“…Such trade-offs should be linked to negative genetic correlations of larval fitness traits across hosts, favouring the evolution of hostspecialised genotypes that could theoretically lead to speciation (see Via, 1990;Jaenike, 1990). However, negative across-host genetic correlations of larval performance traits have rarely been found (e.g., recently Ueno et al, 1997;Keese, 1998;Lazarevic et al, 1998;discussed in, among others, Rausher, 1988discussed in, among others, Rausher, , 1992Jaenike, 1990;Via, 1990;Thompson, 1994;Joshi & Thompson, 1995;Fry, 1996). Recently, it has been proposed that negative genetic correlations may not be necessary for host specialisation to evolve, but that a significant statistical interaction of larval performance traits between insect genotypes and host plant could suffice (Fry, 1996).…”