(PDF) Patterns of Local Community Composition Are Linked to ...mented (Uetz and Hosek 2015), and (c) a time-calibrated phylogeny is available that includes 1,262 species (Zheng and Wiens - DOKUMEN.TIPS (2024)

(PDF) Patterns of Local Community Composition Are Linked to ...mented (Uetz and Hosek 2015), and (c) a time-calibrated phylogeny is available that includes 1,262 species (Zheng and Wiens - DOKUMEN.TIPS (1)

Patterns of Local Community Composition Are Linked to

Large-Scale Diversification and Dispersal of Clades

John J. Wiens*

Department of Ecology and Evolutionary Biology, University of Arizona, Tucson, Arizona 85721

Submitted February 19, 2017; Accepted September 12, 2017; Electronically published December 20, 2017

Dryad data: http://dx.doi.org/10.5061/dryad.9bs8n.

abstract: At any location, a group of organismsmay be representedby several clades. What determines which clades will dominate localcommunities in terms of their species richness? Here, this relativelyneglected question is addressed by analyzing 166 local assemblagesof snakes distributed globally. For most regions, local assemblagesare dominated by clades with higher global-scale diversification ratesand more frequent dispersal into each region, and not by clades thathave been present in that region longer. This result contrasts withmany other studies of local richness (in other organisms), which showstrong impacts of regional colonization time on overall local speciesrichness of clades. Furthermore, even though local assemblages are as-sembled independently on different continents, most regions haveconverged on similar patterns of proportional richness. Specifically,a few rapidly diversifying clades dominate most communities aroundthe world. The high diversification rates of these clades are then linkedto their high dispersal rates. Similar patterns may occur in manygroups, such as plants, frogs, salamanders, birds, and mammals.

Keywords: biogeography, community, diversification, phylogeny,snakes, species richness.

Introduction

One of the most basic descriptors of a community is howmany species it contains of each taxonomic group. For ex-ample, a community may have 10 species of one genus andonly two of another. Such a pattern becomes particularly in-teresting when it is repeated across communities within aregion and across regions globally. What might explainwhy some clades often dominate local communities (interms of species richness) while others do not? Surprisingly,this fundamental question about the basic structure of com-munities has not been a major focus of recent ecological re-search. Nevertheless, much research has focused on impor-tant related topics, such as patterns of overall community

richness (e.g., Ricklefs and Schluter 1993; Harrison andCornell 2008), phylogenetic structure of communities (e.g.,Cavender-Bares et al. 2009), and the relative abundance ofspecies rather than the relative richness of clades (e.g., Hub-bell 2001; ter Steege et al. 2013). There is also an extensiveliterature on the ratio of species to genera in communitiesand regions (e.g., Elton 1946; Strong 1980; Enquist et al. 2002;Krug et al. 2008).A prominent study by Ricklefs and Renner (2012) ad-

dressed another related question: is richness of clades cor-related across sites? These authors examined trees fromseven tropical forest sites in South America, Asia, and Af-rica. They found that species richness per family was posi-tively correlated in pairwise comparisons between sitesacross regions. They concluded that this pattern was ex-plained by traits of families that are conserved across re-gions. However, they did not directly address the relativerichness of different clades across communities, nor did theytest explanations for these patterns.The relative richness of clades in local communities has

several potential (nonexclusive) explanations (table 1). Ingeneral, richness patterns will be directly determined bythe processes that change species numbers: dispersal, ex-tinction, and speciation (e.g., Ricklefs 1987). These pro-cesses should also directly determine the relative richnessof clades in local communities. Different ecological mecha-nisms (e.g., competition, niche divergence) may then drivethe processes of speciation, extinction, and dispersal to ul-timately explain relative richness. However, the relative im-portance of these ecological mechanisms will depend onhow these three processes (speciation, extinction, dispersal)directly determine species richness patterns.Perhaps the simplest explanation for relative richness pat-

terns is that clades that are present in a region longer will havemore time to build up higher regional richness (e.g., throughspeciation within the region), leading to higher proportionalrichness of these clades locally, relative to clades arriving inthe region more recently (table 1). This time-for-speciationeffect is demonstrably important for explaining local richness

* E-mail: [emailprotected].

Am. Nat. 2018. Vol. 191, pp. 184–196. q 2017 by The University of Chicago.0003-0147/2018/19102-57572$15.00. All rights reserved.DOI: 10.1086/695495

vol . 1 9 1 , no . 2 the amer ican natural i st february 20 18

(PDF) Patterns of Local Community Composition Are Linked to ...mented (Uetz and Hosek 2015), and (c) a time-calibrated phylogeny is available that includes 1,262 species (Zheng and Wiens - DOKUMEN.TIPS (2)

patterns within clades, both among regions and among hab-itats within regions. The time hypothesis has been directlysupported in groups including treefrogs, salamanders, andglassfrogs (Wiens et al. 2011; Kozak and Wiens 2012; Hutteret al. 2013) and indirectly supported in others, such as dam-selflies (Brown et al. 2000), turtles (Stephens and Wiens2003), and bats (Stevens 2006). However, it is unclear whetherthis time-for-speciation effect (i.e., differences in when cladesarrived) explains the relative richness of different clades at thelocal scale within regions.

A second explanation is that clades dominate local as-semblages because of the frequency of their dispersal ratherthan the timing of dispersal (table 1). For example, a clademight build up high richness in a region and local commu-nities throughmultiple colonization events by different spe-cies in the same clade. However, the frequency of dispersalalone seems unlikely to be the sole explanation for propor-tional richness patterns. For example, a clade cannot colo-nize a regionmultiple times unless many species are presentin adjacent regions, a pattern that must also be explained.Thus, this factor seems more likely to work in conjunctionwith others (e.g., within-region speciation).

A third explanation is that clades diversify at differentrates, which then translates into differences in relative rich-ness locally (table 1). The estimated net diversification rateof a clade (speciation2 extinction) is related to its richnessand age (e.g., Magallón and Sanderson 2001): older cladeswith few species have lower rates, and younger clades withmany species have higher rates. Clearly, a clade with a verylow diversification rate will have few species globally and inlocal communities. However, a clade with many species anda high global diversification rate need not dominate com-

munities across its geographic range. For example, it mighthave high local richness where it originated but lower rich-ness in regions colonized more recently. Furthermore, di-versification rates might vary strongly among geographic re-gions, such that global-scale rates are uninformative forpredicting local-scale richness patterns. Few studies (if any)have tested whether diversification rates explain the relativerichness of clades in local communities. In a similar vein,Heard and Cox (2007) analyzed the “diversity skewness” ofprimate communities (i.e., tree imbalance, potentially reflect-ing higher diversification rates in some clades) but did notrelate diversification rates to the proportional richness ofclades in local communities.A closely related explanation is that local-scale species

interactions determine global-scale diversification patterns,instead of the other way around (e.g., as discussed by Har-rison and Cornell 2008). For example, one clade might con-sistently have more species than another because it out-competes species from other clades at the local scale (e.g.,a clade that monopolizes more diverse resources locallymight prevent other clades from diversifying in associationwith use of different resources). This dominancemight thentranslate into a higher diversification rate globally. There isevidence that sympatry between clades can lower their di-versification rates (e.g., Kozak and Wiens 2010; Silvestroet al. 2015). However, these past studies did not relate thesepatterns of diversification to patterns of dominance of par-ticular clades in local communities. These latter two expla-nations (diversification vs. interactions) are potentially dif-ficult to distinguish. Nevertheless, establishing a link betweendiversification rates and proportional richness patterns in lo-cal communities is necessary to determine whether either ex-

Table 1: Summary of hypotheses to explain why some clades have higher species richness in local sites (i.e., communities orassemblages) and the patterns that they predict

Hypothesis Explanation Prediction

Colonization time Local sites in a given region will be dominated byclades that arrived in the region earlier and havehad more time for speciation within the region,increasing the clade’s regional and local speciesrichness relative to other clades

Strong relationship between the time that each cladehas been present in a region (i.e., based on theearliest colonization estimated by biogeographicreconstructions) and its mean proportional localrichness across sites in the region

Dispersal frequency Local sites will be dominated by clades that dispersedinto the region more frequently, increasing theclade’s regional and local species richness relativeto other clades

Strong relationship between the number of times thateach clade has independently colonized the region(i.e., from biogeographic reconstructions) and itsmean local proportional richness across sites inthe region

Diversification rate Local sites will be dominated by clades that havehigher diversification rates, increasing theirregional and local species richness relative to otherclades; numerous factors may explain the differ-ences in diversification rates, potentially includinglocal-scale species interactions

Strong relationship between the diversification rateof each clade and its mean proportional richnessacross local sites in the region

Richness of Clades in Local Communities 185

(PDF) Patterns of Local Community Composition Are Linked to ...mented (Uetz and Hosek 2015), and (c) a time-calibrated phylogeny is available that includes 1,262 species (Zheng and Wiens - DOKUMEN.TIPS (3)

planation is relevant at all. For example, if the proportionalrichness of clades at local sites in a region are determinedby when each clade arrived in that region (not diversificationrates), then neither of these hypotheses is relevant (i.e., impactof global diversification on local richness vs. impact of localinteractions on global diversification). Therefore, I considerthis latter explanation part of the diversification rate hypoth-esis.

These three hypotheses (table 1) need not be mutuallyexclusive, and different explanations might apply to differ-ent clades or to the same clade in different regions. For ex-ample, local richness patterns in clades with poor dispersalability and/or in geographically isolated regions (e.g., Aus-tralia, Madagascar) might tend to be dominated by coloni-zation times and/or high diversification rates of endemicclades, whereas patterns in clades with high dispersal ratesor in less isolated regions (e.g., Asia, Middle America) mightshow a stronger influence of dispersal frequency or global-scale diversification rates of widespread clades.

Here, I test the causes of proportional richness patternsin local assemblages in snakes. Snakes are a diverse andglobally distributed group, consisting of ∼24 families and∼3,600 described species (Uetz andHosek 2015). Snakes of-fer a useful model system because (a) data on local speciescomposition is available for many sites (see below), (b) broad-scale geographic ranges of species are relatively well docu-mented (Uetz and Hosek 2015), and (c) a time-calibratedphylogeny is available that includes 1,262 species (Zheng andWiens 2016). These ingredients are particularly useful forassessing the roles of dispersal and diversification in localrichness patterns.

In this study, I compile and analyze data on local richnessof snakes from 166 sites distributed among 12 global bio-geographic regions. I first determine the relative proportionof species from each family-level clade at each site. For eachregion, I estimate the mean proportional richness of cladesacross sites. I also perform biogeographic reconstructionsacross the phylogeny to determine when and how ofteneach clade colonized each region. I then use phylogeneticregression to test whether clades that colonized a regionearlier (or more frequently) dominate local communities(or “assemblages”) in that region (colonization time anddispersal frequency hypotheses; table 1). I then use dataon global richness and ages of clades to estimate their di-versification rates. For each region, I then use phylogeneticregression to test whether the mean proportional local rich-ness of clades is related to their global-scale diversificationrates (diversification rate hypothesis; table 1). I also testsome of the factors that might influence global-scale diversi-fication rates of clades, including their geographic extent(number of regions a clade occurs in) and inferred numberof dispersal events (number of shifts between regions for aclade).

Material and Methods

Local Assemblages

The literature was searched (repeatedly over several years)for studies including snake faunas of local sites (e.g., Duell-man 2005) and for species lists from national parks, na-ture reserves, and similar sites (e.g., using Google searches).Sites and references are given in appendix A; appendixes A–N are available in the Dryad Digital Repository: http://dx.doi.org/10.5061/dryad.9bs8n (Wiens 2017a). Usable listsof species were generated for 166 local assemblages (app. B).Marine sites were not considered. Sites were excluded if theyhad low richness relative to other sites in the same regionand similar habitat (potentially due to incomplete sampling)but quantifying sampling effort was not generally possible.Sites were preferred that were relatively small and containeda single habitat type (e.g., desert, tropical rain forest) butmultiple microhabitats (e.g., forests, ponds, streams). Siteswere also targeted to encompass asmany different subregionsand habitat types as possible within each region. In a few re-gions (e.g., Australian Wet Tropics, Eastern Palearctic), fewstudies of local sites were found, but detailed species-levelrangemaps were available instead. In these cases, one ormorearbitrary locations within the region were selected, and localspecies composition was inferred based on these range maps.Two potential concerns might be raised regarding the

sizes of sites. First, many sites contained multiple micro-habitats and (in some cases) habitats. Thus, these may notbe local sites or communities in the same sense as for ses-sile organisms (e.g., plants). However, a given individualsnake can utilize multiple microhabitats (e.g., tree, ground,pond), and many species can occur in multiple habitats(e.g., both grassland and forest). Thus, the larger scale for lo-cal community used here should be appropriate for snakes(alternatively, one could simply replace the word “commu-nity” with “assemblage” or another less standard term).Similarly, local sites differed in area, and larger sites mighthave more species due to greater area. However, the focushere was on proportional richness of clades at each site,not overall richness (overall richness was not compared be-tween sites). Nevertheless, data on area (ha) were obtainedfor many sites (app. A), and area showed no relationshipwith richness (n p 111, r2 p 0:007, P p :39). Therefore,results should not be biased by site area. Furthermore,the units of analysis here were clades and their mean pro-portional species richness among all the sites within a re-gion. Therefore, any potential issues with any particularsite should have limited impact on the analysis of cladesin a region. Clearly, it would be preferable to have stan-dardized sampling of multiple sites in all regions, but suchdata are unavailable for snakes (and most other organisms)at the global scale. It is also unclear how heterogeneity insite size or sampling effort would actually bias the results

186 The American Naturalist

(PDF) Patterns of Local Community Composition Are Linked to ...mented (Uetz and Hosek 2015), and (c) a time-calibrated phylogeny is available that includes 1,262 species (Zheng and Wiens - DOKUMEN.TIPS (4)

to favor one hypothesis over another (again, the units ofanalysis here are clades and not individual sites). Finally,the spatial scale used here is similar to that used in previ-ous studies that showed a strong impact of time on overalllocal richness (e.g., Wiens et al. 2011; Kozak and Wiens2012; Hutter et al. 2013), in contrast to the results here.

A total of 24 clades were used to quantify proportionalrichness of clades (fig. 1; apps. C, D; table D1). Some fam-ilies were not present in any sampled communities (e.g.,Acrochordidae, Bolyeriidae, Calabariidae, Xenophidiidae)and were therefore not included. Other families were in-cluded and treated as separate clades (17 total), except forthe hyperdiverse Colubridae. For colubrids, species wereassigned to seven subfamilies, which some authors treatas separate families (e.g., Uetz and Hosek 2015). Amongthese 24 higher taxa, two were nonmonophyletic in thetree used (Zheng andWiens 2016). First, two genera usuallyplaced in Lamprophiidae (Oxyrhabdium, Micrelaps) werecloser to Elapidae than other Lamprophiidae. Second, threegenera usually placed in Colubrinae (Ahaetulla, Chrysopelea,Dendrelaphis) grouped with Grayiinae instead. Because ofuncertainty about their placement, these five genera wereexcluded from subsequent analyses.

Species numbers for each clade at each site were used toestimate each clade’s local proportional richness (i.e., thenumber of species in that clade divided by overall localsnake richness; values for each assemblage are in app. C).The mean proportional richness for each clade across allassemblages in a region was then estimated (app. D; ta-ble D1). Proportional richness values in each local assem-blage were generally similar to these mean values for the re-gion (see below and table 2). Furthermore, excluding a fewsites that differed from the mean values should have littleimpact on the mean values. Clades present in a region,but absent at a given site, were given a value of zero whenestimating mean proportions across sites. Estimating aclade’s mean local richness among sites based only on sitesin which it was present in a region would overestimate itsmean local richness among sites across the region (e.g.,imagine a clade present at a frequency of 25% but occur-ring at only one site in the region). However, clades absententirely from a region were excluded. It would make littlesense to include, for example, clades present only in theOld World in an analysis of South American communities:their diversification rates would be irrelevant to local rich-ness and the overall importance of dispersal would be in-flated (e.g., numerous Old World clades would have zerodispersal events to NewWorld regions and zero local rich-ness in every region and community). A few species of un-certain clade assignment (e.g., the colubrine and lampro-phiid genera mentioned above) were also excluded.

A total of 12 standard ecoregions (exact definitions fol-low Pyron andWiens 2013) were used: Tropical SouthAmer-

ica (n p 24 sites), Temperate South America (n p 5),Middle America (n p 19), West Indies (n p 6), Nearctic(n p 21), Afrotropical (n p 23), Madagascar (n p 6),Western Palearctic (n p 23), Eastern Palearctic (n p 5),South Asia (n p 10), Southeast Asia (n p 11), and Aus-tralasia (n p 13). Because separate analyses were conductedfor each region, unequal numbers of sites between regions(or other differences) should have no impact. Similarly,the fact that fewer or different ecoregions could be recog-nized (e.g., 11, but excluding reptiles, in Holt et al. 2013)should also have little effect, since results were generallysimilar among regions. In fact, some of the few regions withmore unusual richness patterns (e.g., West Indies, Temper-ate South America) were not recognized as distinct by Holtet al. (2013), which suggests that following their classifica-tion strictly would ignore relevant variation.A set of analyses was performed to explicitly test whether

the mean proportional richness calculated across the sitesfor each region reflected the patterns of richness for mostsites within that region. Specifically, a nonparametric Spear-man’s rank correlation was done between the mean pro-portional richness of clades for the region and the propor-tional richness of clades for each site in that region. Theresults (table 2; app. C) show that richness patterns in mostsites in most regions are significantly correlated with themean richness pattern for the region. The major exceptionwas those regions with very few clades (n p 4). These re-gions showed similar mean correlations to other regions buthigher P values. There were also some cases in which siteswere not significantly correlated with themean regional pat-tern because of variation within a region. For example, inthe Australasian region, elapids dominate most communi-ties in Australia but not in New Guinea. Again, eliminatingoutlier sites that differed from the mean regional patternshould have very little impact. Furthermore, subdividing re-gions would be difficult and should not overturn the mainresults.The estimates of mean proportional richness of clades

within a region might potentially be biased by sites thatare geographically close to one another. However, multiplesites within the same country were generally in differentgeographic subdivisions (e.g., different states) or differentparts of these subdivisions (e.g., different parts of states).In those cases in which sites were relatively close to eachother (e.g., same national park), it was clear that these sitesdid not dominate the mean proportional richness patternsfor the region, because their r values (indicating similarityto the regional mean values; see above) were not unusuallyhigh (app. C), typically with one or more sites below theaverage r value for the region (table 2).Additional analyses were performed to address how re-

sults were influenced by the choice of clades, either treat-ing all Colubridae as one clade (18 clades overall) or all

Richness of Clades in Local Communities 187

(PDF) Patterns of Local Community Composition Are Linked to ...mented (Uetz and Hosek 2015), and (c) a time-calibrated phylogeny is available that includes 1,262 species (Zheng and Wiens - DOKUMEN.TIPS (5)

Figure 1: Summary of mean proportional richness of different snake groups in local communities in each of 12 regions (above), phylogenyof snake clades included here (lower left), and global richness of each clade (lower right). To simplify visualization, some clades are groupedinto nonmonophyletic categories (noncolubroids, noncolubrine colubroids), since many clades in these categories have low proportionalrichness. However, dipsadines dominate among noncolubrine colubroids in the West Indies and Middle and South America, whereaslamprophiids dominate in Africa and Madagascar. Snake images from the author (top, Boa constrictor; middle, Crotalus willardi; bottom,Rhinocheilus lecontei). Map from freevectormaps.com.

(PDF) Patterns of Local Community Composition Are Linked to ...mented (Uetz and Hosek 2015), and (c) a time-calibrated phylogeny is available that includes 1,262 species (Zheng and Wiens - DOKUMEN.TIPS (6)

colubroids as one clade (12 clades). Results were broadlysimilar to the main analyses using 24 clades (app. D), butrelationships were often nonsignificant given the reducedsample sizes of clades (apps. E, F). Therefore, the main anal-yses used 24 clades.

Two other factors might seem relevant but are actuallysomewhat tangential to the main focus of the study. First,species abundances were not analyzed, because the focushere is on proportional richness of clades and not relativeabundance of species. Furthermore, quantitative data onspecies abundance were not generally available, and mostsnake species tend to be similarly uncommon (e.g., relativeto lizards; Duellman 2005). Second, the relationship be-tween local and regional richness (e.g., Harrison and Cor-nell 2008) was not examined because the focus here is onthe relative richness of different clades locally, not the over-all richness of communities. Overall richness of communi-ties is the focus of studies of local versus regional richness.

Biogeographic Analyses

In order to reconstruct the timing and number of coloni-zation events for each region for each clade, dispersal eventsamong the 12 regions were estimated across the tree of1,262 species (app. G). Each species was first assigned toone or more of the regions listed above (app. H), using dis-tributional information from Uetz and Hosek (2015). TheC11 version of LaGrange (Ree and Smith 2008) was thenimplemented with the software RASP (Reconstruct Ances-tral State in Phylogenies; Yu et al. 2015). The oldest coloni-zation time for each region was estimated based on the

oldest branch that was unambiguously reconstructed asoccurring in that region, following standard practice insimilar studies (e.g., Wiens et al. 2011; Kozak and Wiens2012; Hutter et al. 2013). Following the standard threshold,a branch was unambiguously assigned to a region whenthat region’s proportional likelihood was ≥0.87. When like-lihoods were lower, the branch was assigned to the regionwith the higher proportional likelihood. If proportionallikelihoods were 1 for two regions for the same branch(e.g., given widespread species), then the colonization datewas assigned based on the next (more recent) unambiguousbranch. The colonization time was considered themidpointage of the earliest branch on which the clade occurred inthat region. When a clade’s ancestor was inferred to bepresent in a region prior to splitting within the clade, theclade’s crown group age was used (a conservative estimateof its time in the region). The stem group age was usedfor monotypic groups (which lack a crown age). From thesebiogeographic reconstructions, the following values werealso estimated: (a) the number of times a clade was inferredto have colonized a given region (apps. D–F), (b) the totalnumber of dispersal events among the 12 regions for eachclade (app. I; table I1), and (c) the total number of regionsin which a clade occurs (app. I; table I2).Given that the species-level sampling in the phylogeny is

incomplete, some colonization events may be missed. Thus,in a few cases, a clade was recorded in a community, but acolonization event for that region for that clade was lackingin the tree. In these cases, one colonization event (of un-known age) was considered to have occurred. However, thisshould have little impact on the analyses overall, since these

Table 2: Summary of results for Spearman’s rank correlation between each clade’s proportional richness at each site and the meanproportional richness across sites in a region

Regions No. cladesSites with significant correlation withregional pattern/total sites in region

Mean correlations across all sites in region

r P

Tropical South America 10 23/24 .882 .0114Temperate South America 4 0/5 .910 .1197Middle America 11 19/19 .822 .0120West Indies 4 0/6 .800 .1898Nearctic 7 15/21 .817 .0655Afrotropical 10 22/23 .840 .0176Western Palearctic 8 16/23 .778 .0634Eastern Palearctic 8 4/5 .829 .0343South Asia 13 8/10 .764 .0198Southeast Asia 15 11/11 .797 .0038Australasia 7 7/13 .728 .1458

Notes: Full results for each site are given in appendix C, available in the Dryad Digital Repository: http://dx.doi.org/10.5061/dryad.9bs8n (Wiens 2017a). Thenumber of sites with a significant correlation (P ! :05) is shown, along with mean r and P value across all sites in the region. Note that P values depend on thenumber of clades in a region and not the number of sites. Thus, there are no significant correlations in regions with only four clades. Madagascar is not listed herebecause it has only three clades.

Richness of Clades in Local Communities 189

(PDF) Patterns of Local Community Composition Are Linked to ...mented (Uetz and Hosek 2015), and (c) a time-calibrated phylogeny is available that includes 1,262 species (Zheng and Wiens - DOKUMEN.TIPS (7)

cases involved colonization events that yielded relativelyfew species in the region overall.

Diversification Analyses

The net diversification rate for each clade was estimated us-ing the standard method-of-moments estimator (r) forstem group ages (Magallón and Sanderson 2001, eq. [6]).The stem group estimator was utilized because it (a) in-corporates the entire history of the clade (i.e., from whena clade first splits from its sister taxon), avoiding potentialartifacts associated with ancient clades with very recentcrown groups; (b) is fully robust to incomplete taxon sam-pling within clades; (c) allows monotypic and species-poorclades to be included; and (d) should be more accurate thancrown group estimators, given that the accuracy of theseestimators increases with clade age (Kozak and Wiens 2016)and that crown ages are younger than stem ages. Net diver-sification (r) was estimated based on the clade age (t), the to-tal number of described species (n; from Uetz and Hosek2015), and an assumed relative extinction fraction (ε), suchthat r p ln(n(12 ε)1 ε)=t. Here, ε corrects for the biasedinclusion of clades (extant only) across the entire tree. Fol-lowing standard practice, three values for ε (0.0, 0.5, 0.9)were used. However, different values had limited impacton the overall results, and only results from the intermediatevalue (0.5) are presented in the main text.

Some authors have stated that this rate estimator as-sumes constant diversification rates and will therefore beinaccurate without a strong positive relationship betweenclade age and richness (e.g., Rabosky et al. 2012). However,these authors did not actually show that these methods gaveinaccurate estimates under these conditions. Recent simu-lations show that these estimators can be accurate (i.e.,strong correlations between true and estimated rates) re-gardless of relationships between clade age and richness(Kozak and Wiens 2016; Meyer and Wiens, forthcoming).Furthermore, the stem group estimator will correctly assignhigh rates to young clades with many species (and lowerrates to older clades with fewer species), regardless of varia-tion in instantaneous rates within clades.

Some readers might be tempted to suggest that it is circu-lar to use diversification rates to address richness patterns,given that diversification rates are estimated (in part) fromspecies richness. However, this study is analyzing patternsof proportional local richness of clades within regions,which is clearly not the same thing as the global richnessof the clade used to estimate its net diversification rate. Fur-thermore, even if this study were focused on the global rich-ness of clades, it is not inevitable that species richness anddiversification rates will be correlated among clades, asshown by simulations (Kozak and Wiens 2016) and empir-ical analyses (Scholl and Wiens 2016). For example, fast

rates in young, species-poor clades can potentially uncou-ple diversification rates and richness.Diversification rates for families were also estimated us-

ing BAMM 2.3.0 (Rabosky 2014), as described in appen-dix J. However, this gave problematic results overall. Spe-cifically, diversification rate estimates from BAMM andthe net rate estimators were significantly correlated, but dif-ferent rate estimates from BAMM for the same clade werenot (i.e., rates estimated for a clade from the whole tree orfrom that clade in isolation often gave different rate esti-mates, even though these rate estimates should be identical;for details, see app. K; tables K1–K4). Similar results wereobtained for snake clades by Meyer and Wiens (forthcom-ing). Simulation results (Moore et al. 2016; Meyer and Wiens,forthcoming) also suggest that BAMM can give inaccurate di-versification rate estimates. Therefore, the results are based onthe net rate estimator described above, not BAMM. A crit-icism (Rabosky et al. 2017) of the simulation study by Mooreet al. (2016) does not justify using BAMM here, because theempirical results here show that BAMM gives misleadingestimates in the real world for these data.In theory, one could estimate the diversification rate as-

sociated with each colonization of each region by each clade.However, this would require assigning each species in eachregion to a colonization event, and taxon sampling in thetree was inadequate for this purpose. Thus, the analyses im-plicitly assume that variation in diversification rates amongregions within a single clade do not obscure the impact ofglobal-scale rates. The overall results of this study generallysupport this assumption, showing a strong impact of global-scale diversification rates on patterns of proportional rich-ness of clades within regions (table 3).Many factors may potentially influence diversification

rates of clades, including local-scale ecology (e.g., diet, micro-habitat), large-scale patterns of climatic and geographic dis-tribution, and rates of change in ecological, morphological,and genomic traits (recent review in Wiens 2017b). Two po-tentially important and interrelated factors are the geographicextent of a clade and its rate of dispersal to new regions. Spe-cies richness and area are often related, potentially throughthe effects of area on diversification (e.g., Rosenzweig 1995).Dispersal to new areasmay open up new opportunities for di-versification that might otherwise be constrained by compe-tition in the ancestral region (e.g., Schluter 2000). Therefore,analyses were performed to test whether these two factors(geographic extent, dispersal rate) were related to global-scale diversification rates of clades (see below).

Statistical Analyses

Phylogenetic generalized least squares regression (PGLS) inthe R package caper (Orme 2013) was used to test whetherthe mean proportional species richness of clades in each re-

190 The American Naturalist

(PDF) Patterns of Local Community Composition Are Linked to ...mented (Uetz and Hosek 2015), and (c) a time-calibrated phylogeny is available that includes 1,262 species (Zheng and Wiens - DOKUMEN.TIPS (8)

gion were explained by (a) their oldest colonization time inthe region, (b) the number of times the clade colonized theregion, and/or (c) the net diversification rate of the cladeglobally. Each of these three independent variables wastested separately in each region, using the proportional rich-ness of clades as the dependent variable. However, Mada-gascar was excluded because too few clades occur there(n p 3). A reduced tree was used for each analysis in eachregion (including only clades occurring there). For PGLSanalyses, branch lengths were transformed based on thelikelihood-estimated value of phylogenetic signal (l; Pagel1999), and the k and d transformations were each set to 1(following standard practice). In addition to pairwise analy-ses, multiple regression analyses were performed when twoor more variables showed significant (or nearly significant)relationships with proportional richness. An analysis in-cluding all three variables was also performed for each re-gion, and the relative contribution of each variable to thismodel was evaluated using standardized partial regressioncoefficients (following Moen and Wiens 2017). However,the three-variable model often had poorer fit than those withfewer variables. Model fit was compared using the Akaikeinformation criterion (AIC). Relationships among the threepredictor variables were also tested for each region usingPGLS. Data for each region for each analysis are providedin appendixes D–F; trees including all clades are providedin appendixes L–N.

Some of the factors that might help explain variation indiversification rates among clades were also tested. PGLSanalyses were used to test whether diversification rates (de-pendent variables) were higher in clades that occurred inmore regions and had more dispersal events (independent

variables) and whether these two values (number of regions,dispersal events) were positively related (with number ofregions as the dependent variable and dispersal events asindependent). These analyses were conducted on the tree of24 clades (app. L), with ε p 0:5. The total number of dis-persal events and the total number of regions in which cladesoccur were log transformed (i.e., log10(variable1 0:5)) to im-prove linearity and normality. Data for these variables aregiven in appendix I.The absolute number of dispersal events is potentially in-

fluenced by the number of species in a clade. After all, aclade with only one species might have only a limited num-ber of dispersal events. Therefore, analyses were conductedusing two simple estimates of dispersal rate: the number ofdispersal events divided by the stem age of the clade (eventsper unit of time) and the number of dispersal events dividedby the number of species in the clade (events per species).These variables were also tested for their relationships withdiversification rates.

Results

Overall Patterns

Patterns of mean proportional local richness, phylogeny,and global richness of clades are summarized in figure 1.The proportional richness of clades at local sites was themain focus of this study, and for each region, the patternsof richness at individual sites were generally strongly corre-lated with the mean values for the region (table 2). Overall,the mean proportional local richness of clades was moststrongly related to the global diversification rates of clades

Table 3: Results for the phylogenetic generalized least squares regression between each clade’s proportional richness at local sites(mean across sites in a region; dependent variable) and three independent variables: the timing of the first colonization of thatclade in the region, the clade’s net diversification rate (for ε p 0:5), and the number of colonization events by that clade intothe region (log transformed)

Regions No. clades

Time of first colonization Diversification rate No. colonization events

r2 P r 2 P r2 P

Tropical South America 10 .076 .545 .461 .018 .552 .007Temperate South America 4 .023 .954 .814 .103 .902 .051Middle America 11 .407 .032 .524 .005 .503 .012West Indies 4 .139 .756 .879 .064 .764 .134Nearctic 7 .112 .570 .322 .189 .642 .022Afrotropical 10 .000 1.000 .480 .015 .011 .917Western Palearctic 8 .154 .395 .452 .054 .272 .188Eastern Palearctic 8 .020 .887 .311 .145 .556 .023South Asia 13 .096 .347 .493 .003 .464 .004Southeast Asia 15 .008 .907 .487 .001 .368 .007Australasia 7 .114 .632 .293 .299 .000 .998

Note: All three variables were tested across all clades occurring in each region. Significant results are in boldface. The sample size is the number of clades ineach region. Multiple regression results are in appendix D, table D2, available in the Dryad Digital Repository: http://dx.doi.org/10.5061/dryad.9bs8n (Wiens2017a).

Richness of Clades in Local Communities 191

(PDF) Patterns of Local Community Composition Are Linked to ...mented (Uetz and Hosek 2015), and (c) a time-calibrated phylogeny is available that includes 1,262 species (Zheng and Wiens - DOKUMEN.TIPS (9)

in 7 of the 11 included regions, and diversification ratesexplained 145% of the variation in proportional richness in8 regions (table 3). The number of colonization events byeach clade into the region explained the most variation inproportional richness in the other four regions and 145%of this variation in seven regions (table 3). The timing ofcolonization did not explain the most variation in propor-tional richness (or 145%) in any region and had a signifi-cant effect on proportional local richness in only one region(Middle America; table 3). Diversification rates and pro-portional richness of clades were significantly related in 5of 11 regions (table 3), including the most species-rich re-gions globally (Tropical South America, Middle America,Afrotropical, South Asia, Southeast Asia). Two additionalregions showed strong relationships (r2 1 0:80) that werenonsignificant given the few clades present (n p 4; Tem-perate South America, West Indies). Four other regions(also with fewer clades) did not show significant relation-ships (Australasia, Nearctic, Eastern Palearctic, Western Pa-learctic). Colonization events per clade were significantlyrelated to their proportional richness in seven regions (ta-ble 3; all but Afrotropical, Australasian, Western Palearctic,and West Indian regions).

Comparison of the AIC of single and multiple regressionmodels for each region showed that models with one pre-dictor variable often had the best fit (app. D; table D2), withsome regions showing colonization events as the most im-portant variable (Tropical and Temperate South America,Nearctic, Eastern Palearctic) and others supporting diversi-fication rates instead (e.g., Australia, Southeast Asia, WestIndies). In Middle America, a model including diversifica-tion rates and time had the best fit, although the improve-ment in fit was negligible (1.061 AIC units) relative to thenext best model (including diversification rates alone). Sim-ilarly, for the Afrotropical andWestern Palearctic regions, amodel including all three variables had the best fit, but againthe improvement in fit was very small (!1.2 AIC units) rel-ative to the next best model (diversification rates alone). InSouth Asia, a model including all three variables had thebest fit by a more substantial margin (3.808 AIC units).Across all regions, models that included both diversifi-cation rates and number of colonization events generallyexplained similar amounts of variation in proportionalrichness relative to models including one (or both) of thesevariables separately. Thus, for example, diversification ratesand number of colonization events might each separatelyexplain ∼50% of the variation in proportional richness, buta model including both often explained !60%. These twovariables were strongly related among clades both globally(see below) andwithinmany regions (app. D; table D3).Mul-tiple regression analyses including all three variables si-multaneously showed that diversification rates and num-ber of colonization events combined always explained the

majority of variation in proportional richness of clades foreach region, but with diversification rates having a consis-tently stronger effect than the number of colonization events(app. D; table D2).These overall patterns were generally similar using alter-

native estimates of diversification rates (app. D; tables D4,D5) and alternative ways of subdividing clades (apps. E,F). Specifically, mean proportional local richness of cladeswas most strongly linked to diversification rates and thenumber of colonization events, whereas the timing of col-onization of each region had generally little impact. Whenconsidering only 12 clades, the number of colonization eventswas the dominant factor.Clade diversification rates (24 clades, ε p 0:5) were sig-

nificantly related to their total number of dispersal eventsamong regions (r2 p 0:546, P ! :001) and the total numberof regions where they occurred (r2 p 0:469, P ! :001). Thenumber of regions and the number of dispersal events werestrongly related (r2 p 0:872, P ! :001). There was also astrong relationship between the dispersal rate (events/time)and diversification rate of clades (r2 p 0:560, P ! :001), butnot between diversification rate and the per-species dis-persal rate (r2 p 0:045, P p :319). These latter results sug-gest that accelerated diversification is coupled with acceler-ated dispersal but that the ratio of species (and, possibly,speciation) to dispersal is similar across clades. Thus, it ap-pears that many clades have similar numbers of species perdispersal event but that some clades may have exceptionaldiversification rates because they disperse rapidly and re-peatedly among regions (although more diversification mayalso facilitate more dispersal).

Patterns among Clades

These overall results can be explained in terms of the dynam-ics of specific clades. All regions contain communities withboth “primitive” snake lineages (including blind snakes, boas,pythons, and relatives) and “advanced” snakes (Colubroidea,including the highly venomous elapids, viperids, and thehyperdiverse Colubridae). Primitive snake lineages failed todominate communities in almost all regions (fig. 1), eventhough they have been present in many regions longer thanany colubroid clades. These noncolubroid clades generallyhave lower diversification rates (0–0.090, ε p 0:5; app. K).In contrast, colubroids dominate communities in almost

every region (fig. 1). Mean proportional local richness ofColubroidea was 172% in almost every region and 190%in many (Nearctic, Eastern and Western Palearctic, MiddleAmerica, Tropical and Temperate South America; app. F).The sole exception was theWest Indies (43%). These sharedpatterns across continents arose largely through indepen-dent evolution of species on different continents, ratherthan dispersal of a set of widespread species shared across

192 The American Naturalist

(PDF) Patterns of Local Community Composition Are Linked to ...mented (Uetz and Hosek 2015), and (c) a time-calibrated phylogeny is available that includes 1,262 species (Zheng and Wiens - DOKUMEN.TIPS (10)

communities in different regions. For example, among the1,262 species included in the biogeographic analyses (app. H),most species (76.1%) are confined to a single region, 21.2%occur in only two regions, 1.8% occur in three regions, andless than 1% occur in four or more regions.

Remarkably, one rapidly diversifying clade within Colu-broidea (Colubrinae; r p 0:131, where r is the diversifica-tion rate) was the most species rich in local communitiesin most regions across the globe, regardless of when they ar-rived, including the Palearctic, Asia, Nearctic, and MiddleAmerica (but absent in Temperate South America, Mada-gascar, and the West Indies). Yet, even among colubroidlineages, colubrines were frequently not the first lineagesto colonize these regions. Colubrines also showed numerouscolonization events into almost all regions where they occur(mean p 14:9, range p 22 26; app. D). They also showthe most rapid dispersal rate (per unit time) of any snakeclade (4.43 events per million years, more than twice thatof the next fastest clade). In contrast, their per-species dis-persal rate (0.273 events per species) is almost identical tothe average value across all 24 clades (mean p 0:272). Col-ubrines show extensive dispersal along with within-regionspeciation.

Three important exceptions to the pattern of colubrinedominance were in Tropical South America (dominatedby the rapidly diversifying dipsadines, r p 0:124, with col-ubrines a distant second), Africa (dominated by lampro-phiids, r p 0:105, with colubrines a close second), andAustralasia (dominated by elapids, with a moderately fastdiversification rate, r p 0:117). In all three cases, colubrinesarrived after the currently dominant colubroid clade in theregion (and by substantial margins in the Australasian andAfrotropical regions).

Intriguingly, two other colubroid lineages (elapids, vi-perids) have dispersed almost as widely as colubriness butwere generally less diverse locally (fig. 1). Viperids occuron all continents with snakes, except Australasia. Viperidsfrequently co-occur with colubrines but are generally lessdiverse locally. Nevertheless, they colonized several regionsbefore colubrines (e.g., Afrotropical, Eastern Palearctic,Western Palearctic, South Asia), often with numerous col-onization events. Similarly, elapids occur on all continentswith snakes and often co-occur with colubrines. Elapidscolonized before colubrines in the Eastern Palearctic andSoutheast Asia but are less diverse locally in both. Colu-brines colonized before elapids in Afrotropical, WesternPalearctic, South Asia, Nearctic, Middle America, and Trop-ical South America.

Discussion

This study addresses a fundamental but underexplored topicin ecology: the relative species richness of different clades in

local communities. The results from snakes showed that lo-cal communities within each region were generally domi-nated by clades with higher diversification rates globallyand that colonized the region more frequently, not by cladesthat colonized the region earlier. Moreover, there were strongrelationships between diversification rates and how fre-quently (and how rapidly) clades dispersed, as well as be-tween diversification rates and the geographic extents ofclades.These results are surprising on two main levels. First, the

results here on relative local richness between clades differstrikingly from studies of local richness patterns withinclades, which showed that the relative timing of coloni-zation of each region strongly impacted local richness,whereas diversification rates did not (e.g., Wiens et al. 2011;Kozak and Wiens 2012; Hutter et al. 2013). Second, theseresults show that global-scale diversification rates stronglyimpact local-scale richness patterns. Within a region, it isintuitive that a rapidly diversifying clade can generate manyspecies, which can then outnumber species of other cladesregionally and at local sites. Remarkably, the results hereshow that a rapidly diversifying clade can spread globallyand dominate almost everywhere it occurs, regardless ofhow long other clades have been there. This is the case withcolubroid snakes overall, and especially with the Colubri-nae. Colubrines have spread repeatedly to almost every con-tinent (typically colonizing each region 110 times) and arethe richest clade in many regions, including some regionswhere they arrived later than many other snake clades. How-ever, there were also a few cases in which colubrines arrivedlater than other colubroid clades, and these other colubroidclades dominated those regions instead (elapids in Austral-asia, dipsadines in Tropical South America, lamprophiidsin Africa).How general are these patterns? This is difficult to say

without similar analyses in other groups. Nevertheless,many well-known groups of organisms are also numericallydominated by relatively young clades with high richness,rapid diversification rates, and near-global geographic ex-tent (e.g., angiosperm plants [Magallón et al. 2015], placen-tal mammals [Meredith et al. 2011], passerine birds [Eric-son et al. 2014], neobatrachian frogs [Roelants et al. 2007;Wiens 2007]). This numerical dominance is almost cer-tainly reflected in local richness patterns in most global re-gions. Therefore, it seems likely that the patterns found herein snakes may apply to many other organisms.Intriguingly, in salamanders and frogs, it appears that the

proportional local richness of clades is also explained bydifferences in their diversification rates, even though differ-ences in local richness among regions are explained primar-ily by colonization time. For example, plethodontids dom-inate salamander communities in those regions with thehighest salamander richness (Appalachia, Central America;

Richness of Clades in Local Communities 193

(PDF) Patterns of Local Community Composition Are Linked to ...mented (Uetz and Hosek 2015), and (c) a time-calibrated phylogeny is available that includes 1,262 species (Zheng and Wiens - DOKUMEN.TIPS (11)

Petranka 1998; Kozak and Wiens 2012), and plethodontidshave the highest diversification rate in salamanders (Gómez-Rodríguez et al. 2015). Yet, patterns of local plethodontidrichness among regions are explained by time rather than di-versification rates (Kozak and Wiens 2012). Similarly, infrogs, local richness patterns are dominated by the cladeNeobatrachia (based on sites from Moen et al. 2016), a cladewith very high diversification rates relative to other frogs(e.g., Roelants et al. 2007; Wiens 2007). Nevertheless, anal-yses within frog families show strong impacts of time on lo-cal richness patterns but not of diversification rates (e.g.,Wiens et al. 2011; Hutter et al. 2013).

Many patterns in biology depend on the scale at whichthey are examined. Thus, examining relative richness ofclades at different phylogenetic or geographic scales mightreveal somewhat different patterns. Importantly, the spatialscale for local sites used here is similar to that in studiesshowing time of colonization (rather than diversificationrates) as the main driver of overall local richness (e.g., Wienset al. 2011; Kozak and Wiens 2012; Hutter et al. 2013). Dif-ferent patterns might emerge at a smaller spatial scale. Sim-ilarly, the timing of colonization might be more importantthan diversification rates over shorter temporal and phylo-genetic scales (e.g., Hutter et al. 2017; Pontarp and Wiens2017), especially in groups that have not yet become globallywidespread. Conversely, biogeographic timing might be lessimportant at deeper timescales (although the number of col-onization events might remain important).

An important question arising from this research is: whatexplains the differences in diversification rates and coloniza-tion frequency among these snake clades? At least two (non-exclusive) explanations seem likely. First, intrinsic factors insome clades might explain their rapid radiation. Second,local-scale species interactions might determine the global-scale patterns of diversification. There may also be somecombination of these first two explanations. The resultshere do not resolve this question but offer some relevant ob-servations. The rapid radiation of colubroid snakes is wellknown (e.g., Stanley 1979), but the causes have remainedunclear. The results here show that colubroid snakes havenot simply diversified rapidly overall. Instead, only certaincolubroid clades have high diversification rates, and thoseclades now dominate most communities worldwide (colu-brines, dipsadines, lamprophiids). Many other colubroidclades of similar age have failed to radiate as successfullyand rapidly (e.g., pareatids, xenodermatids, calamariines,grayiines, pseudoxenodontines, sibynophiines).

The ecological and evolutionary lability of these threecolubroid clades might explain their success. First, the re-sults here show a strong relationship between diversifica-tion rates of snake clades and how often they have dispersedand to how many regions (and their dispersal rate per unittime). Thus, many colubroid lineages that failed to diversify

rapidly are largely confined to single continents (e.g., Asiafor xenodermatids, pareatids, homalopsids, and calama-riines). Furthermore, many of these clades are confinedlargely to mesic habitats in tropical and subtropical regions(Pough et al. 2016). Climatic niche conservatism may limittheir ability to spread across regions and continents. Thisexplanation may also apply to many noncolubroid lineages,most of which are confined to mesic tropical and subtrop-ical regions (e.g., aniliids, anomalepidids, cylindrophiids,uropeltids, xenopeltids). The rapidly diversifying lineagesoccur in both tropical and temperate habitats and both me-sic and arid conditions (app. D). Second, the rapidly radiat-ing clades are ecologically diverse on niche axes that mightbe important for local-scale niche differentiation and re-source partitioning. Thus, colubrines, dipsadines, and lam-prophiids each collectively include sets of species spanningdiverse microhabitats (e.g., aquatic, arboreal, terrestrial, bur-rowing), having diverse diets (e.g., invertebrates, vertebrates),and with a wide range of body sizes (data in, e.g., Ashe et al.2002; Duellman 2005; Pough et al. 2016; Bars-Closel et al.2017). In contrast, many clades that failed to diversify rapidlyin terms of species richness also failed to diversify ecologi-cally. For example, many slowly diversifying snake cladesare primarily burrowers (e.g., aniliids, cylindrophiids, lep-totyphlopids, loxocemids, xenopeltids, uropeltids) or aquatic(grayiines, homalopsids; Bars-Closel et al. 2017). Future stud-ies should test whether ecological diversification drove speciesdiversification of these hyperdiverse colubroid clades. It willbe especially important to disentangle whether these cladesare more species rich because they are ecologically diverseor are ecologically diverse because they are more species rich(i.e., by using estimates of rates of change in ecological var-iables).The results here show other intriguing patterns that

should be explored in future studies. One particularly inter-esting pattern involves the viperids and elapids, two fami-lies that possess remarkable venom systems and includemost of the snake species that are dangerous to humans(Pough et al. 2016). The results show that elapids and vi-perids are often depauperate in local communities (fig. 1),even though they are as widespread as colubrines (and some-times arrived in particular regions earlier). Yet, elapids dom-inate Australia, where colubrines arrived much later. Thesepatterns suggests that competitionmight limit the diversifica-tion and local richness of these venomous snake clades. Theresults also show that some of the most rapidly diversifyingclades repeatedly diversified in sympatry with each other(e.g., dipsadines and colubrines inMiddle and Tropical SouthAmerica; lamprophiids and colubrines in Africa), in contrastto the general idea that sympatry between clades constrainsdiversification (e.g., Schluter 2000).In summary, this study provides an initial exploration of

a fundamental but neglected aspect of communities: the rel-

194 The American Naturalist

(PDF) Patterns of Local Community Composition Are Linked to ...mented (Uetz and Hosek 2015), and (c) a time-calibrated phylogeny is available that includes 1,262 species (Zheng and Wiens - DOKUMEN.TIPS (12)

ative richness of co-occurring clades. The results show thatproportional richness of clades locally is explained primar-ily by global-scale diversification rates of clades and theirfrequent dispersal into each region, rather than the timingof colonization. Further testing the generality and causes ofthese patterns will be an important area for future research.

Acknowledgments

I am very grateful to Verônica Thiemi-Tsutae de Sousa for allher extensive work on this project. I thank Tania Hernández-Hernández for assistance calculating standardized partial re-gression coefficients. I thank Yannis Michalakis, David Col-lar, and two anonymous reviewers for helpful comments onthe manuscript. I thank the National Science Foundation(DEB 1655690) for support.

Literature Cited

Ashe, J., K. Howell, S. Spawls, and R. Drewes. 2002. A field guide tothe reptiles of East Africa. Academic Press, London.

Bars-Closel, M., T. Kohlsdorf, D. S. Moen, and J. J. Wiens. 2017. Di-versification rates are more strongly related to microhabitat thanclimate in squamate reptiles (lizards and snakes). Evolution 71:2243–2261.

Brown, J. M., M. A. McPeek, and M. L. May. 2000. A phylogeneticperspective on habitat shifts and diversity in the North AmericanEnallagma damselflies. Systematic Biology 49:697–712.

Cavender-Bares, J., K. H. Kozak, P. V. A. Fine, and S. W. Kembel.2009. The merging of community ecology and phylogenetic biol-ogy. Ecology Letters 12:693–715.

Duellman, W. E. 2005. Cusco Amazónico: the lives of amphibiansand reptiles in an Amazonian rainforest. Comstock, Ithaca, NY.

Elton, C. S. 1946. Competition and the structure of animal commu-nities. Journal of Animal Ecology 15:54–68.

Enquist, B. J., J. P. Haskell, and B. H. Tiffney. 2002. General patternsof taxonomic and biomass partitioning in extant and fossil plantcommunities. Nature 419:610–613.

Ericson, P. G. P., S. Klopfstein, M. Irestedt, J. M. T. Nguyen, and J. A. A.Nylander. 2014. Dating the diversification of the major lineages ofPasseriformes (Aves). BMC Evolutionary Biology 14:8.

Gómez-Rodríguez, C., A. Baselga, and J. J. Wiens. 2015. Is climaticniche width related to diversification rate? Global Ecology and Bio-geography 24:383–395.

Harrison, S. P., and H. V. Cornell. 2008. Toward a better under-standing of the regional causes of local community richness. Ecol-ogy Letters 11:969–979.

Heard, S. B., and G. H. Cox. 2007. The shapes of phylogenetic treesof clades, faunas, and local assemblages: exploring spatial patternin differential diversification. American Naturalist 169:E107–E118.

Holt, B. G., J.-P. Lessard, M. K. Borregaard, S. A. Fritz, M. B. Araujo,D. Dimitrov, P. H. Fabre, et al. 2013. An update of Wallace’s zoo-geographic regions of the world. Science 339:74–78.

Hubbell, S. P. 2001. A unified neutral theory of biodiversity and bio-geography. Princeton University Press, Princeton, NJ.

Hutter, C. R., J. M. Guayasamin, and J. J. Wiens. 2013. Explaining An-dean megadiversity: the evolutionary and ecological causes of glass-frog elevational richness patterns. Ecology Letters 16:1135–1144.

Hutter, C. R., S. M. Lambert, and J. J. Wiens. 2017. Rapid diversifi-cation and time explain amphibian species richness at differentscales in the Tropical Andes, Earth’s most biodiverse hotspot.American Naturalist 190:828–843.

Kozak, K. H., and J. J. Wiens. 2010. Accelerated rates of climatic-nicheevolution underlie rapid species diversification. Ecology Letters13:1378–1389.

———. 2012. Phylogeny, ecology, and the origins of climate-richnessrelationships. Ecology 93:S167–S181.

———. 2016. Testing the relationships between diversification, spe-cies richness, and trait evolution. Systematic Biology 65:975–988.

Krug, A. Z., D. Jablonski, and J. W. Valentine. 2008. Species-genus ra-tios reflect a global history of diversification and range expansion inmarine bivalves. Proceedings of the Royal Society B 275:1117–1123.

Magallón, S., S. Gómez-Acevedo, L. Sánchez-Reyes, and T. Hernández-Hernández. 2015. A metacalibrated time-tree documents the earlyrise of flowering plant phylogenetic diversity. New Phytologist 207:437–453.

Magallón, S., and M. J. Sanderson. 2001. Absolute diversificationrates in angiosperm clades. Evolution 55:1762–1780.

Meredith, R. W., J. E. Janecka, J. Gatesy, O. A. Ryder, C. A. Fisher,E. C. Teeling, A. Goodbla, et al. 2011. Impacts of the Cretaceousterrestrial revolution and KPg extinction on mammal diversifica-tion. Science 334:521–524.

Meyer, A. L. S., and J. J. Wiens. Forthcoming. Estimating diversifica-tion rates for higher taxa: BAMM can give problematic estimatesof rates and rate shifts. Evolution.

Moen, D. S., H. Morlon, and J. J. Wiens. 2016. Testing convergenceversus history: convergence dominates phenotypic evolution forover 150 million years in frogs. Systematic Biology 65:146–160.

Moen, D. S., and J. J. Wiens. 2017. Microhabitat and climatic-nichechange explain patterns of diversification among frog families.American Naturalist 190:29–44.

Moore, B. R., S. Höhna, M. R. May, B. Rannala, and J. P. Huelsenbeck.2016. Critically evaluating the theory and performance of Bayesiananalysis of macroevolutionary mixtures. Proceedings of the Na-tional Academy of Sciences of the USA 113:9569–9574.

Orme, D. 2013. The caper package: comparative analysis of phy-logenetics and evolution in R. R package, version 0.5.2.

Pagel, M. 1999. Inferring the historical patterns of biological evolu-tion. Nature 401:877–884.

Petranka, J. W. 1998. Salamanders of the U.S. and Canada. Smith-sonian Institution, Washington, DC.

Pontarp, M., and J. J. Wiens. 2017. The origin of species richnesspatterns along environmental gradients: uniting explanationsbased on time, diversification rate, and carrying capacity. Journalof Biogeography 44:722–735.

Pough, F. H., R.M. Andrews,M. L. Crump, A.H. Savitzky, K. D.Wells, andM. C. Brandley. 2016. Herpetology. 4th ed. Sinauer, Sunderland, MA.

Pyron, R. A., and J. J. Wiens. 2013. Large-scale phylogenetic analysesreveal the causes of high tropical amphibian diversity. Proceedingsof the Royal Society B 280:20131622.

Rabosky, D. L. 2014. Automatic detection of key innovations, rate shifts,and diversity-dependence on phylogenetic trees. PLoSONE 9:e89543.

Rabosky, D. L., J. S. Mitchell, and J. Chang. 2017. Is BAMM flawed?theoretical and practical concerns in the analysis of multi-rate di-versification models. Systematic Biology 66:477–498.

Rabosky, D. L., G. J. Slater, and M. E. Alfaro. 2012. Clade age andspecies richness are decoupled across the eukaryotic tree of life.PLoS Biology 10:e1001381.

Richness of Clades in Local Communities 195

(PDF) Patterns of Local Community Composition Are Linked to ...mented (Uetz and Hosek 2015), and (c) a time-calibrated phylogeny is available that includes 1,262 species (Zheng and Wiens - DOKUMEN.TIPS (13)

Ree, R. H., and S. A. Smith. 2008. Maximum likelihood inference ofgeographic range evolution by dispersal, local extinction, andcladogenesis. Systematic Biology 57:4–14.

Ricklefs, R. E. 1987. Community diversity: relative roles of local andregional processes. Science 235:167–171.

Ricklefs, R. E., and S. S. Renner. 2012. Global correlations in tropicaltree species richness and abundance reject neutrality. Science335:464–467.

Ricklefs, R. E., and D. Schluter. 1993. Species diversity in ecologicalcommunities: historical and geographical perspectives. Universityof Chicago Press, Chicago.

Roelants, K., D. J. Gower, M. Wilkinson, S. P. Loader, S. D. Biju, K.Guillaume, L. Moriau, and F. Bossuyt. 2007. Global patterns of di-versification in the history of modern amphibians. Proceedings ofthe National Academy of Sciences of the USA 104:887–892.

Rosenzweig, M. L. 1995. Species diversity in space and time. Cam-bridge University Press, Cambridge.

Schluter, D. 2000. The ecology of adaptive radiation. Oxford Univer-sity Press, Oxford.

Scholl, J. P., and J. J. Wiens. 2016. Diversification rates and speciesrichness across the Tree of Life. Proceedings of the Royal SocietyB 283:20161335.

Silvestro, D., A. Antonelli, N. Salamin, and T. B. Quental. 2015. Therole of clade competition in the diversification of North Americancanids. Proceedings of the National Academy of Sciences of theUSA 112:8684–8689.

Stanley, S. M. 1979. Macroevolution: pattern and process. Freeman,San Francisco.

Stephens, P. R., and J. J.Wiens. 2003. Explaining species richness fromcontinents to communities: the time-for-speciation effect in emydidturtles. American Naturalist 161:112–128.

Stevens, R. D. 2006. Historical processes enhance patterns of diver-sity along latitudinal gradients. Proceedings of the Royal Society B273:2283–2289.

Strong, D. R. 1980. Null hypotheses in ecology. Synthese 43:271–285.ter Steege, H., N. C. A. Pitman, D. Sabatier, C. Baraloto, R. P. Salamao,

J. E. Guevara, O. L. Phillips, et al. 2013. Hyperdominance in theAmazonian tree flora. Science 342:1243092.

Uetz, P., and J. Hošek, eds. 2015. The Reptile Database. AccessedFeb. 18, 2015. http://www.reptile-database.org.

Wiens, J. J. 2007. Global patterns of diversification and species rich-ness in amphibians. American Naturalist 170(suppl.):S86–S106.

———. 2017a. Data from: Patterns of local community compositionare linked to large-scale diversification and dispersal of clades.American Naturalist, Dryad Digital Repository, http://dx.doi.org/10.5061/dryad.9bs8n.

———. 2017b. What explains patterns of biodiversity across theTree of Life? BioEssays 39:1600128.

Wiens, J. J., R. A. Pyron, and D. C. Moen. 2011. Phylogenetic originsof local-scale diversity patterns and the causes of Amazonianmegadiversity. Ecology Letters 14:643–652.

Yu, Y., A. J. Harris, C. Blair, and X. He. 2015. RASP (ReconstructAncestral State in Phylogenies): a tool for historical biogeography.Molecular Phylogenetics and Evolution 87:46–49.

Zheng, Y., and J. J. Wiens. 2016. Combining phylogenomic andsupermatrix approaches, and a time-calibrated phylogeny for squa-mate reptiles (lizards and snakes) based on 52 genes and 4,162 spe-cies. Molecular Phylogenetics and Evolution 94:537–547.

Associate Editor: David C. CollarEditor: Yannis Michalakis

Snakes belonging to the family Colubridae from the Jinggang Mountains of China. Left, a Sichuan mountain keelback snake (Opisthotropislatouchii). Right, a yellow-banded big-tooth snake (Lycodon flavozonatus). Photo credit: John J. Wiens.

196 The American Naturalist

(PDF) Patterns of Local Community Composition Are Linked to ...mented (Uetz and Hosek 2015), and (c) a time-calibrated phylogeny is available that includes 1,262 species (Zheng and Wiens - DOKUMEN.TIPS (2024)
Top Articles
Latest Posts
Article information

Author: Ray Christiansen

Last Updated:

Views: 5723

Rating: 4.9 / 5 (49 voted)

Reviews: 88% of readers found this page helpful

Author information

Name: Ray Christiansen

Birthday: 1998-05-04

Address: Apt. 814 34339 Sauer Islands, Hirtheville, GA 02446-8771

Phone: +337636892828

Job: Lead Hospitality Designer

Hobby: Urban exploration, Tai chi, Lockpicking, Fashion, Gunsmithing, Pottery, Geocaching

Introduction: My name is Ray Christiansen, I am a fair, good, cute, gentle, vast, glamorous, excited person who loves writing and wants to share my knowledge and understanding with you.