Wood density is traditionally determined by a volumetric method that is accurate but expensive for large-scale sampling. A new device called the Resistograph was investigated for rapid assessment of relative wood density of live trees in progeny trials. Fourteen full-sib families of loblolly pine (Pinus taeda L.) produced by a six-parent half-diallel mating design were tested at four sites. For each family, wood density was measured with the traditional volumetric method and then compared with the Resistograph readings (amplitude). Amplitude had weak (0.29) to moderate (0.65) phenotypic correlations with wood density on an individual-tree basis over the four sites. The family mean correlation between the two measurements, however, was much stronger (0.92). The additive genetic correlation between the two measures was also high (0.95). Individual-tree breeding values of amplitude yielded more accurate rankings than phenotypic values. The rankings of the parental, general-combining abilities were identical for the two measures. Both wood density and amplitude were under strong genetic control at the family level (full-sib family heritability (h2fs) = 0.95 for wood density and h2fs = 0.85 for amplitude). The efficiency of using the Resistograph as a means of indirect selection for improvement of wood density was 87% at the family level. Results from this study suggest that the Resistograph could be used reliably and efficiently to assess relative wood density of live trees for selection in tree improvement programs. The method is rapid, nondestructive, and much cheaper than the traditional volumetric method.