Several biochemical studies have suggested that in Drosophila melanogaster the two common allozymes of G6PD differ in their in vitro activities and thermal stabilities. Yet, it remains to be shown that these characterizations reflect actual in vivo differences and are not artifacts of the biochemical approach. In this study it is shown that in vivo activity differences must exist between these two variants. This conclusion arises from the observation that the viability of flies bearing a low activity allele of 6PGD is strongly dependent on the genotype at the G6PD (Zw) locus, whereas no measurable difference in viability can be detected between Zw genotypes in a normal activity 6PGD background. These viability interactions are in the direction predicted by the reported in vitro activities of the allozymes and the proposed deleterious effects of 6-phosphogluconate accumulation.--In addition, a genetic scheme is used that uncouples and quantifies the effects of viability modifiers in the region of the Zw locus, while homogenizing 98% of the X chromosome. The viability of different Zw genotypes is measured by examining whole chromosome viabilities relative to the FM6 balancer chromosome. The advantages of this particular scheme are discussed.