A SIMPLE METHOD FOR ESTIMATING THE NUMBER OF SEED CONES ON INDIVIDUAL BLACK SPRUCE

A procedure for quickly and conveniently predicting the number of seed cones on individual black spruce (Picea mariana (Mill.) B.S.P.) trees in a seed production area is proposed. As one approaches a tree to be sampled, a quick visual estimate is made of the tree's seed cones. When this estimate suggests that the number of seed cones on the tree, N, is less than 13, it is faster to count them. But when N greater than or equal to 13, it is faster to measure the position, LbPOS, of the branch bearing the lowest seed cone, and to estimate the tree's seed-cone yield by taking the antilogarithm of 1.08 + ln(LbPOS - 0.125) and subtracting 1.0 from the result. The age of the tree at the time the seed cones were produced was incorporated in this model as an additional explanatory variable. A time study showed that on heavily bearing trees, a full count could take 80 times as long as this alternative procedure. The drawback to this alternative is the inaccuracy introduced by taking estimates rather than full counts. When this alternative procedure was applied to the sampled trees, the average relative bias, mean square error, and mean relative deviation in estimating N were 0.270, 16 100 seed cones(2), and 0.605 per tree, respectively. Although this study suggests the cutoff number of cones to be counted at N = 13, one may prefer spending more time to gather more accurate estimates. For this, a table provides the corresponding time and accuracy statistics for Various other possible choices of the maximum number of seed cones at which full counts can be conducted.