Bivariate Modelling of the Long-Short Axis System for the Study of
Eccentric Growth and Its Effect on Wood Specific Gravity Variation
Abstract
The specific gravity of wood (WSG), a variable property, has a strong
influence on the quality of the wood. Eccentric growth contributes to
the variation of WSG within a tree but is poorly understood. To clarify
the effects of eccentric growth on WSG variation, two opposite sides of
the same tree, representing a differential growth rate, were used for
the bivariate modelling of the long-short-axis system. The WSG values on
the long and short axes were considered two distinct traits, and their
common biological architecture was examined. By jointly modelling the
long-short axis, correlations between aspect (spatial correlation) and
contemporaneous correlation (within aspect) were examined. The variances
of WSG in the long-short axis were different and non-zero. The spatial
correlation was -0.03 in the outer wood modules and 0.52 in the core
wood modules, averaging 0.25 for the entire WSG profile between the
long-short axis. The contemporaneous correlation ranged from 0.79 to
0.86. The results suggest a spatial correlation between the long-short
axis and this spatial correlation change with the developmental stage of
the tree. The bivariate mixed model method provides the underlying
physiological, genetic, and environmental mechanisms related to the
long-short axis to explain, clarify, and predict the behaviour of
complex wood formation associated with eccentric growth. When young,
wood formed on opposite sides may be controlled by the same gene set,
but in the outer wood shell, the cambium can be controlled by different
and unrelated gene sets that drive WSG variation between aspects. It may
be that cambium evolves to meet the needs of a changing crown
architecture or local environment over time.