Sheep Health & Production

Introduction | How far behind the stud is the commercial producer's flock? | Selection of replacement females in commercial flocks | Choosing replacement rams from a stud for a commercial flock | General references

Return to Sheep Health & Production Index

Introduction

The aim in this chapter is to discuss the type of practical genetics issues that the veterinarian is likely to encounter in commercial sheep production. The design of breeding programs is not one of these and the reader is referred to one of several excellent texts on this subject, including Veterinary Genetics by F Nicholas, for further information.

Genetic structure of the wool-sheep industry

Commercial flocks

The term 'commercial' is used here to describe flocks in which rams are not bred but are purchased from other flocks as required. Commercial flocks include all-wether flocks, in which no rams are used, and breeding flocks where replacement ewes are bred, purchased or a combination of both. The distinguishing feature of a commercial flock is that all lambs are castrated, usually at two to eight weeks of age.

Ram-breeding flocks

In the simplest terms, rams are bred in studs and sold to the owners of commercial flocks. While in reality this is largely true, there are a few practices which require that we broaden the description of the wool-sheep seedstock industry.

Not all suppliers of wool-sheep rams are studs. To qualify as a stud flock requires that the individual stud sheep are registered with a breed society. For Australian Merinos and Poll Merinos, the relevant society is the Australian Association of Stud Merino Breeders Limited. Registered studs produce about 80% of the rams used in the Australian wool industry ­ and the remainder are supplied from non-registered flocks. Of these, one of the largest and most influential is the Australian Merino Society (AMS), which operates a nucleus breeding scheme based on ram breeder co-operatives and is a very significant ram supplier in WA particularly.

In order to cover all ram suppliers in this chapter I will use the term ram-breeding flock (RBF) to include stud and non-registered flocks which supply rams to other breeders.

The price of rams

Ram-breeding flocks supply rams to other RBFs and to commercial growers. The greatest number of rams go to commercial flocks but the best (and most expensive) rams are either retained in the RBF where they were bred or sold to other ram-breeders. The rams that commercial growers buy are called flock rams and they are sold in two or more grades. Top grade flock rams often sell in the range $700 - $1400, while lower grade rams may sell for $300 to $700.

Semen sales

RBFs also sell semen from rams, usually from the higher performing ones. Frozen semen is too expensive for commercial flock owners to use to breed ewes and wethers. The AMS, on the other hand, has a scheme by which high performing rams are used in a number of commercial flocks with AI and fresh semen. The cost of this procedure is sufficiently low to be economical for commercial flock owners.

Home-bred rams

Some commercial flock owners, rather than buying flock rams, breed their own by selecting some of their best ewes and mating them to a high merit ram, either through natural mating or with purchased frozen semen. The rams which sire the flock rams are usually bred in RBFs, rather than by the commercial flock owners.

Table 5.1 : Genetic parameters of some important traits of Merino sheep[1]

Genetic implications of the different flock types

Different genetic principles apply to each of these different types of flocks. First, for RBFs, the principles required to make rapid genetic improvement are well established in quantitative genetic theory. Matters of flock structure, generation length, nucleus flock size, the use of objective measurement, determination of breeding objectives and the use of selection indices all have to be optimised. These principles are well described in a number of other texts and will not be described here.

The key genetic difference between commercial flocks and RBFs

It should be noted, however, that one of the key factors which determine the type of genetic principles applied in RBFs is that rams are bred from both ewes and rams which were themselves sired by selected rams in the previous generation. If it were not the case that the dams of the rams were themselves sired by the previous generation of selected sires, then genetic improvement would cease after a few generations.

The key genetic strategy in commercial flocks using purchased flock rams

For commercial flocks, the relevant genetic principles are quite different from those of RBFs and it these principles which will be discussed here. In short, the key strategy which commercial flock owners who buy flock rams must get correct is to identify and use the RBF which excels in traits which best coincide with their commercial wool-producing goals. Hence, these commercial growers need to know how to identify suppliers of rams of high genetic merit for the traits of relevance to them. Information is now available which facilitates that process of identification, and it will be discussed further below.

The key genetic strategy in commercial flocks using home-bred flock rams

For commercial flock owners who breed their own flock rams from purchased rams or semen, identification of the best individual rams (rather than best RBF) is the key strategy. For reasons which will become clearer below, the purchase of semen and the use of AI holds the most promise for successful identification of top sires.

Genetic decisions are key business decisions for commercial growers

For all commercial wool growers, their success in identifying the best sources of genetic material is of overwhelming importance to the profitability of their wool-sheep enterprise. It is not an overstatement to say that the difference between a good genetic decision and a bad one can be the difference between strong profits or serious financial losses.

In discussing the genetic principles for commercial flock operators, we will first establish some facts about the genetic merit of their flocks.

How far behind the stud is the commercial producer's flock?

A proof that a commercial flock is two generations behind its ram supplier

Consider a commercial flock which is supplied with rams from stud X, and has been for many years. Typically for flock rams, rams are purchased from around the average for phenotypic merit in the stud flock. Stud X is improving in genetic merit at b units per year.

The commercial flock has a stable flock composition, and the same number of new rams are purchased from the stud each year. In any year, the lambs that are born have received half their genetic merit from the rams. We can relate the genetic merit of the commercial flock lambs to the stud flock lambs. If the average age of the rams is 3 years, then the half of the commercial flock lambs' genes coming from the rams is equivalent in genetic merit to the stud's lambs 3 years before.

The commercial flock ewes, however, did not come from the stud, but were bred in the commercial flock from rams bought from the stud several years before. To further complicate the description of the genetic merit of the ewes, remember that their dams were also bred in the flock, from rams purchased from the stud several years earlier again! And so on ....

The easiest way to relate the merit of the ewes in the commercial flock to the stud flock is to consider the genetic merit of subsequent generations of ewes. A generation on the ewes' side (GL_ewes)is defined as the average age of the dams when the progeny are born.

For the commercial flock ewes, half of their genes came from the rams in use in the commercial flock one generation ago. The other half came from their dams, for whom half the genes came from the sires in use 2 generations ago. The other quarter came from their grand-dams, half of whose genes came from the sires in use 3 generations ago. In summary, the genes in the present ewe flock came from

where Gen1 to Gen5 represents the sires in use 1, 2, 3, 4, 5 generations ago.

If the difference in genetic merit in each generation is the same (assumes a constant rate of genetic improvement in the stud flock), we can sum the sources of genetic merit for the ewes. Let the merit in the sires in use one generation ago be M and the improvement in each generation be a, then, the genetic merit of the ewes is

which approaches an upper limit of M - 2a, which is the merit of the sires three generations ago.

The sires in use one generation ago had a mean genetic merit of M, so the merit of the progeny born to the current generation of sires and dams is the average of their parents, ie

(M + M - 2a)/2, which equals M - a.

The generation length on the rams' side may be different from that on the ewes' side, so it must be calculated separately. Remember that b is the improvement in the stud per year. The average merit of the commercial ram flock, therefore, must be GL_ramsH b behind the merit of the lambs born in the stud in the same year, where GL_rams is the average age of the rams when the lambs are born. If the merit of the stud lambs this year is S, then

M = S - GL_rams H b

We can also describe a in terms of b and the generation length of the ewes, GL_ewes.

a = GL_ewes H b

So, the merit of the progeny, which eas shown to be M ­ a,

= S - GL_rams H b - GL_ewesH b

= S - b H (GL_ewes + GL_rams)

As GL_rams + GL_ewes is twice the generation length of the flock,

the merit of the progeny in the commercial flock equals S - 2 H b H GL_flock

Thus, the progeny of the commercial flock lags two generations behind those of the stud flock.

An example to illustrate the differences in genetic merit

Assume rams are purchased from a stud which is making genetic improvement at the rate of 1 unit per year. The two year old rams in the commercial flock have a merit of 105 units, but the stud has continued to make progress such that the genetic merit in the current year's lambs in the stud is 107 (Table 5.2).

Table 5.2 The difference in genetic merit of ewes and rams in a commercial flock, based on an arbitrary value of 100 units for the current years' lambs and assuming the stud selling rams to the commercial flock is improving at 1 unit per year.

From Table 5.2:

The average age of rams when lambs are born = 3 years

The average age of ewes when lambs are = 4 years

The generation length in flock = 3.5 years

The average genetic merit of the progeny is the average of the parents (104 and 96), which is 100. The genetic merit of the lambs in the commercial flock is 7 units, or the result of two generation lengths of improvement, behind the stud.

Extending the generation length of the flock, by keeping ewes or rams longer, increases the genetic lag between the commercial flock and the stud. If attempts are made to shorten the GL, it does not matter whether the ewes or the rams are sold at younger ages, the effects are the same.

It is not necessarily economic to cull rams at younger ages to move the flock closer to the stud. For example, in the flock above, selling rams after two joinings would shorten the GL by 0.25 years. The merit of the commercial flock would rise, gradually, by 0.25 units. If one unit is worth 1% of clean fleece weight per head, the improvement in productivity from the ewe flock would be, over 2,000 ewes, about $125 per year. A similar result would be obtained from the wether flock, if there is one.

The cost of doing so would be the cost of buying 15 rams per year, rather than 10. If rams cost $400 each, the extra cost is $2,000 per year.

Note also, a corollary of this is that the ewes in a commercial flock are always two ewe generations behind the rams of the same age, in genetic merit.

Selection of replacement females in commercial flocks

In commercial flocks where breeding is practised at least partly to maintain the female breeding flock, young ewes are selected to replace aged ewes which are culled and ewes which have died. If there are more young ewes available for selection than are required to maintain the breeding flock, then selection can be practised on the young ewe flock to ensure that the best ewes are retained and the worst ewes are culled, or at least, not used for breeding.

The act of selection is practised following either a visual appraisal, an objective evaluation of one or more characteristics of each young ewe or a combination of both. In commercial sheep flocks the selection may be as simple as choosing the biggest, or as complicated as weighing all or most fleeces and measuring the average fibre diameter of a sample of the fleece. It makes sense to retain the most productive sheep, particularly if the traits under selection have a genetic basis, because these sheep will probably be more productive for the rest of their lives in the flock than the culled ones, and their progeny can be expected to inherit half of the genetic superiority of their dams.

The genetic aspects of this process are frequently used to justify selection activities which are quite expensive. While most producers acknowledge that the sires contribute the greatest amount to genetic improvement, this is often based on the assumption that selection intensities for rams are higher than can be obtained with normal reproductive rates in commercial flocks. Unfortunately this is not the only limitation of selection of females within commercial flocks and it is important to examine all of the limitations to ensure that excessive value is not placed on the practice and time and money wasted.

In this section we will examine the genetic and financial implications of ewe selection in commercial flocks. The conclusions are, of course, applicable also to commercial cattle herds or any other animal population where males are introduced from an outside source.

What is the response to selection of young ewes in commercial flocks?

Current generation gains

In self-replacing sheep flocks, selection is usually practised when the sheep are aged 1 to 1.5 years (hogget age), and often at the time of their first adult shearing. Typically, 60% to 90% of the young ewes are retained, the exact proportion depends on the flock manager's strategies for culling cast-for-age ewes, the reproductive rate in the flock and the survival rate of young sheep from weaning to adulthood. If we assume that selection is applied in some structured way, and is not random, the selected (and retained) ewes will be, on average, superior to the unselected flock, ie, all the ewe hoggets before selection.

Generally, the selected ewes will be shorn several more times before being cast for age. If the characteristic under selection is a wool trait (such as wool weight or fibre diameter) and, if that trait has a usefully high repeatability, at least some of the superiority of the selected group will be demonstrated at each subsequent shearing. Unless the repeatability is 1, however, not all of the superiority demonstrated at the hogget shearing will be realised at subsequent shearings. Generally, repeatability lies between the values for heritability of the trait, and 1 (Falconer and Mackay 1996).

For example, consider a flock of hogget ewes which is fleece weighed at shearing. Based on their greasy fleece weights, 30% are culled and sold off-shears. The mean greasy fleece weight of the retained portion is likely to be 0.25 kg heavier than the mean greasy fleece weight of the entire ewe hogget flock. This increase in the mean value of the selected portion over the mean value of the entire group measured is called the selection differential (s). Assuming the repeatability of greasy fleece weight is 0.6, we expect that, at future shearings, this group of selected ewes will produce 0.15 kg more wool (0.6 x 0.25) than would have been the case had the 30% been culled at random with respect to fleece weight.

So, for one round of selection, the flock owner is rewarded with extra wool production at each of the subsequent shearing events in the life of the selected ewes. The increase in wool production in the lifetime of the selected ewes is called the current generation gain, and equals the product of the repeatability of the trait (r) and the selection differential (s). If we call the current generation gain the phenotypic response (R_p) to selection, then

This represents the reward to the producer of practising efficient selection of replacement ewes. The selection process itself has a cost but, if it accurately identifies the most productive sheep in the flock, the value of R_p in years 1, 2, 3 and 4 could exceed the cost of selection in year 0.

Current generation gains also have value if practised on wethers. For reasons which will become clearer below, this is more likely to be profitable if a significant percentage of the young wethers are culled from the flock after the hogget shearing, and if wethers are retained for several years beyond hogget age.

Genetic gains

For many traits of interest to producers, the superiority of the selected sheep has a genetic component, half of which they will pass on to their progeny. If we call the genetic component of the superior merit of the selected ewes the genetic response to selection (R_g), we know that it equals the product of the heritability of the trait (h²) and the selection differential (s). Thus

and half of R_g will be passed onto the progeny. The merit of the next generation will therefore be 0.5 x R_g higher than it would have been had the dams not been selected. This additional merit is passed onto the progeny regardless of the contribution of the ram.

To return to the fleece weighing example, assume the progeny are subject to fleece weighing and selection at their hogget shearing, like their dams were. The result of the two generations of selection is additive, and results in selected ewes with merit of R_p + 0.5 R_g compared to the baseline merit of the first and unselected generation. The genetic component of R_p is, again, R_g, half of which is passed onto their progeny, which therefore have a merit of 0.5 x (R_g + 0.5 R_g), or 0.75 x R_g.

With each generation of breeding and selection of the ewes, the contribution to merit made by selection of the previous generation of ewes is halved. Gradually, over a number of generations and assuming selection is continuously practised, the genetic merit of the flock approaches R_g. It does this by getting half way there in the first generation, three quarters there in the next, seven eighths in the third, fifteen sixteenths in the next, and so on.

Eventually the merit of the flock will be very close to R_g. On top of this, selection of the ewe hoggets will continue to give current generation gains. The maximum value for the merit of the breeding flock after several generations of continuous selection is R_p + R_g. The maximum value for the unselected sheep (lambs, hoggets, wethers if not selected themselves) is R_g.

Unlike genetic selection in ram breeding flocks, this form of selection has a ceiling and the merit of the flock is only sustained at the high level if selection is continued. Should it be stopped, the accumulated gains will slip away over several generations.

Why is it so?

Conceptually, it seems unreasonable that genetic selection applied to ewes results in limited and relatively small improvements in productivity. For some of us at least, the reason is difficult to visualise, particularly when we have grasped an understanding of the contribution that ram and ewe selection can make within ram-breeding flocks. What is the difference here?

The difference is that the selected ewes are not the dams of the next generation of rams. Thus, their 'good' genes are passed on to their progeny after 'dilution' from unrelated rams, and the process of dilution continues with each generation. Effectively, the ewe selection process takes one step forward, the mating to unrelated rams takes us half a step back. The process comes into equilibrium when the accumulated gains made by ewe selection equals the genetic response to current generation selection. At this point, the descendants of selected ewes have accumulated one 'step', the selection round takes another 'step', then the mating halves the two 'steps' back to one, which is passed on to the next generation.

In ram breeding flocks, by contrast, selection of ewes results in genetic merit being passed on to their sons, which are the sires of the next generation. This is the key difference between ram breeding flocks and commercial flocks.

How great will the gains be?

If the selection proceeds for several generations, and the selection intensity applied each year is the same, the genetic component gradually approaches h²s. If we assume the heritability of GFW is 0.35, the maximum value for the genetic response is 0.0875 kg. Thus, we can predict the maximum responses which will be achieved in a flock following the procedures of this example, after practising the methods for several generations (Table 5.3).

Table 5.3 The increase in greasy fleece weight in a commercial flock as a consequence of long-term and continuous selection of ewe hoggets. (3% annual mortality assumed)

In wool producing flocks, the genetic merit is also expressed in wethers and, if they are selected and culled similarly, they can also demonstrate current generation gains on top of genetic changes inherited from their dams.

The discussion to date has focussed on responses to selection of ewes and has ignored the effect that the rams are having in the flock. This is valid because the response to selection of females in the commercial flock is independent of the effects due to the rams. Rates of genetic improvement in ram-breeding flocks for traits such as GFW could be 1% per annum or more, if sound genetic principles are practised (Atkins 1997), These rates of gain are passed on to client commercial flocks which can, therefore, make gains due to genetic improvement in the ram breeding flock at the same rate; say 0.04 kg GFW per year. The gain from ewe selection reported in Table 5.3, therefore, of 0.24 kg of GFW is not insignificant - it is equivalent to six years of improvement from the ram source (Figure 5.1). It has been pointed out previously (McGuirk 1976) that selection of ewes in commercial flocks can make gains in the first five years that it is practised which compare favourably to gains passed on to the commercial flock by genetic improvement programs in the stud supplying rams to the commercial flock. Unlike the improvement from the ram source, however, it does not continue beyond the level shown and, as soon as ewe selection ceases, the advantage in productivity starts to decline back towards the level of the unselected flock.

Figure 5.1 Comparison of selection of ewes in a commercial flock to the effects of improvement (at 1% per year) in the ram breeding flock supplying rams to the commercial flock

The economic consequences of selecting ewe hoggets on fibre diameter.

The steady state - where selection has been practised for many years.

The economic benefit of selecting ewe hoggets depends on a number of factors, the most important amongst them being the cost and accuracy of the test, the selection intensity which can be applied, the number of animals in the flock benefiting from the genetic effects of selection and the value of the trait under selection. In recent years premiums for fibre fineness have increased markedly and the cost of fibre diameter testing has fallen, presenting the possibility of cost-effective selection of ewe hoggets on fibre fineness in commercial flocks.

In a recently reported study, a computer model of a sheep flock was used to estimate the economic consequences of ewe hogget selection for fibre fineness for a range of flock characteristics. The principal variables examined were weaning rate, which determined the selection intensity which could be applied, and the age at which wethers were culled, which determined the number of progeny of selected ewes which expressed the genetic superiority per ewe hogget subject to measurement and selection.

The results are presented as breakeven costs for fibre diameter measurement when the relationship between fibre diameter and price is -$1 per micrometer (Table 5.4). With the assumption of a linear relationship between diameter and price, the breakeven costs of testing are directly proportional to the value of a one micrometer change in fibre diameter. For example, if a one micrometer decrease in fibre diameter increases wool value by $1 per kg then the breakeven cost of ewe hogget selection in a flock with a high weaning rate (90%) and with wethers retained to 4.5 years is $14.04. If the premium fell to $0.50 per kg, the breakeven cost would be $7.02.

The results are also directly proportional to the within-flock standard deviation of fibre diameter. In the results shown below, the standard deviation used was 1.73μm (Atkins 1989). Other assumptions of the modelled flock are given in the original publication (Abbott 2001).

Genetic and phenotypic correlations between clean fleece weight and fibre diameter are built into the model. The predicted effect of selecting ewe hoggets for fibre fineness is to reduce clean fleece weight. The predictions in Table 5.4 are based on a value for wool of $10 per kg clean, so that the $1 per micrometer change in value represents a 10% micron premium. If the effect on clean fleece weight were ignored, the breakeven costs would increase by 15% to 16%. At lower micron premiums (say 5%) the importance of the correlated response in clean fleece weight becomes relatively greater.

Table 5.4 The breakeven cost for fibre diameter testing of ewe hoggets in a commercial flock (the cost below which testing is profitable) per $1 of fineness premium

Table 5.4 demonstrates that at low weaning rate the economic benefits of selection are small - simply because so few ewe hoggets can be culled that there is little change in flock mean fibre diameter. At high weaning rates, such as 100% when about 32% of ewe hoggets can be culled, the breakeven costs are $12 to $16.70 per hogget, depending on how many wethers are retained in the flock. For example, if testing costs $3 per hogget, including all measurement, identification and labour costs, there will be a profit of $9 to $13.70 per hogget measured. Thus, in a flock with 1000 ewe hoggets and weaning rates of 100%, fibre diameter testing of all ewe hoggets and culling on the results will lead to profits of $9 000 to $13 700 per year, depending on the age at which wethers are culled from the flock. Note, this result is dependent also on the micron premium both as a percentage of the wool value, and in absolute terms.

The reason for the increased benefit of selecting ewe hoggets as weaning rate increases is the additional selection intensity which can be applied. In the modelled flock, 10% of the ewe hoggets are culled on traits unrelated to fibre diameter. With a weaning rate of 70%, only 2.5% of the remaining hoggets can be culled, giving a selection intensity (i) of 0.06. With weaning rates of 80%, 90% and 100%, culling rates after the initial 10% are 15%, 24% and 32% and the resultant selection intensities 0.27, 0.41 and 0.52 respectively.

With selection intensities of around 0.3 to 0.5 (say 0.4) and an assumed standard deviation of fibre diameter of 1.73, the predicted R_g and R_p values are 0.35 µm and 0.48 µm respectively. The adult breeding ewes benefit from both R_g and R_p, while the young sheep and wethers benefit from only R_g. With an average effect across the flock of about 0.5 µm, how can the return be so great? The answer is that there is about 30 kg (clean) of wool produced for each hogget measured and selected. A 0.5 nm reduction (at $1 per µm) over 30 kg is worth about $15, as predicted in Table 5.4 for the higher weaning rate flocks.

The reason for the increased benefit of selecting ewe hoggets as the proportion of wethers in the flock increases is that more wool is produced from the flock for each ewe hogget selected. The ratio changes from about 22 kg per ewe hogget when wethers are all sold at 1.5 years, to about 35 kg per ewe hogget when wethers are retained to 4.5 years. Although the wethers receive only R_g, the fact that the benefit of each act of selection is applied to so much more wool without any additional cost outweighs the fact that ewes (with both R_g and R_p) form a smaller part of the flock.

Starting out - the delay in reaping financial rewards from selection of ewes

The benefits of selection of ewe hoggets seems clear from Table 5.4. Given moderate to high reproductive rates and a high premium for fibre fineness, there appear to be profits to be made from selection on fibre diameter provided the cost per ewe hogget selected is less than, say, $5 per hogget. Should producers, therefore, be encouraged to test every ewe hogget for fibre diameter and base culling decisions on the results?

The answer is "possibly"! Caution should be expressed because there will be several years' delay before reaping the financial reward promised in Table 5.4, but the cost of testing has to be paid every year, from year 0.

Table 5.5 Cash flows associated with ewe hogget selection for fibre diameter where weaning rate is 90% and wethers are retained to 3.5 years. Fibre diameter testing is assumed to cost $3 per sheep tested. Other assumptions as for Table 5.4. Costs and incomes per 1000 ewe hoggets.

Table 5.5 shows that R_p reaches its maximum value once the breeding flock is composed entirely of ewes which were selected on the basis of hogget fibre diameter, but R_g is continuing to increase usefully even nine years after selection commenced and, at 10 years, is about 90% of its maximum value. Assuming a cost of measurement of $3 per hogget, the procedure is breakeven after the wool is sold in year 4.

For each individual producer contemplating ewe hogget selection a table like Table 5.5 should be constructed, with values specific to each particular flock and the predicted premiums for fibre fineness in the future.

Choosing replacement rams from a stud for a commercial flock

Traditionally, there has been very little useful comparative information provided by studs, so that commercial producers have found it very difficult to decide objectively which is the best source of rams for them. In many cases, this choice has been based on family allegiances.

Fortunately, this situation is gradually changing.

Merino bloodline performance

First, there has been an increase in the number and size of wether trials. These trials involve the running of groups of wethers from different studs as a single flock. Some trials involve so few animals as to be practically worthless. However, many trials are run on a sufficient scale to provide extremely valuable comparative information. Dr Kevin Atkins, a sheep geneticist with NSW Agriculture, has recently analysed 48 state-wide trials conducted between 1981 and 1991, involving a total of 895 teams representing 133 Merino bloodlines (studs). The value in Dr Atkins' analysis is that he actually publishes the comparative results for those studs with sufficient data to make their results meaningful. His minimum criteria for publication is that the standard error (the standard deviation divided by the square root of the number of observations) for CFW must be less than 3% of the mean. An Agfact (A3.3.33) reports the results from the 53 studs which met this criterion (see www.agric.nsw.gov.au/Sheep/ and go to Merino breeding).

Many different traits are measured in these trials, including CFW, fibre diameter, body weight, and a series of subjectively-assessed wool-quality traits: style (coded 1 to 5), staple length (coded 1 to 3), colour (coded 1 to 3), and tenderness (coded 1 to 4). The average performance for each trait in each stud is published, and is made freely available to commercial producers. The analysis shows that there are quite substantial differences between studs in various traits, which means that it is well worthwhile for producers to use this information to seek and use the best ram supplier for their individual objectives. They still have the difficult task of deciding which combination of traits is best for their conditions, but at least they now have very useful comparative information available to help them make a sensible decision.

Central test sire evaluation (CTSE)

A second reason for the increased availability of comparative information available to producers is the increasing popularity of sire evaluation (SE) schemes, which involve the comparison of individual rams, on the basis of their progeny's performance. In other words, SE involves progeny testing of individual rams. Some progeny tests are conducted on individual studs (on-farm SE), while others are conducted at a central site usually run by a government department or a university (central-test SE). Increasingly, different on-farm SE schemes are becoming 'linked' together, and are also becoming linked with central-test SE schemes. These linkages are created by the use of AI of a small number of 'reference' or 'link' sires in all participating flocks. The reference sires are usually of exceptional merit.

The big advantage of linkages between different SE flocks is that once a linkage is established, it is then possible to compare all the sires used in all linked flocks. The results of these comparisons are then published in a standard format, usually in terms of estimated breeding values (EBVs). Both stud and commercial breeders can then evaluate the relative merits of different rams, and make an informed decision about where they should buy their replacement rams.

The influence of both wether trials and SE schemes will undoubtedly increase in the future, and it is important that you are aware of both activities. It is also important that you should be aware of their limitations. One of the major limitations at present is that not all traits that contribute to profitability are measured in all trials and schemes; and even if most of the important traits are measured, there is no simple practical way by which breeders can combine information on different traits into a single measure of overall profitability. (There is no theoretical hindrance to producing a single measure of overall profitability; it is just that in the Australian wool-sheep industry, the necessary facilities have not yet been set up for doing this. However, we are someway along this path already, and further progress will be made in the next few years.)

What is an EBV?

EBVs are relative things

An estimated breeding value (EBV) is just that ­ an estimate of the amount of genetic merit that an animal is likely to pass on to its offspring. EBVs are always relative to some base level. In its simplest form, an EBV is quoted for an animal based on a comparison to its peers ­ those animals born in the same flock at roughly the same time and exposed to the same (as much as possible) environment from birth to the age at which measurement is made (typically 1 to 1½ years for sheep).

If flocks, or years within flocks, can be genetically linked, the average merit of the flocks can also be compared so that two or more animals from different flocks and/or born in different years can be compared ­ effectively by comparing each animal to the average of its own flock and then comparing the averages of the two flocks. This will be discussed more, later.

Calculating a simple EBV for a ram

To resume discussion of the simple case, consider a RBF in which 1000 ram lambs are born and raised to 1½ years. At shearing at that age their fleeces are weighed and, for each animal, a greasy fleece weight (GFW) is recorded. The distribution of GFW for the 1000 animals forms a bell curve ­ representing a normal distribution. The mean value is 4 kg, the standard deviation (SD) is 0.6 kg. The best ram has a fleece weight of 5.8 kg, three standard deviations above average. If this animal is selected by the breeder to be a sire of the next generation, the selection differential (s) for that animal is 1.8 kg (Figure 5.2).

In order to predict the effect of using this ram as a sire, rather than selecting a ram at random with respect to GFW, consider a ram which was average for GFW. For him, s = 0. What is the effect, in the next generation, of using the high fleece weight ram rather than the ram which is average for GFW? We expect his progeny to be better but how much?

First, we know thathe will pass on only ½ of his genes to his offspring, the other ½ coming from the dam.

Second, he will only pass on the genetic component of his superiority. His phenotypic superiority was +1.8 kg but published heritability estimates tell us that only 35% of that superiority is likely to be genetic. Thus, we predict his breeding value = 0.35 x 1.8 = 0.63 kg. This is his EBV.

His progeny we expect to receive on average half that amount and therefore to have fleeces 0.315 kg heavier than those sired by the average ram.

The accuracy of an EBV

How likely is it that any one progeny of the superior ram will be as good as predicted? There is certainly no guarantee, and the uncertainty arises from two sources.

First, all of the ram's progeny get half of the sire's genes but not every progeny will get exactly half of the good genes ­ those genes that contribute to the sires phenotypic superiority. Some will get more than half of the good ones, some less. On average, his progeny will get half of his genes or, stated another way, the average merit of a large group of his progeny (large enough to avoid chance effects) will reflect half the merit of their sire. While this is satisfactory in breeding programs where selection is practised again on the next generation, it is not sufficiently reliable to assume that every son or daughter of a pair of champions is a champion!

Second, the original estimate of the ram's breeding value may have been inaccurate. The explanation for this is that, for that one individual, the estimate that 35% of his phenotypic superiority was genetic may have been incorrect. The figure of 35% for h² is calculated from experimental observations on large numbers of animals but it is only a 'best guess' for any one individual. That ram may have been blessed by a particularly outstanding environment as a young sheep, leading to a phenotypic superiority which was all due to his good luck! It may be that his true breeding value is close to zero.

Because it is known that there is uncertainty associated with the estimate of an individual's breeding value, the accuracy of the estimate is also often published. This allows the calculation of confidence limits around the prediction of the progeny's merit. Accuracies are reported on a 0 ­ 1 scale and are effectively the correlation between the animal's phenotype and its true, but unknown, breeding value. In the simple, one-trait case we have discussed here, the accuracy is h, the square root of the heritability - 0.6 or 60% in the case of GFW.

Making an index of one or more traits using EBVs

In breeding programs, owners of RBFs generally want to improve a number of traits simultaneously. For example, it is common in wool-sheep breeding programs to wish to increase CFW and simultaneously reduce FD. So, a ram's EBV for both traits are of interest. How then should they be combined? The answer is relatively simple ­ each EBV should be weighted according to its economic value and then the two weighted EBVs, now in units of dollars (or whatever currency was used), are added together. For example, if a one kg change in fleece weight will increase an animal's fleece value by $10 and a one µm fall in FD will increase it by $4, arational index is I = 10.EBV_CFW + 4.EBV_FD.

EPVs and EPDs

Note that EBVs are halved before predicting the effect of a parent on his or her progeny. In fact, if the dam's EBV is known the predicted merit of the progeny is the average of the EBVs of both parents. Some people, finding it tedious to halve an EBV, have turned to estimated progeny values (EPVs) or expected progeny differences (EPDs) which have half the value of an EBV, relieving the bother of halving the sire's EBV when calculating the expected progeny merit.

General references

Agfacts

A3.3.1. Fleece measurement services for NSW sheep breeders

A3.3.2. Selecting rams and ewes for wool production

A3.3.7. Sheep breeding: systems of sheep selection

A3.3.8.Sheep breeding: inbreeding, line breeding and crossbreeding

A3.3.33. Merino bloodlines: the differences

Proceedings of the Australian Association of Animal Breeding and Genetics. Seventh Conference (1988); Eighth Conference (1990); Ninth Conference (1991), Tenth Conference (1992).

Abbott KA (1994) Cost-benefit evaluation of artificial insemination for genetic improvement of wool-producing sheep Aust vet J 71 p 353

Atkins KD, Semple S, Casey A and Hygate L (1992) Variation in production traits between Merino bloodlines -NSW wether comparisons 1981-1991 Report 1, Animal Industries Report, Agricultural Research and Veterinary Centre, NSW Agriculture, Orange.

Atkins KD (1993) Benefits of genetic improvement to the Merino wool industry Wool Technology and Sheep Breeding 41 p 257

Brash LD (1995) Advanced breeding techniques for wool sheep improvement Wool Technology and Sheep Breeding 42 p 327

Brash LD, Wray NR and Goddard ME (1995) Use of MOET in Merino breeding programs: a practical and economic appraisal Animal Production (in press)

Brien FD (1993) Beyond Woolplan Wool Technology and Sheep Breeding 41 p 281

Brien FD, Kearins RD and Maxwell WMC (eds) (1993) Merino Sire Evaluation in Australia: Recommendations on Sire Evaluation (Sire Referencing) for the Australian Merino Sheep Industry Woolplan Management Committee, Wool Research and Development Corporation, Melbourne.

Hammond K, Graser H-U and McDonald CA (eds) (1992) Animal Breeding: the Modern Approach Postgraduate Foundation in Veterinary Science, Sydney.

Kearins RD (ed) (1988) Objective Measurement and Sheep Breeding (Proceedings of the Sheep and Wool Seminar and Refresher Course, Orange, April, 1988). Sheep and Wool Section, NSW Department of Agriculture, Orange.

Maxwell WMC and Brien FD (eds) (1988) A Ram Breeder's Guide to Woolplan Information Services Unit, SA Department of Agriculture, Adelaide.

McGuirk BJ (1987) Merino Improvement Programs in Australia Australian Wool Corporation, Melbourne.

Nicholas FW (1987) Veterinary Genetics Clarendon Press, Oxford.

Figure 5.2 A normal distribution. The horizontal axis contains the values for a trait, such as fleece weight. The vertical axis contains values for the numbers of animals with each value for the trait. The animals with the highest value for the trait are at the right hand end of the distribution, in the red-coloured area P which refers to the proportion selected. The mean value (eg fleece weight) of the selected proportion is s, the selection differential and represents the average superiority of the selected group over the entire unselected population. If s is divided by the standard deviation of the trait we get i, in units of standard deviations. We also refer to i as the selection intensity, and i can be predicted from tables for a given value of P. The cut-off point for the selected population is x, which can also be predicted from tables.

Table 5.6 Truncated normal distribution. P = percentage of the population selected; x = the closest value to the mean for the selected percentage; i = the mean value for the selected percentage. Values for x and i assume a standard deviation of 1.

Values for x and i can be found in tables or, in Microsoft Excel ®, the following expressions can be used : x = -1*NORMSINV(P) and i = EXP(-0.5*x^2)/2.5066283/P, where P has values between 0.00001 and 0.99999.

[1] Atkins KD 1987