Uncoupling ecological innovation and speciation in sea snakes (Elapidae, Hydrophiinae, Hydrophiini)

Transcription

1 doi: /j x Uncoupling ecological innovation and speciation in sea snakes (Elapidae, Hydrophiinae, Hydrophiini) K. L. SANDERS*, MUMPUNI &M.S.Y.LEE*à *Darling Building, School of Earth and Environmental Sciences, University of Adelaide, Adelaide, SA, Australia Museum Zoologi Bogor, Puslit Biologi-LIPI, Cibinong, Indonesia àearth Sciences Section, South Australian Museum, North Terrace, Adelaide, SA, Australia Keywords: diversification rate; Hydrophiini; phylogenetic analysis; sea snakes. Abstract The viviparous sea snakes (Hydrophiini) are by far the most successful living marine reptiles, with 60 species that comprise a prominent component of shallow-water marine ecosystems throughout the Indo-West Pacific. Phylogenetically nested within the 100 species of terrestrial Australo-Melanesian elapids (Hydrophiinae), molecular timescales suggest that the Hydrophiini are also very young, perhaps only 8 13 Myr old. Here, we use likelihood-based analyses of combined phylogenetic and taxonomic data for Hydrophiinae to show that the initial invasion of marine habitats was not accompanied by elevated diversification rates. Rather, a dramatic three to six-fold increase in diversification rates occurred at least 3 5 Myr after this transition, in a single nested clade: the Hydrophis group accounts for 80% of species richness in Hydrophiini and 35% of species richness in (terrestrial and marine) Hydrophiinae. Furthermore, other co-distributed lineages of viviparous sea snakes (and marine Laticauda, Acrochordus and homalopsid snakes) are not especially species rich. Invasion of the oceans has not (by itself) accelerated diversification in Hydrophiini; novelties characterizing the Hydrophis group alone must have contributed to its evolutionary and ecological success. Introduction Species richness varies dramatically across taxa and regions, and identifying the organismal and environmental drivers of this disparity is a major goal for evolutionary biology. Patterns of clade species richness have often been assumed to reflect variation in rates of speciation and extinction, which together determine net diversification rate. Studies of dated molecular phylogenies have suggested extreme variation in diversification rates across clades and or over time in groups such as fish (Meyer, 1993; Baldwin & Sanderson, 1998), beetles Correspondence: Kate L. Sanders, Darling Building, School of Earth and Environmental Sciences, University of Adelaide, Adelaide, SA 5005, Australia. Tel.: ; fax: ; kate.sanders@adelaide.edu.au (Barraclough et al., 1998), birds (Grant & Grant, 2002; Lovette et al., 2002; Moyle et al., 2009), amphibians (Wiens et al., 2006), lizards (Rabosky et al., 2007) and plants (Klak et al., 2003; Hughes & Eastwood, 2006). Insights into the underlying drivers of diversification are most likely to be gained from very young groups, where relatively limited environmental histories have been experienced and intrinsic diversification rates are less likely to be dampened because of ecological niche saturation (e.g. Meyer, 1993; Hughes & Eastwood, 2006; Moyle et al., 2009). The viviparous sea snakes (Hydrophiini) are a very recent radiation that comprises 90% of living marine reptiles (Heatwole, 1999; Sanders et al., 2008). Around 60 species in 16 genera occupy a wide range of mostly shallow-water marine habitats throughout the Indo-West Pacific, reaching peak diversity in the 2685

2 2686 K. L. SANDERS ET AL. Indo-Australian biodiversity hotspot (Heatwole, 1999). Phylogenetic studies support the monophyly of viviparous sea snakes and nest them within a viviparous clade of the 100 terrestrial species of endemic Australasian elapids (Keogh et al., 1998; Scanlon & Lee, 2004; Lukoschek & Keogh, 2006; Sanders et al., 2008). The eight species of amphibious sea kraits (Laticauda) are an independent marine lineage that is sister to the terrestrial Australasian elapids plus Hydrophiini; these three groups together form the Hydrophiinae (Keogh et al., 1998; Scanlon & Lee, 2004; Sanders et al., 2008). Hydrophiini forms two well-defined clades on the basis of morphological (Smith, 1926; Voris, 1977; Rasmussen, 1997, 2002) and molecular data (Lukoschek & Keogh, 2006): an Aipysurus lineage of nine, predominantly Australasian reef species in two genera, and a Hydrophis lineage of 47 mostly Southeast Asian species in 11 genera. Additionally, three monotypic genera endemic to Australasian coastal habitats appear to have branched off relatively early along the Hydrophis lineage (McDowell, 1969; Lukoschek & Keogh, 2006). Molecular divergence dating suggests that the Hydrophiinae shared their most recent common ancestor 8 22 (Sanders & Lee, 2008) or (Kelly et al., 2009) million years ago (Ma). The highly diverse viviparous sea snakes (Hydrophiini) are deeply nested within Hydrophiinae, with their common ancestor dated at 5 8 Ma (Sanders et al., 2008). This is consistent with a rapid recent radiation (Lukoschek & Keogh, 2006). The high species diversity of Hydrophiini has sometimes been associated with colonization of a novel habitat (e.g. Dunson, 1975; Lillywhite et al., 2008); however, it is notable that none of the earliest branching lineages of Hydrophiini are especially diverse. Rather, the high species richness in the clade is accounted for by a single nested subgroup, the Hydrophis group (e.g. Minton & da Costa, 1975; Voris, 1977; Lukoschek & Keogh, 2006), which arose perhaps only within the last 3 6 Myr (Sanders et al., 2008). This pattern, and the observation that other marine snake lineages (within Acrochordus, Laticauda and Homalopsidae) are not especially species rich (see Heatwole, 1999), suggests that invasion of the oceans alone did not trigger rapid diversification in snakes. We use likelihood-based, lineage-through-time methods to assess variation in net diversification rates among viviparous sea snakes and their terrestrial relatives. These methods have been increasingly applied and elaborated (see previous references and Ricklefs, 2007) but rely on a close positive relationship between clade age and species richness through time. This assumption will be violated if there are ecological limits on total species diversity of clades, so that disparity in clade size will not accurately reflect variation in speciation and extinction rates. This limitation is rarely acknowledged, or rigorously evaluated (Rabosky, 2009), but is addressed here for Hydrophiinae. Methods Phylogeny and divergence times We used a published, dated tree for Hydrophiinae constructed using Bayesian relaxed clock analysis (BEAST: Drummond & Rambaut, 2006) of 5800 base pairs from seven mitochondrial and nuclear genes for 49 taxa (Sanders et al., 2008); this sampling includes all genera of terrestrial forms and taxa representing the Aipysurus, Hydrophis and intermediate Hydrelaps lineages of sea snakes. This tree had to be dated using secondary calibrations, as robust direct fossil calibrations within Hydrophiinae, or indeed Elapidae, do not yet exist. The inherent uncertainties in secondary calibrations (Graur & Martin, 2004) were incorporated into the analyses by using the 95% posterior intervals (rather than the point estimates) from the original study to inform prior distributions. The resulting posteriors were broadly consistent with one another and the priors, indicating that the divergence dates obtained were not problematic. Relationships within many genera have not been resolved (because of short branches and limited taxon sampling), making it difficult to determine the species diversity of each terminal (clade represented by each sampled species). However, the monophyly of every genus except Neelaps is well supported based on morphological and molecular data (e.g. Hutchinson, 1990; Scanlon & Lee, 2004; Lukoschek & Keogh, 2006; Sanders et al., 2008). Thus, this tree was pruned (using MacClade: Maddison & Maddison, 2005) to yield a genus-level tree of 34 terminals. Because of the likely polyphyly of Neelaps (Sanders et al., 2008), both species were retained as separate terminals. Species numbers were then assigned to each terminal on the pruned tree; these were compiled from the current primary literature (see Fig. 1, and Table S1). Constructing and then pruning a tree from more extensive data is likely to yield more accurate topology and branch lengths (e.g. Zwickl & Hillis, 2002), so we focus on the results obtained from the above-mentioned approach. However, a Bayesian relaxed clock (BEAST) tree constructed from the 34 generic terminals alone is very similar to the pruned tree, in both topology and branch lengths (not shown). All diversification rate analyses used stem ages, i.e. the date when each terminal taxon diverged from its sister taxon. Crown ages for terminal taxa (corresponding to the most recent common ancestor of all extant representatives of a genus or clade) could not be reliably ascertained, either in the pruned tree used here (because of the use of single exemplars of each genus), or in the more complete tree in Sanders et al. (2008). This is because of incomplete taxon sampling and uncertain relationships within many genera: one cannot be certain that the species for which sequence data are available span the basal dichotomy within the genus. We note, however, that using stem rather than crown ages renders

3 Uncoupling ecological innovation and speciation in sea snakes 2687 Fig. 1 BEAST ultrametric tree for Hydrophiinae taken from Sanders et al. (2008) and modified by pruning all but one species for every major lineage. Numbers in brackets indicate the taxonomic diversity of each terminal. Invasion of the marine realm occurred along the stem lineage leading to extant sea snakes, on the branch labelled sea snakes leading from the black dot. The bold branch is the ML estimate of the location of the diversification rate shift under the two-rate model (Lapemis represents the Hydrophis group). Timescale is in millions of years before present. Posterior probabilities of clades are >0.95 unless marked with an asterisk (*). this analysis conservative (with respect to finding elevated rates in the Hydrophis lineage): unpublished molecular and morphological data for viviparous sea snakes suggest that the Hydrophis lineage has a relatively long stem but recent crown divergence compared to most other terrestrial and marine clades of Hydrophiinae (Sanders et al., unpubl. data). Thus, it is likely that using the crown age would greatly reduce the time frame for Hydrophis diversification, leading to even greater inferred speciation rates. Absolute molecular age estimates within Hydrophiinae vary depending on the data and calibrations used (Wüster et al., 2007; Sanders et al., 2008; Kelly et al., 2009). The analyses of shifts in diversification rates described later rely entirely on relative rather than absolute branch lengths (i.e. divergence dates) and are thus robust to uncertainty in age estimates. However, estimates of net diversification rates require dated nodes and were calculated using two alternative sets of absolute age estimates. Constraints on clade diversity Estimates of net diversification rates will be misleading if diversity is independent of time, such as occurs when there are strong ecological constraints regulating the maximum diversity of clades (Rabosky, 2009). Because the diversity data here are composite (phylogenetic plus taxonomic), it was not possible to compare the respective fit of diversity-independent and diversity-dependent diversification models to test for constraints on clade diversity (see Discussion). Instead, we tested for an association between clade age and clade species diversity for all terminal taxa (i.e. do older clades generally have more species?): lack of a positive relationship will confound meaningful estimates of net diversification rate and is a likely indicator of ecological limits on clade diversity through time (Ricklefs et al., 2007; Rabosky, 2009). The regression gradient or coefficient m (parametric) and Spearman s correlation coefficient (nonparametric) were employed. Because aberrant terminal taxa (e.g. a young rapidly diversifying genus, or an ancient depauperate genus) would tend to obscure the agediversity relationship, analyses were performed with four different sets of terminal taxa: the first three excluded likely outliers, and the last included all taxa. Analyses were performed with (1) core Australian terrestrial taxa only (i.e. Oxyuranus-Notechis clade excluding Hydrophiini and early branching (basal) terrestrial genera: Fig. 1); (2) core Australian terrestrial taxa including Hydrophiini (i.e. Oxyuranus-Notechis clade, including Hydrophiini but excluding early branching terrestrial genera); (3) the full

4 2688 K. L. SANDERS ET AL. tree excluding Hydrophiini; and (4) the full tree. Analyses without Hydrophiini (1, 3) were performed because the terrestrial and amphibious (Laticauda) taxa might be differently regulated compared to the fully marine Hydrophiini. Similarly, analyses without the earliest branching lineages (1, 2) were performed because, unlike most core Australian terrestrial taxa and Hydrophiini, basalmost lineages are often low-diversity clades which are either endemic to small islands (e.g. Micropechis, Toxicocalamus), or occupy a restricted niche in a small geographical range (Neelaps calanotus, Neelaps bimaculatus), thus are more likely to have reached ecologically limited diversity levels. Mechanistic tests of these putative limiting factors are required to corroborate the observed age-diversity patterns (see Results). Diversification rates Because of evidence that age and diversity are uncoupled in the earliest branching hydrophiine clades, suggesting they might be diversity-regulated (see Results), diversification analyses were performed on both the full tree, and the Oxyuranus-Notechis clade alone (i.e. with the earliest branching taxa excluded). To test for unexpectedly species-rich or species-poor clades, we used the maximum likelihood estimator of net diversification rate r (speciation minus extinction) that incorporates both the phylogenetic relationships and taxonomic diversity of terminal lineages (Rabosky et al., 2007; Santos et al., 2009). Estimates of r were calculated under two extinction fractions, (a) a pure birth model (no extinction, a = 0) and (b) an arbitrary but high rate of extinction relative to speciation (a = 0.9), using the fitndr_1rate function of the R package LASER (Rabosky, 2006). Clade selection was necessarily somewhat arbitrary, but included most taxonomic and ecological groups. Sea snake clades were the traditionally recognized Aipysurus, Hydrelaps and Hydrophis groups (sensu McDowell, 1972; treated as subfamilies by Kharin (2005)). Although Hydrelaps is monotypic, we did not include this lineage within the Hydrophis clade because it has a distinct (semimarine) ecology and is separated by long molecular branch from the fully marine crown Hydrophis group species (Lukoschek & Keogh, 2006; Sanders et al., unpubl. data). Confidence intervals of 95% for the expected size of clades that have diversified at the net rate observed for the overall radiation under both rates of relative extinction were estimated at time intervals of two million years since the origin of the group using the lambda.stem.ci function in LASER (Rabosky, 2006). We further tested for diversification rate variation by comparing the likelihood of the data under models with (i) equal diversification rates among lineages; and (ii) a two-rate model with a diversification rate shift using LASER. For the latter model, bipartitions are created at every node in turn and the likelihood of the rate for each partition is calculated; the bipartitioning scheme resulting in the highest likelihood is taken as the maximum likelihood estimate of the location of the rate shift (Rabosky et al., 2007). Both analyses were conducted under both extinction fractions (a = 0 and 0.9) using the fitndr_2rate function of LASER (Rabosky, 2006). If the initial marine invasion was accompanied by elevated diversification rates, we would expect a rate shift corresponding to the sea snake stem lineage. Because LASER does not currently implement a three-rate model using composite data, we performed an additional test using MEDUSA in the Geiger library (Alfaro et al., 2009a). This comparative method fits a series of increasingly complex birth death models (including multiple rate shifts) to the combined phylogenetic and taxonomic data using Rabosky et al. s (2007) maximum likelihood estimator. Here, we used an improvement of 4 AIC score units as the cut-off for retaining rate shift models. The above-mentioned analyses depend only on relative (not absolute) branch lengths. We also investigated absolute diversification rates for selected stem clades under minimum and maximum extinction fractions using equation 7 in Magallón & Sanderson (2001). Because of differences in molecular age estimates within Hydrophiinae, diversification rates were calculated using two sets of divergence times: (i) node ages from Sanders et al. (2008), which used two soft-bounded secondary calibrations from Sanders & Lee (2008) (Laticauda vs. Oxyuraninae at Ma and Micropechis vs. other oxyuranines at Ma); and (ii) node ages estimated by re-analysing the Sanders et al. (2008) data matrix using two different secondary calibrations taken from Kelly et al. (2009) (Laticauda vs. oxyuranines at Ma and Notechis vs. Elapognathus at Ma). For the latter analysis, the Bayesian method (BEAST), model parameters, topological constraints, relaxed clock priors and MCMC settings were identical to those used in Sanders et al. (2008). We did not employ the mccrtest.rd function (Rabosky et al., 2007; Santos et al., 2009), which generates full species trees and then prunes them to the match the lineage sampling in the observed tree. This method makes the assumption that lineage sampling is random but, in our case, exemplars were chosen from different genera, potentially increasing terminal (and decreasing internal) branch lengths relative to that expected from a random sample and thus biasing the results. Results When only the core Australian terrestrial taxa (the Oxyuranus-Notechis clade excluding Hydrophiini and early branching terrestrial genera) are considered, there is a positive relationship between terminal clade age and diversity (q = 0.49, P = 0.04; m = 0.53, P = 0.06; Fig. 2). However, when the clade of core Australian terrestrial taxa plus Hydrophiini (the Oxyuranus-Notechis clade) is considered, there is no significant relationship (q = 0.33,

5 Uncoupling ecological innovation and speciation in sea snakes 2689 Fig. 2 The positive relationship between clade species richness and estimated clade age within Hydrophiinae. The regression was calculated for the core Australian terrestrial taxa (black circles) excluding the Hydrophiini (blue/grey circles) (see Analysis 1 in Methods). P = 0.13; m = 0.1, P < 0.92). This is consistent with the interpretation that clade diversity is not strongly limited in the core Australian terrestrial clade, but that the agediversity relationship disappears when the outlying (young but hyper-diverse) Hydrophiini are added. When the full tree of terrestrial and marine Hydrophiinae is considered, there is no significant relationship between clade age and diversity (q = 0.24, P = 0.2; m = 0.32, P = 0.14), and this lack of relationship persists even when the marine Hydrophiini are excluded (q = 0.18, P = 0.29; m = )0.08, P = 0.87). This is consistent with diversity regulation in the earliest branching (basalmost) clades of Hydrophiinae; diversification rate heterogeneity alone is not responsible, because deletion of the Hydrophiini does not restore a significant positive relationship. However, regardless of whether independence of age and diversity in the full hydrophiine tree reflects ecological limits, the strong positive age-diversity relationship in the Notechis-Oxyuranus clade provides confidence that net diversification rates can be meaningfully estimated and compared for sea snakes and their immediate relatives. Comparing observed vs. expected species richness reveals that only two clades are excessively species rich (Fig. 3a, b): the Hydrophiini have higher species richness than expected under both extinction fractions for the full tree, and under zero extinction (but not high extinction) for the Oxyuranus-Notechis clade alone; and the Hydrophis Fig. 3 Extant diversities for marine and terrestrial hydrophiine clades compared with 95% confidence intervals on expected diversity according to stem group age and a constant background diversification rate (r) equal to that of (a) hydrophiines as a whole and (b) the Oxyuranus-Notechis clade alone. Diversification rates were calculated using Rabosky et al. s (2007) combined taxonomic and phylogenetic estimator. The dashed lines are confidence intervals for expected species richness in the absence of extinction [r 0 = for (a) and for (b)], and the solid lines are confidence intervals under a model of high relative extinction [r 0.9 = for (a) and for (b)]. Included are selected hydrophiine genera and monophyletic groups such as the viviparous terrestrial sister clade to sea snakes and the Simoselaps group of burrowing species (Fig. 1 and Table 2).

6 2690 K. L. SANDERS ET AL. Table 1 Model-based analysis of diversification rates in hydrophiine snakes. Analyses of the full hydrophiine tree and the core Oxyuranus-Notechis clade alone reject the constant-rate model in favour of the two-rate model under a = 0 and 0.9 (v 2 13; d.f. = 1; P < 0.001). In both analyses, the branch corresponding to the Hydrophis group (Lapemis) represents the ML estimate of the location of the rate shift. DAIC denotes the difference between the AIC score of a model and the overall best-fitting model. r hy and r refer to diversification rates for the tree partition that does and does not include the Hydrophis group, respectively. All hydrophiines Oxyuranus-Notechis clade Extinction model Constant-rate Two-rate Constant-rate Two-rate log L (DAIC) Net rate log L (DAIC) Net rate log L (DAIC) Net rate log L (DAIC) Net rate a =0 )142.3 (15.4) r = )132.7 (0) r = 0.253; r hy = )89.7 (11.5) r = )81.9 (0) r = 0.269; r hy = a = 0.9 )151.9 (10.2) r = )144.8 (0) r = 0.056; r hy = )96.3 (9.1) r = )89.8 (0) r = 0.055; r hy = Table 2 Stem diversification rates for marine and terrestrial hydrophiine clades estimated using Magallón & Sanderson (2001), equation 7) in the absence of extinction (r 0) and under a high relative extinction rate (r 0.9). Diversification rates are shown for two sets of age estimates: (i) node ages from Sanders & Lee (2008) and Sanders et al. s (2008); and (ii) node ages estimated by calibrating the Sanders et al. (2008) matrix using alternative secondary calibrations from Kelly et al. (2009) (see Methods). Calibration scheme 1 Calibration scheme 2 Clade n t (stem) r 0 r 0.9 t (stem) r 0 r 0.9 Hydrophiinae Hydrophiini Viviparous sister group to sea snakes Hydrophis group Aipysurus group Pseudonaja + Oxyuranus Toxicocalamus + Ogmodon Demansia Laticauda Simoselaps group Rhinoplocephalus group Suta group is excessively species rich under both extinction fractions for the full and reduced trees. The significant results for Hydrophiini appear to be entirely driven by the included Hydrophis group, as the other lineages of Hydrophiini (Aipysurus, Emydocephalus and Hydrelaps) are not excessively species rich. No other clades are unexpectedly species-rich or species-poor (although this test has low power to detect species-poor lineages; see Rabosky et al., 2007). When single- and two-rate models are evaluated using LASER, the data reject the single-rate model in favour of the two-rate model under both extinction fractions (P < 0.001; Table 1). The ML estimate of the diversification rate shift occurs along the branch leading to the Hydrophis group. ML estimates of r for the Hydrophis group compared to that for all other lineages suggest that this clade has diversified at a net rate that is 3 (a =0)to6(a = 0.9) times higher than that observed for the rest of Hydrophiinae, including other marine lineages. The additional test using MEDUSA to fit models with three (or more) rates found no support for a three-rate model over a two-rate model using both the full tree and the Oxyuranus-Notechis clade alone (DAIC 2 and 1.5, respectively). Absolute diversification rates (per lineage per million years) vary depending on the method used and the calibration scheme (Table 2). However, in all analyses, the Hydrophis group has by far the highest rate, estimated at (a = 0.9) to (a = 0) events per lineage per million years under calibration scheme 1 (stem age of 4.9 Ma) and (a = 0.9) to (a = 0) events per lineage per million years under calibration scheme 2 (stem age of 8.7 Ma). Rates for most terrestrial clades, and the Aipysurus, Emydocephalus and Hydrelaps marine lineages, are comparable to the overall rate of (a = 0.9) to (a = 0) under calibration scheme 1, and (a = 0.9) to (a = 0) under calibration scheme 2. The amphibious Laticauda has the lowest rate of all, estimated at (a = 0.9) to (a = 0) under calibration scheme 1 and (a = 0.9) to (a = 0) under calibration scheme 2. Discussion Despite being a very young and diverse group, and the only fully marine squamates to have existed in the last 30 Myr (Lukoschek & Keogh, 2006; Sanders et al., 2008),

7 Uncoupling ecological innovation and speciation in sea snakes 2691 the viviparous sea snakes (Hydrophiini) do not appear to have undergone an initial burst of diversification immediately after entering the marine environment. While it is tempting to associate the proliferation of Hydrophiini with ecological release into marine habitats (e.g. Lillywhite et al., 2008), the observation that most basal branching lineages of Hydrophiini are low-diversity suggests that marine habits alone do not provide a general explanation for high species-richness in Hydrophiini. This is quantitatively demonstrated here. Comparisons of clade diversity show that Hydrophiini as a whole is excessively species rich; however, this is evidently because of diversification within an included subclade, the Hydrophis group (only the Hydrophis group is excessively species rich; Aipysurus, Emydocephalus and Hydrelaps are not). Consistent with this result, a two-rate diversification model best explains the data over one-rate and three-rate models and places the maximum likelihood location for a rate increase on the branch leading to the Hydrophis group at least 3 5 Myr after the shift to the marine realm. This pattern contrasts with a large number of radiations in which speciation rates were apparently highest during the early stages of diversification within a new geographical area or adaptive zone (e.g. Moore & Donoghue, 2007; Rabosky & Lovette, 2008; Phillimore & Price, 2008; Burbrink & Pyron, 2010). However, some temporal lag between the initial invasion of a new realm and successful exploitation of vacant marine niches might be expected, and several recent studies have noted a lack of correspondence between diversification rate shifts and putative key morphological innovations (e.g. Donoghue, 2005; Alfaro et al., 2009a,b; Steeman et al., 2009; Slater et al., 2010). Net diversification analyses can be misleading if clade diversity is regulated by ecological limits to the extent that it becomes uncoupled from variation in speciation and extinction rates. This possibility has not been sufficiently addressed in many studies, but is unlikely to explain the current results. First, the results hold across analyses using all taxa, and using only taxa which show a positive age-diversity relationship. The strong positive relationship between age and species diversity in the terrestrial Notechis-Oxyuranus clade suggests that diversity is not constrained in this group and provides confidence that net diversification rates can be meaningfully estimated and compared for sea snakes and their immediate relatives. If clade comparisons are based only on the Notechis-Oxyuranus clade (and the included Hydrophiini), the results remain qualitatively unchanged. Secondly, Hydrophiinae and the included Hydrophiini are extremely young groups that occupy very broad distributions and a variety of niches (Greer, 1997; Voris, 1972; Voris & Voris, 1983), so are less likely to have reached their diversity limits compared to the much older and or geographically restricted clades considered by Rabosky (2009) (but see Seehausen, 2006). It is also unlikely that the results are a taxonomic artefact caused by underestimation of diversity in the non-hydrophis lineages, and or overestimation in the Hydrophis group. For instance, the relatively accessible and medically important terrestrial hydrophiines (e.g. Skinner et al., 2005) and the primarily Australasian marine Aipysurus (Lukoschek et al., 2007) have received more recent taxonomic attention than the Hydrophis group. Moreover, the most detailed molecular evaluation of the Hydrophis group confirmed that nearly all sampled species are reciprocally monophyletic (Lukoschek & Keogh, 2006), and there are several potential cryptic species contained in geographically widespread and morphologically variable Hydrophis species (Smith, 1926; Rasmussen, 1993; Sanders et al., unpubl. data). It thus seems unlikely that that the results could be explained by the Hydrophis group containing proportionally fewer undescribed species than the terrestrial and Aipysurus hydrophiine lineages. The disparity in species richness between the two sister clades of sea snakes is 9 vs. 48, which is significant under the test of Rabosky et al. (2007). Relationships within the Hydrophis group remain intractible (Voris, 1977; Lukoschek & Keogh, 2006) because of high species diversity, short times since and between speciation events, and likely problems of incomplete lineage sorting. Thus, the important caveat needs to be made that it is not currently possible to partition diversity within the Hydrophis group; if this could be performed, the results might further reveal that increased diversification rates characterize only particular clades within the group and not the group as a whole. The lack of resolution among extremely short branches at the base of the Hydrophis group (Lukoschek & Keogh, 2006; Sanders et al., unpubl. data) is consistent with the interpretation that increased diversification rates characterize the entire Hydrophis group, rather than some part of it, but this needs to be tested once Hydrophis group phylogeny is better resolved. Given that diversification rates have increased greatly in only a single clade of marine snakes, it is difficult to statistically test potential underlying drivers (e.g. Lynch, 2009). However, that several Hydrophiini lineages have diversified in broadly the same ecological context and region, yet only Hydrophis has speciated explosively, suggests that traits specific to the Hydrophis lineage are likely to be implicated. Possible factors are greater adaptive phenotypic plasticity, and or novel innovations, that have enabled the Hydrophis group to respond rapidly to the available ecological opportunities for diversification. Hydrophis group species exhibit high degrees of trophic specialization and sympatry. Assemblages can comprise up to 11 species and are dominated by dietary specialists that display striking variation in head shape and head body proportions suggestive of ecomorphological divergence (Glodek & Voris, 1982; Voris & Voris, 1983; Heatwole, 1999). Particularly

8 2692 K. L. SANDERS ET AL. notable are numerous, varyingly microcephalic specialists on burrowing eels and goby-like fishes prey resources largely unexploited by the mostly generalist Aipysurus fish eaters and two egg-eating Emydocephalus (Glodek & Voris, 1982; Voris & Voris, 1983). Hydrophis group species also tend to have substantially larger geographical ranges (compared with most Aipysurus- Emydocephalus species and all intermediate lineages (Hydrelaps, Ephalophis and Parahydrophis): see Heatwole, 1999), which might reduce extinction and provide greater opportunity for isolation by distance and or biogeographic barriers to promote ecological divergence (e.g. Rosenzweig, 1995; Seehausen, 2006). Young age, marked differentiation in resource use and tropic morphology, and high levels of sympatry are traits that are often associated with rapid adaptive radiations (Schluter, 2000). However, while ecological differentiation is expected to elevate intrinsic speciation rates, it would also increase the limits to total species diversity ( carrying capacity ) of clades over longer timescales. Additional analyses using species-level phylogenies are needed to distinguish the roles of these and other lineage-specific factors in generating the exceptional diversity of Hydrophis group sea snakes. Acknowledgments We are grateful to eresearchsa for access to supercomputer resources, and Andrew F. Hugall and Dan Rabosky for advice and discussion. This work is supported by an Australian Research Council grant to KL Sanders and MSY Lee. References Alfaro, M.E., Santini, F., Brock, C., Alamillo, H., Dornburg, A., Rabosky, D.L., Carnevale, G. & Harmon, L.J. 2009a. Nine exceptional radiations plus high turnover explain species diversity in jawed vertebrates. Proc. Nat. Acad. Sci. USA 106: Alfaro, M.E., Brock, C.D., Banbury, B.L. & Wainwright, P.C. 2009b. Does evolutionary innovation in pharyngeal jaws lead to rapid lineage diversification in labrid fishes? BMC Evol. Biol. 9: 255. Baldwin, B.G. & Sanderson, M.J Age and rate of diversification of the Hawaiian silversword alliance (Compositae). Proc. Natl Acad. Sci. USA 95: Barraclough, T.G., Vogler, A.P. & Harvey, P.H Revealing the factors that promote speciation. Philos. Trans. R. Soc. Lond. B Biol. Sci. 353: Burbrink, F.T. & Pyron, R.A How does ecological opportunity influence rates of speciation, extinction and morphological diversification in New World ratsnakes (tribe Lampropeltini)? Evolution 64-4: Donoghue, M.J Key innovations, convergence, and success: macroevolutionary lessons from plant phylogeny. Paleobiology 31: Drummond, A.J. & Rambaut, A BEAST v1.4. Available from URL Dunson, W.A Adaptions of sea snakes. In: The Biology of Sea Snakes (W. Dunson, ed.), pp University Park Press, Baltimore. Glodek, G.S. & Voris, H.K Marine snake diets: prey composition, diversity and overlap. Copeia 1982: Grant, P.R. & Grant, B.R Adaptive radiation of Darwin s finches. Am. Sci. 90: Graur, D. & Martin, W Reading the entrails of chickens: molecular timescales of evolution and the illusion of precision. Trends Genet. 20: Greer, A The Biology and Evolution of Australian Snakes. Surrey Beatty & Sons, Sydney. Heatwole, H Sea Snakes. University of New South Wales Press, Sydney. Hughes, C. & Eastwood, R Island radiation on a continental scale: exceptional rates of plant diversification after uplift of the Andes. Proc. Natl Acad. Sci. USA 103: Hutchinson, M.N The generic classification of the Australian terrestrial elapid snakes. Mem. Queensl. Mus. 28: Kelly, C.M.R., Barker, N.P., Villet, M.H. & Broadley, D.G Phylogeny, biogeography and classification of the snake superfamily Elapoidea: a rapid radiation in the late Eocene. Cladistics 25: Keogh, J.S., Shine, R. & Donnellan, S Phylogenetic relationships of terrestrial Australo-Papuan elapid snakes (Subfamily Hydrophiinae) based on cytochrome b and 16S rrna sequences. Mol. Phylogenet. Evol. 10: Kharin, V.E A check-list of sea snakes (Serpentes: Laticaudidae, Hydrophiidae) of the World oceans. Izv. TINRO 140: [in Russian]. Klak, C., Reeves, G. & Hedderson, T Unmatched tempo of evolution in Southern African semi-desert ice plants. Nature 427: Lillywhite, H.B., Sheey, C.M. & Zaidan, F Pitviper scavenging at the intertidal zone: an evolutionary scenario for the invasion of the sea. Bioscience 58: Lovette, I.J., Bermingham, E. & Ricklefs, R.E Cladespecific morphological diversification and adaptive radiation in Hawaiian songbirds. Proc. Biol. Sci. 269: Lukoschek, V. & Keogh, J.S Molecular phylogeny of sea snakes reveals a rapidly diverged adaptive radiation. Biol. J. Linn. Soc. 89: Lukoschek, V., Waycott, M. & Marsh, H Phylogeography of the olive sea snake, Aipysurus laevis (Hydrophiinae) indicates Pleistocene range expression around northern Australia but low contemporary gene flow. Mol. Ecol. 16: Lynch, V.J Live-birth in vipers (Viperidae) is a key innovation and adaptation to global cooling during the Cenozoic. Evolution 63: Maddison, D.R. & Maddison, W.P MacClade. Sinauer Associates, Sunderland. Magallón, S. & Sanderson, M.J Absolute diversification rates in angiosperm clades. Evolution 55: McDowell, S.B Notes on the Australian sea-snake Ephalophis greyi M. Smith (Serpentes: Elapidae, Hydrophiinae) and the origin and classification of sea snakes. Zool. J. Linn. Soc. 48: McDowell, S.B The genera of sea snakes of the Hydrophis group (Serpentes, Elapidae). Trans. Zool. Soc. London 32:

9 Uncoupling ecological innovation and speciation in sea snakes 2693 Meyer, A Phylogenetic relationships and evolutionary processes in East African cichlid fishes. Trends Ecol. Evol. 8: Minton, S.A. & da Costa, M.S Serological relationships of sea snakes and their evolutionary implication. In: The Biology of Sea Snakes (W. Dunson, ed.), pp University Park Press, Baltimore. Moore, B.R. & Donoghue, M.J Correlates of diversification in the plant clade dipsacales: geographic movement and evolutionary innovations. Am. Nat. 170: S28 S55. Moyle, R.G., Filardi, C.E., Smith, C.E. & Diamond, J Explosive Pleistocene diversification and hemispheric expansion of a great speciator. Proc. Natl Acad. Sci. USA 106: Phillimore, A.B. & Price, T.D Density dependent cladogenesis in birds. PLoS Biol. 6: e71 (DOI: /journal.pbio ). Rabosky, D.L LASER: a maximum likelihood toolkit for detecting temporal shifts in diversification rates. Evol. Bioinform. Online 2: Rabosky, D.L Ecological limits and diversification rate: alternative paradigms to explain the variation in species richness among clades and regions. Ecol. Lett. 12: Rabosky, D.L. & Lovette, I.J Density dependent diversification in North American wood warblers. Proc. Biol. Sci. 275: Rabosky, D.L., Donnellan, S.C., Talaba, A.L. & Lovette, I.J Exceptional among-lineage variation in diversification rates during the radiation of Australia s most diverse vertebrate clade. Proc. Biol. Sci. 274: Rasmussen, A.R The Status of the Persian Gulf Sea Snake Hydrophis lapemoides (Gray, 1849) (Serpentes, Hydrophiidae), Bull. Nat. Hist. Mus. London (Zool.) 59: Rasmussen, A.R Systematics of the sea snakes: a critical review. Symp. Zool. Soc. London 70: Rasmussen, A.R Phylogenetic analysis of the true aquatic elapid snakes Hydrophiinae (sensu Smith et al., 1977) indicates two independent radiations into water. Steenstrupia 27: Ricklefs, R.E Estimating diversification rates from phylogenetic information. Trends Ecol. Evol. 22: Ricklefs, R.E., Losos, J.B. & Townsend, T.M Evolutionary diversification of clades of squamate reptiles. J. Evol. Biol. 20: Rosenzweig, M.L Species Diversity in Space and Time. Cambridge University Press, Cambridge. Sanders, K.L. & Lee, M.S.Y Molecular evidence for a rapid late-miocene radiation of Australasian venomous snakes. Mol. Phylogenet. Evol. 46: Sanders, K.L., Lee, M.S.Y., Leys, R., Foster, R. & Keogh, J.S Molecular phylogeny and divergence dates for Australasian elapids and sea snakes (Hydrophiinae): evidence from seven genes for rapid evolutionary radiations. J. Evol. Biol. 21: Santos, J.C., Coloma, L.A., Summers, K., Caldwell, J.P., Richard Ree, R. & Cannatella, D.C Amazonian amphibian diversity is primarily derived from late Miocene Andean lineages. PLoS Biol. 7: e Scanlon, J.D. & Lee, M.S.Y Phylogeny of Australasian venomous snakes (Colubroidea, Elapidae, Hydrophiinae) based on phenotypic and molecular evidence. Zool. Scr. 33: Schluter, D The ecology of Adaptive Radiation. Oxford University Press, Oxford. Seehausen, O African cichlid fish: a model system in adaptive radiation research. Proc. Biol. Sci. 273: Skinner, A., Donnellan, S.C., Hutchinson, M.N. & Hutchinson, R.G A phylogenetic analysis of Pseudonaja (Hydrophiinae, Elapidae, Serpentes) based on mitochondrial DNA sequences. Mol. Phylogenet. Evol. 37: Slater, G.J., Price, S.A., Santini, F. & Alfaro, M.E Diversity versus disparity and the radiation of modern cetaceans. Proc. Biol. Sci. 277: Smith, M Monograph of the sea snakes (Hydrophiidae). Taylor & Francis, London. Steeman, M.E., Hebsgaard, M.B., Fordyce, R.E., Ho, S.Y.W., Rabosky, D.L., Nielsen, R., Rahbek, C., Glenner, H., Martin, V., Sørensen, M.V. & Willerslev, E Radiation of Extant Cetaceans Driven by Restructuring of the Oceans. Syst. Biol. 58: Voris, H.K The role of sea snakes (Hydrophiidae) in the trophic structure of coastal ocean communities. J. Mar. Biol. Assoc. India 14: Voris, H.K A phylogeny of the sea snakes (Hydrophiidae). Fieldiana, Zool. 70: Voris, H.K. & Voris, H.H Feeding strategies in marine snakes: an analysis of evolutionary, morphological, behavioural and ecological relationships. Am. Zool. 23: Wiens, J.J., Engstrom, T.N. & Chippendale, P.T Rapid diversification, incomplete isolation, and the speciation clock in North American salamanders (genus Plethodon): testing the hybrid swarm hypothesis of rapid radiation. Evolution 60: Wüster, W., Crookes, S., Ineich, I., Mane, Y., Pook, C.E., Trape, J.-F. & Broadley, D.G The phylogeny of cobras inferred from mitochondrial DNA sequences: evolution of venom spitting and the phylogeography of the African spitting cobras (Serpentes: Elapidae: Naja nigricollis complex). Mol. Phylogenet. Evol. 45: Zwickl, D.J. & Hillis, D.M Increased taxon sampling greatly reduces phylogenetic error. Syst. Biol. 51: Supporting information Additional Supporting Information may be found in the online version of this article: Table S1 Major genera and species groups within Hydrophiinae. As a service to our authors and readers, this journal provides supporting information supplied by the authors. Such materials are peer-reviewed and may be re-organized for online delivery, but are not copy-edited or typeset. Technical support issues arising from supporting information (other than missing files) should be addressed to the authors. Received 1 March 2010; revised 31 August 2010; accepted 3 September 2010