The analysis of reproductive records with use of labelled sequences and its application to a grazing experiment

Abstract

A new method is described for analysing sequential results from animals in a reproductive situation. The elements of each sequence are labelled according to whether the animal is pregnant or not, and lactating or not; the resulting labelled sequences are used for analysis, and animals are compared only when they are in the same state at the same time. It is suggested that such a method obviates the necessity for correcting for the weight of the foetus, and thus enables animals which are not in exact reproductive synchrony to be directly compared. The method involves the computation of a dissimilarity matrix, which can be subjected to classificatory or ordination techniques and subsequent statistical analyses.

The method is applied to a grazing experiment in the dry tropics, and is shown to produce novel and informative results. The fertility and seasonal changes in liveweight of cows grazing a Townsville stylo-spear grass pasture were examined over 32 months. There were four randomized blocks of six treatments, which were factorial combinations of two stocking rates and three rates of superphosphate.

Superphosphate significantly affected fertility and liveweight, and the highest fertility was obtained at the intermediate fertilizer rate. Stocking rate affected liveweights only. There were greater liveweight differences between the four reproductive states than between treatments within a reproductive state. The largest treatment differences occurred in nonpregnant, non-lactating cows grazing fertilized as against unfertilized pastures.