.245  0.027    DnC010  6  0.678  0.680  0.622  0.582  0.004    DnD007  5  0.432  0.403  0.371  0.783  0.044    Mean  7.6 (5.0)  0.477 (0.092)  0.519 (0.083)  0.488 (0.081)    0.082 (0.038)    Total  186.3        0.964    Population  Locus  A  HO  HE  PIC  Excl  Null  TutunupA  DnA105  4  0.502  0.537  0.467  0.725  0.039  N = 245  DnA117  5  0.465  0.487  0.433  0.745  0.023    DnA129  9  0.498  0.763  0.730  0.441  0.214    DnB008  14  0.755  0.763  0.745  0.405  0.005    DnB114  13  0.769  0.807  0.786  0.358  0.019    DnC010  6  0.457  0.438  0.386  0.779  0.030    DnD007  4  0.624  0.718  0.663  0.541  0.073    Mean  7.9 (4.2)  0.581 (0.051)  0.645 (0.058)  0.601 (0.063)    0.058 (0.027)    Total  62.9        0.985    Gale  DnA105  5  0.616  0.626  0.575  0.619  0.003  N = 294  DnA117  4  0.158  0.242  0.229  0.874  0.200    DnA129  10  0.224  0.370  0.349  0.793  0.252    DnB008  5  0.399  0.432  0.406  0.751  0.046    DnB114  18  0.830  0.878  0.864  0.245  0.027    DnC010  6  0.678  0.680  0.622  0.582  0.004    DnD007  5  0.432  0.403  0.371  0.783  0.044    Mean  7.6 (5.0)  0.477 (0.092)  0.519 (0.083)  0.488 (0.081)    0.082 (0.038)    Total  186.3        0.964    A, number of alleles per locus; HO, observed heterozygosity; HE, expected heterozygosity; PIC, polymorphic information content; Excl, exclusion probability for the second parent; Null, estimated frequency of null alleles; N, total number of plants sampled, including adult plants and seedling progeny. Standard deviation is given after allele mean numbers. Standard error is in parentheses. View Large Outcrossing rate Multi-locus outcrossing in both populations was high (0.942 in Gale and 0.965 in TutunupA) with a small amount of inbreeding from self-pollination and biparental inbreeding (Table 2). The fixation index of mother trees was low in both populations, but was slightly higher in the Gale population (0.013 ± 0.004) than in TutunupA (0.005 ± 0.014). The fixation index of seedlings was also higher in the Gale population (FO = 0.116) than TutunupA population (FO = 0.107). The estimate of biparental mating was nearly twice as high in the TutunupA population (0.113 ± 0.038) than in the Gale population (0.060 ± 0.025). Correlation of selfing among loci indicated 88.5 ± 3.2% of inbreeding in the Gale population was from self-pollination compared to 77.3 ± 7.7% in the TutunupA population. Estimates of the correlation of paternity (rp(m)) indicated the number of pollen donors (Nep) per maternal plant at the Gale population (20.8) was much higher than at TutunupA (5.3). Estimates of mean co-ancestry in adult plants were similar in both populations (TutunupA = −0.008, Gale = 0.001). In seedlings, estimates of mean co-ancestry were higher than that expected for half-sibs, indicating some seedlings within families had the same fathers, which is consistent with estimates of correlated paternity. The TutunupA families were 25% (0.157–0.125/0.125) more closely related than complete half sibs, and this value was 18% for Gale families. Table 2. Estimates of mating system parameters for populations of Banksia nivea ssp. uliginosa Parameter  TutunupA  Gale  Parental F estimate  0.005 (0.014)  0.013 (0.004)  tm  0.965 (0.017)  0.942 (0.018)  1 − tm  0.035  0.058  ts  0.852 (0.049)  0.882 (0.026)  tm − ts  0.113 (0.038)  0.060 (0.025)  rsloc  0.773 (0.077)  0.885 (0.032)  rsfam  0.070 (0.017)  0.092 (0.036)  rp(m)  0.190 (0.036)  0.048 (0.009)  rp(s) − rp(m)  −0.390 (0.036)  −0.248 (0.019)  1/rp(m)  5.26  20.83  Co-ancestry  0.157  0.148  Parameter  TutunupA  Gale  Parental F estimate  0.005 (0.014)  0.013 (0.004)  tm  0.965 (0.017)  0.942 (0.018)  1 − tm  0.035  0.058  ts  0.852 (0.049)  0.882 (0.026)  tm − ts  0.113 (0.038)  0.060 (0.025)  rsloc  0.773 (0.077)  0.885 (0.032)  rsfam  0.070 (0.017)  0.092 (0.036)  rp(m)  0.190 (0.036)  0.048 (0.009)  rp(s) − rp(m)  −0.390 (0.036)  −0.248 (0.019)  1/rp(m)  5.26  20.83  Co-ancestry  0.157  0.148  tm, multilocus population outcrossing rate; ts, single locus population outcrossing rate; rsloc, correlation of selfing among loci; rsfam, correlation of selfing among families (normalized variance of selfing); rp(m), multilocus correlation of paternity (fraction of siblings sharing the same father); rp(s), single locus correlation of paternity; Co-ancestry, Loiselle’s kinship coefficient. Standard error in parentheses estimated based on 500 bootstrap re-samplings. View Large Table 2. Estimates of mating system parameters for populations of Banksia nivea ssp. uliginosa Parameter  TutunupA  Gale  Parental F estimate  0.005 (0.014)  0.013 (0.004)  tm  0.965 (0.017)  0.942 (0.018)  1 − tm  0.035  0.058  ts  0.852 (0.049)  0.882 (0.026)  tm − ts  0.113 (0.038)  0.060 (0.025)  rsloc  0.773 (0.077)  0.885 (0.032)  rsfam  0.070 (0.017)  0.092 (0.036)  rp(m)  0.190 (0.036)  0.048 (0.009)  rp(s) − rp(m)  −0.390 (0.036)  −0.248 (0.019)  1/rp(m)  5.26  20.83  Co-ancestry  0.157  0.148  Parameter  TutunupA  Gale  Parental F estimate  0.005 (0.014)  0.013 (0.004)  tm  0.965 (0.017)  0.942 (0.018)  1 − tm  0.035  0.058  ts  0.852 (0.049)  0.882 (0.026)  tm − ts  0.113 (0.038)  0.060 (0.025)  rsloc  0.773 (0.077)  0.885 (0.032)  rsfam  0.070 (0.017)  0.092 (0.036)  rp(m)  0.190 (0.036)  0.048 (0.009)  rp(s) − rp(m)  −0.390 (0.036)  −0.248 (0.019)  1/rp(m)  5.26  20.83  Co-ancestry  0.157  0.148  tm, multilocus population outcrossing rate; ts, single locus population outcrossing rate; rsloc, correlation of selfing among loci; rsfam, correlation of selfing among families (normalized variance of selfing); rp(m), multilocus correlation of paternity (fraction of siblings sharing the same father); rp(s), single locus correlation of paternity; Co-ancestry, Loiselle’s kinship coefficient. Standard error in parentheses estimated based on 500 bootstrap re-samplings. View Large Paternity assignment and pollen dispersal Paternity assignment identified most likely fathers for 275 progeny (70%) across the two populations. The TutunupA population had a high outcrossing rate between individuals within the population (72.5%), 23% of pollen was assumed to come from outside the population and 4.5% of seedlings were the result of selfing (Table 3). The Gale population had a slightly lower within-patch mating rate of 64.8%, 30% of the seedlings were not assigned paternity from the genotyped plants within the patch and 5.1% of seedlings were assigned to the mother plants. Cryptic pollen flow was relatively high in both populations, 0.493 and 0.680 for the TutunupA and Gale populations, respectively, meaning estimates of pollen immigration may be underestimates by up to 32 and 36% for TutunupA and Gale populations, respectively. Table 3. Selfing and outcrossing rates for seedlings from two populations of Banksia nivea ssp. uliginosa assessed using paternity assignment Population  Mother  Number of progeny  Selfing rate (%)  Outcrossing within sampled patch (%)  Apparent pollen immigration to patch (%)  TutunupA  1  20  0  50  50    5  20  0  100  0    10  20  5  75  20    14  20  0  85  15    23  20  5  85  10    33  20  5  65  30    34  20  10  75  15    36  20  5  65  30    37  20  5  90  5    42  20  10  35  55    Mean  20  4.5  72.5  23  Gale  1  15  0  40  60    5  20  5  80  15    17  20  5  65  30    18  18  11  78  11    30  20  15  50  35    31  20  0  80  20    40  20  0  90  10    71  20  5  65  30    83  20  0  50  50    95  20  10  50  40    Mean  19.3  5.1  64.8  30.1  Population  Mother  Number of progeny  Selfing rate (%)  Outcrossing within sampled patch (%)  Apparent pollen immigration to patch (%)  TutunupA  1  20  0  50  50    5  20  0  100  0    10  20  5  75  20    14  20  0  85  15    23  20  5  85  10    33  20  5  65  30    34  20  10  75  15    36  20  5  65  30    37  20  5  90  5    42  20  10  35  55    Mean  20  4.5  72.5  23  Gale  1  15  0  40  60    5  20  5  80  15    17  20  5  65  30    18  18  11  78  11    30  20  15  50  35    31  20  0  80  20    40  20  0  90  10    71  20  5  65  30    83  20  0  50  50    95  20  10  50  40    Mean  19.3  5.1  64.8  30.1  View Large Table 3. Selfing and outcrossing rates for seedlings from two populations of Banksia nivea ssp. uliginosa assessed using paternity assignment Population  Mother  Number of progeny  Selfing rate (%)  Outcrossing within sampled patch (%)  Apparent pollen immigration to patch (%)  TutunupA  1  20  0  50  50    5  20  0  100  0    10  20  5  75  20    14  20  0  85  15    23  20  5  85  10    33  20  5  65  30    34  20  10  75  15    36  20  5  65  30    37  20  5  90  5    42  20  10  35  55    Mean  20  4.5  72.5  23  Gale  1  15  0  40  60    5  20  5  80  15    17  20  5  65  30    18  18  11  78  11    30  20  15  50  35    31  20  0  80  20    40  20  0  90  10    71  20  5  65  30    83  20  0  50  50    95  20  10  50  40    Mean  19.3  5.1  64.8  30.1  Population  Mother  Number of progeny  Selfing rate (%)  Outcrossing within sampled patch (%)  Apparent pollen immigration to patch (%)  TutunupA  1  20  0  50  50    5  20  0  100  0    10  20  5  75  20    14  20  0  85  15    23  20  5  85  10    33  20  5  65  30    34  20  10  75  15    36  20  5  65  30    37  20  5  90  5    42  20  10  35  55    Mean  20  4.5  72.5  23  Gale  1  15  0  40  60    5  20  5  80  15    17  20  5  65  30    18  18  11  78  11    30  20  15  50  35    31  20  0  80  20    40  20  0  90  10    71  20  5  65  30    83  20  0  50  50    95  20  10  50  40    Mean  19.3  5.1  64.8  30.1  View Large At TutunupA, the observed pollen dispersal distance ranged from 0.4 to 78.1 m with a mean distance per mother plant of 19.7 ± 4.4 m. A majority of mating events (67%) occurred with pollen from plants within 10 m of a mother plant (Fig. 3A). In the Gale population, the observed pollen dispersal ranged from 1.0 to 75.0 m with a mean distance per mother plant of 30.8 ± 2.7 m. The majority of mating events (73%) occurred with pollen from plants located within 40 m of a mother plant (Fig. 3B). The distribution of observed and potential pollen dispersal distance within both populations was similar, except for the closest distance category (0−10 m) where distance to observed pollen parents was significantly higher than potential pollen parents in both populations (TutunupA: χ2 = 106.1, d.f. = 9, P < 0.001; Gale: χ2 = 69.0, d.f. = 9, P < 0.001). Figure 3. View largeDownload slide Comparison of potential (light bars) and observed (dark bars) pollen dispersal distance within the (A) TutunupA and (B) Gale populations in Banksia nivea ssp. uliginosa. Potential dispersal distance is the average distance from each materal plant to all potential fathers, and observed dispersal distance is the average distance to the assigned fathers based on paternity assignment. Pollen disperal distance is measured in 10-m classes to 100 m. Asterisks indicate significant differences between observed and expected pollen dispersal distance, P < 0.001. Figure 3. View largeDownload slide Comparison of potential (light bars) and observed (dark bars) pollen dispersal distance within the (A) TutunupA and (B) Gale populations in Banksia nivea ssp. uliginosa. Potential dispersal distance is the average distance from each materal plant to all potential fathers, and observed dispersal distance is the average distance to the assigned fathers based on paternity assignment. Pollen disperal distance is measured in 10-m classes to 100 m. Asterisks indicate significant differences between observed and expected pollen dispersal distance, P < 0.001. Analysis of pollen dispersal based on the Neighbourhood model gave the following estimates for the TutunupA and Gale populations, respectively: selfing rates of 7.1 ± 2.2 and 4.5 ± 0.0%, shape of dispersal kernel (bp) of 1.3 and 0.9, mean distances of pollen dispersal kernel (dp) of 26.9 and 30.9 m, mean within-neighbourhood pollen dispersal distances of 13.5 and 20.4 m and percentages of pollen immigration (mp) coming from outside the sampling area of 18.6 ± 3.6 and 28.5 ± 0.0%. The TWOGENER AMOVA indicated high levels of global pollen pool differentiation within sampled populations, ΦST (Gale) = 0.224, ΦST (TutunupA) = 0.331, indicating that pollinators were sampling different pollen pools within the populations. DISCUSSION Our investigation has shown that the pattern of pollen dispersal within two populations of Banksia nivea ssp. uliginosa is consistent with expectations based on the size and density of population patches. We found more restricted dispersal within the linear, smaller and more clumped population, consistent with theoretical expectations of the influence of size and spatial arrangements of plants. Although present in a naturally fragmented landscape, B. nivea ssp. uliginosa showed high outcrossing, with a small amount of selfing in the large population, and mating between relatives in the small population. There was evidence of reasonable levels of pollen immigration from the adjacent populations supplementing the within-population mating. Predominant outcrossing with low level of selfing Our analysis shows that B. nivea ssp. uliginosa is predominantly outcrossing, with a small amount of selfing. This is consistent with the outcrossing rate of most studied banksias that ranges from 0.670 to 0.950 (Scott, 1980; Coates et al., 2007; Collins, Walsh & Grey, 2008; Llorens et al., 2012), although the rate observed in B. nivea ssp. uliginosa is at the high end of this range. In this study, a higher fixation index in seedlings than in adult plants indicates that selection occurs during establishment of adult plants. Removal of inbred seed through selection protects populations against potential genetic diversity loss, and our study shows that both studied populations of B. nivea ssp. uliginosa have high levels of genetic diversity despite occurring in a naturally fragmented landscape where genetic diversity is expected to be reduced. The mean co-ancestry values for the adult plants in both populations were not different from zero, indicating negligible relatedness among the established adult plants, and little effect of population size or density. This low relatedness is reflected in the level of genetic diversity of these B. nivea ssp. uliginosa populations (A = 7.7, HO = 0.529), which is comparable to other banksia species with more widespread distributions, such as B. brownii (A = 3.4, HO = 0.472; Coates, McArthur & Byrne, 2015), B. attenuata (A = 5.8, HE = 0.581−0.778; He et al., 2008) and B. sphaerocarpa ssp. sphaerocarpa (A = 11.5, HO = 0.670; Nistelberger et al., 2015). Nonetheless, the mean co-ancestry estimates of seedlings in both populations were higher than that expected from complete half sibs, consistent with estimates of some correlated paternity. A theoretical consequence of small population size is that there are fewer mates available leading to higher selfing and higher correlated paternity in such populations (Young et al., 1996; Honnay & Jacquemyn, 2007; Aguilar et al., 2008; Eckert et al., 2010). Indeed, population size has been shown to have an effect on outcrossing and on correlated paternity in species with mixed mating systems, such as the shrubs Calothamnus quadrifidus (Yates et al., 2007a) and B. sphaerocarpa ssp. sphaerocarpa (Llorens et al., 2012). In B. nivea ssp. uliginosa, the inbreeding level was low in both populations, indicating no effect of different population size, probably a consequence of the high outcrossing rate and low relatedness of plants. However, there were some differences between the populations, as inbreeding at TutunupA was mainly due to the low effective number of fathers, higher mating between relatives and higher number of progeny sharing the same father, while inbreeding in Gale was mostly caused by selfing. Influence of plant distribution on pollen dispersal Although the number of plants in the population is important in terms of mate availability, plant density also affects pollinator movement, with greater pollinator movement between plants over greater distances in less dense populations (Levin & Kerster, 1974). The TutunupA population showed most mating events occurred within 10 m of a mother plant while mating in the Gale population occurred up to 40−50 m from a mother plant. Subsequently, pollen dispersal kernel at TutunupA had an exponential shape while was a fat-tailed dispersal kernel at Gale. This difference in dispersal kernel is consistent with a shorter mean within-neighbourhood pollen distance in the TutunupA population in comparison with the Gale population. This difference is also likely to be related to the different spatial arrangement of plants within the populations, as plant density is known to affect pollinator movement and to affect different pollinators differently (Levin & Kerster, 1974; Ashley, 2010). For example, in a study of mating in Grevillea macleayana, bees tended to move within plants while birds showed more between-plant flights than bees (Whelan, Ayre & Beynon, 2009), and analysis of pollen dispersal in the bird- and mammal-pollinated B. sphaerocarpa ssp. sphaerocarpa showed patterns that were highly correlated with plant density (Llorens et al., 2012). Banksia nivea ssp. uliginosa is primarily pollinated by small mammals with some contribution from birds, but there is little known about mammal pollinator movement. The honey possum, Tarsipes rostratus, is the most likely main pollinator, as it is a common visitor to banksia flowers, and although these animals have been observed not move more than 30 m even over several months (Garavanta, Wooller & Richardson, 2000; Bradshaw et al., 2007), the present study suggests these mammals may move further in populations with lower plant density. Bird pollinators, such as honeyeaters, are likely to facilitate long-distance pollen dispersal of B. nivea ssp. uliginosa as they have been observed carrying pollen of banksia species (Wooller, 2001, 2002), and have been attributed to contribute to pollen dispersal of C. quadrifidus between fragments over 5 km (Byrne et al., 2007). Different spatial arrangements of plants in this study may also influence bird pollinator movement, similar to the increased movement of bird pollinators between C. quadrifidus plants in denser populations (Yates et al., 2007a). Both populations showed a high level of pollen pool differentiation, indicating plants were accessing different pollen pools within the population. This may be due to different plants flowering at different times or flowers on plants becoming receptive at different times during the season. Flowering in B. nivea ssp. uliginosa occurs over several months within populations, over at least 1 month in large plants and over several days in an inflorescence. This is similar to another banksia species in which anthesis varied between 7 and 33 days (Saffer, 2004; Wooller & Wooller, 2003). Despite having a clumped distribution of plants, the TutunupA population had a more highly differentiated pollen pool and received pollen from fewer mates than the larger and less dense population at Gale, suggesting that population size rather than density has a greater effect on maintaining pollen pool homogeneity in this species. This is consistent with a study of B. sphaerocarpa ssp. sphaerocarpa where larger population size was correlated with lower pollen pool differentiation and greater number of mates (Llorens et al., 2012). Interestingly, this is in contrast to a similar study on the bird-pollinated C. quadrifidus, which showed lower pollen pool heterogeneity in the denser population, which is probably explained by concurrent flowering among plants in this species (Byrne et al., 2007). Estimation of pollen immigration into the sampled patches was variable across mother plants in the studied populations. Estimates of 23–30% of progeny (and up to 55 and 66% [for TutunupA and Gale, respectively] based on estimates of cryptic pollen flow) were sired from pollen derived from outside the sampled patches, consistent with expectations of pollination by mobile pollinators such as mammals and birds. Although this pollen immigration was from nearby plants within the broader populations, levels of pollen immigration have been shown to be extensive in fragmented landscapes in south-western Australia, with values ranging from 3 to 64% over distances of up to 5 km (Byrne et al., 2007, 2008; Sampson & Byrne, 2008; Krauss et al., 2009; Llorens et al., 2012; Millar et al., 2012; Sampson et al., 2014), and influenced by a range of variables, including isolation, size, density and pollinator type. The accumulation of these data on pollen immigration demonstrates that pollen dispersal is not limited in fragmented landscapes and provides a means of achieving genetic connectivity among populations. CONCLUSIONS Despite predominant pollination by a non-flying mammal that is likely to have a limited foraging range, we found high outcrossing and minimal selfing in the populations of B. nivea ssp. uliginosa studied here, and up to 30% of progeny were derived from mating with fathers outside the sampled patches. Pollen dispersal was affected by adult plant distribution and density as expected from predominantly mammal pollination, although high levels of pollen immigration are consistent with some contribution from bird pollination. Animal visitation is essential for pollination, so maintaining a viable population of pollinators is important in ensuring pollination and gene flow in fragmented landscapes. An effective pollinator community is also dependent on maintenance of the species pool it forages on, so maintaining habitat context for rare species is an important component of management programmes. ACKNOWLEDGMENTS We would like to thank three anonymous reviewers for their comments on the manuscript. We thank Bronwyn Macdonald for assistance in the laboratory, Margaret Langley and Belinda Newman for assistance with sample collection, Melissa Millar for help with the ArcView program and Jane Sampson for advice on analysis. The project was supported by funding from the South West Catchments Council. REFERENCES Aguilar R, Ashworth L, Galetto L, Aizen MA. 2006. Plant reproductive susceptibility to habitat fragmentation: review and synthesis through a meta-analysis. Ecology Letters  9: 968– 980. Google Scholar CrossRef Search ADS   Aguilar R, Quesada M, Ashworth L, Herrerias-Diego Y, Lobo J. 2008. Genetic consequences of habitat fragmentation in plant populations: susceptible signals in plant traits and methodological approaches. Molecular Ecology  17: 5177– 5188. Google Scholar CrossRef Search ADS   Amos B. 2000. NEWPAT, a general paternity program, 5thedn . Available at:http://www.zoo.cam.ac.uk/zoostaff/amos/ newpat.htm. (accessed 19 May 2009). Ashley MV. 2010. Plant parentage, pollination, and dispersal: How DNA microsatellites have altered the landscape. Critical Reviews in Plant Sciences  29: 148– 161. Google Scholar CrossRef Search ADS   Bennett A. 2003. Habitat fragmentation. In: Attiwill P, Wilson B, eds. Ecology: an Australian perspective . Melbourne: Oxford University Press, 440– 456. Bradshaw SD, Phillips RD, Tomlinson S, Holley RJ, Jennings S, Bradshaw FJ. 2007. Ecology of the Honey possum, Tarsipes rostratus, in Scott National Park, Western Australia. Australian Mammalogy  29: 25– 38. Google Scholar CrossRef Search ADS   Brown A, Thomson-Dans C, Marchant N, eds. 1998. Western Australia’s threatened flora . Perth: Department of Conservation and Land Management. Burbidge AH, James SH. 1991. Postzygotic seed abortion in the genetic system of Stylidium (Angiospermae: Stylidiaceae). Journal of Heredity  82: 319– 328. Google Scholar CrossRef Search ADS   Byrne M, Elliott CP, Yates C, Coates DJ. 2007. Extensive pollen dispersal in a bird-pollinated shrub, Calothamnus quadrifidus, in a fragmented landscape. Molecular Ecology  16: 1303– 1314. Google Scholar CrossRef Search ADS   Byrne M, Elliott CP, Yates CJ, Coates DJ. 2008. Maintenance of high pollen dispersal in Eucalyptus wandoo, a dominant tree of the fragmented agricultural region in Western Australia. Conservation Genetics  9: 97– 105. Google Scholar CrossRef Search ADS   Byrne M, Macdonald B, Francki M. 2001. Incorporation of sodium sulfite into extraction protocol minimizes degradation of Acacia DNA. Biotechniques  30: 742– 744, 748. Byrne M, Parrish TL, Moran GF. 1998. Nuclear RFLP diversity in Eucalyptus nitens. Heredity  81: 225– 232. Google Scholar CrossRef Search ADS   Cavanagh T, Pieroni M. 2006. The dryandras . Melbourne: Australian Plants Society (SGAP Victoria) Inc. and Wildflower Society of Western Australia Inc. Chybicki IJ, Burczyk J. 2010. NM+: software implementing parentage-based models for estimating gene dispersal and mating patterns in plants. Molecular Ecology Resources  10: 1071– 1075. Google Scholar CrossRef Search ADS   Coates DJ, Byrne M. 2005. Genetic variation in plant populations: assessing cause and pattern. In: Henry RJ, ed. Plant diversity and evolution: genotypic and phenotypic variation in higher plants . Boston: CABI Publishing, 139– 164. Google Scholar CrossRef Search ADS   Coates DJ, McArthur SL, Byrne M. 2015. Significant genetic diversity loss following pathogen driven population extinction in the rare endemic Banksia brownii (Proteaceae). Biological Conservation  192: 353– 360. Google Scholar CrossRef Search ADS   Coates DJ, Sampson JF, Yates CJ. 2007. Plant mating systems and assessing population persistence in fragmented landscapes. Australian Journal of Botany  55: 239– 249. Google Scholar CrossRef Search ADS   Collins BG, Walsh M, Grey J. 2008. Floral development and breeding systems of Dryandra sessilis and Grevillea wilsonii (Proteaceae). Australian Journal of Botany  56: 119– 130. Google Scholar CrossRef Search ADS   Dixon K, Dixon B, Krauss S. 2003. Kings Park and Botanic Gardens (BGPA) Science Directorate research proposal for the rescue of four rare and endangered species at BHP Beenup minesite . Perth: BGPA. Dow BD, Ashley MV. 1998. High levels of gene flow in bur oak revealed by paternity analysis using microsatellites. Journal of Heredity  89: 62– 70. Google Scholar CrossRef Search ADS   Dudash MR, Fenster CB. 2000. Inbreeding and outbreeding depression in fragmented populations. In: Young AG, Clarke GM, eds. Conservation biology 4: genetics, demography and viability of fragmented populations . Cambridge: Cambridge University Press, 35– 45. Google Scholar CrossRef Search ADS   Eckert CG, Kalisz S, Geber MA, Sargent R, Elle E, Cheptou PO, Goodwillie C, Johnston MO, Kelly JK, Moeller DA, Porcher E, Ree RH, Vallejo-Marín M, Winn AA. 2010. Plant mating systems in a changing world. Trends in Ecology & Evolution  25: 35– 43. Google Scholar CrossRef Search ADS   Ellstrand NC. 2014. Is gene flow the most important evolutionary force in plants? American Journal of Botany  101: 737– 753. Google Scholar CrossRef Search ADS   English V. 1999. Interim recovery plan no. 44: Shrubland association on Southern Swan Coastal Plain Ironstone (Busselton Area) (Southern Ironstone Association) Interim recovery plan 1999–2002 . Wanneroo: Department of Conservation and Land Management Western Australian Threatened Species and Communities Unit. Excoffier L, Smouse PE, Quattro JM. 1992. Analysis of molecular variance inferred from metric distances among DNA haplotypes: application to human mitochondrial DNA restriction data. Genetics  131: 479– 491. Frankham R, Ballou JD, Briscoe DA. 2002. Introduction to conservation genetics . Cambridge: Cambridge University Press. Google Scholar CrossRef Search ADS   Garavanta CAM, Wooller RD, Richardson KC. 2000. Movement patterns of honey possums, Tarsipes rostratus, in the Fitzgerald River National Park, Western Australia. Wildlife Research  27: 179– 183. Google Scholar CrossRef Search ADS   George AS. 1996. New taxa and a new infrageneric classification in Dryandra R.Br. (Proteaceae: Grevilleoideae). Nuytsia  10: 313– 408. Haddad NM, Brudvig LA, Clobert J, Davies KF, Gonzalez A, Holt RD, Lovejoy TE, Sexton JO, Austin MP, Collins CD, Cook WM, Damschen EI, Ewers RM, Foster BL, Jenkins CN, King AJ, Laurance WF, Levey DJ, Margules CR, Melbourne BA, Nicholls AO, Orrock JL, Song DX, Townshend JR. 2015. Habitat fragmentation and its lasting impact on Earth’s ecosystems. Science Advances  1: e1500052. Google Scholar CrossRef Search ADS   Hardy OJ, Vekemans X. 2002. SPAGeDi: a versatile computer program to analyse spatial genetic structure at the individual or population levels. Molecular Ecology Notes  2: 618– 620. Google Scholar CrossRef Search ADS   He T, Lamont BB, Krauss SL, Enright NJ, Millar BP. 2008. Covariation between intraspecific genetic diversity and species diversity within a plant functional group. Journal of Ecology  96: 956– 961. Google Scholar CrossRef Search ADS   Honnay O, Jacquemyn H. 2007. Susceptibility of common and rare plant species to the genetic consequences of habitat fragmentation. Conservation Biology  21: 823– 831. Google Scholar CrossRef Search ADS   Krauss SL, He T, Barrett LG, Lamont BB, Enright NJ, Miller BP, Hanley ME. 2009. Contrasting impacts of pollen and seed dispersal on spatial genetic structure in the bird-pollinated Banksia hookeriana. Heredity  102: 274– 285. Google Scholar CrossRef Search ADS   Levin DA, Kerster HW. 1974. Gene flow in seed plants. In: Dobzhansky T, Hecht MK, Steere WC, eds. Evolutionary biology, Vol. 7 . Boston: Springer US, 139– 220. Google Scholar CrossRef Search ADS   Llorens TM, Byrne M, Yates CJ, Nistelberger HM, Coates DJ. 2012. Evaluating the influence of different aspects of habitat fragmentation on mating patterns and pollen dispersal in the bird-pollinated Banksia sphaerocarpa var. caesia. Molecular Ecology  21: 314– 328. Google Scholar CrossRef Search ADS   Loiselle BA, Sork VL, Nason J, Graham C. 1995. Spatial genetic structure of a tropical understory shrub, Psychotria officinalis (Rubiaceae). American Journal of Botany  82: 1420– 1425. Google Scholar CrossRef Search ADS   Lowe A, Harris S, Ashton P. 2004. Ecological genetics: design, analysis, and application . Carlton: Blackwell Science Ltd. Marshall TC, Slate J, Kruuk LE, Pemberton JM. 1998. Statistical confidence for likelihood-based paternity inference in natural populations. Molecular Ecology  7: 639– 655. Google Scholar CrossRef Search ADS   Mast AR, Thiele K. 2007. The transfer of Dryandra R.Br. to Banksia L.f. (Proteaceae). Australian Systematic Botany  20: 63– 71. Google Scholar CrossRef Search ADS   Mendonça AH, Russo C, Melo ACG, Durigan G. 2015. Edge effects in savanna fragments: a case study in the cerrado. Plant Ecology & Diversity  8: 493– 503. Google Scholar CrossRef Search ADS   Millar MA, Byrne M. 2008. Characterization of polymorphic microsatellite DNA markers in Banksia nivea, formerly Dryandra nivea. Molecular Ecology Resources  8: 1393– 1394. Google Scholar CrossRef Search ADS   Millar MA, Byrne M, Nuberg IK, Sedgley M. 2012. High levels of genetic contamination in remnant populations of Acacia saligna from a genetically divergent planted stand. Restoration Ecology  20: 260– 267. Google Scholar CrossRef Search ADS   Nei M. 1977. F-statistics and analysis of gene diversity in subdivided populations. Annals of Human Genetics  41: 225– 233. Google Scholar CrossRef Search ADS   Nelson‐Tunley M, Morgan‐Richards M, Trewick SA. 2016. Genetic diversity and gene flow in a rare New Zealand skink despite fragmented habitat in a volcanic landscape. Biological Journal of the Linnean Society  119: 37– 51. Google Scholar CrossRef Search ADS   Nistelberger HM, Coates DJ, Llorens TM, Yates CJ, Byrne M. 2015. A cryptic genetic boundary in remnant populations of a long-lived, bird-pollinated shrub Banksia sphaerocarpa var. caesia (Proteaceae). Biological Journal of the Linnean Society  115: 241– 255. Google Scholar CrossRef Search ADS   O’Connell LM, Mosseler A, Rajora OP. 2006. Impacts of forest fragmentation on the reproductive success of white spruce (Picea glauca). Canadian Journal of Botany  84: 956– 965. Google Scholar CrossRef Search ADS   Ouborg NJ, Vergeer P, Mix C. 2006. The rough edges of the conservation genetics paradigm for plants. Journal of Ecology  94: 1233– 1248. Google Scholar CrossRef Search ADS   Pannell JR, Fields PD. 2014. Evolution in subdivided plant populations: concepts, recent advances and future directions. New Phytologist  201: 417– 432. Google Scholar CrossRef Search ADS   Peakall R, Smouse PE. 2006. GENALEX 6: genetic analysis in Excel. Population genetic software for teaching and research. Molecular Ecology Notes  6: 288– 295. Google Scholar CrossRef Search ADS   R Core Team. 2017. R: A language and environment for statistical computing . Vienna: R Foundation for Statistical Computing. Available at http://www.R-project.org/. (accessed 18 December 2017). Reed DH. 2005. Relationship between population size and fitness. Conservation Biology  19: 563– 568. Google Scholar CrossRef Search ADS   Richards CM. 2000. Genetic and demographic influences on population persistence: gene flow and genetic rescue in Silene alba. In: Young AG, Clarke GM, eds. Conservation biology 4: genetics, demography and viability of fragmented populations . Cambridge: Cambridge University Press, 271– 292. Google Scholar CrossRef Search ADS   Ritland K. 2002. Extensions of models for the estimation of mating systems using n independent loci. Heredity  88: 221– 228. Google Scholar CrossRef Search ADS   Saffer VM. 2004. Are diel patterns of nectar production and anthesis associated with other floral traits in plants visited by potential bird and mammal pollinators? Australian Journal of Botany  52: 87– 92. Google Scholar CrossRef Search ADS   Sampson JF, Byrne M. 2008. Outcrossing between an agroforestry plantation and remnant native populations of Eucalyptus loxophleba. Molecular Ecology  17: 2769– 2781. Google Scholar CrossRef Search ADS   Sampson JF, Byrne M, Yates CJ, Gibson N, Thavornkanlapachai R, Stankowski S, MacDonald B, Bennett I. 2014. Contemporary pollen-mediated gene immigration reflects the historical isolation of a rare, animal-pollinated shrub in a fragmented landscape. Heredity  112: 172– 181. Google Scholar CrossRef Search ADS   Sampson JF, Hankinson M, McArthur S, Tapper S, Langley M, Gibson N, Yates C, Byrne M. 2015. Long-term ‘islands’ in the landscape: low gene flow, effective population size and genetic divergence in the shrub Hakea oldfieldii (Proteaceae). Botanical Journal of the Linnean Society  179: 319– 334. Google Scholar CrossRef Search ADS   Scott JK. 1980. Estimation of the outcrossing rate for Banksia attenuata R.Br. and Banksia menziesii R.Br. (Proteaceae). Australian Journal of Botany  28: 53– 59. Google Scholar CrossRef Search ADS   Sebbenn AM. 2006. Sistema de reprodução em espécies arbóreas tropicais e suas implicações para a seleção de árvores matrizes para reflorestamentos ambientais. In: Higa AR, Silva L, eds. Pomares de sementes de espécies florestais nativas . Curitiba: FUPEF, 193– 198. Segelbacher G, Cushman SA, Epperson BK, Fortin M-J, Francois O, Hardy OJ, Holderegger R, Taberlet P, Waits LP, Manel S. 2010. Applications of landscape genetics in conservation biology: concepts and challenges. Conservation Genetics  11: 375– 385. Google Scholar CrossRef Search ADS   Smouse PE, Dyer RJ, Westfall RD, Sork VL. 2001. Two-generation analysis of pollen flow across a landscape. I. Male gamete heterogeneity among females. Evolution  55: 260– 271. Google Scholar CrossRef Search ADS   Whelan RJ, Ayre DJ, Beynon FM. 2009. The birds and the bees: pollinator behaviour and variation in the mating system of the rare shrub Grevillea macleayana. Annals of Botany  103: 1395– 1401. Google Scholar CrossRef Search ADS   Wooller RD, Wooller SJ. 2003. The role of non-flying animals in the pollination of Banksia nutans. Australian Journal of Botany  51: 503– 507. Google Scholar CrossRef Search ADS   Wooller SJ, Wooller RD. 2001. Seed set in two sympatric banksias, Banksia attenuata and B. baxteri. Australian Journal of Botany  49: 597– 602. Google Scholar CrossRef Search ADS   Wooller SJ, Wooller RD. 2002. Mixed mating in Banksia media. Australian Journal of Botany  50: 627– 631. Google Scholar CrossRef Search ADS   Yates C, Coates DJ, Elliott CP, Byrne M. 2007a. Composition of the pollinator community, pollination and the mating system for a shrub in fragments of species rich kwongan in south-west Western Australia. Biodiversity and Conservation  16: 1379– 1395. Google Scholar CrossRef Search ADS   Yates CJ, Elliott CP, Byrne M, Coates DJ, Fairman R. 2007b. Seed production, germinability and seedling growth for a bird-pollinated shrub in fragments of kwongan in south-west Australia. Biological Conservation  136: 306– 314. Google Scholar CrossRef Search ADS   Young A, Boyle T, Brown T. 1996. The population genetic consequences of habitat fragmentation for plants. Trends in Ecology & Evolution  11: 413– 418. Google Scholar CrossRef Search ADS   © 2018 The Linnean Society of London, Biological 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) http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Biological Journal of the Linnean Society Oxford University Press

Population density and size influence pollen dispersal pattern and mating system of the predominantly outcrossed Banksia nivea (Proteaceae) in a threatened ecological community

Download PDF
12 pages

Loading next page...
 
/lp/ou_press/population-density-and-size-influence-pollen-dispersal-pattern-and-gWWpebCbcI

References (64)

Publisher
Oxford University Press
Copyright
© 2018 The Linnean Society of London, Biological Journal of the Linnean Society
ISSN
0024-4066
eISSN
1095-8312
DOI
10.1093/biolinnean/bly050
Publisher site
See Article on Publisher Site

Abstract

Abstract Gene flow is a critical component of plant mating systems and influences population fitness. However, pollen dispersal can be highly variable and influenced by natural and anthropogenic fragmentation. Gene flow through pollen dispersal was investigated in two populations of contrasting size and habitat context in Banksia nivea ssp. uliginosa, a rare species in the Busselton ironstone threatened ecological community of Western Australia with a naturally fragmented distribution. Paternity analysis was conducted with seven microsatellite loci to determine mating system parameters and patterns of pollen dispersal. Outcrossing was high in both populations with a similar level of selfing for both populations despite differences in population size, density and vegetation matrix. Most mating occurred within 10 m of a mother plant in the small, clumped population, while more dispersed mating, up to 50 m from a mother plant, was recorded in the large, less dense population. Our results show that population density and size are important influences on mating system parameters and level of pollen dispersal. gene flow, fragmentation, microsatellite, paternity analysis, pollinator movement INTRODUCTION Gene flow through pollen dispersal is necessary to maintain long-term population viability (Segelbacher et al., 2010), and is largely dependent on the mechanisms of dispersal or mode of reproduction and the degree of connectivity among populations (Ellstrand, 2014; Lowe, Harris & Ashton, 2004; Pannell & Fields, 2014). Fragmentation creates physical barriers by increasing isolation among patches and reducing total habitat area (Frankham, Ballou & Briscoe, 2002). Small, irregularly shaped fragments can exacerbate edge effects on populations within a fragment (Bennett, 2003), expose fragments to an increased risk of disease or exotic species invasion (Mendonça et al., 2015), affect pollen dispersal and the behaviour of pollinators (Byrne et al., 2007; Yates et al., 2007a; Llorens et al., 2012) and change community dynamics (Haddad et al., 2015). Roads and fences can alter the shape and distribution of small and degraded remnant populations to form linear strips with a high edge to area ratio. In addition, small population fragments may receive fewer pollinator visits because they do not provide enough flowering plants to attract or support pollinators (Yates et al., 2007a). Fragmentation impacts pollinators differently depending on several factors, such as number of fragments, size and spatial pattern of fragments, time since fragmentation, dispersal ability of the species, diversity of resources for pollinators, and the environment around the fragments and its impact on foraging range (Aguilar et al., 2006). These factors may lead to a reduction or loss of pollinators, and/or changed patterns of pollinator behaviour, causing reduced reproductive success of plant species that depend on these pollinators (Reed, 2005). Isolation and small population size also increase the chance of self-pollination, mating between related individuals and correlated mating as the result of fewer mates (Ouborg, Vergeer & Mix, 2006; Yates et al., 2007b), as well as reduction in pollen immigration (Byrne et al., 2007). Due to the large number of factors influencing both historical and contemporary gene flow in fragmented landscapes, it is not surprising that a range of responses of plant populations to fragmentation has been shown (Young, Boyle & Brown, 1996; Honnay & Jacquemyn, 2007; Aguilar et al., 2008; Eckert et al., 2010). Limited gene flow can result in genetic drift or differential selection in local environments leading to genetic differentiation between populations (Coates & Byrne, 2005). An increase in selfing in normally outcrossed species may lead to reduced genetic diversity and increased differentiation between populations (Coates, Sampson & Yates, 2007). Increases in inbreeding can lead to expression of lethal or harmful recessive mutations that reduce rates of growth and flowering, decrease pollen quantity, lower seed production, reduce fecundity through low germination success, and increase mortality (Burbidge & James, 1991; Dudash & Fenster, 2000; Richards, 2000; O’Connell, Mosseler & Rajora, 2006). However, some studies have shown no significant difference in estimates of outcrossing rate in large or small populations in fragmented landscapes (Coates et al., 2007; Yates et al., 2007a). One possible explanation for this is that bird pollinators transport pollen between populations and buffer small populations from reduced mate availability through increased effective population size (Yates et al., 2007a). Another explanation is the presence of post-zygotic seed abortion mechanisms involving deleterious recessive genes that selectively decrease progeny numbers but increase the multilocus outcrossing rate (Burbidge & James, 1991). In addition, species in naturally fragmented populations may be adapted to this landscape context and be less affected by anthropogenic fragmentation (Sampson et al., 2014, 2015; Nelson-Tunley, Morgan Richards & Trewick, 2016). The Busselton ironstone ecological community of Western Australia provides a suitable landscape context to investigate the effects of natural and anthropogenic fragmentation on gene flow through pollen dispersal. This community occurs on ironstone soils that are associated with shallow seasonal inundation as a consequence of the impermeable nature of subsurface layers of ironstone and associated heavy soil (English, 1999). The plant community on these ironstone soils is listed as a Critically Endangered Threatened Ecological Community due to its restricted occurrence (Fig. 1) and there are also several rare endemic species restricted to the community. The community has been affected by land clearing, dieback disease (particularly affecting species in the family Proteaceae), weed invasion, grazing, altered fire regimes, hydrological change and salinity (English, 1999). Figure 1. View largeDownload slide Vegetation layer and presence of ironstone in the Busselton area of Western Australia showing natural fragmentation of ironstone soil patches and anthropogenic fragmentation of vegetation through clearing. Remaining vegetation is shown in grey, cleared land is shown in white, and ironstone soils are outlined in red. Location of the Tutunup and Gale populations indicated by arrows. Note Banksia nivea ssp. uliginosa only occurs in 15 populations on some ironstone soils with remaining vegetation. Figure 1. View largeDownload slide Vegetation layer and presence of ironstone in the Busselton area of Western Australia showing natural fragmentation of ironstone soil patches and anthropogenic fragmentation of vegetation through clearing. Remaining vegetation is shown in grey, cleared land is shown in white, and ironstone soils are outlined in red. Location of the Tutunup and Gale populations indicated by arrows. Note Banksia nivea ssp. uliginosa only occurs in 15 populations on some ironstone soils with remaining vegetation. Banksia nivea ssp. uliginosa (A.S. George) A.R. Mast & K.R. Thiele (swamp honeypot) is one such species that is restricted to the ironstone habitat around Busselton (George, 1996; Mast & Thiele, 2007) and has a naturally fragmented population distribution (Fig. 1). It is isolated from the common subspecies by >100 km. While the ecological characteristics of B. nivea (formerly Dryandra nivea) are well described by George (1996) and Cavanagh & Pieroni (2006), little is known about its reproductive biology. A pollinator exclusion experiment has shown that mammals are the predominant pollinator with some bird pollination (R. Thavornkanlapachai, unpubl. data). Based on primarily mammal pollination with some bird pollination, we predict that patterns of pollen dispersal both within and between populations would be influenced by population size and isolation, with lower pollen dispersal distance in smaller road verge populations compared to larger populations in nature reserves. This study aimed to test the hypothesis that pollen dispersal in B. nivea ssp. uliginosa in fragmented populations is consistent with population characteristics of size and density, through examining the mating system and pollen dispersal in two populations of different size and density in a naturally fragmented landscape. MATERIAL AND METHODS Study species and study sites Banksia nivea ssp. uliginosa has a patchy distribution in two main areas, east of Busselton and on the Scott River Plain, approximately 220 and 260 km, respectively, south of Perth, Western Australia (George, 1996). It grows as a mounded shrub up to 1.5 m tall and 1.5 m across. Flowers are yellowish brown with maroon-coloured style and green pollen presenter, have a mousey odour, and are well hidden within the plant (Brown, Thomson-Dans & Marchant, 1998; Cavanagh & Pieroni, 2006). Flowering occurs between July and September. Each inflorescence has 86.7 ± 2.9 flowers on average (R. Thavornkanlapachai, unpubl. data). Although some follicles open to release seeds periodically, many remain on the plant for a long period time. If plants die after a disturbance, such as fire, mass recruitment occurs (Dixon, Dixon & Krauss, 2003). This study was undertaken in two populations of differing size and habitat context. The location of sampled plants in each population is shown in Figure 2. One population is located on the northern side of Tutunup Road, south-east of Busselton. Adjacent to this area, there is a larger number of plants on the southern side of the road. The size of the TutunupA population is 45 plants, and the majority occur in dense clumps scattered along the road verge with an average number of 21.7 ± 11.4 plants in a radius of 10 m. The overall density is 86 plants/ha, although this overall value is not representative of the plant distribution due to the presence of two more isolated plants, and the density in the main distribution is 144 plants/ha (Fig. 2A). The other population is located in a moderately sized nature reserve at Gale Road. This population is large and dispersed with an average number of 7.5 ± 3.1 plants in a radius of 10 m, and is isolated from other populations (Fig. 2B). The estimated population size is > 500 plants with a density of 134.2 plants/ha that is representative of the plant distribution. Figure 2. View largeDownload slide Location of plants sampled for paternity analysis in two populations of Banksia nivea ssp. uliginosa: (A) Gale population and (B) TutunupA population. ●, sampled plant;⋆, plant sampled for seed; number identifies mother plants; lines represent roads. Figure 2. View largeDownload slide Location of plants sampled for paternity analysis in two populations of Banksia nivea ssp. uliginosa: (A) Gale population and (B) TutunupA population. ●, sampled plant;⋆, plant sampled for seed; number identifies mother plants; lines represent roads. Plant material and genotyping Leaf samples were collected from all 45 plants in the TutunupA population, and 101 plants sampled from a 106 × 71-m patch in the centre of the reserve on Gale Road. Twenty infructescences produced in 2008 from each of ten randomly selected mature plants were collected in both populations. For each plant, 25 seeds were extracted, sterilized in 1% sodium hypochlorite solution for 10 min, and rinsed in sterile water four times. The seeds were placed on 7% Amy Media bacteriological agar in 9-cm Petri dishes, wrapped with Parafilm, and incubated at 15 °C with a 12-h dark/light cycle. Seedlings were grown for at least 5 weeks prior to DNA extraction. Both leaf and seedling samples were freeze dried for 2 days. DNA from leaf samples was extracted using the CTAB-PVP method (Byrne, Parrish & Moran, 1998) with sodium sulphite added to the extraction buffer (Byrne, Macdonald & Francki, 2001). DNA from seedlings was extracted using a small-scale version of the Doyle & Doyle (1990) method. DNA samples were normalized to 20 ng/µL and genotyped for seven microsatellite loci, DnA105, DnA129, DnA117, DnB008, DnB114, DnC010 and DnD007, following Millar & Byrne (2008). Modifications were made as follows: addition of another microsatellite locus, DnC010; forward primer sequence 5′-TGTTGCATGTTGGGATAAGG-3′ and reverse primer sequence 5′-TGCATCCAGCCTTGTCATAG-3′; and MgCl2 concentration was adjusted to 1.0 mM (DnA105, DnA117, DnC010, DnD007), 1.5 mM (DnB008, DnB114) or 2.0 mM (DnA129). Data analysis CERVUS 3.0 was used to obtain maximum likelihood-based paternity assignment and measure genetic diversity in the populations [number of alleles per locus, observed and expected heterozygosity, and polymorphic information content (PIC)] (Marshall et al., 1998). Paternity assignment used the following parameters: 10000 simulated offspring; number and proportion in blankets of candidate fathers of 45 (1.0) and 101 (0.1) for TutunupA and Gale, respectively; proportion of loci typed 0.98 and 0.99; inbreeding rates of 0.035 and 0.058; proportion of mistypes of 0.01; minimum number of loci of 4; and confidence level for assignment of 80%. Differences in measures of genetic diversity between populations, as well as between parents and progenies, were statistically tested using Friedman’s rank sum and Wilcoxon’s signed-rank tests implemented in the R version 3.4.3 statistical package (R Core Team, 2017). NEWPAT was also used in conjunction with CERVUS 3.0 as it can identify multiple-matching male parents in addition to having specific algorithms for dealing with mismatches potentially caused by null alleles (Amos, 2000). In cases where two possible fathers with the same likelihood were identified, the closest father by distance to the mother plant was selected as the most likely father. Cryptic pollen flow, or the probability that a seed was assigned to a pollen donor within a population when the true parent is located outside of the population, was calculated as 1 − (P2n) where P2 is the total exclusion probability for the second parent (Dow & Ashley, 1998). Pollen dispersal distances (between mother plant and pollen donor) were determined based on paternity assignment and the number of pollen donors for each 10-m distance class was plotted. Mating system parameters were estimated using MLTR v3.0 (Ritland, 2002) based on the expectation maximization (EM) method for determining maternal and pollen allele frequencies and the Newton Raphson method for joint maximum likelihood estimation of the multi-locus outcrossing rate (tm) and single-locus outcrossing estimate averages over all loci (ts). Estimates of bi-parental inbreeding (tmts), correlation of paternity (rp), correlation of t among loci (rs), correlation of selfing among loci within progeny arrays (rsloc) and correlation of selfing among families (rsfam) were also obtained. Standard errors were computed based on 500 bootstraps. The correlation of paternity was translated to provide estimates of the effective number of pollen donors using the equation Nep = 1/rp (Ritland, 2002). The standard error was used to determine whether mating parameters were significantly lower than one or greater than zero. Pollen and ovule frequencies were not pooled. The fixation index of seedlings (FO) was estimated, FO=1−(HO/HE), where HO a direct estimate of observed heterozygosity from the seed calculated in GENALEX v6.5 (Peakall & Smouse, 2006) and HE is expected heterozygosity, manually calculated using HE=1−∑​p2, where p is the frequency of each allele of a given locus in the pollen pool (Nei, 1977). Mean co-ancestry of adult plants was estimated as Loiselle’s kinship coefficient (Loiselle et al., 1995) using SPAGeDi (Hardy & Vekemans, 2002). Mean co-ancestry within families was manually calculated as Θ = 0.125(1+Fm)[4s + (tm2 +stmrs )(1+rpm)], where Fm is the fixation index of mother trees, s is the selfing rate, estimated as s =1−tm, rs is correlation of selfing among families, and rpm is multilocus correlation of paternity (Sebbenn, 2006). We inferred pollen-mediated gene dispersal patterns by modelling parentage probabilities of offspring using the neighbourhood model as implemented in NM+ v1.1 (Chybicki & Burczyk, 2010). NM+ uses a maximum likelihood approach to estimate the proportion of offspring resulting from selfing (s), pollen immigration from outside the sampling area (mp) and shape parameter of pollen (bp). Initial parameters were set as follows: neighbourhood of both populations to infinite (to sample all individuals in the sampling area), error rates to 0.01, pollen immigration rate (mp) and selfing rate (s) following the values in Table 3, and initial value for bp to one. Average pollen dispersal distance (dp) was calculated from the average distance of assigned father to mother plant. The level of pollen immigration, to, was estimated as to= 1 − (ti + s), where s is the proportion of selfed progeny and ti is the proportion of outcross events originating from pollen sources within the population. The heterogeneity of the pollen pools sampled from seedlings of each maternal tree was analysed using the TWOGENER method (Smouse et al., 2001) implemented in GENALEX v6.5 (Peakall & Smouse, 2006). This analysis is based on analysis of molecular variance (AMOVA) (Excoffier, Smouse & Quattro, 1992) and identifies the effects of fathers on pollen pool structure (Smouse et al., 2001). RESULTS Microsatellite variation The microsatellite loci were highly variable in the sampled populations of B. nivea ssp. uliginosa with a total of 55 and 53 alleles detected across the seven loci in the TutunupA and Gale populations, respectively (Table 1). The average numbers of alleles per locus (A) were similar in the TutunupA and Gale populations (7.9 ± 4.2 and 7.6 ± 5.0, respectively). Observed heterozygosity (HO) in the TutunupA population was high (0.581 ± 0.051) but not significantly different from that in the Gale population (0.477 ± 0.09; W = 33, P = 0.318). There were no significant differences in observed or expected heterozygosity for parent and progenies in either population (TutunupA: χ2 = 0.110, d.f. = 6, P > 0.05; Gale: χ2 = 0.094, d.f. = 6, P > 0.05). Estimates of null allele frequencies were lower than 0.1, except for DnA129 in both populations and DnA117 in the TutunupA population. As mismatches generated by null alleles are treated as typing errors in paternity assignment, null alleles with low frequencies do not have a significant effect. The polymorphic information content of the loci in both populations was moderate and not significantly different (W = 35, P = 0.209). Table 1. Allelic diversity values for microsatellite markers in two populations of Banksia nivea ssp. uliginosa Population  Locus  A  HO  HE  PIC  Excl  Null  TutunupA  DnA105  4  0.502  0.537  0.467  0.725  0.039  N = 245  DnA117  5  0.465  0.487  0.433  0.745  0.023    DnA129  9  0.498  0.763  0.730  0.441  0.214    DnB008  14  0.755  0.763  0.745  0.405  0.005    DnB114  13  0.769  0.807  0.786  0.358  0.019    DnC010  6  0.457  0.438  0.386  0.779  0.030    DnD007  4  0.624  0.718  0.663  0.541  0.073    Mean  7.9 (4.2)  0.581 (0.051)  0.645 (0.058)  0.601 (0.063)    0.058 (0.027)    Total  62.9        0.985    Gale  DnA105  5  0.616  0.626  0.575  0.619  0.003  N = 294  DnA117  4  0.158  0.242  0.229  0.874  0.200    DnA129  10  0.224  0.370  0.349  0.793  0.252    DnB008  5  0.399  0.432  0.406  0.751  0.046    DnB114  18  0.830  0.878  0.864  0.245  0.027    DnC010  6  0.678  0.680  0.622  0.582  0.004    DnD007  5  0.432  0.403  0.371  0.783  0.044    Mean  7.6 (5.0)  0.477 (0.092)  0.519 (0.083)  0.488 (0.081)    0.082 (0.038)    Total  186.3        0.964    Population  Locus  A  HO  HE  PIC  Excl  Null  TutunupA  DnA105  4  0.502  0.537  0.467  0.725  0.039  N = 245  DnA117  5  0.465  0.487  0.433  0.745  0.023    DnA129  9  0.498  0.763  0.730  0.441  0.214    DnB008  14  0.755  0.763  0.745  0.405  0.005    DnB114  13  0.769  0.807  0.786  0.358  0.019    DnC010  6  0.457  0.438  0.386  0.779  0.030    DnD007  4  0.624  0.718  0.663  0.541  0.073    Mean  7.9 (4.2)  0.581 (0.051)  0.645 (0.058)  0.601 (0.063)    0.058 (0.027)    Total  62.9        0.985    Gale  DnA105  5  0.616  0.626  0.575  0.619  0.003  N = 294  DnA117  4  0.158  0.242  0.229  0.874  0.200    DnA129  10  0.224  0.370  0.349  0.793  0.252    DnB008  5  0.399  0.432  0.406  0.751  0.046    DnB114  18  0.830  0.878  0.864  0.245  0.027    DnC010  6  0.678  0.680  0.622  0.582  0.004    DnD007  5  0.432  0.403  0.371  0.783  0.044    Mean  7.6 (5.0)  0.477 (0.092)  0.519 (0.083)  0.488 (0.081)    0.082 (0.038)    Total  186.3        0.964    A, number of alleles per locus; HO, observed heterozygosity; HE, expected heterozygosity; PIC, polymorphic information content; Excl, exclusion probability for the second parent; Null, estimated frequency of null alleles; N, total number of plants sampled, including adult plants and seedling progeny. Standard deviation is given after allele mean numbers. Standard error is in parentheses. View Large Table 1. Allelic diversity values for microsatellite markers in two populations of Banksia nivea ssp. uliginosa Population  Locus  A  HO  HE  PIC  Excl  Null  TutunupA  DnA105  4  0.502  0.537  0.467  0.725  0.039  N = 245  DnA117  5  0.465  0.487  0.433  0.745  0.023    DnA129  9  0.498  0.763  0.730  0.441  0.214    DnB008  14  0.755  0.763  0.745  0.405  0.005    DnB114  13  0.769  0.807  0.786  0.358  0.019    DnC010  6  0.457  0.438  0.386  0.779  0.030    DnD007  4  0.624  0.718  0.663  0.541  0.073    Mean  7.9 (4.2)  0.581 (0.051)  0.645 (0.058)  0.601 (0.063)    0.058 (0.027)    Total  62.9        0.985    Gale  DnA105  5  0.616  0.626  0.575  0.619  0.003  N = 294  DnA117  4  0.158  0.242  0.229  0.874  0.200    DnA129  10  0.224  0.370  0.349  0.793  0.252    DnB008  5  0.399  0.432  0.406  0.751  0.046    DnB114  18  0.830  0.878  0.864  0.245  0.027    DnC010  6  0.678  0.680  0.622  0.582  0.004    DnD007  5  0.432  0.403  0.371  0.783  0.044    Mean  7.6 (5.0)  0.477 (0.092)  0.519 (0.083)  0.488 (0.081)    0.082 (0.038)    Total  186.3        0.964    Population  Locus  A  HO  HE  PIC  Excl  Null  TutunupA  DnA105  4  0.502  0.537  0.467  0.725  0.039  N = 245  DnA117  5  0.465  0.487  0.433  0.745  0.023    DnA129  9  0.498  0.763  0.730  0.441  0.214    DnB008  14  0.755  0.763  0.745  0.405  0.005    DnB114  13  0.769  0.807  0.786  0.358  0.019    DnC010  6  0.457  0.438  0.386  0.779  0.030    DnD007  4  0.624  0.718  0.663  0.541  0.073    Mean  7.9 (4.2)  0.581 (0.051)  0.645 (0.058)  0.601 (0.063)    0.058 (0.027)    Total  62.9        0.985    Gale  DnA105  5  0.616  0.626  0.575  0.619  0.003  N = 294  DnA117  4  0.158  0.242  0.229  0.874  0.200    DnA129  10  0.224  0.370  0.349  0.793  0.252    DnB008  5  0.399  0.432  0.406  0.751  0.046    DnB114  18  0.830  0.878  0.864  0.245  0.027    DnC010  6  0.678  0.680  0.622  0.582  0.004    DnD007  5  0.432  0.403  0.371  0.783  0.044    Mean  7.6 (5.0)  0.477 (0.092)  0.519 (0.083)  0.488 (0.081)    0.082 (0.038)    Total  186.3        0.964    A, number of alleles per locus; HO, observed heterozygosity; HE, expected heterozygosity; PIC, polymorphic information content; Excl, exclusion probability for the second parent; Null, estimated frequency of null alleles; N, total number of plants sampled, including adult plants and seedling progeny. Standard deviation is given after allele mean numbers. Standard error is in parentheses. View Large Outcrossing rate Multi-locus outcrossing in both populations was high (0.942 in Gale and 0.965 in TutunupA) with a small amount of inbreeding from self-pollination and biparental inbreeding (Table 2). The fixation index of mother trees was low in both populations, but was slightly higher in the Gale population (0.013 ± 0.004) than in TutunupA (0.005 ± 0.014). The fixation index of seedlings was also higher in the Gale population (FO = 0.116) than TutunupA population (FO = 0.107). The estimate of biparental mating was nearly twice as high in the TutunupA population (0.113 ± 0.038) than in the Gale population (0.060 ± 0.025). Correlation of selfing among loci indicated 88.5 ± 3.2% of inbreeding in the Gale population was from self-pollination compared to 77.3 ± 7.7% in the TutunupA population. Estimates of the correlation of paternity (rp(m)) indicated the number of pollen donors (Nep) per maternal plant at the Gale population (20.8) was much higher than at TutunupA (5.3). Estimates of mean co-ancestry in adult plants were similar in both populations (TutunupA = −0.008, Gale = 0.001). In seedlings, estimates of mean co-ancestry were higher than that expected for half-sibs, indicating some seedlings within families had the same fathers, which is consistent with estimates of correlated paternity. The TutunupA families were 25% (0.157–0.125/0.125) more closely related than complete half sibs, and this value was 18% for Gale families. Table 2. Estimates of mating system parameters for populations of Banksia nivea ssp. uliginosa Parameter  TutunupA  Gale  Parental F estimate  0.005 (0.014)  0.013 (0.004)  tm  0.965 (0.017)  0.942 (0.018)  1 − tm  0.035  0.058  ts  0.852 (0.049)  0.882 (0.026)  tm − ts  0.113 (0.038)  0.060 (0.025)  rsloc  0.773 (0.077)  0.885 (0.032)  rsfam  0.070 (0.017)  0.092 (0.036)  rp(m)  0.190 (0.036)  0.048 (0.009)  rp(s) − rp(m)  −0.390 (0.036)  −0.248 (0.019)  1/rp(m)  5.26  20.83  Co-ancestry  0.157  0.148  Parameter  TutunupA  Gale  Parental F estimate  0.005 (0.014)  0.013 (0.004)  tm  0.965 (0.017)  0.942 (0.018)  1 − tm  0.035  0.058  ts  0.852 (0.049)  0.882 (0.026)  tm − ts  0.113 (0.038)  0.060 (0.025)  rsloc  0.773 (0.077)  0.885 (0.032)  rsfam  0.070 (0.017)  0.092 (0.036)  rp(m)  0.190 (0.036)  0.048 (0.009)  rp(s) − rp(m)  −0.390 (0.036)  −0.248 (0.019)  1/rp(m)  5.26  20.83  Co-ancestry  0.157  0.148  tm, multilocus population outcrossing rate; ts, single locus population outcrossing rate; rsloc, correlation of selfing among loci; rsfam, correlation of selfing among families (normalized variance of selfing); rp(m), multilocus correlation of paternity (fraction of siblings sharing the same father); rp(s), single locus correlation of paternity; Co-ancestry, Loiselle’s kinship coefficient. Standard error in parentheses estimated based on 500 bootstrap re-samplings. View Large Table 2. Estimates of mating system parameters for populations of Banksia nivea ssp. uliginosa Parameter  TutunupA  Gale  Parental F estimate  0.005 (0.014)  0.013 (0.004)  tm  0.965 (0.017)  0.942 (0.018)  1 − tm  0.035  0.058  ts  0.852 (0.049)  0.882 (0.026)  tm − ts  0.113 (0.038)  0.060 (0.025)  rsloc  0.773 (0.077)  0.885 (0.032)  rsfam  0.070 (0.017)  0.092 (0.036)  rp(m)  0.190 (0.036)  0.048 (0.009)  rp(s) − rp(m)  −0.390 (0.036)  −0.248 (0.019)  1/rp(m)  5.26  20.83  Co-ancestry  0.157  0.148  Parameter  TutunupA  Gale  Parental F estimate  0.005 (0.014)  0.013 (0.004)  tm  0.965 (0.017)  0.942 (0.018)  1 − tm  0.035  0.058  ts  0.852 (0.049)  0.882 (0.026)  tm − ts  0.113 (0.038)  0.060 (0.025)  rsloc  0.773 (0.077)  0.885 (0.032)  rsfam  0.070 (0.017)  0.092 (0.036)  rp(m)  0.190 (0.036)  0.048 (0.009)  rp(s) − rp(m)  −0.390 (0.036)  −0.248 (0.019)  1/rp(m)  5.26  20.83  Co-ancestry  0.157  0.148  tm, multilocus population outcrossing rate; ts, single locus population outcrossing rate; rsloc, correlation of selfing among loci; rsfam, correlation of selfing among families (normalized variance of selfing); rp(m), multilocus correlation of paternity (fraction of siblings sharing the same father); rp(s), single locus correlation of paternity; Co-ancestry, Loiselle’s kinship coefficient. Standard error in parentheses estimated based on 500 bootstrap re-samplings. View Large Paternity assignment and pollen dispersal Paternity assignment identified most likely fathers for 275 progeny (70%) across the two populations. The TutunupA population had a high outcrossing rate between individuals within the population (72.5%), 23% of pollen was assumed to come from outside the population and 4.5% of seedlings were the result of selfing (Table 3). The Gale population had a slightly lower within-patch mating rate of 64.8%, 30% of the seedlings were not assigned paternity from the genotyped plants within the patch and 5.1% of seedlings were assigned to the mother plants. Cryptic pollen flow was relatively high in both populations, 0.493 and 0.680 for the TutunupA and Gale populations, respectively, meaning estimates of pollen immigration may be underestimates by up to 32 and 36% for TutunupA and Gale populations, respectively. Table 3. Selfing and outcrossing rates for seedlings from two populations of Banksia nivea ssp. uliginosa assessed using paternity assignment Population  Mother  Number of progeny  Selfing rate (%)  Outcrossing within sampled patch (%)  Apparent pollen immigration to patch (%)  TutunupA  1  20  0  50  50    5  20  0  100  0    10  20  5  75  20    14  20  0  85  15    23  20  5  85  10    33  20  5  65  30    34  20  10  75  15    36  20  5  65  30    37  20  5  90  5    42  20  10  35  55    Mean  20  4.5  72.5  23  Gale  1  15  0  40  60    5  20  5  80  15    17  20  5  65  30    18  18  11  78  11    30  20  15  50  35    31  20  0  80  20    40  20  0  90  10    71  20  5  65  30    83  20  0  50  50    95  20  10  50  40    Mean  19.3  5.1  64.8  30.1  Population  Mother  Number of progeny  Selfing rate (%)  Outcrossing within sampled patch (%)  Apparent pollen immigration to patch (%)  TutunupA  1  20  0  50  50    5  20  0  100  0    10  20  5  75  20    14  20  0  85  15    23  20  5  85  10    33  20  5  65  30    34  20  10  75  15    36  20  5  65  30    37  20  5  90  5    42  20  10  35  55    Mean  20  4.5  72.5  23  Gale  1  15  0  40  60    5  20  5  80  15    17  20  5  65  30    18  18  11  78  11    30  20  15  50  35    31  20  0  80  20    40  20  0  90  10    71  20  5  65  30    83  20  0  50  50    95  20  10  50  40    Mean  19.3  5.1  64.8  30.1  View Large Table 3. Selfing and outcrossing rates for seedlings from two populations of Banksia nivea ssp. uliginosa assessed using paternity assignment Population  Mother  Number of progeny  Selfing rate (%)  Outcrossing within sampled patch (%)  Apparent pollen immigration to patch (%)  TutunupA  1  20  0  50  50    5  20  0  100  0    10  20  5  75  20    14  20  0  85  15    23  20  5  85  10    33  20  5  65  30    34  20  10  75  15    36  20  5  65  30    37  20  5  90  5    42  20  10  35  55    Mean  20  4.5  72.5  23  Gale  1  15  0  40  60    5  20  5  80  15    17  20  5  65  30    18  18  11  78  11    30  20  15  50  35    31  20  0  80  20    40  20  0  90  10    71  20  5  65  30    83  20  0  50  50    95  20  10  50  40    Mean  19.3  5.1  64.8  30.1  Population  Mother  Number of progeny  Selfing rate (%)  Outcrossing within sampled patch (%)  Apparent pollen immigration to patch (%)  TutunupA  1  20  0  50  50    5  20  0  100  0    10  20  5  75  20    14  20  0  85  15    23  20  5  85  10    33  20  5  65  30    34  20  10  75  15    36  20  5  65  30    37  20  5  90  5    42  20  10  35  55    Mean  20  4.5  72.5  23  Gale  1  15  0  40  60    5  20  5  80  15    17  20  5  65  30    18  18  11  78  11    30  20  15  50  35    31  20  0  80  20    40  20  0  90  10    71  20  5  65  30    83  20  0  50  50    95  20  10  50  40    Mean  19.3  5.1  64.8  30.1  View Large At TutunupA, the observed pollen dispersal distance ranged from 0.4 to 78.1 m with a mean distance per mother plant of 19.7 ± 4.4 m. A majority of mating events (67%) occurred with pollen from plants within 10 m of a mother plant (Fig. 3A). In the Gale population, the observed pollen dispersal ranged from 1.0 to 75.0 m with a mean distance per mother plant of 30.8 ± 2.7 m. The majority of mating events (73%) occurred with pollen from plants located within 40 m of a mother plant (Fig. 3B). The distribution of observed and potential pollen dispersal distance within both populations was similar, except for the closest distance category (0−10 m) where distance to observed pollen parents was significantly higher than potential pollen parents in both populations (TutunupA: χ2 = 106.1, d.f. = 9, P < 0.001; Gale: χ2 = 69.0, d.f. = 9, P < 0.001). Figure 3. View largeDownload slide Comparison of potential (light bars) and observed (dark bars) pollen dispersal distance within the (A) TutunupA and (B) Gale populations in Banksia nivea ssp. uliginosa. Potential dispersal distance is the average distance from each materal plant to all potential fathers, and observed dispersal distance is the average distance to the assigned fathers based on paternity assignment. Pollen disperal distance is measured in 10-m classes to 100 m. Asterisks indicate significant differences between observed and expected pollen dispersal distance, P < 0.001. Figure 3. View largeDownload slide Comparison of potential (light bars) and observed (dark bars) pollen dispersal distance within the (A) TutunupA and (B) Gale populations in Banksia nivea ssp. uliginosa. Potential dispersal distance is the average distance from each materal plant to all potential fathers, and observed dispersal distance is the average distance to the assigned fathers based on paternity assignment. Pollen disperal distance is measured in 10-m classes to 100 m. Asterisks indicate significant differences between observed and expected pollen dispersal distance, P < 0.001. Analysis of pollen dispersal based on the Neighbourhood model gave the following estimates for the TutunupA and Gale populations, respectively: selfing rates of 7.1 ± 2.2 and 4.5 ± 0.0%, shape of dispersal kernel (bp) of 1.3 and 0.9, mean distances of pollen dispersal kernel (dp) of 26.9 and 30.9 m, mean within-neighbourhood pollen dispersal distances of 13.5 and 20.4 m and percentages of pollen immigration (mp) coming from outside the sampling area of 18.6 ± 3.6 and 28.5 ± 0.0%. The TWOGENER AMOVA indicated high levels of global pollen pool differentiation within sampled populations, ΦST (Gale) = 0.224, ΦST (TutunupA) = 0.331, indicating that pollinators were sampling different pollen pools within the populations. DISCUSSION Our investigation has shown that the pattern of pollen dispersal within two populations of Banksia nivea ssp. uliginosa is consistent with expectations based on the size and density of population patches. We found more restricted dispersal within the linear, smaller and more clumped population, consistent with theoretical expectations of the influence of size and spatial arrangements of plants. Although present in a naturally fragmented landscape, B. nivea ssp. uliginosa showed high outcrossing, with a small amount of selfing in the large population, and mating between relatives in the small population. There was evidence of reasonable levels of pollen immigration from the adjacent populations supplementing the within-population mating. Predominant outcrossing with low level of selfing Our analysis shows that B. nivea ssp. uliginosa is predominantly outcrossing, with a small amount of selfing. This is consistent with the outcrossing rate of most studied banksias that ranges from 0.670 to 0.950 (Scott, 1980; Coates et al., 2007; Collins, Walsh & Grey, 2008; Llorens et al., 2012), although the rate observed in B. nivea ssp. uliginosa is at the high end of this range. In this study, a higher fixation index in seedlings than in adult plants indicates that selection occurs during establishment of adult plants. Removal of inbred seed through selection protects populations against potential genetic diversity loss, and our study shows that both studied populations of B. nivea ssp. uliginosa have high levels of genetic diversity despite occurring in a naturally fragmented landscape where genetic diversity is expected to be reduced. The mean co-ancestry values for the adult plants in both populations were not different from zero, indicating negligible relatedness among the established adult plants, and little effect of population size or density. This low relatedness is reflected in the level of genetic diversity of these B. nivea ssp. uliginosa populations (A = 7.7, HO = 0.529), which is comparable to other banksia species with more widespread distributions, such as B. brownii (A = 3.4, HO = 0.472; Coates, McArthur & Byrne, 2015), B. attenuata (A = 5.8, HE = 0.581−0.778; He et al., 2008) and B. sphaerocarpa ssp. sphaerocarpa (A = 11.5, HO = 0.670; Nistelberger et al., 2015). Nonetheless, the mean co-ancestry estimates of seedlings in both populations were higher than that expected from complete half sibs, consistent with estimates of some correlated paternity. A theoretical consequence of small population size is that there are fewer mates available leading to higher selfing and higher correlated paternity in such populations (Young et al., 1996; Honnay & Jacquemyn, 2007; Aguilar et al., 2008; Eckert et al., 2010). Indeed, population size has been shown to have an effect on outcrossing and on correlated paternity in species with mixed mating systems, such as the shrubs Calothamnus quadrifidus (Yates et al., 2007a) and B. sphaerocarpa ssp. sphaerocarpa (Llorens et al., 2012). In B. nivea ssp. uliginosa, the inbreeding level was low in both populations, indicating no effect of different population size, probably a consequence of the high outcrossing rate and low relatedness of plants. However, there were some differences between the populations, as inbreeding at TutunupA was mainly due to the low effective number of fathers, higher mating between relatives and higher number of progeny sharing the same father, while inbreeding in Gale was mostly caused by selfing. Influence of plant distribution on pollen dispersal Although the number of plants in the population is important in terms of mate availability, plant density also affects pollinator movement, with greater pollinator movement between plants over greater distances in less dense populations (Levin & Kerster, 1974). The TutunupA population showed most mating events occurred within 10 m of a mother plant while mating in the Gale population occurred up to 40−50 m from a mother plant. Subsequently, pollen dispersal kernel at TutunupA had an exponential shape while was a fat-tailed dispersal kernel at Gale. This difference in dispersal kernel is consistent with a shorter mean within-neighbourhood pollen distance in the TutunupA population in comparison with the Gale population. This difference is also likely to be related to the different spatial arrangement of plants within the populations, as plant density is known to affect pollinator movement and to affect different pollinators differently (Levin & Kerster, 1974; Ashley, 2010). For example, in a study of mating in Grevillea macleayana, bees tended to move within plants while birds showed more between-plant flights than bees (Whelan, Ayre & Beynon, 2009), and analysis of pollen dispersal in the bird- and mammal-pollinated B. sphaerocarpa ssp. sphaerocarpa showed patterns that were highly correlated with plant density (Llorens et al., 2012). Banksia nivea ssp. uliginosa is primarily pollinated by small mammals with some contribution from birds, but there is little known about mammal pollinator movement. The honey possum, Tarsipes rostratus, is the most likely main pollinator, as it is a common visitor to banksia flowers, and although these animals have been observed not move more than 30 m even over several months (Garavanta, Wooller & Richardson, 2000; Bradshaw et al., 2007), the present study suggests these mammals may move further in populations with lower plant density. Bird pollinators, such as honeyeaters, are likely to facilitate long-distance pollen dispersal of B. nivea ssp. uliginosa as they have been observed carrying pollen of banksia species (Wooller, 2001, 2002), and have been attributed to contribute to pollen dispersal of C. quadrifidus between fragments over 5 km (Byrne et al., 2007). Different spatial arrangements of plants in this study may also influence bird pollinator movement, similar to the increased movement of bird pollinators between C. quadrifidus plants in denser populations (Yates et al., 2007a). Both populations showed a high level of pollen pool differentiation, indicating plants were accessing different pollen pools within the population. This may be due to different plants flowering at different times or flowers on plants becoming receptive at different times during the season. Flowering in B. nivea ssp. uliginosa occurs over several months within populations, over at least 1 month in large plants and over several days in an inflorescence. This is similar to another banksia species in which anthesis varied between 7 and 33 days (Saffer, 2004; Wooller & Wooller, 2003). Despite having a clumped distribution of plants, the TutunupA population had a more highly differentiated pollen pool and received pollen from fewer mates than the larger and less dense population at Gale, suggesting that population size rather than density has a greater effect on maintaining pollen pool homogeneity in this species. This is consistent with a study of B. sphaerocarpa ssp. sphaerocarpa where larger population size was correlated with lower pollen pool differentiation and greater number of mates (Llorens et al., 2012). Interestingly, this is in contrast to a similar study on the bird-pollinated C. quadrifidus, which showed lower pollen pool heterogeneity in the denser population, which is probably explained by concurrent flowering among plants in this species (Byrne et al., 2007). Estimation of pollen immigration into the sampled patches was variable across mother plants in the studied populations. Estimates of 23–30% of progeny (and up to 55 and 66% [for TutunupA and Gale, respectively] based on estimates of cryptic pollen flow) were sired from pollen derived from outside the sampled patches, consistent with expectations of pollination by mobile pollinators such as mammals and birds. Although this pollen immigration was from nearby plants within the broader populations, levels of pollen immigration have been shown to be extensive in fragmented landscapes in south-western Australia, with values ranging from 3 to 64% over distances of up to 5 km (Byrne et al., 2007, 2008; Sampson & Byrne, 2008; Krauss et al., 2009; Llorens et al., 2012; Millar et al., 2012; Sampson et al., 2014), and influenced by a range of variables, including isolation, size, density and pollinator type. The accumulation of these data on pollen immigration demonstrates that pollen dispersal is not limited in fragmented landscapes and provides a means of achieving genetic connectivity among populations. CONCLUSIONS Despite predominant pollination by a non-flying mammal that is likely to have a limited foraging range, we found high outcrossing and minimal selfing in the populations of B. nivea ssp. uliginosa studied here, and up to 30% of progeny were derived from mating with fathers outside the sampled patches. Pollen dispersal was affected by adult plant distribution and density as expected from predominantly mammal pollination, although high levels of pollen immigration are consistent with some contribution from bird pollination. Animal visitation is essential for pollination, so maintaining a viable population of pollinators is important in ensuring pollination and gene flow in fragmented landscapes. An effective pollinator community is also dependent on maintenance of the species pool it forages on, so maintaining habitat context for rare species is an important component of management programmes. ACKNOWLEDGMENTS We would like to thank three anonymous reviewers for their comments on the manuscript. We thank Bronwyn Macdonald for assistance in the laboratory, Margaret Langley and Belinda Newman for assistance with sample collection, Melissa Millar for help with the ArcView program and Jane Sampson for advice on analysis. The project was supported by funding from the South West Catchments Council. REFERENCES Aguilar R, Ashworth L, Galetto L, Aizen MA. 2006. Plant reproductive susceptibility to habitat fragmentation: review and synthesis through a meta-analysis. Ecology Letters  9: 968– 980. Google Scholar CrossRef Search ADS   Aguilar R, Quesada M, Ashworth L, Herrerias-Diego Y, Lobo J. 2008. Genetic consequences of habitat fragmentation in plant populations: susceptible signals in plant traits and methodological approaches. Molecular Ecology  17: 5177– 5188. Google Scholar CrossRef Search ADS   Amos B. 2000. NEWPAT, a general paternity program, 5thedn . Available at:http://www.zoo.cam.ac.uk/zoostaff/amos/ newpat.htm. (accessed 19 May 2009). Ashley MV. 2010. Plant parentage, pollination, and dispersal: How DNA microsatellites have altered the landscape. Critical Reviews in Plant Sciences  29: 148– 161. Google Scholar CrossRef Search ADS   Bennett A. 2003. Habitat fragmentation. In: Attiwill P, Wilson B, eds. Ecology: an Australian perspective . Melbourne: Oxford University Press, 440– 456. Bradshaw SD, Phillips RD, Tomlinson S, Holley RJ, Jennings S, Bradshaw FJ. 2007. Ecology of the Honey possum, Tarsipes rostratus, in Scott National Park, Western Australia. Australian Mammalogy  29: 25– 38. Google Scholar CrossRef Search ADS   Brown A, Thomson-Dans C, Marchant N, eds. 1998. Western Australia’s threatened flora . Perth: Department of Conservation and Land Management. Burbidge AH, James SH. 1991. Postzygotic seed abortion in the genetic system of Stylidium (Angiospermae: Stylidiaceae). Journal of Heredity  82: 319– 328. Google Scholar CrossRef Search ADS   Byrne M, Elliott CP, Yates C, Coates DJ. 2007. Extensive pollen dispersal in a bird-pollinated shrub, Calothamnus quadrifidus, in a fragmented landscape. Molecular Ecology  16: 1303– 1314. Google Scholar CrossRef Search ADS   Byrne M, Elliott CP, Yates CJ, Coates DJ. 2008. Maintenance of high pollen dispersal in Eucalyptus wandoo, a dominant tree of the fragmented agricultural region in Western Australia. Conservation Genetics  9: 97– 105. Google Scholar CrossRef Search ADS   Byrne M, Macdonald B, Francki M. 2001. Incorporation of sodium sulfite into extraction protocol minimizes degradation of Acacia DNA. Biotechniques  30: 742– 744, 748. Byrne M, Parrish TL, Moran GF. 1998. Nuclear RFLP diversity in Eucalyptus nitens. Heredity  81: 225– 232. Google Scholar CrossRef Search ADS   Cavanagh T, Pieroni M. 2006. The dryandras . Melbourne: Australian Plants Society (SGAP Victoria) Inc. and Wildflower Society of Western Australia Inc. Chybicki IJ, Burczyk J. 2010. NM+: software implementing parentage-based models for estimating gene dispersal and mating patterns in plants. Molecular Ecology Resources  10: 1071– 1075. Google Scholar CrossRef Search ADS   Coates DJ, Byrne M. 2005. Genetic variation in plant populations: assessing cause and pattern. In: Henry RJ, ed. Plant diversity and evolution: genotypic and phenotypic variation in higher plants . Boston: CABI Publishing, 139– 164. Google Scholar CrossRef Search ADS   Coates DJ, McArthur SL, Byrne M. 2015. Significant genetic diversity loss following pathogen driven population extinction in the rare endemic Banksia brownii (Proteaceae). Biological Conservation  192: 353– 360. Google Scholar CrossRef Search ADS   Coates DJ, Sampson JF, Yates CJ. 2007. Plant mating systems and assessing population persistence in fragmented landscapes. Australian Journal of Botany  55: 239– 249. Google Scholar CrossRef Search ADS   Collins BG, Walsh M, Grey J. 2008. Floral development and breeding systems of Dryandra sessilis and Grevillea wilsonii (Proteaceae). Australian Journal of Botany  56: 119– 130. Google Scholar CrossRef Search ADS   Dixon K, Dixon B, Krauss S. 2003. Kings Park and Botanic Gardens (BGPA) Science Directorate research proposal for the rescue of four rare and endangered species at BHP Beenup minesite . Perth: BGPA. Dow BD, Ashley MV. 1998. High levels of gene flow in bur oak revealed by paternity analysis using microsatellites. Journal of Heredity  89: 62– 70. Google Scholar CrossRef Search ADS   Dudash MR, Fenster CB. 2000. Inbreeding and outbreeding depression in fragmented populations. In: Young AG, Clarke GM, eds. Conservation biology 4: genetics, demography and viability of fragmented populations . Cambridge: Cambridge University Press, 35– 45. Google Scholar CrossRef Search ADS   Eckert CG, Kalisz S, Geber MA, Sargent R, Elle E, Cheptou PO, Goodwillie C, Johnston MO, Kelly JK, Moeller DA, Porcher E, Ree RH, Vallejo-Marín M, Winn AA. 2010. Plant mating systems in a changing world. Trends in Ecology & Evolution  25: 35– 43. Google Scholar CrossRef Search ADS   Ellstrand NC. 2014. Is gene flow the most important evolutionary force in plants? American Journal of Botany  101: 737– 753. Google Scholar CrossRef Search ADS   English V. 1999. Interim recovery plan no. 44: Shrubland association on Southern Swan Coastal Plain Ironstone (Busselton Area) (Southern Ironstone Association) Interim recovery plan 1999–2002 . Wanneroo: Department of Conservation and Land Management Western Australian Threatened Species and Communities Unit. Excoffier L, Smouse PE, Quattro JM. 1992. Analysis of molecular variance inferred from metric distances among DNA haplotypes: application to human mitochondrial DNA restriction data. Genetics  131: 479– 491. Frankham R, Ballou JD, Briscoe DA. 2002. Introduction to conservation genetics . Cambridge: Cambridge University Press. Google Scholar CrossRef Search ADS   Garavanta CAM, Wooller RD, Richardson KC. 2000. Movement patterns of honey possums, Tarsipes rostratus, in the Fitzgerald River National Park, Western Australia. Wildlife Research  27: 179– 183. Google Scholar CrossRef Search ADS   George AS. 1996. New taxa and a new infrageneric classification in Dryandra R.Br. (Proteaceae: Grevilleoideae). Nuytsia  10: 313– 408. Haddad NM, Brudvig LA, Clobert J, Davies KF, Gonzalez A, Holt RD, Lovejoy TE, Sexton JO, Austin MP, Collins CD, Cook WM, Damschen EI, Ewers RM, Foster BL, Jenkins CN, King AJ, Laurance WF, Levey DJ, Margules CR, Melbourne BA, Nicholls AO, Orrock JL, Song DX, Townshend JR. 2015. Habitat fragmentation and its lasting impact on Earth’s ecosystems. Science Advances  1: e1500052. Google Scholar CrossRef Search ADS   Hardy OJ, Vekemans X. 2002. SPAGeDi: a versatile computer program to analyse spatial genetic structure at the individual or population levels. Molecular Ecology Notes  2: 618– 620. Google Scholar CrossRef Search ADS   He T, Lamont BB, Krauss SL, Enright NJ, Millar BP. 2008. Covariation between intraspecific genetic diversity and species diversity within a plant functional group. Journal of Ecology  96: 956– 961. Google Scholar CrossRef Search ADS   Honnay O, Jacquemyn H. 2007. Susceptibility of common and rare plant species to the genetic consequences of habitat fragmentation. Conservation Biology  21: 823– 831. Google Scholar CrossRef Search ADS   Krauss SL, He T, Barrett LG, Lamont BB, Enright NJ, Miller BP, Hanley ME. 2009. Contrasting impacts of pollen and seed dispersal on spatial genetic structure in the bird-pollinated Banksia hookeriana. Heredity  102: 274– 285. Google Scholar CrossRef Search ADS   Levin DA, Kerster HW. 1974. Gene flow in seed plants. In: Dobzhansky T, Hecht MK, Steere WC, eds. Evolutionary biology, Vol. 7 . Boston: Springer US, 139– 220. Google Scholar CrossRef Search ADS   Llorens TM, Byrne M, Yates CJ, Nistelberger HM, Coates DJ. 2012. Evaluating the influence of different aspects of habitat fragmentation on mating patterns and pollen dispersal in the bird-pollinated Banksia sphaerocarpa var. caesia. Molecular Ecology  21: 314– 328. Google Scholar CrossRef Search ADS   Loiselle BA, Sork VL, Nason J, Graham C. 1995. Spatial genetic structure of a tropical understory shrub, Psychotria officinalis (Rubiaceae). American Journal of Botany  82: 1420– 1425. Google Scholar CrossRef Search ADS   Lowe A, Harris S, Ashton P. 2004. Ecological genetics: design, analysis, and application . Carlton: Blackwell Science Ltd. Marshall TC, Slate J, Kruuk LE, Pemberton JM. 1998. Statistical confidence for likelihood-based paternity inference in natural populations. Molecular Ecology  7: 639– 655. Google Scholar CrossRef Search ADS   Mast AR, Thiele K. 2007. The transfer of Dryandra R.Br. to Banksia L.f. (Proteaceae). Australian Systematic Botany  20: 63– 71. Google Scholar CrossRef Search ADS   Mendonça AH, Russo C, Melo ACG, Durigan G. 2015. Edge effects in savanna fragments: a case study in the cerrado. Plant Ecology & Diversity  8: 493– 503. Google Scholar CrossRef Search ADS   Millar MA, Byrne M. 2008. Characterization of polymorphic microsatellite DNA markers in Banksia nivea, formerly Dryandra nivea. Molecular Ecology Resources  8: 1393– 1394. Google Scholar CrossRef Search ADS   Millar MA, Byrne M, Nuberg IK, Sedgley M. 2012. High levels of genetic contamination in remnant populations of Acacia saligna from a genetically divergent planted stand. Restoration Ecology  20: 260– 267. Google Scholar CrossRef Search ADS   Nei M. 1977. F-statistics and analysis of gene diversity in subdivided populations. Annals of Human Genetics  41: 225– 233. Google Scholar CrossRef Search ADS   Nelson‐Tunley M, Morgan‐Richards M, Trewick SA. 2016. Genetic diversity and gene flow in a rare New Zealand skink despite fragmented habitat in a volcanic landscape. Biological Journal of the Linnean Society  119: 37– 51. Google Scholar CrossRef Search ADS   Nistelberger HM, Coates DJ, Llorens TM, Yates CJ, Byrne M. 2015. A cryptic genetic boundary in remnant populations of a long-lived, bird-pollinated shrub Banksia sphaerocarpa var. caesia (Proteaceae). Biological Journal of the Linnean Society  115: 241– 255. Google Scholar CrossRef Search ADS   O’Connell LM, Mosseler A, Rajora OP. 2006. Impacts of forest fragmentation on the reproductive success of white spruce (Picea glauca). Canadian Journal of Botany  84: 956– 965. Google Scholar CrossRef Search ADS   Ouborg NJ, Vergeer P, Mix C. 2006. The rough edges of the conservation genetics paradigm for plants. Journal of Ecology  94: 1233– 1248. Google Scholar CrossRef Search ADS   Pannell JR, Fields PD. 2014. Evolution in subdivided plant populations: concepts, recent advances and future directions. New Phytologist  201: 417– 432. Google Scholar CrossRef Search ADS   Peakall R, Smouse PE. 2006. GENALEX 6: genetic analysis in Excel. Population genetic software for teaching and research. Molecular Ecology Notes  6: 288– 295. Google Scholar CrossRef Search ADS   R Core Team. 2017. R: A language and environment for statistical computing . Vienna: R Foundation for Statistical Computing. Available at http://www.R-project.org/. (accessed 18 December 2017). Reed DH. 2005. Relationship between population size and fitness. Conservation Biology  19: 563– 568. Google Scholar CrossRef Search ADS   Richards CM. 2000. Genetic and demographic influences on population persistence: gene flow and genetic rescue in Silene alba. In: Young AG, Clarke GM, eds. Conservation biology 4: genetics, demography and viability of fragmented populations . Cambridge: Cambridge University Press, 271– 292. Google Scholar CrossRef Search ADS   Ritland K. 2002. Extensions of models for the estimation of mating systems using n independent loci. Heredity  88: 221– 228. Google Scholar CrossRef Search ADS   Saffer VM. 2004. Are diel patterns of nectar production and anthesis associated with other floral traits in plants visited by potential bird and mammal pollinators? Australian Journal of Botany  52: 87– 92. Google Scholar CrossRef Search ADS   Sampson JF, Byrne M. 2008. Outcrossing between an agroforestry plantation and remnant native populations of Eucalyptus loxophleba. Molecular Ecology  17: 2769– 2781. Google Scholar CrossRef Search ADS   Sampson JF, Byrne M, Yates CJ, Gibson N, Thavornkanlapachai R, Stankowski S, MacDonald B, Bennett I. 2014. Contemporary pollen-mediated gene immigration reflects the historical isolation of a rare, animal-pollinated shrub in a fragmented landscape. Heredity  112: 172– 181. Google Scholar CrossRef Search ADS   Sampson JF, Hankinson M, McArthur S, Tapper S, Langley M, Gibson N, Yates C, Byrne M. 2015. Long-term ‘islands’ in the landscape: low gene flow, effective population size and genetic divergence in the shrub Hakea oldfieldii (Proteaceae). Botanical Journal of the Linnean Society  179: 319– 334. Google Scholar CrossRef Search ADS   Scott JK. 1980. Estimation of the outcrossing rate for Banksia attenuata R.Br. and Banksia menziesii R.Br. (Proteaceae). Australian Journal of Botany  28: 53– 59. Google Scholar CrossRef Search ADS   Sebbenn AM. 2006. Sistema de reprodução em espécies arbóreas tropicais e suas implicações para a seleção de árvores matrizes para reflorestamentos ambientais. In: Higa AR, Silva L, eds. Pomares de sementes de espécies florestais nativas . Curitiba: FUPEF, 193– 198. Segelbacher G, Cushman SA, Epperson BK, Fortin M-J, Francois O, Hardy OJ, Holderegger R, Taberlet P, Waits LP, Manel S. 2010. Applications of landscape genetics in conservation biology: concepts and challenges. Conservation Genetics  11: 375– 385. Google Scholar CrossRef Search ADS   Smouse PE, Dyer RJ, Westfall RD, Sork VL. 2001. Two-generation analysis of pollen flow across a landscape. I. Male gamete heterogeneity among females. Evolution  55: 260– 271. Google Scholar CrossRef Search ADS   Whelan RJ, Ayre DJ, Beynon FM. 2009. The birds and the bees: pollinator behaviour and variation in the mating system of the rare shrub Grevillea macleayana. Annals of Botany  103: 1395– 1401. Google Scholar CrossRef Search ADS   Wooller RD, Wooller SJ. 2003. The role of non-flying animals in the pollination of Banksia nutans. Australian Journal of Botany  51: 503– 507. Google Scholar CrossRef Search ADS   Wooller SJ, Wooller RD. 2001. Seed set in two sympatric banksias, Banksia attenuata and B. baxteri. Australian Journal of Botany  49: 597– 602. Google Scholar CrossRef Search ADS   Wooller SJ, Wooller RD. 2002. Mixed mating in Banksia media. Australian Journal of Botany  50: 627– 631. Google Scholar CrossRef Search ADS   Yates C, Coates DJ, Elliott CP, Byrne M. 2007a. Composition of the pollinator community, pollination and the mating system for a shrub in fragments of species rich kwongan in south-west Western Australia. Biodiversity and Conservation  16: 1379– 1395. Google Scholar CrossRef Search ADS   Yates CJ, Elliott CP, Byrne M, Coates DJ, Fairman R. 2007b. Seed production, germinability and seedling growth for a bird-pollinated shrub in fragments of kwongan in south-west Australia. Biological Conservation  136: 306– 314. Google Scholar CrossRef Search ADS   Young A, Boyle T, Brown T. 1996. The population genetic consequences of habitat fragmentation for plants. Trends in Ecology & Evolution  11: 413– 418. Google Scholar CrossRef Search ADS   © 2018 The Linnean Society of London, Biological 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)

Journal

Biological Journal of the Linnean Society – Oxford University Press

Published: May 10, 2018

There are no references for this article.