Quantitative studies of hybridization in wild plant populations

Abstract

The methods for studying populations containing hybrids have been discussed and the quantative methods for describing and analysing such populations have been explored. These techniques mostly depend upon the properties of normal univariate r multivariate distributions. If the size of a character is normally distributed in a population free of hybrids, its deviation from normality in other populations may indicate the presence of hybrids. Most methods of analysis require that the characters measured are uninfluenced by the environment. A method of using pairs of correlated meaurements when each is influenced by the environment has been proposed. The techniques described have been applied to the study of introgression in the genera Juncus and Primula. In the genus Juncus some evidence was found for the introgression of J. Effusus (meadow form) into J. inflexus; the data further suggested that J. effuses contained at least two ecotypes, a meadow and a woodland form. Because of these, evidence of introgrssion of J. effuses by J. inflexus was not sought. In the genus Primula the data suggested that slight reciprocal introgression of P. Vulgaris and P. veris had occurred. Plants resembling the artificial Fl hybrid between these species were found in the most places where the two grew intermixed: their numbers were always small. Very few putative backcross hybrids were observed. Their scarcity may be real or due to many of them reembling the Fl hydrib or the recurrent parent, as they are known to from artificial backcross families. The absence of hybrids other than the apparent Fl in the field may be due to their inability to compete with their parents. Artifical backcross families are easily secured and they are relatively fertile, so it would appear that external and not internal barriers seperate the species.