Abstract Cropping system effects on population levels of the cassava hornworm, Erinnyis ello L., and cassava stemborer, Chilomima clarkei (Amsel) were studied in three growing seasons between 1983 and 1985 in the Department of Tolima, Colombia. In the first trial, herbivore density was compared between cassava/maize intercrops and monocultures grown at two different densities. In subsequent experiments, cassava monoculture and intercrops with maize and cowpea were grown at one cassava density. In the first trial, hornworm and stemborer numbers per square metre were higher in denser plantings of cassava. However, pest density per plant was greater in sparse stands. Effects of intercropping on hornworm incidence varied with cassava and maize varieties with reduced levels in some combinations but not others. Stemborers were lower in intercrops than in monoculture; however, this cropping system advantage disappeared after intercrop harvest. Cassava varieties differed in levels of both herbivores but varietal mixtures had no overall beneficial effect vis‐à‐vis pest levels in pure stands. Nevertheless, distinct herbivore preferences for different varieties suggest the possibility of trap cropping in regions where these pests cause significant yield losses. Keywords: Cassavaintercroppingvarietal mixturesErinnyis elloChilomima clarkei