When theory predicts which phenotypes are well adapted to a given environme
nt, the data do not always match the predictions. Host-plant selection by h
erbivorous insects is one such example. Herbivorous insects often appear to
make poor choices about where their offspring should develop. New evidence
presented by Scheirs et al. suggests that adult insects can choose oviposi
tion sites that enhance their own long-term fitness at the expense of their
individual offspring. This suggests that herbivorous insects might be genu
inely bad mothers, that host choice is nonetheless adaptive, and that theor
y needs to incorporate new assumptions about host effects on adult performa
nce.