Physiological host specificity of microsporidia as an indicator of ecological host specificity.

For most groups of biological control agents the relationship between laboratory (physiological) host range and the host range in the field (ecological host range) has not been explored empirically. The objective of our study was to investigate this relationship using the North America gypsy moth, Lymantria dispar, as a model nontarget host for microsporidia from native North American Lepidoptera. The gypsy moth, L. dispar, a native of Europe, has been established in North America for nearly 130 years and presumably exposed to many species of microsporidia from sympatric native Lepidoptera. Nevertheless, microsporidia have never been observed in North American populations of L. dispar. We conducted traditional laboratory feeding experiments using microsporidia from 20 lepidopteran host species and 1 coleopteran host species against L. dispar. Microsporidia from 18 native hosts infected L. dispar larvae. Although some of the infections were not typical of infections in the indigenous natural hosts, mature spores were produced in most of these infections. Horizontal transmission experiments, based on exposure of uninfected L. dispar larvae to infected L. dispar larvae, demonstrated that the microsporidia were far more host specific than the direct feeding experiments suggested. Of the three microsporidian biotypes that were horizontally transmitted between the nontarget L. dispar larvae, all were transmitted at very low levels. The results of our experiments provide additional evidence that the ecological host specificity of terrestrial microsporidia is much narrower than the physiological host specificity. Our studies establish the validity of using nonindigenous insect species with long-term data sets on natural enemies associated with them as a tool for testing hypotheses about host specificity. Copyright 1998 Academic Press.