ESTIMATION OF VARIANCE COMPONENTS OF INTERNODE LENGTH IN A CLONED POPULATION OF MIMULUS GUTTATUS

Abstract

a plant is vegetatively reproduced, or cloned, so that the same genotype can I;evelop in more than one environment, differences in a given character between the individuals of the clone reflect the effect of the different environments on the expression of the character. If a population of plants is cloned and randomly grown in a nonuniform environment, it is possible to estimate environmental and total genetic components of variation by evaluating the variation within and between clones. BURTON and DEVANE (1953) computed environmental and total genetic components of variance, and broad-sense heritability, for seven characters using replicated clones of several ecotypes of tall fesque. COOPER (1959) further refined the technique by using parent-off spring regression to estimate narrow-sense (additive) heritability of Lolium ear emergence, and by comparing broad-sense and narrow-sense heritabilities for evidence of nonadditive effects. It is the purpose of this study to further explore variance component estimation using clones, and to compare such estimates to independent estimates using more conventional designs. MATERIALS A N D METHODS Mimulus guttatus Fisch., the yellow monkey flower commonly found growing on stream banks and seeps in western North America, was chosen for these experiments. It is an outcrossing perennial which clones in nature, and has proven to be an excellent research organism. The logarithm of the length of the six inter-nodes above the first flowering node was the quantitative character chosen for study (Figure 1). This character develops late enough so that differences in the condition of the vegetative parts used for cloning did not strongly influence the development of the character (LIBBY and JUND 1962). Analysis of a 105-plant population sample indicated that the logarithm of this measurement is normally distributed (x2 = 3.35, x2 95 = 15.51 : 8 df), has negligible skewness (t = 0.06, t 9 5 = 1.96), and slight but nonsignificant kurtosis (t =-0.81, t.95 =-1.96), while analyses of other samples indicated that the variance was uncorrelated with the mean level.