Aaron M. Sullivan

ON TEMPORAL VARIATION AND CONFLICTING SELECTION PRESSURES: A TEST OF THEORY USING NEWTS

Most studies that examine conflicting selection pressures hold resources and risks constant, despite their ubiquitous fluctuation. Since little is known about the conse- quences of neglecting this variation, we examined the temporal response of male red-spotted newts, Notophthalmus viridescens, to conflicting female pheromones and damage-release alarm chemicals signaling predation. After a single exposure in both the laboratory and field, males were attracted to female odor and avoided conspecific alarm chemicals. Re- sponse to these combined cues depended on time after exposure, with males initially avoid- ing, and then being attracted to, the cue combination. This response shift was due to the resource and risk declining at different rates, and female odor accelerating male recovery from antipredator behavior. In the laboratory, males suppressed activity when exposed to alarm chemicals alone but increased their activity when female odor was added. Iterative exposures through the breeding season revealed that, as male mate search activity declined, male avoidance of alarm chemicals increased, but alarm chemical production appeared unchanged. Thawing dates differed between ponds of the same and different populations, which offset levels of mate search activity and consequently alarm chemical avoidance. As a result, simultaneous pond surveys made it appear as though there was geographic variation in reproductive and predator-avoidance behaviors. However, when thawing dates were aligned, the time courses of reproductive and predator-avoidance behaviors for the ponds coincided, demonstrating that observed site differences were predominantly due to different behavioral onsets, which would have gone overlooked had the larger temporal scale not been considered. These results indicate that temporal variation can be easily mistaken for geographic variation in behavior, increasing the potential for data interpretation errors. These studies underscore the importance of considering temporal variation when examining conflicting selection pressures.

Behavioural Responses by Red-backed Salamanders to Conspecific and Heterospecific Cues

Chemical cues released from injured prey are thought to indicate the proximity of a predator or predation event, and therefore, an area of elevated predation risk. Prey often avoid chemical cues released from injured heterospecifics, but there is little evidence to determine whether this is due to homologous cues among phylogenetically related species, or avoidance of injured syntopic species that experience predation from the same predators. The purpose of this study was to examine the response of terrestrial red-backed salamanders ( Plethodon cinereus ) to chemical cues from non-injured and injured members of their prey guild that vary in their relatedness to P.cinereus . In the laboratory, P.cinereus avoided chemical cues from injured conspecifics, injured and non-injured slimy salamanders ( P.glutinosus ), and injured confamilial dusky salamanders ( Desmognathus ochrophaeus ). Red-backed salamanders did not avoid rinses from non-injured conspecifics and dusky salamanders, or cues from injured and non-injured earthworms ( Lumbricus sp .), a more distantly related prey guild member. These results cannot be fully explained by either phylogenetic relatedness (among plethodontid salamanders) or prey guild membership alone. We suggest that a combination of these factors, and perhaps others, likely influenced the evolution of heterospecific alarm cue avoidance in the red-backed salamander.

A complex, cross-taxon, chemical releaser of antipredator behavior in amphibians

Prey species show diverse antipredator responses to chemical cues signaling predation threat. Among terrestrial vertebrates, the red-backed salamander, Plethodon cinereus, is an important species in the study of these chemical defenses. During the day and early evening, this species avoids rinses from garter snakes, Thamnophis sirtalis, independent of snake diet, but late at night, avoids only those rinses from garter snakes that have recently eaten P. cinereus. We tested whether the selective, late-night response requires the ingestion or injury of salamanders. In three experiments, we tested P. cinereus for their responses to separate or combined rinses from salamanders (undisturbed, distressed, and injured P. cinereus) and snakes (unfed, earthworm fed, and salamander-fed T. sirtalis). When paired against a water control, only rinses from salamander-fed snakes were avoided. When salamander treatments (undisturbed or distressed) were combined with the snake treatments (unfed or earthworm-fed) and tested against a water control, the combinations elicited avoidance. When selected treatments were paired against the standard rinse from salamander-fed snakes, only the combined rinses from salamanders and snakes nullified the avoidance response to the standard rinse. These data reveal a prey defense mechanism involving chemical elements from both the predator and prey that does not require injury or ingestion of the prey in the formation of the cue.

Sex differences and seasonal trade-offs in response to injured and non-injured conspecifics in red-spotted newts, Notophthalmus viridescens

Abstract. Injured prey often release alarm chemicals that induce antipredator behaviors in conspecifics. Injured or killed prey most likely release a wide array of chemicals in addition to alarm substances, such as sexual pheromones, which could enhance or compromise antipredator responses. Thus, damage-release cues provide an excellent opportunity to examine the influence of seasonally fluctuating sexual pheromones on antipredator behaviors. We used a series of laboratory and field experiments and meta-analysis to examine seasonal changes and sex differences in the response of red-spotted newts, Notophthalmus viridescens, to the odor of non-injured conspecifics and conspecific tissue extracts, the latter of which presumably contain pheromones of non-injured conspecifics combined with alarm chemicals signaling predation. During the peak of the breeding season, males were attracted to females and multiple males, but did not avoid tissue extracts from either sex. As the breeding season waned, male attraction to females and males decreased, while avoidance of alarm extracts from both sexes concurrently increased. In contrast to male behavior, females were indifferent to both sexes during the breeding season, and showed significant avoidance only of female extract. As the breeding season progressed, females displayed no change in response to treatments. Male and female responses to female rinse and extract differed significantly, but their response to male treatments did not. During the non-breeding season, both males and females were indifferent to the odor of conspecifics and avoided conspecific tissue extracts, with the magnitude of male avoidance greater than that of female avoidance, suggesting sex differences in response to alarm cues in both the breeding and non-breeding seasons. In general, both male and female response to conspecific odor and tissue extracts covaried positively, suggesting that social pheromones can be detected within conspecific macerates and compromise alarm-chemical avoidance. Many of the sex differences in both seasons are likely explained by selection pressures imposed on males to intensely mate search during the breeding season, suggesting that the mating system of newts directly influences predation threat during reproductive activity and may have significant indirect consequences on risk during the non-breeding season.

VARIATION IN THE ANTIPREDATOR RESPONSES OF THREE SYMPATRIC PLETHODONTID SALAMANDERS TO PREDATOR-DIET CUES

Organisms may reduce the risk of predation by responding to chemical cues from predators. Recent research shows that many species vary their antipredator response depending on the diet of the predator. We examined the responses of three plethodontid species of salamander (Plethodon cinereus, Eurycea bislineata, and Desmognathus ochrophaeus) to chemical cues from a shared snake predator (Thamnophis sirtalis). At the time of the study, Eurycea bislineata showed overlap in habitat with Plethodon cinereus and Desmognathus ochrophaeus, but Plethodon cinereus and Desmognathus ochrophaeus showed no overlap with one another. Each salamander species was presented with chemical cues from snakes fed Desmognathus ochrophaeus (TSDo), Eurycea bislineata (TSEb), and Plethodon cinereus (TSPc). Plethodon cinereus avoided both TSPc and TSEb, whereas Eurycea bislineata avoided only TSEb. Conversely, Desmognathus ochrophaeus did not avoid any cues from the predator, regardless of the diet of the snake. When we analyzed activity data, we discovered that Plethodon cinereus showed higher activity levels when exposed to TSPc than to the other cues. Individual Eurycea bislineata did not vary their activity to the three treatments. Lastly, Desmognathus ochrophaeus, which did not avoid any of the cues from the predator, were more active in response to TSDo and TSEb than to TSPc. These results show that phylogenetically related prey species may employ a variety of antipredator behaviors and suggest that discrimination of predator diet-cues may be linked to the degree of microhabitat overlap among the different prey species at the time of our study. Our study also highlights the importance of using multiple response variables when examining antipredator behavior.

Nocturnal shift in the antipredator response to predator-diet cues in laboratory and field trials

Prey species may react to predator chemical traces in the environment with a variety of antipredator behaviors (Weldon, 1990; Chivers and Smith, 1998; Kats and Dill, 1998). Such responses to predator chemical cues might reduce predation risk, but also may result in lost foraging or mating opportunities (Lima, 1998a,b), so it is not surprising that some species adjust their responses based on the degree of perceived predation threat. These modifications may be based on chemical information gathered directly from the predator or its recent prey (Madison et al., 1999a; Chivers and Mirza, 2001). The mosaic of chemical products released at a predation site, and dispersed by the predator, could allow nearby prey to assess predation threat and fine-tune their responses. Among the chemical cues released from predators during and after a predation event, those associated with predator diet have emerged as an important factor in predator assessment by both aquatic (Chivers and Mirza, 2001) and terrestrial prey species (Madison et al., 1999a,b). However, in most circumstances it is unclear whether the active components of the cue are from the prey, the predator, or some combination of prey and predator cues (e.g., Madison et al, 2002). As one example of this complexity, the interaction between the terrestrial red-backed salamander, Plethodon cinereus, and its garter snake predator, Thamnophis sirtalis, shows that during the day salamander avoidance of snake cues occurs independent of snake diet, but late at night avoidance is apparently restricted to cues from snakes feeding on red-backed salamanders (Madison et al. 1999a,b). One limitation of these studies is that few diet treatments were tested, leaving the possibility that alternative diets might also elicit late night avoidance.

THE ONTOGENY OF CHEMICALLY-MEDIATED ANTIPREDATOR BEHAVIOURS IN NEWTS (NOTOPHTHALMUS VIRIDESCENS): RESPONSES TO INJURED AND NON-INJURED CONSPECIFICS

Summary Responses to alarm chemicals from injured prey may ine uence predation risk and foraging success of receivers and senders, while learning can ine uence the strength of these responses. Thus, it is important to know when in ontogeny prey produce and detect alarm substances and how learning shapes their response, but surprisingly little is known about either of these topics. We assessed when in the life of red-spotted newts, Notophthalmus viridescens, alarm chemicals are produced and detected by comparing adult versus eft (terrestrial juveniles) and larval responses to rinses and tissue extracts from individuals in each life-history stage. To evaluate the ine uence of experience in larvae exposed to conspecie c alarm substances and rinses from adults known to cannibalize larvae, we compared the response of naive larvae, which had no prior experience with alarm chemicals or predators, to experienced larvae, which were likely to have experienced alarm chemicals and predators in their native pond. Larvae were indifferent to larval rinses and extracts, but reduced their activity in response

Decline in avoidance of predator chemical cues: Habituation or biorhythm shift?

Studies of the behavior of captive animals are widespread in the literature. Such practices may occur because animals are not seasonally available, do not occur near the home institution of the researcher, or because it is much easier to control for multiple variables in the laboratory. In addition, because of animal rarity or inaccessibility, repeat testing procedures are often used on small numbers of animals, potentially causing pseudoreplication problems (Ramirez et al., 2000). It is reasonable to expect that the laboratory environment at some point will alter the natural responsiveness of animals to particular stimuli, or result in behavior that may be an artifact of stress, housing, diet, and the testing apparatus (Hennig and Dunlap, 1978; Jarvi, 1990). Through differential mortality, captivity can also create laboratory populations of atypical animals most tolerant of captive conditions (Navas and Gomes, 2001). Even when captivity effects are detected, there are few opportunities or rewards for investigators to report shortcomings in their own methodology or in correcting long-standing methods widely in use. Field validation of results is one way of detecting possible laboratory artifact (e.g., Sullivan et al., 2002). However, field studies may not be possible, may not be directly comparable to laboratory studies, and may introduce other experimental effects (Rohr and Madison, 2001; Rohr et al., 2002). An alternative approach to detecting captivity effects is to record data on changes in behavior of captive animals through time (Rohr et al., 2003). In seven years of studies, we have documented a highly predictable avoidance response in the red-backed salamander, Plethodon cinereus, to body rinses from garter snakes, Thamnophis sirtalis, that have been feeding on P. cinereus (standard TSPC stimulus; Madison et al., 1999a,b; McDarby et al., 1999; Maerz et al., 2001; Madison et al., 2002; Sullivan et al., 2002, 2003). While we always get avoidance of TSPC, no such avoidance occurs in response to body rinses from garter snakes feeding on earthworms, goldfish, or other salamander species when these tests are conducted late at night (Madison et al., 1999a,b; Sullivan et al., 2003), so avoidance is not simply to nitrogenous wastes. The response to TSPC occurs regardless of the individual adult garter snake used