Genetic-parameter estimates and parental breeding-value predictions were compared from open-pollinated and control-pollinated progeny populations of Eucalyptus globulus and two populations of E. nitens. For E. globulus there were two types of open-pollinated populations (native stand open-pollinated and seed orchard open-pollinated) and two types of control-pollinated populations (intra-provenance and interprovenance full-sib families). For E. nitens there were two populations, a seed orchard open-pollinated population and intra-provenance full-sib families. Progeny tests were established across multiple sites and 2-year height and diameter were measured and volume calculated. Genetic parameters from native stand open-pollinated E. globulus were unlike the parameters from the other three E. globulus populations; heritability estimates were severely inflated, presumably due to high levels, and possibly differential levels, of inbreeding depression relative to the other populations. Estimates of dominance variance in the E. globulus full-sib populations were high, but were zero in the E. nitens population. Correlations among parental breeding values, predicted using data from the different populations, were generally low and non-significant, with two exceptions: predictions from the two E. globulus full-sib populations were significantly correlated (r=0.54, P = 0.001), as were predictions from the E. nitens seed orchard OP and full-sib population (r = 0.61, P = 0.08). There was some indication that superior parents of E. globulus native stand open-pollinated families also tended to have above-average breeding values based on the performance of intra-provenance full-sib offspring. The consequences of these results for exploitation of base-population collections from native stands are discussed.

The growth of E. globulus and E. nitens pollen tubes in styles of E. globulus was examined in order to elucidate the site of the unilateral barrier to hybridisation. Pollen tubes of E. nitens failed to grow the full length of the larger E. globulus style. E. globulus pollen tubes grew an average of 1.4 mm per day for the first 4 days, compared with 0.8 mm per day for pollen tubes of E. nitens. From days 4 to 14, the growth of E. nitens pollen tubes slowed to an average of 0.2 mm per day and virtually no growth occurred after day 14. In contrast, E. globulus pollen tubes grew through the style and into the ovary between days 5 and 14. By day 28, at about the time of style abscission, E. nitens tubes had grown only 6 mm, well short of the full length of the E. globulus style (9-10 mm). A similar difference in growth was obtained in vitro where E. nitens pollen tubes were significantly shorter than those of E. globulus. A comparison also including E. ovata, E. urnigera and E. gunnii indicated a significant correlation between style length and in vitro pollen tube length. It is argued that the unilateral cross-incompatibility between E. globulus and E. nitens is due to a structural barrier arising from an inherent limit to pollen tube growth which is associated with pistil size.