Some Problems in the Study of Anuran Pairing Patterns

Abstract

come to resemble one another. In some instances they do so closely, in others sufficiently only to recall an unpleasant experience by means of some striking feature which they possess in common. From treatments like that of Ford we feel that the message delivered is that a predator rapidly learns to avoid distasteful models, in principle after a single trial, and thereafter hardly ever touches a similar looking insect again. We therefore interpret Ford's assertion, that mullerian mimics all of which are protected, to mean that they are not attacked, i.e. gain full protection when predators have once come into contact with a mullerian mimic. Thus we have never claimed that the probability of a new instar of L. equestris being attacked when presented during trial 5 is equal to that of it being attacked when presented for the first time, and the point Paxton and Harvey make about the sample size used, making it impossible to detect a 25% decrease in attack frequency from trial 1 to 5, is of course correct. However, we do not agree with their claim that trial 4 does not make a suitable comparison with trial 5. On the contrary, we believe that the important comparison to be made is that between the two trials which mark the point where a switch in prey type occurs. Since our experiment was designed to give the birds' experience of the first prey type, and thus a possibility for avoidance learning during four consecutive trials, the relevant comparison to be made is that between trials 4 and 5. Finally, the null hypothesis relevant to our experiment is that the birds do not detect a difference between the two prey types. This null hypothesis could not have been rejected if the attack frequency had decreased or remained the same between trials 4 and 5. However, the statistically significant increase in attack frequency that we found in our experiment forces us to reject the null hypothesis and to conclude that the birds do perceive a difference between the two instars. In conclusion, judging from our data on attack frequencies, there is no evidence to indicate that the birds used in our experiments did generalize between adults and fifth instar larvae of L. equestris.