The Spatial Pattern and Scale of Variation in Eucalyptus globulus ssp Globulus: Variation in Seedling Abnormalities and Early Growth

Variation in seedling abnormalities and 2- and 4-year growth were studied in a trial in north-westem Tasmania established from 594 open-pollinated families from throughout the geographical range of Eucalyptus globulus Labill. ssp, globulus and populations intergrading with other subspecies. Most (77-80%) of the total (phenotypic) variation in growth traits occurred within families. The genetic variation between families within localities (within c. 10 krn; 13-15% of the total variation), between localities within regions (within c. 100-150 km; c: 4%) and between regions (3-8%) was generally highly significant. However, regional and locality components together accounted for only a small proportion of the total phenotypic variation. Little differentiation was detected between separate sampling sites within localities. Estimates of individual narrow-sense heritabilities were markedly higher than previous reports and were 0.38 for conic volume and 0.41 for height at 4 years, assuming a coefficient of relatedness of 0.4 amongst open-pollinated sibs. On average, progenies from the Otway Ranges region were the fastest growing at the test site, followed by those from King Island. Parent trees with high breeding values were concentrated in the Otway Ranges, Strzelecki Ranges and far southeastem Tasmania with the Bass Strait island localities having intermediate frequencies. Forest fragmentation through clearing for agriculture appears to have had a deleterious effect on the quantitative genetic structure of intergrade populations consistent with high levels of inbreeding. Remnant populations tended to have higher levels of severely abnormal seedlings, higher mortality and poorer growth, and higher heritability estimates and variability both within and between families. Advanced generation hybridisation and inbreeding due to long periods of isolation in small, relict populations may also have had similar effects. Populations sampled are, therefore, likely to have markedly different levels of inbreeding which may have inflated differences between localities and may have important consequences for the exploitation of this material for breeding.