Host plant resistance has long been a focus of the CSIRO cotton breeding program. Over the past 8 years we have compared the levels of pest resistance in a variety of cotton cul ti vars which bear either morphological (okra leaf, frego bract), or biochemical factors (high gossypol) for resistance to insects. Some of this work was reported at the last cotton conference (Pitt et al 1992) and elsewhere (Thomson 1987, Wilson and Fitt 1987). As a result of that work, which focussed on resistance to Helicoverpa and mites, several promising lines have been introduced into the CSIRO breeding program. Conventional breeding for pest resistance makes small incremental improvements in the tolerance of varieties to insect feeding. With the advent of genetically engineered cotton the stage is set for quantum leaps in pest resistance, through for example the introduction of the Bt toxin genes. However, this does not mean conventional approaches are no longer useful. Any change which can be made to the plant to make it less atttracti.ve to pests or more tolerant to damage will only enhance the value of genetically engineered traits by providing a stronger, more stable basis on which to manage those genes