Seed traits and phylogeny explain plant distribution at large geographic scale

Understanding the mechanisms that shape the geographic distribution of plant species is a central theme of 15 biogeography. Although seed mass, seed dispersal mode and phylogeny have long been suspected to affect species distribution, the link between the sources of variation of these attributes and their joint effects to the distribution of seed plants remain poorly documented. This study aims to quantify the joint effects of key seed traits and phylogeny on the species’ distribution. We collected seed mass and seed dispersal mode from 1,616 species of seed plants representing 554 genera of 130 families. We used 5,639,009 specimens to calculate species range size through ArcGIS10.2. Phylogenetic 20 generalized least squares regression modeling and variation partitioning were performed to estimate the joint effects of seed mass, seed dispersal mode and phylogeny on species distribution. We found that species range size was constrained by seed dispersal mode and phylogeny. Seed mass and its intraspecific variation were also important in limiting species distribution, but their effects were different among species with different dispersal modes. Variation partitioning revealed that seed mass, https://doi.org/10.5194/bg-2020-186 Preprint. Discussion started: 29 June 2020 c © Author(s) 2020. CC BY 4.0 License.

have the same time to dispersal, may co-occur and experience similar selection pressures in similar habitats, e.g., adaptive niche convergence (Losos, 2008;Grossenbacher et al., 2015). Alternatively, phylogenetic relationships can influence other ecological processes (e.g., niche partitioning in overlapping habitats) and variation in life-history traits, seed traits included, which in turn influence the distribution of species (Moles et al., 2005). Thus a species' age or degree of relatedness to other https://doi.org/10.5194/bg-2020-186 Preprint. Discussion started: 29 June 2020 c Author(s) 2020. CC BY 4.0 License. attachment to an animal body), and anemochory (dispersal by wind) according to the morphological features of their seeds or fruits (Pé rez-Harguindeguy et al., 2013). For example, seeds or fruits with wings, hairs or pappus were considered wind diffused; seeds or fruits with hooks, spines or barbs were dispersed through exozoochory; seeds or fruits with anaril or flesh offering a succulent reward for consumers were classified as endozoochory; and seeds or fruits lacking modifications pertaining to the other three categories were classed as autochory (unassisted dispersal) (Qi et al., 2014).

Construction of phylogenetic tree and statistical analyses
The phylogenetic tree was extracted from a previously published supertree (Qian and Jin, 2016) using the 'S.PhyloMaker' function in R package phytools, which was based on the APG classification of flowering plants (Zanne et al., 2014). The https://doi.org/10.5194/bg-2020-186 Preprint. Discussion started: 29 June 2020 c Author(s) 2020. CC BY 4.0 License. 'multi2d' function in the ape package was used to randomly resolve polytomies in the phylogenetic tree. To test the phylogenetic signal in species distribution, 'phylosig' function in R package phytools was used to calculate Pagel's Lambda; 115 this value ranges between 0 and 1. A value of 0 means that the evolution of the trait is independent of phylogeny, and a value of 1 indicates that trait evolution follows Brownian motion. Any value of Lambda significantly higher than zero can be regarded as a phylogenetic signal approaching Brownian motion to a different degree (Arè ne et al., 2017).
Because closely related species tend to have similar traits and interspecific analyses can be compromised by phylogenetic correlation (Lynch, 1991), a phylogenetic generalized least squares (PGLS) regression was used to determine 120 the partial effects of seed mass and seed mass variability on the range size of species for each dispersal mode, respectively.
Collinearity inspection of the above models was based on a variance inflation factor (VIF); VIF larger than four suggests collinearity (Fox, 2002). In addition, the 'rda' function in ade4 package was used to partition the variation of a species' range size explained by seed mass, seed mass variability, dispersal mode and genus (regarded as phylogeny). Finally, species range sizes were also analyzed by Tukey's HSD tests to assess potential differences between dispersal modes. Before the analysis, 125 the values of species range size were log10-transformed to meet the assumptions of normality. All statistical analyses in this study were conducted using R3.5.1 (R Core Team, 2018).

Effects of phylogeny and dispersal modes on species' range size
The phylogenetic signal of species range size was significant (Lambda = 0.515, P< 0.001). The range sizes of 130 phylogenetically related species were more similar than that for unrelated species (Fig.2). When compared to species dispersed by other modes, exozoochorous species had the largest mean distributional range size compared, while anemochorous species had the smallest mean distributional range size (Fig.3, Table A1).

Joint effects of key seed traits and phylogeny on species' distribution
Variation partitioning showed that the collective effects of seed mass, seed mass variability, dispersal mode and phylogeny, explained 40.44% of the variance associated with plant species distribution (Fig.5). Seed traits (including seed mass, seed mass variability and dispersal mode) independently explained 15.70% of the variance in species distributional range size, 145 while phylogeny explained 32.73% of the variation associated with species distributional range size (Fig.5); 7.99% of the explained variance was shared between seed traits and phylogeny (Fig.5).

The relationship between phylogeny and species distributional range size
We found a significant phylogenetic signal associated with species range size (Fig.2). This result suggest that closely related 150 species have a more similar geographic distribution range than more distantly related species. Our study corroborates similar results found in other studies (e.g., Hunt et al., 2005;Martin and Husband, 2009), but does not support those of Webb and Gaston (2003), where the range size of closely related species were not more similar to each other than expected by chance alone, and the correlation between species distributional range size and phylogeny was overestimated. This discrepancy may be due to the evolutionary history of the various taxa involved as well as the heritability of their life-history traits, which can play a critical role in the establishment and persistence of species, and thus their distributional range sizes (Angert and Schemske, 2005;Umaña et al., 2018). Seed traits associated with range size can also gradually change over evolutionary time, which can in turn increase the range of a species' distribution (Blomberg et al., 2003). Furthermore, the distribution range of a species can be influenced by its ecological tolerances associated with life-history traits (Geber and Griffen, 2003;Canham et al., 2018), which can also have a heritable component. Our results imply that the geographic distribution of 160 related species may have similar response patterns to climate change at the regional scale, due in part, to phylogenetic constraints on species distribution. Here, it seems likely that closely related species have evolved similar seed traits that result in shared adaptative strategies to climate change, although this causal mechanism requires further empirical study in the field.

165
We found that exozoochorous species generally had larger distributional range sizes (Fig.3). This result is consistent with other studies (e.g., Dupré and Ehrlé n, 2002) and can be explained by the longer dispersal distance of exozoochorous species compared to those having other modes of dispersal. Exozoochory is usually mediated by large mammals and birds, and has a greater dispersal distance than most other dispersal modes (Vittoz and Engler, 2007;Chen et al., 2019b). This finding suggests that dispersal modes can have varying effects on the distributional range size of seed plant species.

170
We also found a significant relationship between seed mass and species range size in exozoochorous and anemochorous species after controlling for phylogeny (Fig.4, Table A2). Our results are, in part, distinct from previous studies that found a significant relationship between seed mass and species distributional range size (Morin and Chuine, 2006; Procheş et al., 2012), due perhaps, to differences in the taxonomic composition of our study compared to those previous. For example, because our study is taxonomically broad, we were able to show that the effect of seed mass on seed dispersal distance may 175 be different among species that have different dispersal modes. Based on our results for exozoochorous and anemochorous species, seed mass is a key factor for dispersal distance, but for autochorous and endozoochorous species, dispersal mode is more important than seed mass for dispersal distance. Given that small-and large-seeded species have been shown to adapt https://doi.org/10.5194/bg-2020-186 Preprint. Discussion started: 29 June 2020 c Author(s) 2020. CC BY 4.0 License.
to different habitats of heterogeneous environments (Silvertown, 1989), it seems likely that the autochorous and endozoochorous species in our study may experience trade-offs between competition ability and dispersal ability through 180 seed mass variation (Chen et al., 2018), which could result in a similar effect for seed mass on species distributional range size at large geographic scales.
We found significant relationships between intraspecific seed mass variability and species distributional range size ( Fig.4). This result implies that species with a large amount of variation in seed mass have a greater adaptability to heterogeneous environments and can occupy more sites. It is interesting to note that Sides and Sloat (2014) found that 185 species with greater intraspecific variation in specific leaf area (SLA) have wider ecological breadth. Because of its potential role in modulating the responses of plant species to environmental changes, greater intraspecific functional variability may enable species to adjust to a wider range of competitive and abiotic conditions (Sides and Sloat, 2014;Manna et al., 2019).
Furthermore, variability in seed mass optimizes plant functioning in a specific environment, and the requirements for optimal seed mass functioning may differ in different environments (Rozendaal et al., 2006). Hence, our results suggest that plant 190 species with a high intraspecific variation in seed mass have more potential adaptive strategies in highly heterogeneous environments.
Plastic responses of seed mass to heterogeneous environments may be related to molecular signals at a single gene or across the entire genome (Nicotra et al., 2010), which also has the potential to influence the distribution of species (Savolainen et al., 2007). Distributional patterns of plant species may reflect the fact that individuals within a species have 195 different levels of genetic variation that are also associated with seed mass, thus facilitating a species to adapt to a broad spectrum of environments (Völler et al., 2012). It is worth noting that the regression slopes of seed mass variability were statistically different among dispersal modes in our study. This could be explained by different seed mass allocation strategies among dispersal modes (Chen et al., 2018;Chen et al., 2019a).