Artificial selection in plants and animals.

Abstract

shall be dealing in this paper with the results of selection experiments in plant said animal populations, partly from the point of view of the effectiveness of the selection in actually producing improvement, but rather more in terms of the formation that these experiments give us about the architecture of quantitative genetic variation in different kinds of characters. I shall be dealing almost entirely vith outbreeding species. As I know very little about plants I shall have to restrict myself to those plant species which behave like animals. A theory of the short-term effects of selection has existed for almost thirty years. Fundamental to this is the concept of a basic random outbreeding population; and the theory, in giving a static statistical description of the population at present, allows an immediate prediction of the response of the population to selection pressure. Strictly speaking this prediction is valid only for a single generation and its usefulness over several generations will vary from population to population. In essence this is because the prediction of changes of mean in the population are in |terms of the genetic and phenotypic variance at the time of selection, but the theory cannot predict the change of the variances though these must be expected to change as well as the mean as selection proceeds. Experimental examination of the validity of this theory rests almost entirely on experiments with laboratory animals. I would name particularly the experiments of Mather on Drosophila and of various workers in Edinburgh on both Drosophila and mice. It is a rather odd fact that the group of workers, mostly in the United States, concerned with the development of the theory were in fact primarily concerned experimentally with inbreeding and crossing experiments on animals, and as a result there are very few critical selection experiments in larger animals which were started more than about ten years ago. The notable exception to this is Lerner’s well-documented selection experiment for egg production in poultry. There were of course many statistical analyses of the genetic variation in populations of plants and animals and results of these suggested that the response to selection in many characteristics of economic importance should be fairly rapid. Recent selection experiments have confirmed this and show that the wide extent of potential genetic variation that we know to exist in our laboratory populations and also in wild populations of animals is also found in populations of domestic plants and animals.