?Even more than most disciplines in the Animal Sciences, quantitative genetics is dependent upon models. Models, by definition, are abstractions of reality. Invariably they require simplifying assumptions, which should be but sometimes are not clearly specified. One thesis of this article, illustrated by examples, is that many of the assumptions upon which animal breeding theory and practice are based are not valid. Some proportion of research resources should be devoted to challenging or verifying those assumptions and following up those areas of enquiry suggested by the outcome of such research. A further thesis is that the selection of topics and priorities for animal breeding research should be a matter of choice by individual scientists and should not be determined by steering committees or directed by administrative fiat. Hopefully, the resultant mutation, cross-fertilization, assortment, recombination and selection of ideas that would result would bestow upon our discipline higher fitness from multiple-peak epistasis, and minimal danger of extinction (or petrification) from over-specialization. A final thesis is that true creativity by research scientists should be nurtured and rewarded and that work in traditional areas of breeding and quantitative genetics should be continued-but done better.
