RES EARCH

EVOLUTION Approach
We constructed a genomic backbone phylogeny
The macroevolutionary singularity of snakes for 1018 species of squamates, sequencing an
average of 4.4 Mb across 4945 loci per taxon
Pascal O. Title1,2,3†, Sonal Singhal3,4†, Michael C. Grundler3, Gabriel C. Costa3,5, R. Alexander Pyron6,7, (figs. S1 and S2 and data S1). This phylogeny
Timothy J. Colston7,8, Maggie R. Grundler3,9,10, Ivan Prates3, Natasha Stepanova3, Marc E. H. Jones11,12,13, was then used as a scaffold upon which we
Lucas B. Q. Cavalcanti14, Guarino R. Colli15, Nicolas Di-Poï16, Stephen C. Donnellan17, Craig Moritz18, added additional species from GenBank, yield-
Daniel O. Mesquita14, Eric R. Pianka19‡, Stephen A. Smith20, Laurie J. Vitt21, Daniel L. Rabosky3*† ing a species-level, dated phylogeny containing
6885 of the 10,759 squamate taxa (fig. S3 and
Snakes and lizards (Squamata) represent a third of terrestrial vertebrates and exhibit spectacular data S2). We then time-calibrated a phyloge-
innovations in locomotion, feeding, and sensory processing. However, the evolutionary drivers of this nomic tree subsampled to 134 tips that spanned
radiation remain poorly known. We infer potential causes and ultimate consequences of squamate 31 fossil calibrations (figs. S4 and S5 and data S1)
macroevolution by combining individual-based natural history observations (>60,000 animals) with and used the resulting node dates as secondary
a comprehensive time-calibrated phylogeny that we anchored with genomic data (5400 loci) from calibrations for our full tree (fig. S6).
1018 species. Due to shifts in the dynamics of speciation and phenotypic evolution, snakes have With this tree, we quantified macroevolution-
transformed the trophic structure of animal communities through the recurrent origin and diversification ary dynamics across a range of ecological, mor-
of specialized predatory strategies. Squamate biodiversity reflects a legacy of singular events phological, and environmental traits. Traits were

Downloaded from https://www.science.org at Museu Paraense Emlio Goeldi on February 22, 2024
that occurred during the early history of snakes and reveals the impact of historical contingency collated from a variety of sources and augmented
on vertebrate biodiversity. with primary natural history data from our field-
and museum-based research programs (17), in-

F
cluding a dietary dataset (n = 68,547 records)
rom a phylogenetic perspective, snakes are spect to ecology, morphology, and biogeography from preserved stomach contents and field ob-
indisputably nested within lizards (1, 2) (8–10). These observations suggest that evolu- servations that spanned 1314 species of snakes
and are simply one example of a particu- tionary dynamics in snakes are qualitatively and lizards. For each species, we then computed
larly species-rich and cosmopolitan group different from those in lizards. If so, the origin a simple statistical index of net innovation, Y,
of “scaled reptiles” (Squamata). Yet, unlike of snakes is potentially consistent with a Simp- defined as the absolute difference in phenotype
lizards, snakes engage with human emotions in sonian view of macroevolution (11, 12), whereby between the focal species and the inferred an-
a visceral manner unmatched by almost any major biodiversity expansions occur through cestral state for all extant squamates (17). We cal-
other group of organisms and for this reason qualitative phase shifts into new adaptive zones. culated indices for net trophic innovation (Ydiet),
have played important cultural roles in human These phase shifts can be conceptualized as morphological innovation in skull shape (Yskull),
societies (3). Most readily known for their lack macroevolutionary “singularities”: patterns of presacral vertebral count (Yvert), body elongation
of limbs and unique prey-capture strategies, rapid change across multiple organismic and (Yelong), and chemosensory processing (Ychem).
snakes exhibit an incredible degree of ecomor- ecological axes that, when viewed retrospec- For the same traits, we also computed an index
phological diversity and specialization. The tively through the prism of geological time, of absolute, branch-specific change along indi-
~4000 extant snake species include shovel- are sufficiently clustered together so as to seem vidual branches (fig. S7).
snouted burrowers that hunt desert scorpions, virtually instantaneous. The term “singular” To estimate evolutionary rates, we developed
slender arboreal predators that prey on tree also refers to the fact that these transforma- a tip-rate metric for univariate and multivariate
snails, and paddle-tailed marine forms that tions typically appear unpredictable from prior phenotypic data that is similar to the widely
probe reef crevices for fish eggs and eels. How- character states and phylogenetic position alone; used “diversification rate” (DR) statistic for spe-
ever, more than 25 clades of lizards have in- such transitions have been documented within ciation rate (18); we refer to this metric as the
dependently evolved limblessness (4), and birds (13, 14), mammals (15), and other taxa “TR” statistic (17). We then used a variance-
other lineages also evolved dietary specializa- (16). Here, we characterize the tempo and mode partitioning analysis to identify nodes that best
tion, venom, highly mobile skulls, and/or ad- of ecological and morphological innovation account for species-level phenotypic variation
vanced chemoreception (5–7)—all attributes across squamate reptiles to address the role [canonical phylogenetic ordination (CPO) (19)].
typically associated with snakes. This conver- of macroevolutionary singularities in generat- We also characterized heterogeneity in phe-
gence raises fundamental questions about how ing large-scale patterns of lizard and snake notypic macroevolution with multiprocess
and why particular traits have influenced squa- biodiversity. In particular, we test the extent evolutionary models that allowed for shifts in
mate diversification more generally. to which phenotypic shifts have predictable both (trait) state and evolutionary rates, in
Viewed across multiple traits, snakes none- consequences for evolutionary diversification which “state shifts” are defined as large jumps
theless appear distinct from lizards with re- across squamates. in phenotype on individual branches (17).

1
Department of Ecology and Evolution, Stony Brook University, Stony Brook, NY 11794, USA. 2Environmental Resilience Institute, Indiana University, Bloomington, IN 47408, USA. 3Museum of
Zoology and Department of Ecology and Evolutionary Biology, University of Michigan, Ann Arbor, MI 48109, USA. 4Department of Biology, California State University, Dominguez Hills, Carson, CA
90747, USA. 5Department of Biology and Environmental Sciences, Auburn University at Montgomery, Montgomery, AL 36117, USA. 6Department of Biological Sciences, The George Washington
University, Washington, DC 20052, USA. 7Department of Vertebrate Zoology, National Museum of Natural History, Smithsonian Institution, Washington, DC, 20560, USA. 8Biology Department,
University of Puerto Rico at Mayagüez, Mayagüez 00680, Puerto Rico. 9Department of Environmental Science, Policy, and Management, University of California, Berkeley, Berkeley, CA 94720,
USA. 10Museum of Vertebrate Zoology, University of California, Berkeley, Berkeley, CA 94720, USA. 11Science Group: Fossil Reptiles, Amphibians and Birds Section, Natural History Museum,
London SW7 5BD, UK. 12Research Department of Cell and Developmental Biology, University College London, London WC1E 6BT, UK. 13Biological Sciences, University of Adelaide, Adelaide, SA
5005, Australia. 14Departamento de Sistemática e Ecologia, Universidade Federal da Paraíba, João Pessoa, Paraíba 58051-900, Brazil. 15Departamento de Zoologia, Universidade de Brasília,
Brasília, Distrito Federal 70910-900, Brazil. 16Institute of Biotechnology, Helsinki Institute of Life Science, University of Helsinki, 00014 Helsinki, Finland. 17South Australian Museum, North
Terrace, Adelaide, SA 5000, Australia. 18Research School of Biology, The Australian National University, Canberra, ACT 2600, Australia. 19Department of Integrative Biology, The University of
Texas at Austin, Austin, TX 78712, USA. 20Department of Ecology and Evolutionary Biology, University of Michigan, Ann Arbor, MI 48109, USA. 21Sam Noble Museum and Department of Biology,
University of Oklahoma, Norman, OK, USA.
*Corresponding author. Email: [email protected]
†These authors contributed equally to this work.
‡Deceased

Title et al., Science 383, 918–923 (2024) 23 February 2024 1 of 6

RES EARCH | R E S E A R C H A R T I C L E

Limb-reduced lizards Snakes

Value
Higher
A B C D E F G H K
Median
J
Lower L
Net innovation ( )

diet

elong

Downloaded from https://www.science.org at Museu Paraense Emlio Goeldi on February 22, 2024
skull

TR diet
Rates (TR)

TR elong

TR skull
Million years ago

50 F L
E A Dibamidae G Anguiformes
K
100 J B Gekkota H Iguania
B D
C Scincoidea I Snakes
150 G
D Teioidea J Scolecophidia
H
A C E Amphisbaenia K Alethinophidia
200 F Lacertoidea L Colubriformes

Fig. 1. Phenotypic innovation and evolutionary rates across squamates. rates of evolution for innovation in skull shape (Yskull), elongation (Yelong),
Net innovation (Y) and evolutionary rate (TR) are shown across 246 squamate and diet composition (Ydiet); these results are even more striking for colubriform
clades (bottom), subsampled from our full time-calibrated phylogeny of snakes (node L). Tip values are residuals from tree-wide median value;
6885 species (fig. S3). Snakes show greater phenotypic innovation and faster additional traits are shown in fig. S8.

Results Rates for morphological and ecological traits lution: The rate of trophic niche evolution was
Phenotypic innovation across snakes and lizards generally increased early in the evolutionary 3.2-fold faster in snakes than in lizards, and snake
Snakes have undergone large transformations history of snakes (Fig. 1). Snakes occupy distinct diets have diverged much more from the ances-
along multiple phenotypic axes (Figs. 1 and 2 morphological space and exhibit elevated rates tral squamate diet relative to lizards (Fig. 1).
and figs. S7 to S10) relative to other squamates. of morphological evolution relative to lizards, Across a range of traits and rates, both CPO
These shifts distinguish them both quantita- as evidenced by their greater body elongation and multiprocess models reveal that much of the
tively and qualitatively from the ancestral squa- (mean snake elongation index and rate >6× phenotypic variation across snakes and lizards
mate phenotype and from their closest living and >24× that of lizards; fig. S11), higher ver- can be explained by a single event that oc-
(nonsnake) relatives. Estimates of net innova- tebral counts (mean count and rate for snakes curred early in the evolutionary history of snakes
tion range from 1.6- to 3-fold higher in snakes >5× and >80× that of lizards; figs. S11 and S12), (Fig. 2 and figs. S14 and S15). Multiprocess
relative to lizards (Fig. 1 and fig. S8). These re- highly distinctive skull shape [mean snake rate phenotypic models generally recovered a phase
sults suggest that the origin or early diversification >3× that of lizards; fig. S11 (9)], and greater rate shift in both rate and state associated with
of snakes was associated with massive shifts in of body mass evolution (snake rate >12× that the ancestor of extant snakes or an associated
traits associated with feeding, locomotion, and of lizards; fig. S13). Patterns of morphological early divergence within snakes (Fig. 2 and
sensory processing. evolution are mirrored by ecological trait evo- figs. S16 and S17), implying both decoupled

Title et al., Science 383, 918–923 (2024) 23 February 2024 2 of 6

RES EARCH | R E S E A R C H A R T I C L E

Diet Elongation Skull Mass Climate

Clade
Snakes
Alethinophidia
Model fitting
1.0 1.0 1.0 1.0 1.0
Variance explained

Canonical phylogenetic Other

0.8 ordination 0.8 0.8 0.8 0.8
(cumulative)

Relative support
0.6 0.6 0.6 0.6 0.6 Higher
0.4 0.4 0.4 0.4 0.4 Lower
0.2 0.2 0.2 0.2 0.2
0.0 0.0 0.0 0.0 0.0
1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5
Rank order Rank order Rank order Rank order Rank order

Fig. 2. Phase shifts associated with the early evolution of snakes explain explained by discrete phase shifts on single branches, as inferred from two
trait variation across extant squamates. (Top row) Multiprocess models distinct unsupervised learning methods. The total variance explained by five
generally find strong support for major phase shifts in both trait state and rates largest shift nodes (top row) is shown as triangles; results from stepwise
associated with all snakes (red) or alethinophidian snakes (orange; 88% of snake canonical phylogenetic ordination are shown as squares. Individual node effects

Downloaded from https://www.science.org at Museu Paraense Emlio Goeldi on February 22, 2024
diversity). (Bottom row) Cumulative trait variance across extant squamates account for much less variance in climate (far right) relative to other traits.

rate dynamics in snakes relative to other squa- Previous analyses of trophic macroevolution genetic data were not available [fig. S23, (17)].
mates and also a major shift (jump) in pheno- across lizards (7) have formulated the “deep- Snakes show substantially elevated rates of
typic state. These findings are consistent with history” hypothesis of squamate diets, in which speciation relative to lizards [mean lizard
those of several recent studies that demon- resource use is deeply conserved, in contrast lCLaDS = 0.08, lBAMM = 0.09; snake lCLaDS = 0.18,
strated phase shifts in the tempo and mode to dynamic patterns of trophic niche evolution lBAMM = 0.21 (CLaDS, cladogenetic diversifica-
of cranial evolution associated with the origin observed in some vertebrate radiations (21–23). tion rate shift model; BAMM, Bayesian analysis
and early radiation of snakes (8, 20). Our diet analyses recovered elements of this of macroevolutionary mixtures)] (Fig. 4 and figs.
hypothesis (7, 24) (fig. S20), but this legacy S24 and S25D), including other well-studied
Trophic structure of squamate biodiversity effect is dwarfed by the trophic shifts and ex- lizard clades thought to be the result of rapid
Snakes occupy much greater dietary space rel- pansions that have occurred in snakes throughout radiation (e.g., Anolis and sphenomorphine
ative to lizards (Fig. 3A), and their diets are the Cenozoic (25). The origin of alethinophidian skinks) (Fig. 4, C and D). Approximately 45%
generally nonoverlapping owing to the dispro- snakes (all snakes but blindsnakes; 88% of snake of the variance in tip-level speciation rates
portionate consumption of vertebrate prey by species; Fig. 2) accounts for nearly 40% of the across squamates is explained by the node
snakes (Fig. 3B and figs. S18 and S19). Snakes variance in species-level diet state across squa- spanning colubriform snakes (lCLaDS = 0.20;
also use an aquatic and semiaquatic prey base mates (fig. S14G). This shift, accompanied by lBAMM = 0.23; figs. S14 and S26). Outside of
that is largely untouched by lizards (separation an overall acceleration in the rate of trophic snakes, the most important rate shift involved
along diet PC 2, Fig. 3, A and B, and fig. S18) niche evolution in colubriform snakes (Fig. 1, the ancestor of the species-rich Liolaemus lizard
(PC, principal component). Although snakes as TRdiet), led to a dramatic expansion of the clade from montane and arid South America
a whole use a greater diversity of food resources squamate trophic universe. Overall, our results (~260 species), which accounts for ~11% of
than lizards, the diets of individual snake spe- highlight the power of primary natural history the variance in tip-level rates. For higher taxa,
cies are consistently narrower than those of data obtained from museum specimens and speciation rate is correlated with species rich-
lizards (Fig. 3D). Snakes have evolved a range field observations to reveal major shifts in the ness (n = 119 subfamilies; Spearman’s r =
of phenotypes to specialize on viscous prey ecological structure of the biota. 0.63) (Fig. 4C), but this positive correlation is
(worms, slugs), slippery prey (eggs, eels, and driven by the presence of very small clades
caecilians), and noxious and dangerous prey Climate and life-history evolution with low speciation rates. For clades with at
(e.g., venomous snakes, mammals, centipedes, Climatic niche and life history traits do not least 50 species—collectively accounting for
and scorpions), as well as snails and other appear to be evolving along distinct evolution- 96% of global squamate diversity—there is
“armored” resources (Fig. 3, A and B). ary trajectories in snakes relative to lizards, thus effectively no relationship between richness
A small number of lizards (e.g., some moni- departing from patterns observed for morphol- and rate (n = 42, r = 0.03). Similar decoupling
tor lizards, Heloderma, and Lialis) feed heavily ogy and diet (Fig. 2). We find high lability of a between speciation rates and diversity patterns,
on vertebrates, mainly exploiting prey already key life-history trait (parity mode), and in cli- both for taxonomic groups and geographic re-
used by snakes. By contrast, finer partitioning matic niche parameters (figs. S8 and S9). Despite gions, has now been observed in a wide range
of dietary categories among arthropod-feeding some phylogenetic signal in climatic niche param- of taxa (18, 26, 27).
squamates (Fig. 3C) reveals that snakes and eters (figs. S14 and S15), there is little evidence that We then tested whether previously proposed
lizards consume largely nonoverlapping re- such conservatism persists across large phylo- key innovations predict speciation rates across
sources, with blindsnakes feeding almost ex- genetic scales (figs. S21 and S22). Variation in squamates more generally. We found that indi-
clusively on insect brood—larval and pupal ants climatic niche across squamates thus cannot vidual traits had minimal ability to explain
and termites in particular—and with several be explained by deep divergences in their evo- speciation-rate variation (Fig. 4F and fig. S25E),
other snake lineages specializing on spiders, lutionary past (Fig. 2 and figs. S14L and S15R). even for those traits that have evolved multiple
centipedes, and scorpions. Conversely, arthropod- times across squamates [maximum coefficient
feeding lizards consume far more adult insects Speciation rates in lizards and snakes of determination (R2) = 0.01]. To the extent that
spanning multiple taxonomic categories, and We quantified speciation rates across squamates, these traits contributed to snake diversification,
lizard niche breadths are broader than those using phylogenetic imputation to infer place- their influence is statistically confounded with
of arthropod-feeding snakes (Fig. 3C). ments for the 3872 squamate species for which a macroevolutionary singularity associated with

Title et al., Science 383, 918–923 (2024) 23 February 2024 3 of 6

RES EARCH | R E S E A R C H A R T I C L E

A B Prey items
Anolis
punctatus Plants
Thecadactylus Varanus Crotalus Lizards
More
terrestrial prey

5 rapicauda gouldii cerastes

Snakes Insects
Arachnids
& centipedes
Decapods
Worms &
mollusks
Fishes
0
Storeria
Amphibians
PC 2 (14.2%)

occipitomaculata

Thamnophis Squamates
radix
Mammals &
birds
Lizards
Snakes

Downloaded from https://www.science.org at Museu Paraense Emlio Goeldi on February 22, 2024
5
Species'
dietary
breadth Enyalioides
palpebralis
aquatic prey

3
Nerodia
Snakes
6
9 fasciata
More

12
Lizards
10 5 0 5
More More
invertebrate prey PC 1 (53.5%) vertebrate prey

C Snake taxa within typical lizard dietary space (n = 72) D All squamate taxa in the diet dataset (n = 1314)
Adult insects
Arachnids
Amerotyphlops Centipedes
reticulatus Lizards
Insect larvae, pupae, eggs
Density
40 20 0 20
Excess prey utilization relative to lizards (%)
Sonora
semiannulata
ta Snakes Data points
Lizards

Tantilla Snakes
nigriceps
0 5 10 15 2 4 6 8 10 12
Species' dietary breadth Species' dietary breadth

Fig. 3. Fundamental ecological differences between snakes and lizards. number of species assigned to the group. (C) Apparent dietary overlap between
(A) Multidimensional trophic niche as inferred from 68,547 individual squamates some snakes and lizards (A) masks differences in the ways that these groups
(stomach contents, field observations). Individual points correspond to species use invertebrate prey. Arthropod-feeding snakes tend to eat spiders, centipedes,
(n = 1314); point size denotes dietary-niche breadth. (B) Similarity path through and scorpions (e.g., Sonora semiannulata and Tantilla nigriceps), or soft-bodied
diet space as inferred by principal graph analysis, illustrating taxonomic structure larval insects (e.g., Amerotyphlops reticulatus), and they also consume far fewer
of resource differences between snakes and lizards. Nodes (circles) are adult insects relative to lizards. (D) Dietary-niche breadths are much greater
groupings of species that consume similar resources; colors represent fractional for lizards than for snakes, even after accounting for intraspecific sampling
resource use of each inferred node, and node sizes are proportional to the variability.

the origin of snakes (Fig. 4E and fig. S27) (28). dynamics that occurred in association with the (20). However, our results highlight the chal-
We also considered several biogeographic pre- origin of snakes and with their most successful lenge of linking specific innovations to the
dictors of speciation rate, finding no evidence constituent clade (Colubriformes, 3100 species) overall “success” of the snake phenotype. Rep-
for associations with latitude (fig. S28) or bio- (Fig. 1), a radiation that shows no sign of slowing licated innovations in our dataset, including
geographic region (fig. S29). Similar to other ver- toward the present (25). Numerous hypotheses chemosensory innovation, limb loss, and body
tebrate clades (18, 26), the latitudinal gradient in have been proposed to account for the dramatic elongation, show no predictable effects on
squamate diversity appears largely decoupled evolutionary radiation of snakes, including shifts speciation rate (Fig. 4). Moreover, counter-
from speciation rates, and geographic variability in feeding mechanisms, extreme cranial kinesis, factual experiments have occurred throughout
in speciation rate is largely a function of the rela- body elongation, limb reduction, and venom squamate evolutionary history that support a
tive fraction of snakes versus lizards in the biota. (8, 25, 29, 30). Furthermore, snakes show sev- decoupling between innovation per se and
eral genomic attributes, such as transposable macroevolutionary regime shifts. For example,
Conclusions element proliferation (31) and metabolic protein lizard (nonsnake) lineages that have evolved
Global squamate biodiversity has been heavily redesign (32), that might contribute to evolu- potent venom (Heloderma) have scarcely diver-
shaped by a singularity in macroevolutionary tionary versatility of the general snake body plan sified, and perhaps the most extreme instance

Title et al., Science 383, 918–923 (2024) 23 February 2024 4 of 6

RES EARCH | R E S E A R C H A R T I C L E

A Lizard clades including

Snakes limb-reduced taxa

Colubriform snakes

Dipsadines Colubrines
"True" geckos Australian Anoles Liolaemines Non- NW Cobras,
sphenomorphines colubriforms pit vipers sea snakes Speciation
& allies rate
Higher
C D E F
Lizards Non-colubriforms Liolaemus
Colubriform snakes melanops Median
Hydrophis
Hydro Typical Limb-reduced Snakes
0.3 1.00 0.05

Speciation rate (log CLaDS)

major
m lizards lizards
Speciation rate (CLaDS)

Speciation rate (CLaDS)

Cobras, sea snakes & allies 0.6 Lower
New World pit vipers 0.04

Phylogenetic R 2
Dipsadines 0.30
0.2 Liolaemines
0.4 0.03

Downloaded from https://www.science.org at Museu Paraense Emlio Goeldi on February 22, 2024
Colubrines
0.10
0.02
0.1 Anoles "True" geckos 0.2
0.03 0.01
Sphenomorphines
0.0
0.00
0 250 500 750 1000 1250 rds kes rds es les es rms kes is) 300 1000 3000 sis ode ory ateode ape unt phy ape iet unt
liza na liza phin anolaeminbrifo sna droph ine m ns lim m /sh co ra sh dl co
all all s ed mor lio −coluriform (Hy ll k rity ose c ing ize imb eogkull
Species richness duc no Elongation index (log) u br
a
b −re sphe noncolub nake
s sk pahem oragdy s git/l g s rte
limalian s c f o di ve
s t r sea b
Au

Fig. 4. Speciation rates in snakes are consistently elevated relative to rates across all snakes and lizards and for several key groups. (E) Speciation rates
lizards. (A) Subsampled phylogeny of 246 clades (as shown in Fig. 1) illustrating for individual species as a function of body elongation. Snakes are elongate and
(B) speciation rates (CLaDS) relative to tree-wide median rate. Rates for each have fast rates of speciation (D), but elongate and/or limb-reduced lizards show no
tip represent the mean rate for all species assigned to each terminal clade. trends toward faster rates. Phylogenetic distribution of limb reduction is indicated
(C) Species richness and tip-averaged speciation rate for major clades within by blue branches in the phylogeny shown in (A). (F) Phylogenetic regression analysis
squamates (subfamily level). (D) Median and interquartile range for tip speciation reveals minimal effect of any phenotypic traits on speciation rates (R2 < 0.01).

of cranial kinesis in a nonsnake squamate oc- (e.g., sexual selection and trophic flexibility) of the biota through time may be partly at-
curs in a species-poor group of legless geckos might facilitate comparison of putative causal tributable to phenotypic changes that confer
(33). Conversely, one lineage of snakes—the processes or agents across diverse branches of increased survivorship at the species or clade
splitjaw snakes (Bolyeridae)—has further in- the Tree of Life. Nonetheless, the action of these level (40, 41). Determining why snakes and
creased the flexibility of the alethinophidian and other innovations may be highly contingent other groups—from bats to beetles, and from
trophic apparatus through dramatic skeletal on geographic and other circumstances peculiar angiosperms to acanthomorph fishes—have
innovations not seen in any other vertebrates to individual clades (35, 36). experienced such dramatic shifts in evolution-
(34). Yet this innovation was unaccompanied The view of phenotypic evolution emerging ary tempo and mode is, as Darwin put it, an
by any substantive ecological or species diver- from squamates and other taxa (15, 20, 37) is “abominable mystery.” Yet the major features
sification, and the lineage is currently repre- that many of the most spectacular changes in the of biological diversity may owe more to these
sented by a tiny clade of endangered or recently history of life have involved macroevolutionary revolutionary events than to the causal struc-
extinct island endemics (Casarea dussumieri singularities: rare, transformative events that are tures we infer by extrapolating contemporary
and Bolyeria multocarinata). unpredictable from evolutionary time, phylo- microevolution through geological time.
Snakes thus appear to represent an ongoing genetic position, and other factors. In this study,
adaptive radiation characterized by sustained the recurrent evolution of snake-like traits in REFERENCES AND NOTES

ecological and species diversification that was other groups of nonsnake lizards has failed to 1. C. L. Camp, Bull. Am. Mus. Nat. Hist. 48, 289–480
(1923).
triggered ultimately by one or more unknown generate comparable evolutionary outcomes, 2. T. Townsend, A. Larson, E. Louis, J. R. Macey, Syst. Biol. 53,
and perhaps unknowable singular events. The suggesting that replicated patterns of adaptive 735–757 (2004).
principal challenge in resolving the causal basis radiation observed in many groups of organisms 3. H. W. Greene, Snakes: The Evolution of Mystery in Nature
(Univ. of California Press, 1997).
of such singularities involves a lack of phyloge- (35, 38) may be superseded at deeper phyloge- 4. J. J. Wiens, M. C. Brandley, T. W. Reeder, Evolution 60, 123–141
netic replication. By focusing on traits specific netic scales by systemic shifts that ultimately (2006).
to squamates (e.g., cranial kinesis and venom), give rise to higher taxa (11). The extent to which 5. K. Schwenk, Brain Behav. Evol. 41, 124–137 (1993).
6. K. Metzger, in Topics in Functional and Ecological
evolutionary biologists are necessarily limited such shifts are qualitatively distinct from those Vertebrate Morphology, P. Aerts, K. D’Aout, A. Herrel,
to a small number of statistically independent associated with young radiations is likely to R. Van Damme, Eds. (Shaker Publishing, Düren and
data points or contrasts (28). However, some remain a challenging problem. Perhaps most Maastricht, 2002), pp. 15–46.
7. L. J. Vitt, E. R. Pianka, Proc. Natl. Acad. Sci. U.S.A. 102,
hypotheses, particularly those associated with importantly, the pruning effect of extinction,
7877–7881 (2005).
genomic change or molecular evolutionary rates through the selective elimination of ecologi- 8. A. Watanabe et al., Proc. Natl. Acad. Sci. U.S.A. 116,
(20), could be tested across vastly greater cally and phenotypically intermediate clades, 14688–14697 (2019).
phylogenetic scales (e.g., all vertebrates or may exacerbate the distinctiveness of clades 9. F. O. Da Silva et al., Nat. Commun. 9, 376 (2018).
10. U. Roll et al., Nat. Ecol. Evol. 1, 1677–1682 (2017).
metazoans). Likewise, the development of taxon- that comprise the “survivors” that we observe 11. G. G. Simpson, Tempo and Mode in Evolution (Columbia
independent ontologies to describe key processes in the present day (39). Thus, transformation Univ. Press, 1944).

Title et al., Science 383, 918–923 (2024) 23 February 2024 5 of 6

RES EARCH | R E S E A R C H A R T I C L E

12. M. J. Landis, J. G. Schraiber, Proc. Natl. Acad. Sci. U.S.A. 114, 37. C. Venditti, A. Meade, M. Pagel, Nature 479, 393–396 (2011). Council (C.M.); Australian Research Council grant DE130101567
13224–13229 (2017). 38. D. L. Mahler, T. Ingram, L. J. Revell, J. B. Losos, Science 341, (M.E.H.J.); National Council for Scientific and Technological Development
13. C. R. Cooney et al., Nature 542, 344–347 (2017). 292–295 (2013). grant 304715/2021-2 (D.O.M.); Fundação de Apoio à Pesquisa do Estado
14. G. Navalón, A. Bjarnason, E. Griffiths, R. B. J. Benson, Nature 39. D. N. Reznick, R. E. Ricklefs, Nature 457, 837–842 (2009). da Paraíba-PB (D.O.M.); Coordenação de Aperfeiçoamento de Pessoal
611, 306–311 (2022). 40. D. Jablonski, Science 231, 129–133 (1986). de Nível Superior-PRINT UFPB grant AUXPE 88881.312082/2018-01
15. A. Goswami et al., Science 378, 377–383 (2022). 41. G. C. Williams, Natural Selection: Domains, Levels, and (D.O.M.); NSF Division of Environmental Biology grants 0415430,
16. J. W. Valentine, D. Jablonski, D. H. Erwin, Development 126, Challenges (Oxford Univ. Press, 1992). 9200779, and 9505518 (L.J.V.), and a postdoctoral fellowship with the
851–859 (1999). 42. P. O. Title et al., Data from: The macroevolutionary singularity Environmental Resilience Institute at Indiana University (P.O.T.).
17. Materials and methods are available as supplementary materials. of snakes, Dryad (2023); https://doi.org/10.5061/ Author contributions: Project conceptualization and planning:
18. W. Jetz, G. H. Thomas, J. B. Joy, K. Hartmann, A. O. Mooers, dryad.p5hqbzkvb. D.L.R., P.O.T., S.S., M.C.G., and G.C.C. Data collection: S.S., T.JC., M.R.G.,
Nature 491, 444–448 (2012). 43. P. O. Title et al., Data from: The macroevolutionary singularity G.C.C., L.B.Q.C., G.R.C., N.S., M.E.H.J., P.O.T., D.L.R., M.C.G., and L.J.V.
19. N. P. Giannini, Syst. Biol. 52, 684–695 (2003). of snakes, Version v1, Zenodo (2023); https://doi.org/ Funding acquisition: D.L.R., R.A.P., G.C.C., G.R.C., and S.S. Resources:
20. T. R. Simões, O. Vernygora, M. W. Caldwell, S. E. Pierce, 10.5281/zenodo.8250112. D.L.R., R.A.P., G.C.C., L.B.Q.C., G.R.C., N.D.P., S.C.D., C.M., D.O.M., E.R.P.,
Nat. Commun. 11, 3322 (2020). S.A.S., and L.J.V. Method development: M.C.G., D.L.R., and P.O.T. Data
21. O. Seehausen, C. E. Wagner, Annu. Rev. Ecol. Evol. Syst. 45, ACKN OWLED GMEN TS
analysis: P.O.T., S.S., M.C.G., D.L.R., and G.C.C. Data visualization: P.O.T.,
621–651 (2014). S.S., M.C.G., I.P., and D.L.R. Writing of primary draft: D.L.R., P.O.T., and
22. D. Schluter, The Ecology of Adaptive Radiation (Oxford Univ. We thank G. Schneider and F. Kraus for sample access, P. Skipwith S.S. Editing, commenting on, and approving final draft: All authors but
Press, 2000). and S. Potter for logistical support, and A. Devault, J. Enk, E.R.P. (deceased). Competing interests: The authors declare that
23. T. Price, Speciation in Birds (W. H. Freeman, 2008). M. R. Marchan-Rivadeneira and the team at UMich ARC for they have no competing interests. Data and materials availability:
24. T. J. Colston, G. C. Costa, L. J. Vitt, Biol. J. Linn. Soc. Lond. 101, technical support. We also thank A. Davis Rabosky, members of Accession numbers (NCBI) for genomic data are available in
476–486 (2010). the Rabosky Lab, and two anonymous reviewers for comments supplemental data S1. Data and scripts used in data processing,
25. M. C. Grundler, D. L. Rabosky, PLOS Biol. 19, e3001414 and suggestions that substantially improved this manuscript. analysis, and visualization are available on Dryad (42), Zenodo (43),

Downloaded from https://www.science.org at Museu Paraense Emlio Goeldi on February 22, 2024
(2021). Approximately 80% of the dietary records in our dataset came from and GitHub (https://github.com/macroevolution/squamata). License
26. D. L. Rabosky et al., Nature 559, 392–395 (2018). preserved voucher specimens, and we thank the curatorial staff of the information: Copyright © 2024 the authors, some rights reserved;
27. L. F. Henao Diaz, L. J. Harmon, M. T. C. Sugawara, E. T. Miller, many natural history collections worldwide that made this research exclusive licensee American Association for the Advancement of
M. W. Pennell, Proc. Natl. Acad. Sci. U.S.A. 116, 7403–7408 possible. For photographs, we thank T. Schramer, R. Recoder, Science. No claim to original US government works. https://www.
(2019). M. T. Rodrigues, J. Grummer, J. Farquhar, J. Crowe-Riddell, and science.org/about/science-licenses-journal-article-reuse
28. J. C. Uyeda, R. Zenil-Ferguson, M. W. Pennell, Syst. Biol. 67, A. Davis Rabosky. Pictures by R. Brown and B. Trapp were modified to
1091–1109 (2018). compose manuscript figures. Funding: This work was supported by SUPPLEMENTARY MATERIALS
29. B. G. Fry et al., Mol. Cell. Proteomics 7, 215–246 (2008). NSF Division of Environmental Biology grant 1754398 (D.L.R. and S.S.); the
science.org/doi/10.1126/science.adh2449
30. H. W. Greene, Integr. Comp. Biol. 23, 431–441 (2015). Packard Fellowship for Science and Engineering (D.L.R.); Coordenação
Materials and Methods
31. G. I. M. Pasquesi et al., Nat. Commun. 9, 2774 (2018). de Aperfeiçoamento de Pessoal de Nível Superior (G.R.C.); the National
Figs. S1 to S44
32. T. A. Castoe et al., Proc. Natl. Acad. Sci. U.S.A. 110, Council for Scientific and Technological Development (G.R.C.); the
Tables S1 to S5
20645–20650 (2013). FAPDF (G.R.C.); California State University, Dominguez Hills RSCA
References (44–249)
33. F. C. Patchell, R. Shine, J. Herpetol. 20, 59–64 (1986). (S.S.); NSF Division of Environmental Biology grant 1754425 (G.C.C.);
MDAR Reproducibility Checklist
34. T. H. Frazzetta, Am. Nat. 104, 55–72 (1970). NSF Division of Biological Infrastructure grant 0905765 (R.A.P.) and
Data S1 to S3
35. C. E. Wagner, L. J. Harmon, O. Seehausen, Nature 487, 1930030 (S.A.S.); NSF Division of Environmental Biology grant 1441719
366–369 (2012). (R.A.P.); the J. William Fulbright Fellowship (T.J.C.); NSF Division of Submitted 22 February 2023; accepted 2 January 2024
36. A. de Queiroz, Syst. Biol. 51, 917–929 (2002). Environmental Biology grant 1501711 (T.J.C.); Australian Research 10.1126/science.adh2449

Title et al., Science 383, 918–923 (2024) 23 February 2024 6 of 6