THE SPATIAL PATTERN AND SCALE OF VARIATION IN EUCALYPTUS-GLOBULUS SSPGLOBULUS - VARIATION IN SEEDLING ABNORMALITIES AND EARLY GROWTH

Variation in seedling abnormalities and 2- and 4-year growth were stud ied in a trial in north-western Tasmania established from 594 open-pol linated 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 grow th traits occurred within families. The genetic variation between fami lies within localities (within c. 10 lan; 13-15% of the total variatio n), between localities within regions (within c. 100-150 km; < 4%) and between regions (3-8%) was generally highly significant. However, reg ional and locality components together accounted for only a small prop ortion of the total phenotypic variation. Little differentiation was d etected between separate sampling sites within localities. Estimates o f individual narrow-sense heritabilities were markedly higher than pre vious 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-polli nated sibs. On average, progenies from the Otway Ranges region were th e fastest growing at the test site, followed by those from King Island . Parent toes with high breeding values were concentrated in the Otway Ranges, Strzelecki Ranges and far southeastern Tasmania with the Bass Strait island localities having intermediate frequencies. Forest frag mentation through clearing for agriculture appears to have had a delet erious effect on the quantitative genetic structure of intergrade popu lations consistent with high levels of inbreeding. Remnant populations tended to have higher levels of severely abnormal seedlings, higher m ortality and poorer growth, and higher heritability estimates and vari ability both within and between families. Advanced generation hybridis ation and inbreeding due to long periods of isolation in small, relict populations may also have had similar effects. Populations sampled ar e, therefore, likely to have markedly different levels of inbreeding w hich may have inflated differences between localities and may have imp ortant consequences for the exploitation of this material for breeding .