The practical dog breeder, striving to produce a bench show winner or a phenomenal field trial performer, often shrugs his shoulders skeptically at the experience and precepts of breeders of other kinds of thoroughbred stock, and he is inclined to pooh-pooh the theoretical work of scientists. If the principles of variation, heredity, and selection do not apply to dogs as they do to other animals, then scientific dog breeding, in the accepted meaning of the words, is chimerical. The dog breeders' indifference to these things is, in this case, thoroughly reasonable and well justified. If, however, the reverse is true, then dog breeders merely handicap themselves by their failure to understand and apply the principles of genetics to their problems.
"The proof of the pudding," says the old saw, "is in the eating." The proof of the dog breeder's pudding is in the puppies. The kennel that consistently turns out home-bred winners, either on the bench or in the field, is avowedly successful. The dog show fancier's ideal is the bench champion: the hunting dog enthusiast's ideal is the field trial champion. This furnishes a definite standard of accomplishment of the object at which all serious dog breeders are aiming. Moreover, the records of the wins at shows and trials are available for reference. The records of wins plus the pedigree records of the stud books furnish us with a handy gauge by means of which we can either prove or disprove the efficiency of scientific principles applied to practical breeding problems and also, a good medium by means of which the work of the biologist can be translated into the everyday terms of the breeder.
I know very well the objections that can be raised against the stud books as a source of material for studies of this kind. In the first place, there is no guarantee that the pedigrees recorded are correct. There have been, of course, some cases of deliberate fraud in the registering of dogs, but ninety-nine dog fanciers out of a hundred would, I am sure, as soon steal pennies from a beggar as fake a pedigree record. Calamity howlers to the contrary, the men and women who breed dogs are at heart sportsmen and sportswomen, and for every story of crooked dealing one hears gossiped at the ringside, one can hear ten of such scrupulous honesty that it puts to shame the "butcher, the baker, and the candlestick maker." The deliberate errors in the pedigree records are so infrequent that for practical purposes they can be discounted.
More common is a distortion of the records through personal fancies or prejudice. For example, the stud books would be but a poor place from which to collect data on Boston Terrier colorings. Many dogs must be entered as "dark brindle, evenly marked with white," which are not, strictly speaking, this very popular color and much desired markings. Such errors are impossible of detection, and the sole solution lies in the selection for study of subjects that do not bear any taint of this kind. Most common of all are the typographical errors. These creep into any book, and, it is to be regretted, they are specially frequent in records of this kind, made up, as they are, from blanks filled in by many different persons writing in all degrees of legibility. Nor do the fantastic names bestowed on some dogs help this trouble. These most common errors, however, are seldom such that they affect the results of statistical studies.
Granting the weight of these objections, there is, however, much that a breeder can learn from studies of pedigrees. Nor is this, especially if one be working with his own favorite variety, the dry-as-dust work it might be supposed to be. The main objects of the studies I have made here have been to throw light upon the direct connection existing between the scientific work of the biologist and the practical work of the breeder, and to suggest studies that serious breeders can make with advantage each in his own breed.
Whenever two or three dog fanciers go hobby horse riding together the old question about in-breeding is sooner or later almost sure to be discussed. We have all heard some careful breeders praise such a system of breeding to the skies, and others, equally competent and successful, denounce it bitterly. The general public has the idea that thoroughbred stock, and most especially thoroughbred dogs, have been weakened in constitution and dulled in intellect by continued in-breeding. Many dog owners, while they will stoutly deny the ill effects of such mating, will almost in the same breath exclaim that there is altogether too much in-breeding. The whole subject has been badly muddled by a loose use of the term in-breeding, and by very hazy notions on the part of every one concerned as to just how common true in-breeding is among dogs.
In-breeding means nothing more nor less than the crossing of the blood of one individual. There are only three possible ways in which this can be accomphshed.
1. By breeding a sire to his own daughter, as expressed in a pedigree of the following general type:
2. By breeding a dam to her own son:
3. By breeding together full brother and full sister:
These, and only these, are true in-breeding. Bearing this in mind, any dog breeder will at once recognize that a great deal of what is really line breeding passes in current talk as true in-breeding. Such inaccuracy in the use of simple terms is decidedly unfortunate. The perfectly true statement that continued inbreeding results in degeneration of both physical and mental powers is transformed into a foolish bugaboo if every line bred dog is to be considered in-bred.
Just what proportion of thoroughbred dogs are in-bred and what proportion are line bred? Here is a question to which the stud book can furnish the answer. For the purpose I have selected two terrier breeds, the Scottish and the Airedale Terriers. These belong to the same general family, but the Scottie is one of the oldest breeds in this division, while the Airedale is a comparatively recently manufactured variety. In each breed one hundred average dogs were selected, fifty of each sex and ten from each stud book for the past ten years, so as to get a fair general average of the whole breed. In like manner, a hundred champions were picked out at random from each variety. The results were as follows:
Including both in-breeding and the primary cross of line breeding only 7 per cent. of the Scottish terriers are closely bred, and but 13 per cent. of the Airedales. Plainly, there must be a great deal of exaggerated talk about in-breeding. I will say frankly that, although I am tolerably familiar with Scottish Terrier breeding, I expected to find more than seven among one hundred average dogs closely bred; and an experienced and successful Airedale breeder to whom I put the question thought "about a third (33 per cent.) of the Airedales are close bred."
Those who advocate in-breeding as the solution of the dog breeder's problems are wont to claim that the majority of the champions and great dogs have been produced by this system of breeding. The records do not confirm this belief. Of the Scottish terrier champions only one in ten is close bred (including inbreeding and the primary cross of line breeding), and only one in eight of the Airedales. It is interesting to note that in both breeds the number of line bred champions greatly exceeds those that are truly in-bred.
In accomplishing the result desired by dog breeders, i.e., the winning of the honorable title, line breeding has proved to be, at least in the case of these two representative breeds of Terriers, tremendously more effective. It is also interesting for a breeder to note that in-bred dogs winning their championships have been only half as numerous as in-bred dogs in the average lot. In other words, an in-bred dog has only half the chance of becoming a champion that is enjoyed by an out-bred animal. In the case of the line bred dogs, however, the chances of winning the right to the prefix "Ch." seem to be double those of an average out-bred dog. Very evidently line breeding is the best friend of the Scottie and Airedale fancier. This is, of course, exactly what one might have foretold by a broad application of the principles of inheritance, remembering especially the behavior of all variations and the law of ancestral heredity.
One remarks at once in this table that the Airedales are more closely bred throughout than are the Scottish Terriers. I will hazard a guess, without dignifying it by calling it even a probable explanation, and suppose that this is the result of close breeding being forced on the Airedale fanciers because theirs is a recently manufactured variety and has therefore less variety in blood lines.
The very general recognition afforded to families and strains by dog fanciers, and the very commonness of such expressions as "the winning strain" and "bred in the purple" show that the prepotency of certain blood lines has been so great as to be apparent without any serious statistical study of pedigrees. In several varieties, the line of descent has been traced out through the strain that has proved to be preëminently successful in the show ring or the field. H. E. Packwood in his admirable monograph, "The Show Collie" goes to considerable pains to show that the vast majority of the finest specimens of this breed have been closely bred to the strain started by Old Cockie. The chart of the family he gives is as follows:
Somewhat similar tables have been worked out for other breeds, notably wire Fox Terriers, Airedales, Scottish Terriers, and Pomeranians. More elaborate records were analyzed by the late Major Taylor in his valuable work on the performances and pedigrees of the Setters and Pointers. Professor Davenport of the University of Illinois has made an exhaustive statistical study of the ancestry and track records of American trotting horse sires and their get. These studies demonstrate strikingly that the greatest improvement in speed has come through a few exceptional sires, all more or less closely related in blood.
Adapting Davenport's methods to a study of dogs, substituting for the time record of the trotter the winning of a championship as a standard of accomplishment, I have analyzed the ancestry and get of all Airedale Terrier dogs who have won the honorable title either in Great Britain or the United States. This data, when analyzed, bears out the general conclusion that most of the improvement in a breed of dogs, as in the case of a breed of horses, has been due to a few exceptional sires, all of the same family. It is interesting to note that the line of descent worked out by me from stud book statistics is identical with the family tree prepared empirically by Messrs. Ralph W. Condee and Earle J. Woodward. Statistical study coincides in results with the practical experience of breeders.
To January, 1913, eighty Airedale dogs and sixty-nine bitches had won their championship in England and this country. Of these eighty dog champions, who any fancier knows will have better-than-average opportunities at stud, fifty-three never sired a championship winner of either sex. Despite the exceptional stud opportunities of all champions, only twenty-seven of them sired champions. Moreover, of these twenty-seven champion sires of champions but thirteen were the sire of more than one champion. But, and this is significant, these thirteen champions sired forty-nine of the hundred and forty-nine championship winners. In other words, one-third of the total number of champions of the breed were sired by one-third of the dog champions, while only a half of all champions were sired by any champion. The thirteen exceptional champion sires got 33 per cent. of the champions, while the sixty-seven other dog champions between them sired only 22 per cent. of all champions. These figures are only roughly approximate, but they show impressively the tremendous influence for good of these thirteen exceptional breeding individuals.
In the second generation, of the eighty dog champions only twenty-four appear as grandsires of champions, and of this twenty-four, only ten are grand-sires of four or more champions. The fact that almost without exception the thirteen champions who sired two or more champions are also among those who were grand-sire of four or more champions is confirmatory evidence of their prepotency. The exceptional sires are also the exceptional grandsires. The following table shows the champions in the ancestry and get of these exceptional breeding individuals:
Without tracing out all the relationships—and every dog in this table is related—it is worth while to point out that all are descendants of Cholmondely Briar. The three greatest producers of the lot are Master Briar, grandson of Cholmondely Briar, his son Clonmel Marvel, and Crompton Oorang, who is by a son of Master Briar out of a daughter of Clonmel Monarch.
This same table is very useful in checking certain beliefs more or less popular among dog breeders. For example, there is the old kennel proverb that "a great sire is famous for his daughters." The table does not bear this out. The great Airedale sires have produced thirty championship winning sons to twenty championship winning daughters, or half again as many. Since the proportion of dog to bitch champions in the whole breed is only eight to seven, the great sires have been distinctly above the breed average in their male get who have won championships.
In the second generation, the proportion of grandsons winning their championships is even greater. The exceptional grand-sires produced seventy-two grandsons and fifty-three grand-daughters who were able to annex the title. Moreover, sons of champions produced eighty champions of both sexes, fortyfive dogs and thirty-five bitches, while the daughters of champions produced forty-five champions, twenty-seven dogs, and eighteen bitches. Except that both sons and daughters of champions produced more champion sons this is not a fair comparison, since the opportunities of the dogs are vastly greater than those of the bitches. In fact, when we consider the comparatively few number of puppies that any bitch can produce during her lifetime, the daughters of exceptional sires have made a wonderful showing and, compared with the average progeny of the average brood bitch, they surely fall under the class of the exceptional breeding individual. Theoretically, an ideal matron would be one whose sire and two grandsires were among the exceptional sires. Such a bitch could be very reasonably expected to prove the dam of puppies above the average class of the breed.
Certain characteristics in cattle, poultry, rabbits, guinea-pigs, and human beings have been proved to follow Mendelian inheritance, and it is a very natural hypothesis that certain points in thoroughbred dogs do likewise. Eye color, wire and smooth coats, and coat color are those characters that, from what has been discovered to be true in other animals, we should most naturally expect to behave in the typically Mendelian ratio when contrasting factors are crossed in dogs.
That black and brown coloring do in fact exhibit the same behavior in transmission among dogs as among mice and guinea pigs was forecast by the experiments of Professor A. Lang and Dr. A. L. Hagendoorn. This has recently been corroborated by the pedigree studies of coat color in Pointers by C. C. Little of the Bussey Institution. The Pointer was chosen because the breed is well established and exhibits at the same time marked and easily recognized differences in color, and yet there is no fancy, placing a premium on any particular shade. The results of Little's studies, which were first published in The Journal of Heredity, are set forth in the following table:
Little concludes from his study of these matings and the results obtained that there are two factors at work producing the colors in Pointers: "1. B. The factor for black pigment which is absent in brown (liver) animals. E. The factor for the extension of the brown and black pigment in the hair. In the presence of this factor, animals have spots of brown or black pigment on a white ground. In its absence the colored spots are yellow of various shade." This allows the following combinations, resulting in the four different visible types:
Little makes practical application of this analysis of coloring in the following suggestions to Pointer breeders: "In order to purify a strain from black individuals it is only necessary to go on breeding liver to yellow or yellow to yellow indefinitely. As long as this is carefully done there should be no blacks produced. Yellows from two liver parents may safely be crossed with liver-colored animals. If, however, any other yellows are used, the breeder may expect a certain number of black young among the progeny. To obtain a pure black strain is not so simple, requiring for its certain completion a separate breeding test for each black individual by crossing it with yellows coming from two liver parents. If among the progeny any liver or yellow young are found, it is certain that the black in question is not of the formula BBEE and will, therefore, not breed true."
This is a splendid example of the help that the practical dog breeder can get from the scientific work of the trained biologist. In the past, dog breeders have regarded such work as purely theoretical and of little or no use to them, largely because the data and experiments of these investigators of breeding problems have not been made with dogs. On the other hand, because the material is expensive and the time required for practical experimentation comparatively long, biologists have not employed dogs in their work on the principles of genetics. Recently there have been indications of a change in this condition. Dog breeders are beginning to awaken to the opportunities they have neglected in shutting their eyes to scientific help for their practical work. Biologists are extending their studies to include analysis of the statistical data of the kennel club stud books and direct experiments with thoroughbred dogs. Both breeders and biologists must profit by such a movement.