Abstract:
Animal Production in Australia Vol. 15 MATERNAL INFLUENCE ON THE TEMPERAMENT OF BOS INDICUS CROSS COWS G. FORDYCE*+ and M.E. GODDARD* SUMMARY The behaviour of cows during per rectal pregnancy diagnosis was observed on six occasions. Six scores were recorded for which high values indicate undesirable behaviour. The scores declined with increasing cow age and with decreasing body condition. Cows which had had a new ear tag inserted immediately prior to examination had slightly higher scores. The scores were moderately repeatable but had low heritabilities. There may be a maternal environment effect on this behaviour. INTRODUCTION Temperament is considered, by cattle producers, to be an important trait of cattle (Elder et al. 1980a and 1980b). Both genotype (Hearnshaw et al. 1979, Fordyce et al. 1982) and experience (Grandin 1980) influence temperament. Boorman (pers. comm.) suggested that cows have a strong influence on the temperament of their calves. This influence could be genetic and/or from the experience gained by the calf while with its mother. This paper reports on genetic and environmental influences on the behaviour of cows while being examined in a crush. MATERIALS AND METHODS This experiment used the Droughtmaster cow herd at the CSIRO research station 'Lansdown', which is situated 60 kilometres south of Townsville in a dry tropical environment. Pregnancy diagnosis by per rectal examination of these cows was carried out in June or -July each year from 1978 to 1981 and also in October 1979 and 1980. Cows were held individually in a crush without headbailing for examination and several aspects of the behaviour of each cow were The scoring system was the same as that described by Fordyce et al. recorded. (1982) for the CRUSH and RACE tests. The vigour of movement (MOV) was recorded on a 7-point scale from 1 to 7; the degree of audible respiration (BLO) was scored on a 4-point scale from 0 to 1.5; kicking (KIC), bellowing (BEL) and kneeling down (DOW) were scored 1 if they occurred and 0 if they did not occur and DOW = 2 if a cow laid down. The sum of these 5 scores seems to give the best overall measure of the animal's reaction to the situation and formed the TEM score. For each score, lower values indicate more docile animals. At the same A total of time body condition was assessed on a g-point scale (Holroyd 1978). 3 452 observations were made, spread roughly evenly over the six occasions, on a total of 889 cows. At the October 1979 observation, 249 out of 540 cows had new ear tags inserted while they were in the race immediately before the crush. In 1981 the cows were divided into nine groups and examined one group at a time over a period of days. The statistical analysis presented some difficulties because, on a least squares model, year, age and cow are confounded and because pedigree information was available on only some cows. Therefore a range of models was used. To estimate fixed effects the six behaviour scores were analysed by a mixed model * + Graduate School of Tropical Veterinary Science, James Cook University of North Queensland, TownsvillelQld 4811. Present Address: Qld Dept Primary Industries, Swan's Lagoon Beef Cattle Research Station, Millaroo,Qld 4807. 345 Animal Production in Australia Vol. I5 procedure in which year-group, age, ear-tagging, and a regression on body condition were fixed effects, and cow was a random effect with an assumed repeatability of 0.4. The repeatability of each score was calculated from a least squares analysis which took account of effects due to year-group, ear-tagging, year of birth, condition and cow within year of birth. Because both single and multiple sire mating has been used at 'Lansdown', the sire was known for only 444 cows with a total of 1 852 observations representing 63 sires. Heritability estimates were made from the between-sire variance components estimated from a model including effects due to year-group, ear-tagging, condition, age (a quadratic regression), sire and cow within sire. There were 280 pairs of dams and daughters with records available and these The scores were were used to calculate weighted dam-daughter covariances. corrected for the fixed effects estimated in the first analysis and the mean corrected score of each cow calculated. The weighting factor used was mm/(m + X) (n + A) where m = no. of observations on the dam, n = no. of This observations on the daughter and A = (1 - repeatability)/repeatability. weighting gi-ves an approximately maximum likelihood estimate. RESULTS AND DISCUSSION The fixed effects (other than year-group) for the six scores are given in Table 1. The MOV, BLO and TEM scores decreased with increasing cow age, probably Inserting a because older cows had had more experience in the handling routine. new ear tag causes a small amount of pain, but cows which had this experience The effect shortly before being scoredshowed only a small increase in scores. Cows in better condition had higher was significant for the MOV and TEM scores. This probably reflects generally depressed activity MOV, BLO and TEM scores. in cows in poor condition rather than a specific association with temperament. This is consistent with the finding of decreased temperament scores in cattle However, in young cattle with nematode infections by Fordyce et al. (1982). we have found low negative correlations between temperament scores and weight gain (Fordyce and Goddard, unpublished data). There were significant In 1981 the groups in 1981. but there were no consistent due to some variation in the differences between groups were tested differences between standards used for years and between measurement in three different sets of yards, yards. This variation may be scoring the behaviour. Bellowing, kicking and lying down occurred rarely, so the analyses of the BEL, KIC and DOW scores are rather imprecise. None of the effects were significant but there was a tendency for the BEL and DOW scores to decline with age. All the scores except KIC were moderately repeatable (Table of individual correlations showed that the repeatabilities were higher for scores taken four months apart than for scores taken apart. Thus although all cows improve with age, the ranking of fairly stable over several years. Examination 2). only slightly three years the cows stays 34 6 Animal Production in Australia Vol. 15 TABLE 1 Least-squares constants for factors affecting temperament TABLE 2 Repeatability, heritability and dam-daughter correlations TABLE 3 Between-and within-animal correlations The within-cow correlations (Table indicating that the errors involved in of the errors in scoring other traits. MOV and BLO is very high, and the other 3) except those involving TEM, are low scoring one trait are largely independent The between-cow correlation between correlations are positive supporting 347 Animal Production in Australia Vol. 15 our assumption that these six scores are all influenced by some underlying trait which might be called 'temperament'. The heritabilities and dam-daughter correlations in Table 2 are on the basis of single observations. The heritabilities are low, and lower than those of Fordyce et al. (1982). The test situation used here is comparable to their CRUSH test, which appeared to give lower heritability estimates than the RACE and BAIL tests. Also their results were for young cattle 9-22 months old, and it may be that with increasing age and experience, environmental variation becomes more important. The significant dam-daughter correlations for MOV and TEM are in contrast to the zero heritability estimations and suggest that cows have a non-genetic influence on the behaviour of their offspring which persists until the offspring are mature. In 1981 the cows were tested in small groups and an attempt was made to standardise testing conditions within each group. A separate analysis of the 1981 scores yielded higher estimates of heritabilities and dam-daughter correlations (e.g. for MOV they were 0.15 and 0.17) although the significance levels were the same as those given in Table 2. Culling cows on the basis of scores such as those used here would successfully remove bad cows from the herd but would lead to little genetic progress. Genetic progress would probably be greater if based on the average at several scores from other tests (e.g. the BAIL test), carried out at a younger age and under standardised conditions. If the maternal environment effect of cows on their offspring is substantiated by further research, this would be an additional reason for culling bad temperament cows. There was no carrel ation betwe en temperament scores and pregnancy status. Polled and dehorned cows had higher scores than horned cows. ACKNOWLEDGEMENTS We would like to thank Mr. R. D ixon and his staff at 'Lansdown' for their co- operation in the collection of th ese data. REFERENCES WATERS, K.S., DUNWELL , G.H., EMMERSON F.R. AN, J.F., MORRI S, d KNOTT, S.G. ( 1980a). Aust. Vet. J. X 2 5. 56: ~ KEARNAN, J.F., WATERS , K-S., DUNWELL, G-H., EMMERS ON, F.R., KNOTT Aust. Vet. J. X 2 19. 56: d MORRIS, R.S. (1980b) GODDARD, M.E. and SE1 FERT, G.W. (1982 . Pr 0 NC. Aus t. Soc. Anim. 3 29. Int. J. Stud Anim. Prob. 1: 1. 80 >HEARNSHAW, BA RLOW, R. and WANT, G. (1979). Proc. Aust. Assoc. Anim. Breed. 1 64. HOLROYD, R.G. (1978). In 'Beef Cattle Production in the Tropics', p-233, edito rs R.M. Murray and K.W. Entwistle. (James Cook University Press: Townsville ELDER, J.K., R.S. an ELDER, J.K., S.G. an FORDYCE, G., Prod. GRANDIN, T 348