Journal of Chemical Ecology, Vol. 18, No. 10, 1992
NAIVE OPHIOPHAGUS LIZARDS RECOGNIZE A N D AVOID VENOMOUS SNAKES USING CHEMICAL CUES
JOHN
A. P H I L L I P S *
and ALLISON
C. A L B E R T S
Center for Reproduction of Endangered Species Zoological Society of San Diego Box 551, San Diego, California 92112 (Received March 9, 1992; accepted June 1, 1992) Abstract--Monitor lizards prey on snakes. Conversely, venomous snakes prey on juvenile monitor lizards. Immediately after hatching, monitor lizards are naive to all prey items, thus correct assessment of snake prey is paramount for survival. Experiments were conducted to determine how hatchling monitor lizards (Faranus albigularis) with no previous exposure to snakes reacted to sympatric venomous and nonvenomous snakes. Hatchling lizards attacked harmless snakes, but avoided venomous species. Lizards readily accepted meat from skinned snakes, regardless of species, When invertebrate prey covered with skin segments from venomous snakes were restrained from moving, they were usually investigated by tongue-flicking and rejected. Unrestrained skin-covered prey, however, were generally attacked and eaten without prior evaluation by tongue-flicking. Attack was inhibited in trials in which unrestrained prey were tongue-flicked, suggesting that chemical cues contained in snake skins mediate avoidance of venomous snakes. Selection for the ability to perceive snake integumental chemicals may be especially strong in species that both consume and are consumed by snakes. Key Words--Chemical cues, Varanus, snakes, lizards, monitor lizard.
INTRODUCTION
A v a r i e t y o f v e r t e b r a t e p r e y react to the s c e n t p r o d u c t s o f s n a k e s ( W e l d o n a n d B u r g h a r d t , 1979; W e l d o n , 1991), a n d c e r t a i n p r e y u s e c h e m o r e c e p t i o n to disc r i m i n a t e s n a k e s t h a t are t h e i r p r e d a t o r s f r o m t h o s e that are h a r m l e s s ( W e l d o n , 1982; Dial et al., 1989). T h e c h e m i c a l s e n s e s also aid s o m e p r e d a t o r s in t h e i r *To whom correspondence should be addressed. 1775 0098~0331/92/1000-1775506.50/0 9 1992 Plenum Publishing Corporation
!776
PHILLIPS AND ALBERTS
identification of snake prey (Weldon and Schell, 1984). However, the responses by a species to chemical cues from organisms that are potentially either their predators or their prey have not been examined systematically. As a genus, monitor lizards consume a variety of prey (King and Green, 1979; Losos and Greene, 1988) and chemoreception is extremely important in their foraging ecology (Auffenberg, 1978; Pianka, 1986; Cooper, 1989). Although invertebrates account for the majority of the natural diet, a substantial percentage of the annual caloric intake is obtained from vertebrate prey, including snakes (Shine, 1986; Branch, 1991). Conversely, as juveniles, many species of monitors are preyed upon by snakes (Shine, 1989; Shine and Slip, 1990). The white-throated savanna monitor, Varanus albigularis, is a large ( > 5 kg) lizard that inhabits a variety of southern African habitats, including the arid savanna. In this environment, dramatic increases in vegetational growth and animal populations occur during the rainy season (Berry and Louw, 1982). As with most reptiles, hatching of snakes and monitor lizards occurs midway through the wet season (Branch, 1988). Growth of recently hatched snakes and lizards during this period is rapid, with most species amassing substantial fat deposits required for survival during the subsequent eight months of drought. Juvenile lizards and snakes are favored prey items of numerous species of birds and mammals, and lizards and snakes also prey upon one another (Branch, 1988). Snakes that eat lizards and those that do not are likely to be visually similar at a distance, thus necessitating close-range investigation by lizards attempting to distinguish them. Because snakes are generally formidable predators, any misclassification of a predator species as prey is potentially very costly. In this study, we demonstrate the chemical discrimination of snake predators from snake prey by naive white-throated savanna monitor hatchlings.
METHODS AND MATERIALS
Five gravid female white-throated savanna monitors were caught and housed in a 50-m2 outdoor enclosure at the Etosha Ecology Institute, Etosha National Park, in north-central Namibia. Oviposition occurred within two weeks of capture, and eggs were artificially incubated. The females were released at the site of their capture. Twenty-seven of 120 hatchlings (five or six hatchlings from each clutch) were retained in captivity; the remaining hatchlings were released in the area where their mother had been caught. After the umbilical scar had completely sealed, the retained hatchlings were released in the outdoor enclosure. The hatchlings were exposed to human contact for 2 hr each day, and exhibited no discemible aggressive or defensive behavior to human presence in the enclosure. Hatchlings were offered daily a variety of invertebrates (grasshoppers, crickets, land snails, centipedes, scorpions) that are
N A I V E L IZ AR D S I D E N T I F Y BY O D O R V E N O M O U S SNAKES
1777
known to be common prey of this species of monitor lizard (W.R. Branch, personal communication; J.A. Phillips, unpublished data). Prey items were offered with forceps individually or were placed in the enclosure en masse. Water was available ad libitum. Hatchlings were fed until satiated. All animal care was in compliance with the guidelines of the Zoological Society of San Diego Animal Care and Use Committee. Experiment 1. At six weeks of age each hatchling (total length 220-250 mm) was offered unrestrained grasshoppers, and land snails, grasshoppers, and corn crickets restrained by forceps, on each of two occasions. Trials were begun between 1300 and 1400 hr, at which time the lizards had been actively searching for food for several hours. The sequence of tongue-flicking, aggressive and/or defensive postures, and attack and/or retreat were videotaped and analyzed for each prey item. At seven weeks of age, hatchlings were assigned to one of three groups (nine hatchlings per group). Members of each group were individually exposed to either a whole fresh (less than 2 h old) road-killed sand snake (Psammophis leightoni), road-killed spitting cobra (Naja nigricollis), or road-killed homed adder (Bitis caudalis) at two-day intervals. The order of presentation of the three snakes was randomized among hatchling groups, and a total of nine different snake carcasses (three from each species) were used in the presentations. The behavioral responses of individual hatchling lizards to the road-killed snakes were videotaped and analyzed as above. Experiment 2. At eight weeks of age, the hatchlings were assigned to one of five groups, with one to two hatchlings from each clutch represented in each group (numbers of hatchlings in each of the five groups were 5, 5, 5, 6, 6). Each member of the five groups of hatchlings was tested individually for its response to fresh road-killed sand snakes, spitting cobras, and homed adders under five treatment conditions: (I) whole snake, (II) 50-mm intact snake sections, (III) 50-mm sections of snake with the skin removed, (IV) live grasshopper covered with a coat of snake skin restrained from jumping by a pair of forceps (Figure 1), and (V) live unrestrained grasshopper covered with a coat of snake skin. The snakes used in the experiment were 700-930 mm total length. Treatments were presented to the five hatchling groups in a Latin-square design, thus the order of exposure to the five treatments differed for each group of hatchlings. Each of the three snake species was tested under each of the five treatment conditions in random order at 72- to 96-h intervals. Trials were begun between 1300 and 1400 hr. Statistics. Repeated measures analysis of variance (ANOVA) was used to examine variation in the rate of tongue-flicking in response to invertebrate prey and venomous and nonvenomous snakes in experiment 1. In experiment 2, Latin-square ANOVA was used to compare the percentage (arcsine square-root transformed) of hatchlings in each of the five groups that attacked snakes under
1778
PHILLIPS AND ALBERTS
FIG. I. Pamphagid grasshopper (approximately 50-ram body length) covered with a coat of skin from a homed adder.
the five treatment conditions. Differences in response to each of the three snake species were assessed using Fisher's protected least significant difference tests (Carmer and Swanson, 1973).
RESULTS AND DISCUSSION
In experiment 1, hatchlings exhibited tongue-flicking, a behavior that probably delivers odorants to the vomeronasal organ (Halpem, 1987), when exposed to invertebrate prey (Table 1). In most cases, two to four tongue-flicks were emitted before a prey item was attacked. However, most attacks on unrestrained grasshoppers occurred in the absence of tongue-flicking. Tongue-flicking rates were significantly lower in response to unrestrained moving grasshoppers than to the other invertebrate prey items presented (F = 132.48; df = 3,53; P -0.0001), suggesting that visual cues may be more important than chemical cues in eliciting attack of moving prey. In no instance were aggressive or defensive behaviors observed in response to live invertebrate prey. Sand snakes do not eat monitor lizards but are relatively common prey
1779
NAIVE LIZARDS IDENTIFY BY ODOR VENOMOUS SNAKES
TABLE 1. RESPONSES OF
27
HATCHLING SAVANNA MONITORS TO INVERTEBRATES AND W H O L E R O A D - K I L L E D SNAKES a
Invertebrates Land snail (54) Corn cricket (54) Restrained grasshopper (54) Unrestrained grasshopper (54) Snakes Sand snake (27) Spitting cobra (27) Horned adder (27)
Percent of hatchlings exhibiting defensive/ aggressive behaviors
Mean tongue flicks
Percent attack
4.33 4.35 3.94 0.18
100 100 100 100
0 0 0 0
0 0 0 0
0 0 0 0
0 0 0 0
4.27 4.84 4.55
100 22 11
0 100 100
0 93 100
0 52 81
0 100 0
Lateral
Hiss
Slap
Vent
aSample size in parentheses. Each invertebrate prey type was offered on two different occasions. Lateral compression of the body (lateral), hissing vocalizations (hiss), and tail slap (slap) are aggressive displays, whereas vent dragging (vent) is a territorial behavior.
items o f adult savanna monitors. The stomach contents o f 18 road-killed savanna monitors revealed two sand snakes, among other prey, accounting for about 10% o f the total caloric value o f the stomach contents (J.A. Phillips, personal observation). Sand snakes were universally attacked by hatchlings in a manner similar to invertebrate prey (Table 1). After initial attack, the hatchlings exhibited no hesitation in attempts to consume the sand snakes. No aggressive or defensive behaviors were noted in response to sand snakes. The spitting cobra and the homed adder are highly venomous snakes that eat hatchling monitor lizards (Branch, 1988). Similar to other potential prey items, tongue-flicking by hatchling lizards occurred immediately after exposure to these two species. There was no difference in the rate o f tongue-flicking in response to the three species o f snakes ( F = 1.16; df = 2,78; P = 0.32). However, in only nine o f 54 cases was attack o f venomous snakes observed (Table 1), and in no instance did the hatchlings attempt to consume venomous snake carcasses. Regardless o f whether snakes were attacked, all hatchlings exhibited one or more defensive and/or aggressive behaviors toward the venomous snakes (Table 1). The behavioral responses exhibited by hatchlings were specific to each o f the two venomous snake species. In particular, posterior adpression (vent dragging), apparent in every encounter with the spitting cobra, was never elicited when hatchlings were exposed to h o m e d adders. Hatchling savanna
1780
PHILLIPS AND ALBERTS
monitors defecate at latrine sites, and when exiting these sites, often perform vent dragging (J.A. Phillips, personal observation). When other hatchlings investigate areas where vent dragging has occurred, aggressive posturing is frequently observed, suggesting that vent dragging may function in territorial marking. In experiment 2, the hatchlings responded to whole bodies of the three snake species in a manner similar to that of experiment 1. Latin-square A N O V A revealed no effects of order of presentation (F = 1.001, df = 4,12, P > 0.50), no differences in response among the five hatchling groups (F = 1.001, df = 4,12, P > 0.50), and no differences in response across the five treatments (F = 1.001, df = 4,12, P > 0.50) for sand snakes (Figure 2). These results contrasted with those for the two venomous snake species. For spitting cobras, there were no effects of order of presentation (F = 0.56, df = 4,12, P > 0.50), no differences among hatchling groups (F = 0.28, df = 4,12, P > 0.50), but significant differences across treatments (F = 6.84, df = 4,12, P < 0.01). For homed adders, there were also no effects of order of presentation (F = 1.62, d f = 4,12, P > 0.20), no differences among hatchling groups (F = 0.78 df = 4, 12, P > 0.50), but significant differences across treatments (F = 48.10, df = 4,12, P < 0.001). For spitting cobras and homed adders, Fisher's protected least significant difference tests revealed that hatchlings were significantly more likely to attack skinned body sections and skin-covered unrestrained grasshoppers than whole snakes, sectioned snakes, or skin-covered restrained grasshoppers (Figure 2). Hatchlings avoided sectioned snakes and skin-covered restrained grasshoppers 100
~
8o
~
60
=N
40
N O Q.
20
9 [] []
sand snake spitting cobra horned adder
0
II
III Treatment
IV
FIG. 2. Mean percentage (+SE) of hatchling savanna monitors in five groups that attacked (I) whole road-killed snakes, (II) snake sections, (III) skinned snake sections, (IV) a restrained grasshopper covered with a segment of snake skin, or (V) an unrestrained grasshopper covered with a segment of snake skin. Three snake species were presented, the sand snake, and two highly venomous species, the spitting cobra and the homed adder.
NAIVE LIZARDS I D E N T I F Y BY O D O R V E N O M O U S SNAKES
1781
at a rate similar to that for whole snakes. Hatchlings attacked unrestrained grasshoppers covered with skin from either spitting cobras or homed adders in 49 of 54 trails. The five hatchlings that rejected skin-covered grasshoppers exhibited tongue-flicking, whereas all hatchlings that attacked grasshoppers did so without tongue-flicking. This indicates that chemical, rather than visual, cues associated with snake skins mediate avoidance of venomous snakes. Past studies with scincid lizards have shown higher levels of tongue-flicking toward snakes that are their predators than those that are not (Cooper, 1990). Skinks are insectivorous, thus harmless snakes represent neither predator nor prey and are not expected to elicit a strong response. The lack of a difference in tongue-flicking rate to the three snake species in the present study probably reflects the fact that all of the snake species tested interact with savanna monitors. Although spitting cobras and homed adders are predators and sand snakes are prey, predator detection and foraging are both ecological contexts in which tongue-flicking occurs (Burghardt, 1970; Halpem and Kubie, 1983; Simon, 1983; Duvall et al., 1990). Naive hatchling savanna monitors not only differentiated venomous from nonvenomous snakes in their propensity to attack them, they also reacted differently toward the two species of venomous snakes. Homed adders (Viperidae) and spitting cobras (Elapidae) possess divergent hunting strategies and chemically distinctive venoms (Visser, 1967). For these reasons, it is not unexpected that responses by their prey will differ. The tendency for some hatchlings to attack cobras and all to perform vent dragging behavior, which may be associated with aggression in savanna monitors, suggests that cobras may be less dangerous than adders, at least to these lizards. The long, stiff fangs of adders, capable of inflicting a deeper and more serious bite than those of cobras, may partially explain the differential treatment of the two predatory snake species. The rejection of invertebrate prey bearing skins of venomous snakes and the indiscriminant acceptance of skinned snake meat suggest that savanna monitor hatchlings use skin-borne chemicals to identify snakes. That lizards having no previous experience with snakes exhibit these responses indicates that they are innate. Because snakes represent both predators and prey for savanna monitors, the ability to discriminate among them potentially functions in two different adaptive contexts. Foraging efficiency is increased when appropriate prey items are identified prior to attack, and survivorship is enhanced when potential predators are detected and avoided with minimal physical contact. Previous experiments suggest that the movement of prey is important in the feeding behavior of reptiles (Drummond, 1985). Chemical cues also stimulate feeding behavior in snakes, and naive hatchlings recognize and attack their species-typical prey on the basis of chemical cues (Burghardt, 1966, 1968). Our results for monitor lizards indicate that visual cues associated with prey movement can override chemical cues. These visual responses may be learned, as
1782
PHILLIPS AND ALBERTS
the tested lizards had been fed invertebrate prey since hatching. A hierarchy of cues involving multiple sensory modalities and both learned and unlearned behavior is probably used by lizards in the recognition of sympatric snake species and assessment of their relative suitability as prey. Acknowledgments--We thank the Department of Nature Conservation of Namibia, and especially the Namibian scientists at Etosha National Park for their permission and help in carrying out this research. We are grateful to Janet Phillips for excellent technical help and to Joel Berger for useful advice on experimental design. Invaluable comments on earlier drafts of the manuscript were provided by Don Lindburg, Don Owings, and Paul Weldon. This work was supported by an NSF Conservation and Restoration Biology grant (BNS 90-00100) and the Institute of Museum Services (IC-00348-90).
REFERENCES AUFFENBERG,W. 1978. Social and feeding behavior in Varanus komodoensis, in pp. 301-331, N. Greenberg and P.D. MacLean (eds.). Behavior and Neurology of Lizards. U.S. Department of Health, Education, and Welfare, Washington, D.C. BERRY, H.H., and Louw, G.N. 1982. Nutritional balance between grassland productivity and large herbivore demand in the Etosha National Park. Madoqua 12: 141-150. BRANCH, W.R. 1988. South African Red Data Book: Reptiles and Amphibians. South African Scientific Programs Report, CSIR, Pretoria. BRANCH, W.R. 1991. The Regenia registers of 'Gogga' Brown (1869-1909) "Memoranda on a species of monitor or Varan." Mertensiella 2: 57-110. BURGItARDT, G.M. 1966. Stimulus control of the prey attack response in naive garter snakes. Psychon. Sci. 4:37-38. BURGHARDT,G.M. 1968. Chemical preference studies on newborn snakes of three sympatric species of Natrix. Copeia 1968:732-737. BUROHARDT, G.M. 1970. Chemical perception in reptiles, in pp. 241-308, J.W. Johnston, D,G. Moulton, and A. Turk, (eds.). Advances in Chemoreception. Appleton-Century-Crofts, New York. CARMER, S.G., and SWANSON, M.R. 1973. An evaluation of ten pairwise multiple comparison procedures by Monte Carlo methods. J. Am. Stat. Assoc. 68:66-74. COOPER, W.E., JR. 1989. Prey odor discrimination in the varanoid lizards Heloderma suspectum and Varanus exanthematicus. Ethology 81:250-258. COOPER, W.E., JR. 1990. Chemical detection of predators by a lizard, the broad-headed skink (Eumeces laticeps). J. Exp. Zool. 256:162-167. D~AL, B.E., WELDON, P.J., and CURTIS, B. 1989. Chemosensory identifieation of snake predators (Phyllorhynchus decurtatus) by banded geckos (Coleonyx variegatus). J. Herpetol. 23:224229. DRUMMOND, H. 1985. The role of vision in the predatory behaviour of natrieine snakes. Anita. Behav. 1985:206-215. DUVALL, D., CHISZAR,D., HAYES, W.K., LEONHARDT,J.K., and GOODE, M.J. 1990. Chemical and behavioral ecology of foraging in prairie rattlesnakes (Crotalus viridis viridis). J. Chem. Ecol. 16:87-101. HALPERN,M. 1987. The organization and function of the vomeronasal system. Annu. Rev. Neurosci. 10:325-362. I-IALVERN,M., and KUBIE,J.L. 1983. Snake tongue flicking behavior: Clues to vomeronasal system
NAIVE LIZARDSIDENTIFY BY ODOR VENOMOUSSNAKES
1783
functions, in pp. 45-72, D. Miiller-Schwarze and R.M. Silverstein (eds.). Chemical Signals in Vertebrates, Vol. 3. Plenum Press, New York. KING, D., and GREEN, B. 1979. Notes on the diet and reproduction of the sand goanna, Varanus gouldii rosenbergi. Copeia 1979:64-70. Losos, J.B., and GREENE, H.W. 1988. Ecological and evolutionary implications of diet in monitor lizards. Biol. J. Linn. Soc. 35:379-407. PIANKA, E.R. 1986. The Ecology and Natural History of Desert Reptiles. Princeton University Press, Princeton University Press, Princeton, New Jersey. SHINE, R. 1986. Food habits, habitats and reproductive biology of four sympatric species of varanid lizards in tropical Australia. Herpetologica 42:346-360. SHINE, R. 1989. Constraints, allometry, and adaptation: Food habits and reproductive biology of Australian brownsnakes (Pseudonaja: Elapidae). Herpetologica 45:195-207. ShiNE, R., and SLIP, D.J. 1990. Biological aspects of the adaptive radiation of Australasian pythons (Serpentes: Boidae). Herpetologiea 46:283-290. SIMON, C.A. 1983. A review of lizard chemoreeeption, in pp. 119-133, R.B. Huey, E.R. Pianka, and T.W. Schoener (eds.). Lizard Ecology: Studies of a Model Organism. Harvard University Press, Cambridge, Massachusetts. VISSER, J. 1967. Poisonous Snakes of Southern Africa and the Treatment of Snake Bite. Timmons, Cape Town. WELDON, P.J. 1982. Responses to ophiophagus snakes by snakes of the genus Thamnophis. Copeia 1982:788-794. WELbON, P.J. 1991. Responses by vertebrates to chemicals from predators, inpp. 500-521, D. MacDonald, D. Mfiller-Schwarze, and S.E. Natynczuk (eds.). Chemical Signals in Vertebrates, Vol. 5. Oxford University Press, Oxford. WELOON, P.J., and BURGHARDT,G.M. 1979. The ophiophage defensive response in crotaline snakes: Extension to new taxa. J. Chem. Ecol. 5:141-151. WELOON, P.J., and SCHELL,F.M. 1984. Responses by king snakes (Lampropeltis getulus) to chemicals from colubrid and crotaline snakes. J. Chem. Ecol. 10:1509-1520.
NAIVE OPHIOPHAGUS LIZARDS RECOGNIZE A N D AVOID VENOMOUS SNAKES USING CHEMICAL CUES
JOHN
A. P H I L L I P S *
and ALLISON
C. A L B E R T S
Center for Reproduction of Endangered Species Zoological Society of San Diego Box 551, San Diego, California 92112 (Received March 9, 1992; accepted June 1, 1992) Abstract--Monitor lizards prey on snakes. Conversely, venomous snakes prey on juvenile monitor lizards. Immediately after hatching, monitor lizards are naive to all prey items, thus correct assessment of snake prey is paramount for survival. Experiments were conducted to determine how hatchling monitor lizards (Faranus albigularis) with no previous exposure to snakes reacted to sympatric venomous and nonvenomous snakes. Hatchling lizards attacked harmless snakes, but avoided venomous species. Lizards readily accepted meat from skinned snakes, regardless of species, When invertebrate prey covered with skin segments from venomous snakes were restrained from moving, they were usually investigated by tongue-flicking and rejected. Unrestrained skin-covered prey, however, were generally attacked and eaten without prior evaluation by tongue-flicking. Attack was inhibited in trials in which unrestrained prey were tongue-flicked, suggesting that chemical cues contained in snake skins mediate avoidance of venomous snakes. Selection for the ability to perceive snake integumental chemicals may be especially strong in species that both consume and are consumed by snakes. Key Words--Chemical cues, Varanus, snakes, lizards, monitor lizard.
INTRODUCTION
A v a r i e t y o f v e r t e b r a t e p r e y react to the s c e n t p r o d u c t s o f s n a k e s ( W e l d o n a n d B u r g h a r d t , 1979; W e l d o n , 1991), a n d c e r t a i n p r e y u s e c h e m o r e c e p t i o n to disc r i m i n a t e s n a k e s t h a t are t h e i r p r e d a t o r s f r o m t h o s e that are h a r m l e s s ( W e l d o n , 1982; Dial et al., 1989). T h e c h e m i c a l s e n s e s also aid s o m e p r e d a t o r s in t h e i r *To whom correspondence should be addressed. 1775 0098~0331/92/1000-1775506.50/0 9 1992 Plenum Publishing Corporation
!776
PHILLIPS AND ALBERTS
identification of snake prey (Weldon and Schell, 1984). However, the responses by a species to chemical cues from organisms that are potentially either their predators or their prey have not been examined systematically. As a genus, monitor lizards consume a variety of prey (King and Green, 1979; Losos and Greene, 1988) and chemoreception is extremely important in their foraging ecology (Auffenberg, 1978; Pianka, 1986; Cooper, 1989). Although invertebrates account for the majority of the natural diet, a substantial percentage of the annual caloric intake is obtained from vertebrate prey, including snakes (Shine, 1986; Branch, 1991). Conversely, as juveniles, many species of monitors are preyed upon by snakes (Shine, 1989; Shine and Slip, 1990). The white-throated savanna monitor, Varanus albigularis, is a large ( > 5 kg) lizard that inhabits a variety of southern African habitats, including the arid savanna. In this environment, dramatic increases in vegetational growth and animal populations occur during the rainy season (Berry and Louw, 1982). As with most reptiles, hatching of snakes and monitor lizards occurs midway through the wet season (Branch, 1988). Growth of recently hatched snakes and lizards during this period is rapid, with most species amassing substantial fat deposits required for survival during the subsequent eight months of drought. Juvenile lizards and snakes are favored prey items of numerous species of birds and mammals, and lizards and snakes also prey upon one another (Branch, 1988). Snakes that eat lizards and those that do not are likely to be visually similar at a distance, thus necessitating close-range investigation by lizards attempting to distinguish them. Because snakes are generally formidable predators, any misclassification of a predator species as prey is potentially very costly. In this study, we demonstrate the chemical discrimination of snake predators from snake prey by naive white-throated savanna monitor hatchlings.
METHODS AND MATERIALS
Five gravid female white-throated savanna monitors were caught and housed in a 50-m2 outdoor enclosure at the Etosha Ecology Institute, Etosha National Park, in north-central Namibia. Oviposition occurred within two weeks of capture, and eggs were artificially incubated. The females were released at the site of their capture. Twenty-seven of 120 hatchlings (five or six hatchlings from each clutch) were retained in captivity; the remaining hatchlings were released in the area where their mother had been caught. After the umbilical scar had completely sealed, the retained hatchlings were released in the outdoor enclosure. The hatchlings were exposed to human contact for 2 hr each day, and exhibited no discemible aggressive or defensive behavior to human presence in the enclosure. Hatchlings were offered daily a variety of invertebrates (grasshoppers, crickets, land snails, centipedes, scorpions) that are
N A I V E L IZ AR D S I D E N T I F Y BY O D O R V E N O M O U S SNAKES
1777
known to be common prey of this species of monitor lizard (W.R. Branch, personal communication; J.A. Phillips, unpublished data). Prey items were offered with forceps individually or were placed in the enclosure en masse. Water was available ad libitum. Hatchlings were fed until satiated. All animal care was in compliance with the guidelines of the Zoological Society of San Diego Animal Care and Use Committee. Experiment 1. At six weeks of age each hatchling (total length 220-250 mm) was offered unrestrained grasshoppers, and land snails, grasshoppers, and corn crickets restrained by forceps, on each of two occasions. Trials were begun between 1300 and 1400 hr, at which time the lizards had been actively searching for food for several hours. The sequence of tongue-flicking, aggressive and/or defensive postures, and attack and/or retreat were videotaped and analyzed for each prey item. At seven weeks of age, hatchlings were assigned to one of three groups (nine hatchlings per group). Members of each group were individually exposed to either a whole fresh (less than 2 h old) road-killed sand snake (Psammophis leightoni), road-killed spitting cobra (Naja nigricollis), or road-killed homed adder (Bitis caudalis) at two-day intervals. The order of presentation of the three snakes was randomized among hatchling groups, and a total of nine different snake carcasses (three from each species) were used in the presentations. The behavioral responses of individual hatchling lizards to the road-killed snakes were videotaped and analyzed as above. Experiment 2. At eight weeks of age, the hatchlings were assigned to one of five groups, with one to two hatchlings from each clutch represented in each group (numbers of hatchlings in each of the five groups were 5, 5, 5, 6, 6). Each member of the five groups of hatchlings was tested individually for its response to fresh road-killed sand snakes, spitting cobras, and homed adders under five treatment conditions: (I) whole snake, (II) 50-mm intact snake sections, (III) 50-mm sections of snake with the skin removed, (IV) live grasshopper covered with a coat of snake skin restrained from jumping by a pair of forceps (Figure 1), and (V) live unrestrained grasshopper covered with a coat of snake skin. The snakes used in the experiment were 700-930 mm total length. Treatments were presented to the five hatchling groups in a Latin-square design, thus the order of exposure to the five treatments differed for each group of hatchlings. Each of the three snake species was tested under each of the five treatment conditions in random order at 72- to 96-h intervals. Trials were begun between 1300 and 1400 hr. Statistics. Repeated measures analysis of variance (ANOVA) was used to examine variation in the rate of tongue-flicking in response to invertebrate prey and venomous and nonvenomous snakes in experiment 1. In experiment 2, Latin-square ANOVA was used to compare the percentage (arcsine square-root transformed) of hatchlings in each of the five groups that attacked snakes under
1778
PHILLIPS AND ALBERTS
FIG. I. Pamphagid grasshopper (approximately 50-ram body length) covered with a coat of skin from a homed adder.
the five treatment conditions. Differences in response to each of the three snake species were assessed using Fisher's protected least significant difference tests (Carmer and Swanson, 1973).
RESULTS AND DISCUSSION
In experiment 1, hatchlings exhibited tongue-flicking, a behavior that probably delivers odorants to the vomeronasal organ (Halpem, 1987), when exposed to invertebrate prey (Table 1). In most cases, two to four tongue-flicks were emitted before a prey item was attacked. However, most attacks on unrestrained grasshoppers occurred in the absence of tongue-flicking. Tongue-flicking rates were significantly lower in response to unrestrained moving grasshoppers than to the other invertebrate prey items presented (F = 132.48; df = 3,53; P -0.0001), suggesting that visual cues may be more important than chemical cues in eliciting attack of moving prey. In no instance were aggressive or defensive behaviors observed in response to live invertebrate prey. Sand snakes do not eat monitor lizards but are relatively common prey
1779
NAIVE LIZARDS IDENTIFY BY ODOR VENOMOUS SNAKES
TABLE 1. RESPONSES OF
27
HATCHLING SAVANNA MONITORS TO INVERTEBRATES AND W H O L E R O A D - K I L L E D SNAKES a
Invertebrates Land snail (54) Corn cricket (54) Restrained grasshopper (54) Unrestrained grasshopper (54) Snakes Sand snake (27) Spitting cobra (27) Horned adder (27)
Percent of hatchlings exhibiting defensive/ aggressive behaviors
Mean tongue flicks
Percent attack
4.33 4.35 3.94 0.18
100 100 100 100
0 0 0 0
0 0 0 0
0 0 0 0
0 0 0 0
4.27 4.84 4.55
100 22 11
0 100 100
0 93 100
0 52 81
0 100 0
Lateral
Hiss
Slap
Vent
aSample size in parentheses. Each invertebrate prey type was offered on two different occasions. Lateral compression of the body (lateral), hissing vocalizations (hiss), and tail slap (slap) are aggressive displays, whereas vent dragging (vent) is a territorial behavior.
items o f adult savanna monitors. The stomach contents o f 18 road-killed savanna monitors revealed two sand snakes, among other prey, accounting for about 10% o f the total caloric value o f the stomach contents (J.A. Phillips, personal observation). Sand snakes were universally attacked by hatchlings in a manner similar to invertebrate prey (Table 1). After initial attack, the hatchlings exhibited no hesitation in attempts to consume the sand snakes. No aggressive or defensive behaviors were noted in response to sand snakes. The spitting cobra and the homed adder are highly venomous snakes that eat hatchling monitor lizards (Branch, 1988). Similar to other potential prey items, tongue-flicking by hatchling lizards occurred immediately after exposure to these two species. There was no difference in the rate o f tongue-flicking in response to the three species o f snakes ( F = 1.16; df = 2,78; P = 0.32). However, in only nine o f 54 cases was attack o f venomous snakes observed (Table 1), and in no instance did the hatchlings attempt to consume venomous snake carcasses. Regardless o f whether snakes were attacked, all hatchlings exhibited one or more defensive and/or aggressive behaviors toward the venomous snakes (Table 1). The behavioral responses exhibited by hatchlings were specific to each o f the two venomous snake species. In particular, posterior adpression (vent dragging), apparent in every encounter with the spitting cobra, was never elicited when hatchlings were exposed to h o m e d adders. Hatchling savanna
1780
PHILLIPS AND ALBERTS
monitors defecate at latrine sites, and when exiting these sites, often perform vent dragging (J.A. Phillips, personal observation). When other hatchlings investigate areas where vent dragging has occurred, aggressive posturing is frequently observed, suggesting that vent dragging may function in territorial marking. In experiment 2, the hatchlings responded to whole bodies of the three snake species in a manner similar to that of experiment 1. Latin-square A N O V A revealed no effects of order of presentation (F = 1.001, df = 4,12, P > 0.50), no differences in response among the five hatchling groups (F = 1.001, df = 4,12, P > 0.50), and no differences in response across the five treatments (F = 1.001, df = 4,12, P > 0.50) for sand snakes (Figure 2). These results contrasted with those for the two venomous snake species. For spitting cobras, there were no effects of order of presentation (F = 0.56, df = 4,12, P > 0.50), no differences among hatchling groups (F = 0.28, df = 4,12, P > 0.50), but significant differences across treatments (F = 6.84, df = 4,12, P < 0.01). For homed adders, there were also no effects of order of presentation (F = 1.62, d f = 4,12, P > 0.20), no differences among hatchling groups (F = 0.78 df = 4, 12, P > 0.50), but significant differences across treatments (F = 48.10, df = 4,12, P < 0.001). For spitting cobras and homed adders, Fisher's protected least significant difference tests revealed that hatchlings were significantly more likely to attack skinned body sections and skin-covered unrestrained grasshoppers than whole snakes, sectioned snakes, or skin-covered restrained grasshoppers (Figure 2). Hatchlings avoided sectioned snakes and skin-covered restrained grasshoppers 100
~
8o
~
60
=N
40
N O Q.
20
9 [] []
sand snake spitting cobra horned adder
0
II
III Treatment
IV
FIG. 2. Mean percentage (+SE) of hatchling savanna monitors in five groups that attacked (I) whole road-killed snakes, (II) snake sections, (III) skinned snake sections, (IV) a restrained grasshopper covered with a segment of snake skin, or (V) an unrestrained grasshopper covered with a segment of snake skin. Three snake species were presented, the sand snake, and two highly venomous species, the spitting cobra and the homed adder.
NAIVE LIZARDS I D E N T I F Y BY O D O R V E N O M O U S SNAKES
1781
at a rate similar to that for whole snakes. Hatchlings attacked unrestrained grasshoppers covered with skin from either spitting cobras or homed adders in 49 of 54 trails. The five hatchlings that rejected skin-covered grasshoppers exhibited tongue-flicking, whereas all hatchlings that attacked grasshoppers did so without tongue-flicking. This indicates that chemical, rather than visual, cues associated with snake skins mediate avoidance of venomous snakes. Past studies with scincid lizards have shown higher levels of tongue-flicking toward snakes that are their predators than those that are not (Cooper, 1990). Skinks are insectivorous, thus harmless snakes represent neither predator nor prey and are not expected to elicit a strong response. The lack of a difference in tongue-flicking rate to the three snake species in the present study probably reflects the fact that all of the snake species tested interact with savanna monitors. Although spitting cobras and homed adders are predators and sand snakes are prey, predator detection and foraging are both ecological contexts in which tongue-flicking occurs (Burghardt, 1970; Halpem and Kubie, 1983; Simon, 1983; Duvall et al., 1990). Naive hatchling savanna monitors not only differentiated venomous from nonvenomous snakes in their propensity to attack them, they also reacted differently toward the two species of venomous snakes. Homed adders (Viperidae) and spitting cobras (Elapidae) possess divergent hunting strategies and chemically distinctive venoms (Visser, 1967). For these reasons, it is not unexpected that responses by their prey will differ. The tendency for some hatchlings to attack cobras and all to perform vent dragging behavior, which may be associated with aggression in savanna monitors, suggests that cobras may be less dangerous than adders, at least to these lizards. The long, stiff fangs of adders, capable of inflicting a deeper and more serious bite than those of cobras, may partially explain the differential treatment of the two predatory snake species. The rejection of invertebrate prey bearing skins of venomous snakes and the indiscriminant acceptance of skinned snake meat suggest that savanna monitor hatchlings use skin-borne chemicals to identify snakes. That lizards having no previous experience with snakes exhibit these responses indicates that they are innate. Because snakes represent both predators and prey for savanna monitors, the ability to discriminate among them potentially functions in two different adaptive contexts. Foraging efficiency is increased when appropriate prey items are identified prior to attack, and survivorship is enhanced when potential predators are detected and avoided with minimal physical contact. Previous experiments suggest that the movement of prey is important in the feeding behavior of reptiles (Drummond, 1985). Chemical cues also stimulate feeding behavior in snakes, and naive hatchlings recognize and attack their species-typical prey on the basis of chemical cues (Burghardt, 1966, 1968). Our results for monitor lizards indicate that visual cues associated with prey movement can override chemical cues. These visual responses may be learned, as
1782
PHILLIPS AND ALBERTS
the tested lizards had been fed invertebrate prey since hatching. A hierarchy of cues involving multiple sensory modalities and both learned and unlearned behavior is probably used by lizards in the recognition of sympatric snake species and assessment of their relative suitability as prey. Acknowledgments--We thank the Department of Nature Conservation of Namibia, and especially the Namibian scientists at Etosha National Park for their permission and help in carrying out this research. We are grateful to Janet Phillips for excellent technical help and to Joel Berger for useful advice on experimental design. Invaluable comments on earlier drafts of the manuscript were provided by Don Lindburg, Don Owings, and Paul Weldon. This work was supported by an NSF Conservation and Restoration Biology grant (BNS 90-00100) and the Institute of Museum Services (IC-00348-90).
REFERENCES AUFFENBERG,W. 1978. Social and feeding behavior in Varanus komodoensis, in pp. 301-331, N. Greenberg and P.D. MacLean (eds.). Behavior and Neurology of Lizards. U.S. Department of Health, Education, and Welfare, Washington, D.C. BERRY, H.H., and Louw, G.N. 1982. Nutritional balance between grassland productivity and large herbivore demand in the Etosha National Park. Madoqua 12: 141-150. BRANCH, W.R. 1988. South African Red Data Book: Reptiles and Amphibians. South African Scientific Programs Report, CSIR, Pretoria. BRANCH, W.R. 1991. The Regenia registers of 'Gogga' Brown (1869-1909) "Memoranda on a species of monitor or Varan." Mertensiella 2: 57-110. BURGItARDT, G.M. 1966. Stimulus control of the prey attack response in naive garter snakes. Psychon. Sci. 4:37-38. BURGHARDT,G.M. 1968. Chemical preference studies on newborn snakes of three sympatric species of Natrix. Copeia 1968:732-737. BUROHARDT, G.M. 1970. Chemical perception in reptiles, in pp. 241-308, J.W. Johnston, D,G. Moulton, and A. Turk, (eds.). Advances in Chemoreception. Appleton-Century-Crofts, New York. CARMER, S.G., and SWANSON, M.R. 1973. An evaluation of ten pairwise multiple comparison procedures by Monte Carlo methods. J. Am. Stat. Assoc. 68:66-74. COOPER, W.E., JR. 1989. Prey odor discrimination in the varanoid lizards Heloderma suspectum and Varanus exanthematicus. Ethology 81:250-258. COOPER, W.E., JR. 1990. Chemical detection of predators by a lizard, the broad-headed skink (Eumeces laticeps). J. Exp. Zool. 256:162-167. D~AL, B.E., WELDON, P.J., and CURTIS, B. 1989. Chemosensory identifieation of snake predators (Phyllorhynchus decurtatus) by banded geckos (Coleonyx variegatus). J. Herpetol. 23:224229. DRUMMOND, H. 1985. The role of vision in the predatory behaviour of natrieine snakes. Anita. Behav. 1985:206-215. DUVALL, D., CHISZAR,D., HAYES, W.K., LEONHARDT,J.K., and GOODE, M.J. 1990. Chemical and behavioral ecology of foraging in prairie rattlesnakes (Crotalus viridis viridis). J. Chem. Ecol. 16:87-101. HALPERN,M. 1987. The organization and function of the vomeronasal system. Annu. Rev. Neurosci. 10:325-362. I-IALVERN,M., and KUBIE,J.L. 1983. Snake tongue flicking behavior: Clues to vomeronasal system
NAIVE LIZARDSIDENTIFY BY ODOR VENOMOUSSNAKES
1783
functions, in pp. 45-72, D. Miiller-Schwarze and R.M. Silverstein (eds.). Chemical Signals in Vertebrates, Vol. 3. Plenum Press, New York. KING, D., and GREEN, B. 1979. Notes on the diet and reproduction of the sand goanna, Varanus gouldii rosenbergi. Copeia 1979:64-70. Losos, J.B., and GREENE, H.W. 1988. Ecological and evolutionary implications of diet in monitor lizards. Biol. J. Linn. Soc. 35:379-407. PIANKA, E.R. 1986. The Ecology and Natural History of Desert Reptiles. Princeton University Press, Princeton University Press, Princeton, New Jersey. SHINE, R. 1986. Food habits, habitats and reproductive biology of four sympatric species of varanid lizards in tropical Australia. Herpetologica 42:346-360. SHINE, R. 1989. Constraints, allometry, and adaptation: Food habits and reproductive biology of Australian brownsnakes (Pseudonaja: Elapidae). Herpetologica 45:195-207. ShiNE, R., and SLIP, D.J. 1990. Biological aspects of the adaptive radiation of Australasian pythons (Serpentes: Boidae). Herpetologiea 46:283-290. SIMON, C.A. 1983. A review of lizard chemoreeeption, in pp. 119-133, R.B. Huey, E.R. Pianka, and T.W. Schoener (eds.). Lizard Ecology: Studies of a Model Organism. Harvard University Press, Cambridge, Massachusetts. VISSER, J. 1967. Poisonous Snakes of Southern Africa and the Treatment of Snake Bite. Timmons, Cape Town. WELDON, P.J. 1982. Responses to ophiophagus snakes by snakes of the genus Thamnophis. Copeia 1982:788-794. WELbON, P.J. 1991. Responses by vertebrates to chemicals from predators, inpp. 500-521, D. MacDonald, D. Mfiller-Schwarze, and S.E. Natynczuk (eds.). Chemical Signals in Vertebrates, Vol. 5. Oxford University Press, Oxford. WELOON, P.J., and BURGHARDT,G.M. 1979. The ophiophage defensive response in crotaline snakes: Extension to new taxa. J. Chem. Ecol. 5:141-151. WELOON, P.J., and SCHELL,F.M. 1984. Responses by king snakes (Lampropeltis getulus) to chemicals from colubrid and crotaline snakes. J. Chem. Ecol. 10:1509-1520.
