Absence of adaptive learning from the oviposition foraging behaviour of a checkerspot butterfly

Abstract

A population of Edith's checkerspot butterfly, Euphydryas editha, laid eggs on two morphologically distinct host-plant species, one of which it had recently colonized. Search behaviour of two groups of females was compared in the field. ‘Naive’ females were experimentally deprived of both flight experience and host encounter, while ‘experienced’ females were observed without intervention. Three analyses failed to show evidence for adaptive learning that might have allowed the butterflies to increase their efficiency of host finding with experience: (1) proportions of alights on hosts were not different between the naive and experienced groups, (2) rates of alighting on hosts did not differ between the naive and experienced groups, and (3) probabilities of alighting on hosts did not increase during searching bouts. Both groups searched efficiently for their traditional host species, Pedicularis semibarbata(dwarf lousewort), and inefficiently for the novel host, Collinsia torreyi(blue-eyed Mary). There was also no evidence for behavioural differences among individuals of either group. The distribution among individuals of alight frequencies on C. torreyi appeared Poisson, with a single mode near the frequency expected from random alighting. This finding is consistent with the hypothesis that no females recognized C. torreyi as a host in flight, and that differences between butterflies in observed alight frequencies stemmed simply from random variation in the composition of the vegetation over which they flew. Thus, the absence of adaptive learning impeded the evolutionary incorporation of the novel host into the diet. These results complement previous findings that E. editha from the same population failed to learn to accept particular hosts after alighting and are in contrast to other studies of insect foraging, all of which have shown that learning is an important component of foraging behaviour, causing search efficiency to improve with experience. It is suggested that E. editha could be a model ‘non-learning’ insect in comparative studies that empirically test the role of learning in resource choice.