Artificial disruptive selection for early height growth in Patagonian Cypress

APARICIO, ALEJANDRO; PASTORINO, MARIO JUAN; GALLO, LEONARDO ARIEL

Alcalá de Henares

Conferencia; IUFRO Conference “Population Genetics and Genomics of Forest Trees: from Gene Function to Evolutionary Dynamics and Conservation”; 2006

IUFRO

Artificial selection against the intermediate phenotypes is not frequent in forest tree species. However, selection of the best and the worst types can significantly contribute to different research lines in advanced stages of  a breeding program, such as the formation of a segregating population for the detection of QTLs. A breeding program has been recently initiated for Austrocedrus chilensis (D. Don) Pic. Ser. et Bizzarri, the Patagonian Cypress, which is the most important productive conifer of the temperate forests in Argentina. Their marginal populations from the steppe have been found to be particularly relevant for breeding as well as for conservation purposes. In order to analyze differences in the shape of the early height growth curves among progenies and among natural populations of this species,  a greenhouse trial with a randomize complete block design was established under controlled conditions of  temperature, humidity and fertilization with 158 progenies corresponding to 10 marginal populations from the steppe (N = 3318 seedlings). Height from the cotyledons insertion till the terminal bud of the main axis was measured nine times across the whole second period of growth activity. Based on the final height a disruptive selection was performed at two levels: 1) best 100 and worst 100 individuals of the whole trial, irrespective of their origin or progeny belonging; 2) 21 individuals of each of the best 4 and worst 4 progenies of each origin. A height growth model was adjusted for each group, and differences among the adjusted models for each group at the two levels were tested by means of the Akaike Criterium of Information (AICc). Mean final height of best 100 and worst 100 was 338 ± 45 mm and 126 ± 18 mm respectively (significantly different, p < 0,0001; with 512 mm maximum and 65 mm minimum absolute values). In all groups, height growth was best described by second grade polynomial models of the form: log10h = a + b.t – c.t2 (h = height [mm]; t = time [days]), with significant differences in initial height (a parameter) and growth rate (b and c parameters). An additional observation is useful to illustrate the dispersion of data even within groups: when separating best and worst groups within each group at the level of analysis 2) (progenies selection in each origin), worst group of the best 4 progenies resulted similar to the best one of the worst 4 progenies of the same origin. Although differences in the initial height account for a great part of the differences among the growth curves of groups at the same level, their differences in the growth rate arise as a character of deeper and more permanent significance, since it describes a continuing process instead of a state.