- Email: [email protected]
Behavioural responses of Mallorcan midwife toad tadpoles to natural and unnatural snake predators
Anim. Behav., 1998, 55, 207–214
Behavioural responses of Mallorcan midwife toad tadpoles to natural and unnatural snake predators RICHARD A. GRIFFITHS*, LAURENT SCHLEY*, PENNY E. SHARP*, JAYNE L. DENNIS* & ALVARO ROMA u N† *The Durrell Institute of Conservation & Ecology, Research School of Biosciences, University of Kent at Canterbury †Fons Ferreret, Campanet, Mallorca (Received 14 April 1997; initial acceptance 12 May 1997; final acceptance 11 June 1997; MS. number: 5517)
Abstract. The activity levels of Mallorcan midwife toad tadpoles, Alytes muletensis, were compared in two natural torrent pools which differed in their use by predatory viperine snakes, Natrix maura. Activity levels were lower in a pool regularly used by snakes than they were in a snake-free pool, but were reduced in both pools when snakes were experimentally introduced in nylon bags. In the presence of snakes, however, activity was more suppressed in the pool that was usually snake-free. Corresponding reductions in activity were also observed when tadpoles were treated with chemical cues from Mallorcan N. maura in a gravitational flow-through system. However, tadpoles failed to respond to chemical cues from other species of amphibian-eating snakes, or even to those from N. maura collected from a different population in mainland Spain. As none of the snakes used had previously eaten midwife toads, the responses cannot be related to previous diet, and seem to be specific to those N. maura from the island of Mallorca. As viperine snakes were probably introduced to Mallorca about 2000 years ago, the evolution of anti-predator behaviour in midwife toad tadpoles must have occurred relatively ? 1998 The Association for the Study of Animal Behaviour recently. Prey organisms possess a wide variety of behavioural strategies to minimize the risk of being eaten (Lima & Dill 1990). In amphibians, these behaviours include forming aggregations, shifting habitat use, increasing time in refuges and reducing activity (e.g. Kats et al. 1988; Sih et al. 1988; Griffiths & Denton 1993; Jackson & Semlitsch 1993). There may be costs associated with such behaviours, however, such as reductions in foraging efficiency and survival (Lawler 1989; Skelly 1992). The effectiveness of anti-predator strategies therefore relies on the prey being able to distinguish between real predators and those that pose little or no threat. Visual, tactile or chemical cues all provide information concerning the identity of a potential predator, and may be used in isolation or in combination, depending on the Correspondence: R. A. Griffiths, DICE, Research School of Biosciences, University of Kent, Canterbury CT2 7NJ, U.K. (email: [email protected]). L. Schley is now at the School of Biological Sciences, University of Sussex, Falmer, Brighton, Sussex BN1 9QG, U.K. 0003–3472/98/010207+08 $25.00/0/ar970596
environment. In murky water, for example, chemical cues would seem to provide the most reliable information (Magurran 1989), and this is borne out by studies on a variety of aquatic organisms which have shown that predators can be identified using chemosensory information alone (e.g. Weldon 1990; Dodson et al. 1994; Chivers et al. 1996). Although chemical cues may provide information that is predator-specific, studies of fish, amphibians and invertebrates have shown that the chemicals produced depend upon the diet of the predator. Prey may therefore respond only to those predators labelled with ‘alarm pheromone’, acquired through consuming conspecifics of the prey (Howe & Harris 1978; Crowl & Covich 1990; Mathis & Smith 1993). Although the basis of these responses seems to be genetic (e.g. Kats et al. 1988; Van Damme et al. 1995), a degree of conditioning to the chemical stimulus may also be needed, as several species possess inherited traits that are subsequently reinforced through experience (e.g. Magurran 1990; Semlitsch & Reyer
? 1998 The Association for the Study of Animal Behaviour
207
208
Animal Behaviour, 55, 1
1992; Mathis & Smith 1993). Indeed, naive prey may not respond at all to a natural predator unless the alarm pheromone is also present (Magurran 1989). Island faunas frequently provide unique opportunities for the study of the evolution of interactions between predators and their prey. The effects of introduced animals on island-dwelling species that have otherwise evolved in isolation from predators are often devastating. Lacking any natural defence strategies against the predators, such populations may be driven to extinction (e.g. Atkinson 1989; Diamond 1989). Equally, introduced predators may provide a strong selection pressure under which new strategies that minimize the risk of predation may evolve. The endemic midwife toad of Mallorca, Alytes muletensis, was known only from fossil evidence and thought to be extinct until a few toads were rediscovered in the remote gorges in the northwest of the island in the late 1970s (Mayol & Alcover 1981). A number of introduced predators may have contributed to its decline. Foremost among these is the amphibian-eating viperine snake, Natrix maura. Although the Quaternary fauna of the Balearic islands is well known, no fossil snakes have ever been found on Mallorca. Present-day snake populations are therefore probably descendants of those introduced to Mallorca by the Romans for religious purposes (Alcover & Mayol 1981). Viperine snakes may therefore have contributed to the eradication of Mallorcan midwife toads from much of lowland Mallorca, such that the toads are now confined to mountain gorges which are less accessible to snakes (Tonge 1986). Although there is a successful captive breeding and conservation programme for the toad (Tonge & Bloxam 1991; Bloxam & Tonge 1995), viperine snakes are still considered a threat to toads in a number of areas. This raises the question of whether the Mallorcan midwife toad has evolved defence strategies to minimize the risk of predation by snakes during the 2000 years or so since snakes were introduced to the island. In this paper, we describe the results of experiments that tested the behavioural responses of Mallorcan midwife toad tadpoles to snake predators in both natural pools and in the laboratory. Specifically, we test the hypotheses that (1) toad tadpoles have evolved behaviours that minimize the risk of predation by introduced snakes; (2) these responses are chemically mediated; and (3)
these responses are specific to viperine snakes that have been introduced to Mallorca.
METHODS Field Experiments We carried out field observations of the responses of tadpoles to snakes at two natural torrent pools in the Serra de Tramuntana, northwest Mallorca, during May–June 1996. Pool A was discovered as a toad breeding site in 1993, and no snakes have been observed in the area during 34 subsequent visits to the site. This pool was located at the base of a dry waterfall at an altitude of about 770 m. It measured 4 m in diameter and shelved to a maximum depth of 150 cm at the base of the waterfall. At Pool B viperine snakes were observed preying on tadpoles up to and including the day before observations started. This pool was situated about 8.5 km northwest of Pool A at an altitude of 300 m, and was also located in an otherwise dry torrent bed. The pool measured 8 m long by 2 m wide and had a maximum depth of 70 cm. Both pools had clear water and bare rock substrates, and were devoid of vegetation in the shallow areas. Pool A contained about 740 tadpoles and Pool B 310 tadpoles (Schley 1996). All tadpoles were free-swimming, feeding independently, and had a mean total length& of 46.0&7.85 mm (Pool A) and 41.8&13.33 mm (Pool B; N=100 in both cases). We used two methods to monitor tadpole activity in each pool. First, we monitored the activity of 10 randomly chosen tadpoles by measuring the amount of time each individual spent swimming during 1 min. This measure (‘individual tadpole activity’) was expressed as the mean number of seconds spent swimming per min. Second, we placed a green plastic transect measuring 50 cm long by 1.5 cm wide on the bottom of the pool, and counted the number of tadpoles crossing the transect during a 10-min period. Transect crossings were expressed as the percentage of the total number of tadpoles in the pool swimming across the transect per min. We carried out observations over a period of 6 days at each pool, at 1700 and 1900 hours, as previous work has shown that tadpoles are most active at these times (Schley 1996). Analyses were
Griffiths et al.: Responses of tadpoles to snakes performed on the mean values of the two observation periods. For the first 3 days, we monitored tadpole activity in the absence of viperine snakes. We then repeated the observations after deliberately conditioning the water in each pool with cues from viperine snakes. This was done by placing four viperine snakes in nylon mesh bags and suspending them in the middle of the pool (making sure that the snakes could surface for air). The mesh bags allowed visual, chemical and tactile cues from the snakes to pervade the water body, but prevented direct predation of the tadpoles by the snakes. We placed the bags in the pool at 1200 hours and removed them immediately after observations were completed at 1900 hours. Laboratory Experiments To investigate the nature and specificity of cues provided by snakes, we compared the responses of naive tadpoles to chemical cues from different snake predators in the laboratory. We conducted four experiments, each time using a different snake predator to condition the water. These were (1) Mallorcan N. maura, (2) Spanish N. maura, (3) English N. natrix, and (4) Asiatic Elaphe rufodorsata. All of these snakes are semi-aquatic and feed on amphibians and fish in the wild. The specimens used had all been collected at least 2 months before the experiments, and were fed dead fish and ranid frog tadpoles in captivity. The Mallorcan N. maura were collected from an area that did not contain midwife toads, so it is unlikely that they had consumed the latter even prior to capture. To obtain snake-conditioned water, we placed the snakes individually in water reservoirs at least 2 h before the experiments, and removed them after completing the observations on tadpole behaviour (see below). All Mallorcan midwife toad tadpoles used in the experiments were from a population that had been bred in captivity for several generations and had no previous experience of predators. We tested tadpoles individually in circular pans containing 650 ml of water which were connected to a gravitational flow-through system, similar to that described by Semlitsch & Gavasso (1992). Reservoirs containing 9 litres of either unconditioned or snake-conditioned water were positioned on a shelf above the test chambers, which they fed at a flow rate of 0.13 litres/min. We tested
209
tadpoles in two sequential trials, and changed the source of water flow between trials. We initially treated tadpoles with unconditioned water, and then changed the water source to either (1) snakeconditioned water (‘snake-test’), or (2) more unconditioned water from a different reservoir (‘control’). Water flow from the different reservoirs was controlled using a series of valves which allowed an uninterrupted flow of water between tests and avoided any disturbance to the tadpoles. We observed the tadpoles for two 10-min periods in each experiment. Initially, we observed all the tadpoles in unconditioned water (having first allowed them to acclimatize to flowing water conditions for 5 min), and we then repeated the observations under either ‘snake-test’ or ‘control’ conditions. We observed the tadpoles every 30 s for 10 min, and scored each individual as either actively swimming, immobile on the bottom or immobile on the surface or side. Thus each trial lasted approximately 25 min. The scores were converted to the percentage time engaged in each of the three behaviours during 10 min, and we compared behaviour before and after transfer to snake-test water using a two-tailed Wilcoxon signed-ranks test. In the first round of experiments we randomly allocated tadpoles to either snake-test or control conditions. All tadpoles that were initially used as controls were observed again under snake-test conditions and vice versa, so in total we observed each tadpole for four 10-min periods. Thirty-two tadpoles were used in each trial. No tadpoles were tested more than once with a snake predator, and no tadpoles were tested against different snake predators. Experiments were carried out between 1300 and 1700 hours in a laboratory subject to natural photoperiods and at a temperature of 18–22)C. The tadpoles used were all freeswimming and feeding independently, and measured 43.9&9.63 mm (X&; N=125). Ethical Note All tadpoles used in the experiments were bred in captivity as part of a wider conservation and research programme for the Mallorcan midwife toad. The design of these experiments was arrived at after detailed discussions with the conservation authorities in Mallorca. They gave full support to the work as it resulted in information that would help them manage the toad sites more effectively.
Animal Behaviour, 55, 1
In the field, there were no obvious effects on tadpole survival. In the laboratory, all tadpoles survived the experiments and were returned to individual holding boxes at the end of the tests. Although all three species of snakes used in the experiments are highly aquatic, nylon bags and water reservoirs were designed so that they could surface for air at all times. The snakes were left in the experimental enclosures for a maximum of 8 h (field experiments) or 3 h (laboratory experiments), and at the end of the tests were kept in captivity for further observational work. None of the snake species used are endangered.
20
(a)
18 Tadpole activity (s/min)
210
16 14 12 10 8 6 4 2 0
RESULTS Responses of Tadpoles to Snakes in Natural Pools Tadpole activity levels were much lower in Pool B (used regularly by snakes) than in Pool A (snake-free). In the absence of snakes, the tadpoles spent approximately 27% and 16% of the time actively swimming in Pools A and B, respectively. When snakes were experimentally introduced into the two pools, activity fell dramatically in both cases, but the degree of the change differed between the two pools. In Pool A, activity was reduced by over 50%, compared with a less than 25% reduction in Pool B (Fig. 1). These trends were confirmed by analysis of variance; significant pool#snake interactions indicated that the response to the introduction of snakes differed between the pools (Table I). The two methods of assessing tadpole activity gave very similar results and were strongly correlated (r10 =0.84, P<0.001). Specificity of Cues Produced by Snakes The tadpoles usually spent less than 30% of the time swimming, and spent most time immobile on the bottom or sides of the pans. However, tadpoles treated with water conditioned by Mallorcan N. maura reduced their activity and increased their immobility on the bottom (Table II). No such responses were observed when tadpoles were treated with water conditioned by N. maura from mainland Spain, or with water conditioned by more distantly related amphibianeating snakes. Control experiments in which tadpoles were tested twice in unconditioned water
Crossings per min (%)
0.5
(b)
0.4 0.3 0.2 0.1 0
Before After Pool A
Before After Pool B
Figure 1. The effect of viperine snakes on tadpole activity in Pools A (snakes absent) and B (used regularly by snakes). Activity before and after snakes were introduced is expressed as (a) mean number of seconds active per min, N=10; and (b) percentage of the total number of tadpoles crossing the transect per min. Bars show X& across 3 days of observations.
revealed no changes in tadpole behaviour. Thus changes in behaviour cannot have been a function of sequential testing. Immobility on the surface or sides was unaffected by chemical cues from snakes (all Wilcoxon signed-ranks tests: ). DISCUSSION In natural ponds and streams, aquatic organisms are presented with a cocktail of chemical cues from potential predators and food sources. This may explain why very few studies have demonstrated anti-predator behaviours in natural aquatic systems (but see Petranka et al. 1987; Kats et al. 1988; Feminella & Hawkins 1994;
Griffiths et al.: Responses of tadpoles to snakes Table I. Summary of analyses of variance comparing tadpole activity in Pools A (snakes absent) and B (used regularly by snakes), and before and after the introduction of viperine snakes df Pools Snakes Pools#snakes Error Total
1 1 1 8 11
Individual activity
Transect crossings
11.4** 139.5*** 71.5***
8.0* 23.4** 10.8*
Activity was scored as mean number of seconds active per min, N=10 (‘individual activity’); and percentage of the total number of tadpoles crossing the transect per min (‘transect crossings’). Transect crossings were converted to proportions and arcsine transformed prior to analysis. Values shown are F-ratios. *P<0.05; **P<0.01; ***P<0.001.
Holumuzki 1995). In the present study, the depression in activity of midwife toad tadpoles in response to viperine snakes in the field corresponded with the results obtained in the laboratory. This suggests that the predator signal must have been very strong under natural conditions. The results also showed that the behavioural response of tadpoles to snake predators can be chemically mediated and is specific to those snakes that have been introduced to the island of Mallorca. Moreover, as none of the tadpoles tested had experienced predators before, the behaviours observed must have been inherited rather than learnt through previous encounters with predators. Previous research has shown that toad tadpoles respond only to those predator species that they would naturally experience in the wild (Kiesecker et al. 1996). As in other species, however, the response may depend upon the predator having previously consumed the prey species and therefore being chemically labelled by it (e.g. Wilson & Lefcort 1993). In the present experiments all of the snakes used were fed a similar diet that did not include Mallorcan midwife toads, and it is very unlikely that the N. maura from Mallorca had previously fed on this species. The results therefore show that Mallorcan midwife toads display inherited anti-predator behaviour that is independent of the diet of the predator. Moreover, the lack of response to N. maura from the Spanish mainland shows that the behaviour is not only
211
specific to viperine snakes, but also specific to the viperine snakes on the island of Mallorca. This raises the question of the nature of the chemical cue produced by snakes. Snakes release pheromones through the skin and cloaca, and lizards can distinguish between different species of snake on the basis of skin chemicals alone (Dial et al. 1989). These chemicals include non-volatile lipids which may be long lasting (Crews & Garstka 1982). Skin lipids vary considerably in composition between different vertebrate taxa, and even between individuals of the same species. This means that different snakes may possess their own unique ‘chemical signature’ (Mason 1992). Clearly, further work is needed to determine whether such chemical signatures are populationspecific. Although the Mallorcan midwife toad is endemic to Mallorca, three closely related species occur on the Iberian peninsula where they too may be subject to predation by viperine snakes. Although the responses of these species to snakes are unknown, it is possible that anti-snake behavioural responses were present in the ancestral toads when they first colonized the Balearics from the Iberian peninsula, which was probably about 4–5 million years ago (Arntzen & Garcia-Paris 1995). If so, the behaviour would have needed to be retained within the Mallorcan population for several million years prior to the introduction of viperine snakes around 2000 years ago. As reduced activity incurs a cost in the form of reduced growth and development (e.g. Lawler 1989; Skelly 1992), it seems unlikely that such anti-predator responses could have been retained for so long in the absence of selection pressures. Moreover, the lack of response to viperine snakes from mainland Spain supports the view that the anti-predator response is an adaptation that has evolved since the introduction of viperine snakes to Mallorca. Although anti-predator behaviour may contribute to the persistence of Mallorcan midwife toads in the face of snake predation, how can such a trait persist in a population such as at Pool A, where there appear to be no snakes? Viperine snakes may range widely, and the colonization of a midwife toad breeding pool by snakes may largely be due to chance. Indeed, it may take only a single snake to eliminate an entire cohort of tadpoles from one pool. Mallorcan midwife toads are excellent climbers, and can negotiate vertical
Swimming Immobile on bottom
Swimming Immobile on bottom
Swimming Immobile on bottom
Swimming Immobile on bottom
Mallorcan Natrix maura
Spanish Natrix maura
English Natrix natrix
Asiatic Elaphe rufodorsata
23 31
24 27
28 28
25 31
8.8 58.6
8.5 62.0
29.1 44.4
15.2 57.1
Unconditioned water
12.4 56.1
10.3 59.1
22.3 37.5
8.9 72.2
Snakeconditioned water
75 235
125 147
117 150
49* 69**
Wilcoxon T
23 25
23 28
27 29
24 28
N for test
11.3 49.9
7.4 59.2
20.6 50.8
17.9 59.2
Unconditioned water
10.0 51.0
8.2 61.5
20.3 46.7
17.4 54.9
Unconditioned water
Control
111 156
124 181
165 172
135 140
Wilcoxon T
Data show the mean percentage time spent engaged in two behaviours: (1) before and after switching to snake-conditioned water (Snake-test conditions); and (2) before and after switching between two sources of unconditioned water (Control conditions). In each of the four experiments, 32 tadpoles were used but N for test is less than 32 as a result of tied data. *P<0.01, **P<0.001 by the Wilcoxon signed-ranks test (two-tailed). All other paired comparisons are not significant.
Behaviour
Snake
N for test
Snake-test
Table II. Effect of snake-conditioned water on tadpole behaviour in a gravitational flow-through system
212 Animal Behaviour, 55, 1
Griffiths et al.: Responses of tadpoles to snakes cliffs and waterfalls, which are physical barriers to colonization of some toad pools by snakes. These snake-free pools may be able to provide toads that subsequently recolonize pools where snake predation has resulted in local extinction. Equally, if there is reciprocal transfer of toads from snakeaccessible pools to snake-free pools, anti-predator behaviour is likely to be retained within the metapopulation as a whole, even if predation is not occurring within a particular sub-population (e.g. Pool A). From a conservation point of view, it is reassuring that Mallorcan midwife toads have retained natural anti-predator behaviour through several generations of captive breeding. Captivebred toads that are reintroduced to the wild should therefore be able to minimize predation risk from snakes by continuing to express such behaviours. Owing to its precarious status, however, removal of viperine snakes from breeding pools must continue to be part of the population management strategy for the Mallorcan midwife toad. ACKNOWLEDGMENTS This work was supported by AAT-Garten und-Teichfreunde Luxemburgs; by a NERC Advanced Research Fellowship to Richard Griffiths and by an ASAB vacation scholarship to Penny Sharp. Jersey Wildlife Preservation Trust, Sarah Bush and Leigh Gillett provided captive-bred tadpoles; Stuart Worth and Derek Thompson provided snakes from Spain and England, respectively. Sarah Bush, Jose-Maria Roma´n, Alex Forteza, Pere Vicenc, Joan Mayol and the Consellaria de Medi Ambient provided logistical support for field work in Mallorca. Ernest Mendoza assisted with the translation of foreign papers, and Leigh Gillett and two anonymous referees provided helpful comments which improved the manuscript. REFERENCES Alcover, J. A. & Mayol, J. 1981. Espe`cies relı´quies d’amfibis i de re`ptils a les Balears i Pitiu¨ses. Boll. Soc. Hist. Nat. Balears., 25, 151–167. Arntzen, J. W. & Garcia-Paris, M. 1995. Morphological and allozyme studies of midwife toads (genus Alytes), including the description of two new taxa from Spain. Contrib. Zool., 65, 5–34.
213
Atkinson, I. 1989. Introduced animals and extinctions. In: Conservation for the Twenty-first Century (Ed. by D. Western & M. Pearl), pp. 54–75. Oxford: Oxford University Press. Bloxam, Q. M. C. & Tonge, S. J. 1995. Amphibians: suitable candidates for breeding-release programmes. Biodiv. Conserv., 4, 636–644. Chivers, D. P., Wisenden, B. D. & Smith, R. J. F. 1996. Damselfly larvae learn to recognize predators from chemical cues in the predator’s diet. Anim. Behav., 52, 315–320. Crews, D. & Garstka, W. R. 1982. The ecological physiology of a garter snake. Scient. Am., 247, 158– 168. Crowl, T. A. & Covich, A. P. 1990. Predator-induced life-history shifts in a freshwater snail. Science, 247, 949–951. Dial, B. E., Weldon, P. & Curtis, B. 1989. Chemosensory identification of snake predators (Phyllorhynchus decurtatus) by banded geckos (Coleonyx variegatus). J. Herpetol., 23, 224–229. Diamond, J. 1989. Overview of recent extinctions. In: Conservation for the Twenty-first Century (Ed. by D. Western & M. Pearl), pp. 37–41. Oxford: Oxford University Press. Dodson, S. I., Crowl, T. A., Peckarsky, B. L., Kats, L. B., Covich, A. P. & Culp, J. M. 1994. Non-visual communication in freshwater benthos: an overview. J. N. Am. benthol. Soc., 13, 268–282. Feminella, J. W. & Hawkins, C. P. 1994. Tailed frog tadpoles differentially alter their feeding behavior in response to non-visual cues from four predators. J. N. Am. benthol. Soc., 13, 310–320. Griffiths, R. A. & Denton, J. 1993. Interspecific associations in tadpoles. Anim. Behav., 44, 1153–1157. Holomuzki, J. R. 1995. Oviposition sites and fishdeterrent mechanisms of two stream anurans. Copeia, 1995, 607–613. Howe, N. R. & Harris, L. G. 1978. Transfer of the sea anemone pheromone, anthopleurine, by the nudibranch Aeolidia papillosa. J. Chem. Ecol., 4, 551–561. Jackson, M. E. & Semlitsch, R. D. 1993. Paedomorphosis in the salamander Ambystoma talpoideum: effects of a fish predator. Ecology, 74, 342–350. Kats, L. B., Petranka, J. W. & Sih, A. 1988. Antipredator defenses and the persistence of amphibian larvae with fishes. Ecology, 69, 1865–1870. Kiesecker, J. M., Chivers, D. P. & Blaustein, A. R. 1996. The use of chemical cues in predator recognition by western toad tadpoles. Anim. Behav., 52, 1237– 1245. Lawler, S. P. 1989. Behavioural responses to predators and predation risk in four species of larval anurans. Anim. Behav., 38, 1039–1047. Lima, S. L. & Dill, L. M. 1990. Behavioral decisions made under the risk of predation: a review and prospectus. Can. J. Zool., 68, 619–640. Magurran, A. E. 1989. Acquired recognition of predator odour in the European minnow (Phoxinus phoxinus). Ethology, 82, 216–223. Magurran, A. E. 1990. The inheritance and development of minnow anti-predator behaviour. Anim. Behav., 39, 834–842.
214
Animal Behaviour, 55, 1
Mason, R. 1992. Reptilian pheromones. In: Biology of the Reptilia. 18. Hormones, Brain and Behavior (Ed. by C. Gans & D. Crews), pp. 114–228. Chicago: University of Chicago Press. Mathis, A. & Smith, R. J. F. 1993. Fathead minnows, Pimephales promelas, learn to recognize northern pike, Esox lucius, as predators on the basis of chemical stimuli from minnows in the pike’s diet. Anim. Behav., 46, 645–656. Mayol, J. & Alcover, J. A. 1981. Survival of Baleaphryne Sanchiz and Adrover, 1979 (Amphibia: Anura: Discoglossidae) on Mallorca. Amphib.-Reptil., 3/4, 343–345. Petranka, J. W., Kats, L. B. & Sih, A. 1987. Predatorprey interactions among fish and larval amphibians: use of chemical cues to detect predatory fish. Anim. Behav., 35, 420–425. Schley, L. 1996. Activity patterns of tadpoles and anti-predator behaviour of tadpoles and postmetamorphic Mallorcan midwife toads Alytes muletensis. M.Sc. thesis, University of Kent. Semlitsch, R. D. & Gavasso, S. 1992. Behavioural responses of Bufo bufo and Bufo calamita tadpoles to chemical cues of vertebrate and invertebrate predators. Ethol. Ecol. Evol., 4, 165–173. Semlitsch, R. D. & Reyer, H.-U. 1992. Modification of anti-predator behaviour in tadpoles by environmental conditioning. J. Anim. Ecol., 61, 353–360.
Sih, A., Petranka, J. W. & Kats, L. B. 1988. The dynamics of prey refuge use: a model and tests with sunfish and salamander larvae. Am. Nat., 132, 463–483. Skelly, D. K. 1992. Field evidence for a cost of behavioral antipredator response in a larval amphibian. Ecology, 73, 704–708. Tonge, S. 1986. Collecting the Mallorcan midwife toad. Oryx, 20, 74–78. Tonge, S. J. & Bloxam, Q. M. C. 1991. The breeding programme for the Mallorcan midwife toad Alytes muletensis at the Jersey Wildlife Preservation Trust. Dodo, 27, 146–156. Van Damme, R., Bauwens, D., Thoen, C., Vanderstighelen, D. & Verheyen, R. F. 1995. Responses of naive lizards to predator chemical cues. J. Herpetol., 29, 38–43. Weldon, P. J. 1990. Responses by vertebrates to chemicals from predators. In: Chemical Signals in Vertebrates, vol. 5 (Ed. by D. W. MacDonald, D. Muller-Schwarze & R. M. Silverstein), pp. 501–521. New York: Plenum Press. Wilson, D. J. & Lefcort, H. 1993. The effect of predator diet on the alarm response of red-legged frog, Rana aurora, tadpoles. Anim. Behav., 46, 1017–1019.
Behavioural responses of Mallorcan midwife toad tadpoles to natural and unnatural snake predators RICHARD A. GRIFFITHS*, LAURENT SCHLEY*, PENNY E. SHARP*, JAYNE L. DENNIS* & ALVARO ROMA u N† *The Durrell Institute of Conservation & Ecology, Research School of Biosciences, University of Kent at Canterbury †Fons Ferreret, Campanet, Mallorca (Received 14 April 1997; initial acceptance 12 May 1997; final acceptance 11 June 1997; MS. number: 5517)
Abstract. The activity levels of Mallorcan midwife toad tadpoles, Alytes muletensis, were compared in two natural torrent pools which differed in their use by predatory viperine snakes, Natrix maura. Activity levels were lower in a pool regularly used by snakes than they were in a snake-free pool, but were reduced in both pools when snakes were experimentally introduced in nylon bags. In the presence of snakes, however, activity was more suppressed in the pool that was usually snake-free. Corresponding reductions in activity were also observed when tadpoles were treated with chemical cues from Mallorcan N. maura in a gravitational flow-through system. However, tadpoles failed to respond to chemical cues from other species of amphibian-eating snakes, or even to those from N. maura collected from a different population in mainland Spain. As none of the snakes used had previously eaten midwife toads, the responses cannot be related to previous diet, and seem to be specific to those N. maura from the island of Mallorca. As viperine snakes were probably introduced to Mallorca about 2000 years ago, the evolution of anti-predator behaviour in midwife toad tadpoles must have occurred relatively ? 1998 The Association for the Study of Animal Behaviour recently. Prey organisms possess a wide variety of behavioural strategies to minimize the risk of being eaten (Lima & Dill 1990). In amphibians, these behaviours include forming aggregations, shifting habitat use, increasing time in refuges and reducing activity (e.g. Kats et al. 1988; Sih et al. 1988; Griffiths & Denton 1993; Jackson & Semlitsch 1993). There may be costs associated with such behaviours, however, such as reductions in foraging efficiency and survival (Lawler 1989; Skelly 1992). The effectiveness of anti-predator strategies therefore relies on the prey being able to distinguish between real predators and those that pose little or no threat. Visual, tactile or chemical cues all provide information concerning the identity of a potential predator, and may be used in isolation or in combination, depending on the Correspondence: R. A. Griffiths, DICE, Research School of Biosciences, University of Kent, Canterbury CT2 7NJ, U.K. (email: [email protected]). L. Schley is now at the School of Biological Sciences, University of Sussex, Falmer, Brighton, Sussex BN1 9QG, U.K. 0003–3472/98/010207+08 $25.00/0/ar970596
environment. In murky water, for example, chemical cues would seem to provide the most reliable information (Magurran 1989), and this is borne out by studies on a variety of aquatic organisms which have shown that predators can be identified using chemosensory information alone (e.g. Weldon 1990; Dodson et al. 1994; Chivers et al. 1996). Although chemical cues may provide information that is predator-specific, studies of fish, amphibians and invertebrates have shown that the chemicals produced depend upon the diet of the predator. Prey may therefore respond only to those predators labelled with ‘alarm pheromone’, acquired through consuming conspecifics of the prey (Howe & Harris 1978; Crowl & Covich 1990; Mathis & Smith 1993). Although the basis of these responses seems to be genetic (e.g. Kats et al. 1988; Van Damme et al. 1995), a degree of conditioning to the chemical stimulus may also be needed, as several species possess inherited traits that are subsequently reinforced through experience (e.g. Magurran 1990; Semlitsch & Reyer
? 1998 The Association for the Study of Animal Behaviour
207
208
Animal Behaviour, 55, 1
1992; Mathis & Smith 1993). Indeed, naive prey may not respond at all to a natural predator unless the alarm pheromone is also present (Magurran 1989). Island faunas frequently provide unique opportunities for the study of the evolution of interactions between predators and their prey. The effects of introduced animals on island-dwelling species that have otherwise evolved in isolation from predators are often devastating. Lacking any natural defence strategies against the predators, such populations may be driven to extinction (e.g. Atkinson 1989; Diamond 1989). Equally, introduced predators may provide a strong selection pressure under which new strategies that minimize the risk of predation may evolve. The endemic midwife toad of Mallorca, Alytes muletensis, was known only from fossil evidence and thought to be extinct until a few toads were rediscovered in the remote gorges in the northwest of the island in the late 1970s (Mayol & Alcover 1981). A number of introduced predators may have contributed to its decline. Foremost among these is the amphibian-eating viperine snake, Natrix maura. Although the Quaternary fauna of the Balearic islands is well known, no fossil snakes have ever been found on Mallorca. Present-day snake populations are therefore probably descendants of those introduced to Mallorca by the Romans for religious purposes (Alcover & Mayol 1981). Viperine snakes may therefore have contributed to the eradication of Mallorcan midwife toads from much of lowland Mallorca, such that the toads are now confined to mountain gorges which are less accessible to snakes (Tonge 1986). Although there is a successful captive breeding and conservation programme for the toad (Tonge & Bloxam 1991; Bloxam & Tonge 1995), viperine snakes are still considered a threat to toads in a number of areas. This raises the question of whether the Mallorcan midwife toad has evolved defence strategies to minimize the risk of predation by snakes during the 2000 years or so since snakes were introduced to the island. In this paper, we describe the results of experiments that tested the behavioural responses of Mallorcan midwife toad tadpoles to snake predators in both natural pools and in the laboratory. Specifically, we test the hypotheses that (1) toad tadpoles have evolved behaviours that minimize the risk of predation by introduced snakes; (2) these responses are chemically mediated; and (3)
these responses are specific to viperine snakes that have been introduced to Mallorca.
METHODS Field Experiments We carried out field observations of the responses of tadpoles to snakes at two natural torrent pools in the Serra de Tramuntana, northwest Mallorca, during May–June 1996. Pool A was discovered as a toad breeding site in 1993, and no snakes have been observed in the area during 34 subsequent visits to the site. This pool was located at the base of a dry waterfall at an altitude of about 770 m. It measured 4 m in diameter and shelved to a maximum depth of 150 cm at the base of the waterfall. At Pool B viperine snakes were observed preying on tadpoles up to and including the day before observations started. This pool was situated about 8.5 km northwest of Pool A at an altitude of 300 m, and was also located in an otherwise dry torrent bed. The pool measured 8 m long by 2 m wide and had a maximum depth of 70 cm. Both pools had clear water and bare rock substrates, and were devoid of vegetation in the shallow areas. Pool A contained about 740 tadpoles and Pool B 310 tadpoles (Schley 1996). All tadpoles were free-swimming, feeding independently, and had a mean total length& of 46.0&7.85 mm (Pool A) and 41.8&13.33 mm (Pool B; N=100 in both cases). We used two methods to monitor tadpole activity in each pool. First, we monitored the activity of 10 randomly chosen tadpoles by measuring the amount of time each individual spent swimming during 1 min. This measure (‘individual tadpole activity’) was expressed as the mean number of seconds spent swimming per min. Second, we placed a green plastic transect measuring 50 cm long by 1.5 cm wide on the bottom of the pool, and counted the number of tadpoles crossing the transect during a 10-min period. Transect crossings were expressed as the percentage of the total number of tadpoles in the pool swimming across the transect per min. We carried out observations over a period of 6 days at each pool, at 1700 and 1900 hours, as previous work has shown that tadpoles are most active at these times (Schley 1996). Analyses were
Griffiths et al.: Responses of tadpoles to snakes performed on the mean values of the two observation periods. For the first 3 days, we monitored tadpole activity in the absence of viperine snakes. We then repeated the observations after deliberately conditioning the water in each pool with cues from viperine snakes. This was done by placing four viperine snakes in nylon mesh bags and suspending them in the middle of the pool (making sure that the snakes could surface for air). The mesh bags allowed visual, chemical and tactile cues from the snakes to pervade the water body, but prevented direct predation of the tadpoles by the snakes. We placed the bags in the pool at 1200 hours and removed them immediately after observations were completed at 1900 hours. Laboratory Experiments To investigate the nature and specificity of cues provided by snakes, we compared the responses of naive tadpoles to chemical cues from different snake predators in the laboratory. We conducted four experiments, each time using a different snake predator to condition the water. These were (1) Mallorcan N. maura, (2) Spanish N. maura, (3) English N. natrix, and (4) Asiatic Elaphe rufodorsata. All of these snakes are semi-aquatic and feed on amphibians and fish in the wild. The specimens used had all been collected at least 2 months before the experiments, and were fed dead fish and ranid frog tadpoles in captivity. The Mallorcan N. maura were collected from an area that did not contain midwife toads, so it is unlikely that they had consumed the latter even prior to capture. To obtain snake-conditioned water, we placed the snakes individually in water reservoirs at least 2 h before the experiments, and removed them after completing the observations on tadpole behaviour (see below). All Mallorcan midwife toad tadpoles used in the experiments were from a population that had been bred in captivity for several generations and had no previous experience of predators. We tested tadpoles individually in circular pans containing 650 ml of water which were connected to a gravitational flow-through system, similar to that described by Semlitsch & Gavasso (1992). Reservoirs containing 9 litres of either unconditioned or snake-conditioned water were positioned on a shelf above the test chambers, which they fed at a flow rate of 0.13 litres/min. We tested
209
tadpoles in two sequential trials, and changed the source of water flow between trials. We initially treated tadpoles with unconditioned water, and then changed the water source to either (1) snakeconditioned water (‘snake-test’), or (2) more unconditioned water from a different reservoir (‘control’). Water flow from the different reservoirs was controlled using a series of valves which allowed an uninterrupted flow of water between tests and avoided any disturbance to the tadpoles. We observed the tadpoles for two 10-min periods in each experiment. Initially, we observed all the tadpoles in unconditioned water (having first allowed them to acclimatize to flowing water conditions for 5 min), and we then repeated the observations under either ‘snake-test’ or ‘control’ conditions. We observed the tadpoles every 30 s for 10 min, and scored each individual as either actively swimming, immobile on the bottom or immobile on the surface or side. Thus each trial lasted approximately 25 min. The scores were converted to the percentage time engaged in each of the three behaviours during 10 min, and we compared behaviour before and after transfer to snake-test water using a two-tailed Wilcoxon signed-ranks test. In the first round of experiments we randomly allocated tadpoles to either snake-test or control conditions. All tadpoles that were initially used as controls were observed again under snake-test conditions and vice versa, so in total we observed each tadpole for four 10-min periods. Thirty-two tadpoles were used in each trial. No tadpoles were tested more than once with a snake predator, and no tadpoles were tested against different snake predators. Experiments were carried out between 1300 and 1700 hours in a laboratory subject to natural photoperiods and at a temperature of 18–22)C. The tadpoles used were all freeswimming and feeding independently, and measured 43.9&9.63 mm (X&; N=125). Ethical Note All tadpoles used in the experiments were bred in captivity as part of a wider conservation and research programme for the Mallorcan midwife toad. The design of these experiments was arrived at after detailed discussions with the conservation authorities in Mallorca. They gave full support to the work as it resulted in information that would help them manage the toad sites more effectively.
Animal Behaviour, 55, 1
In the field, there were no obvious effects on tadpole survival. In the laboratory, all tadpoles survived the experiments and were returned to individual holding boxes at the end of the tests. Although all three species of snakes used in the experiments are highly aquatic, nylon bags and water reservoirs were designed so that they could surface for air at all times. The snakes were left in the experimental enclosures for a maximum of 8 h (field experiments) or 3 h (laboratory experiments), and at the end of the tests were kept in captivity for further observational work. None of the snake species used are endangered.
20
(a)
18 Tadpole activity (s/min)
210
16 14 12 10 8 6 4 2 0
RESULTS Responses of Tadpoles to Snakes in Natural Pools Tadpole activity levels were much lower in Pool B (used regularly by snakes) than in Pool A (snake-free). In the absence of snakes, the tadpoles spent approximately 27% and 16% of the time actively swimming in Pools A and B, respectively. When snakes were experimentally introduced into the two pools, activity fell dramatically in both cases, but the degree of the change differed between the two pools. In Pool A, activity was reduced by over 50%, compared with a less than 25% reduction in Pool B (Fig. 1). These trends were confirmed by analysis of variance; significant pool#snake interactions indicated that the response to the introduction of snakes differed between the pools (Table I). The two methods of assessing tadpole activity gave very similar results and were strongly correlated (r10 =0.84, P<0.001). Specificity of Cues Produced by Snakes The tadpoles usually spent less than 30% of the time swimming, and spent most time immobile on the bottom or sides of the pans. However, tadpoles treated with water conditioned by Mallorcan N. maura reduced their activity and increased their immobility on the bottom (Table II). No such responses were observed when tadpoles were treated with water conditioned by N. maura from mainland Spain, or with water conditioned by more distantly related amphibianeating snakes. Control experiments in which tadpoles were tested twice in unconditioned water
Crossings per min (%)
0.5
(b)
0.4 0.3 0.2 0.1 0
Before After Pool A
Before After Pool B
Figure 1. The effect of viperine snakes on tadpole activity in Pools A (snakes absent) and B (used regularly by snakes). Activity before and after snakes were introduced is expressed as (a) mean number of seconds active per min, N=10; and (b) percentage of the total number of tadpoles crossing the transect per min. Bars show X& across 3 days of observations.
revealed no changes in tadpole behaviour. Thus changes in behaviour cannot have been a function of sequential testing. Immobility on the surface or sides was unaffected by chemical cues from snakes (all Wilcoxon signed-ranks tests: ). DISCUSSION In natural ponds and streams, aquatic organisms are presented with a cocktail of chemical cues from potential predators and food sources. This may explain why very few studies have demonstrated anti-predator behaviours in natural aquatic systems (but see Petranka et al. 1987; Kats et al. 1988; Feminella & Hawkins 1994;
Griffiths et al.: Responses of tadpoles to snakes Table I. Summary of analyses of variance comparing tadpole activity in Pools A (snakes absent) and B (used regularly by snakes), and before and after the introduction of viperine snakes df Pools Snakes Pools#snakes Error Total
1 1 1 8 11
Individual activity
Transect crossings
11.4** 139.5*** 71.5***
8.0* 23.4** 10.8*
Activity was scored as mean number of seconds active per min, N=10 (‘individual activity’); and percentage of the total number of tadpoles crossing the transect per min (‘transect crossings’). Transect crossings were converted to proportions and arcsine transformed prior to analysis. Values shown are F-ratios. *P<0.05; **P<0.01; ***P<0.001.
Holumuzki 1995). In the present study, the depression in activity of midwife toad tadpoles in response to viperine snakes in the field corresponded with the results obtained in the laboratory. This suggests that the predator signal must have been very strong under natural conditions. The results also showed that the behavioural response of tadpoles to snake predators can be chemically mediated and is specific to those snakes that have been introduced to the island of Mallorca. Moreover, as none of the tadpoles tested had experienced predators before, the behaviours observed must have been inherited rather than learnt through previous encounters with predators. Previous research has shown that toad tadpoles respond only to those predator species that they would naturally experience in the wild (Kiesecker et al. 1996). As in other species, however, the response may depend upon the predator having previously consumed the prey species and therefore being chemically labelled by it (e.g. Wilson & Lefcort 1993). In the present experiments all of the snakes used were fed a similar diet that did not include Mallorcan midwife toads, and it is very unlikely that the N. maura from Mallorca had previously fed on this species. The results therefore show that Mallorcan midwife toads display inherited anti-predator behaviour that is independent of the diet of the predator. Moreover, the lack of response to N. maura from the Spanish mainland shows that the behaviour is not only
211
specific to viperine snakes, but also specific to the viperine snakes on the island of Mallorca. This raises the question of the nature of the chemical cue produced by snakes. Snakes release pheromones through the skin and cloaca, and lizards can distinguish between different species of snake on the basis of skin chemicals alone (Dial et al. 1989). These chemicals include non-volatile lipids which may be long lasting (Crews & Garstka 1982). Skin lipids vary considerably in composition between different vertebrate taxa, and even between individuals of the same species. This means that different snakes may possess their own unique ‘chemical signature’ (Mason 1992). Clearly, further work is needed to determine whether such chemical signatures are populationspecific. Although the Mallorcan midwife toad is endemic to Mallorca, three closely related species occur on the Iberian peninsula where they too may be subject to predation by viperine snakes. Although the responses of these species to snakes are unknown, it is possible that anti-snake behavioural responses were present in the ancestral toads when they first colonized the Balearics from the Iberian peninsula, which was probably about 4–5 million years ago (Arntzen & Garcia-Paris 1995). If so, the behaviour would have needed to be retained within the Mallorcan population for several million years prior to the introduction of viperine snakes around 2000 years ago. As reduced activity incurs a cost in the form of reduced growth and development (e.g. Lawler 1989; Skelly 1992), it seems unlikely that such anti-predator responses could have been retained for so long in the absence of selection pressures. Moreover, the lack of response to viperine snakes from mainland Spain supports the view that the anti-predator response is an adaptation that has evolved since the introduction of viperine snakes to Mallorca. Although anti-predator behaviour may contribute to the persistence of Mallorcan midwife toads in the face of snake predation, how can such a trait persist in a population such as at Pool A, where there appear to be no snakes? Viperine snakes may range widely, and the colonization of a midwife toad breeding pool by snakes may largely be due to chance. Indeed, it may take only a single snake to eliminate an entire cohort of tadpoles from one pool. Mallorcan midwife toads are excellent climbers, and can negotiate vertical
Swimming Immobile on bottom
Swimming Immobile on bottom
Swimming Immobile on bottom
Swimming Immobile on bottom
Mallorcan Natrix maura
Spanish Natrix maura
English Natrix natrix
Asiatic Elaphe rufodorsata
23 31
24 27
28 28
25 31
8.8 58.6
8.5 62.0
29.1 44.4
15.2 57.1
Unconditioned water
12.4 56.1
10.3 59.1
22.3 37.5
8.9 72.2
Snakeconditioned water
75 235
125 147
117 150
49* 69**
Wilcoxon T
23 25
23 28
27 29
24 28
N for test
11.3 49.9
7.4 59.2
20.6 50.8
17.9 59.2
Unconditioned water
10.0 51.0
8.2 61.5
20.3 46.7
17.4 54.9
Unconditioned water
Control
111 156
124 181
165 172
135 140
Wilcoxon T
Data show the mean percentage time spent engaged in two behaviours: (1) before and after switching to snake-conditioned water (Snake-test conditions); and (2) before and after switching between two sources of unconditioned water (Control conditions). In each of the four experiments, 32 tadpoles were used but N for test is less than 32 as a result of tied data. *P<0.01, **P<0.001 by the Wilcoxon signed-ranks test (two-tailed). All other paired comparisons are not significant.
Behaviour
Snake
N for test
Snake-test
Table II. Effect of snake-conditioned water on tadpole behaviour in a gravitational flow-through system
212 Animal Behaviour, 55, 1
Griffiths et al.: Responses of tadpoles to snakes cliffs and waterfalls, which are physical barriers to colonization of some toad pools by snakes. These snake-free pools may be able to provide toads that subsequently recolonize pools where snake predation has resulted in local extinction. Equally, if there is reciprocal transfer of toads from snakeaccessible pools to snake-free pools, anti-predator behaviour is likely to be retained within the metapopulation as a whole, even if predation is not occurring within a particular sub-population (e.g. Pool A). From a conservation point of view, it is reassuring that Mallorcan midwife toads have retained natural anti-predator behaviour through several generations of captive breeding. Captivebred toads that are reintroduced to the wild should therefore be able to minimize predation risk from snakes by continuing to express such behaviours. Owing to its precarious status, however, removal of viperine snakes from breeding pools must continue to be part of the population management strategy for the Mallorcan midwife toad. ACKNOWLEDGMENTS This work was supported by AAT-Garten und-Teichfreunde Luxemburgs; by a NERC Advanced Research Fellowship to Richard Griffiths and by an ASAB vacation scholarship to Penny Sharp. Jersey Wildlife Preservation Trust, Sarah Bush and Leigh Gillett provided captive-bred tadpoles; Stuart Worth and Derek Thompson provided snakes from Spain and England, respectively. Sarah Bush, Jose-Maria Roma´n, Alex Forteza, Pere Vicenc, Joan Mayol and the Consellaria de Medi Ambient provided logistical support for field work in Mallorca. Ernest Mendoza assisted with the translation of foreign papers, and Leigh Gillett and two anonymous referees provided helpful comments which improved the manuscript. REFERENCES Alcover, J. A. & Mayol, J. 1981. Espe`cies relı´quies d’amfibis i de re`ptils a les Balears i Pitiu¨ses. Boll. Soc. Hist. Nat. Balears., 25, 151–167. Arntzen, J. W. & Garcia-Paris, M. 1995. Morphological and allozyme studies of midwife toads (genus Alytes), including the description of two new taxa from Spain. Contrib. Zool., 65, 5–34.
213
Atkinson, I. 1989. Introduced animals and extinctions. In: Conservation for the Twenty-first Century (Ed. by D. Western & M. Pearl), pp. 54–75. Oxford: Oxford University Press. Bloxam, Q. M. C. & Tonge, S. J. 1995. Amphibians: suitable candidates for breeding-release programmes. Biodiv. Conserv., 4, 636–644. Chivers, D. P., Wisenden, B. D. & Smith, R. J. F. 1996. Damselfly larvae learn to recognize predators from chemical cues in the predator’s diet. Anim. Behav., 52, 315–320. Crews, D. & Garstka, W. R. 1982. The ecological physiology of a garter snake. Scient. Am., 247, 158– 168. Crowl, T. A. & Covich, A. P. 1990. Predator-induced life-history shifts in a freshwater snail. Science, 247, 949–951. Dial, B. E., Weldon, P. & Curtis, B. 1989. Chemosensory identification of snake predators (Phyllorhynchus decurtatus) by banded geckos (Coleonyx variegatus). J. Herpetol., 23, 224–229. Diamond, J. 1989. Overview of recent extinctions. In: Conservation for the Twenty-first Century (Ed. by D. Western & M. Pearl), pp. 37–41. Oxford: Oxford University Press. Dodson, S. I., Crowl, T. A., Peckarsky, B. L., Kats, L. B., Covich, A. P. & Culp, J. M. 1994. Non-visual communication in freshwater benthos: an overview. J. N. Am. benthol. Soc., 13, 268–282. Feminella, J. W. & Hawkins, C. P. 1994. Tailed frog tadpoles differentially alter their feeding behavior in response to non-visual cues from four predators. J. N. Am. benthol. Soc., 13, 310–320. Griffiths, R. A. & Denton, J. 1993. Interspecific associations in tadpoles. Anim. Behav., 44, 1153–1157. Holomuzki, J. R. 1995. Oviposition sites and fishdeterrent mechanisms of two stream anurans. Copeia, 1995, 607–613. Howe, N. R. & Harris, L. G. 1978. Transfer of the sea anemone pheromone, anthopleurine, by the nudibranch Aeolidia papillosa. J. Chem. Ecol., 4, 551–561. Jackson, M. E. & Semlitsch, R. D. 1993. Paedomorphosis in the salamander Ambystoma talpoideum: effects of a fish predator. Ecology, 74, 342–350. Kats, L. B., Petranka, J. W. & Sih, A. 1988. Antipredator defenses and the persistence of amphibian larvae with fishes. Ecology, 69, 1865–1870. Kiesecker, J. M., Chivers, D. P. & Blaustein, A. R. 1996. The use of chemical cues in predator recognition by western toad tadpoles. Anim. Behav., 52, 1237– 1245. Lawler, S. P. 1989. Behavioural responses to predators and predation risk in four species of larval anurans. Anim. Behav., 38, 1039–1047. Lima, S. L. & Dill, L. M. 1990. Behavioral decisions made under the risk of predation: a review and prospectus. Can. J. Zool., 68, 619–640. Magurran, A. E. 1989. Acquired recognition of predator odour in the European minnow (Phoxinus phoxinus). Ethology, 82, 216–223. Magurran, A. E. 1990. The inheritance and development of minnow anti-predator behaviour. Anim. Behav., 39, 834–842.
214
Animal Behaviour, 55, 1
Mason, R. 1992. Reptilian pheromones. In: Biology of the Reptilia. 18. Hormones, Brain and Behavior (Ed. by C. Gans & D. Crews), pp. 114–228. Chicago: University of Chicago Press. Mathis, A. & Smith, R. J. F. 1993. Fathead minnows, Pimephales promelas, learn to recognize northern pike, Esox lucius, as predators on the basis of chemical stimuli from minnows in the pike’s diet. Anim. Behav., 46, 645–656. Mayol, J. & Alcover, J. A. 1981. Survival of Baleaphryne Sanchiz and Adrover, 1979 (Amphibia: Anura: Discoglossidae) on Mallorca. Amphib.-Reptil., 3/4, 343–345. Petranka, J. W., Kats, L. B. & Sih, A. 1987. Predatorprey interactions among fish and larval amphibians: use of chemical cues to detect predatory fish. Anim. Behav., 35, 420–425. Schley, L. 1996. Activity patterns of tadpoles and anti-predator behaviour of tadpoles and postmetamorphic Mallorcan midwife toads Alytes muletensis. M.Sc. thesis, University of Kent. Semlitsch, R. D. & Gavasso, S. 1992. Behavioural responses of Bufo bufo and Bufo calamita tadpoles to chemical cues of vertebrate and invertebrate predators. Ethol. Ecol. Evol., 4, 165–173. Semlitsch, R. D. & Reyer, H.-U. 1992. Modification of anti-predator behaviour in tadpoles by environmental conditioning. J. Anim. Ecol., 61, 353–360.
Sih, A., Petranka, J. W. & Kats, L. B. 1988. The dynamics of prey refuge use: a model and tests with sunfish and salamander larvae. Am. Nat., 132, 463–483. Skelly, D. K. 1992. Field evidence for a cost of behavioral antipredator response in a larval amphibian. Ecology, 73, 704–708. Tonge, S. 1986. Collecting the Mallorcan midwife toad. Oryx, 20, 74–78. Tonge, S. J. & Bloxam, Q. M. C. 1991. The breeding programme for the Mallorcan midwife toad Alytes muletensis at the Jersey Wildlife Preservation Trust. Dodo, 27, 146–156. Van Damme, R., Bauwens, D., Thoen, C., Vanderstighelen, D. & Verheyen, R. F. 1995. Responses of naive lizards to predator chemical cues. J. Herpetol., 29, 38–43. Weldon, P. J. 1990. Responses by vertebrates to chemicals from predators. In: Chemical Signals in Vertebrates, vol. 5 (Ed. by D. W. MacDonald, D. Muller-Schwarze & R. M. Silverstein), pp. 501–521. New York: Plenum Press. Wilson, D. J. & Lefcort, H. 1993. The effect of predator diet on the alarm response of red-legged frog, Rana aurora, tadpoles. Anim. Behav., 46, 1017–1019.