## LSU Historical Dissertations and Theses

1991

Dissertation

#### Degree Name

Doctor of Philosophy (PhD)

Paul E. Humes

#### Abstract

Data on eight different two breed combinations of calves produced in two locations (Louisiana and Nebraska) were analyzed to determine the magnitude of genotype x location interactions for gestation length (GESTLEN), birth weight (BWT), pre-weaning average daily gain (ADG) and weaning weight (WWT). Sire breed (SB), sire/SB and cow breed (CB) interactions with location (LOC) were the primary sources of variation of interest in the analysis of variance. The SB x LOC interaction was significant for all traits whereas the CB x LOC interaction was important only for ADG. The sire/SB x LOC interaction accounted for a significant amount of variation for ADG and WWT. Within LOC, heritability estimates were.47,.26,.30 and.26 for GESTLEN, BWT, ADG and WWT respectively. Across LOC, heritability estimates were.37,.19, and.02, respectively, for GESTLEN, BWT and WWT. A negative sire component was obtained for ADG in the across LOC analysis. Genetic correlations between GESTLEN and BWT (r$\sb{\rm g}$ = 49), BWT and WWT (r$\sb{\rm g}$ =.54), and ADG and WWT (r$\sb{\rm g}$ = 96) were all positive; the remaining were negative. The phenotype correlation was positive between GESTLEN and BWT (r$\sb{\rm p}$ =.32) small and positive between GESTLEN and ADG (r$\sb{\rm p}$ =.05) and small but negative between GESTLEN and WWT (r$\sb{\rm p}$ = -.04). Genetic correlations of the same traits across location were obtained to determine the magnitude of sire x LOC interactions. After correcting for heterogeneity of environmental variances, genetic correlations between LOC of 1.10,.76, -.05 and.12 were obtained for GESTLEN, BWT, ADG and WWT, respectively. Three methods of detecting the nature of genotype by environment interactions were illustrated for reproductive traits in beef cattle. The Azzalini and Cox (1984) procedure seemed preferable for it allowed the determination of the direction of qualitative interactions. Although the usual graphical representation of the means remained the method of choice in visual appraisal of a genotype x location interaction, two graphical display techniques were studied. One method related to the original data (Biplot), and an alternative to the analysis of variance for fixed effects (Analysis of means). They generally were in close agreement with the Analysis of Variance procedure.

100

COinS