In Eq. 1, y is the vector of individual-tree observations, β
is the vector of fixed effects associated to y by the incidence
matrix X, and such that the matrix X is of full column rank.
For each trial, the fixed effects were: overall means,
replicates, and genetic group to account for the means of
the different provenances of parent trees. Fifteen genetic
groups were formed according to provenance (see above,
Table 1). The random vector a contained the additive
genetic effects (or breeding values) of individual trees and
is related to y by the incidence matrix Z. The expectation of
a is 0, and the covariance matrix is A σA
2 where A is the
additive tree-level relationship matrix (Henderson 1984) for
the trial trees and their known ancestors, and σA
2 is the
additive genetic variance. The A matrix was modified to
take into account 25% average selfing in E. viminalis (see
below genetic parameters). There were only a maximum of
80 bulk trees (assumed to be unrelated) for the variable
HT2 in the MAYO trial, so no extra independent additive
variance was assumed for these individuals and were
included in A with their parents unknown