Estimating the Number of Filled Seeds per Cone of Western Hemlock from Coastal British Columbia

Unopened western hemlock (Tsuga heterophylla) cones from coastal British Columbia were tested for the relationship between total filled seeds per cone and the number of sound seeds exposed by slicing through the center of the cone's long axis, and with cone length. Predictive linear relationships of total number of filled seeds per cone (TFS), based on number of filled seeds cut on the face through the cone's long axis (CC), were found in a cone sample bulked from a number of trees. The regression based on CC explained about 63% of the variation in TFS. More accurate estimates of TFS were found where cone length, and exponential factors of CC and cone length, were included in the analysis. Analyses of cones from seven parents did not find statistically significant trends of TFS on CC in all cases, and differed in slope in most cases. Sample sizes (cones) to estimate TFS to a predetermined level of precision were much larger than cones needed to accept or reject a tree for study. West. J. Appl. For. (11)2:44-49.