Behavioral Ecology and Sociobiology

, Volume 64, Issue 7, pp 1117–1123

Risk level of chemical cues determines retention of recognition of new predators in Iberian green frog tadpoles

Authors

    • Departamento de Ecología EvolutivaMuseo Nacional de Ciencias Naturales, CSIC
  • Pilar López
    • Departamento de Ecología EvolutivaMuseo Nacional de Ciencias Naturales, CSIC
  • José Martín
    • Departamento de Ecología EvolutivaMuseo Nacional de Ciencias Naturales, CSIC
Original Paper

DOI: 10.1007/s00265-010-0927-y

Cite this article as:
Gonzalo, A., López, P. & Martín, J. Behav Ecol Sociobiol (2010) 64: 1117. doi:10.1007/s00265-010-0927-y

Abstract

In aquatic environments, many prey rely on chemosensory information from injured (alarm cues) or stressed conspecifics (disturbance cues) to assess predation risk. Alarm cues are considered as a sign of higher risk than disturbance cues. These cues could be used by prey to learn potential new predators. In this study, we tested whether Iberian green frog tadpoles (Pelophylax perezi) exhibited antipredator responses to alarm and disturbance cues of conspecifics and whether tadpoles could associate new predators with alarm or disturbance cues. Tadpoles reduced their activity in the presence of disturbance cues, but only weakly when compared with their response to alarm cues. Also, tadpoles learned to recognize new predators from association with alarm or disturbance cues. However, the period of retention of the learned association was shorter for disturbance than alarm cues. Our results indicate that tadpoles are able to modify their antipredatory behavior according to (1) the degree of risk implied by the experimental cues (2) their previous experience of chemical cues of the predator.

Keywords

Alarm cuesDisturbance cuesMemoryPredation riskPredator recognitionFrogs

Introduction

In aquatic habitats, diverse prey organisms rely on chemosensory information to assess local predation risk (Smith 1992; Chivers and Smith 1998). Aquatic environments are ideal for the solution and dispersal of chemical cues, which are especially useful in turbid water and highly structured habitats or for species with poor developed visual senses (Wisenden 2000). Antipredatory or defensive responses occur to chemical cues released by predators, but also to chemical cues released by other prey (i.e., alarm signals; Kats and Dill 1998).

Chivers and Smith (1998) divided chemical alarm signaling systems into two general categories based on the point in the predation sequence when cues are emitted. Damage-released alarm cues are those chemicals released by prey animals only upon being captured by a predator. In contrast, disturbance cues are chemicals that are released by senders that have been disturbed or stressed, but not captured by a predator. Damage-released alarm cues have been found in a variety of aquatic organisms, including amphibians (Pfeiffer 1966; Hews and Blaustein 1985; Wilson and Lefort 1993; general review in Chivers and Smith 1998). In fish, alarm cues are typically localized in the epidermis of the prey and are released only following mechanical damage to the skin, as would occur during a predation event (Chivers and Smith 1998). However, such information is unavailable for amphibian’s alarm cues.

Chemical disturbance cues have been found in a high variety of aquatic animals, including amphibians (Kiesecker et al. 1999; Bryer et al. 2001), but they have received less attention than alarm cues. Iowa darters, Etheostoma exile, increase antipredator behavior when exposed to chemical cues of disturbed conspecifics (Smith 1979; Wisenden et al. 1995). Both crayfish, Orconectes virilis, and hermit crabs, Calcinus lavimanus, increase antipredator behavior when exposed to chemical cues released by disturbed conspecifics (Hazlett 1985, 1990b). In addition, O. virilis responds to disturbance cues produced by numerous heterospecifics, including other crayfish species, leeches, newts, and fish (Hazlett 1985, 1989, 1990a). This suggests that disturbance cues have a commonality of composition across taxa, which may include ammonium (Hazlett 1990a, b). This agrees with the finding that red-legged frog tadpoles release ammonium (NH4+) upon being disturbed by a predator and that conspecific tadpoles respond with antipredator behavior when exposed experimentally to a pulse of ammonium (Kiesecker et al. 1999). Ammonium may be excreted from the gills or in the urine during periods of increased metabolic activity that is required for effective escape or when prey are exposed to some stressful situation such as the presence of a predator. Release of disturbance cues may not be intentional, but may represent a normal physiological process to which other individuals have become sensitive (Mirza and Chivers 2002). If the disturbance cue is really ammonium, which is the excretion product in most aquatic animals, this would explain why organisms are able to react to disturbance cues from heterospecifics.

The release and detection of alarm and disturbance cues have important implications for predator–prey interactions. Prey animals that detect alarm or disturbance cues have an early warning of the presence of a predator and may be able to avoid an encounter by leaving the area (Jordão 2004) or by reducing movements (Mirza and Chivers 2002) and becoming cryptic. Early detection of a predator’s presence may allow the prey to increase vigilance, which will probably result in an improved chance of survival should the encounter escalate to an attack (Hews 1988; Mathis and Smith 1993).

However, according to the threat-sensitive predator avoidance hypothesis (Helfman 1989), natural selection should favor individuals that take action appropriate to the magnitude of threat, which would require an accurate discrimination of the current level of risk that each predator poses. Thus, the response of prey to alarm cues should be stronger than the response to disturbance cues, if the latter are perceived as indicating lower level risk than alarm cues (Chivers and Smith 1998). Disturbance cues are thought to be low-level indicators of risk to which prey animals respond with antipredator behavior, but the detection of disturbance cues can provide a survival benefit during an encounter with a predator (Mirza and Chivers 2002).

Disturbance cues might facilitate learning to recognize potential predators (Chivers and Smith 1998). Mirza and Chivers (2002) showed that damage-released alarm cues could facilitate learned recognition of predators. Ferrari et al. (2008) showed than juvenile rainbow trout use the disturbance cues as a warning cue, but are unable to use disturbance cues to learn to recognize new predators. To our knowledge, there have been no studies of the capacity of amphibians to retain the association of a new predator with chemical distress cues. As disturbance cues are low level indicators of threat, we hypothesized that if tadpoles are able to associate new predators with disturbance cues, tadpoles should react to the new predator chemical cues more weakly than to alarm cues, decreasing their activity to a lesser degree. We further hypothesized that tadpoles exposed to disturbance cues retain the memory of the predator for less time than tadpoles that have associated the predator with alarm cues.

In our experiments, we first tested whether Iberian green frog tadpoles (Pelophylax perezi) exhibited an antipredator response. We then measured the magnitude of this response to damage-released alarm cues from injured conspecifics and to chemical cues from disturbed conspecifics. In an additional experiment, we examined whether tadpoles were able to learn to associate chemical cues from non-predatory exotic fish species with predation risk (to which tadpoles can not be genetically predisposed). The non-predatory fish was associated artificially with predation risk by the simultaneous presence of conspecific alarm cues or disturbance cues. Finally, we tested whether each type of cue resulted in different periods of retention of the learned association.

Methods

Study animals

Iberian green frog tadpoles, P. perezi (length, \( \overline X \) ± SE = 5.4 ± 0.2 cm, Gosner’s stage, 25; see Gosner 1960) were collected by netting at several small ponds in Collado Mediano (Madrid, central Spain). Tadpoles were housed in groups of five at El Ventorrillo Field Station, located 10 km from the capture area, in a plastic aquaria (49 × 29 × 25 cm lwh) with 5 L of water at ambient temperature and under a natural photoperiod. They were fed with commercial fish flakes every day.

We obtained non-predatory zebra danio fish (Brachyodanio rerio) from a commercial dealer to be used as the source of neutral scent. Before and after the experiment, fishes were maintained in a large filtered aquarium and regularly fed commercial fish flakes.

All animals were healthy during the trials and all maintained or increased their original body mass. The experiments were performed under license from the “Consejería de Medioambiente de la Comunidad de Madrid” (the Environmental Agency of the local Government of Madrid). Procedures conformed to recommended guidelines for use of live amphibians in laboratory research (ASIH 2004).

Preparation of chemical stimuli

Alarm cues of tadpoles were prepared from three tadpoles (length, \( \overline X \) ± SE = 4.2 ± 0.1 cm). Tadpoles were made hypothermic by placing them in 4°C for 20 min and then euthanized by a quick blow to the head to avoid suffering (ASIH 2004). We did not use a chemical anesthetic because these chemicals may interfere with natural chemical cues in subsequent trials. An extract was prepared by putting the dead tadpoles in a clean disposable plastic dish and macerating them in 3,000 mL of distilled water. The stimulus water was then filtered through absorbent paper to remove solid particles and immediately frozen in 10-mL portions until used (Woody and Mathis 1998).

Disturbance cues from tadpoles were prepared using 20 tadpoles that were placed into a 200-mL aquarium and stressed by simulating a 30-s predator attack with a wooden bird model that imitated a little egret, Egretta garzetta (beak, 5 cm). The predator attacks consisted of moving the wooden model around the aquarium to simulate ten attempts to catch the tadpoles with the beak. Care was taken not to touch or damage any of the tadpoles. Since we did not know the period of degradation of the disturbance cues, we prepared them immediately before each experiment and quickly used them. Ten different tadpoles were used each time and then kept separately. All 20 tadpoles used (ten for each experiment) behaved normally 1 h after being stressed.

The fish chemical stimulus was prepared by placing ten zebra danio fishes (2.6 ± 0.2 cm) into a 10-L aquarium with clean water for 3 days. This aquarium was aerated, but not filtered. Fishes were not fed during this short period to avoid contaminating water with food odor. Thereafter, water was drawn from the aquaria and frozen in 10-mL portions for use in experiments. Fishes were returned and fed in their large home aquaria. We prepared control water in an identical manner but without placing fish in the aquarium (Woody and Mathis 1998).

Experiment 1: responses of tadpoles to alarm and disturbance cues

To determine whether Iberian green frog tadpoles were able to detect disturbance cues, we conducted an experiment in which they were exposed to (1) chemical cues from conspecifics disturbed with the wooden bird model, (2) alarm cues from conspecifics, or (3) clean water as a control for basic activity levels.

Tadpoles were tested individually in gray, U-shaped troughs (101 × 11.4 × 6.4 cm lwh) sealed at both ends with plastic caps. We marked the internal part of each trough with four crossing lines that created five subdivisions of equal surface area. We filled each trough with 3 L of clean water that was obtained from a mountain spring that lacked fish. Water temperature was 20°C. We placed clear plastic over each trough to isolate the system from air movements in the testing room (see Rohr and Madison 2001). Each trial lasted 1 h and consisted of a 30-min pre-stimulus period and a 30-min post-stimulus period separated by a stimulus introduction. We assigned 10 mL of test solutions (water, alarm cues, or disturbance cues) to one end of each trough (right or left) by stratified randomization. We placed a single tadpole in each trough and waited 5 min for habituation. Then, we started the pre-stimulus period. Immediately after the pre-stimulus period ended, we carefully pipetted the test solution. The post-stimulus period began upon conclusion of pipetting. During both the pre- and the post-stimulus periods, we recorded from a blind the quadrant occupied by each tadpole at 1-min intervals. We calculated levels of activity from the number of lines crossed by each tadpole during the observation period (Rohr and Madison 2001; Gonzalo et al. 2007). We used 45 tadpoles, 15 assigned randomly to each treatment. Tadpoles were tested once and then discarded. Because all individual tadpoles used in the experiment were observed at least once in all of the subdivisions of the trough, we were confident that all tadpoles were exposed to the chemical stimuli.

For each trial, we calculated levels of activity as the difference between the numbers of line crossings between the post- and pre-stimulus periods. Positive values indicated increased movement following addition of the stimulus; negative values indicate decreased activity. Data were logarithmically transformed and then tested with a one-way ANOVA with independent treatments. Post hoc multiple comparisons were made using Tukey’s pairwise comparisons (Sokal and Rohlf 1995).

Experiment 2: retention of the learned association

We designed this experiment to determine (1) whether tadpoles were able to learn to recognize novel predators by associating the predator chemical cues with disturbance cues from conspecifics and (2) the duration of predator recognition if learning occurred. We randomly assigned 12groups of five tadpoles to three different treatments (control, alarm, and disturbance). On the first day of the experiment, tadpoles from the control treatment were exposed to the fish chemical cues alone mixed with clean water. Simultaneously, tadpoles from the alarm treatment were exposed to both the scent of the fish and conspecific chemical alarm cues, thus simulating the cues from a predatory fish that was eating a conspecific tadpole. Similarly, tadpoles from the disturbance treatment were exposed to both the scent of the fish and conspecific disturbance cues, simulating the cues from a predatory fish that stressed conspecific tadpoles by attacking, but not capturing them. Previous studies showed that tadpoles conditioned with the mix of fish and alarm cues were able to recognize the fish chemical cues alone as coming from a predator 2 days later (Gonzalo et al. 2007) and were able to remember the predators at least 9 days after being conditioned by alarm cues (Gonzalo et al. 2009). Thus, tadpoles from the alarm group were considered to be conditioned.

To assess the duration of predator recognition, tadpoles from the three treatments were tested with the fish chemical cues alone in clean water on days 1, 3, 6, or 9 after the initial conditioning. The experiment was conducted using the same procedure as in experiment 1. All tadpoles from the three groups were tested with 10 mL of fish scent. Each trial lasted 1 h and consisted of a 30-min pre-stimulus period and a 30-min post-stimulus period separated by a stimulus introduction. Individuals of the three groups were tested in parallel, and observations were carried out from a blind. Each day, we tested 15 individuals from each of the three groups (control, alarm, and disturbed). These tadpoles were chosen randomly from several groups. After the trial, they were kept separately and not used in subsequent trials. We did not use the same individual tadpoles in more than one test because tadpoles could progressively learn the irrelevance of the non-dangerous predator (Hazlett 2003) or habituate to the predator cue, so that the results might not reflect duration of memory alone.

For each trial, we calculated levels of activity as the difference between the numbers of line crossings between the pre- and post-stimulus periods as described above. Data were logarithmically transformed and then tested by general linear modeling (Grafen and Hails 2002). Day (i.e., days from the initial conditioning event) and treatment were independent variables. We included the interactions between variables in the model to test differences in effects of the treatments among days. Post hoc multiple comparisons were made using Tukey’s pairwise comparisons (Sokal and Rohlf 1995).

Results

Experiment 1: responses of tadpoles to alarm and disturbance cues

There were significant differences among treatments in changes in average activity (one-way ANOVA, F2,42 = 51.5, P < 0.0001; Fig. 1). Tadpoles from the control group were significantly more active during the post-stimulus period relative to the pre-stimulus period than tadpoles from the disturbance and the alarm groups (Tukey’s tests, P < 0.001 in both cases). Also, tadpoles from the alarm group had significantly greater reduction of movements in the post-stimulus period than tadpoles from the disturbance group (P < 0.005). Therefore, alarm cues seemed to be considered by tadpoles as a more dangerous signal than disturbance cues (Fig. 1).
https://static-content.springer.com/image/art%3A10.1007%2Fs00265-010-0927-y/MediaObjects/265_2010_927_Fig1_HTML.gif
Fig. 1

Mean (±SE) activity level (i.e., difference between the numbers of line crossing between the pre- and post-stimulus periods) of tadpoles exposed to conspecific disturbance cues, water alone, or conspecific alarm cues

Experiment 2: retention of the learning association

Effects of conditioning differed significantly among the three groups (F2,168 = 36.71, P < 0.0001; Fig. 2). Post-stimulus activity of tadpoles relative to pre-stimulus activity increased significantly over time since the initial conditioning event (day effect: F3,168 = 5.60, P < 0.002). However, the interaction between factors was significant (F6,168 = 3.77, P < 0.002; Fig. 2). One day after the initial conditioning event, tadpoles from the two experimental groups significantly decreased activity in the post-stimulus period in comparison with control tadpoles (Tukey’s test, P < 0.001 in both cases), but post-stimulus changes in activity did not differ significantly between the two experimental groups (P = 0.99). However, 3 days after the conditioning event, tadpoles from the disturbance group increased their relative post-stimulus activity to the level of the “control” group (P = 0.99) and maintained this activity level on days 6 and 9 (P ≥ 0.9 in both cases). Tadpoles conditioned with alarm cues maintained their low relative post-stimulus activity level in comparison with the control and disturbance groups throughout the experiment (P ≤ 0.03 in all cases; Fig. 2).
https://static-content.springer.com/image/art%3A10.1007%2Fs00265-010-0927-y/MediaObjects/265_2010_927_Fig2_HTML.gif
Fig. 2

Mean (±SE) activity level (i.e., difference between the numbers of line crossing between the pre- and post-stimulus periods) of experimental and control tadpoles several days after the initial conditioning. Cues used were from non-predatory fish alone, fish cues mixed with conspecific alarm cues, or fish cues mixed with conspecific disturbance cues

Discussion

Our results indicate that Iberian green frog tadpoles are able to modify their antipredatory behavior according to the degree of risk associated with the cue to which they were exposed and according to their previous experience of chemical cues of the predator. The results of the first experiment show that the tadpoles display antipredator behaviors (i.e., a reduction in activity) in response to chemical cues released from disturbed conspecifics. This is a typical antipredatory response by Iberian green frog tadpoles in the presence of alarm cues (Gonzalo et al. 2007). The results also showed that reduction of activity levels is stronger in response to alarm cues than to disturbance cues. According to the threat-sensitive predator avoidance hypothesis, prey species should behave flexibly in relation to varying degrees of predator threat and, consequently, leave more time for other activities when the threat is low (Helfman 1989). If tadpoles react weakly to disturbance cues, it is possible that they assess risk based on the nature of the chemical cues and adaptively balance the costs and benefits of predator avoidance. Often the intensity of an animal’s antipredator response reflects the level of threat posed by the predator (Helfman 1989; Chivers et al. 2001). The type of chemical cue that an animal detects may be used to mediate the intensity of the antipredator response. The threat-sensitive predator avoidance hypothesis predicts that as the threat indicated by chemical cue decreases, the intensity of the antipredator response decreases (Helfman 1989). Disturbance cues are thought to indicate low risk, whereas alarm cues indicate high risk (Chivers and Smith 1998). Therefore, if tadpoles can asses the risk level, they should react differently to predators labeled with different types of chemical cues, as we have found in our experiment. Another explanation might be that the presence of disturbance cues elicits heightened vigilance, which could result in a decrease of activity similar to the typical antipredatory response (Ferrari et al. 2008). Differing concentrations of disturbance and alarm cues could be an alternative explanation for the differences in activity reduction, but in natural habitats, the concentration of alarm cues when a tadpole is trapped by a predator should be higher than the concentration of disturbance cues released by a tadpole that is stressed by a predator, but not captured. Therefore, differences in reduction of activity are more likely a consequence of the different nature of the two cues rather than a consequence of a different cues concentration.

The ability to learn and memorize cues indicating the presence of potential predators may be especially important for the survivorship of prey species that are likely to encounter a high variety of predators while they are in the aquatic life historical phase (Gonzalo et al. 2007). The results of the second experiment showed that tadpoles remember the fish chemical cues for 9 days if they had been conditioning by alarm cues. Tadpoles also reacted to the fish chemicals if they had been conditioning by disturbance cues, which implies that tadpoles are able to learn and remember at least for 1 day the potential predator labeled with the disturbance cues. However, tadpoles apparently forgot learned association with disturbance cues quickly because they behaved similarly as control tadpoles in all subsequent tests.

In natural conditions, tadpoles continually receive information about their environment and must filter this information to focus on those aspects most important to survival (Dukas 2002). This information includes a wide range of mixed chemical signal, (e.g., alarm cues, disturbance cues, predator chemical cues, non-predator cues). However, prey often do not have complete information about their environment and must estimate predation risk based on available information. This could lead them to either over- or underestimate risk (van der Veen 2002). In our experiments, we gave tadpoles incomplete or unreliable information regarding the identity of a new predator, allowing them to respond only to chemical cues from implying different levels of threat from a new potential predator. An antipredator response to learned predator chemical cues during the first day after the conditioning event, even if potential risk level is overestimated, could be adaptive. The differences in assessed risk level between the two types of cues (disturbance vs. alarm) was clear when we tested the tadpoles’ memory of learned association. Because tadpoles did not have additional information about the actual risk from the predator, decisions about antipredator behavior should be informed by recent experiences (Turner et al. 2005; Ferrari and Chivers 2006). Thus, a predator that is considered more dangerous because it is labeled with alarm cues should be remembered longer and elicit stronger antipredatory behaviors than a predator labeled with disturbance cues. In contrast to our results, juvenile trout are unable to associate disturbance cues with a new predator (Ferrari et al. 2008). This difference between studies might be explained by the differences in the time of exposure to the predator chemicals before the conditioning event. We found activity reduction in tadpoles 24 h after the conditioning, but responses had disappeared after 72 h, However, Ferrari et al. (2008) did not test fish until 48 h after their conditioning. Thus, it remains possible that the trout were able to retain the association between predator cues and disturbance cues only for a few hours. Alternatively, the learning capacities of tadpoles and trout may differ qualitatively.

Although disturbance cues have not been studied as well as the alarm cues, their involvement in decision making by tadpoles about activity level is evident. Mirza and Chivers (2002) demonstrated that the survival of juvenile brook charr was enhanced during encounters with predators if brook charr were previously conditioned with disturbance cues. Ferrari et al. (2008) argued that this result could be due to increased vigilance in response to disturbance cues. However, because the brook charr were tested a few hours after the conditioning, they may have retained the association between the two cues. Similarly, our data show that tadpoles are able to associate new predators with disturbance cues within a few hours and to retain the association between the chemical cues and the predator until the next day. However, tadpoles do not retain the association as long as they do when they have learned a predator cue by an association with alarm cues.

In a previous experiment, Iberian green frog tadpoles behaved flexibly toward a varying degree of predator threat: Some snakes normally perceived as low-risk predators were perceived as more threatening if tadpoles had been previously exposed to snake chemical cues mixed with alarm cues (Gonzalo et al. 2007). Our current data support the idea that Iberian green frog tadpoles behaved according to the threat-sensitive predator avoidance hypothesis when they were exposed to a predator labeled with disturbance cues or alarm cues. Alarm cues alone were perceived as a sign of high risk, whereas disturbance cues elicited a weaker reaction. Also, our results indicated that Iberian green frog tadpoles are able to modify their antipredatory behavior according to their previous experience of chemical cues of the predator. Thus, fish were perceived to be high-risk predators if the tadpoles had been previously exposed to fish cues mixed with alarm cues, but the same fish species was perceived as a dangerous predator only for 1 day if tadpoles had been exposed to fish cues mixed with disturbance cues. We have shown that there is a large difference in the duration of effects of two types of chemical cues, implying different degrees of threat on the antipredator behavior of the tadpoles.

Acknowledgments

We thank Carlos Cabido and two referees for helpful comments, Dr. William Cooper for his valuable help with the English, and ‘El Ventorrillo’ MNCN Field Station for the use of their facilities. Financial support was provided by a MEC-FPU grant to A.G. and by the projects MEC-CGL2005-00391/BOS and MCI-CGL2008-02119/BOS.

Copyright information

© Springer-Verlag 2010