The significance of transgenic cotton in the pest control strategy adopted by the

Australian industry makes the management of resistance to the Cry1Ac toxin of Bacillus

thuringiensis essential. Accordingly, the CRDC supported the selection of resistance in

H. armigera so that the extent and nature of the threat could be estimated before it

actually occurred in the field.

During selection of the original H. armigera strain (BX) in which Bt resistance was first

detected, resistance rose to 300-fold and then declined to stabilise at approximately 80-

fold. To assess whether the decline in resistance was due to a loss in general vigour

associated with inbreeding or rather was the result of fitness costs associated with the

resistance, the BX strain was outcrossed to a susceptible laboratory strain and re-selected as the IS strain. With increasing selection pressure, the IS strain reached a resistance

ratio exceeding 800-fold. It was clear from this experiment that the decline in resistance

of the BX strain was the result of inbreeding. However, it was not clear whether the very

high resistance detected in the IS strain was merely the result of improved vigour. As

higher levels of resistance in Cry1Ac-resistant diamondback moth can be associated with

more than one resistance gene (Ferr6 and Van Rie, 2002, Ann. Rev. Entomol . 47, 501-

503), the higher resistance in H. armigera might similarly be indicative of the presence

of a secondary resistance gene (or genes). It is important that we understand all the

resistance options available to H. armigera because we cannot be certain that the order in

which resistance genes arise will be the same in the field as in the laboratory.

Fitness cost associated with resistance is an important factor in determining the

parameters required for an effective refuge strategy. When the fitness of the BX strain

was assessed against a susceptible laboratory strain, we recorded a significant delay in

development that might impede the efficiency of the resistance strategy. However, it

would not be reasonable to rely too heavily on these data because of the different genetic

backgrounds of the resistant and susceptible lines used in that experiment. We

backcrossed the resistance allele into the susceptible line so that we now have resistant

(ISOC) and susceptible lines that share a common genetic background, to the extent that

they are >93% genetically similar. We are, therefore, now in a much better position to

assess the fitness cost associated with the lower level of resistance to Cry 1Ac in

.