Genotype x year (G x Y) and genotype x crop (G x C) interactions may be confounded in sugar cane when data is obtained from the plant crop and/or from ratoon crops in successive years. A technique to minimise the effects of confounding of the G x C interaction with the G x Y interaction is presented. The mean yield of cane from neighbouring farms was used as an indicator of the year (environment) effect. It was used to re-analyse yields from three experiments comparing six genotypes grown over a plant crop and three ratoons. Two experiments were grown under rainfed conditions and one experiment was irrigated. The confounded interactions were partitioned into a linear contrast on the farm yields (a year effect) and a crop effect that was the remainder. When the yields adjusted for the farm yield effects were compared with original yields of cane, yields of plant and first ratoon were reduced and those of third ratoon were increased. There were only minor changes in the ranking of genotypes on yield. It was concluded that the re-analysis using farm yield showed that the confounding effect of years on the interpretation of genotype x crop effects was samll.