ABSTRACT
The eight environments form a crude (4 within 2)×2 factorial, in which years are crossed with two locations in each version of coastal-inland. In the Scatter plot matrices, the relationships among the attributes are very clear for the overall mean, but much more diverse for the other contrasts. The difference between the early and late maturity classes is most apparent for the within-coastal contrast and the interaction of the effect with the year effect. Another way to look at re-environmenting is to treat C1 as a princo and regress each of the original environments on this for each attribute. For an individual genotype, little information except for doubly princoed results, princoed both on attributes and environments.
