Demand for genetically improved seed for establishment of eucalypt plantations has increased the need to intensively manage seed orchards, including predicting current and future seed availability. This requires developing a technique to reliably estimate an individual tree's seed crop. A possible approach is to sample a limited number of capsules from a given point in the crown, estimate the number of seeds per capsule and multiply this figure by the estimated number of capsules in the crown. This experiment aimed to identify the most representative sampling point in the crown of the tree, and evaluate the 'cut-test1, a technique currently in use for estimating the number of seeds per capsule. Capsules were harvested from 12 points around the crowns of 20 sample trees from a Eucalyptus globulins seed orchard and the number of seed per capsule estimated. The weight of extracted seed and the number of viable seed from 20 capsules at each point in the crown were measured. These results were then compared with the total yield of capsules and seed from each tree. Mean capsule weight increased up the crown, but there were no other significant differences in capsule or seed parameters between positions in the crown. The cut-test underestimated the number of seeds per capsule, and was not a reliable predictor of actual numbers of seed per capsule. Total capsule and seed yields were highly variable between trees. Further work is required to develop a technique to reliably estimate the number of capsuleson the tree. Estimates of the number of seed per capsule are best made by harvesting a limited number of capsules from the zone of the crown carrying the bulk of the capsules, and extracting the seed. Family differences need lo be more closely evaluated, and the feasibility of applying a mean family value for the number of seeds per capsule should be evaluated.
展开▼