Increased residual variance at developmental switch points: Statistical artifact or indicator of genotypic influence?