Chapter V - Selection Picture

The dog breeder is merely a spectator of variation and heredity. However much he may know of the causes of these highly important factors in his breeding operations: however deeply he may be interested in the results, his direct control over either cause or result is nil. He is quite powerless to exert the least bit of influence over the combination of the heredity units in the germ plasm of the stud dog or of the brood bitch he breeds to him.

He can, however, select which two germ plasms are to be combined. In doing this, as he does every time he mates two dogs, he ceases to be a mere passive on-looker. He becomes the active, directing force. His will, in a measure, replaces natural selection. Obviously, his selection will have a most vital influence on the breed with which he is dealing. His skill and knowledge, or his carelessness and his ignorance, as the case may be, have effects much more far reaching than the immediate success or failure of his own kennels. Many breeders do not seem to realize the great responsibility that rests upon their shoulders.

The principles of variation and of heredity, those two important laws governing all breeding, must always be remembered during a consideration of the very practical problems of selection. Bearing this in mind, the solution of the problems of selection will be discovered in the answers to the following questions: What is the true object of selection? What can and what cannot be accomplished by selection? How can the results possible through selection be best accomplished?

What is the true object of selection?

We have already seen that all variations swing close to the race average, and heredity always tends to keep them there. These variations may be on one side of the average of the race, or they may be on the other—they may be, from a fancier's point of view, either favorable or unfavorable. But, excepting sudden mutations, all cluster about the race average, and just as many will be on one side as the other. The curve of normal variability illustrates this graphically. The principle of regression in heredity is always exerting its drag to keep the get of exceptional parents, which may be exceptionally good or exceptionally bad from the breeder's standards, closer to the average of the race than were the individual parents.

It is very evident that the only way the breeder can make any important, permanent headway is to bring the average of his own strain closer to the ideal expressed in the Standard than it is to the average of the race. In this way, and only in this way, can the drag of the race be lessened, and this drag is the breeder's worst enemy. Until he can overcome it, his breeding can only be partially successful. To overcome it, by raising the average of his own strain, is the true object of all selection.

Just what this drag of the race means to a breeder is expressed in mathematical form by Galton's law of ancestral heredity. This law was first derived from statistics. It has subsequently proved to be substantially correct. Pearson, by taking into consideration the individual variation in each generation, has worked out a more complete mathematical expression of the same principle, but for practical purposes Galton's figures are sufficiently accurate. It is very obvious, from what we know of variation and heredity, that inheritance is not alone from the immediate parents, but also from the parent's parents "even to the third and fourth generation." Galton's law, as stated in the following table, expresses in per cent. the effective heritage contributed by each of the first six generations and by every individual in those generations.

Generation Number of
Influence of
Influence of
1 2 50. 25.
2 4 25. 6.25
3 8 12.5 1.56
4 16 6.25 0.39
5 32 3.125 0.10
6 64 1.5625 0.024

These figures give true meaning to the drag of the race. They put before us vitally the advantages of lessening this drag, which should be the true object of all selection. Yet how very few keep this object before them. Most are trying, by hook or crook, to breed a champion. They rush cross lots, forgetting that "the longest way round is the shortest way home." Instead of trying to breed a chance champion, they should strive to bring their own strain gradually closer to the ideal. In other words, they should work to purify the ancestry of their strain, so that the inheritance from all ancestors may not only be similar in type, but also as close as possible to the ideal. Such a strain would be invaluable. To be sure, every puppy bred would not be a complete expression of the ideal—in fact the ideal dog had never been seen, and probably never will be—but that every litter would contain pups close enough to that ideal to win is beyond cavil.

Can this be accomplished by selection?

Theoretically it can. Practically, there are great difficulties to be overcome. In several strains, if not all the points that make up the modern dog, at least some of them have been brought to high and uniform perfection. The underjaws of Bulldogs bred in Mr. Walter Jefferies' Stone Kennels are famous the world over. "For legs and feet go to Redmond" is a byword among English Fox Terrier breeders. Llewellyn bred English Setters are universally accredited with dash and speed. A few other instances might be cited, but it is sad they are so few.

The great difficulty in dog breeding lies in the very great number of points that make up the dog. It would not be difficult to pick out one or two points and to establish a strain that would average close to the Standard. It is, however, a very different matter to bring all the multitudinous points of the modern show or field trial dog to the ideal. Nevertheless, it is possible, for variations are always being presented for selection. In working for the ideal dog a breeder sometimes feels as if he were trying to drain the sea dry with a sieve. This is not so, though it is almost as if he were attempting this task with an after dinner coffee spoon. It is theoretically possible to establish a strain that would turn out champions to order, even if the practical difficulties are great.

A glance at the table of Galton's law of ancestral heredity gives a foundation for this statement. It will be noticed that the influence of the individual ancestor becomes less with astounding rapidity. The individual contribution of a dog in the sixth generation is only 0.024 per cent. A somewhat involved mathematical proof has been worked out to show that after six generations of careful and continued selection a certain character will invariably breed true. No further selection for that point is necessary, provided no dogs which will deteriorate the inheritance for that point are introduced into the strain. In other words, six generations of selection in the case of our old example of the length of a Fox Terrier's head would raise the average from six to seven inches, and that new length would be indefinitely maintained so long as adverse selection was not brought to bear. The very practical advantages of possessing a strain of Fox Terriers whose length of head was fluctuating about the ideal average rather than an inch below it needs no comment.

The mathematical study of variation and heredity is of inestimable value, and serious breeders should acquaint themselves with the formulae which, however, have no place in a little handbook. Without going into the mathematics involved, it may be said that it has been established that any character consistently bred for can be fixed in six generations, but that there will be but a very slight reduction in the variation. We could increase the length of Fox Terrier heads in a strain from six to seven inches, but when the new average was established the number of dogs varying about this new point would be the same as formerly varied about the old average.

This seems a refutation of the common belief that a strain will die out. Except for adverse selection, either intentional or unintentional, there is no reason why, if fertility can be maintained, any family or strain should wear out, deteriorate, or become extinct. In actual practice there have been many cases where, in spite of all that could be done, a strain has degenerated. What has happened in most such cases is that the breeder has been forced to have recourse to outside blood to keep up the stamina and fertility of his strain. Adverse selection has been forced upon him willy-nilly.

Furthermore, it seems at first sight that if variability is not materially reduced by continued selection, a type once established should not only breed true, but be capable of indefinite further improvement. Raise the average length of head in a strain of Fox Terriers to seven inches, then set a new ideal of eight inches and repeat the process. It sounds reasonable, if it is true that variation is not reduced. The figures show that this cannot be done below 85 per cent. of the original variation, but there are other limits set to what can be done by selection. There is a definite mechanical limit, for example, to the length of leg and the weight it will support. There is a physiological limit to the work that can be done by the vital organs, such as the size of the heart and the amount of blood it can pump. These mechanical and physiological limits make it impossible to breed a Great Dane up to the size of a shire horse.

The breeder can expect that intelligent, continued selection will change type in any desired direction, and that new type will breed true after six generations of continued selection. He cannot, however, expect to accomplish any material reduction in the amount of variation. So far as the opportunity that variation always presents for further selection is concerned, the breeder will always have material available, but there are mechanical and physiological limits beyond which no amount of selection can ever be carried. However, in all probability, these limits have not been reached, except possibly in the size of the very large and the very small breeds.

If the true object of selection is to lessen the drag of the race and if careful continued selection can change type, but not reduce variation, how can the dog breeder most quickly and effectively accomplish his results? What are the principles involved in rational selection?

First and foremost, the breeder must know the points of the dog he is breeding. "An eye for livestock" is a common phrase. Some people are blessed with it: others seem to lack it. Several years ago, two friends visited my kennels to pick a Scottish Terrier puppy. Neither visitor knew the points of the breed, and color and markings were in this case little help in distinguishing between the youngsters. The elder could not tell one puppy from another: the younger could invariably pick out any one of the dozen odd with ease. The one had no "eye for livestock." The other, provided she learned the points of the breed, was a born judge. If a breeder has an "eye" for a dog he is indeed blessed. If he has not this desirable faculty, he should do all in his power to cultivate it.

Once a breeder has thoroughly learned the points of his breed, he must "keep his eye in," as the saying is, by attending shows. It is suicidal for a breeder to bury himself in his own kennels. He must visit shows, see and study good dogs, distinguish their excellencies and their defects. Otherwise, he will be unable to tell which of his puppies are ducklings and which are swans. The homebred puppy very often appears to be a worldbeater in his own run. In the show ring, however, his proud owner often discovers faults that he never before dreamed existed.

But the breeder who returns to his kennels with a new ideal of perfection after every bench show that he visits will never accomplish a thing. Once the points of the breed are learned, a breeder must set up before him an ideal toward which he must always work. To change that ideal every six weeks is to insure failure. True, the type that wins at the bench show to-day is often very different from the type that was winning five years ago. This places a great burden on the already overloaded shoulders of the breeder. It is hard work to establish any given type or any given set of characters. To be continually shifting that type and changing the characters makes lasting improvement impossible. From all that has been said of the true object of selection and from what we know of ancestral heredity, it is plain that to be forever tinkering with the ideal and at the same time to expect to fix that ideal in a strain are two things as incompatible as oil and water.

Besides knowing the points of his breed and having a fixed and definite ideal to breed for, the breeder must know the history of his variety. He should know what they were originally used for, and, if a manufactured variety, what breeds were employed in their manufacture. Even more important is a knowledge of the good and bad points of the different individual dogs of the past. He must appreciate the prevailing faults of his breed. He must know just what certain bloodlines stand for and what they mean in good and bad points. He must realize what strains are predominant and in what ways their predominance will be beneficial or harmful to his own stock. This knowledge is absolutely essential to a successful, intelligent selection. The practical application of the principles of variation and heredity, so essential to a breeder's success, can only be made by means of this knowledge. Yet how very often do we see men and women attempting to breed from dogs whose very sire and dam are absolutely unknown to them save in name only.

The sire and the dam are, so far as heredity is concerned, equipotent. The germ cells of each carry a complete set of all heredity units, as is evidenced by the nature of Mendelian inheritance. Either sire or dam may transmit to any puppy of either sex any or all of their own characteristics. The parent, however, whose heredity units for any particular point are purest will, in general, predominate over the parent whose inheritance for this same point is mixed. This explains the prepotency of an in-bred dog. A Bulldog, for example, in-bred to a strain noted for wonderful underjaws and layback, big flat skull, but with bad ears and a long tail can naturally be expected not only to possess these peculiar excellencies and faults himself, but also to pass them on to his get. The figures in Galton's law of ancestral heredity will prove to be a rough guide by which to judge the prepotency of an in-bred dog or bitch. Since they express this in figures, they are easy of application. Breeders must always remember that in-breeding is a double-edged sword. The faults or defects in any in-bred dog or strain will be intensified and passed on just as surely as any good points.

Quite aside from in-breeding, there is another feature of predominance that is highly important. In most breeds of thoroughbred dogs the most casual study of pedigrees impresses one with the great importance of certain dogs and certain families. In all breeds that I have studied, including Airedales, Scottish Terriers, Chow Chows, and field trial Pointers, I have found that most of the great dogs have been bred either in or close up to certain families. Wheeler in the Show Collie traces the main line of Collie blood through the various important dogs, each a direct descendant of Old Cockie. Mr. Ralph W. Condee has traced out the line of the main line of blood in Airedale Terriers, which has come down to us from old Champion Brush.

These interesting family trees are not works of the imagination, and any one can readily convince himself of the great importance of these main lines by selecting at random any important show winner of these breeds and seeing how very quickly his pedigree runs into the main line of blood. In fact, if his sire is not directly in the main line, the chances are that his sire's or his dam's sire will be. In the chapter on Pedigree Studies these two charts, and others, are given in full.

In most breeds it is safe to assume that there is a prepotent strain. A few phenomenal dogs and their get have been responsible for most of the improvement that has been made. It is impossible for a breeder to tell offhand whether or not any given dog is really a great sire, or merely a popular show dog being used extensively at stud. It will be well worth while to study the pedigrees of his own breed to determine what family has produced the great dogs. If a stud dog is bred in the right strain it is safe to assume that he stands just so much better chance of being a truly valuable stud dog than one outside of this line.

Although the influence of sire and dam are equal and prepotency may be displayed by either sex, still in a numerical sense the sire is more important. An average dog at stud will be the sire of fifty puppies to any one whelped by a bitch. His influence on the future generations of the breed is obviously the greater in direct proportion to the number of pups he gets. This must not, however, be taken in any way as an excuse for the old fallacy that "any old bitch with a pedigree is a good brood bitch." Quite the reverse is the truth. For the breeder, his own brood bitches are every bit as important as the stud dogs to which he breeds them. Both sire and dam are equal in any individual mating. The greater importance of the sire only applies broadly to the whole breed, but both for the advantage of the individual breeder and the good of the whole breed it would be well if it could be determined whether or not a certain dog is an exceptionally good sire during his lifetime, and not, as is so often the case, after he is dead. Sires should always be selected on their records as producers of good puppies, not on their records as bench show specimens or field trial winners.

In selecting the individuals for breeding stock a distinction should be made between faults and defects. A fault is positive: a defect is negative, though the difference is often mainly one of degree. Never breed to a dog that is downright faulty in any one point no matter how great may be his other excellencies is an old-time rule that is sound and sensible. Among the faults specially to be eschewed in any breed are those of conformation and soundness. Bad pasterns, cow hocks, deafness, blindness—except when any of these are caused by overwork or accident—should be shunned. Over-shot or under-shot, except, of course, in breeds where these formations of the jaws and teeth are required; or bad shoulders must be viewed with keen suspicion. Such positive faults, once they are introduced into a strain are hard to eradicate.

In studying pedigrees so that they may be made a practical aid to selection, the breeder must remember that while the importance of back blood is very great, still this importance lies mainly in the early generations. To go back five or six generations before finding a good dog in a pedigree is foolish. Take the very extreme example of when a good dog appears so many as ten times in the sixth generation. Even then his influence would only be 0.24, or less than one quarter of one per cent. If the individuals between the first and this sixth generation are inferior or commonplace or doubtful, such a pedigree, however beautiful it may look, is not a good one for breeding purposes.

To sum up: the real purpose of selection is to bring the type of the breed as close as possible to the ideal as set forth in the Standard. To do this for any length of time raises the quality of the inheritance of every pup since all ancestors are closer to the average of the ideal than to the average of the race. The drag of the race in this way becomes a help and ceases to be a hindrance. Selection will fix a character or type in six generations, but no amount of selection can materially reduce variation. There are, however, limits, mechanical and physiological, to what can be accomplished by selection, but in dog breeding these limits have not yet been reached. In order to make a wise selection, the breeder must know his breed, its points, its characteristics, its history, and, that his selections may be cumulatively effective, he must set up before him an ideal and always breed toward it.

Next chapter...
To this book's index page
Previous chapter...