Body measurements as selection criteria for growth in South African hereford cattle

Birth weight (BW), weaning weight (WW), yearling weight (YW) and seven body measurements (cannon bone length at birth (CB); hip height at weaning (HHW); hip height at yearling (HHY); body length at weaning (BLW); body length at yearling (BLY); scrotum circumference at weaning (SCW) and scrotum circumference at yearling (SCY), measured in a Hereford stud (1497 animals) over a period of 19 years, were used in estimating genetic (co)variances, heritabilities and correlations. Nine different multitrait animal model evaluations were carried out whereby (co)variance components were estimated using the REML VCE 3.0 package. Estimates from different evaluations were pooled, weighing each estimate by the inverse of the sampling variance to calculate weighted mean variance ratios among the different traits. Generally, structural traits tended to have lower heritability estimates (CB = 0.24; HHW = 0.28; HHY = 0.33; BLW = 0.22; BLY= 0.14 ) when compared with literature values , while estimates for production traits compared well with estimates reported for the South African National Evaluation. Weights and linear body measurements were positively correlated (0.4 0.9). Negative correlations were found for scrotum, circumference at weaning and yearling with BW and CB respectively. SCW and BW: -0.54; SCY and BW: -0.65; SCW and CB: -0.66; SCY and CB: -.58. It is concluded that CB could serve as an early indicator of mature size and weights at different ages.


Introduction
Herefords in South Africa increased in size from an average Frame Score 3 (119.38cm) to an average of 5 (129.54cm) at 15 months of age over a period of approximately 15 years (Hereford Breeders Society of Southern Africa, 1997).The past fifteen years of Hereford cattle breeding in South Africa were directed towards selection for larger framed animals with a higher growth rate.Hereford breeders used hip height as an additional measurement to weight in assisting them in the selection process (Hereford Breeders Society of Southern Africa, 1983) .Body measurements have been of recurring interest in beef cattle selection and breeding programmes.Measurements such as hip height or wither height were used to study skeletal development.BROWN et al. (1983) and GOSEY (1984) reported that hip height is a trait indicating maturity at an early age, followed by height at withers and width of Shoulders.Skeletal growth in growing animals takes priority for nutrients over muscle growth and fat deposition (BERG and BUTTERFIELD, 1976).Regardless of plane of nutrition, frame size of animals of the same age and sex should increase according to their genetic potential.Studies indicate that length of the cannon bone (MEYER, 1995) and hip height (GILBERT et al., 1993) could be used as indication of maturity at an early age.Reports on genetic parameters for body length and height are limited.Scrotal circumference and its relationship with fertility, has been studied extensively (NEELY et al., 1982;SMITH and BRINKS, 1989;MORRIS et al., 1992).Scrotal circumference is favourably related to both fertility in the male and its female offspring.Therefore it is of interest to evaluate the relationship of scrotal circumference, as indicator of reproduction, with other linear body measurements and weights.Hereford breeders are taking body measurements and weights at different ages and production stadia.It should therefore be established which measurements provide the most accurate and best Information for effective selection.The aim of this study was therefore to estimate the heritabilities and genetic correlations for the different measurements in order to evaluate these as selection criteria for growth and reproduction in Hereford cattle in South Africa.

Material and Method
Data for this study were obtained from a registered polled Hereford stud near Vryburg in the North-West Province of South Africa.The herd was established in 1978 and participated in the South African National Beef Cattle Performance Testing Scheme (KÖSTER, 1992).

Herd management
The herd numbers were maintained at an average of 250 breeding cows.Cows and calves were kept on natural pastures throughout the year.Natural pastures consist of a semi-arid Bush Savanna, classified by ACOCKS (1988) as the eastern Kalahari Thorn-veld Proper.Summer and winter supplementation of minerals, protein and energy were provided.The mating seasons were between mid-December to March and again from June to July.Approximately one third of the cows were artificially inseminated.Calves were weaned at an average age of 205 days.Bull and heifer calves were separated after weaning.Young bulls received a growth meal (14% protein, 72% TDN, one percent of live weight) from weaning to year old, while heifers were kept on natural pasture with supplementation.Calves with physical defects were culled at weaning.Cows were evaluated on reproductive and calf Performance.Emphasis was placed on selection for a leaner type of Hereford with an average Frame Score between 6 and 7 (137.16cm -142.24cm at eighteen months of age).For this purpose Hereford semen was imported from the USA, since 1982.

Data description
Including all pedigrees the data set consisted of 1497 records.A total of 51 sires and 376 dams were represented.The data were edited to eliminate incomplete records and outliers.The twinning rate was low, 1.5% of the data, and were excluded.All pedigrees were checked to ensure parents were born at appropriate times before their offspring.

Statistical methods
The GLM (General Linear Model) procedure of SAS (1995) was used to determine which fixed effects had a significant influence on the data.Age of dam was classified into three categories, i.e. dams younger than 30 months; dams 31-92 months of age and dams older than 92 months.As calves received creep feeding that could have had a cannulation effect it is expected that age of dam had little effect after birth.Animals were grouped according to the number born per month over the 19 years namely; January to April, May to August and September to December.A year-season concatenation was established and fitted as a fixed effect in the model for all traits, to account for possible interactions.Sex and age when weighed were also considered as fixed effects.A series of univariate REML estimations (MEYER, 1991) fitting animal modeis with füll relationships, were carried out.The appropriate mixed modeis, with the best likelihood fits, were decided upon for multitrait analyses, therefore in all analyses the direct-maternal covariance was ignored.Due to the data structure a multiple trait analysis, including all the traits, could not be performed, therefore the following nine analysis were carried out: Variance components were estimated with the REML VCE 3.0 package (GROENEVELD, 1994), approximating the first derivatives by finite differences.Optimization was done using the UNCMIN optimizer on the Cholesky factor.In Table 1 the final model fitted for variance component estimation for each trait is indicated.Since a series of (co)variance components and variance component ratios were obtained from different multitrait analyses, these (co) variances and ratios were pooled, weighing each estimate by the inverse of its sampling variance (SE 2 ) (KOOTS et al., 1994), as follows : hVoled^Xhi/lSE^jVXl/lSE^) 2

I=:I
Results and Discussion Characteristics of the data for all traits with phenotypic means and Standard deviations are summarised in Table 2.  Data strueture for birth weight (BW), weaning weight (WW), yearling weight (YW), cannon bone length at birth (CB), hip height at weaning (HHW), hip height at yearling (HHY), body length at weaning (BLW), body length at yearling (BLY), scrotal circumference at weaning (SCW) and at yearling (SCY) (Strukturangaben für Geburtsgewicht, Absetzgewicht; Jährlingsgewicht, Mittelfußknochenlänge bei Geburt, HUfthöhe beim Absetzen und als Jährling, Körperlange beim Absetzen und als Jährling, Hodenumfang beim Absetzen und als Jährling)      the off diagonals-Heritability estimates of the growth traits (Table 3) compared well with the National Genetic Evaluation of the South African Hereford (BW: h 2 = 0.283, m 2 = 0.108; WW: h 2 = 0.109, m 2 = 0.111 and YW: h 2 = 0.109, m 2 = 0.093) (MUSTERT, 1997unpublished data).Except for the direct heritability of BW, estimates found in this study were lower than the average estimates found in the literature (KOOTS et al., 1994).Contrary to the results of MEYER (1995), CB was not subjeeted to maternal effects.The direct heritability of CB compared well with that estimated by MEYER (1995) of 0.30 on Australian Polled Herefords.Literature estimates varied from 0.08 and 0.11 for length of rear and fore cannon in live animals to 0.48 for length of metacarpal bone of cattle slaughtered at fixed weight (WILSON et al., 1976).GILBERT et al. (1993) reported a heritability of 0.3 for CB measured at weaning.
Structural traits tended to be highly heritable.Estimates reported in the literature for height varied from 0.36 at weaning (MEYER, 1995) to 0.43 and 0.57 for Angus and Hereford at one year of age (GILBERT et al., 1993).The direct heritability found in this study at weaning for height was lower than most estimates, but still indicates moderate heritability.The estimate for hip height at yearling in this study was 0.33, also lower than the 0.59 reported by BULLOCK and BERTRAND (1993).Maternal heritabilities for hip height at weaning (0.10) and yearling (0.15) were low.This supports the findings of MEYER (1995) who indicated that maternal effects have little influence on hip height (0.04).
Length of body, at weaning and yearling, had lower direct heritabilities than that for hip height at the same ages, while the estimates for maternal genetic effects were low.
The maternal environmental effect had a greater influence on BLW (c 2 = 0.07) than the maternal genetic effect (m 2 = 0.02), while at yearling the reverse was found (c 2 = 0.04 and m 2 = 0.06) (Table 3).Direct heritabilities for yearling scrotal circumference of 0.53 for Hereford and 0.16 for Brangus were reported by KRIESE et al. (1990).TOELLE and ROBISON (1985) estimated heritabilities of 0.08 and 0.44 at weaning and yearling respectively.KEETON et al. (1996) found heritabilities of 0.46 for SC at weaning.Direct heritabilities estimated in this study were lower for SC at both weaning (0.29) and yearling (0.25).
Positive genetic correlations were found between BW and hip height (0.48 with HHW; 0.5 with HHY) and BW and body length (0.4 with BLW; 0.59 with BLY).The genetic correlation estimates of weight (WW and YW) with body length and height were highly positive.Estimates for weights with body length ranged from 0.66 (WW x BLW) to 0.84 (YW x BLY) and for weights with hip height ranging from 0.73 (YW x HHW) to 0.82 (YW x HHY).This suggests that selection for weight should favour taller and longer animals.A higher correlation was found (0.9) between CB and hip height, than between BW and hip height at weaning and year old.Cannon bone should therefore be a good early predictorof the height ,of the animal at both ages.Cannon bone was also highly correlated (0.88) with BW, but only moderately correlated with WW (0.53) and YW (0.48).The results of this study confirmed the findings of MEYER (1995) that showed CB as a good predictor of mature size.According to MEYER (1995) genetic correlations between CB and mature size ranged from 0.6 to 0.7.Scrotal circumferences at weaning and year old were negatively correlated (-0.54 and -0.65) with BW.Low positive correlations were found for SC with WW (0.29) and YW ( 0.37).KNIGHTS et al. (1984) reported low, but positive genetic correlations between weight at birth, at weaning and at year old with SC, suggesting that bulls superior for growth should sire sons with a larger scrotal circumference.High negative correlations were found between SC at weaning (-0.66) and yearling (-.58) with cannon bone.This indicates that selection for fast maturing animals will not necessarily lead to larger SC.COULTER and FOOTE (1977) reported a curvilinear relationship between SC and body weight and that the testes tend to reach mature size more rapidly.It is therefore not possible to conclude that selection for heavier animals will necessarily result in change in scrotal circumference.Body height and length at weaning and year old seem to have little effect on scrotal circumference.As scrotal circumference is recognized as a good indicator of fertility, it is important to pay more attention to possible negative and or positive correlations between scrotal circumference and body weight in long term selection for heavier bulls.

Conclusion
Maternal effects were evident in body measurements, even after weaning.As expected, ample evidence was found of positive genetic correlations among traits associated with skeletal development.Maturity height end-points had a negative genetic correlation with SC (CB, HHY) whereas height pre-weaning had no detrimental effect.Preweaning growth had no detrimental effect on SC, but there is reason for concern if selection is based on growth end-points.

ERNST RITTER, Dummerstorf
was not included in the model for the specific trait.

Table 1
Final model used for variance component estimation for each trait (Modell der Varianzkomponentenschätzung

Table 3
Weighted heritabilities (on diagonal) and genetic correlations (above diagonal) for all the traits (Gewogene

Table 4 (
continued) MARLE-KÖSTER et al.: Body measurements as selection criteria for growth in South African hereford cattleMATERNAL GENETIC VARIANCES AND CO-VARIANCES