381.qxd

10/25/2007

3:13 PM

Page 263

Chemoecology 17: 263–266 (2008) 0937-7409/07/040263-4 © Birkhäuser Verlag, Basel, 2007 DOI 10.1007/s00049-007-0381-0

CHEMOECOLOGY

Short communication Degradation of chemical alarm cues under natural conditions: risk assessment by larval woodfrogs Maud C.O. Ferrari, François Messier and Douglas P. Chivers Department of Biology, University of Saskatchewan, Saskatoon, SK S7N 5E2 Canada

Summary. Many aquatic species use chemosensory information to assess predation risk. The cues used in such risk assessment can come either from the predator (predator odour) or from injured prey (alarm cues). The information conveyed through chemicals may, however, be inaccurate both spatially and temporally, as chemicals may persist in the environment long after the predator is gone. Thus, the level of accuracy of the cues for risk assessment may depend on the persistency of the chemicals in the habitat. Here, we investigated the persistency of alarm cues of a larval amphibian, the woodfrog (Rana sylvatica) in a ephemeral pond, their natural habitat. We introduced either alarm cues or control water in enclosed sleeves (∼10 L) installed in the pond. The sleeve water was then sampled after 5 min and every two hours for eight hours. We used the behavioural response of woodfrog tadpoles to alarm cues as a bioassay to assess how long the alarm cues persisted in the environment. We found that tadpoles responded with an antipredator response to the pond water containing alarm cues 5 min after the injection of the cues in the sleeves but did not respond to that same pond water after two hours. Our results indicate that biodegradation and/or photodegradation of alarm cues in natural habitats might occur relatively quickly as the loss of a response to the cues in our experiment was independent of a dilution effect. This contrasts with previous laboratory results indicating that chemicals may be active after several hours. Key words. Alarm cues – degradation – persistency – field experiment – risk assessment – larval woodfrog – Rana sylvatica

Introduction A pre-requisite for effective antipredator responses is that prey animals are able to recognize potential threats (Lima and Dill 1990). The information that prey use in such risk assessment has received considerable attention from ecologists (Chivers and Smith 1998). In aquatic systems, prey animals have been shown to use chemosensory information to recognize potential danger. The source of the chemicals can arise from either the predators themselves (i.e., predator Correspondence to: Maud C.O. Ferrari, e-mail [email protected]

odours, Kats and Dill 1998) or alternatively from cues released by injured prey (i.e. alarm cues, Chivers and Smith 1998; Wisenden and Chivers 2005). Chemical cues provide prey animals with a valuable source of information particularly in conditions in which other sensory modalities, such as vision, are limited. Such conditions may occur at night, in turbid water or high structured habitats. However, chemical cues indicating predation risk may not reliably indicate the true threat that the prey are exposed to if chemicals persist in the environment after the predator has left the area. The disconnection between perceived and actual risk has received little attention from chemical ecologists. We know of only one published study that indirectly tested the persistency of alarm cues. In fact, Hazlett (1999) concluded that alarm cues of crayfish (Orconectes virilis) may stay active for at least 6 hours. However, this experiment was done under artificial conditions and may not be ecologically relevant. Information on the persistency of chemical cues under natural conditions is vital if we are to begin to understand how accurate the information is that prey use in risk assessment. Here, we investigated the persistency of amphibian alarm cues under natural conditions. Given that the chemical structure of woodfrog alarm cues is not known, we used the responses of woodfrog tadpoles to alarm cues as a bioassay to determine the persistency of the cues. After injecting a biologically relevant concentration of alarm cues in enclosed areas of a small ephemeral pond, we sampled the pond water after 5 min and every 2 hours for 8 hours. The behavioural responses of the tadpoles would indicate whether or not the alarm cues are still active and detectable. We chose to inject the cues in enclosed areas, to be able to differentiate a dilution effect (the solution becoming too diluted might explain the absence of antipredator response to the alarm cues) from a degradation effect.

Methods Study species Woodfrogs (Rana sylvatica) are commonly found across Canada. Explosive breeders, the adults usually lay their eggs within a couple of weeks in a given pond, in late April and early May at our site location. The eggs hatch usually after 1-2 weeks and the larvae metamorphose within a few weeks. Woodfrogs are locally migrant, and are usually found within several hundred meters from the

381.qxd

10/25/2007

264

3:13 PM

Page 264

M.C.O. Ferrari, F. Messier and D.P. Chivers

breeding site. The juveniles reach sexual maturity after 2 years (Duellman & Trueb 1994). Larval woodfrog, like many species of larval amphibians, possess chemicals (hereafter referred to as alarm cues) which elicit antipredator responses in nearby conspecifics (Chivers et al. 1999, Hews 1988, Hews & Blaustein 1985, Petranka 1989). The cues are released following mechanical damage, as would occur during a predatory attack. Behavioural responses to alarm cues include a reduction in activity which is the same response shown to cues from predators (Chivers et al. 1999, Ferrari et al. 2007). Ferrari et al. (unpublished data) showed that woodfrog tadpoles displayed a threat-sensitive response to alarm cues, increasing the intensity of their antipredator response when exposed to increased concentrations of alarm cues. Water and test subjects Five weeks prior to starting the experiment, a 1900-l tub was filled with well water and left outdoors. The water was seeded with aquatic plants (sedges: Carex spp, slough grass, horsetail: Equisetum spp.), zooplankton and phytoplankton from a local pond using a fine mesh dip net. This ensured that the holding and testing water contained a full array of algae and plankton but no cues from potential predators present in the local pond water. This water is hereafter referred to as ‘well water’. Eight woodfrog egg clutches were collected in early May from a pond in central Alberta, and transferred into a pool (60 cm diameter) containing pond water and aquatic plants. The pool was positioned on the pond to equalize the temperature of the pool water with the pond water. After hatching, the tadpoles were raised for two weeks. Rabbit food was provided to supplement the algae already present in the pool. Experimental protocol The goal of this experiment was to test the speed of degradation of amphibian alarm cues in a natural pond habitat. We used the behavioural response of woodfrog tadpoles to alarm cues as a bioassay to assess the persistency of alarm cues in the pond water. Two days before the start of the experiment, eight rectangular sleeves (white food-grade plastic: 22 × 22 × 20 cm) were installed in a local pond, a known breeding site for wood frogs. The eight sleeves were installed 1.5 m from the north shore of the pond and were all exposed to the same light conditions. At each site, the sleeves were pushed into the mud, to obtain an enclosed volume of the pond containing a variety of micro-organisms, plants and insects. The sleeves did not have any bottom, thus the water in the sleeves was in contact with the substrate. The top of the sleeves were 1-2 cm above the water surface. Larval woodfrogs often occur in large aggregations of several hundred tadpoles. Insects predators, like beetles or larval dragonflies, often injure several tadpoles in an attempt to catch one (personal observation). Likewise, vertebrate predators such as salamanders and snakes can rapidly capture many tadpoles in succession. Consequently, the amount of alarm cues released in one area may be substantial. Hence, in our experiment, we injected the cues of 20 crushed tadpoles to obtain a biologically relevant concentration of alarm cues per sleeve. Sleeves were set in pairs, one randomly chosen receiving 60 mL of well water and the other one receiving 60 mL of alarm cues made from the 20 crushed tadpoles (using a mortar and pestle) in 60 mL of well water. The first three pairs of sleeves were sampled shortly after introducing the stimuli (+5 min), and every two hr for eight hr. The time of stimuli introduction was offset by ∼ 40 min among the pairs (0920 hr, 0955 hr and 1035 hr). This allowed the video recording of 12 tadpoles for each of the two cues, and an acclimation period of 30 min for the next groups to be tested. Thirty min after recording the last group of tadpoles (+8 hr, third pair of sleeves), we injected fresh well water and alarm cues in the fourth pair of sleeves and two groups of 12 tadpoles were tested immediately (+5 min) for each of the two cues. This was done to ensure that tadpoles displayed a response to alarm cues at this time of day. The absence of behavioural response in the first three sleeves later in the day may be due to a lack of behavioural response to alarm cues (due to time

CHEMOECOLOGY of day effect), or to a lack of detection of the cues by the tadpoles (degradation effect). Behavioural assay and testing procedure Larval anuran amphibians, including woodfrog tadpoles, have been shown to decrease activity in response to predation cues (Chivers & Mirza 2001). Hence, a line was drawn in the middle of the testing cups (0.5 L) and the number of line crosses was counted during the observation periods. We considered that a tadpole crossed a line when its entire body was on the other side of the line. Prior to testing, single tadpoles were transferred into a testing cup containing well water and left to acclimate for 30 min. The trials consisted of a 4-min pre-stimulus followed by a 4-min poststimulus injection period during which the behaviour of the tadpole was recorded (number of line crosses). The two periods were separated by a 30-sec injection period, during which the contents of a five mL syringe were emptied slowly on the side of the cup to minimize disturbance. Tadpoles were exposed to five mL of water from a sleeve containing either well water or alarm cues. Tadpoles were tested in groups of 12, hence we used a 60 mL syringe to withdraw water from the sleeves. The syringe was placed towards the bottom of the sleeve and filled while pulling the syringe towards the surface, to ensure the sampling of cues from different positions in the water column. Statistical analysis First, we assessed whether there was a time of day effect (i.e., sleeve effect) in the responses of tadpoles to cues. The poststimulus data obtained from tadpoles tested with cues from each of the four sleeves at +5 min were analyzed using a one-way nested ANCOVA, using cue (water or alarm cues) as the main factor, the time of day as the nesting factor and the pre-stimulus data as covariables. Second, we used the data from the first three pairs of sleeves to assess the effect of time on the response of tadpoles to water and alarm cues. Because the results of the previous analysis revealed no effect of time of day (i.e., no difference among sleeves), the data for the three pairs of sleeves were pooled (3 × 12 tadpoles for water and 3 × 12 tadpoles for alarm cues). To investigate at which time the tadpoles stopped showing an alarm response to alarm cues, we performed a one-way ANCOVA at each time (+5 min, +2 hr, +4 hr, +6 hr and +8 hr), using cue (water or alarm cues) as the main factor and pre-stimulus data as covariables. Three a posteriori repeated-measures ANCOVAs were performed on the alarm cues data only to compare the intensities of response of tadpoles within the first 4 hr. For each ANCOVA, the covariables followed the assumption of no-interaction with the main factor.

Results The results of the nested ANCOVA revealed that cue type affected the responses of tadpoles, with tadpoles exposed to sleeve water containing alarm cues decreasing activity significantly more than tadpoles exposed to sleeve water containing water alone (F1,5 = 18.1, P = 0.008). Time of day, however, did not affect the response of the tadpoles (F5,75 = 1.4, P = 0.217). The results of the time-series revealed that at + 5 min, tadpoles responded to the sleeve water containing alarm cues with a higher intensity response than they did to the sleeve water containing water alone (F1,68 = 21.8, P < 0.001, figure 1). However, at +2hr, we failed to find a difference in the intensity of response of tadpoles to sleeve water containing alarm cues or water (F1,67 = 0.03, P = 0.866). The same results were found at +4hr (F1,66 = 0.8, P = 0.374), at +6hr (F1,69 = 3.4, P = 0.067) and at +8hr (F1,66 = 0.03, P = 0.851, figure 1). A posteriori repeated-measures revealed that tadpoles did not

10/25/2007

3:13 PM

Page 265

Vol. 17, 2007

Degradation of alarm cues under field conditions 10 5 0 −5

b

−10 −15 −20 −25

265

Fig. 1 Mean (± S.E.) change in movement from the pre-stimulus baseline of tadpoles exposed to sleeve water containing well water (solid bars) or alarm cues (empty bars) at different times after the injection of the cues in the sleeves.

∗ Mean change in movement from the prestimulus period

381.qxd

ab a

+ 5 min

+ 2 hr

+ 4 hr

differ in their intensities of antipredator response to alarm cues at +5 min and +2hr (F1,30 = 0.17, P = 0.684), or at +2hr and +4hr (F1,28 = 0.87, P = 0.360). However, the tadpoles responded with a stronger intensity response to alarm cues at +5 min than +4hr (F1,29 = 4.8, P = 0.035).

Discussion The results of our study demonstrate that larval wood frogs exhibit an antipredator response to a solution of alarm cues five minutes after the solution was released into a natural pond. However, after the stimulus remained in the pond for a few hours, the tadpoles failed to exhibit a response, when compared to a water control. Four hours following injection of the cues into the pond, the tadpoles showed a statistically lower intensity response than the tadpoles showed five minutes after the injection into the pond. The response of the tadpoles to cues after two hours post-injection was intermediate (not different from either the five minute or four hour post-injection). Despite hundreds of studies examining the responses of aquatic organisms (invertebrates, fishes, amphibians) to alarm cues and predator odours, we know of no studies examining how fast the cues breakdown under natural conditions. In one study, Hazlett (1999) showed that crayfish alarm cues persisted for more than six hours. However, in this experiment, the odour was collected in clean dechlorinated tap water. Such as design may overestimate how long the cues persist given that the absence of sunlight could reduce photodegradation and the lack of bacterial fauna could limit biodegradation. In our experiment, all of the sleeves were exposed to natural sunlight, consequently, the cues could have been broken down through photodegradation, in addition to biodegradation. Note that the loss of response was not due to the dilution of the alarm cues, as the cues were injected in an enclosed space. Moreover, tadpoles respond to fresh alarm cues throughout the day (our results and unpublished data). Hence, degradation of alarm cues is the only explanation for the loss of response. Information on how fast alarm cues and predator odours degrade under natural conditions is very valuable for researchers trying to understand risk assessment by prey. Visual information about predation risk is spatially and temporally reliable (i.e. the prey can see the exact location of the

+ 6 hr

+ 8 hr

predator in real time). Chemical cues are often not as reliable either spatially or temporally. The predator may have captured a prey and left the area some time ago. Different concentrations of cues could indicate that the threat is either closer or further away (distance indicator) or that the stimulus is fresh or partially degraded (temporal indicator) (Ferrari et al. 2006). Given the ubiquitous nature of chemosensory risk assessment by prey, it is surprising that we know almost nothing about how long the cues persist under natural conditions. Our work provides an initial framework for future studies. Based on our results, we suggest that researchers conduct studies over shorter time intervals. Comparative studies examining the persistence of cues among different taxa are worth pursuing. Is there selection for prey to have alarm cues that persist for different time periods? Researchers should also consider whether alarm cues degrade at different rates depending on environmental conditions. Differences in water chemistry could influence how long the chemicals persist. Likewise, researchers should consider the effects of temperature and water clarity (amount of dissolved organic carbon) on the persistence of alarm cues.

Acknowledgements We thank Jean & Glen Chivers for their help and support and for letting us invade their home and wetlands for the duration of our field season. Research funding was provided to M. Ferrari through the R. J. F. Smith Memorial Scholarship, and to F. Messier and D. Chivers through the Natural Sciences and Engineering Research Council of Canada. All work reported was in accordance with the Animal Care Committee Protocol # 20060014 from University of Saskatchewan.

References Chivers DP, Mirza RS (2001) The importance of predator-diet cues in the responses of larval woodfrogs to fish and invertebrate predators. J. Chem. Ecol. 27: 45–51 Chivers DP, Smith RJF (1998) Chemical alarm signaling in aquatic predator/prey interactions: a review and prospectus. Ecoscience 5: 338–352

381.qxd

10/25/2007

266

3:13 PM

Page 266

M.C.O. Ferrari, F. Messier and D.P. Chivers

Chivers DP, Kiesecker JM, Wildy EL, Belden LK, Kats LB, Blaustein AR (1999) Avoidance response of juvenile anurans to cues of injured conspecifics and predators. Journal of Herpetology 33: 472–476 Duellman WE, Trueb L (1994) Biology of amphibians. John Hopkins University Press Ferrari MCO, Messier F, Chivers DP (2007) First documentation of cultural transmission of predator recognition by larval amphibians. Ethology 113: 621–62 Ferrari MCO, Messier F, Chivers DP (2006) The nose knows: minnows determine predator proximity and density through detection of predator odours. Animal Behaviour 72: 927–932 Hazlett BA (1999) Responses to multiple chemical cues by the crayfish Orconectes virilis. Behaviour 136: 161–177 Hews DK (1988) Alarm response in larval western toads, Bufo boreas: release of larval chemicals by a natural predator and its effect on predator capture efficiency. Animal Behaviour 36: 25–33

CHEMOECOLOGY Hews DK, Blaustein AR (1985) An investigation of the alarm response in Bufo boreas and Rana cascadae tadpoles. Behav. Neural Biol 43: 47–57 Kats LB, Dill LM (1998) The scent of death: chemosensory assessment of predation risk by prey animals. Ecoscience 5: 361–394 Lima SL, Dill LM (1990) Behavioral decisions made under the risk of predation: a review and prospectus. Canadian Journal of Zoology 68: 619–640 Petranka JW (1989) Response of toad tadpoles to conflicting chemical stimuli: predator avoidance versus “optimal’’ foraging. Herpetologica 45: 283–292 Wisenden BD, Chivers DP (2005) The ubiquitous bouquet: the role of public chemical information in antipredator behaviour in fishes. Fish Chemoreception. pp. 259–278. (Ladich F, Collins SP, Moller P, Kapoor BG, eds), Science Publisher

Received 7 May 2007; accepted 7 July 2007. Published Online First 17 September 2007.

/chemo