Genetic and phenotypic parameters for core wood density of Pinus taeda L. were estimated for ages ranging from 5 to 25 years at two sites in southern United States. Heritability estimates on an individual-tree basis for core density were lower than expected (0.20–0.31). Age–age genetic correlations were higher than phenotypic correlations, particularly those involving young ages. Age–age genetic correlations were high, being greater than 0.75. Age–age genetic correlations had a moderately linear relationship, while age–age phenotypic correlations had a strong linear relationship with natural logarithm of age ratio. Optimum selection age for core density was estimated to be 5 years when calculations were based on both genetic and phenotypic correlations. However, age 5 was the youngest examined in this study and optimum selection age may be younger than 5. Generally, the optimum selection age was robust to changes in breeding phase and assumptions concerning age-related variation in heritability estimates. Early selection for core density would result in a correlated increase in earlywood density but little progress in latewood density or latewood proportion at maturity.

Inheritance of growth curves is critical for understanding evolutionary change and formulating efficient breeding plans, yet has received limited attention. Growth curves, like other characters that change in concert with development, often have higher heritability than age-specific traits. This study compared genetic parameters of height-growth curves with those of age-specific heights for a conifer, Pinus taeda L. Growth curves were fitted with: (1) a linear regression model, and (2) a non-linear model based on Richards' function using two sources of height data: two six-parent diallel tests assessed at age 2 to 10 years and two tests from a nested mating design with 222 parents assessed at 1 to 25 years. Additive genetic control of growth-curve parameters was moderate (h(2) = 0.06 to 0.26) and slightly lower than that for age-specific heights. Additive variance exceeded dominance variance for rate and shape parameters, but not for the asymptote. Genetic correlations among growth-curve parameters were high. Early selection on height was as efficient as selection on growth-curve parameters.

No abstract