Prey odor discrimination by ingestively naive coachwhip snakes(Masticophis flagellum)

86 Chemoecology 1 (1990) 86-91 Prey odor discrimination by ingestively naive coachwhip snakes (Masticophisflagellum) William E. Cooper, Jr. 1, Donald G. Buth 2, and Laurie J. Vitt 2' * Department of Biology, Auburn University at Montgomery, Montgomery, Alabama 36117, USA 2 Department of Biology, University of California at Los Angeles, Los Angeles, California 90024-1606, USA Received July 26, 1990 / Revision accepted November 22, 1990 Summary Ingestively naive hatchling coachwhip snakes (Masticophis flagellum) detected integumentary chemicals from several potential prey species and discriminated them from chemical stimuli from other animals and from distilled water, strongly suggesting a genetic basis for these abilities. The strongest responses were to lizard and snake stimuli, which form a major part of the diet. Variable responses to chemical cues from other taxa are discussed. Responses by coachwhip snakes to prey chemicals appear to be highly Introduction The ability to detect and recognize prey by chemical cues is highly developed in many squamate reptiles, especially snakes (e.g., Burghardt 1970a, 1977, 1980; Chiszar & Scudder 1980; Cooper 1989a,b; Cooper & Vitt 1989). Responsiveness to chemical prey stimuli is highly specific to prey type. For garter snakes, genus Thamnophis, detailed studies have demonstrated selective responsiveness to chemical stimuli derived from the integument of preferred prey and geographic variation in responsiveness correlated with variation in prey preferences (Burghardt 1969, 1970a; Arnold 1977, 1981a). A genetic contribution to these selective responses to chemical cues has been demonstrated for several snakes by studies of ingestively naive neonates and hatchlings and by estimation of heritability from comparisons among litters. Previously unfed newborn snakes of several colubrid species differ dramatically in response strength elicited by chemical cues from various prey types, each snake species reacting strongly to chemical cues from its own principal prey types (Burghardt 1966, 1967, 1968, 1975; Morris & Loop 1969; Burghardt & Abeshaheen 1971; Herzog & Burghardt 1974; Gove & Burghardt 1975; Henderson et al. 1983; Von Achen & Rakestraw 1984). Comparisons within and between litters of Thamnophis elegans revealed significant genetic variance. The heritabilities were fairly low (0.25-0.35) and showed little or no geographic variation (Arnold 1981a). specific, as suggested by the stronger reaction to vomodors of sympatric than of allopatric lizard species. The highly developed use of chemical cues by the diurnal, visually oriented eoachwhip snake emphasizes the general importance of chemical senses to predation by nonvenomous snakes, regardless of the involvement of vision. Key words prey odor, behavior, heritability, Reptilia, Squamata, Serpentes, Colubridae, Masticophis flagellum flicks its tongue, molecules that adhere to the lingual surface are carried into the mouth when the tongue returns to its sheath. The molecules arc carried to vomeronasal ducts at the roof of the mouth and thence to vomeronasal epithelia containing the chemosensory neurons (Halpern & Kubie 1980). Taste is probably not needed for prey odor discrimination prior to biting because the tongue is devoid of taste buds (Schwenk 1985) and by its movements would not appear to distribute chemicals widely over the oral epithelium. Olfaction is not necessary because blocking it does not affect chemical release of attack on prey in garter snakes (Burghardt & Hess 1968). Tongue-flicking behavior provides a bioassay of responsiveness to chemical stimuli, most likely those being assessed by vomerolfaction. The primary chemical sense snakes use to detect prey is vomerolfaction (Cooper & Burghardt 1990b), which is mediated by the vomeronasal organ. When a snake In this study we extend the findings regarding genetically programmed responsiveness to chemical stimuli arising from natural foods by juvenile snakes to a diurnal, visually oriented colubrid, the coachwhip snake, Masticophisflagellum. This snake is a fast, agile, diurnal, and actively foraging predator that is widely distributed in the southern United States and in Mexico (Wright & Wright 1957; Groves 1971; Stebbins 1985). Such a snake provides an interesting test of the generality of reliance on chemical cues among colubrids. If a species that is highly visually oriented in prey detection and pursuit is sensitive to chemical cues from prey, it seems likely that the reliance is universal or nearly so. The diet of M. flagellum includes lizards, snakes, insects (mainly orthopterans), small mammals (primarily rodents), birds, bird eggs, and occasional amphibians and young turtles (Wright & Wright 1957; Wylie 1974; Stebbins 1985). We examined the ability of ingestively naive hatchling M. flagellum to detect and respond selectively to chemical prey cues. Tongue-flicking © Georg Thieme Verlag Stuttgart • New York * Present address: Oklahoma Museum of Natural History, University of Oklahoma, Norman, Oklahoma 73019, USA Prey odor d i s c r i m i n a t i o n by coachwhip snakes and attack behavior were studied by presenting chemical stimuli, either integumentary chemical stimuli of various foods or controls, to snakes on cotton-tipped applicators. Material and methods Animals and maintenance Sixteen M. flagellum piceus (Van Denburgh) hatched on or about 3-4 September, 1988 from eggs of a lemale (red phase) collected in April, 1988 near Lake Perris in Riverside County, California. They were kept in a cage without food at the University of California, then transported to Auburn University at Montgomery, where they arrived on 21 September, 1988. Each was housed individually in a 31 × 17 x 10 cm plastic shoebox containing wood shavings (Aspen). After the first experiment and observations of responses to chemical stimuli derived from Cnernidophorus tigris and mice, an undercage heater afforded the opportunity for behavioral thermoregulation. The light cycle was the natural cycle for the vicinity supplemented by fluorescent lighting for 10 hours per day. Water was available ad libitum. Experimental methods and design Responses of the ingestively naive snakes to chemical stimuli from several potential food and control substances were investigated. The chemical stimuli were prepared as follows. First, the cotton tip of a 15 cm wooden applicator was dipped in distilled water (at ambient temperature). This produced the odorless control stimulus used to gauge responses to the experimental situation, inctuding the applicator and moisture. Next, the moistened cotton was rolled over the skin of a potential prey, including surfaces on the head, dorsum, and ventrum. In each trial, the top of the cage was lifted gently and the cotton tip of the applicator slowly moved to a position ca 1 cm anterior to the snake's snout. The number of tongue-flicks directed to the applicator in 60 seconds was recorded. In the first experiment, tongue-flicks were recorded only until the snake moved away. In subsequent work, if the snake moved slowly, apparently searching, the applicator was moved slowly in front of it and tongue-flicks continued to be recorded unless the snake fled. If a snake bit an applicator, the attack and its latency in seconds were recorded and the trial was terminated. If a snake moved away from the applicator, latency to the movement (s) was recorded. The same litter of snakes was used in all experiments and tests. Three experiments were performed to compare response to chemical stimuli from natural foods and other potential food items. In experiment 1 the snakes' responses to chemical stimuli from domestic cricket (Acheta domesticus), chicken (Gallus gallus), golden shiner (Notemigonus crysoleucas), an iguanid lizard (Callisaurus draconoides), and distilled water were tested in a randomized blocks (repeated measures) design. Only the lizard was likely to be a frequent natural food. Each snake responded once to each of the five stimuli in randomized sequence. No serious departures from counterbalancing occured: Numbers of trials on the first two test days were 8 for cricket, 7 for chicken, 4 for golden shiner, 5 for the lizard, and 6 for water. Trials were conducted between 1200 and 1600 h CST from 30 September to 6 0 c t o b e r , C h e m o e c o l o g y 1 (1990) 1988. Each snake was tested once per day. Ambient temperature was 25-26 °C, lower than optimal, but high enough that the snakes were alert and active. Experiment 2 tested responses to distilled water and to oders of several reptiles and amphibians: the colubrid snakes Chionactis occipitalis and Lampropeltis getulus californiae, the iguanid lizard Uta stansburiana, and the anuran Acris creptitans. All of these species except A. crepitans are sympatric with M. flagellum. The lizard and C. occipitalis are potential prey species; L. getülus is a predator of snakes. The experimental design and testing sequence and procedures were as in experiment 1. Although all conditions were not represented an equal number of times on each testing day, departures from equal representation did not favor increased response rates toward the stimuli that elicited the greatest responsiveness. Number of trials on the first two days were 5 for C. occipitalis, 2 for L. getulus, 5 for U. stansburiana, 5 for A. crepitans, and 6 for water. Ambient temperature was 27-28°C on the testing dates 10-14 October, 1989. In these and later tests, the snakes were fully active and warmer than air temperature (as revealed by several measurements of cloacal temperatures in the low 30 s°C immediately following trials). Experiment 3 was a preliminary test of the specificity of chemical responsiveness by snakes to sympatric species of lizards. Responses by the snakes to cgemical stimuli from a sympatric lizard, Callisaurus draconoides, and an allopatric lizard belonging to the same family, Anolis sagrei, were compared in experiment 3. Anolis and Callisaurus are confamilial, but only distantly related. No anoles occur near the coachwhip population from which the mother of ~~hehatchlings was collected. Thirteen snakes responded to each of the two stimuli in counterbalanced sequence (with the exception that the thirteenth snake responded first to the anole stimulus). Data were collected on 24 October, 1989 with approximately 1 h between successive trials for each snake. Ambient temperature was 27 °C, but cloacal temperatures of three individuals measured after the experiment were 30-32 °C. We additionally recorded responses to three other stimuli. Responses to chemical stimuli derived from the teiid lizard CnemMopkorus tigris were tested on 7 0 c t o b e r , 1989 and fl'om domestic mouse (Mus rnusculus) on 90ctober. Ambient temperature was 25°C. Responses to conspecific coachwhip snakes (littermates) were recorded on 17 October at an ambient temperature of 26 °C. Because air temperatures (Cooper & Vitt 1986), hunger levels, time between trials, and experience with applicators and odors varied among experiments and the other observations, results are directly comparable only within each experiment. Therefore statistical tests were limited to comparisons of responses to different stimuli within experiments, for which temperatures and intertrial intervals were similar and differences in experience and hunger were controlled by counterbalancing or randomization. Strict control of air temperatures proved impossible for the entire duration of experiments, but the temperatures were adequate for the snakes to be active and to allow detection of differential responses during each experiment. The undercage heaters were added before experiment two because it seemed likely that warmer snakes might have greater motivation to feed and thus 87 88 C h e m o e c o l o g y 1 (1990) C o o p e r et al. Responsesin experiment 1 by ingestively naive hatcMing Masticophis flage//um to [ntegumentarychemicals from the iguanid l[zard Callisaurus draconoides, the goldon shiner (Notemigonus crysoleucas), chicken, domestic cr[cket, and disti[led water. Rangesfor tongue-flick attack Table 1 score (TFAS) and tongue-flicks disagree because tongue-flicks are presented only for snakes that did not bite I TFAS ;timulus Mean SE Range Tongue-fiicks (60 s) Mean SF ùizard ;hiner ;hicken ;ricket Vater 34.3 14.1 6.6 4.7 5.3 8.4 6.7 1.9 1.0 1.6 10-107 3-107 1- 30 1-16 1 25 22.9 8.0 7.9 3.6 3.6 be more likely to bite applicators. Biting and perhaps tongueflicking frequencies might be expected to increase with hunger levels. However, because increasing experience with applicators not leading to feeding might counteract that effect, trends between observations and experiments were unpredictable. Data analysis Data for snakes that failed to tongue-flick in all five conditions were discarded due to a lack of vomeronasal activation. The primary response variable was tongue-flick attack score (TFAS), a composite variable combining influences of number of tongue-flicks, attack on the applicator, and latency to attack. Number of tongue-flicks in the 60 s test interval is a good bioassay of chemosensory response to the stimuli unless biting occurs. When a snake bites, however, the trial is terminated, precluding further tongue-flicks. Predatory attacks on applicators included extensive contact between the mouth and the cotton and prolonged holding of the applicator. These bites clearly differed from defensive strikes in which the snakes struck, rapidly released or did not bite, and again assumed a defensive posture. Only predatory bites were recorded. Biting at short latency represents rapid chemical discrimination, biting and latency are heavily weighted in TFAS, which provides an overall index of response strength. If a snake does not bite, TFAS is simply the number of tongueflicks in the trial. If it bites, TFAS = maximum number of tongue-flicks emitted by any snake in any trial of the entire study + (60 - latency to bite). TFAS has been used often and meets theoretical criteria for a desired measure of feeding response strength combining tongue-flicking and attack (e.g., Burghardt 1967, 1970a; Cooper & Burghardt 1990a). Additional variables analyzed were number of tongue-flicks and orientation time, the latter being the number of seconds from the beginning of the trial until the snake moved away from the applicator (maximum = 60 s). However, if the snake bit, an orientation time of 60 s was assigned arbitrarily for that trial. All TFAS and tongue-flick data were tested for homogeneity of variance by Hartley's Fmax tests prior to parametric analysis (Winer 1962). If variances were heterogeneous, the data were tested parametrically if logarithmic transformation yielded homogeneity. Otherwise, they were tested nonparametrically. For experiment 1, TFAS was tested by ANOVA for a single-variable experiment with repeated measures followed by individual comparisons by Newman-Keuls 8.t 1.7 3.3 0.9 1.1 Range Orientation time (s) Mean SE Range 13 50 4-18 2-30 1- 8 1-1 t 58.4 46.8 41.7 48.1 44.5 1.6 5.5 7.0 5.7 6,1 36-60 2-60 3-60 2 60 4 60 Bites 3 1 0 0 0 procedures (Winer 1962). Number.of tongue-flicks by snakes that completed all trials without fleeing or attacking (n = 8) and orientation time (n = 7 after elimination of individuals for which all ranks were tied) were examined by Friedman twoway anova followed by multiple comparison tests (Hollander & Wolfe 1973). For experiment 2, statistical tests were as in experiment 1 except that a sign test was used in addition for comparison of orientation times (Siegel 1956). For experiment 3, differences in number of tongue-flicks among stimuli were tested by a t test for correlated (paired) samples (Winer 1962). TFAS and orientation time data were tested using the Wilcoxon matched-pairs signed-ranks test (Siegel 1956). All tests were two-tailed unless otherwise noted, with alpha = 0.05. The statistical tests must be considered to represent the larger population only approximately due to the likelihood that behaviotal responses of littermates are correlated. Results Experiment 1 Fifteen snakes responded in all conditions. Although seven of these moved away from the applicator before the end of the trial in one or more conditions, there were pronounced differences in responses among stimuli. All 15 snakes tongue-flicked in all conditions. Four snakes bit applicators, three bearing stimuli from C. draconoides and one from N. crysoleucas. TFAS varied substantially among conditions (Table 1). Variances for TFAS (n= 15) were heterogeneous (Fmax = 77.76; d f = 5, 14; P < 0.01), but were rendered homogeneous by logarithmic transformation (Fmax= 1.71; d l = 5, 14; P > 0.10). The chemical stimulus effect was highly significant (F = 14.62; d f = 4 , 52; P < 0.001). Individual comparisons revealed significantly greater response to chemical stimuli from C. draconoides than to the other stimuli ( P < 0.01 each). In addition, shiner odors elicited significantly higher TFAS than did distilled water or cricket odors (P<0.05). No other differences were significant. Numbers of tongue-flicks (Table 1) were elevated in response to chemical stimuli derived from C. draconoides. Tongue-flick rate showed a rank ordering identical to that for TFAS except that the rates were the same for cricket stimuli and distilled water. Analysis of tongue-flicks among individuals that completed all five trials without attacking or moving away from the applicator by a Friedman tests revealed signifieant variation among conditions (X~=18.83, d f = 4 , P<0.001). The only significant differences among pairs of conditions were that odors of C. draconoides elicited a signif- P r e y o d o r d i s c r i m i n a t i o n b y c o a c h w h i p snakes icantly higher tongue-flicking rate than did cricket stimuli (P<0.05), chicken stimuli (P<0.05), and distilled water (P<0.001). The rank difference between response to C. draconoides and minnow stimuli approached, but did not attain, statistical significance despite the similarity in mean numbers of tongue-flicks for minnow and chicken stimuli (P>0.05). Orientation times varied little among stimuli (Table 1). Eight of the 15 individuals oriented to all stimuli for the entire 60-second trial. Even after their data were eliminated, there was no difference in orientation time among stimuli (X 2 = 6.66, d f = 4, P > 0.10). Numbers of bites did not differ between any pair of stimuli. However, all four bites were elicited by the two stimuli having the greatest mean TFAS and number of tongue-flicks. If bites are equiprobable for all stimuli, the total proportion of bites delivered to any two stimuli is 0.40. Thus, the binomial probability that all four bites would be directed to the two stimuli that elicited the most tongue-flicks is 0.026 (one-tailed test justified by the expectation that repeated tongue-flicking and biting represent different levels of responsiveness to food stimuli). Experiment 2 Eleven individuals responded in all conditions, but all of them moved away from the applicator in less than than 60 s in at least one condition. All snakes tongueflicked in all conditions, but none attacked a swab. Response strength differed markedly among stimuli (Table 2). In the absence of biting, TFAS and number of tongue-flicks were identical. Variances of these variables were homogeneous (Fmax = 3.26; d f = 5, 10; P > 0.05). TFAS varied significantly among conditions (F = 8.07; d f = 4, 40; P < 0.005), the strongest responses being elicited by the three reptilian stimuli. TFAS did not differ significantly between distilled water and chemical stimuli from A. crepitans, but was significantly greater for odors of L. getulus and U. stansburiana than for either distilled water or A. crepitans (all P < 0 . 0 1 except the comparison between distilled water and U. stansburiana, for which P < 0.05). Chemical stimuli derived from C. occipitalis elicited significantly greater TFAS than did those from A. crepitans ( P < 0.01) and distilled water ( P < 0.05, by a one-tailed test which is appropriate because water was the odorless control stimulus). C h e m o e c o l o g y 1 (1990) Table 2 Responsesin experiment 2 by ingestively naive hatchling Masticophis flagellum to integumentary chemicats from the iguanid lizard Uta stansburiana, the snakes Chionactis occipitalis and Lamprooeltis getulus, the flog Acris crepitans, and distilled water. In the absence of bites, Tongue-flick attack score (TFAS) and number of tongue-flicks are identical Orientation Tongue-flicks time Is) (60 s) Mean SE Range Mean SE Stimulus Ran e i Uta stansburiana Chionactis occipitalis Lamprope/tis getu/us Acris crepitans Water 47.1 37.7 46.7 14.0 23.5 7.0 7.6 6.3 42 616 9-72 6-69 6-70 1-39 2-71 62.6 56.7 56.3 37.2 49.6 4.7 3.3 2.2 75 518 1024- i3 P<0.005). For number of tongue-flicks, the variances were homogeneous (Fmax = 2.10; df = 2, 12; P > 0.05). Significantly more tongue-flicks were emitted in response to odors of C. draconoides (t = 3.62, d f = 12, P < 0 . 0 1 ) . Orientation time differed significantly between conditions (T = 1, n = 7, P < 0.05), being greater in response to odors of C. draconoides. However, this difference would disappear if snakes that bit the applicator were not assigned the 60 s orientation time. Four snakes bit applicators bearing chemical stimuli from C. draconoides; none bit in response to anole stimuli ( P = 0.06, sign test). Addifional Observations Odors of the teiid lizard C. tigris eticited substantial responses. One snake attacked an applicator. TFAS values were higher than those elicited by any nonreptilian stimulus (30.1+7.1, range 10-95). This was true also for tongue-flicks (24.2 + 3.9, range 10-51). Orientation time was (57.6+ 1.9, range 37-60). Mouse odors elicited no bites, relatively low TFAS and number of tongue-flicks (14.7-+ 1.9, range = 9-25 each), and short orientation times (47.2-+4.9, range 14-60). Responses by coachwhip snakes to conspecific chemicals stimuli were also relatively weak for the temperature studied, with no bites, low TFAS and number of tongueflicks (21.8-+ 4.9, range 2-58 each), and short orientation time (50.2-+4.5, with range 14-60). Discussion Orientation times varied little among the reptilian chemical stimuli and water, but were somewhat lower for the frog stimuli (Table 2). All eleven snakes abandoned the applicator before 60 s in at least one condition. The numbers of snakes having orientation times less than 60 s for the various stimuli were 6 for A. crepitans, 1 for C. occipitalis, and 3 each for the others. Only the difference in orientation time between A. crepitans and C. occipitalis approached significance ( P = 0.06, 1-tailed sign test). Experiment 3 Chemical stimuli from the sympatric lizard C. draconoides elicited much stronger predatory responses than did those derived from the allopatric lizard A. sagrei (Table 3). Variances for TFAS were heterogeneous (Fmax = 8.68; d f = 2, 12; P < 0 . 0 1 ) . A nonparametric test demonstrated that TFAS was significantly greater for stimuli from C. draconoides than from A. sagrei (Wilcoxon's T = 5 , n = 1 3 , Discrimination of chemieal prey stimuli The coachwhip snakes responded most strongly to lizard odors in experiment 1 and to the chemical stimuli from three reptilian species, including two snakes and one lizard, in experiment 2 (Fig. 1). These findings for TFAS and number of tongue-flicks were consistent in experiments 1 and 2. Because data for stimuli labelled with an asterisk in Figure 1 were collected at lower temperature than the rest, somewhat lower tongue-flicking rates and TFAS are expected for those data than for data collected at higher temperature (Cooper & Vitt 1986). Thus, the snakes also showed strong response, including an attack, to odors of another sympatric lizard, C. tigris. The six chemical stimuli eliciting the greatest TFAS were from sympatric squamate reptiles, including five different species. That squamates form a large part of the diet of M. flagellum (Wright & Wright 1957; Wylie 1974) confirms an unlearned sensitivity to integumentary chemical stimuli from natural food taxa. 89 90 C o o p e r et al. C h e m o e c o l o g y 1 (1990) Table 3 Responsesin experiment 3 by ingestively naive hatchling Masticophisflage//umto integumentary chemicals from two liizard species, the sympatric Callisaurusdraconoides(CD) and the allopatric Ano/issagrei(AS) T7AS Stimulus Mean SE Range Tongue-flicks (60 s) Mean SE CD AS 52.9 12.7 9,5 3.2 9-107 3- 47 28.7 12.7 4.7 3.2 Range SE Range 6-56 3 47 52.9 41.2 3.6 7.1 19-60 4-60 4 0 is not known to eat fish in the field. Coachwhip snakes from xeric habitats in the Southwest probably rarely encounter fish. Although not statistically comparable to other responses, response to mouse stimuli was weaker than to reptilian stimili, but similar to those accorded golden shiner stimuli. It may represent positive chemosensory investigation, but is merely suggestive. Achetadomest/cus*L Distilled woter, exp.1" Ga#usgellus* Anol/ssagte~ Acris crep/tans Notem/~onuscrysoleüca$~ Musmusculus* Mast/¢oph/sflagellum L\ \\\\\\\\\\\\\ \\\ \ ~ I \\\\\\\\\\~ Distilled water, exp. 2 IX\\\\\\\ / ù ~ Cnem/~ophorüs tlgr/s L\\\\\\'~ \\\\\~\\\\\\\\\\\~ Chionactisocc/~o/Tol/:~F\\\\\~ Lampropeltisgetu/us k\\\\\~ Ut« stonsbur/ona L\\\\\~ X~~~~\\\\\\\\\\\\\\\\\\\\~ Callisourusdroconoides,exp. I*[\\\\\\'~ \\\\\\\\\\\\\\\N 1 Call/saurüsdracono/des,exp. 3 I[~, ~, \"\ \- \, \\ \\\ \ \ \ \ \ \ \ ~ \ \ \ \ \ \ \ N i i o Bites Orientation time(s) Mean I 10 , l l 20 l , l 30 TFAS l l 40 i 50 I 60 Fig. 1 Tongue-flickattack scores (TFAS) elidted by chemicaJ stimuli from various species and control stimuli by ingestiveiy naive hatchling Masticophisflagel/um*- indicates that data were cot[ected at 25-26 °C. Snakes were several degrees warmer in the remaining tests Although bites were infrequent, the almost exclusive direction of bites to stimuli from sympatric lizards indicates strong response to these stimuli, consistent with the findings for tongue-flieks and TFAS. Orientation time was a rauch less sensitive and reliable index of response strength than either tongue-flicking or TFAS. Responses to odors of C. occipitalis were presumably predatory. However, motivation underlying the high tongue-flicking rates to chemicals ffom the ophiophagous L. getulus could indicate a specific sensitivity to predator odors. In two species of gatter snakcs, genus Thamnophis (Weldon 1982), ehemical stimuli flora L. getulus induce increased tongue-flicking rates. Crotaline snakes perform defensive body-bridging behavior in response to chemical cues from ophiophagous snakes (Bogert 1941; Weldon & Burghardt 1979). Failure of juvenile coachwhip snakes to show a pronounced tongue-flicking response to conspecific chemical stimuli suggests that the snakes did not respond to them as food stimuli. However, this is merely suggestive because the data are not statistically comparable to other responses. Among the remaining odor sources, only the golden shiner and mouse evoked potentially important reactions. The response to fish odor may be a laboratory artifact, probably due to high concentration or odor intensity, because M. flagellum The unresponsiveness by hatchling M. flagellum to frog odors agrees with the infrequency of anurans in the diet, but somewhat greater responses to chemical stimuli of chicken, mouse, and especially cricket were expected from dietary data for Texas populations (Wylie 1974). Possible explanations include 1) stimu]ation of initial predatory episodes for birds, small mammals, and orthopterans visua]ly or by a combination of chemical and visual cues, 2) scarcity of these taxa in the diet of the California population, and 3) specificity of response by coachwhips to chemical aspects of local representatives of these taxa. Genetic contribution to prey odor discrimination The strong responses to chemical stimu]i (presumably vomodors - perceived qualities of stimuli detected by vomerolfaction - Cooper & Burghardt 1990b) from lizards and certain snakes by M. flagellum that had never eaten nor observed others eating suggest that inital responsiveness to chemical stimuli from these taxa is genetica]ly based. This finding agrees with that for all colubrid speeies studied (cited above). Because the initia! response to C. draconoides odors in experiment 1 was rauch stronger than that to anole stimuli in experiment 3, despite the lower temperature and hunger level, the difference in responsiveness to anole and C. draconoides odors in experiment 3 cannot be attributed to prior experience with the latter odor. If prior experience without feeding had any effect, it would have been expected to decrease TFAS to C. draconoides odors. The stronger reaction of coachwhip snakes to chemical stimuli from the sympatric C. draconoides than from the allopatric A. sagrei is compatible with data showing geographie variation in chemosensory response correlated with local diets (Burghardt 1970b; Arnold 198la,b). Although naive hatchlings have the capacity to respond selectively to chemical stimuli of reptilian prey, the responses by hatehlings to chemical cues may differ from those of older snakes due to either learning or ontogeny. Some colubrids have the eapacity to rapidly acquire strong responses to ehemical aspects of other prey that may be encountered during an individual's lifetime (Fuchs & Burghardt 1971; Arnold 1978; see also Burghardt 1978). It is also possib]e that hatchlings might (adaptively) not respond to chemical stimuli from food eaten by adults, yet too large or dangerous for hatchlings to consume. Prey odor discrimination by coachwhip snakes Relative importance o f visuai and ehemical eues in snake predation The coachwhip snake is generally considered to be one o f the m o s t visually oriented N o r t h A m e r i c a n colubrids, hut it is quite sensitive to chemical prey cues, as are another highly visual terrestrial species, C o l u b e r constrictor (Herzog & B u r g h a r d t 1974), and two arboreal diurnal species, O p h e o d r y s vernalis (Burghardt 1967) a n d U r o m a c e r f r e n a t u s ( H e n d e r s o n et al. 1983). Natricines are highly chemically oriented, but even they r e s p o n d to visual as well as chemical prey cues (Arnold 1978; Burghardt & D e n n y 1983; D r u m m o n d 1985). Thus, colubrids studied to date rely o n b o t h chemical and visual cues in predation. Chemical prey cues may be i m p o r t a n t to p r e d a t i o n by colubrids regardless o f the i m p o r t a n c e o f visual cues in foraging, o f diel activity cycle, and o f degree o f terrestriality or arboreality. Acknowledgements Partial s u p p o r t for this work was provided by the Biology D e p a r t m e n t s at out institutions. G o r d o n M. Burghardt read the first d r a f t o f the m a n u s c r i p t and provided valuable c o m m e n t s . We are grateful to M a r k Arvizu who provided the gravid female coachwhip snake and to Allison Collins-Rainboth and Steve Secor for their assistance. References Arnold SJ (1977) Polymorphism and geographic variation in the feediing behavior of the gartet snake Thamnophis elegans. Science 197:676678 Arnold SJ (1978) Some effects of early experience on feeding responses in the common garter snake, Thamnophis sirtalis. Anim Behav 26:455462 Arnold SJ (1981a) Behavioral variation in natural populations. I. Phenotypic, genetic and environmental correlations between chemoreceptive responses to prey in the garter snake, Thamnophis elegans. Evolution 35:489-509 Arnold SJ (1981b) The microevolution of feeding behavior. Pp 409-453 in Kamil AC & Sargent TD (eds) Foraging Behavior: Ecological, Ethological, and Psychological Approaches. New York: Garland STPM Press Bogert CM (1941) Sensory cues used by rattlesnakes in their recognition of ophidian enemies. Ann NY Acad Sci 41:329-343 Burghardt GM (1966) Stimulus control of the prey attack response in naive garter snakes. Psychon Sci 4:37-38 Burghardt GM (1967) Chemical-cue preferences of inexperienced snakes: comparative aspects. Science 157:718-721 Burghardt GM (1968) Chemical preference studied on newborn snakes of three sympatric species of Natrix. Copeia 1968:732-737 Burghardt GM (1969) Comparative prey-attack studies in newborn snakes of the genus Thamnophis. Behaviour 33:77-114 Burghardt GM (1970a) Chemical perception in reptiles. Pp 241-308 in Johnston JW, Monlton DG, Turk A (eds) Advances in Chemoreception, Vol. I. Communication by Chemical Signals. New York: Appleton-Century-Crofts Burghardt GM (1970b) Intraspecific geographical variation in chemical food cue preferences of newborn garter snakes (Tharnnophis sirtalis). Behaviour 36:246-257 Burghardt GM (1975) Chemical prey preference polymorphism in newborn gartet snakes Thamnophis sirtalis. Behaviour 52:202-225 Burghardt GM (1977) The ontogeny, evolution, and stimulus control of fedding in human and reptils. Pp 253-275 in Kare M, Maller O (eds) The Chemical Senses and Nutrition. New York: Academic Press Burghardt GM (1978) Behavioral ontogeny in reptiles: whence, whither, and why? Pp 149-174 in Burghardt GM, Bekoff M (eds) The Develop- Chemoecology 1 (1990) ment of Behavior: Comparative and Evolutionary Aspects. New York: Garland STPM Press Burghardt GM (1980) Behavioral and stimulus correlates of vomeronasal functioning in reptiles: feeding, grouping, sex, and tongue use. Pp 275-301 in Müller-Schwarze D, Silverstein RM (eds) Chemical Signals: Vertebrates and Aquatic Invertebrates. New York: Plenum Press Burghardt GM, Abeshaheen JP (1971) Responses to chemical stimuli of prey in newly hatched snakes of the genus Elaphe. Anim Behav 19:486-489 Burghardt GM, Denny D (1983) Effects of prey movement and prey odor on feeding on gatter snakes. Z Tierpsychol 62:329-347 Burghardt GM, Hess EH (1968) Factors influencing the chemical release of prey attack in newborn snakes. J Comp Physiol Psychol 66:289295 Chiszar D, Scudder KM (1980) Chemosensory searching by rattlesnakes during predatory episodes. Pp 125-139 in Müller-Schwarze D, Silverstein RM (eds) Chemical Signals: Vertebrates and Aquatic Invertebrates. New York: Plenum Press Cooper WE Jr (1989a) Absence of prex odor discrimination by iguanid and agamid lizards in applicator tests. Copeia 1989:472-478 Cooper WE Jr (1989b) Prey odor discrimination in the varanoid lizards Heloderma suspectum and Varanus exanthematicus. Ethology 81 : 250258 Cooper WE Jr, Burghardt GM (1990a) A comparative analysis of scoring methods for chemical discrimination of prey by squamate reptiles. J Chem Ecol 16:45-65 Cooper WE Jr, Burghardt GM (1990b) Vomerolfaction and vomodor. J Chem Ecol 16:103-105 Cooper WE Jr, Vitt LJ (t986) Thermal dependence of tongue-flicking and comments on use of tongue-flicking as an index of squamate behavior. Ethology 71:177-186 Cooper WE Jr, Vitt LJ (1989) Prey odor discrimination by the broad-headed skink, (Eumeces laticeps). J Exp Zool 249:i1-16 Drummond H (1985) The role of vision in the predatory behaviour of natricine snakes. Anim Behav 33: 206-215 Fuchs JL, Burghardt GM (1971) Effects of early feeding experience on the responses of gatter snakes to food chemicals. Learn Motiv 2:271-279 Gove D, Burghardt GM (1975) Responses of ecologically dissimilar populations of the water snake, Natrix s. sipedon, to chemical cues from prey. J Chem Ecol 1:25-40 Groves JD (1971) Correlation of certain ophidian sensory modalities with gross brain proportions. J. Herpetol 5:200-204 Halpern M, Kubie JL (1980) Chemical access to vomeronasal organs of gartet snakes. Physiol Behav 24:367-371 HHendersonRW, Binder MH, Burghardt GM (1983) Responses of neonate hispaniolan vine snakes (Uromacerfrenatus) to prey extracts. Herpetologica 39:75-77 Herzog HA Jr, Burghardt GM (1974) Prey movement and predatory behavior of juvenile western yellow-belIiedracers, Coluber constrictor mormon. Herpetologica 30:285-289 Hollander M, Wolfe DA (1973) Nonparametric Statistical Methods. New York: John Wiley and Sons Morris DD, Loop MS (1969) Stimulus control of prey attack in naive rat snakes: a species duplication. Psychon Sci 15:141-142 Schwenk K (1985) On the occurrence, distribution, and functional significance of taste buds in lizards. Copeia 1985:99-101 Siegel S (1956) Nonparametric Statistics for the Behavioral Sciences. New York: McCraw-Hill Stebbins RC (1985) A Field Guide to Western Reptiles and Amphibians, second edition. Boston: Houghton Mifflin Co Von Achen PH, Rakestraw JL (1984) The role of chemoreception in the prey selection of neonate reptfles. Pp 163-172 in Seigel RA, Hunt LE, Knight JL, Zuschlag NL (eds) Vertebrate Ecology and Systematics - a Tribute of Henry S. Fitch. Lawrence, Kansas: University of Kansas, Museum of Natural History Weldon PJ (1982) Responses to ophiophagous snakes by snakes of the genus Thamnophis. Copeia 1982:788-794 Weldon P J, Burghardt GM (1979) The ophiophage defensive response in crotaline snakes: extension to new taxa. J Chem Ecol 5:141-151 Winer BJ (1962) Statistical Principles in Experimental Design. New York: McCraw-Hill Wright AH, Wright AA (1957) Handbook of Snakes of the United States and Canada, vol. 1. Ithaca, New York: Cornelt Univ. Press Wylie SR (1974) A comparison of the ecological niches of coachwhips, Masticophisflagellum (Shaw), and racers, Coluber «onstrietor Linnaeus, in Texas. Unpublished M.A. thesis, Univ. Texas at Arlington 91