Published in Holstein Journal
October, 1990

The Science Behind the Art of Mating Cows

Jack C.M. Dekkers
Centre for Genetic Improvement of Livestock
Animal & Poultry Science, University of Guelph

In the field of dairy cattle breeding, the expression, "the art of breeding cows", is often heard. Certainly breeding cows does require talent, and some may have more talent than others. However, today's budding painters attend art school to learn the methods and essentials of the art of painting. In the same way and regardless of talent, a farmer can benefit from an understanding of the scientifically proven principles underlying the art of mating cows. The purpose of this article is to outline some of these principles.

Breeding Goals Are Fundamental to Success

Every person involved in dairy cattle breeding is aware of the difficulties involved with mating cows. Decisions that need to be made involve both selection of sires and arranging matings between selected sires and cows. No realistic decisions can be made without having a clear-cut objective or goal. Breeding goals will be the first topic of this article.

The purpose of all breeding decisions is to increase profitability of the herd. When contemplating which sires to breed your current cows and heifers to, the immediate concern is the group of heifer calves you expect to result from these matings. After all, they will enter the milking line-up three years from now. Thus, breeding decisions should ensure that future returns from this group of heifers are as high as possible.

A measure of future profit expected from a heifer calf is called her total genetic merit. Total genetic merit is the combined effect of several traits of economic importance, such as milk, fat and protein yield, type, calving ease, etc. An animal's total genetic merit is calculated as a weighted sum of its genetic indexes, or proofs in the case of sires (eg. 2xFAT + 2xPROTEIN + 1xTYPE). Putting proper weights on traits is a complicated matter, on which research is currently under way at the University of Guelph. What can be said at this point is that, although weights may differ from herd to herd, primary emphasis should be on the yield traits. This applies to commercial herds, as well as breeders who aim to provide stock for the commercial dairy industry.

Sire Selection Is The Most Important Breeding Decision

Now that we have defined our objective (i.e. maximizing average total genetic merit of the progeny group), we can start talking about how this goal can be achieved. There are two tools that we can use for this: selection and mating. Selection mainly relates to which sires to use in your herd (i.e. what semen should I buy?), while mating deals with which of these sires to mate to which cow (or heifer). Sire selection is certainly the most important of these two tools. Sires should be selected by ranking them on their total genetic merit index, and using the ones at the top. Proper and intense sire selection will ensure use of the best genetic material and a high total genetic merit of the progeny group. Current efforts by several A.I. studs to pool semen from top sires, such that it is available to farmers across Canada at realistic prices, are very helpful in selecting the best sires. Ideally, only the top 20 to 25 bulls in Canada on a total genetic merit index should be used to breed cows.

How to Mate Your Cows

After the best sires have been selected, the question is how to mate these sires to your cows. Basically, three types of mating schemes can be distinguished: random mating, positive assortative mating, and disassortative mating. With random mating, the selected sires are assigned to cows without any planning. Mating sires and cows that are similar in total genetic merit, or individual traits, is called positive assortative mating. This amounts to mating the best to the best, and the lowest sires to the poorest cows. Disassortative mating is better known as corrective mating. This type of mating relates mainly to the type traits, when, for example, a cow with curved legs is mated to a bull that tends to sire straight legs. However, it may also be practiced for production, when we allocate a bull that is particularly high in protein yield to cows that are relatively low in protein production.

The question is whether a mating strategy other than random mating improves the average total genetic merit of our group of heifers, over and above what we already have achieved by using the best sires in the first place. The answer to this is "yes potentially", particularly if there is: a) dominance genetic variation; b) danger of inbreeding; c) potential for improving uniformity of the resulting heifers; or d) non-linear economic values for some traits. I will return with an explanation of each of these conditions shortly.

The important result to consider at this point is that if none of these four conditions exists, mating the selected sires randomly to your cows will do as good as any type of mating scheme. This is illustrated in Table 1, where five sires are mated to five cows. I am using fat yield here as an example, but it would hold for any other trait, as well as for total genetic merit. An estimate of the genetic value of the heifer resulting from mating Sire A to Cow I is equal to the sire's proof plus the cow's genetic index (+19). This number indicates the returns we can expect from fat production by a daughter of Sire A and Cow I. Similarly, the genetic value of daughters from the other four matings can be predicted. As explained above, our goal is to make the average total merit of the progeny group (or in this case their fat production) as high as possible. For the matings in Table 1, which are more or less at random, the average genetic value of the progeny will be +19.4. In fact, this will be the average no matter how you arrange the matings (give it a try). Therefore, in this case, random mating is as good as any kind of mating.

TABLE 1. Expected progeny from random mating
of 5 sires to 5 cows for fat yield
Sire Sire's
Cow Cow's
genetic value
A + 13 I +6 +19
B +16 II +8 +24
C +18 III +5 +23
D +12 IV +3 +15
E +15 V +1 +16
Average +19.4

However, as mentioned before, there are cases where that wouldn't be true. Let's go through those one by one.

a) Dominance genetic variation. If there is dominance genetic variation (over and above the additive genetic variation), the expected genetic value of progeny is not equal to the sire's proof plus the cow's index, but may be lower or higher. This phenomena is better known as "nicking", and was discussed thoroughly by Tempelman and Burnside in an earlier article in this series (August 1989, Holstein Journal). They concluded that dominance genetic effects are important and that they should be used in assigning matings, if estimates of dominance genetic effects are available. The latter is the problem. Currently we don't know beforehand whether Sire A will nick better with Cow I or Cow II, or better with daughters of Sire X or Sire Y. Therefore, the best current estimate of progeny genetic value is the sire's proof plus the cow's index. We'll get to know more about nicking in the next few years as full sister families get larger through ET and other reproductive technologies.

b) Inbreeding. With use of a relatively small number of sires in the population, the danger of inbreeding increases. Inbreeding leads to inbreeding depression: a decrease in production, reproduction and overall fitness. Therefore, reducing inbreeding by avoiding mating of relatives will be beneficial. Animals can, of course, be related to different degrees. Research by Miglior and Burnside at the Centre for Genetic Improvement of Livestock has shown that we shouldn't worry too much about inbreeding below 10%. As a reference, mating a sire to his half sister results in 12.5% inbreeding. Therefore, mate Sire A in Table 1 to Cow II instead of Cow I, if Cow I is his half sister.

c) Improving uniformity. It is often claimed that corrective mating is beneficial because it improves uniformity of the group of heifers. Although the real economic benefit of uniformity can be questioned, it is undoubtedly one of the goals of many breeders. Uniformity can be measured by the degree of variation in performance among the resulting heifers. Table 2 shows the amount by which this variation can be reduced by corrective mating. Only under the most ideal circumstances for corrective mating will the reduction in variation be more than 8%. Under practical circumstances, the reduction will not be larger than 5%. This increase in uniformity will, unfortunately, hardly be noticeable in the barn. Therefore, uniformity is not a goal that can be attained by corrective mating. Breeders may also ask themselves if unformity is the wisest course to steer. For example, it might be more advantageous from a business angle to have the top ranking cows for genetic indexes for milk, fat and protein, than a uniform group of cows.

TABLE 2. Percent reduction in phenotypic variance
among progeny due to corrective mating
corrective mating
50% Effective
corrective mating
Heritability .2 .4 .2 .4
No Selection 2% 8% 1% 4%
50% Selected 1% 3% .5% 1.5%

d) Non-linear economic values. For most traits, especially the production traits, economic values are said to be linear.What is meant by this is that every point difference in genetic index represents the same difference in expected returns. Or, in practical terms, the difference in economic benefit between Cows I and II in Table 1 is expected to be the same as that between Cows III and IV. Both pairs have a difference in cow index for fat yield of 2 BCA points, which will translate into a differences in fat production of about 8 kg.

However, not all traits behave this way, especially some of the type traits. A good example is set of rear legs. This trait has a so-called intermediate optimum, which means that genetic indexes at either extreme (curved and straight) are undesirable, while a value in the middle is desirable and has the highest economic value. This trait is, therefore, said to have a non-linear economic value. The difference between linear and non-linear values is further illustrated in Figure 1.

Now for a trait like set of rear legs, you want to assign the matings in such a way that the progeny have genetic indexes that represent the highest economic merit. This means that their genetic indexes have to be intermediate. A simplified example of the way in which we can assign such matings is in Table 3, where we want to mate Sires A and B each to one cow. It should be clear from the economic merit column in Table 3 that Sire A should be mated to Cow II and Sire B to Cow I. Those matings result in an average economic merit of progeny of +19, instead of +2.5 for the alternative matings.

However, two cautions should be placed by this procedure of assigning matings for traits with non-linear economic values. The first is that the cow's genetic index, instead of her phenotype, should be used when deciding what kind of sire a particular cow needs. A cow with straight legs does not necessarily pass that type of legs on to her daughter. The cow's genetic index is twice as accurate as her type evaluation as a predictor of how she'll breed. So use genetic indexes when making mating decisions.

TABLE 3. Expected progeny from mating 2 sires
to 2 cows for a trait with an intermediate economic optimum.
Sire Sire's
Cow Cow's
A +4 I +3 +7 +5
II -5 -1 +20
B -5 I +3 -2 +18
II -5 -10 0

The second caution relates to the emphasis placed on finding a bull that "corrects" a fault in a given cow. As mentioned before, production traits are the traits of major importance in almost all herds. Make sure that you don't sacrifice too much on the production side when trying to find a bull who sires straight legs for a cow with curved legs.


Although this article may not have given you all the answers on how to mate your cows, I hope it has provided you with insight into the issues and complexities involved. However, one of the most important aspects of mating cows, and also its major difficulty, remains setting your breeding goal by putting proper emphasis on traits of economic importance. No matter how fancy and sophisticated your matting scheme is, it will not get you anywhere if you are heading in the wrong direction.