As discussed in the preceding section, the SCA effect for the cross between parents A and B is common between the two reciprocal ways of making the cross SCA(AxB) = SCA(BxA). If we have reciprocal crosses in the data, we can model the reciprocal SCA effect, which is due to differences between AxB and BxA. Furthermore, some part of the reciprocal effect may be due to differences in the GCA effect of a parent when used as male and female. We can augment the diallel model to include reciprocal effects of both GCA and SCA. To do this, we need to add a column to the data set for an indicator variable that distinguishes reciprocal crosses. This indicator variable can be a covariate set to 1 for crosses in one direction and -1 for crosses in the other direction, and it is arbitrary as to which cross direction is coded as 1.