Factors associated with growth in beef cattle

New Zealand Journal of Agricultural Research ISSN: 0028-8233 (Print) 1175-8775 (Online) Journal homepage: http://www.tandfonline.com/loi/tnza20 Factors associated with growth in beef cattle P. J. Brumby, D. K. Walker & R. M. Gallagher To cite this article: P. J. Brumby, D. K. Walker & R. M. Gallagher (1963) Factors associated with growth in beef cattle, New Zealand Journal of Agricultural Research, 6:6, 526-537, DOI: 10.1080/00288233.1963.10420010 To link to this article: https://doi.org/10.1080/00288233.1963.10420010 Published online: 05 Jan 2012. Submit your article to this journal Article views: 2425 View related articles Citing articles: 10 View citing articles Full Terms & Conditions of access and use can be found at http://www.tandfonline.com/action/journalinformation?journalcode=tnza20

526 FACTORS ASSOCIATED WITH GROWTH IN BEEF CATTLE By P. J. BlUMBY, D. K. WALKER, AND R. M. GALLAGHER Ruakura Anhnal Research Station, Department of Agriculture, Hamilton i Receiced 18 [anua.v 1963) ABSTRACT The following conclusions were reached from the remits of an analysis of the growth of 1,072 young Aberdeen Angus bulls bred and reared in pedigree herds, an analysis of the growth of 900 steers and heifers bred in one herd, and a study of the relation between the milk Yield of a heterogeneous herd of 25 cows and the growth rate of their calves, when all animals were maintained under pastoral conditions:- l. The correlation between weaning weight and the darn's milk supply when weaning takes place at six to eight months of age is 0.7. 2. The correlation between weaning weight and weight at 21 months is 0.5-0.6, and the correlation between weaning weight and subsequent rate of gain from weaning to 21 months of age is negligible for bulls and about--o.2 for steers and heifers. 3. There appears to be little correlation between winter and summer liveweight gain. 4. The heritability of growth rate is of the order of 0.4. 5. With the possible exception of eye muscle area, the correlated responses in major carcass traits associated with increased growth rate appear favourable, INTRODUCTION Recent work on performance recording in beef cattle carried out in the U.S.A. and Britain has aroused considerable interest amongst both breeders and geneticists. The results published so far show that the method promises to be a valuable one for improving the efficiency of beef production. Most publications have discussed results obtained with cattle fed in "feed lots" or with stall-fed animals; relatively little is known of the growth of cattle fed entirely under grassland conditions. In 1955 a long-term project was started at Ruakura with the main objective of studying the improvement of growth rate by selective breeding in beef cattle. Other problems associated with liveweight gain in cattle are also under investigation in association with this major project. This paper reports results arising from these latter investigations; the results of the selective breeding programme involving selection for liveweight gain will be described in a subsequent paper. N.Z. ]. agric. Res. 6: 526-37

P. J. BRUMBY, D. K. WALKER, AND R. M. GALLAGHER 527 Briefly stated, the points considered to date may be summarised as follows:- 1. The relation of weight at 21 months to weaning weight and post-weaning liveweight gain. 2. Factors influencing weaning weight. 3. The correlation of winter and summer rates of gain. 'to The relation of rate of weight gain to various carcass traits. MATERIALS AND METHODS The investigations were made by analyses of existing reco:ds of body weights in the selected lines of cattle and in the seven pedigree herds participating in the selection experiment referred to above. The base population of 300 Aberdeen Angus cows in the selection herd from which the high and low gain selection lines were started consisted of a normal mixed age herd which had been bred by the usual technique of buying replacement bulls, selected for conformation, from pedigree breeders. This herd was kindly provided by H.M. Prisons Department and located at Waikeria Youth Centre near Ruakura. The pedigree herds associated with the selection experiment were normal herds ranging in size from 60 to 500 cows. The breeding policy was one of buying replacement bulls from fellow breeders. Young bulls were selected from these herds for use in the selection herd on the basis of their individual rate of gain in relation to the average of the herd in which they were bred. All animals in both the pedigree herds and the selection herd were fed under the free grazing conditions characteristic of New Zealand pastoral farming whereby meadow hay is the only supplement provided during the usual winter deficiency of pasture. Calving was confined to the early spring months. Only in special groups of cattle (Section 2) were calves identified and weighed at birth. The normal practice with most cows was to group them according to month of calving and then to identify, and weigh (and castrate the male calves) at six to eight weeks of age. Weaning occurred at an average age of 180 days. The optimum weight to kill Aberdeen Angus cattle in New Zealand appears to be in the vicinity of 1,000-1,100 lb liveweight. If cattle are heavier they incur an increasing feed cost associated with increasing fat deposition; if lighter than 1,000 Ib the increased overheads of breeding stock become appreciable (MacDonald 1958). Because of the seasonal nature of pasture growth it is desirable that fattening cattle be sold prior to the onset of the winter feed shortage; thus the objective of the beef farmer will be to produce an animal weighing 1,000-1,100 10 liveweight within 21 months of birth. The measurement of liveweight at 21 months is therefore a critical assessment of the success of a beef production programme. At least four liveweight measurements were made routinely on each animal in each herd. i.e., in March at weaning, in May at the onset of winter, in September at the end of winter, and again in May at approximately 21 months of age. In addition to this material some data, derived from a further experiment in which the milk yields of various beef cows were measured by weighing calves before and after suckling, are included.

528 Factors associated with growth in beef cattle RESULTS 1. The relation of weight at 21 months to weaning weight and postweaning liveweight gain Data for this analysis were derived from two sources. The liveweight measurements made on young bulls bred and reared in the seven pedigree herds comprised the first source. Four successive annual calf crops were weighed and the body weights analysed within years and farms on a within-sire basis. In non-orthogonal data of this type the computation of the between-year component is a rather clumsy exercise and it has been omitted from this analysis. TABLE 1. A nalyses of variance of weaning weight (1), 21 month weight (2), and gain from weaning to 21 months (3) of young bulls in breeders' herds Source d.i. I M.S. (1) 1v1.S. (2) M.S. (3) I COy (1) (2) -- I Herds 24 64.558 383,545 312,566 67,768 Sires 1C4 11:461 19.609 6,684 12,191 Within sires, 943 2,917 6,716 3,601 I 3,017 - x 1b 537 I 970 433 --- Weighted means \ Prorenv/ sire 3.0 Sire.i/herd 4.6 Table 1 summarises the analyses of variance made on weaning weight, 21 month weight, and gain from weaning to 21 months. It is apparent that the sire components for weaning weight and 21 month weight are unrealistically large for they lead to heritabilities of around 100 per cent. This observation highlights the fact that environmental and management practices in pedigree herds, such as assortative mating, mating young bulls to young cows, and differences in sire group calving dates, inflate the observed differences between sires in their apparent breeding values. This complication to the interpretation of the analysis is less likely to influence the difference between two weights taken on the same animal at successive stages of growth (see Discussion) ; thus the analysis for liveweight gain is likely to be a better guide to the components, of variance due to herds and bulls. The weighted mean number of progeny/sire was eight, with 4.6 sires/herd, leading to a heritability estimate for liveweight gain of 0.38 ± 0.1. TABLE 2. Regression and correlation coefficients (within sires) for young bulls Regression coefficient (lb) Correlation coefficient 21 month weight on weaning wei rht 1.03 0.63 G:cin from weaning to 21 months on weaning weight 0.03 0.02

P. J. BRUMBY, D. K. WALKER, AND R. M. GALLAGHER 529 Based upon within-sire estimates of mean squares and mean products, the resulting regression and correlation coefficients of 21 month weight, and of gain from weaning to 21 months, on weaning weight, were completed, and are summarised in Table 2. The regression of 21 month weight on weaning weight proved to be of the order of one, i.e., for every ] lb increase in weaning weight there is a corresponding 1 lb increase in 21 month weight, the correlation between weaning weight and 21 month weight being 0.68. As expected from this result the relation between weaning weight and subsequent rate of gain proved small; in other words, heavy weaners made almost the same weight gain to 21 months of age as did the light weaners. TABLE 3. Analyses of variance of weaning weight (1),21 month weight (2), and gain from weaning to 21 months (3) for heifers and steers I Source d.f, M.S. (1)! M.S. (2) ------ - I M.S. (3)! COy (I) (2) Heifers (8.9/sire) Between sirca 27 2,702 6.943 5,348 2,191 withn years Within sire; 238 1,797 3,124 2,450 II 1.230 x Ib 395 722 I 327 Steers (8.9/sire) Between sires 27 4.609 12,744 6,774 5,303 within years Within sires 239 2,295 4,252 2,969 1.810 x lb 423 868 145 The second series of data derives from the selection herd of 300 cows m which liveweight measurements were made on all steer and heifer progeny. In these data as with the data from bull breeding herds, the between-sire variance is biased by virtue of the selection of sires as either fast or slow gain animals, but in contrast with the pedigree herd data the cows these bulls mated were chosen at random. Tables 3 and 4 summarise the analyses of variance and covariance and the regression and correlation estimates made on the weaning weights, 21 month weights, and gain from weaning to 21 months, for steer and heifer progeny. TABLE 4. Regression and correlation coefficients within sires for heifers and steers Heifers Regression coefficient (lb) Correlation coefficient Steers Regression coefficient (lb) Correlation coefficient 21 month weight on weaning weight 0.68 0.51 0.80 0.56 Cain from weaning to 21 month on weanin-; wei zht 0.32 ± 0.07-0.20 0.21 ± 0.07-0.18

530 Factors associated with growth in beef cattle Because of the selection of the sires used in this herd the sire component of the variation in these animals will be somewhat increased for both rates of gain as well as actual liveweights; in consequence, heritability figures derived from these data will be likely to over-estimate the true value. The values actually obtained were 0.48 for the heifers and 0.50 for the steers. Within-sire covariances, however, lead to reasonable estimates of regression and correlation coefficients. A comparison of the estimates of these coefficients based upon bull breeding herds (Table 2) with those based on the steer and heifer progeny in the selection experiment (Table 4 i reveals an interesting difference between young bulls and growing steers and heifers. In the former case weaning weight and subsequent gain to 21 months are unrelated; in the latter case the large wcancrs gain a little less weight to 21 months of age than do light weanen. It seems Iikcly that this situation comes about simply because of the earlier maturity associated with heifers and steers compared with bulls. Difference in the environment of the bulls compared with the heifers and steers may also be involved. It may be said, then, that the heritability of liveweight gain from weaning to 21 months of age is of the order of 0.4, that sire influences on weaning weight and 21 month weight respectively are appreciable, that the correlation of weaning weight and 21 month weight is of the order of 0.5-0.6, and th : correlation of weaning weight and subsequent gain to 21 months is small. Because of the significant regression and correlation between weaning weight and 21 month weight, weaning weights assume considerable economic importance. It is convenient, then, to partition the growth of an animal into a pre-weaning and post-weaning growth phase and to study factors affecting each. 2. Factors influencing weaning weights Of the many possible influences operating to affect weaning weight the nutritional status of the calf is undoubtedly a most important one. Previous studies (see Mason 1951) have indicated that some 50 per cent of the variation in weaning weight may be accounted for by differences in the milk production of the dams. One of the authors of this paper!d.k.\'\t.) recently undertook an extensive experimental investigation into the milk yields of various breeds of beef cattle. The detailed results of this investigation will provide the subject matter of a further paper, but a brief outline of some results of this work is relevant to this discussion. Milk yields of cows were estimated by the weighing of calves immediately before and immediately after suckling, which occurred twice daily. The cattle involved were a heterogeneous cojlection of several breeds and crosses, each of which was sired by a bull of the Aberdeen Angus breed and then mated to an Angus bull; thus all calves were at least i-bred Angus cattle. The relation between liveweights of the calf and milk yield of the dam is illustrated in Figs. 1 to 6. Figs. 1, 2, and 3 illustrate the relation between liveweight and milk production at successive stages of growth, while Figs. 4. 5, and 6 illustrate the effect that milk consumption has on liveweight gain at successive ages in the calf.

P. J. BRUMBY, D. K. WALKER, AND R. M. GALLAGHER 531 140.!,p 120 <; {IOO FIG. I.....,..... b : 0-17 r,,0 93 80 I FIG. 4 70 "'! ooj. 50 ;; r ;0............ b : 0 II : 078 g 260, 240,i 220! '0 ZOO 180 i 160 400 500 600 700 Milk Consumption 0-6 Weeks (Ib) FIG 2....1..... b: 010 r : 0 90 1.000 1,200 1,400 1,600 Milk Consumption 0-12 Weeks (Ib) I FIG 5 400 500 600 700 Milk Ccnsumottcn 0-6 Weeks Ob) --, - 110 -- I IOO I - 90 j.... "'80 :.. I 160..,. '2J I I I I, 500 600 700 800 900 1,000 Milk Consumption 6 12 Weeks (Ib)..... o "'70. b=006 :;420 1 ;t380. ȧ 340 i 300 FIG. J........\.. b 005 r o-n -5 180 160! c 140 8.g,120 FIG 6........ b 002 r " 0-31 2,000 2,500 3,000 Milk Consumpttor't 0-24 Weeki (Ib) 900 1.200 1,500 1.800 Milk Consumption 12-24 Wllk. UbI Figs. 1-3.- -The effect of milk consumption on liveweight. Figs. 4-6.-The effect of milk consumption on livcweight gain. The correlation between liveweight at birth and total milk yield to 12 weeks of age in these data proved to be non-significant with a value of 0.15. Any possible effect that birth weight has on the observed milk yield of the dam appears to arise through an appetite effect, with larger calves at birth demanding and consuming slightly more milk than smaller calves. There appears to be no common factor increasing both birth weight and milk production, for when the birth weights of 200 dairy heifers calving at Ruakura were correlated with their dams' milk yield the correlation proved to be negligible.(brumby unpublished data). It appears probable that a substantial relation between birth weight and milk production would exist if the young calf was incapable of consuming all its dam's production, but it is interesting to note from Fig. 1 that the limit to a young calf's capability appears to be in excess of 16 lb /day, a production figure unlikely to be attained by many beef cows.

532 Factors associated with g/owth in beet cattle Figs. I, 2, and 3 illustrate the marked influence that quantitative deficiencies in milk consumption have upon the growth of the young calf; at the early stages of growth this is almost absolute but it declines gradually with increasing age. This latter point is made clear by Figs. 4, 5, and 6, which illustrate the declining dependence of the liveweight gain of the growing calf upon its milk consumption, the regression of Jiveweight gain on milk consumption declining markedly with increasing age, while the deviations from the regression line increase with age. Interestingly enough the regression of weight gain on milk consumption appears linear over the range of milk yield encountered, an observation with considerable practical implications in the production of veal. At weaning the correlation between the liveweight of the calf and its dam's milk production proved to be 0.7; in other words, some 50 per cent of the variation in weaning weight may be attributed to differences in milk consumption. Because the correlation between milk production and liveweight declines with increasing age the influence of the sire upon weaning weight will be partly dependent upon the time weaning takes place; with early weaning sire effects may be expected to be smaller than would be the case with later weaning. 3. The correlation of winter and summer rates of gain Post-weaning growth may be conveniently considered in two main phases: the growth made during the winter period, when feed is generally limited; and the subsequent spring, summer, and autumn growth, when feed is usually plentiful. This approach raises the question of whether animals with good gain under plentiful feeding also exhibit good relative gains under the limited feed supplies associated with winter. In an attempt to answer this problem, an analysis of mean squares and mean products between and within sires within years was made of winter gain and subsequent gain to 21 months. This analysis is summarised in Tahle 5. TABLE.'1. Analysis of mean squares and mean products of winter gain (1), and surnmer-riut.umn gain (2), for steers and heifers.f_'_l I Cov. M.S. Source M.S. (2) Correlation coefficient - -- n c'fers Between sires 27 2,288 1.536 128 0.07 within years Within sires 229 1,113 1,300-142 -0.12 Mean gain (lb) 93 191 St ee rs Between sire) 27 1,836 '1.213-19 0.00 within years Within sires 241 1,000 1,890-302 -0.22 Mean gain (lb) 118 272 I

P. J. BRUMBY, D. K. WALKER, AND R. M. GALLAGHER 533 It is interesting to note that in spite of the limited liveweight gains achieved during the winter period substantial between-sire influences remained in the gain of the different progeny groups during this period. The covariance between winter and summer gain is surprisingly small, and it appears that rate of gain in the winter bears little relation to subsequent summer gains. This low correlation may be attributed in part to compensating growth and in part to appreciable errors in the estimates of liveweight gain, particularly in the winter growth period. To what extent it may be attributed to an interaction between individual genotypes and the differing pasture environments associated with winter and summer remains to be clarified. TABLE 6, Analysis of variance of carcass traits of steers (241 degrees of freedom for individuals, 27 degrees for sires) (for mean values see Table 8) M.S.w. M.S.b. Carcass weight Fat cover* Eye muscle Length of side Total points Carcass weight gain/day * (sum of C + D measurements) 1,880 40.2 1.78 1.20 88 0.0039 3,010 70.0 2.91 3.06 293 0.0115 4. The relation of weight gain to various carcass traits The final point to be considered here is the relation of rate of liveweight gain to various carcass traits. Table 6 summarises the analysis of variance within and between sires in the variation in cold dressed weight, fat cover, eye muscle area, length of side, and total carcass points of the steer progeny of the first 30 sires used. Carcass measurement and pointings were made in accordance with the New Zealand carcass evaluation system as derived by Kneebone et al. (1950). The actual age at slaughter in these animals averaged 860 days. A somewhat disconcerting feature of these results may be observed in the appreciable sire component in cold carcass weight, for an attempt was made to kill all animals each year at approximately the same liveweight. Because of the size of the sire component in carcass weight it is not surprising to find appreciable sire effects in the carcass measurements made. This difficulty might be overcome by using an analysis of the covariance of each carcass assessment on carcass weight, but such an adjustment based on the limited data available involves appreciable errors and may be used as a rough guide only of the likely magnitude of sire components. The crux of this question, however, is not really the magnitude as such of the sire components in the various carcass measurements but rather it involves the problem of correlated responses: in what way does the carcass change as genetic potential in growth changes? Analyses of the covariance between growth rate (expressed as lb of carcass/ day) and several carcass traits resulted in the phenotypic and

534 Factors associated with growth in beef cattle TABLE 7. Phenotypic and genetic correlations between carcass weight gain and carcass traits (r.p.) (r.g.) Weight gain X fat cover Weight gain X eye muscle Weight gain X length side Weight gain X total points 0.14 0.20 0.28 0.03-0.53-0.24 0.00 0.21 genetic correlations shown in Table 7. As is the case with pigs (Smith, King, and Gilbert 1962) there appears to be a negative genetic correlation between fat cover and growth rate and between growth rate and eye muscle area. This observation is interesting in that negative correlations between both fat cover and eye muscle area imply a difference in the conformation and perhaps more bone in the fast growing animals. It should be remembered, however, that the standard errors of these genetic correlations are large; they are likely to be of the order of 0.3-0.4 in these data, so that the suggestion of these negative relations between growth rate, eye muscle area, and fat cover is best treated with reserve at this stage. A comparison of the mean value of the carcass measurements on the progeny of the "fast" sires as compared with the "slow" sires reinforces this point, for the mean eye muscle area of the progeny from the "fast" sires is slightly greater than that from the slow sires in spite of the apparent negative genetic correlation (Table 8). TABLE 8. Carcass a.pp raisal of the steer progeny of 30 sires selected cs either fast or slow gain animals Eye Carcass Fat* Length I. muscle weight cover side area (lb) (mm) (in.) (in.) \ Total points Fast 622 24.74 I 11.22 49.60 54.33 Slow 630 26.47 11.13 49.69 53.33 ---.* Sum of C and D measurements. DISCUSSION Several interesting points arise on the collection of data relating to beef cattle growth under pastoral conditions. The time of weaning, nutritional levels at various stages of growth, and the point at which animals are killed, each provide problems on which diverse opinions are frequently expressed. Two distinct considerations arise in discussing these problems. The first is that the experimental design must be such as to reduce to a minimum complications in the interpretation of the actual results obtained. As a corollary of this it follows that the design must be such as to maximise group differences of interest The second point is that the design must be such as to minimise the dangers of extrapolation

P. J. BRUMBY, D. K. VVALKER, AND R. M. GALLAGHER 535 of conclusions from the experiment to field conditions and to this end it is desirable that the conditions of the experiment parallel industry conditions to the greatest extent possible. Three possible killing points in the life of a group of animals are obvious: at a definite age, at a definite liveweight, or at a constant degree of "finish". The arguments in favour of killing according to an eye appraisal of finish centre about the consequent uniformity of the dressed carcass, particularly in fat cover. This argument may be sound for bullocks which have undergone an appreciable store period, but it appears of little importance in considering young fast growing animals, for data collected by Mason (unpublished) together with further unpublished data from this Station suggest that the variation in fat cover is not influenced to any appreciable degree by the technique adopted in selecting two- to three-year-old animals for slaughter. Killing at a constant finish seems, therefore, to have little to recommend it, for the resultant carcasses vary in age, wcight, and fat cover. Killing at a constant age controls one variable in this triumvirate, while killing at a constant liveweight minimises the carcass weight variable; thus the choice would seem to be between these two, depending upon which one maximises the difference between animals in rate of gain.... s: C' Q) Fig. T 7.-The relation between rate of liveweight gain and criteria of slaughter. Time The broad issues involved in this choice may be illustrated by reference to Fig. 7. Suppose that the two curves represent the average growth curves for two animals. If these two animals are killed at a constant age, i.e. at point T, then the difference in rate of gain between the

536 Factors associated with growth in beef cattle two (Dl) at point W T Now IS 8W T W, the W T -I- 8t Dl D2 If the animals are killed at a constant weight, i.e., difference in rate of gain, i.e., D2, is given by W8t T:J. 8W T T:J W8t 8W 8t- T W T dw dt W dw T d. F Thus D 1 will be greater than D2 when dt Wexcee s umty. or European breeds of cattle under reasonable conditions of management the value of is likely to be of the order of 0.6, thus only in the case of very fast grown animals is Dl likely to exceed D2. In general terms it may be suggested that the optimum slaughter technique is indicated by the relation between the instantaneous growth rate and weight for age. In practice, killing at a constant liveweight is usually indicated. This generalisation concerning the optimum technique for slaughter selection is based on simplified theoretical considerations only, and the conclusion may well require modification in practice. It may be that the onset of winter makes it desirable to kill the animals remaining in experimental groups irrespective of weight; it may be that where the major experimental objective concerns carcass development and conformation, killing at a constant age is necessary fully to analyse a problem; it may be that price fluctuations make it necessary to kill all animals at one time to provide finance for subsequent experiments. Considerations such as these, then, must inevitably modify the choice of the technique of selection for killing. A major factor inhibiting the adequate interpretation of many of the genetical data in this study arises through the confounding of progeny group differences in the pedigree herds with mating and field management practices. For this reason it is not possible to obtain accurate estimates of heritabilities and genetic correlations from these data. The estimate obtained for the heritability of rate of gain, being based on a difference in two weights in the same animal, is probably the most reliable one, for the differences in environment between progeny groups which inflate the between-sire variance are terminated at weaning, after which time all young bulls are run as one mob; hence differences in two weights on all bulls are not likely to be biased to the same degree as absolute liveweights by the environmental differences encountered at earlier stages of growth. The non-random selection of sires for the selection herd similarly increases the apparent between-sire variation in both absolute liveweights and rates of gain. However, the methods of selection used for choosing the selection herd sires are necessarily imprecise, being based on two, three, or four field weighings only and then after the young bulls had been driven appreciable distances to stock yards and held there for some hours. In consequence, it seems likely that the between-sire variance is not seriously inflated by the actual selection of sires, and the apparent heritability estimates obtained from an analysis of half-sib groups in the

P. J. BRUMBY, D. K. WALKER, AND R. M. GALLAGHER 537 selection herd do not greatly overestimate the true values. In this respect it is worth noting that the actual selection experiment is to be modified by closing the selection herds and selecting within them, for the inaccurate selection differentials computed from the limited weighing possible in stud herds, together with the complications arising through possible genetic differences between studs, make the interpretation of the present selection results unnecessarily hazardous. Correlations between various traits were broken down to betweenand within-sire estimates. The former are influenced by the various factors inflating between-sire variances, and, for this reason, only the within-sire estimates have been reported in any detail. Total correlations differ little from these latter estimates. The results obtained in this study are in good general agreement with those reported in other countries. (For an excellent review of the work to 1950 see Mason 1951). Thus it is shown that weaning weight is largely dependent upon the dam's milk supply, that weaning weight and subsequent weights are correlated but that weaning weight has no great influence on post-weaning gains. The low correlation observed between winter and summer gains is of interest and is a point requiring further exploration. The heritability of growth rates and body size appears high and is probably of the order of 0.4. With the possible exception of eye muscle area the correlated responses in carcass quality appear generally favourable. In essence there appear to be two main requirements for successful beef production: a heavy milking dam that produces a heavy weaner, and a sire that produces a fast rate of livcweight gain in the post-weaning phase. As the latter characteristic has a high heritability, improvement within a breed, based upon the performance testing of young bulls, appears relatively straightforward. The improvement of milk production in beef breeds may well be more complicated, but the use of a dairy cross in the production of beef breeding cows provides an obvious starting point. REFERENCES KNEEBONE, H.; MARKS, T..: McMEEKAN, C. P.; WALKER. D. E. 1950: N.Z. ]. Sci. Tech. 31A: 3. MACDONALD, M. A. 1958: "Beef cattle production." p. 5. Massey Agricultural College, Palmerston North. 59 pp. MASON, 1. L. 1951: An:m. Breed. Abstr. 19: 1. SMITH, C.; KING, J. W. B.; GILBERT, N. 1962: Anim. Prod. 4: 128.