THESIS BEHAVIORAL EXPERIMENTS
RESPONSE OF SALAMANDERS TO CHEMICAL STIMULI FROM PREDATORS IN NATURAL HABITATS
CALEB HICKMAN
This was submitted to the Herpetologists' League for the Jaeger Travel Award (summer 2002). Because of the limitations on length, this work does not fully express my thesis.
Chemical cues often are used in locating prey and mates, assessing competitors, and avoiding predators (Kats and Dill, 1998). Use of chemical signals has been documented in many laboratory studies, but tests under natural conditions have proved exceedingly difficult (Mathis et al., 1995; Mason et al., 1998).
Chemical cues can be effectively used in areas where other types of sensory information are unreliable. For example, animals that rely heavily on visual cues are at a disadvantage in habitats that are dark or where there are obstructions such as vegetation or sediments (e.g., Downes and Shine, 1998).
This paper focuses on the use of chemical cues by salamanders for avoidance of predation. Salamanders are particularly interesting subjects because they occur in a wide range of habitats. Many species occur in murky, vegetated ponds, while others live in streams where the water can be clearer, but where predators often are cryptic. Adults of most species are terrestrial, occupying cover (rocks, logs, leaf litter) where visual information is limited. All are mostly active at night. Numerous laboratory studies have indicated that the most common responses to chemical stimuli from predators are either reduced activity (e.g., Mathis and Vincent, 2000) or flight (e.g., Madison et al., 1999).
My study quantified antipredator responses of salamanders to chemical stimuli from predators in three different habitats, and compared responses to stimuli from predatory and nonpredatory heterospecifics. I tested larval ringed salamanders, Ambystoma annulatum, in a pond, adult neotenic graybelly salamanders, Eurycea multiplicata griseogaster, in a stream, and adult southern red-backed salamanders, Plethodon serratus, in a forest.
METHODS
For all three experiments, the general approach was to examine responses of salamanders to chemical stimuli from predators under natural conditions. In each experiment, salamanders were exposed to chemical stimuli from a predator, a nonpredatory heterospecific, and a blank control (Table 1), and each individual was tested only once. Treatments randomly assigned and coded so that the experiments were blind. All trials were conducted after dark when salamanders in all three habitats are active. Each experiment presented different technological challenges, which are explained below.
Protocol: Aquatic Trials
Experiments in aquatic habitats can fail if researchers wading in the water unduly disturb the salamanders. The first methodological challenge for the aquatic experiments was to devise a technique for identifying and observing individuals without eliciting disturbance effects. In the pond experiment, I located focal larvae (Ambystoma annulatum) that were active around the edge of the pond and observed them from the bank. In the stream experiment, salamander (Eurycea multiplicata) activity was not sufficient along the stream bank for observations. However, as the streambed dries up near the end of the summer, the salamanders follow the water table by burrowing in the rocky substrate. At this time, I dug holes (105 cm ± 10 cm) in the streambed until I reached the water table. At night, salamanders would become active on the surface of the exposed substrates in these holes and I could observe their behavior with only minimal disturbance.
The second challenge was to introduce the stimuli at a standard depth and distance from the salamander while minimizing the potential for disturbance. I constructed a stimulus injection apparatus consisting of a 60-mL syringe connected to polyethylene tubing (105-cm) that was attached to a bamboo splint. The stimulus (50 mL per trial) was introduced approximately 20 cm in front of the focal animal.
The third technical challenge was that aquatic habitats are heterogeneous with different flow rates and back flows depending on microhabitat features. I added 1 drop of green food coloring to each aliquot so that I could observe the exact time at which the stimuli contacted the head of the test animal.
Response variables were chosen based on the results of previous lab studies (Mathis, Murray, and Hickman, unpublished data) and on my preliminary observations. I recorded latency to move for the pond experiment, and latency to flee (move 21 cm) for the stream experiment. Timing began when the stimuli contacted the focal salamander’s head.
Protocol: Terrestrial Trials
The major technical challenge for the terrestrial experiment was observing responses of animals that are hidden either under leaf litter or cover objects. My methods followed the protocol recently described by Sullivan et al. (2002). In laboratory tests, terrestrial salamanders reduce their activity when exposed to chemical stimuli from predatory snakes (Maerz et al., 2001). I removed focal salamanders from under cover objects, and identified them according to individual markings. I then sprayed the substrate beneath the cover objects with the assigned chemical stimulus and replaced the salamanders. Prey are abundant in the damp leaf litter, and salamanders typically leave their cover objects at night to forage (Jaeger, 1980). Reduced activity in response to increased predation risk should result in increased fidelity toward cover objects. I reexamined cover objects for presence of salamanders after two hours.
RESULTS
For Ambystoma annulatum in the pond experiment, there was a significant difference among treatments (Kruskal-Wallis ANOVA: H = 13.45, P = 0.001; Fig. 1A). Latencies to move were significantly longer for the newt (predator) treatment than for either the tadpole (nonpredator) (Q = 4.06, P < 0.025) or blank (Q = 4.83, P < 0.005) treatments. Tadpole and blank treatments did not differ (Q = 0.77, P > 0.50).
For Eurycea multiplicata in the stream experiment, there also was a significant difference among treatments (H =7.59, P = 0.022; Fig. 1B). Latencies were significantly shorter in response to sculpin (predator) treatment than for either the stoneroller (nonpredator) (Q = 16.00, P < 0.001) or the blank (Q = 15.21, P < 0.001) treatments. Stoneroller and blank treatments did not differ (Q = 0.29, P > 0.50).
Sample sizes were somewhat low for Plethodon serratus in the terrestrial experiment (7 salamanders per treatment). Data for the two control treatments were similar (number of recaptures: blank = 1; skink = 2), so we combined these two categories into a single “nonpredator” category. Significantly more recaptures were made for salamanders exposed to the predator (snake) stimuli (71%) than for salamanders exposed to the nonpredator stimuli (25%) (Fisher’s Exact Test, P < 0.05).
DISCUSSION
Clearly, salamanders in this study gave appropriate antipredator responses to chemical stimuli from predators under natural conditions. This result was true for salamanders in three genera (Ambystoma, Eurycea, Plethodon), with three predators (newts, fish, snakes), and in three different natural environments (pond, stream, and terrestrial). The success of these experiments required development of methods that allowed observation of salamanders in low-visibility habitats, caused minimal disturbance to focal animals, and allowed fine control of stimulus presentation.
I also found that salamanders distinguished between stimuli from predatory and nonpredatory heterospecifics. A few other studies have documented responses to chemical stimuli from predators under natural conditions (Kats et al., 1988; Petranka and Hayes, 1998; Sullivan et al., 2002), but none addressed whether responses were specific to predatory stimuli.
Although salamanders in all three experiments responded to the predatory stimuli with antipredator behavior, there were differences among the species in the form of the response (reduced activity versus flight). Both Amybystoma and Plethodon responded to predatory stimuli by reducing activity. In both experiments, the stimuli were from predators with active foraging modes (newts and snakes). Reduced activity can increase survival of prey exposed to active predators because movement tends to draw the predator’s attention. Moreover, reduced activity should be particularly effective in habitats with low visibility. In contrast, the typical response of Eurycea was flight. In Ozark streams, sculpin are cryptic, ambush predators and are inefficient swimmers. For sculpin, rapid flight is probably a more effective antipredator strategy than reduced activity, particularly when the water is clear allowing good visibility. Studies of animals in their natural habitats should lead to a better understanding of the adaptive value of antipredator responses.
TABLE 1.¾Stimulus animals used to provide the predatory and nonpredatory chemical stimuli for the three experiments. For experiments 1 and 2, chemical stimuli were generated by placing the animals in water for 96-120 hours, with volume of water controlled for a standard number of mL per gram of stimulus animal. For experiment 3, stimulus animals were randomly placed in 400 mL beakers for 48 hours and then the beaker was rinsed in 100 mL of water. All samples of stimulus water were frozen and defrosted together at the field site. For each experiment, dechlorinated tap water was frozen as a blank control.
N/ Body mass
Experiment 1: Ambystoma annulatum (pond)
Predator: Notophthalmus viridescens (central newt) 18/ 1.7 ± 0.31g
Nonpredator: Rana sphenocephala (frog tadpoles) 16/ 2.5 ± 0.63g
Experiment 2: Eurycea multiplicata griseogaster (stream)
Predator: Cottus carolinae (banded sculpin) 4/ 9.6 ± 4.55g
Nonpredator: Campostoma pullum (central stoneroller) 5/ 9.1 ± 2.24g
Experiment 3: Plethodon serratus (terrestrial)
Predator: Diadophis punctatus (ringneck snake) 5/ 5.2 ± 1.94g
Nonpredator: Eumeces fasciatus (skink) 5/ 4.7 ± 2.81g
FIGURE LEGEND:
Fig. 1. Eurycea multiplicata experiment: Latency in seconds to flee farther than 21cm when exposed to kairomones from blank control, non-predatory stoneroller, and to a predatory sculpin. Letters “a” and “b” are used to show significant differences among treatments according to nonparametric multiple comparisons tests. These data are represented by bar charts indicating mean latency to flee greater than 21cm with bars showing standard error.
Fig. 2. Eurycea multiplicata experiment: Latency in seconds to complete a burrow when exposed to kairomones from blank control, non-predatory stoneroller, and to a predatory sculpin. Letters “a” and “b” are used to show significant differences among treatments according to nonparametric multiple comparisons tests. These data are represented by bar charts indicating mean latency to complete a burrow with bars showing standard error.
Fig. 3. Ambystoma annulatum experiment: Latency in seconds to move when exposed to kairomones from blank control, non-predatory tadpole, and to a predatory newt. Letters “a” and “b” are used to show significant differences among treatments according to nonparametric multiple comparisons tests. These data are represented by bar charts indicating mean latency to move with bars showing standard error.
Above Fig.'s– Data means ± SE. P-value is for a Kruskal-Wallis ANOVA and different letters represent significant differences according to nonparametric multiple comparison tests (Zar, 1984).
Fig. 4. Plethodon serratus experiment: Percentage of salamanders recaptured under cover objects after being exposed to kairomones from blank control, non-predatory skink, and to a predatory snake for two hours. Both blank and skink are combined into a single “non-predator” category which represents 14% recaptured salamanders. Significantly more recaptures were made for the predator category (71%). Letters “a” and “b” are used to show significant differences among treatments according to percentages.
LITERATURE CITED
DOWNES, S., AND R. SHINE. 1998. Sedentary snakes and gullible geckos: predator-prey coevolution in nocturnal rock-dwelling reptiles. Animal Behaviour 55:1373-1385.
JAEGER, R.G. 1980. Fluctuations in prey availability and food limitation for a terrestrial salamander. Oecologia 44:335-341.
KATS, L. B., AND L. M. DILL. 1998. The scent of death: Chemosensory assessment of predation risk by prey animals. Ecoscience 5:361-394.
KATS, L.B., J.W. PETRANKA, AND A. SIH. 1988. Antipredator defenses and the persistence of amphibian larvae with fishes. Ecology 69:1865-1870.
MADISON, D.M., J.C. MAERZ, AND J.D. MCDARBY. 1999. Chemosensory responses of salamanders to snake odors: freeze and flight contingencies. Pp. 508-516. In: R.E. Johnston, D. Müller-Shwarze, P. Sorensen (Eds.), Advances in chemical communication in vertebrates. Kluwer Academic/Plenum Press, New York, New York, U.S.A.
MATHIS, A., R. G. JAEGER, W. H. KEEN, P. K. DUCEY, S. C. WALLS, AND B. W. BUCHANAN. 1995. Aggression and territoriality by salamanders and a comparison with the territorial behaviour of frogs. Pp. 633-676. In: H. Heatwole and B. K. Sullivan (Eds.), Amphibian Biology, Vol. 2, Social Behaviour. Chipping Norton, NSw, Australia.
MAERZ, J.C., N.L. PANEBIANCO, AND D.M. MADISON. 2001. The effects of predator chemical cues and behavioral biorhythms on the foraging activity of terrestrial salamanders. Journal of Chemical Ecology 27:1333-1334.
MASON, R. T., D.P. CHIVERS, A. MATHIS, AND A.R. BLAUSTEIN. 1998. Bioassay methods for amphibians and reptiles. Pp. 271-325. In K.F. Haynes and J.G. Millar (Eds.), Methods in chemical ecology. Kluwer Academic, Norwell Massachusetts, U.S.A.
MATHIS, A., AND F. VINCENT. 2000. Differential use of visual and chemical cues in predator recognition and threat-sensitive predator-avoidance responses by larval newts (Notophthalmus viridescens). Canadian Journal of Zoology 78:1646-1652.
PETRANKA, J., AND L. HAYES. 1998. Chemically mediated avoidance of a predatory odonate (Anax junius) by American toad (Bufo americanus) and wood frog (Rana sylvatica) tadpoles. Behavioral Ecology and Sociobiology 42:263-271.
SULLIVAN, A.M., J.C. MAERZ, AND D.M. MADISON. 2002. Anti-predator response of red-backed salamanders (Plethodon cinereus) to chemical cues from garter snakes (Thamnophis sirtalis): laboratory and field experiments. Behavioral Ecology and Sociobiology 51:227-233.
ZAR, J.H. 1984. Biostatistical Analysis, 2nd ed. Prentice-Hall, Englewood Cliffs, New Jersey.
Send E-Mail to: caleb@sevilleta.unm.eduCopyright © 2005 Caleb Hickman. All Rights Reserved