Ovi position preference and larval performance in Polygonia c‐album (Lepidoptera: Nymphalidae): the choice between bad and worse

As Thompson ( 1988a) remarks, the relationship between adult oviposition preference and offspring performance is the crux of the problem of understanding the evolution of insect/plant associations. Females often display a hierarchy of preferences for different hosts, so that they use a host lower in the hierarchy when the preferred species is not available (e.g. Wiklund, 1981; Thompson, 1988b). The preference hierarchy may be more or less well correlated with larval performance; here used as a composite term for offspring growth rate, survival and reproduction (Thompson, 1988a). A good correlation suggests that plant characteristics, including plant chemistry and nutritional value, is the most important factor that influences larval performance, limits host plant range and promotes specialization on only a few host plants the general pattern in insects. This is the core of both coevolution theory and sequential evolution theory for the evolution of insect-host plant interactions (Ehrlich & Raven, 1964; Jermy, 1984; Ronquist & Nylin, 1990). Conversely, poor correlations can be used as evidence by those who consider other ecological factors, such as selection for enemy-free space, to be of great importance in the evolution of host plant choice and specialization (e.g. Atsatt, 1981; Bernays & Graham, 1988; see also Strong, 1988). Thompson (1988a) and Thompson & Pellmyr (1991) review some of the evidence on preference-performance relationships, and note that the correlations range all the way from good to poor. Surprisingly often, poor correlations are found (e.g. Chew, 1977; Courtney, 1981; Williams, 1983; Penz & Araujo, 1991; Valladares & Lawton, 1991). There is evidence that a number of factors may cause poor correlations between preference and performance (Thompson, 1988a; Thompson & Pellmyr, 1991). These include: (1) selection for enemy-free space;