The activity of Bacillus thuringiensis Berliner against larvae of the gypsy moth, Lymantria dispar (L.), was measured initially on field-collected foliage of 17 host trees and a laboratory host (leaf lettuce) dipped in suspensions of B. thuringiensis subsp. kurstaki. Large differences in larval mortality among hosts were found; the amount of B. thuringiensis required to obtain comparable levels of mortality varied >10-foId among some hosts. White oak, Quercus alba L., and sweetgum, Liquidambar styraciflua L., were selected for further study. Mortality of larvae on foliage treated with B. thuringiensis suspensions of 2 concentrations and held in small cages was higher on white oak than on sweetgum. The amount of B. thuringiensis deposited on white oak and sweetgum foliage, and the amount remaining after 3 d, was measured by extracting, culturing, and counting colonies produced by viable spores. There was a nonsignificant trend toward higher initial deposition of B. thuringiensis on oak than on sweetgum, but only at the high rate of B. thuringiensis. Otherwise, deposition of B. thuringiensis and survival of spores over 3 d did not differ between white oak and sweetgum, and was not closely related to differences in larval mortality. These results indicate that other factors, possibly secondary plant compounds or environmental factors, are involved.