ABSTRACT
The first 6-vector will be a composite of all attributes, and is naturally complemented by the residuals from regression for each of the individual attributes upon it. The linear combinations of the original attributes for the princo and deviations from regressions were then used to obtain a re-attribution of all genotypes 1-43. The relative scores on the new attributes hardly differed from those obtained from the analyses described in the previous paragraph. For attributes inherited dispersedly, the differences between genotypes will tend to lie more or less near a single line in attribute space. Orthogonality is sought and the close relations to original attributes are only possible for a few of the six new coordinates. A further approach would be to regress ‘days to flowering’ out of each attribute in turn, again getting ‘one plus six’.
