brief communications circumvents constraints on prey size Instead of swallowing a victim whole, -eating tear off bite-sized pieces.

or who are unable to take bites These two of crab-eating snake out of their food, the size of the food have not merely retained a common Fitem that can be consumed is con- feeding mechanism of a crustacean-eating strained by the maximal size of the mouth ancestor. The feeding mechanism used by opening (gape) — snakes are an example captive F. leucobalia (nǃ12) to eat both of gape-limited predators and they usually hard-shelled (nǃ50) and soft-shelled swallow their prey whole1,2. Here we (nǃ6) differed from that used by describe unique feeding behaviours in two G. prevostiana in the mode of attacking closely related species of snake, which cir- prey, how prey was restrained, the usual cumvent their gape limitation by removing orientation for swallowing the crab’s body, and consuming pieces from newly moulted and how pieces were torn from prey. In crabs that are too large to be swallowed contrast, the patterns of coiling shown by intact. This evolutionary innovation is sur- many lineages of constricting snakes are prising, as the needle-like teeth and highly usually evolutionarily conserved6. When mobile bones that facilitate the capture and F. leucobalia consumed hard-shelled crabs, engulfment of large, whole prey by snakes the carapace always remained intact, con- are ill-suited both to cutting and to generat- trary to a suggestion that this snake uses ing large bite forces. its jaws to crush crabs3. By using a method We captured individuals of the homalop- similar to that used for breaking legs off sine snake species Gerarda prevostiana and crabs4, however, captive F. leucobalia did Fordonia leucobalia and examined their gut tear carapaces apart in 50% of trials with contents. We found that G. prevostiana con- very large, soft-shelled crabs. tained the remnants of only freshly moulted We are unaware of any other species of crabs, whereas F. leucobalia contained snake that regularly tears its prey apart so material exclusively from hard-shelled crabs. that it can consume larger prey than could F. leucobalia has several morphological3 and be swallowed whole. The novel feeding some behavioural4 specializations that are mechanism used by G. prevostiana is sur- associated with its unusual diet. However, prising in view of its relatively unremarkable G. prevostiana, the diet and feeding behav- anatomy, which cautions against drawing iour of which have not previously been functional inferences from anatomy in the investigated, lacks the hypertrophied cranial absence of behavioural studies. musculature and short and blunt teeth of The diversity of tropical aquatic snakes in F. leucobalia5. We therefore collected G. pre- correlates well with their range vostiana and a variety of crabs in Singapore of feeding mechanisms and specialisms. and observed the snakes feeding in a dark Within a single mangrove, F. leucobalia and room with infrared video cameras. G. prevostiana consume the same species of Captive G. prevostiana (nǃ15) refused to crab, but partition this resource by relying attack hard-shelled crabs, but bit and rapidly on different moult stages; two other species, swallowed freshly moulted crabs. In 85% of Cantoria violacea and Cerberus rynchops, eat trials (nǃ26), the snake pulled its head only snapping shrimp and fish, respectively7. through a loop in its body and continued to Bruce C. Jayne*, Harold K. Voris†, pull on the crab while coiled around the ani- Peter K. L. Ng‡ mal (Fig. 1a, b). In some trials (35%), snakes *Department of Biological Sciences, University of broke off and consumed up to four crab legs Cincinnati, Ohio, 45221-0006, USA but none of the carapace. The carapace of the e-mail: [email protected] crab was either partially or completely torn †Division of Amphibians and , Field apart in 20% and 32%, respectively, of the Museum of Natural History, 1400 South Lake Shore Figure 1 Feeding behaviour and prey size of the snake Gerarda 22 feedings in which snakes demonstrated Drive, Chicago, Illinois 60605, USA prevostiana. a, b, Infrared video images of G. prevostiana 12.6 s the ‘loop-and-pull’ behaviour (Fig. 1a, b). ‡Raffles Museum of Biodiversity Research, before (a) and 0.6 s after (b) completely tearing apart the carapace Tearing of the carapace was most common Department of Biological Sciences, National of a freshly moulted grapsid crab that was considerably wider than for large crabs, and only occurred when a University of Singapore, Science Drive 4, the snake’s head. Both pieces of the carapace (arrows) were con- loop-and-pull action was used. Singapore 119260 sumed. For videos showing this behaviour, see www.biology.uc.edu/ We also captured two G. prevostiana 1. Cundall, D. & Greene, H. W. in Feeding: Form, Function and faculty/jayne/gerarda_feed2.htm. c, Five appendages (top) from a individuals that had consumed pieces of Evolution in Tetrapod Vertebrates (ed. Schwenk, K.) 293–333 freshly moulted crab (Episesarma versicolor) that was attacked and newly moulted crabs that were much larger (Academic, New York, 2000). partially consumed by a 33-g G. prevostiana snake (middle) in the relative to the snake than any of those used 2. Arnold, S. J. in Snakes: Ecology and Behavior (ed. Collins, J. T.) 87–115 (McGraw-Hill, New York, 1993). wild. The intact, 41-g E. versicolor shown at the bottom is similar in in laboratory trials (Fig. 1c). Consequently, 3. Savitzky, A. H. Am. Zool. 23, 397–409 (1983). size to the crab that was attacked. The white circle (actual ripping crabs apart is unlikely to be inciden- 4. Shine, R. & Schwaner, T. Copeia 1985, 1067–1071 (1985). diameter, 13 mm) is the maximum gape of the snake (determined tal, and probably allows G. prevostiana to 5. Voris, H. K. et al. Copeia (in the press). 6. Greene, H. W. & Burghardt, G. M. Science 200, 74–77 (1978). by inserting cylindrical gauges with 1-mm diameter increments regularly consume crabs with body sizes that 7. Voris, H. K. & Murphy, J. C. J. Nat. Hist. (in the press). into the mouth of the anaesthetized snake). exceed our estimate of its maximal gape. Competing financial interests: declared none.

NATURE | VOL 418 | 11 JULY 2002 | www.nature.com © 2002 Nature Publishing Group 143