Behavioral Ecology Advance Access originally published online on January 25, 2006
Behavioral Ecology 2006 17(3):392-398; doi:10.1093/beheco/arj038
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Responses of domestic chicks (Gallus gallus domesticus) to multimodal aposematic signals
Department of Biology, University of Oslo, P.O. Box 1066 Blindern, N-0316 Oslo, Norway
Address correspondence to H.M. Lampe. E-mail: h.m.lampe{at}bio.uio.no. S.B. Hagen is now at the Department of Biology, University of Tromsø, N-9037 Tromsø, Norway.
Received 7 July 2003; revised 21 December 2005; accepted 2 January 2006.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Many aposematic prey combine their visual warning signals with additional signals. Together, these signals constitute a multimodal or multicomponent warning display. The additional signals are thought to increase the effects of the visual signals on predators. Olfactory signals are much emphasized, but later studies have shown that also auditory signals like the buzzing of certain insects might have multimodal effects. The wasp displays typical visual aposematic signals, black and yellow stripes, but does also emit a characteristic buzzing. We wanted to test if, and in what way, the visual and acoustic display of the wasp has an aversive function on the predators. We therefore conducted a 12-trial discrimination-learning task on inexperienced chicks to study whether there are innate biases toward these signals and how they affect the speed of avoidance learning. We also performed three extinction-learning trials to study how memorable the signals were to the chicks. We show that the visual signals in the display of the wasp contribute to the protection from predators but in different ways; the yellow color had an aversive effect on inexperienced predators, while the striped pattern improved the aversion learning. The sound did not enhance the innate aversions but increased the aversion learning of stripes in green prey.
Key words: aposematism, avoidance learning, chicks, innate biases, multimodal signals, sound, stripes.
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
The theory of warning signals, or aposematism (Poulton, 1890
), describes the phenomenon that unpalatable prey may use conspicuous signals such as bright colors (Roper, 1990) or conspicuous patterns (Schuler and Hesse, 1985) to warn predators about their unprofitability. Studies indicate that avoidance learning in predators improves when the prey is conspicuous (Gittleman et al., 1980; Roper and Redston, 1987). There is also evidence that aversion toward aposematic signals has an innate component (Lindström et al., 1999; Schuler and Hesse, 1985).
Unpalatable prey often combine visual signals with other signals into multimodal or multicomponent warning displays (Rowe, 1999, and references therein). This is thought to increase the efficiency of information transfer by acting on several sense modalities in the predator (Rowe and Guilford, 1999a; Rowe and Skelhorn, 2004). Signals such as odors may not be aversive per se, but have been found to induce unlearned aversions when interacting with aposematic or novel colors (Jetz et al., 2001; Rowe and Guilford, 1996) and speed up the process of avoidance learning (Roper and Marples, 1997; Rowe, 2002). However, studies have also shown negative interactions; one signal may "overshadow" the effect of the other (Guilford and Dawkins, 1991; Roper and Marples, 1997).
Sound is found to be a warning signal to mammals (Dunning and Kruger, 1995; Kirchner and Roschard, 1999), but few studies have examined how the sound in a multimodal display affects bird predators. Rowe and Guilford (1999a), however, found that the buzzing of a bumblebee elicited innate biases toward novel and aposematic colors in chicks. Further, sound may also increase the speed of color discrimination learning (Rowe, 2002). Sound is also found to be an essential part of the multimodal signal of cuckoo chicks and is necessary to elicit full response from their foster parents (Kilner et al., 1999).
In the experiments presented here, we used domestic chicks (Gallus gallus domesticus) in an artificial prey–learning task to study the effect of the signals in the display of the wasp. The main objective of the experiment was to test whether there were multimodal effects of sound in the display. We found the wasp to be an ideal model organism for our purposes because of its characteristic buzzing sound and its display of two well-studied aposematic signals, the yellow color and the black stripes (Edmunds, 1974; Kauppinen and Mappes, 2003). The three signals were combined to create artificial prey displaying one, two, three, or none of the signals in the display of a wasp. The chicks were tested at three levels to allow the investigation of the effects of the signals on several aspects of aversion. On the first level, the chicks' responses to experimentally colored, unpalatable prey in the first trial were used to test for effects of innate biases. On the second level, 12 trials were performed over 3 days to study the avoidance learning of the chicks. On the third level, we wanted to test the memorability of the signals. We performed a three-trial extinction-learning experiment on the first day after the avoidance-learning trials, where chicks were offered prey with the same displays as in aversion learning, but the prey were no longer unpalatable.
The following predictions were made. (1) At all levels, the buzzing sound should enhance the effect of the other signals, so that an aversive signal induces a more aversive response when presented in combination with sound. We expected these enhancement effects to be expressed as interactions between signals in the analyses (Marples and Roper, 1996). (2) In the first trial, the chicks were expected due to innate biases to show more avoidance toward prey displaying aposematic signals than toward prey displaying neutral signals. (3) The rate of avoidance learning was expected to increase when unpalatable prey displayed aposematic signals. (4) Associations made to aposematic signals were expected to last longer than associations made to neutral signals.
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Predators
ROSS 208 domestic chicks (96 days old) of both sexes were used as predators. The chicks arrived in batches of 20–25, delivered from a commercial hatchery. The birds of a batch were housed together in a 60 x 102 x 36–cm3 aluminum cage and individually marked in color codes on the down of wings and head. The floor of the cage was covered with sawdust and heated from above by a 250-W infrared bulb. The chicks were given water and brown-colored chick crumbs ad libitum, also during training and testing (see below).
Prey
Pieces of mealworm (Tenebrio molitor), killed in boiling water were used as prey. Taste-manipulated (unpalatable) prey were soaked in a solution made of 4% quinine hydrochloride in distilled water with 2 g mustard powder per 100 ml for 1–2 h (Rowe and Guilford, 1996). Control prey were soaked for the same period of time in distilled water. The prey were kept in a fridge (4°C) until used in the experiment.
Prey coloration was manipulated by presenting each item under a 2.5 x 1.5–cm2 roof-shaped piece of colored paper. Palatable (control) prey were presented under brown pieces of paper. Brown prey were familiar to the birds as they were used to eating brown-colored chick crumbs in the cage. Unpalatable prey were presented under pieces of (aposematic) yellow- or (nonaposematic) green-colored paper with or without stripes. The stripes were 2 mm in width and painted on the paper with a black pen (Staedtler whiteboard marker), one every half centimeter. These colors were (1) novel to the birds and (2) about equally conspicuous, that is, they looked equally conspicuous to us, and there were only slight differences in reflectance (1.4%, see Hauglund, 2003, for details). This enabled us to test for aversions toward aposematic color per se, independent of prey conspicuousness (Mappes and Alatalo, 1997; Sillén-Tullberg, 1985). Both brown and green have been used as nonaposematic or cryptic colors in similar experiments (e.g., Jetz et al., 2001; Rowe and Guilford, 1996, 1999a).
The visual signals were presented to the chicks both in the presence and absence of the buzzing wing sound of a wasp (Dolichovespula media) produced during flight. The sound was recorded using SONY TC-D5M and Sennheiser ME 67 microphone and played back to the chicks at 66–68 dB with a peak of 72 dB (measured by a Brüel & Kjær sound level meter, type 2219) using a SONY TC-D5M cassette player and Philips SBC 8254 active speakers located one on each side of the experimental apparatus. Chicks assigned to sound treatment groups heard the buzzing as long as the experimental prey was visible.
Experimental chamber
The experimental apparatus was made out of a cardboard box with an inner area of 21.5 x 30.0 cm2 and 25.8-cm-high walls, mounted on top of a 5.5 x 26 x 2.2–cm3 gray polyvinyl chloride (PVC) box (visible through a slit through the middle of the floor of the cardboard box) with four chambers for presentation of prey. Using a sliding lid made of the same gray PVC, with a circular hole (diam = 6.00 cm), the prey could be presented one at a time. The floor was covered by a sheet of paper in a shade of gray slightly lighter than the PVC to make a neutral renewable background.
Training
The chicks were given a total of eight trials of training. To get the chicks used to the feeding and housing conditions, there was no training or testing on the day of arrival. Seven training trials were performed on the first day after the arrival. Before the first experimental trial the second day after arrival, we performed one additional training trial to restore their level of learning gained through the training trials the previous day. Training was performed in the same room and with the same apparatus as in the experiment. The room was acoustically isolated from the room housing the chicks to avoid disturbance.
The chicks were trained in pairs during the first four trials to reduce stress (Gamberale-Stille, 2001; Marples and Roper, 1997). Training trials started by introducing the chicks to palatable mealworms scattered on the cardboard floor and in the four chambers of the PVC box with the lid removed. Trials then proceeded with mealworms more concentrated toward the chambers in each trial. From trial 4, the mealworms were exclusively placed in the four chambers and uncovered one by one using the lid. In trials 5–8, chicks were trained separately. Brown-colored pieces of paper were introduced in trial 6; in trials 7 and 8, only 1–2 mm of the mealworm was visible beneath the paper. In the last four trials (5–8), the chicks were hand held, while the chambers were prepared with four new prey items after one trial of four prey. The chicks were then placed once more in the box and the procedure repeated, so that each trial consisted of eight prey items.
Experimental design and procedure
The experiment was carried out during 6 weeks from December 2001 to March 2002. In each of the 6 weeks, 16 chicks that had been eating during the training trials were chosen for the experiment. The chicks were randomly distributed to eight experimental treatment groups of different signal combinations of unpalatable prey so that each treatment group consisted of two chicks per week (Table 1). The order of testing the chicks was decided at the beginning of each day by organizing the chicks in two heats of eight individuals, with one chick of every experimental treatment group represented in each heat. The order of the chicks in each heat and which heat to be tested first were then randomized. The chicks were tested in the same order throughout the day but in a new order each day.
|
The experimental trials were carried out in the same way as the training trials but with half of the presented prey being unpalatable and experimentally colored. The prey were presented in a fixed order in each trial to minimize within-group variation; the first prey item was unpalatable and partly covered by a piece of experimentally colored paper according to the treatment group of the chick, the second was a control (brown), the third was unpalatable and experimentally colored, and so forth (a total of four unpalatable prey and four control prey per trial). There was no prey visible to the chick as it entered the experimental apparatus. When the birds were offered a prey item, they were given 5 s to initiate attack before the next prey was presented, and a "not touched" response was recorded. When the prey was attacked or eaten, we waited for 5 s before presenting the next prey to avoid wrong associations being made to unpalatability. Prey attacked but not eaten was removed and discarded.
When sound was part of the treatment, it was played simultaneously with the uncovering of the prey item. Sound was only presented with experimentally colored unpalatable prey, not when control prey was visible. As the controls were the same as in the training trials, unpalatable prey displayed novel signals, while controls were familiar in appearance.
A total of 12 experimental trials (four each day) were performed during the 3 days of aversion learning. Each chick received the same treatment in every trial. The response scores from the first experimental trial were used to analyze novelty effects. Three extinction trials were performed similarly to the trials in aversion learning on the fourth day of the experiment. Chicks were offered prey with the same display as they were offered in the aversion trials, but the prey was no longer unpalatable. Control prey was still brown and palatable.
Statistical analysis
Due to the discontinuity of the response variable, sources of variation in the proportion of untouched prey were investigated using logistic regression analysis, except in the analysis of the speed of learning. Here we used a fixed-effects ANOVA because the conditions of ANOVA were fulfilled in this data set (as was concluded from the residual plot). All analyses were performed with a significance level of p < .05, and the normality of the data was confirmed by looking at the residual plots. We focused on the "avoidance of unpalatable prey" as the dependent variable (defined as the number of untouched prey in a trial) because this yields the most direct estimate of the relative degree of protection gained from various signal combinations. However, we also analyzed avoidance of control prey to explore the degree to which the various signals contributed to discriminate between unpalatable and palatable prey. To separate the signals in the analyses, we chose to use "warning color" (present or absent), "stripes" (present or absent), and "sound" (present or absent) as predictor variables instead of the treatment groups. However, as the signals in most treatment groups were presented in combinations, we always included all possible interactions between the factors in the analyses. Moreover, the interactions between the signals express the focal enhancement effect. When interactions between signals were detected, the significance of these was further explored by separately analyzing the relevant subsets of data. These separate analyses represent the effect of signals within different treatment groups.
Some batches of chicks ate more than others, probably due to slight differences in the time of hatching (i.e., age/size) or in the composition of sexes (Jones, 1986). This was controlled for in the analyses by including "batch" as a covariate. Although this variable did show significant effects in some analyses, we have not commented it further. The reason for this is that all treatment groups were represented in all batches, so we assumed that any bias due to which batch the chicks belonged to would affect all treatment groups in a similar way.
Four separate but interrelated analyses were carried out. In the first analysis, we explored the relative role of the various signal combinations in eliciting innate or unlearned prey aversion in the birds. Here we used data only from the first trial, that is, the first few encounters between the birds and the novel prey. In the second analysis we used data for the total number of prey avoided by each bird during the 12 trials to explore the relative role of the various signals and signal combinations in reducing predation rates during the experiment. In the third analysis, we explored the relative role of the various signals and signal combinations in increasing the rate of avoidance learning in the birds. This was done using data on the difference in avoidance at the beginning and at the end of the experiment, that is, the difference in mean avoidance on days 1 and 3. To control for the level of learning acquired the first day, mean avoidance on day 1 was included as a covariate in the analysis. In the fourth analysis, data on the number of experimentally colored prey avoided by each bird during the three extinction trials were used to explore the relative role of the various signals and signal combinations in preventing predation when the prey was no longer unpalatable. Here we included avoidance of unpalatable prey (trials 10–12) from the conditioning trials as a covariate to control the fact that the various treatment groups had different start levels after the avoidance-learning trials.
All analyses were performed in S-plus (Insightful, Seattle, Washington, USA).
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Innate biases
The treatment groups showed great variation in whether they attacked the unpalatable prey in the first trial (Figure 1). Among treatment groups offered prey without stripes, mean avoidance was higher when the prey was buzzing than when it was silent (Figure 1); still the effect of sound was not significant (df = 83, b = 0.04, t = 0.24, ns). Neither was the sound involved in any interactions, which rules out the possibility that sound had significant differential effect on the innate response to the visual aposematic signals.
|
Prey color caused most of the variation in mean avoidance (df = 83, b = 0.49, t = 3.03, p < .01). In accordance with the expectations, this implies a great innate bias against yellow color. The effect of stripes was not significant (df = 83, b = 0.28, t = 1.71, ns), but we found a significant effect of the interaction between stripes and color (df = 83, b = –0.37, t = –2.28, p < .05). When we analyzed the responses of the chicks offered yellow and those offered green prey separately, we found that the response to stripes was different in the two groups of chicks. Among chicks offered yellow prey, there were no significant differences in avoidance between those offered yellow-striped prey and those offered plain yellow prey (Figure 1, right half, df = 39, b = –0.09, t = –0.52, ns). Between chicks offered green-striped prey and those offered plain green prey on the other hand, we found that stripes constituted a significant effect (Figure 1, left half, df = 39, b = 0.66, t = 2.38, p < .05). This implies that the yellow color was the most effective signal to induce avoidance on the chicks' first encounter with the aposematic prey; the stripes did not have additional effect on the prey's protection when the prey was yellow. However, when the prey was green, the stripes significantly increased the protection from the chicks (from 0.17 ± 0.17 to 0.75 ± 0.25 prey avoided, that is, almost up to the same level as avoidance of yellow prey, 0.83 ± 0.21; see Figure 1).
Control prey were eaten at the same rate in all groups.
Aversion-learning trials
Overall effects
There was large variation among treatment groups also in mean avoidance accumulated over the 12 trials (Figure 2). Contrary to expectations, the buzzing sound had a significant negative effect on the avoidance (df = 83, b = –0.06, t = –2.06, p < .05). This implies that the silent prey were avoided more on an overall basis than the buzzing prey. However, the results may have been caused by the treatment group offered green, nonstriped buzzing prey, among which the mean avoidance was as low as 0.86 ± 0.18 prey per trial (compared to the results from chicks offered the plain green prey, which showed a mean avoidance of 1.30 ± 0.24 prey per trial, see Figure 2).
|
As expected, both stripes (df = 83, b = 0.25, t = 8.12, p < .01) and yellow color (df = 83, b = 0.21, t = 6.87, p < .01) contributed significantly to higher overall avoidance (Figure 2). This suggests that both of these signals play important roles in reducing predator attacks on wasps. Further, the interaction between warning color and stripes was significant (df = 83, b = –0.17, t = –5.49, p < .01). Isolated analyses showed that yellow prey was not significantly more avoided with stripes than without (Figure 2, df = 39, b = 0.08, t = 1.94, ns), while among groups offered green prey, the stripes significantly increased the mean avoidance (df = 39, b = 0.44, t = 9.56, p < .01).
We also found that the interaction between sound, stripes, and warning color showed significant effects (df = 83, b = –0.20, t = –6.35, p < .01), which implies that the effect of sound depend on the visual signals the prey displayed. A further look at the responses from chicks offered green prey (Figure 2, left half) showed a strong significance of the interaction between sound and stripes (df = 39, b = 0.24, t = 5.24, p < .01) beside the positive effect of stripes (see above). However, the effect of the buzzing sound was insignificant (df = 39, b = –0.05, t = –1.02, ns). Looking at treatment groups (Figure 2), the green, nonaposematic prey obtained the best protection displaying both stripes and sound (2.03 ± 0.23 prey per trial with sound, 1.58 ± 0.26 prey per trial without sound). Among the green prey without stripes, however, the prey with sound was avoided dramatically less (1.30 ± 0.24 prey per trial with sound, 0.86 ± 0.12 prey per trial without sound).
There was a significant effect of the interaction between sound and stripes (df = 39, b = –0.15, t = –3.40, p < .01) also among the chicks offered yellow prey (Figure 2, right half). However, here the variation was much smaller (Figure 2, left half compared to right half), and neither stripes (mentioned above) nor sound (df = 39, b = –0.08, t = –1.86, ns) exerted any significant effect on the avoidance. Worth mentioning, however, is that the yellow-striped prey seemed to obtain the best protection (Figure 2; 2.13 ± 0.24 prey avoided per trial), while the protection was lowered among yellow-striped prey also displaying sound (Figure 2; 1.70 ± 0.28 prey avoided per trial, approximately the same number as in chicks offered plain yellow prey; 1.69 ± 0.23). These effects were however not confirmed statistically.
Groups of chicks offered yellow prey also ate less of the control prey than chicks offered green prey (df = 83, b = 0.32, t = 4.52, p < .01), while there was no significant difference in the amount of control prey eaten between chicks offered striped prey and chicks offered nonstriped prey (df = 83, b = –0.02, t = –0.24, ns). This implies that the yellow prey made the birds less enthusiastic to eat altogether, while the striped unpalatable prey did not affect the attack rate of the palatable prey. This indicates that the chicks better discriminated unpalatable prey from the controls when the unpalatable prey was striped.
Speed of learning
As the overall effects only described the mean sum of avoided prey per trial during the experiment, we did additional analyses where we analyzed the difference between day 1 and day 3 to see to what degree the signals had improved the chicks' avoidance during the trials (Figure 3). The baseline level of avoidance after day 1 had not unexpectedly a large effect on this difference (df = 1, F = 9.84, p < .01). Neither did the buzzing sound contribute to increase the difference (df = 1, F = 0.11, ns) nor was there any effect of warning color (df = 1, F = 0.01, ns). In fact, the only signal that per se contributed to a higher difference between day 1 and 3 was stripes (df = 1, F = 6.62, p = .01). Together with the overall effects, this implies that both the striped and the yellow prey were avoided on a high rate throughout the experiment, but only the stripes increased the learning rate.
|
However, the results were not straightforward considering the difference in avoidance as the three-way interaction between warning color, stripes, and sound showed significant effects (df = 1, F = 6.22, p < .05). Looking at the treatment groups (Figure 3), chicks offered green-striped buzzing prey showed the largest increase in avoidance (from 1.08 ± 0.23 to 2.78 ± 0.36 prey per trial). Isolated analyses of the groups offered green prey showed that the effect of stripes was significant (df = 1, F = 5.95, p < .05), while there was no significant effect of sound (df = 1, F = 0.07, ns). This implies that although the chicks offered green-striped buzzing prey showed the steepest learning curve, only stripes were a significant contributor to the increase. In the subset of birds offered yellow prey, there were no significant effects of either stripes (df = 1, F = 1.16, ns) or sound (df = 1, F < 0.01, ns). This indicates that although the treatment groups offered yellow prey showed different learning rates, neither buzzing nor stripes contributed to this variance. Further, among treatment groups, the chicks offered plain yellow prey showed the least increase in avoidance (from 1.29 ± 0.24 to 1.92 ± 0.28 prey per trial, Figure 3). Also worth noticing is that the group offered green buzzing prey showed the lowest level of avoidance trough the whole experiment (0.50 ± 0.13 prey per trial day 1 and 1.20 ± 0.17 prey per trial day 3), even lower than the group offered green silent prey (0.71 ± 0.14 prey per trial day 1 and 1.77 ± 0.32 prey per trial day 3). However, these effects were not confirmed statistically.
As in the analyses of the overall effects, we found that chicks offered yellow prey avoided significantly more controls on the third day compared to the first than the chicks offered green prey (df = 1, F = 7.70, p = .01). No significant difference was found in avoidance of control prey between chicks offered striped prey and chicks offered nonstriped prey (df = 1, F = 0.63, ns). Looking at the treatment groups (Figure 3), the groups offered, respectively, yellow-striped prey and yellow-striped buzzing prey showed an increase in avoidance of control prey. This indicates that while most chicks learned to discriminate the good-tasting prey from the bad-tasting ones during the 12 trials, some of the chicks offered yellow prey tended to avoid more of the control prey at the end of the experiment. This again implies that the discrimination-learning process was inhibited.
Extinction learning
On the fourth day of the experiment, the experimentally colored prey was switched from bad tasting to palatable. Sound significantly contributed to decrease the overall avoidance during these three trials (Figure 4, df = 82, b = –0.46, t = –4.60, p < .01). This implies that buzzing prey was accepted as palatable more easily than silent prey. Stripes on the other hand significantly increased the avoidance (Figure 4, df = 82, b = 0.22, t = 2.02, p < .05), while warning color had no significant effect (df = 82, b = –0.17, t = –1.65, ns). The covariate avoidance of unpalatable prey from the learning experiment had, as expected, significant effects on the avoidance of the previously unpalatable prey in extinction learning (df = 82, b = 0.47, t = 10.66, p < .01).
|
Neither were there significant effects of interactions on the avoidance of previously unpalatable prey nor was there any significant effect on the avoidance of controls.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Generally, this study once more confirmed the aposematic strength of yellow color and black stripes. However, we found that yellow was most effective as a signal to trigger innate biases, whereas it was not especially effective as a discriminative stimulus. Stripes were on the other hand less effective on the first encounter but the best contributor to aversion learning. It also seemed to be the most memorable signal. Sound did not fulfill the expectation of enhancing the other aversive signals; it actually lowered the aversion in some instances. As expected from the theory of multimodal warning signals, stripes, color, and sound did enhance the effects of each other in several cases, though there was not enough consistency to draw clear conclusions.
Innate biases
Rowe and Guilford (1999a) showed that an acoustical cue might elicit hidden unlearned biases toward yellow- and novel-colored food without being aversive per se. Based on these results, we expected the sound to enhance the aversive effects of the other signals. However, we did not find the sound to interact significantly with the main aversive signal, that is, warning color or stripes in the first trial. The reason for this lack of effect may be found in the qualities in the sound used; in our study, we used an ordinary flying sound. Possibly, an aggressive sound is a more effective signal in triggering biases. Rowe and Guilford (1999a) used the high-pitched buzzing of bumblebee (Bombus terrestris), recorded as the bumblebee was caught in a net, but suggested that any acoustical signal might have a similar effect. The present study does not support this hypothesis.
Black and yellow stripes have earlier been found to have an innate aversive effect (Schuler and Hesse, 1985). Naive chicks were found to eat more than 10 times less of the yellow and black than the green prey even though both were palatable. We therefore expected to find aversions toward these signals among the naive chicks. As the design of our study allowed us to separate the effect of the signals in the analyses, we were able to show that the yellow color was the most important signal to increase the avoidance of unpalatable prey in the first trial. As both yellow and green were novel signals to the inexperienced chicks, this effect is probably not caused by novelty or neophobia (Marples and Kelly, 1999; Rowe and Guilford, 1999b); the unlearned aversion toward yellow color is most likely caused by an innate bias in the predators against this particular color. This follows the expectations as yellow is a well-known aposematic color (Cott, 1940; Poulton, 1890). The stripes did not show significant effects per se. However, the separate analyses showed that the stripes did increase avoidance in inexperienced chicks but only when the prey was green. A plausible interpretation of this result is that chicks had innate biases also toward stripes, but the aversive effect was overshadowed by the yellow color (Guilford and Dawkins, 1991; Roper and Marples, 1997). This implies that the stripes of yellow wasps have no additional aversive effect on inexperienced predators.
Thus, we did not find evidence for enhancement effects of multimodal displays on inexperienced predators.
Learning effects
A tone not previously known to have any aposematic effect may increase the speed of the discrimination learning of visual signals (Rowe, 2002). We therefore expected that in the groups offered sound as part of the display, the aversion learning would proceed faster, and the buzzing prey would be avoided more through the experiment. However, the only effect of sound found in analyses of learning was that buzzing prey was attacked more on an overall basis than silent prey. This is the opposite of the results found by Rowe (2002). Instead of contributing by being a conditional stimulus, the sound seemed to repress the conditioning.
Why did we get this effect? Studies on rats have shown that stressful noise may function as an activating stimulus, inducing higher speed of food intake (Krebs et al., 1997). Although the sound in the present study was presented at maximum 72 dB, which was not found to affect the chicks per se, it may have aroused similar effect on the chicks. A more speculative hypothesis is that the chicks felt more safe or comfortable with the sound present, a hypothesis with associations to social facilitation (Keeling and Hurnik, 1996). However, more research is necessary to answer these questions.
Although we did not find evidence for enhancement effects of multimodal displays on inexperienced predators, we did see one specific effect in support of the multimodal theory in the learning trials, though not statistically significant. The sound seemed to increase the effect of stripes both in the overall analyses and the speed of learning, but only in green prey. The lack of similar effects in yellow prey may again be explained by the other signals being overshadowed by yellow (Guilford and Dawkins, 1991; Roper and Marples, 1997). However, another explanation may be that birds offered green prey tried out more prey during the first trials because they were not affected by innate aversions and could therefore learn to associate stripes and sound to unpalatability. This is a similar reasoning as made by Roper and Redston (1987), who discussed whether the effect of a conspicuous signal is direct, due to features of the signals, or indirect, due to a higher initial encounter rate. An unpalatable prey that is more easily discovered will give the predator a higher rate of negative reinforcement, given the assumption that it is attacked more often (Roper and Redston, 1987). In the present study, the chicks encountered the yellow and green prey at the same rate, but chicks may have experienced less negative reinforcement from yellow prey than from green prey due to high innate avoidance of yellow prey.
Unexpectedly, the colors of the unpalatable prey also affected the consumption rate of the palatable control prey. In both the overall analyses and the speed of learning analyses, chicks offered yellow unpalatable prey also avoided significantly more of the control prey than those offered green unpalatable prey. In contrast, chicks offered striped prey ate control prey at the same rate as chicks offered nonstriped prey. This suggests that when the chick had been exposed to a yellow-colored prey, attacks on all prey were inhibited, also those not displaying the signals. That is, the fear of yellow became generalized to the brown prey. This also supports the theory that chicks offered yellow-colored prey will be more reluctant to try out prey because of the innate fear, which again makes the learning curve less steep. The fact that there were no differences in the amount of control prey eaten between chicks offered striped prey and nonstriped prey may further explain why stripes increased the avoidance learning; being easy to discriminate from the palatable prey is important in a signal that is aimed to improve learning (Guilford, 1990; Turner, 1975).
Extinction learning
Previous studies have shown that a conditioned conspicuous signal may last longer in a chick than a cryptic signal (Roper and Redston, 1987). This was assumed to be because conspicuous prey was more memorable than cryptic prey. In the present study, we did not characterize the signals according to whether they were conspicuous or not; we expected all aposematic signals to contribute to slower extinction. The result showed that like the conspicuous red signal in the study of Roper and Redston (1987), the stripes were the only signal slowing down extinction and thus probably the most memorable one.
As opposed to stripes, sound increased the speed of extinction learning. This could imply that chicks understood faster that the prey were now palatable when they were buzzing, which is in accordance with the hypothesis that sound could facilitate the conditioning by being an additional conditioned stimulus (Rowe, 2002). In the present study, it seemed that the stripes made a prolonged association especially to unpalatability, while sound facilitated associations generally. However, as a negative effect of sound was found in the overall analysis, what appears to be an increasing effect of sound on extinction may rather show that the sound simply facilitate eating (as discussed under overall effects).
|
CONCLUSIONS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
In conclusion, we failed to find direct support for multimodal effects of the signals or the aversive effect of sound. However, the expectations to the effects of the other two signals displayed by wasps, the stripes and the yellow color, are supported. But instead of being generally aversive or enhancing each other, we found the signals to work in different ways. The yellow color clearly increased the aversion of naive predators toward novel prey, and stripes increased the speed of avoidance learning. As the sacrifice of prey to teach predators may be avoided with innate biases (Edmunds, 1974
), our results suggest that the wasp may gain initial benefits by being yellow; then the stripes may increase the survival of a prey by being a signal that is easily associated with unpalatability because it is memorable (Roper and Redston, 1987), discriminative (Turner, 1975), and probably easy to recognize (Guilford, 1986).
The concepts of multimodal and multicomponent signals have been presented as synonyms in research literature (Rowe, 1999). However, in our study the color and the color pattern had different effects; the yellow color was effective in eliciting innate biases and the black stripes were effective in educating the predators, although they were both visual signals. Thus, this study suggests that multicomponent signals may consist of different components with different effects, not necessarily reaching out to different sense modalities.
|
ACKNOWLEDGEMENTS |
|---|
We thank Øyvind Sjøvik at FFI (Norwegian Defense Research Establishment) for reflection analyses and Tore Slagsvold, Göran Hôgsteth, Michelle Harper, and Nina Holmengen for comments on previous versions of the manuscript. We also thank the staff at the animal housing facilities for taking care of the chicks.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Cott HB, 1940. Adaptive coloration in animals. London: Menthuen.
Dunning DC, Kruger M, 1995. Aposematic sounds in African moths. Biotropica 27:227–231.[CrossRef]
Edmunds M, 1974. Defence in animals. London: Longman.
Gamberale-Stille G, 2001. Benefit by contrast: an experiment with live aposematic prey. Behav Ecol 12:768–772.
Gittleman JL, Harvey PH, Greenwood PJ, 1980. The evolution of conspicuous coloration—some experiments in bad taste. Anim Behav 28:897–899.[CrossRef]
Guilford T, 1986. How do warning colors work—conspicuousness may reduce recognition errors in experienced predators. Anim Behav 34:286–288.
Guilford T, 1990. The evolution of aposematism. In: Insect defenses: adaptive mechanisms and strategies of predators and prey (Evans DL, Schmidt JO, eds). Albany: State University of New York Press; 23–61.
Guilford T, Dawkins MS, 1991. Receiver psychology and the evolution of animal signals. Anim Behav 42:1–14.
Hauglund K, 2003. Responses of domestic chicks (Gallus gallus domesticus) to multimodal aposematic signals (Masters thesis). Oslo: University of Oslo.
Jetz W, Rowe C, Guilford T, 2001. Non-warning odors trigger innate color aversions—as long as they are novel. Behav Ecol 12:134–139.
Jones RB, 1986. Responses of domestic chicks to novel food as a function of sex, strain and previous experience. Behav Process 12:261–271.[CrossRef]
Kauppinen J, Mappes J, 2003. Why are wasps so intimidating: field experiments on hunting dragonflies (Odonata: Aeshna grandis). Anim Behav 66:505–511.[CrossRef]
Keeling LJ, Hurnik JF, 1996. Social facilitation acts more on the appetitive than the consummatory phase of feeding behaviour in domestic fowl. Anim Behav 52:11–15.[CrossRef]
Kilner RM, Noble DG, Davies NB, 1999. Signals of need in parent-offspring communication and their exploitation by the common cuckoo. Nature 397:667–672.[CrossRef]
Kirchner WH, Roschard J, 1999. Hissing in bumblebees: an interspecific defence signal. Insectes Soc 46:239–243.[CrossRef]
Krebs H, Weyers P, Macht M, Weijers HG, Janke W, 1997. Scanning behavior of rats during eating under stressful noise. Physiol Behav 62:151–154.[Medline]
Lindström L, Alatalo RV, Mappes J, 1999. Reactions of hand-reared and wild-caught predators toward warningly colored, gregarious, and conspicuous prey. Behav Ecol 10:317–322.
Mappes J, Alatalo RV, 1997. Effects of novelty and gregariousness in survival of aposematic prey. Behav Ecol 8:174–177.
Marples NM, Kelly DJ, 1999. Neophobia and dietary conservatism: two distinct processes? Evol Ecol 13:641–653.[CrossRef]
Marples NM, Roper TJ, 1996. Effects of novel colour and smell on the response of naive chicks towards food and water. Anim Behav 51:1417–1424.[CrossRef]
Marples NM, Roper TJ, 1997. Response of domestic chicks to methyl anthranilate odour. Anim Behav 53:1263–1270.[Medline]
Poulton EB, 1890. The colours of animals their meaning and use: especially considered in the case of insects. London: Kegan Paul, Trench, Trübner & Co.
Roper TJ, 1990. Responses of domestic chicks to artificially colored insect prey—effects of previous experience and background color. Anim Behav 39:466–473.[CrossRef]
Roper TJ, Marples NM, 1997. Odour and colour as cues for taste-avoidance learning in domestic chicks. Anim Behav 53:1241–1250.[CrossRef][ISI][Medline]
Roper TJ, Redston S, 1987. Conspicuousness of distasteful prey affects the strength and durability of one-trial avoidance-learning. Anim Behav 35:739–747.[CrossRef]
Rowe C, 1999. Receiver psychology and the evolution of multicomponent signals. Anim Behav 58:921–931.[CrossRef][ISI][Medline]
Rowe C, 2002. Sound improves visual discrimination learning in avian predators. Proc R Soc Lond B 269:1353–1357.[CrossRef]
Rowe C, Guilford T, 1996. Hidden colour aversions in domestic clicks triggered by pyrazine odours of insect warning displays. Nature 383:520–522.[CrossRef]
Rowe C, Guilford T, 1999a. The evolution of multimodal warning displays. Evol Ecol 13:655–671.[CrossRef]
Rowe C, Guilford T, 1999b. Novelty effects in a multimodal warning signal. Anim Behav 57:341–346.[CrossRef][ISI][Medline]
Rowe C, Skelhorn J, 2004. Avian psychology and communication. Proc R Soc Lond B 271:1435–1442.[CrossRef][Medline]
Schuler W, Hesse E, 1985. On the function of warning coloration: a black and yellow pattern inhibits prey-attack by naive domestic chicks. Behav Ecol Sociobiol 16:249–255.[CrossRef]
Sillén-Tullberg B, 1985. The significance of coloration per se, independent of background, for predator avoidance of aposematic prey. Anim Behav 33:1382–1384.[CrossRef]
Turner JRG, 1975. Tale of 2 Butterflies. Nat Hist 84:28–37.
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||