Estimation of Genetic Parameters in Open Pollinated Progeny Trials from Plus Trees of Pedunculate Oak (Quercus robur L.) Selected in Posavina and Podravina and Podunavlje Seed Zones

Pedunculate oak (Quercus robur L.) is one of themost important species in Croatia, both ecologically and economically. In the 1950’s natural oak stands in Croatia were evaluated in terms of wood productivity and quality. The results and additional ecological studies delineated three seed zones and seven seed regions (NN107/08).Over the past decades, the periodicity of seed crops has not been regular, which has limited natural regeneration and artificial planting in partial naturally regenerated stands. In order to increase genetic quality and reduce seed periodicity, it was decided to establish clonal seed orchards (Vidaković 1996). Plus trees have been selected in three seed regions, based on ten evaluated traits, and three clonal seed orchards have been established in the areas of Forest Municipality Vinkovci, Našice and Čazma. Potential for genetic gain by selected plus trees can be evaluated by progeny trials in common environmental conditions and because of long lifespan of forest trees, it is crucial to evaluate it as early as possible. Considering biological and ecological characteristics of pedunculate oak, evolutionary factors shaping the genetic structure of its populations and existing research on its genetic variability in Croatia (Perić et al. 2000, Bogdan et al. 2004, 2008, Kajba et al. 2006, Katičić et al. 2010), it can be assumed that most of its genetic variability lies on the intrapopulation, namely individual level. Such a pattern of variability points to considerable possibilities of genetic improvement by individual selection. The aim of this research is to estimate the trends of genetic parameters, heritability and genetic gains for height and diameter growth of pedunculate oak in some Croatian provenance regions. Research area, material and measurements Three open pollinated progeny trials of pedunculate oak plus trees have been established, representing three seed regions:• _ Progeny trial Vukojevački Šikar (TP1), age 2+18 years, established with 21 OP families of selected plus trees from the seed region »srednja Podravina«, control plants from three local populations and one minus tree family.• _ Progeny trial Čazma (TP2), age 2+7 years, established with 25 OP families of plus trees from the seed region »Donja Posavina« and control plants from local populations.• _ Progeny trial Kunjevci (TP3), age 2+7 years, established with 24 OP families of plus trees from the seed region»Gornja Posavina i Pokuplje« and control plants from local populations. The experimental design in TP1 is completely randomized, with five repetitions and in TP2 and TP3 randomized complete block design is applied, with five blocks. In all the trials, families are planted in groups of six, with 2×2 spacing. In TP1 height was measured at the age of 2+3, 2+5, 2+7, 2+8, 2+9, 2+11, 2+15 and 2+18 (H5, H7, H9, H10, H11, H13, H17 and H20). Diameter at breast height was measured at the age of 2+9, 2+11, and circumference at breast height at the age of 2+15 and 2+18 (D11, D13, C17 and C20). In other two trials heights were measured at the age of 2+3, 2+4 and 2+7 (H5, H6 i H9). Mean values of selected families and control heights for all three trials and for circumference in TP1 were calculated for each year of measurements, as well as standard deviations, standard errors and coefficients of variation. Variance analysis was done by GLMprocedure using linear models (1) for TP1 and (2) for TP2 and TP3. Additive and environmental variance components and additive coefficient of variation were calculated using REML method of MIXED procedure in SAS for Windows 8.0, using Equation (3), (4), (5) and (6). Narrow-sense heritabilities were calculated on the basis of individual tree values (hi 2) and family means (hF 2) using equation (7), (8) for TP1 and (9), (10) for TP2 and TP3. The realized gain from the test and expected genetic gains by two possible methods of height selection were calculated by three methods, including the following:• _ Realized gain (R) was calculated as the difference between means of selected plus tree progenies and means of unselected control plants, (Equation 11)• _ Expected genetic gain by individual within provenance mass selection of first generation plus trees at the same age as those represented in the studied trial, (Equation 12)• _ Expected genetic gain by backward selection among first generation plus trees. (Equation 13)ResultsAt the age of 2+18, the average survival in TP1 was 82%. In TP2 and TP3 at the age of 2+9, survival rates were 80% and 94%, respectively. Survival rates per individual families are shown in Fig. 1 (TP1), 4 (TP2) and 6 (TP3). Mean height of selected OP families in TP1 was 8.52 m, and average of control plants was 8.45 m (Fig. 2, Table 2). In TP2, OP families had the average height of 306.8 cm, while control plants reached higher average – 310.4 m (Fig. 5, Table 2). In TP3, control plants with the average height of 122.0 m were significantly lower than the selected OP families, whose mean height was 187.6 cm. (Fig. 7, Table 2). Variance analysis results are shown in Table 1. Variance component caused by family effect for height was statistically significant in trials TP1 and TP3, and non significant in TP2. Variance component from family × block interactionwas significant in both trialswith RCB design and had the highest values in TP2. Fixed block effect was only significant in TP3. Additive variance component had high values throughout most of the researched period in TP1 and TP3, while in TP2 they were low. The opposite stands for environmental variance component. Variance analysis for the trait of circumference in TP1 also showed high values for environmental variance component. Mean height, diameter and circumference values for selected OP families and control plants, as well as estimated genetic parameters for those traits are shown in Table 2 Discussion In TP1 and TP3, the family effect for height was statistically significant for the entire period, showing decreasing trend in TP1 and rising in TP3 (Table 1). The family effect was never significant in TP2, pointing to narrowed individual variability of plus trees in that seed region. The results suggest significant individual genetic variability in two out of three seed regions. Some authors report correlation between flushing phenology and height growth (Baliuckas and Pliura 2003, Perić et al. 2000). Based on that assumption it could be speculated that high additive variability for heights in TP1 and TP2 results from significant individual variabilities for flushing traits in selected plus trees and ecotypical differentiation of their original populations. However, similar research in TP1 showed no correlation between flushing phenology and height growth (Bogdan et al. 2009). Presumably, late frosts in this trial did not cause better competitiveness of the later-flushing plants and trial management in the early years removed competition from weeds, which would put the early-flushers forward. If we keep the assumption of height-flushing phenology correlation, then the results in TP2 point to greater uniformity in flushing of selected plus trees from the corresponding seed region and lack of ecotypic differentiation of its populations.However, high values of environmental variance components and family × block interactions suggest strong environmental influences in that particular trial, which could have led to underestimation of additive variance. High values of additive coefficients of variation in TP1 and TP3 suggest wide genetic variability of selected plus trees and good adaptability of their progeny to changing environmental conditions. Heritabilities High estimated heritabilities in TP1, until the age of 13, result from high component of family effect (Table 1 and 2). Average survival was, in the first years, slightly lower than in TP1, suggesting stronger competition with weeds and negative effect of rodents. Low estimated heritability values for height in TP2 reflect the low family effect variance component, probably caused by negative environmental influences. Genetic gains from selection In all the trials the estimated expected genetic gains by backward selection among first generation plus trees after open-pollinated progeny testing (G2) showed the highest values (Table 2). Realized gains were not very high during the entire period in TP1, showing the decreasing trend in recent years. In TP3 realized gain is improving, due to higher heritabilities and greater differences between selected and control families. Opposite to that, low heritabilities, non significant family variance component and small differences between selected and control families resulted in little, even negative realized gains in TP2