Informations et ressources scientifiques
sur le développement des zones arides et semi-arides

Titre : Genetic relationships between growth and reproduction in cattle in the semi-arid tropics

Auteur : Kirby, Graham William Miles

Université de soutenance : University of Queensland

Grade : Doctor of Philosophy (PhD) 1996

Résumé partiel
The research studies reported in this thesis were made in the context of the genetic improvement of pastoral industry productivity in the semi-arid tropics of Northern Australia, and in the biological context of the evolution of relationships between body size and fitness.

The aim of the research was to test the hypothesis that, among semi-arid tropical cattle populations, genetic associations do not exist between growth or size and female reproduction traits. Thus the expectation were neither that genotypically larger cattle parents could be used to produce offspring with higher female reproductive rates than smaller parents, nor that parental breeds generating high levels of heterosis for progeny growth also would produce high levels of heterosis for progeny female fertility.

This thesis provides information on breed differences, breed additive and heterotic genetic composition, genetic parameters among individual animals within traits, genetic correlations between traits, and direct and correlated responses to short-term selection for weaning and post­ weaning growth rates. The research herd on the Coastal Plains Research Station near Darwin contained various cattle breed, hybrid, individual and selection generation genetic contrasts in the experimental designs. Important supplementary analyses were calculated on published data on various breed contrasts in Australia, Africa and USA.

Particular growth, weight and reproduction traits were arranged for analyses in a sequence of fixed and mixed models. Fixed models were calculated using the Multivariate Generalised Linear Hypothesis module in the SYSTAT program. Mixed models were calculated using the Derivative Free Restricted Maximum Likelihood program for (co)variance estimation and the PEST program for the Best Linear Unbiased Prediction of Estimated Breeding Values. The hypothesis finally was assessed from the pattern of the correlation estimates using the results obtained from the various models and from selection responses.

Présentation