This study, based on 84 commercial logs of radiata pine grown in New Zealand, compares branch size, Pilodyn penetration, and longitudinal stress-wave velocity as methods for structural log selection. Four case studies were performed, each differing in the criterion used to define an acceptable structural log, i.e., different yields of machine-stress-graded F5 and better (AS/NZS 4063: 1992) or MGP8 and better (provisional grade assigned under Australian Machine Graded Pine system). It is clearly shown that for plantation-grown logs, where branches can be large (often exceeding 2 in, diameter), the best results are obtained by combining branch size and sonic grading. This method provided up to 14 percent fewer misclassified logs compared to using branch size or stress-wave velocity alone, and up to 12 percent improvement in structural lumber grade recovery compared to processing unsorted logs. The best results using the Pilodyn were obtained when logs were sorted for structural lumber production from the outerwood only, i.e., where the juvenile core, containing approximately the first 10 growth rings, was intentionally cut into alternative products.