Abstract Morphological traits can determine the ecological niches and performance of ectotherms and structure their distributions along environmental gradients. The thermal melanism hypothesis and Bergmann’s rule describe patterns of body colour luminance and body size along environmental gradients shaped by thermal influences on morphology. However, these patterns have rarely been investigated at the interspecific level for subtropical and tropical mountain environments. In this study, we sampled butterfly assemblages along elevations across three subtropical and tropical locations in China and examined how environmental factors affected body colour luminance and body size. We additionally reconstructed phylogenetic relationships among the sampled butterfly species and investigated morphology–elevation relationships within an evolutionary framework. Butterfly assemblages were consistently darker and larger at higher elevations across three replicate locations. Furthermore, based on a phylogenetic comparative analysis, we found that body colour luminance and body size of butterfly assemblages responded to elevation through both long-term processes and more recent environmental influences. Our findings support the thermal melanism hypothesis and Bergmann’s rule from diverse subtropical and tropical butterfly assemblages, indicating elevation may structure the distributions of tropical species through morphology. The thermal functions of morphology should therefore be considered when investigating species distribution patterns and responses to environmental changes. INTRODUCTION The vulnerability of species to global warming depends on their intrinsic characteristics (sensitivity) and the magnitude of environmental changes across their distributions (exposure) (Williams et al., 2008). Ecological traits including morphological, physiological and phenological features have the potential to buffer climate change threats to species and can therefore contribute to prediction of species and community responses to environmental changes (Lavergne et al., 2010; Buckley & Kingsolver, 2012; Diamond et al., 2012; Garcia et al., 2014; Ficetola et al., 2016). Regional environmental gradients, such as elevation, can create wide environmental variation, which can structure trait variation in species or communities (McGill et al., 2006; Peters et al., 2016). Therefore, investigating trait variation along environmental gradients can reveal mechanisms of community assembly and patterns of ecological processes, such as environmental filtering and species adaptation, critical for understanding biodiversity responses to environmental changes (McGill et al., 2006; Cavender-Bares et al., 2009). Spatial patterns of thermally important morphological traits have been found within and across species along environmental gradients (Zeuss et al., 2014; Moreno Azócar et al., 2015; Peters et al., 2016). Two notable hypotheses illustrate the importance of morphological patterns across environmental gradients: the thermal melanism hypothesis (hereafter TMH) suggests that darker organisms are more efficient in transferring solar energy to body heat gain, a benefit in cool climates (Clusella-Trullas, van Wyk & Spotila, 2007; Clusella-Trullas et al., 2008); and Bergmann’s rule proposes that large body size contributes to heat maintenance, resulting in a general increase in body size or mass in colder climates (Bergmann, 1848; Partridge & French, 1996). While the thermally adaptive basis of morphology remains of debate, with contrasting evidence across taxa and geographical gradients (Ashton & Feldman, 2003; Shelomi, 2012; Peters et al., 2016; De Souza et al., 2017; Stuart-Fox, Newton & Clusella-Trullas, 2017), recent climate warming has been documented as a force in driving distribution shifts through the filtering of species based on body colour and size (Gardner et al., 2011; Sheridan & Bickford, 2011; Zeuss et al., 2014). Therefore, understanding how the TMH and Bergmann’s rule may apply to regional species assemblages along environmental gradients could provide key insights into assessments of climate change vulnerability. As related species tend to share similar ancestral traits and occur in similar environments (and thus may respond similarly to thermal gradients), phylogenetic relatedness must also be accounted for in trait analyses (Cavender-Bares et al., 2009; Wiens et al., 2010; Buckley & Kingsolver, 2012). The TMH and Bergmann’s rule have been largely tested either within species at local scales or based on species occurrence at large geographical scales (Kingsolver, 1983; Zeuss et al., 2014; Pinkert, Brandl & Zeuss, 2017). Little evidence has been shown from field observations at assemblage levels along continuous elevation profiles, especially in the tropics where high levels of biodiversity may be particularly vulnerable to warming (Deutsch et al., 2008; Bonebrake & Deutsch, 2012; Xing et al., 2016). Butterflies represent an ideal system in testing the TMH and Bergmann’s rule for tropical ecototherms (Watt, 1968; Wasserthal, 1975; Kingsolver, 1985, Bonebrake et al., 2010). Previous physiological experiments have demonstrated that colour luminance and body size are important traits in butterfly thermoregulation, and dark colours and large body sizes will significantly raise body temperature when the animal is exposed to the sun in stable environmental conditions (Watt, 1968; Wasserthal, 1975; Kingsolver, 1985; MacLean et al., 2016; Xing et al., 2016). Intraspecific thermal melanism has also been observed in both common garden experiments and along thermal gradients in several butterfly species (Ellers & Boggs, 2002; Davis, Farrey & Altizer, 2005; Karl, Geister & Fischer, 2009). In this study, we aim to test the TMH and Bergmann’s rule on butterflies at the interspecific level along elevational gradients in the tropics. We hypothesized that there would be morphological patterns across elevation gradients for tropical butterflies, and that darker and larger butterflies should be favoured by cool (high elevations) over hot (low elevations) environments. Moreover, we hoped to gain insights into how evolution has shaped these patterns. We investigated forest butterfly assemblages and the environments they experience along elevational gradients in three locations of subtropical and tropical China. Based on field community sampling and specimen measurement, we conducted a species-specific morphological analysis, and applied a phylogenetic framework to account for phylogenetic effects in trait variation. The results offer insights into how thermally important morphological traits may structure species distributions and affect climate change responses of highly diversified subtropical and tropical ectotherm assemblages. MATERIAL AND METHODS Butterfly sampling and environmental data collection We sampled butterfly assemblages in three locations in subtropical and tropical China including Mao’er Mountain National Nature Reserve in Guangxi (25°51′59.42″N, 110°24′46.05″E: subtropical evergreen broadleaf forests), five protected areas in Hong Kong (22°15′0″N, 114°10′0″E: subtropical evergreen broadleaf forests) and Xishuangbanna National Nature Reserve in Yunnan (21°36′42″–58″N, 101°34′26″–47″E: tropical rainforests) (Supporting Information, Fig. S1) during the wet season when most butterflies are active. The three locations were chosen not only because of their high levels of biodiversity in Asia, but also because they are covered with well-preserved forests across large spans of elevational range (Bascombe, Johnston & Bascombe, 1999; Chen, 2001; Wang & Tang, 2012). In Mao’er Mountain National Nature Reserve, field sampling was conducted from 14 to 27 September 2015, covering elevations from 500 to 2000 m. In Hong Kong, field sampling was conducted in five areas (Victoria Peak, Tai Po Kau Nature Reserve, Shingmun Country Park, Pok Fu Lam Reservoir Country Park and Tai Mo Shan Country Park) from 10 June to 8 July 2015, covering elevations from 170 to 673 m. In Xishuangbanna National Nature Reserve, field sampling was conducted in three areas (Xishuangbanna Botanical Garden, Bubeng and Nangongshan station) during 17–31 July 2015 and 22–28 August 2015, covering elevations from 550 to 1500 m. In total, we surveyed 38 days and 88.5 h including 32.5 h within 11 sampling days in Guangxi, 19.5 h within 10 sampling days in Hong Kong and 36.5 h within 17 sampling days in Yunnan (see Table S1 for detailed transect information). Within each location, we collected specimens in late morning and early afternoon on sunny and low wind days by walking 0.5-km-long transects, each within 30 min following standardized Pollard walk methods (Pollard, 1977). We identified species in the field by observation using visual inspection and by catching specimens using hand-nets. We recorded the abundance for each species during sampling. For each individual, we recorded the time it was encountered to track the instant air temperature and solar radiation from our data logger of the transect afterwards (see below). We also noted whether the individual occurred under sunlight by categorizing the spot where the individual was sighted into sun or shade conditions. All specimens collected were mounted for further morphological measurement. Overall, we sampled 184 species from 2500 individuals, including 1227 individuals of 80 species in Guangxi, 554 individuals of 55 species in HK, and 719 individuals of 101 species in Yunnan. To measure the air temperatures and solar radiation butterflies experienced in the forests at each transect, at least one data logger (HOBO UA-002-64, Onset Computer Corporation, Bourne, MA, USA) was placed 1 m above the ground within each transect. The data loggers were programmed to measure ambient air temperature and light intensity at 5-min intervals during the sampling period. Measured light intensity values were converted to solar radiation for further statistical analysis according to the relationship summarized by Johnson & Wilby (2013). We calculated hourly mean air temperatures and solar radiation recorded at each transect and used a multiple linear regression to quantify how hourly mean air temperatures and solar radiation were related to elevation and location. We found that both air temperature and solar radiation varied across locations, but showed a general elevational pattern, with a significant decrease in air temperature (F3,2456 = 2064, P < 0.001) and increase in solar radiation (F3,1226 = 78.32, P < 0.001) with increasing elevation (Fig. S2). Morphological analysis We used specimens from existing collections housed at Xishuangbanna Botanical Tropical Garden (collected in 2014–2015) and South China Agriculture University (collected in 2003–2012), as well as our own collections during field sampling, to acquire the colour luminance and body size information for each species. To measure colour luminance using a standardized and quantitative approach, we photographed each specimen with the same camera (Nikon D5300) under dark conditions using a standardized setting (exposure time: 1/60 s, ISO speed: 125, aperture: F/16) from a fixed distance (45 cm). We took photos of both the dorsal and the ventral sides of the wing for each sample. Previous research has indicated that the region most related to thermoregulation for butterflies is typically the body together with the basal part of the wings (Wasserthal, 1975). We therefore chose the body plus basal one-third of the wing area to analyse colour luminance (similar to Zeuss et al., 2014; Xing et al., 2016). We used Adobe Photoshop CC 2014 software to transfer each image to greyscale using the blur function and then averaged the luminosity [a value between 0 (dark) and 255 (light)] of all pixels covering the chosen region. All specimens were mounted in a standardized manner, with the hind margin of the forewing forming a straight line perpendicular to the body. We used digital calipers (0–100 mm) to measure wingspan (the distance between the tips of the forewings) as a proxy of body size (Sekar, 2012). Although intraspecific variation in traits can be important for some butterfly species (e.g. Ellers & Boggs, 2004), in this study we focused on the morphology across species (interspecific variation) as this is probably the more important component of trait variation at the assemblage level (McGill et al., 2006; Gaston, Chown & Evans, 2008). We averaged morphological values for each species based on collected and existing specimens when multiple specimens were available. Due to limitations in the identification of sex through field observation, we did not record the sex of each individual encountered. We averaged the morphological information across sex when we had specimens for both sexes. Thus, in our analysis, all individuals belonging to one species share the same morphological information. This might cause some bias in our results if sex ratios differ significantly from 1:1 or there are consistent differences across sexes (e.g. female butterflies are typically larger than males; Chou, 1994). However, overall assemblage patterns are probably robust to these effects due to the relatively small magnitude of between-sex variation compared to across-species variation (Zeuss, Brunzel & Brandl, 2017). We obtained dorsal colour luminance for 166 species, ventral colour luminance for 165 species and wingspan for 168 species (Table S2). Phylogenetic reconstruction and analysis Among all the species found in the sampling locations, we obtained sequences from ten genes regions for 139 (out of 184 species recorded in this study) from the National Center for Biotechnology Information (NCBI) (http://www.ncbi.nlm.nih.gov/). We used these sequences for phylogenetic reconstruction (see Fig. S3 for details). We extracted 300 trees from post burn-in trees in one of the Monte Carlo Markoc chain (MCMC) runs. For traits sampled by each tree, we used Pagel’s lambda (Pagel, 1999) to calculate the phylogenetic signal with the phylosig function in the R package phytools (Revell, 2012). Pagel’s lambda assumes the evolution of a trait follows the Brownian motion model, and it lies between zero and one. A Pagel’s lambda closer to zero shows low phylogenetic signal, indicating phylogenetic independence, while a value closer to one shows strong phylogenetic signal, indicating species traits are distributed under Brownian motion (Pagel, 1999).We then rescaled the trait values with Pagel’s lambda using the rescale function in the R package geiger (Harmon et al., 2008). We then applied Lynch’s comparative method (Lynch, 1991) to separate the trait values into phylogenetic components (ancestrally influenced part of the trait) and specific components (species-specific deviation part of the trait), based on 300 trees extracted with the function compar.lynch in the R package ape (Paradis, Claude & Strimmer, 2004). In this way, we calculated 300 phylogenetic components and specific components for each species. Statistical analysis We matched morphological data to environmental information including elevation, location, microclimate (sun/shade) and weather (air temperature, solar radiation) for each individual and pooled data collected from the three locations together. We then ran multiple linear models to test the effects of different environmental factors including weather (solar radiation and ambient temperature) and elevation as continuous factors, and location and sun/shade condition as categorical factors on the morphology of butterfly individuals. We then applied an Akaike information criterion (AIC) for the model selection process to select the best-fit models (lowest AIC value) to choose the most important factors in shaping colour luminance and size of butterflies (Akaike, 1974) using the R package MuMIn (Bartoń, 2013). In addition, we treated the occurrence of a species as a replicate and conducted a multiple linear regression to investigate how colour luminance and body size of the species presented were affected by elevation and location. To control for phylogeny effects on traits, we also applied the same best model we used when investigating the morphology and environment relationship at the individual level to explore how environmental factors shaped the phylogenetic and specific components of the colour luminance and body size of butterfly individuals across elevation. We ran the models based on a subset of the dataset due to limitations in the phylogenetic information for some species (N = 139). We treated each individual as a replicate, but instead of using the actual morphological information as a response variable, we used phylogenetic components and specific components. We ran the same analysis for all 300 pairs of phylogenetic components and specific components, and averaged results of 300 models as the final model result. All analyses were conducted in R v.3.2.5 (R Core Team, 2013). RESULTS Results from model selection suggest that among all the environmental factors, air temperature, solar radiation, elevation and location affected colour luminance of butterfly individuals, while sun/shade condition in addition to those factors affected wingspan (body size) of butterfly individuals (Tables S3 and S4). Both colour luminance and wingspan varied across the three locations, with butterflies in Hong Kong being the darkest and largest (Table 1). The relationship between traits and elevation, however, were consistent among different locations (Figs 1, 2). Overall, using individuals as response units, elevation had significant negative effects on colour luminance of both sides for butterfly individuals, suggesting that individuals were darker with increasing elevation (Table 1 and Fig. 1). In addition, solar radiation positively influenced colour luminance on both dorsal and ventral sides of butterfly individuals. Air temperature also showed positive effects on both sides of butterflies but the effect was only significant for the dorsal side (Table 1). As solar radiation and air temperature represented the microclimate of each transect as well as the temporary weather conditions when each individual was sampled, their effects together indicate that lighter individuals prefer warm and sunny conditions (Table 1). Elevation showed significant positive effects on the wingspan of individuals, suggesting individuals were larger at higher elevation (Table 1 and Fig. 2). While sunny conditions tended to favour large individuals in general, the prevalence of small individuals increased towards stronger solar radiation (Table 1). No significant effects of air temperature were observed on wingspan (Table 1). When using the occurrence of one species as a replicate, we found no significant effect of elevation on luminance of the dorsal wing, but a nearly significant negative effect of elevation on luminance of the ventral wing. We found a significant positive effect of elevation on wingspan. This indicates species with darker ventral wings and larger sizes tend to occur at higher elevations (Table 1). Table 1. Results of best multiple linear regression models for relationships between traits and environmental factors at individual and species levels Traits d.f. Parameter Estimate SE t value Pr(>|t|) Unit: individual Dorsal colour luminance 2272 Solar radiation 0.062 0.023 2.670 0.008** Air temperature 0.625 0.318 1.967 0.049* Elevation −0.008 0.003 −3.152 0.002 ** Hong Kong −45.378 2.369 −19.154 <0.001** Yunnan −19.532 1.679 −11.631 <0.001** Ventral colour luminance 2266 Solar radiation 0.069 0.023 3.000 0.003 ** Air temperature 0.429 0.320 1.340 0.180 Elevation −0.019 0.003 −7.162 <0.001** Hong Kong −59.622 2.380 −25.051 <0.001** Yunnan −25.971 1.687 −15.393 <0.001** Wingspan 2261 Solar radiation −0.042 0.015 −2.794 0.005 ** Air temperature −0.177 0.209 −0.848 0.396 Elevation 0.018 0.002 9.910 <0.001** Sun 2.710 1.042 2.599 0.009 ** Hong Kong 19.120 1.533 12.476 <0.001** Yunnan −7.894 1.073 −7.355 <0.001** Unit: species Dorsal colour luminance 821 Elevation −0.003 0.004 −0.883 0.378 Hong Kong −30.628 4.147 −7.384 <0.001** Yunnan −12.277 3.131 −3.920 <0.001** Ventral colour luminance 811 Elevation −0.007 0.004 −1.804 0.0716 Hong Kong −37.855 4.112 −9.204 <0.001** Yunnan −17.812 3.095 −5.754 <0.001** Wingspan 805 Elevation 0.014 0.002 5.968 <0.001** Hong Kong 20.060 2.560 7.836 <0.001** Yunnan 3.076 1.947 1.580 0.115 Traits d.f. Parameter Estimate SE t value Pr(>|t|) Unit: individual Dorsal colour luminance 2272 Solar radiation 0.062 0.023 2.670 0.008** Air temperature 0.625 0.318 1.967 0.049* Elevation −0.008 0.003 −3.152 0.002 ** Hong Kong −45.378 2.369 −19.154 <0.001** Yunnan −19.532 1.679 −11.631 <0.001** Ventral colour luminance 2266 Solar radiation 0.069 0.023 3.000 0.003 ** Air temperature 0.429 0.320 1.340 0.180 Elevation −0.019 0.003 −7.162 <0.001** Hong Kong −59.622 2.380 −25.051 <0.001** Yunnan −25.971 1.687 −15.393 <0.001** Wingspan 2261 Solar radiation −0.042 0.015 −2.794 0.005 ** Air temperature −0.177 0.209 −0.848 0.396 Elevation 0.018 0.002 9.910 <0.001** Sun 2.710 1.042 2.599 0.009 ** Hong Kong 19.120 1.533 12.476 <0.001** Yunnan −7.894 1.073 −7.355 <0.001** Unit: species Dorsal colour luminance 821 Elevation −0.003 0.004 −0.883 0.378 Hong Kong −30.628 4.147 −7.384 <0.001** Yunnan −12.277 3.131 −3.920 <0.001** Ventral colour luminance 811 Elevation −0.007 0.004 −1.804 0.0716 Hong Kong −37.855 4.112 −9.204 <0.001** Yunnan −17.812 3.095 −5.754 <0.001** Wingspan 805 Elevation 0.014 0.002 5.968 <0.001** Hong Kong 20.060 2.560 7.836 <0.001** Yunnan 3.076 1.947 1.580 0.115 *P < 0.05; **P < 0.01. View Large Figure 1. View largeDownload slide Colour luminance of dorsal (A) and ventral sides (B) of butterflies along elevations across three locations. Each dot represents one individual; darker dots indicate darker species. Solid lines show the mean relationship with standard errors (grey shaded areas) between colour luminance of butterflies and elevation within each location. Figure 1. View largeDownload slide Colour luminance of dorsal (A) and ventral sides (B) of butterflies along elevations across three locations. Each dot represents one individual; darker dots indicate darker species. Solid lines show the mean relationship with standard errors (grey shaded areas) between colour luminance of butterflies and elevation within each location. Figure 2. View largeDownload slide Wingspan of butterflies along elevations across three locations. Each dot represents one individual; larger dots indicate larger species, and darker dots indicate relatively higher frequency of the species. Solid lines show the mean relationship with standard error (grey shaded area) between butterfly wingspan and elevation within each location. Figure 2. View largeDownload slide Wingspan of butterflies along elevations across three locations. Each dot represents one individual; larger dots indicate larger species, and darker dots indicate relatively higher frequency of the species. Solid lines show the mean relationship with standard error (grey shaded area) between butterfly wingspan and elevation within each location. The phylogenetic relationship of butterfly species with phylogenetic information (N = 139) is summarized in Fig. S3. Strong phylogenetic signals were observed in both colour luminance and wingspan of butterfly species with lambda values all higher than 0.8 (closer to 1) (Fig. 3). In addition, for individual colour luminance, we found the same pattern for both phylogenetic and specific components of dorsal and ventral colour luminance, showing a decreased trend (darker) with increased elevation (Table S5 and Fig. 4). We found the same pattern for both phylogenetic and specific components of wingspan, which increased with elevation (Table S5 and Fig. 5). Figure 3. View largeDownload slide Phylogenetic signal (lambda) calculated based on 300 phylogenetic trees for butterfly dorsal colour luminance, ventral colour luminance and wingspan. A lambda value closer to one indicates high phylogenetic signal among species traits. Boxplots show medians with overall variances of lambda values for each trait. Figure 3. View largeDownload slide Phylogenetic signal (lambda) calculated based on 300 phylogenetic trees for butterfly dorsal colour luminance, ventral colour luminance and wingspan. A lambda value closer to one indicates high phylogenetic signal among species traits. Boxplots show medians with overall variances of lambda values for each trait. Figure 4. View largeDownload slide Modelled mean effects with smallest effects (dashed lines) and largest effects (dot-dashed lines) of elevation on phylogenetic (A, C)/specific (B, D) components of colour luminance for butterfly dorsal (A, B)/ventral (C, D) sides in three locations based on 300 phylogenetic trees. Figure 4. View largeDownload slide Modelled mean effects with smallest effects (dashed lines) and largest effects (dot-dashed lines) of elevation on phylogenetic (A, C)/specific (B, D) components of colour luminance for butterfly dorsal (A, B)/ventral (C, D) sides in three locations based on 300 phylogenetic trees. Figure 5. View largeDownload slide Modelled mean effects with smallest effects (dashed lines) and largest effects (dot-dashed lines) of elevation on phylogenetic (A)/specific (B) components of butterfly wingspan in three locations based on 300 phylogenetic trees. Figure 5. View largeDownload slide Modelled mean effects with smallest effects (dashed lines) and largest effects (dot-dashed lines) of elevation on phylogenetic (A)/specific (B) components of butterfly wingspan in three locations based on 300 phylogenetic trees. DISCUSSION We found a consistent pattern across three locations in subtropical and tropical China that butterflies were darker and larger at higher elevations. This observed climatic cline in butterfly morphology provides evidence supporting the TMH and Bergmann’s rule from subtropical and tropical mountain regions. The environment–morphology relationship found in this study is similar to that found in other recent studies on ectotherms across large latitudinal gradients (Zeuss et al., 2014; Moreno Azócar et al., 2015; Peters et al., 2016) and to previous findings at the habitat level in the tropics (Xing et al., 2016), which highlights the role of abiotic factors in influencing the abundance and distribution of insects from subtropical and tropical mountains. The sensitivity of morphological traits to climatic factors indicates potential vulnerability of tropical butterflies to climate change. Solar radiation and ambient temperature could have direct and substantial influences on thermoregulation and activity of ectotherms through behaviour, morphology and the interplay between them (Ohsaki, 1986; Bonebrake et al., 2014; Stuart-Fox, Newton & Clusella-Trullas, 2017). We found that butterflies tended to be lighter-coloured and smaller under strong solar radiation. This could be explained by a limited ability to thermoregulate for lighter and smaller butterflies, which are more dependent on high solar radiation levels for flight due to their slower warm-up rate and higher rate of convective heat loss (Church, 1960; Gilchrist, 1990; Pereboom & Biesmeijer, 2003; Xing et al., 2016). Besides environmental effects, we observed deep evolutionary influences on morphological traits of subtropical and tropical butterfly species. The high lambda value (close to 1) of colour luminance and wingspan showed a strong phylogenetic signal in both of the traits, indicating a tendency for species to retain their ancestor’s morphological characters over time (Harvey & Pagel, 1991; Pagel, 2002). No strong correlation was found between body colour luminance and size (Fig. S4), which suggests that colour luminance and body size may develop independently through thermal adaptation along elevation gradients. The high phylogenetic signal on colour luminance of Asian butterflies is similar to that found for colour luminance of butterflies in Europe (Zeuss et al., 2014), suggesting a conservatism in butterfly colour luminance through evolutionary processes. Similarly, a strong phylogenetic signal (Pagel’s lambda) of colour luminance was also detected among dragonfly species (Pinkert, Brandl & Zeuss, 2017), but not among ant species (Bishop et al., 2016). Therefore, the influence of evolutionary history on colour luminance may vary across different taxa and should be considered when examining environment–colour luminance relationships. As both phylogenetic and specific components of colour luminance and body size responded to elevation, the trait distribution of the butterfly assemblages may result from a mixed effect of conservatism and convergence (Webb et al., 2002). In this case, environmental niches (e.g. cool climates at high elevations) created by environmental factors (elevation in this case) may have been filled as a result of both in situ evolution of species from a certain lineage with morphological traits that fit in the niche and colonization by species of different lineages that share similar traits (Webb et al., 2002; Cavender-Bares et al., 2009). These phylogenetic influences on thermally adaptive traits could render tropical species vulnerable to warming and potentially hinder fast evolving trait adaptation that might otherwise protect these species. Although between-species trait variance could be more useful when investigating traits at assemblage levels (McGill et al., 2006), intraspecific trait variation may provide additional information on trait–environment relationships and improve our understanding of ecological and evolutionary processes (Violle et al., 2012; Classen et al., 2017). Plasticity has been observed in butterfly colour luminance and size as a response to climate (Ellers & Boggs, 2002; Juhász et al., 2016; Klockmann et al., 2016; Kingsolver & Buckley, 2017). Such phenotypic plasticity might be a contributing factor in the significance of environmental effects on specific components in this study. However, in general, our results based on trait means may underestimate the effect of plasticity or sex-specific variation in contributing to the overall assemblage trait distributions due to a lack of individual-specific morphological data. Previous studies have revealed similar climatic niche conservatism in the thermal tolerance of tropical species, suggesting their limited ability to adapt to changing local environmental conditions through physiology (Grigg & Buckley, 2013; García-Robledo et al., 2016). From our results, morphological traits such as body size and colour could be relevant to the thermal performance of organisms in addition to physiology. Moreover, the effects of morphology on the body temperature of ectotherms could be dependent on their thermoregulatory behaviour (Kingsolver, 1985; Bonebrake et al., 2014). The interactive effects and co-adaptation across physiology, morphology and behaviour can be complex, but are necessary in understanding why thermal stress may affect species through their traits (Huey & Bennett, 1987; Clusella-Trullas et al., 2008; Peters et al., 2016). Morphological and behavioural adaptation should be the targets of future study as these responses may mediate warming impacts on tropical species, which are nearer to their thermal upper limit and might be constrained to adapt to higher temperatures through physiological mechanisms (Araújo et al., 2013; Bonebrake et al., 2014; Sunday et al., 2014). In addition, biotic interactions such as sexual selection (Ellers & Boggs, 2003), predation and competition (Brown & Sax, 2004; Bonebrake, 2013; Cheng et al., 2018) could also play important roles in shaping butterfly morphology, but these were not considered in this study. Multiple selection pressures, particularly natural selection and sexual selection, may have trade-off relationships between different morphological functions. Within butterflies, wing melanism has advantages in cool climates but may not necessarily confer benefits in sexual selection (Ellers & Boggs, 2003), and could be constrained by nutrient resources during development (Talloen, Van Dyck & Lens, 2004). Moreover, other melanism functions such as disease resistance (Wilson et al., 2001) and ultraviolet-B protection (Bastide et al., 2014) may alter the elevation–colour luminance relationship from different selection processes, which might explain the insignificance of elevational effects on colour luminance at the species level. Therefore, morphological traits may not be able to easily evolve in one direction to adapt to singular changing environmental factors when other pressures exist, bringing costs to such adaptations (Ellers & Boggs, 2003). In taking a synergistic approach including community assembly, morphological traits and phylogeny, we have demonstrated the role of elevation in shaping the distribution of butterflies through thermally relevant morphological traits in the tropics. The analysis also reveals the importance of evolutionary history in influencing morphological patterns and possibly constraining adaptive responses to climate change. Lastly, this study points to the vulnerability of tropical species when facing climate change and emphasizes the need for greater conservation research efforts on tropical climate change and biodiversity. ACKNOWLEDGEMENTS We thank Gray Williams, Jacintha Ellers, Richard Saunders and three anonymous reviewers for their valuable comments on the manuscript. We also thank Prof. Wang Min for sharing his butterfly specimen collection and supporting our fieldwork in Guangxi. This work was generously supported by a National Geographic Young Explorer Grant (9709-15) and a Research Grants Council General Research Fund award to T.C.B. (HKU 760213). SUPPORTING INFORMATION Additional Supporting Information may be found in the online version of this article at the publisher's web-site: Table S1. Transect of each study site with temperature information from WorldClim during the breeding season (April to October) in 1970–2000. Table S2. Names of sampled species with morphological information (N = 184 species). Table S3. Five best multiple linear regression models for environmental factors and dorsal colour luminance of butterfly individuals. Table S4. Five best multiple linear regression models for environmental factors and wingspan of butterfly individuals. Table S5. Averaged statistical results of best multiple linear regression models for relationships of phylogenetic and specific components with environmental factors based on 300 phylogenetic trees. Figure S1. The three study locations: Hong Kong (22°15′0″N, 114°10′0″E), Yunnan (XSBN: Xishuangbanna National Nature Reserve, 21°36′42″–58″N, 101°34′26″–47″E), Guangxi (MES: Mao’er Mountain Nature Reserve, 25°51′59.42″N, 110°24′46.05″E). Figure S2. Hourly mean air temperature (A) and solar radiation (B) recorded along elevations at the three locations. Boxplots show medians with overall variances of temperature and solar radiation. Figure S3. Bayesian 50% majority-rule consensus tree generated from ten gene regions (COI, COII, 16s, NADH1, NADH5, CAD, EF1α, GAPDH, IDH and wingless). Posterior probabilities are indicated at each node (PP). Among the species found in the sampling locations, we obtained sequences for 139 (out of 184 species recorded in this study) from the National Center for Biotechnology Information (NCBI) (http://www.ncbi.nlm.nih.gov/). Another 21 species of Lepidoptera were additionally selected for phylogenetic reconstruction: Chersonesia rahria, Lamproptera meges, Mathoris loceusalis, Morova subfasciata, Potanthus flavus, Rhodoneura terminalis, Daimio tethys, Striglina cinnamomea, Hasora chromus, Hesperia comma, Danaus plexippus, Ocybadistes walkeri, Thymelicus lineola, Tagiades flesus, Pyrgus malvae, Coeliades forestan, Ctenoptilum vasava, Badamia exclamationis, Thyris fenestrella, Iphiclides podalirius and Polyura nepenthes. Thus, in total we used 155 species of butterfly as in-group species and five species of moths as out-group species. We used mitochondrial gene regions (COI, COII, 16s, NADH1 and NADH5) and combined nuclear genes (CAD, EF1α, GAPDH, IDH and wingless). The sequences were aligned by MAFFT v.7.215 (Katoh & Standley, 2013). Sequences were concatenated and manually edited in Geneious v.4.8.5 (Kearse et al., 2012). The aligned concatenated matrix contained 10689 base pairs. Bayesian inference phylogenetic reconstruction was performed using MrBayes v.3.1.2. (Ronquist & Huelsenbeck, 2003). We used two independent MCMC chains each with 20000000 generations and sampled every 5000th generation. We set a cold Markov chain with the temperature parameter of 0.16, while keeping other priors as default. We adjusted the mean branch length prior to be 0.01 [brlenspr = unconstrained : exponential (100.0)] to reduce the likelihood of stochastic entrapment in local tree length optima (Brown et al., 2010; Marshall, 2010). The resultant standard deviation of split frequencies was < 0.03. The parameters sampled during the MCMC were checked by Tracer v.1.6 (Rambaut et al., 2014) and showed adequate effective sample sizes (ESS > 200). We excluded 1000 initial samples (around 25%) for each MCMC run (burn-in = 1000) from the summary analysis and calculated a 50% majority-rule consensus tree from the post burn-in trees. The posterior probabilities (PP) of each node were summarized (Larget & Simon, 1999) to infer support for individual clades: nodes with PP values ≥ 0.95 were considered well or strongly supported (Yang & Rannala, 1997). Since the incomplete gene information for some species caused relatively high uncertainty in phylogenetic reconstruction, we constrained the phylogenetic relationship with MrBayes according to previous high-level taxa of Lepidoptera phylogeny (Heikkilä et al., 2011; Regier et al., 2013). Figure S4. Correlation between dorsal/ventral colour luminance and wingspan of 164 butterfly species. REFERENCES Akaike H. 1974. A new look at the statistical model identification. IEEE Transactions of Automatic Control 19: 716– 723. Araújo MB, Ferri-Yáñez F, Bozinovic F, Marquet PA, Valladares F, Chown SL. 2013. Heat freezes niche evolution. Ecology Letters 16: 1206– 1219. Google Scholar CrossRef Search ADS PubMed Ashton KG, Feldman CR. 2003. Bergmann’s rule in nonavian reptiles: turtles follow it, lizards and snakes reverse it. Evolution 57: 1151– 1163. Google Scholar CrossRef Search ADS PubMed Bartoń K. 2013. MuMIn: Multi-model inference. R package version 1.9. 13 . Vienna: The Comprehensive R Archive Network (CRAN). Bascombe M, Johnston G, Bascombe F. 1999. The butterﬂies of Hong Kong . London: Academic Press. Bastide H, Yassin A, Johanning EJ, Pool JE. 2014. Pigmentation in Drosophila melanogaster reaches its maximum in Ethiopia and correlates most strongly with ultra-violet radiation in sub-Saharan Africa. BMC Evolutionary Biology 14: 179. Google Scholar CrossRef Search ADS PubMed Bergmann C. 1848. Über die Verhältnisse der Wärmeökonomie der Thiere zu ihrer Grösse . Gottingen: Vandenhoeck und Ruprecht. Bishop TR, Robertson MP, Gibb H, van Rensburg BJ, Braschler B, Chown SL, Foord SH, Munyai TC, Okey I, Tshivhandekano PG, Werenkraut V. 2016. Ant assemblages have darker and larger members in cold environments. Global Ecology and Biogeography 25: 1489– 1499. Google Scholar CrossRef Search ADS Bonebrake TC. 2013. Conservation implications of adaptation to tropical climates from a historical perspective. Journal of Biogeography 40: 409– 414. Google Scholar CrossRef Search ADS Bonebrake TC, Boggs CL, Stamberger JA, Deutsch CA, Ehrlich PR. 2014. From global change to a butterfly flapping: biophysics and behaviour affect tropical climate change impacts. Proceedings of the Royal Society of London B: Biological Sciences 281: 20141264. Google Scholar CrossRef Search ADS Bonebrake TC, Deutsch CA. 2012. Climate heterogeneity modulates impact of warming on tropical insects. Ecology 93: 449– 455. Google Scholar CrossRef Search ADS PubMed Bonebrake TC, Ponisio LC, Boggs CL, Ehrlich PR. 2010. More than just indicators: a review of tropical butterfly ecology and conservation. Biological Conservation 143: 1831– 1841. Google Scholar CrossRef Search ADS Brown JH, Sax DF. 2004. Gradients in species diversity: why are there so many species in the Tropics? In: Lomolino MV, Sax DF, Brown JH, eds., Foundations of biogeography: Classic papers with commentaries . Chicago: University of Chicago Press, 1145– 1154. Buckley LB, Kingsolver JG. 2012. Functional and phylogenetic approaches to forecasting species’ responses to climate change. Annual Review of Ecology, Evolution, and Systematics 43: 205– 226. Google Scholar CrossRef Search ADS Cavender-Bares J, Kozak KH, Fine PV, Kembel SW. 2009. The merging of community ecology and phylogenetic biology. Ecology letters 12: 693– 715. Google Scholar CrossRef Search ADS PubMed Chen MY. 2001. A survey of butterfly resource in Xishuangbanna in Yunnan Province. Journal of Jilin Agriculture University 23: 50– 57. Cheng W, Xing S, Chen Y, Lin R, Bonebrake TC, Nakamura A. 2018. Dark butterflies camouflaged from predation in dark tropical forest understories. Ecological Entomology (in press). doi: 10.1111/een.12499. Chou I. 1994. Monographia rhopalocerorum sinensium . Zhengzhou: Henan Scientific and Technological Publishing House. Church NS. 1960. Heat loss and body temperature of flying insects. II. Heat conduction within the body and its loss by radiation and convection. Journal of Thermal Biology 37: 187– 212. Classen A, Steffan-Dewenter I, Kindeketa WJ, Peters MK. 2017. Integrating intraspecific variation in community ecology unifies theories on body size shifts along climatic gradients. Functional Ecology 31: 768– 777. Google Scholar CrossRef Search ADS Clusella-Trullas S, van Wyk JH, Spotila JR. 2007. Thermal melanism in ectotherms. Journal of Thermal Biology 32: 235– 245. Google Scholar CrossRef Search ADS Clusella-Trullas S, Terblanche JS, Blackburn TM, Chown SL. 2008. Testing the thermal melanism hypothesis: a macrophysiological approach. Functional Ecology 22: 232– 238. Google Scholar CrossRef Search ADS Davis AK, Farrey BD, Altizer S. 2005. Variation in thermally induced melanism in monarch butterflies (Lepidoptera: Nymphalidae) from three North American populations. Journal of Thermal Biology 30: 410– 421. Google Scholar CrossRef Search ADS De Souza AR, Turillazzi S, Lino-Neto J, Santini G. 2017. Colder environments may select for darker paper wasps. Biological Journal of the Linnean Society 120: 700– 704. Deutsch CA, Tewksbury JJ, Huey RB, Sheldon KS, Ghalambor CK, Haak DC, Martin PR. 2008. Impacts of climate warming on terrestrial ectotherms across latitude. Proceedings of the National Academy of Sciences USA 105: 6668– 6672. Google Scholar CrossRef Search ADS Diamond SE, Nichols LM, McCoy N, Hirsch C, Pelini SL, Sanders NJ, Ellison AM, Gotelli NJ, Dunn RR. 2012. A physiological trait-based approach to predicting the responses of species to experimental climate warming. Ecology 93: 2305– 2312. Google Scholar CrossRef Search ADS PubMed Ellers J, Boggs CL. 2002. The evolution of wing color in Colias butterflies: heritability, sex linkage, and population divergence. Evolution 56: 836– 840. Google Scholar CrossRef Search ADS PubMed Ellers J, Boggs CL. 2003. The evolution of wing color: male mate choice opposes adaptive wing color divergence in Colias butterflies. Evolution 57: 1100– 1106. Google Scholar CrossRef Search ADS PubMed Ellers J, Boggs CL. 2004. Functional ecological implications of intraspecific differences in wing melanization in Colias butterflies. Biological Journal of the Linnean Society 82: 79– 87. Google Scholar CrossRef Search ADS Ficetola GF, Colleoni E, Renaud J, Scali S, Padoa-Schioppa E, Thuiller W. 2016. Morphological variation in salamanders and their potential response to climate change. Global Change Biology 22: 2013– 2024. Google Scholar CrossRef Search ADS PubMed Garcia RA, Araújo MB, Burgess ND, Foden WB, Gutsche A, Rahbek C, Cabeza M. 2014. Matching species traits to projected threats and opportunities from climate change. Journal of Biogeography 41: 724– 735. Google Scholar CrossRef Search ADS PubMed García-Robledo C, Kuprewicz EK, Staines CL, Erwin TL, Kress WJ. 2016. Limited tolerance by insects to high temperatures across tropical elevational gradients and the implications of global warming for extinction. Proceedings of the National Academy of Sciences USA 113: 680– 685. Google Scholar CrossRef Search ADS Gardner JL, Peters A, Kearney MR, Joseph L, Heinsohn R. 2011. Declining body size: a third universal response to warming? Trends in Ecology & Evolution 26: 285– 291. Google Scholar CrossRef Search ADS PubMed Gaston KJ, Chown SL, Evans KL. 2008. Ecogeographical rules: elements of a synthesis. Journal of Biogeography 35: 483– 500. Google Scholar CrossRef Search ADS Gilchrist GW. 1990. The consequences of sexual dimorphism in body size for butterfly flight and thermoregulation. Functional Ecology 4: 475– 487. Google Scholar CrossRef Search ADS Grigg JW, Buckley LB. 2013. Conservatism of lizard thermal tolerances and body temperatures across evolutionary history and geography. Biology Letters 9: 20121056. Google Scholar CrossRef Search ADS PubMed Harmon LJ, Weir JT, Brock CD, Glor RE, Challenger W. 2008. GEIGER: investigating evolutionary radiations. Bioinformatics 24: 129– 131. Google Scholar CrossRef Search ADS PubMed Harvey PH, Pagel MD. 1991. The comparative method in evolutionary biology, Vol. 239 . Oxford: Oxford University Press. Huey RB, Bennett AF. 1987. Phylogenetic studies of coadaptation: preferred temperatures versus optimal performance temperatures of lizards. Evolution 41: 1098– 1115. Google Scholar CrossRef Search ADS PubMed Johnson MF, Wilby RL. 2013. Shield or not to shield: effects of solar radiation on water temperature sensor accuracy. Water 5: 1622– 1637. Google Scholar CrossRef Search ADS Juhász E, Végvári Z, Tóth JP, Pecsenye K, Varga Z. 2016. Climate-induced changes in the phenotypic plasticity of the Heath Fritillary, Melitaea athalia (Lepidoptera: Nymphalidae). European Journal of Entomology 113: 104– 112 Google Scholar CrossRef Search ADS Karl I, Geister TL, Fischer K. 2009. Intraspecific variation in wing and pupal melanization in copper butterflies (Lepidoptera: Lycaenidae). Biological Journal of the Linnean Society 98: 301– 312. Google Scholar CrossRef Search ADS Kingsolver JG. 1983. Thermoregulation and flight in Colias butterflies – elevational patterns and mechanistic limitations. Ecology 64: 534– 545. Google Scholar CrossRef Search ADS Kingsolver JG. 1985. Butterﬂy thermoregulation: organismic mechanisms and population consequences. Journal of Research on the Lepidoptera 24: 1– 20. Kingsolver JG, Buckley LB. 2017. Evolution of plasticity and adaptive responses to climate change along climate gradients. Proceedings of the Royal Society of London B: Biological Sciences 284: 20170386. Google Scholar CrossRef Search ADS Klockmann M, Karajoli F, Kuczyk J, Reimer S, Fischer K. 2016. Fitness implications of simulated climate change in three species of copper butterflies (Lepidoptera: Lycaenidae). Biological Journal of the Linnean Society 120: 125– 143. Lavergne S, Mouquet N, Thuiller W, Ronce O. 2010. Biodiversity and climate change: integrating evolutionary and ecological responses of species and communities. Annual Review of Ecology, Evolution, and Systematics 41: 321– 350. Google Scholar CrossRef Search ADS Lynch M. 1991. Methods for the analysis of comparative data in evolutionary biology. Evolution 45: 1065– 1080. Google Scholar CrossRef Search ADS PubMed MacLean HJ, Higgins JK, Buckley LB, Kingsolver JG. 2016. Morphological and physiological determinants of local adaptation to climate in Rocky Mountain butterflies. Conservation Physiology 4: cow035. Google Scholar CrossRef Search ADS PubMed McGill BJ, Enquist BJ, Weiher E, Westoby M. 2006. Rebuilding community ecology from functional traits. Trends in Ecology & Evolution 21: 178– 185. Google Scholar CrossRef Search ADS PubMed Moreno Azócar DL. 2015. Variation in body size and degree of melanism within a lizards clade: is it driven by latitudinal and climatic gradients? Journal of Zoology 295: 243– 253. Google Scholar CrossRef Search ADS Ohsaki N. 1986. Body temperatures and behavioural thermoregulation strategies of three Pieris butterflies in relation to solar radiation. Journal of Ethology 4: 1– 9. Google Scholar CrossRef Search ADS Pagel M. 1999. Inferring the historical patterns of biological evolution. Nature 401: 877– 884. Google Scholar CrossRef Search ADS PubMed Pagel M. 2002. Modelling the evolution of continuously varying characters on phylogenetic trees: the case of hominid cranial capacity. In: MacLeod N, Forey PL, eds. Morphology, shape and phylogeny . London: Taylor & Francis, 269– 286 Google Scholar CrossRef Search ADS Paradis E, Claude J, Strimmer K. 2004. APE: Analyses of Phylogenetics and Evolution in R language. Bioinformatics 20: 289– 290. Google Scholar CrossRef Search ADS PubMed Partridge L, French V. 1996. Thermal evolution of ecotherm body size: why get big in the cold? In: Johnston IA, Bennett AF, eds. Animals and temperature. Phenotypic and evolutionary adaptation . Cambridge: Cambridge University Press, 265– 292. Pereboom JJ, Biesmeijer JC. 2003. Thermal constraints for stingless bee foragers: the importance of body size and coloration. Oecologia 137: 42– 50. Google Scholar CrossRef Search ADS PubMed Peters MK, Peisker J, Steffan-Dewenter I, Hoiss B. 2016. Morphological traits are linked to the cold performance and distribution of bees along elevational gradients. Journal of Biogeography 43: 2040– 2049. Google Scholar CrossRef Search ADS Pinkert S, Brandl R, Zeuss D. 2017. Colour lightness of dragonfly assemblages across North America and Europe. Ecography 40: 1110– 1117. Google Scholar CrossRef Search ADS Pollard E. 1977. A method for assessing changes in the abundance of butterflies. Biological Conservation 12: 115– 134. Google Scholar CrossRef Search ADS R Development Core Team. 2013. R: A language and environment for statistical computing . Vienna: R Foundation for Statistical Computing. Retrieved from http://www.R-project.org/. Revell LJ. 2012. phytools: an R package for phylogenetic comparative biology (and other things). Methods in Ecology and Evolution 3: 217– 223. Google Scholar CrossRef Search ADS Sekar S. 2012. A meta-analysis of the traits affecting dispersal ability in butterflies: can wingspan be used as a proxy? Journal of Animal Ecology 81: 174– 184. Google Scholar CrossRef Search ADS PubMed Shelomi M. 2012. Where are we now? Bergmann’s rule sensu lato in insects. The American Naturalist 180: 511– 519. Google Scholar CrossRef Search ADS PubMed Sheridan JA, Bickford D. 2011. Shrinking body size as an ecological response to climate change. Nature Climate Change 1: 401– 406. Google Scholar CrossRef Search ADS Stuart-Fox D, Newton E, Clusella-Trullas S. 2017. Thermal consequences of colour and near-infrared reflectance. Philosophical Transactions of the Royal Society B: Biological Sciences 372: 20160345. Google Scholar CrossRef Search ADS Sunday JM, Bates AE, Kearney MR, Colwell RK, Dulvy NK, Longino JT, Huey RB. 2014. Thermal-safety margins and the necessity of thermoregulatory behavior across latitude and elevation. Proceedings of the National Academy of Sciences USA 111: 5610– 5615. Google Scholar CrossRef Search ADS Talloen W, Van Dyck H, Lens L. 2004. The cost of melanization: butterfly wing coloration under environmental stress. Evolution 58: 360– 366. Google Scholar PubMed Violle C, Enquist BJ, McGill BJ, Jiang L, Albert CH, Hulshof C, Jung V, Messier J. 2012. The return of the variance: intraspecific variability in community ecology. Trends in Ecology & Evolution 27: 244– 252. Google Scholar CrossRef Search ADS PubMed Wang M, Tang DM. 2012. Butterflies of Guangxi Maoershan National Nature Reserve . Nanning: Guangxi Normal University Press. Wasserthal LT. 1975. The role of butterfly wings in regulation of body temperature. Journal of Insect Physiology 21: 1921– 1930. Google Scholar CrossRef Search ADS Watt WB. 1968. Adaptive significance of pigment polymorphisms in Colias butterflies. I. Variation of melanin pigment in relation to thermoregulation. Evolution 22: 437– 458. Google Scholar CrossRef Search ADS PubMed Webb CO, Ackerly DD, McPeek MA, Donoghue MJ. 2002. Phylogenies and community ecology. Annual Review of Ecology and Systematics 33: 475– 505. Google Scholar CrossRef Search ADS Wiens JJ, Ackerly DD, Allen AP, Anacker BL, Buckley LB, Cornell HV, Damschen EI, Jonathan Davies T, Grytnes JA, Harrison SP, Hawkins BA, Holt RD, McCain CM, Stephens PR. 2010. Niche conservatism as an emerging principle in ecology and conservation biology. Ecology Letters 13: 1310– 1324. Google Scholar CrossRef Search ADS PubMed Williams SE, Shoo LP, Isaac JL, Hoffmann AA, Langham G. 2008. Towards an integrated framework for assessing the vulnerability of species to climate change. PLoS Biology 6: 2621– 2626. Google Scholar CrossRef Search ADS PubMed Wilson K, Cotter SC, Reeson AF, Pell JK. 2001. Melanism and disease resistance in insects. Ecology Letters 4: 637– 649. Google Scholar CrossRef Search ADS Xing S, Bonebrake TC, Tang CC, Pickett EJ, Cheng W, Greenspan SE, Williams SE, Scheffers BR. 2016. Cool habitats support darker and bigger butterflies in Australian tropical forests. Ecology and Evolution 6: 8062– 8074. Google Scholar CrossRef Search ADS PubMed Zeuss D, Brandl R, Brändle M, Rahbek C, Brunzel S. 2014. Global warming favours light-coloured insects in Europe. Nature Communications 5: 3874. Google Scholar CrossRef Search ADS PubMed Zeuss D, Brunzel S, Brandl R. 2017. Environmental drivers of voltinism and body size in insect assemblages across Europe. Global Ecology and Biogeography 26: 154– 165. Google Scholar CrossRef Search ADS © 2018 The Linnean Society of London, Biological Journal of the Linnean Society
Biological Journal of the Linnean Society – Oxford University Press
Published: Mar 1, 2018
It’s your single place to instantly
discover and read the research
that matters to you.
Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.
All for just $49/month
Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly
Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.
Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.
Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.
All the latest content is available, no embargo periods.
“Hi guys, I cannot tell you how much I love this resource. Incredible. I really believe you've hit the nail on the head with this site in regards to solving the research-purchase issue.”Daniel C.
“Whoa! It’s like Spotify but for academic articles.”@Phil_Robichaud
“I must say, @deepdyve is a fabulous solution to the independent researcher's problem of #access to #information.”@deepthiw
“My last article couldn't be possible without the platform @deepdyve that makes journal papers cheaper.”@JoseServera