Can Predation Lead to Narrow Food Specialization in Phytophagous Insects?

Abstract

I can fully agree with Bernays and Graham that: (1) trade-off estimations have not proved advantage of specialists over generalists so far, (2) secondary plant compounds provide proximate but not ultimate cause for narrow host (3) evolutionary arms race analogy cannot fruitfully be applied to insectplant relationships, (4) deterrents do not betoken toxicity of plants, (5) secondary plant compounds cannot be regarded, first and foremost, as defenses against insects, (6) local host abundance is more a prerequisite for specialization than cause of it, and (7) views of specialization as enhancing mate location, reducing mortality from abiotic factors, or providing efficient attachment on plant may be valid only for special cases. However, authors' final conclusion, that predation might be the most important factor pushing insect herbivores toward narrow host range, does not seem to hold. One of their main premises is that host use is labile and, consequently, that current selective pressures may explain evolution of specialized habits. As proof they offer (a) restrictions on host-plant ranges of local populations in polyand oligophagous species, (b) recruitment of insect species to introduced plants, and (c) rapid adaptation of pest species to resistant cultivars. In none of three cases, however, do changes represent basic changes in host range. In cases (a) and (c), presence of a dominant or novel host plant simply selects for a certain genotype that happens to be preadapted to it, or in case (a) observed restriction may represent an experience-induced, and thus temporary, specialization (induced preference) that can occur only within innate host range (Jermy 1987). In case (b) asymptotic nature of recruitment curve itself (Strong et al. 1984) proves that adaptation to novel host is usually not a question of evolutionary time: only insect species recruited are those that harbored already preadapted genotypes. As soon as these are exhausted, recruitment ceases. Thus lability of host use is usually confined within fairly narrow host-plant limits. A further question also arises: how frequently do changes of host range occur in nature? In this connection it must be pointed out that, for example, members of several families of Lepidoptera feed exclusively on species of a single plant family (Ehrlich and Raven 1964); families of scale insects Kermesidae and Dactylopidae live exclusively on Fagaceae and Opuntia spp., respectively (Eastop 1978); a group of species of chrysomelid genus Crioceris, occurring mostly sympatrically in Palearctic region, are restricted to Asparagus spp. (Kaszab 1962); adults of as many as five hispine beetle species may occupy one scroll made of leaves of Heliconia spp. (Strong 1984); etc. Many similar examples can be given. They indicate that in all these insect taxa host range did not change much during process of speciation, which might have taken many millions of years. There is no reason to suppose that in meanwhile biotic and abiotic ecological factors did not change or that selective factors were and are same in whole area of distribution. Thus, conservatism of host range may be as common as lability. Naturally, degree of conservatism is variable and is obviously a fundamental genetic character of each group of insects. The idea propounded by authors that predation may play most important role in evolution host range is challenging, but one has to agree with authors' opinion that experimental studies are needed to prove its validity. The presently available, mostly circumstantial evidence, well reviewed by authors, is meagre. On other hand, it is almost common knowledge that most generalist predators do not hunt by plant species, which would be a prerequisite of leaving certain plant species as enemy-free space. For example, insectivorous birds hunt in forest by structure of trees, but not by their species (Krebs and Davis 1978). Carabid beetles select specific habitats independently of plant species therein (Thiele 1977). In searching for aphids, egg-laying female syrphid flies prefer certain strata of vegetation but do not discriminate among plant species (Bastian 1986). Spiders spin their webs between structures, ants search everywhere, etc. Similarly, specific parasitoids are known to find their hosts by means of kairomones (Birch 1974) as much as by plant-specific cues. The