GENETIC-PARAMETERS AND CORRELATIONS IN TESTS OF OPEN-POLLINATED BLACKSPRUCE FAMILIES IN-FIELD AND RETROSPECTIVE NURSERY TEST ENVIRONMENTS

Open-pollinated families represented in two series (1979 and 1980) of black spruce (Picea mariana (Mill.) B.S.P.) seedling seed orchards and associated field tests were also established in retrospective greenho use and nursery test environments. Height growth was assessed after va rying numbers of growth cycles and compared with performance in the or iginal field tests after 5, 10, and 15 years. The 1979 series consiste ntly produced higher estimates of heritability throughout the 15 years in the field and 6 years in the retrospective nursery test. Phenotypi c and genetic correlations among family means at the various field tes t sites were very strong for the 1979 series material, but generally m uch lower for the 1980 families. Correlations between performance in t he field and retrospective nursery test environments were very strong and highly significant for the 1979 material, with genetic correlation s in excess of 0.8. On the other hand, the 1980 material produced much lower correlations that were often not significant. In both cases, co rrelations with the nursery test environment appeared to peak at age 3 or 4 years. The accelerated greenhouse test environment produced very strong and highly significant correlations with field performance dur ing all measurement intervals far the 1979 series, again with genetic correlations in excess of 0.8. The two accelerated tests of the 1980 m aterial used differing growing regimes; one with artificially extended growing cycles, the other with compressed cycles. The extended-cycle regime produced highly significant phenotypic correlations with field performance in excess of 0.5 and genetic correlations above 0.6, while the compressed-cycle regime had weakly significant phenotypic correla tions of about 0.33 and genetic correlations of about 0.4. The two tes t series present conflicting demonstrations of the impact of early sel ection. The 1979 series suggests that family selection can be done wit h a high degree of reliability after three or four accelerated cycles in a single nursery environment and that family rankings are reliable after only 5 years in the field tests, whereas the 1980 series suggest s that longer test periods are required and that short-term nursery te sts are only effective in identifying the very worst families. Reasons for this conflict and implications for operational breeding programs are discussed.