Journal of Experimental Marine Biology and Ecology 369 (2009) 87–92
Contents lists available at ScienceDirect
Journal of Experimental Marine Biology and Ecology j o u r n a l h o m e p a g e : w w w. e l s e v i e r. c o m / l o c a t e / j e m b e
Cautious cannibals: Behavioral responses of juvenile and adult blue crabs to the odor of injured conspecifics F. Moir, M.J. Weissburg ⁎ School of Biology, 310 Ferst Drive, Georgia Institute of Technology, Atlanta, GA, 30332-0230
a r t i c l e
i n f o
Article history: Received 25 July 2008 Received in revised form 20 October 2008 Accepted 24 October 2008 Keywords: Blue crabs Chemical cues Odor plumes Predation Risk-sensitivity
a b s t r a c t Blue crabs are cannibalistic, and therefore the scent of injured conspecifics represents both a potential food cue, as well as an indicator of predation risk. We examined the response of blue crabs to conspecific odor alone, as well as in mixtures of attractive cues to determine how animals evaluate and respond to this odor. We explicitly manipulated risk-sensitivity based on either animal size (an indicator of susceptibility to predation) or hunger state (susceptibility to starvation) as ways to evaluate theories of risk-allocation, which suggest that decreases in predation risk, or increases in the risk of starvation, ought to result in diminished responses to sensory cues that signal predator presence or activity. Large and small blue crabs were challenged to locate the source of odor plumes consisting of the scent of injured conspecifics (risk cue), attractive food odors (attractive cue), or their mixture (conflicting cue). Neither large nor small blue crabs tracked aversive cues, but large blue crabs consistently tracked conflicting treatments to their source. Responses to conflicting and aversive treatments also involved diminished movement and reduced tracking speed relative to behaviors displayed in attractive plumes. Thus, even cannibalistic crabs seem to respond more prevalently to the apparent predation risk then to food reward, and risk-sensitive behaviors have a likely cost in terms of reduced food intake. Starved animals were more likely than unstarved animals to track conflicting plumes. Both the ontogenic shift and the response of starved animals support the notion that the cost of risk-aversive behaviors results in this strategy being allocated in proportion to the degree of potential risk. Since risk-aversive responses to chemical cues can produce strong effects in communities, the sizedependent nature of these responses in blue crabs may introduce considerable complexity in interactions between blue crabs, their predators, and their prey. © 2008 Elsevier B.V. All rights reserved.
1. Introduction Animals commonly detect potential predators in their environment and respond to this risk with behavioral or morphological changes that reduce their susceptibility (Kats and Dill, 1998; Lima and Dill, 1990; Stachowicz, 2001). This risk-sensitivity not only alters characteristics of the responding individual, but also affects a variety of other community members that interact with these individuals via predation or competition (i.e. trait mediated effects and indirect effects; (Werner and Peacor, 2003)). Current evidence from a variety of systems suggests these indirect effects exert significant impacts on community structure. Although a variety of sensory modalitities mediate risk-sensitivity (e.g. (Bouwma and Hazlett, 2001; Hazlett and McLay, 2005), the use of chemical cues is particularly common, and a staggering variety of aquatic animals sense, and respond to chemical cues emanating from predators or injured conspecifics and occasion-
⁎ Corresponding author. E-mail address:
[email protected] (M.J. Weissburg). 0022-0981/$ – see front matter © 2008 Elsevier B.V. All rights reserved. doi:10.1016/j.jembe.2008.10.026
ally, co-occurring or closely related heterospecifics (Kats and Dill, 1998). For many animals, some of the most lethal encounters are those with members of their own species. Thus, the chemical cue from an injured conspecific may represent an attractive substance indicating food to one individuual, and an indication of potential threat to another. Given that cannibalism generally results from encounters between individuals of disparate size, one might imagine that aversive responses to injured conspecifics are dominant in smaller individuals, whereas larger individuals may exhibit diminished aversion or perhaps, attraction, to such an odor. This speculation on the context-sensitivity of risk-aversion is consistent with the prevailing belief that risk sensitivity is costly (e.g. by diminishing foraging opportunities) and therefore, the degree of response should calibrated to the degree of potential threat (Chivers and Smith, 1998). By example, some prey species respond only to metabolites of specialist predators that recently have consumed that particular type of prey (Chivers and Mirzal, 2001). Prey species often show an ontogenic shift in risk-responsiveness, where older/larger prey show a reduced sensitivity to predator metabolites (Golub and Brown, 2003; Marcus
88
F. Moir, M.J. Weissburg / Journal of Experimental Marine Biology and Ecology 369 (2009) 87–92
and Brown, 2003). Such a dynamic also may characterize responses of cannibalistic species, but this has received scant attention. Blue crabs (Callinectes sapidus) present an attractive system to examine size-specificity in responses to risk. These ecologically and economically important species hunt using chemosensation (Weissburg and Zimmer-Faust, 1993; Zimmer-Faust et al., 1995), and are voracious predators on a variety of prey including conspecifics (Hines, 2003; Hines et al., 1990). Field tests using baited traps containing attractive prey and injured blue crabs or their metabolites confirm that injured animals release compounds that elicit avoidance (Ferner et al., 2005). However, these field studies also show that blue crabs do not universally avoid traps containing putative deterrent compounds, did not assay potential attraction to injured blue crabs or their metabolites, and did not address potential size specificity in risk-sensitivity. Here we use laboratory studies to focus on these latter two issues, and compare tracking behavior of blue crabs in response to attractive prey odors and putative deterrent compounds from injured crabs, singly and in combination. We explicitly examine the context-dependency of risk aversion as functions of both animal size and hunger level. Our results suggest that both of these factors modulate the expression of aversive behaviors, and that risk-aversion in blue crabs is variable and plastic. 2. Methods 2.1. Animal collection and maintenance Adult blue crabs were collected from the waters surrounding the Skidaway Institute of Oceanography in Savannah, GA and from a commercial supplier in Panacea, FL, using standard baited traps. We captured juveniles (b8 cm carapace width) by blocking the escape opening in our commercial traps, or by seining. Crabs were immediately tagged upon receipt at Georgia Tech, measured (width at lateral spines), sexed and placed in holding tanks. Animals were segregated by size whenever possible to minimize lethal encounters. The crabs were kept in 50 gal flow-through tanks equipped with protein skimmers, carbon, particle, and UV filters. Water quality measurements were done weekly to determine, ammonia, phosphorus, nitrate and salinity levels. Water was changed as needed (ca. 25% water change per week) to maintain salinity at 27-30 ppt and minimize the build-up of nitrogenous compounds and phosphates. Animals were kept on a 12:12 D:L cycle with a water temperature of 20± 1 °C. Crabs were fed to satiation (ca. 2 g of shrimp) every other day. Feeding schedules were adjusted to ensure that crabs were starved 24 h before trials. Newly arrived crabs acclimated a minimum of three days before being tested. 2.2. Solution preparation Blue crabs were presented with attractive and aversive compounds in order to examine their response to these individual odor sources as well as a mixture of the two conflicting stimuli. The attractive solution was derived from a food source, and was obtained from 15 g whole shrimp soaked for 1 hour in 2 l of artificial seawater (ASW). The aversive solution was made from 1 injured crab (punctured on the dorsal midcarapace with a 5 mm diameter metal rob) soaked for 3.5 hours in 3 l ASW. ASW used for each solution was obtained directly from the flume's sump to avoid density or chemical differences between the solutions and the surrounding water. Both the attractive and aversive solutions were used previously to examine blue crab chemosensory behavior (Ferner et al., 2005; Keller and Weissburg, 2004). 2.3. Experimental protocol Experiments to visualize C. sapidus behavior were carried out in a 12 m × 0.75 m recirculating salt water flume (Jackson et al., 2007). The
bulk flow velocity was steady at 5 cm s- 1, thus within the range of in situ flow speeds encountered by C. sapidus (Smee and Weissburg, 2006). Water depth was 0.20 m. Stimulus solutions were released from a 4.7 mm inner diameter nozzle located along the midline of the flume 25.4 mm above the top of the bed. A streamlined faring on the nozzle minimized the wake perturbation. Experiments were performed in dim illumination (crabs were not responsive to movements of the investigators beside the flume) to minimize visual cues and facilitate video motion analysis (see below), and tested between approximately 13:00-17:00. Solutions were introduced using a pressure-delivery system with inline flow meters into the nozzle apparatus. Experimental treatments consisted of attractive, aversive and conflicting (attractive + aversive). Attractive and aversive treatments consisted of the appropriate solution paired with ASW, whereas the conflicting treatment consisted of attractive+aversive stimuli. Each solution (ASW, aversive or attractive) was introduced at a rate of 20 ml min- 1. Thus, both the overall release rate (40 ml min-1) and the concentration of individual solutions were constant across the three treatments. Mixing of the solutions was facilitated by an enlarged chamber on the nozzle prior to it narrowing to the nominal diameter at the orifice. Tracking behavior in response to the above treatments was recorded with an overhead video camera and tracks reconstructed using video-motion analysis (see below). Large test subjects were outfitted with a silicone encased battery pack with 2 red lightemitting diodes. We used fluorescent “glow sticks” (2.5 cm long, 4 mm dia.) on smaller crabs. Animals were placed in a small cage 1.5 m downstream of the nozzle and acclimated in for 30 minutes. Chemical solutions were turned on two minutes before the end of the acclimation period to ensure the odor plumes reached the crab at the start of the trial. The trial began when the barrier was removed after 30 minutes, and the animal was allowed to respond freely to the chemical stimulus environment. The trial ended when the crab reached and attempted to grasp a nozzle or moved upstream beyond the nozzle. A trial was ended if 10 minutes elapsed without the crab traveling out of the trial area. At the end of each trial, the tested crab was presented with a small piece of shrimp to confirm the crab's willingness to eat. The trial was not included in the final data if the crab did not grab the shrimp decisively (characterized by a direct movement toward the shrimp and its subsequent capture). The percent of non-motivated animals was roughly 5-10%, and did not vary significantly across treatments. Experimental conditions and animals were randomized as much as possible. Small crabs were tested first because of high mortality rates. Large crabs were randomly netted in the holding tank and tested in that order to avoid stressing an animal by chasing it until it was caught. When possible, individuals from the same tank were not tested sequentially. Each batch of crabs was tested with all treatments, and we used multiple crab sizes and treatments on any given day. Individuals were tested only once, and returned to the field when the experiments were over. 2.4. Data analysis We examined degree of attraction to the source and kinematics of movement for each crab. Measures of attraction included time required to leave the cage once the barrier was lifted (exit time), and whether the animal found the source. We used video motion analysis to quantify the tracks. The position of the two LEDs or glow stick was digitized using a motion analysis system (Motion Analysis VP110). Raw data was acquired at 30 Hz during digitization, but path kinematics were calculated with data down-sampled at a final rate of 5 Hz. The digitized data was used to calculate each crab's speed (normalized for body length), the total amount of time it spent stopped, (defined as velocity b0.25 cm s- 1). We also calculated path linearity as the net to gross displacement ratio or NGDR. This metric
F. Moir, M.J. Weissburg / Journal of Experimental Marine Biology and Ecology 369 (2009) 87–92
ranges from 0 to 1, where 1 indicates a completely straight path from origin to destination. Data on percent tracking was examined with log-liklihood analysis to determine if there were associations between the response frequency (tracking vs. non-tracking vs. not exiting), size and treatment. The smallest crabs we could reliably capture and examine were 7-8 cm carapace width (CW), whereas the largest animals were between 16-17 cm CW. Thus, we sub-divided this range into roughly equal increments and defined small crabs as those less than 12 cm and large crabs as being greater than 12 cm. The resulting mean (±SD) CW for small and large crabs used in our experiments was 8.9 ± 1.6 cm, and 14.1 ± 1.3 cm, respectively (total sample sizes given in the Fig. 1), and there was no systematic variation in the sizes of these groups across treatments. Data on normalized walking speed, NGDR, stop time and exit time of tracking crabs all were normally distributed when examined by a Kolmogorov-Smirnov test, and so we used parametric statistics to examine the influence of size and chemical stimulus treatment on these behaviors. Only 1 small animal tracked to the aversive and conflicting cues, therefore a complete 2way ANOVA was not possible. Accordingly, we examined the behavior of large vs small animals in response to attractive cues, and the behavior of large animals as a function of chemical stimulus treatment. These analyses employed a t-test with significance values adjusted by a Bonferroni correction to account for multiple tests. Exit
89
time of non-tracking animals was not normally distributed since a large number of individuals exposed to aversive and conflicting treatments did not exit the start box in the allotted period. Thus, we examined exit times for non-tracking animals using the nonparametric Kruskal-Wallis test. Since non-parametric tests cannot partition variance, we had to analyze the effect of chemical stimulus separately for large vs. small animals. Comparisons between treatment pairs were subsequently examined by a Mann-Whitney U-test. All analysis were performed with the Systat software package (Systat Inc., Richmond CA.). 2.5. Effect of starvation on risk-aversion Our initial data (see results) indicated injured blue crab metabolites substantially reduced foraging. We therefore examined whether increasing the risk of starvation induced a subsequent shift towards greater tolerance to predation risk. We focused these experiments on small blue crabs given their extreme degree of risk aversion. In addition, small crabs were easy to house individually and respond to food deprivation more quickly. Crabs were housed individually using flow through plastic enclosures (approximately 20 × 20 × 35 cm) placed in our larger tank. These animals were randomly selected from the same groups of animals used in the experiments previously described. Crabs rarely tolerated starvation periods longer than 2 weeks, so we used this as our starvation period. Animals were subsequently tested in the manner described above to determine if tracking frequency to the conflicting cue increased relative to animals in the unstarved state. We succeeded in testing 10 (of 16) animals treated in this way, and we compared tracking success using a 2-way contingency table analysis. We elected to examine only the proportion of tracking vs non-tracking animals (i.e., we combined the no-exit and non-tracking categories) to reduce the number of cells with small sample sizes. 3. Results 3.1. Effect of aversive cues on blue crab tracking
Fig. 1. Response frequency for small and large crabs in response to chemical stimulus treatment. Figure shows the proportion of crabs that stayed in the starting box (stayed), left the box but did not find the source (left), or that successfully tracked to the source (tracked). A. Small crabs. B. Large crabs. Sample sizes are 24, 16, 28 for small crabs in the attractive, aversive and conflicting treatments, respectively. Samples sizes for large crabs are 19, 19, 18 in attractive, aversive and conflicting treatments, respectively.
Animal size, chemical stimulus composition, and their interaction affected the tracking success of blue crabs in response to chemical cues (Fig. 1A and B). Blue crabs presented with plumes containing injured crab metabolites tracked less often and remained in the start-box more often than crabs presented with plumes containing only attractive shrimp metabolites. The loglikelihood analysis shows a significant model effect on tracking frequency (χ2 = 38.8, df = 12, p b 0.0001) suggesting that both chemical composition and size interacted to determine the willingness or ability of crabs to track plumes. Main effects also were significant or marginally so (χ2 = 24.14, df = 4, p b 0.0001; χ2 = 5.5, df = 2, p b 0.064 for treatment and size, respectively. These results indicate that blue crab metabolites act like an aversive cue, generally reducing tracking when presented individually and in combination with prey odors. Examining crab responses by size class and treatment illustrates the aversive effect is particularly intense for small crabs (b12 cm CW), which accounts for the interaction effect. Small crab success rates fell from 38% in the attractive treatment to 6% and 4% in the aversive and conflicting treatments respectively. Success rates of large crabs (N12 cm CW) fell similarly between the attractive and aversive treatments (from 63% to 5%), but a considerable proportion of animals (33%) still tracked in the conflicting treatment. Non-tracking animals generally exited the starting box, particularly in the attractive treatment, and the proportion of crabs that leave is consistently greater than the proportion of crabs that stay in the box. However, a considerable number of animals failed to exit the box in treatments containing injured crab metabolites (aversive and conflicting treatments). The aversive treatment produces the largest proportion of
90
F. Moir, M.J. Weissburg / Journal of Experimental Marine Biology and Ecology 369 (2009) 87–92
individuals staying in the box, and this effect is greater in small vs. large animals. 3.2. Blue Crab Behavior Despite the difference in tracking success between large and small crabs (63% vs. 38%), behavioral parameters during tracking of attractive plumes are not significantly different between size classes. Small crabs took longer to exit than large crabs (Mean ± SE equals 40.29 ± 18.16 s, and 22.67 ± 7.8 s for small and large crabs, respectively, Fig. 2A), although exit times were not significantly different from each other. Both small and large crabs spent very little time stopped while tracking the plume (Mean ± SE: 2.8 s ± 1.72 and 3.79 s ± 2.15, respectively, Fig. 3A). Average normalized speed of small tracking crabs is 0.73 ± 0.18 body lengths s- 1 which is slightly higher than that of large tracking crabs at 0.54 ± 0.07 lengths s- 1 (Fig. 3B).Finally, net-to-grossdisplacement-ratios between small and large crabs were not significantly different (Mean ± SE: 0.67 ± 0.10 and 0.78 ± 0.04, respectively). The propensity of large crabs to track to mixtures of aversive and attractive cues allowed us to examine whether this treatment altered behavior compared to animals tracking attractive plumes. The exit times of these crabs were significantly different between treatments (t = 5.1, p b 0.001, df = 16; Fig. 2A). On average, crabs left the box less
Fig. 3. Analysis of kinematic data for tracking crabs. Figure shows stop times and tracking speed for small and large crabs in each chemical stimulus treatment. Bar gives mean and standard error. A. Time stopped. Sample sizes for large crabs equal 12,1,6 for attractive, aversive and conflicting cues respectively. Sample sizes for small crabs equal 5,1,1 for attractive, aversive and conflicting cues, respectively. B. Tracking speed. For each crab, the walking speed was divided by the carapace width to yield a normalized speed in body lengths s-1. Sample sizes for large crabs equal 12,1,6 for attractive, aversive and conflicting cues respectively. Sample sizes for small crabs equal 5,1,1 for attractive, aversive and conflicting cues, respectively. Sample sizes for kinematic analysis are less than that for tracking success and exit time since poor video quality prevented us from extracting kinematic data from some animals.
Fig. 2. Analysis of exit times. Figure shows exit times for small and large crabs in each chemical stimulus treatment. Bar gives mean and standard error. A. Tracking crabs. Sample sizes for large crabs equal 12,1,6 for attractive, aversive and conflicting cues respectively. Sample sizes for small crabs equal 9,1,1 for attractive, aversive and conflicting cues, respectively. B. Non-tracking crabs. Non-tracking includes crabs that left the box but did not find the source and crabs that stayed in the box. Sample sizes for large crabs equal 7,18,12 for attractive, aversive and conflicting cues respectively. Sample sizes for small crabs equal 15,15,27 for attractive, aversive and conflicting cues, respectively.
quickly in the conflicting treatment than they did in the attractive treatment (Mean ± SE equals 1255 ± 10 s, N = 6; 23 ± 8 s, N = 12, for conflicting vs. attractive treatments, respectively). Average normalized speed decreased significantly (t = 3.2, p = 0.023, df = 16 Fig. 3B) in the conflicting treatment compared to the attractive treatment (Mean ± SE: 0.244 ± 0.086, 0.556 ± 0.072 lengths s- 1, respectively). Although not significantly different (t = 1.5, p = 0.15; Fig. 3A), large crabs spent more time stopped while tracking the conflicting plume than they did in the attractive treatment (Mean ± SE:10.3 ± 4.43 s, 3.79 ± 2.15 s, respectively). Track linearity (NGDR) was statistically indistinguishable across conflicting and attractive treatments (Mean ± SE: 0.83 ± 0.061, 0.78 ± 0.043, respectively; t = 0.69, p N 0.25, df = 16). Average exit times were much higher in crabs that did not successfully find the source, compared to tracking animals, and were significantly related to treatment for both large and small crabs (Fig. 2B; Kruskal-Wallis test Statistic N 18.2, df = 2, p b 0.001 for both comparisons). Average exit times of non-tracking crabs were highest in the aversive treatment for both size classes (Mean ± SE: 375 ± 57 s and 334 ± 53 s, for small and large crabs respectively) and lowest in the attractive treatment (184 ± 50 s and 189 ± 106 s, respectively). For both
F. Moir, M.J. Weissburg / Journal of Experimental Marine Biology and Ecology 369 (2009) 87–92
large and small crabs, exit times for the crabs in aversive and conflicting treatments were both significantly greater than exit times in the attractive cue treatment as indicated by a Mann-Whitney U test (χ2 approximation N 11.6, df = 1, p b 0.001 for all comparisons). Although the aversive cue produced a longer exit time than the conflicting cue, the times were not significantly different as judged by the Mann-Whitney U (χ2 approximation b 1.8, df = 1, p N 0.10 for both comparisons). 3.3. Effect of starvation Starved small blue crabs were more likely to track in the conflicting plume treatment than unstarved individuals. Of the 10 animals that successfully survived the starvation period, 3 tracked (30%) when presented with the plume composed of blue crab and shrimp metabolites. This significantly exceeds the response (4%) of unstarved crabs to this same treatment (χ2 = 5.5, df = 1, p = 0.019). These same animals also were tested after 1 week, and although the lack of independence precludes a formal statistical analysis, the comparison between responses is informative. Notably, not a single animal of these 16 tracked after one week of starvation. Animals appeared sufficiently motivated, since all of them attempted to grab at a pipette which we used to present a small quantity of shrimp solution after testing. These observations suggest that starvation caused risk-averse individuals to forage, despite the presence of injured blue crab metabolites indicating a potential predatory threat. 4. Discussion Our data provide firm evidence that blue crabs rarely interpret metabolites from injured conspecifics as food, but rather, react to these substances with indifference or alarm. Crabs consistently refused to track to plumes wholly (aversive treatment) or partially (conflicting treatment) composed of blue crab metabolites, even as they often located odor sources composed only of shrimp metabolites (attractive treatment). In addition, animals that did not track to the aversive or conflicting plumes seemed to prefer shelter over movement, as evinced by long exit times and the enhanced proportion of animals that did not leave the box (shelter), compared to those behaviors in animals exposed to the attractive plume. Blue crabs tracked in plumes containing both injured blue crab and prey metabolites, even though the scent of injured conspecifics was not attractive when presented singly. Thus, the risk aversion displayed in response to injured blue crabs may be overridden when this aversive cue is paired with attractive odor. Our experiments suggest size-dependent plasticity in the degree to which crabs tolerate risk when foraging. Small crabs very rarely tracked to the conflicting treatment, and in fact, seemed to regard aversive and conflicting treatments as equally unattractive. In contrast, large crabs tracked conflicting plumes with an appreciable frequency that was intermediate between responses to aversive and attractive plumes. The response of blue crabs to plumes containing injured conspecifics is consistent with our prior understanding of riskaversion, but reveals this is a particularly powerful effect even in situations where conspecifics also represent food. It has been argued that animals should generally be more sensitive to riskaversion even when it prevents the expression of feeding because the selective disadvantage of death normally outweighs the advantage of finding food (the life-dinner principle, i.e. (Dawkins and Krebs, 1979)). Metabolites from conspecifics are an exceptionally strong case of the risk-reward tradeoff, as these metabolites represent both food and direct evidence of potential risk. Although examining responses of cannibalistic species to their own metabolites has been done rarely, our evidence that risk trumps reward
91
even in this instance suggests there may be extremely few situations animals resolve the life-dinner tradeoff in favor of a meal. The interaction between animal size and tracking success also is in accordance with current views of risk-sensitivity, which embody the idea that aversive responses entail a cost, and therefore should be employed with an intensity that depends on the degree of risk (Chivers and Smith, 1998; Hazlett, 2000). Blue crabs show a diminished propensity to search when presented with plumes containing both attractive and aversive compounds, take longer to initiate search, and search more slowly. These responses illustrate the potential cost of aversive behaviors in terms of reduced energy intake. Larger animals should be more risk tolerant since size often confers an immunity to predation; juvenile blue crabs seem particularly vulnerable to larger conspecifics (Hines, 2003). The speculation that blue crabs should allocate their foraging time as a function of risk is consistent with our observation that large animals track to food odors in the presence of injured crab metabolites whereas juveniles almost never do. Ontogenic or size-dependent shifts in chemically-mediated risk aversion are well documented in fish (Golub and Brown, 2003; Marcus and Brown, 2003), and crustaceans (Hazlett, 2003) where each of these studies indicates that larger animals are more likely to forage in the presence of aversive chemical cues than are smaller individuals. This issue does not seem to have received much attention in crustaceans despite evidence for size-dependent anti-predator strategies (Wasson and Lyon, 2005). The potential cost of forgoing foraging opportunities escalates as the risk of starvation increases, and a recent review of risk aversion (Chivers and Smith, 1998) cautioned investigators to include analysis and manipulation of hunger level. Elevated tolerance to risk has been seen in fish (Brown and Smith, 1996), and gastropods (Vadas et al., 1994), which conclude that starved animals increase foraging behaviors in the presence of aversive chemical stimuli relative to the responses seen in unstarved individuals. Similarly, we were able to induce small, starved animals to track in the conflicting cue when unstarved animals do so extremely rarely. However, a high level of food deprivation was required before small crabs responded to starvation as opposed to predation risk. We were commonly unable to keep starved small individuals longer than 2 weeks despite the fact that they survive in captivity quite well under normal conditions. The mortal severity of this depravation period is consistent with studies on other crustaceans (Hazlett et al., 1975). Interestingly, related crayfish species vary in their sensitivity to starvation, which may partially explain species-specific demographic differences (Hazlett, 2003). Although we did not perform similar experiments with large crabs, their greater tolerance for risk argues starvation ought to increase their tendency to forage when presented with conflicting cues. Our experiments employed a paradigm of combining attractive with aversive odors and examining the resulting willingness of animals to engage in chemosensory mediated search. This is a particularly relevant scenario for blue crabs and other chemosensory foragers, since they operate in an environment where they are confronted with multiple odor sources each of which may convey specific information about the availability of food, mates or the degree of potential mortal threat. Blue crabs clearly are able to interpret this olfactory mélange to make decisions based on a risk-reward paradigm, and the expression of these effects in natural situations may have important ecological consequences. The general suppression of foraging by metabolites from injured conspecifics has the potential to alter other aspects of community structure. Observations in both marine and freshwater systems indicate that chemical cues can have cascading effects throughout a community by altering the behavior of intermediate-level predators such as blue crabs (Trussel et al., 2003; Turner et al., 2000; Werner and Peacor, 2003). Blue crabs are known to limit the distribution and abundance of other community members, and risk from vertebrate predators can reduce the impact of blue crab predators by suppressing
92
F. Moir, M.J. Weissburg / Journal of Experimental Marine Biology and Ecology 369 (2009) 87–92
their foraging (Eggleston et al., 1984; Hines et al., 1990; Micheli, 1997). Our data show that metabolites from injured blue crabs diminishes both the frequency of search and the amount of movement. Both of these effects could result in habitat-specific activity patterns that change community structure by lessening predation pressure blue crabs exert on other species. Traps baited with menhaden fish (Brevoortia sp.) combined with injured crabs or their metabolites catch fewer blue crabs than traps containing only fish (Ferner et al., 2005) confirming this effect in the field. An important consideration is that flow-induced mixing reduces the navigational ability of blue crabs and other chemosensory foragers (Finelli et al., 2000; Weissburg, 2000; Weissburg and Zimmer-Faust, 1993), so the prevailing flow regime may limit or alter the expression of risk-sensitivity. Such effects have not been well studied, but limited experimental evidence indicates turbulent mixing alters the scales over which animals respond to water-borne substances that convey information regarding predation risk in laboratory and natural settings (Smee and Weissburg, 2006; Smee, et al., 2008). In addition, the blue crab system has the potential for complex interactions resulting from the combination of size specific riskresponsiveness and predation. The greater sensitivity of small blue crabs to conspecific metabolites suggests that foraging and activity patterns of large vs small animals may be quite dissimilar even in the same habitat. Larger blue crabs are able to crush larger bivalve prey, and preferred size of bivalve prey increases with animal size (Arnold, 1984; Seed, 1980). Larger animals may exhibit inverse density dependent functional responses on smaller prey (Eggleston, 1990a, b). Thus, the differential risk-responsiveness of large vs small blue crabs to conspecific metabolites may alter the predation pressure faced by specific size classes of bivalves, potentially creating surprising or counter-intuitive effects. By example, predation on large blue crabs may suppress small crab foraging more completely and therefore release small bivalves from predation. Such an effect may not be obvious by simply examining size-dependencies of pairs of predators and prey without accounting for potential differences stemming from size-specificity in risk aversion. An emerging theme of modern community ecology is that behavioral responses considerably increase the complexity of interactions between even small numbers of species, and the size-specific relationships of predation pressure and risk aversion may be another example of this phenomenon. Acknowledgements Thanks to J. Jackson, J. Hill and M. Watts for animal collection and care, and T. Spellman for assisting in experiments with starved animals. This work was supported by NSF-OCE #0424673 to M.J. Weissburg and D. Webster, and support from the Department of Defense ASSURE (Awards to Stimulate and Support Undergraduate Research Experiences) in conjunction with the National Science Foundation REU program, administered as NSF-OCE #0354033 to M. J. Weissburg and F. Loeffler.[RH] References Arnold, W.S., 1984. The effects of prey size, predator size and sediment composition on the rate of predation of the blue crab Callinectes sapidus Rathbun on the hard clam Mercenaria mercenaria (Linne). Journal of Experimental Marine Biology and Ecology 80, 207–219. Bouwma, P., Hazlett, B.A., 2001. Integration of multiple predator cues by the crayfish Orconectes propinquus. Animal Behaviour 61, 771–776. Brown, G.E., Smith, R.J.F., 1996. Foraging trade-offs in fathead minnows (Pimephales promelas, Osteichthyes,Cyprinidae ): acquired predator recognition in the absence of an alarm response. Ethology 102, 776–785. Chivers, D.P., Smith, R.J.F., 1998. Chemical alarm signaling in aquatic predator-prey systems: A review and prospectus. Ecoscience 5, 338–352. Chivers, D.P., Mirzal, R.S., 2001. Importance of predator diet cues in responses of larval wood frogs to fish and invertebrate predators. Journal of Chemical Ecology 27, 45–51.
Dawkins, R., Krebs, J.R., 1979. Arms races between and within species. Proceedings of the Royal Society of London B. 205, 489–511. Eggleston, D.B., 1990a. Foraging behavior of the blue-crab, Callinectes Sapidus, on juvenile oysters, Crassostrea Virginica - Effects of prey density and size. Bulletin of Marine Science 46, 62–82. Eggleston, D.B., 1990b. Functional-responses of blue crabs Callinectes sapidus rathbun feeding on juvenile oysters Crassostrea virginica (Gmelin) - Effects of predator sex and size, and prey size. Journal of Experimental Marine Biology and Ecology 143, 73–90. Eggleston, D.B., Lipcius, R.N., Hines, A.H., 1984. Density-dependent predation by blue crabs upon infaunal clam species with contrasting distribution and abundance patterns. Marine Ecology Progress Series 85, 55–68. Ferner, M.C., Smee, D.L., Chang, Y.P., 2005. Cannibalistic crabs respond to the scent of injured conspecifics: danger or dinner? Marine Ecology Progress Series 300, 193–200. Finelli, C.M., Pentcheff, N.D., Zimmer, R.K., Wethey, D.S., 2000. Physical constraints on ecological processes: A field test of odor-mediated foraging. Ecology 81, 784–797. Golub, J.L., Brown, G.E., 2003. Are all signals the same? Ontogenetic change in the response to conspecific and heterospecific chemical alarm signals by juvenile green sunfish (Lepomis cyanellus). Behavioral Ecology and Sociobiology 54, 113–118. Hazlett, B.A., 2000. Responses to single and multiple sources of chemical cues in New Zealand crustaceans. Journal of Marine and Freshwater Behavior and Physiology 34, 1–20. Hazlett, B.A., 2003. The effects of starvation on crayfish responses to alarm odor. Ethology 109, 587–592. Hazlett, B.A., McLay, C., 2005. Responses to predation risk: alternative strategies in the crab Heterozius rotundifrons. Animal Behaviour 69, 967–972. Hazlett, B.A., Rubenstein, D., Rittschof, D., 1975. Starvation, energy reserves, and aggression in the crayfish Orconectes virilis (Hagen, 1870) (Decapoda, Cambaridae). Crustaceana 28, 11–16. Hines, A.H., 2003. Ecology of juvenile and adult blue crabs:summary of discussion of research themes and directions. Bulletin of Marine Science 72, 423–433. Hines, A.H., Haddon, A.M., Wiechert, L.A., 1990. Guild structure and foraging impact of blue crabs and epibenthic fish in a subestuary of Chesapeake Bay, vol. 67, pp. 105–126. Jackson, J.L., Webster, D.R., Rahman, S., Weissburg, M.J., 2007. Bed roughness effects on boundary-layer turbulence and consequences for odor tracking behavior of blue crabs (Callinectes sapidus). Limnology and Oceanography 52, 1883–1897. Kats, L.B., Dill, L.M., 1998. The scent of death: Chemosensory assessment of predation risk by prey animals. Ecoscience 5, 361–394. Keller, T.A., Weissburg, M.J., 2004. Effects of odor flux and pulse rate on chemosensory tracking in turbulent odor plumes by the blue crab, Callinectes sapidus. Biological Bulletin 207, 44–55. Lima, S.L., Dill, L.M., 1990. Behavioral decisions made under the risk of predation - a review and prospectus. Canadian Journal of Zoology-Review of Canadian Zoology 68, 619–640. Marcus, J.P., Brown, G.E., 2003. Response of pumpkinseed sunfish to conspecific chemical alarm cues: an interaction between ontogeny and stimulus concentration. Canadian Journal of Zoology-Review of Canadian Zoology 81, 1671–1677. Micheli, F., 1997. Effects of predator foraging behavior on patterns of prey mortality in marine soft bottoms. Ecological Monographs 67, 203–224. Seed, R., 1980. Predator-prey relationships between the Mud Crab Panopeus herbstii, the Blue Crab Callinectes sapidus and the Atlantic Ribbed Mussel Geukensia (= Modiolus) demissa. Estuarine and Coastal Marine Scince 11, 445–458. Smee, D.L., Weissburg, M.J., 2006. Claming up: environmental forces diminish the perceptive ability of bivalve prey. Ecology 87, 1587–1598. Smee, D.L., Ferner, M.C., Weissburg, M.J., 2008. Environmental conditions alter prey reactions to risk and the scales of nonlethal predator effects in natural systems. Oecologia 10.1007/s00442-008-0995-0. Stachowicz, J.J., 2001. Chemical ecology of mobile benthic invertebrates: predators and prey, allies and competitors. In: McClintock, J.B., Baker, B.J. (Eds.), Marine Chemical Ecology. CRC Press, LLC, pp. 157–194. Trussel, G.C., Ewanchuk, P.J., Bertness, M.D., 2003. Trait-mediated effects in rocky intertidal food chains:predator risk cues alter prey feeding rates. Ecology 84, 629–640. Turner, A.M., Bernot, R.J., Boes, C.M., 2000. Chemical cues modify species interactions: the ecological consequences of predator avoidance by freshwater snails. Oikos 88, 148–158. Vadas, R.L., Burrows, M.T., Hughes, R.N., 1994. Foraging strategies of dogwelks, Nucella lapillus (L.): interacting effects of age,diet and chemical cues to the threat of predation. Oecologia 100, 439–450. Wasson, K., Lyon, B.E., 2005. Flight or fight: flexible antipredatory strategies in porcelain crabs. Behavioral Ecology 16, 1037–1041. Weissburg, M.J., 2000. The fluid dynamical context of chemosensory behavior. Biological Bulletin 198, 188–202. Weissburg, M.J., Zimmer-Faust, R.K., 1993. Life and death in moving fluids: Hydrodynamic effects on chemosensory-mediated predation. Ecology 74, 1428–1443. Werner, E.E., Peacor, S.D., 2003. A review of trait-mediated indirect interactions in ecological communities. Ecology 84, 1083–1100. Zimmer-Faust, R.K., Finelli, C.M., Pentcheff, N.D., Wethey, D.S., 1995. Odor plumes and animal navigation in turbulent water flow. A field study. Biological Bulletin 188, 111–116.