One of the most important wood properties determining end use of Pinus radiata is modulus of elasticity (E). Determination of how E can be manipulated by silvicultural management requires information on how factors such as initial planting density and breeding series influence longitudinal variation in this property. Using measurements obtained from 17-year-old P radiata growing on a dryland site in New Zealand, the main and interactive effects of initial stand density, genotype and height along the stem on E was examined. Within the experiment seedlings from three breeding series (850, 870, 268), and cuttings from the 268 series of two physiological ages (one-year-old cuttings and three-year-old cuttings) were grown at initial stand densities ranging from 275 to 2551 stems ha(-1). Modulus of elasticity was significantly influenced by initial stand density (P < 0.0001), genotype (P = 0.003), and their interaction (P = 0.03). Modulus of elasticity increased from 5.1 to 7.0 GPa, between 275 and 2551 stems ha(-1). Across the range of stand densities examined, the 870 breeding series from a different "long internode" breed exhibited the highest average E of 6.6 GPa, which exceeded E for the three-year-old cuttings, 268 breeding series, one-year-old cuttings and 850 breeding series by 5, 10, 13, and 18%, respectively. These latter two series came from a "growth and form" breed. Modulus of elasticity significantly varied longitudinally from 5.1 GPa at the stem base to a maximum of 6.7 GPa at 5 m, or 27% of tree height, before declining to 5.0 GPa at the stem top. The significant interaction found between stand density and height along the stem (P < 0.0001) was attributable to two distinct patterns. The magnitude of longitudinal variation in E (as measured by the difference between maximum and minimum E) significantly increased with increasing stand density ranging from 1.8 GPa at 275 stems ha-1 to 3.1 GPa at 2551 stems ha(-1). In addition, the stem height at which maximum E occurred declined with increasing stand density, from a stem height of 7.7 m at 275 stems ha(-1) to 4.2 m at 2551 stems ha(-1). The significant interaction (P = 0.0091) between genotype and height along the stem was caused by large fluctuations in E within the three-year-old cuttings over the log length. (c) 2007 Elsevier B.V. All rights reserved.