Repeated measures analysis of variance revealed significant differences in total above-ground biomass among the parental types and the two hybrid generations over the first season of growth (repeated measures T = 4, F = 9.29, P < 0.0001, 2/66 df for cross group tested over the variance among genotypes; Fig. 2). There were significant differences in above-ground biomass in the first season of growth for the contrasts between the backcross and parental genotypes (backcross < midparent: F = 11.68, P < 0.0008) and the backcross and F1 genotypes (backcross < F1: F = 6.96, P < 0.0090) but not the F1 and parental genotypes (midparent = F1: F = 0.05, P = 0.8310). Vegetative size at the end of the first season explained a large proportion of the variation in over-winter survival in this (logistic regression: chi square = 5.55, P < 0.0185) and a previous study (Rhode and Cruzan 2004). Mortality of the smallest individuals tended to equalize vegetative sizes among groups for the second season (repeated measures T = 3, F = 1.00, P = 0.3704, 2/213 df). Total reproductive output differed among hybrid and parental genotypes (F = 4.79, P < 0.0091, 2/229 df for cross group tested over the variance among genotypes; Fig. 3), with the

Fig. 2. Means and standard errors for the vegetative component of fitness (estimated aboveground biomass) and phenotypic differences (leaf width/length ratio) for the caroliniana morphotype (C), an advanced-generation hybrid derivative (H), and the F1 (F1) and backcross (BC) hybrids from crosses between C and H of the Piriqueta caroliniana complex. For the C and H parental types and the F1 hybrids, only means and standard errors are shown. For the backcross hybrids both the mean (triangle with the error bar) and individual observations of each genotype (means of field-grown cuttings: triangles without error bars) are shown.

4000

S 3000

Backcross

0.15 0.2 0.25 0.3 Width:Length Ratio

0.35

Fj hybrids reproducing most (F = 9.47 with P < 0.023 and F = 8.75 with P < 0.0043 for Fj vs. parental and backcross genotypes, receptively) and backcross and parental genotypes reproducing equally (F = 0.20, P = 0.6550). It is important to note that the parental genotypes used in this study were inbred (i.e., they were from within-population crosses); therefore, the observed high reproductive output for Fj genotypes was probably partly due to heterosis, which was found to contribute significantly to Fj's fitness in a previous study (Rhode and Cruzan 2004). About half (47.88%) of the back-cross genotypes never reproduced, but some backcrossed plants (16.5%) produced relatively large numbers of flowers and fruits (>10); only 10% of the parental genotypes produced >10 flowers and fruits. Though mean vegetative fitness of parentals was higher than that of hybrids, ten backcrossed genotypes (15% of all backcrosses) had vegetative fitness estimates exceeding that of parentals (Fig. 2).

Was this article helpful?

## Post a comment