Having considered variation, which makes breeding possible, we are led naturally to a study of heredity which makes breeding effective. We have already discussed the germ plasm theory of heredity, the most reasonable explanation for the observed facts, and we must now investigate these facts to see what practical benefits we can derive from the knowledge gained by careful experiments and painstaking researches.
The very simplest case of heredity is the "pure line," or when reproduction is by means of self-fertilization. Although such reproduction is beyond the scope of the dog breeder's operations, still, because it throws much light on the nature of heredity, it will repay us for a moment's investigation.
Professor Johannsen made the original experiments in the study of pure line heredity with peas and beans, and his results have been checked and confirmed by other observers. He took the beans from nineteen plants, each of which had been self-fertilized, as is the usual condition in this variety, and he measured them for variability in weight, breadth, and other details. The whole lot showed a standard variaition, and the curve he plotted from his data was close to the normal variability curve. He kept carefully distinct the seeds from each parent plant, and grew each pure line, or in other words, each lot of seeds from each self-fertilized plant, separately. The plants raised in each pure line produced seed that had each its own distinct average, differing from the general average of the nineteen original parents. Moreover, he divided each pure line into lots according to size, but the same sized seeds from the same pure line produced seeds that were not of their own size, but nearer to the average size of the pure line from which both they and their parents had sprung.
This means that even within a pure line, when the germ plasm is, of course, unaltered and constant in succeeding generations, there is always variation. Variation is inherent in the germ plasm, even when it in itself is unchanged, for seeds larger than the average produce seeds smaller than themselves and closer to the average of the pure line. In other words, individual characteristics, even in a pure line when there is asexual reproduction, do not reproduce in the offspring.
For the breeder this means that no amount of selection in a pure line would ever produce a strain that would regularly have larger seeds than the average. In pure line breeding then, a very definite limit to what can be accomplished solely by selection is set. The great significance of these experiments for the dog breeder is the proof they present that even when the germ plasm is unchanged there is always variation.
In the case of bisexual reproduction, which is the condition under which the dog breeder works, two different germ plasms are united to form the new generation. The germ plasm is no longer a constant, the inheritance is double, from two germ plasms, and the case is obviously more complicated.
In an exhaustive statistical study of the height of 205 parents and their 928 children, Galton was able to analyze the results of bi-sexual reproduction. He arrived at conclu-sions that are of inestimable value to breeders. He found that:
1. Like parents beget unlike offspring, and vice versa, like offspring come from unlike parents. Abnormally tall parents, he found, had tall, short, and medium sized children; while of all the tall children studied, some came from tall, others from short, and others from medium sized parents.
2. Offspring are on the average nearer to the average of the race than their parents. The average height of the children of two parents is not the mean between these parents. When the average height of the two parents is above the average of the race, the children's average will be shorter than their parents, and so closer to the race average.
From the first conclusion it follows that two dogs of very different type may be whelped in the same litter, a fact to which any practical breeder will readily testify. Also, that two dogs of remarkably similar type may have been bred from very different parents, another fact that is corroborated by common experience. The general conclusion emphasizes the value of a good working knowledge of the true meaning of a pedigree and shows the utter foolishness of any attempts to judge the offspring by the parents, or the parents by a consideration of the points of the offspring. Almost daily we see dog fanciers attempt these two impossibilities. It is common to hear a man say, "No, I never saw such-and-such a dog, but judging from his pups he must be a shortbacked one with good legs and feet," or something of that sort; while it is even more usual to hear a man say that "So-and-so has the best eyes and ears of any dog at stud, and he certainly ought to get pups good in these points."
The second conclusion of Galton's studies is the principle of regression, or the drag of the race. To the dog breeder it presents vitally the true value of a pedigree. To return again to our example of the Fox Terrier head, which in the last chapter we assumed would average six inches long with an ideal length of seven inches. It would surely result in disappointment to breed together two dogs with ideal seven inch heads, for regression would bring the average of the resulting puppies back closer to the average of the race, which would be closer to six inches. Conversely, a dog and a bitch with five inch heads would, on the average, produce puppies longer headed than themselves, for the average of the offspring would by the same law approach nearer to the average of the race, which in this case is an inch longer than the immediate parents.
Could a complete refutation of the ideas usually followed by dog breeders be more forcibly expressed? We are so very prone to cant about "like produces like," and so very willing to accept a pedigree, which at best is only a guaranty of purity of blood, as proof positive of uniformity in type.
Plainly, there is but one way to cut loose from the drag of the race. Bring the general race average as close as possible to the ideal expressed in the Standard. In this way, and only in this way, can regression be won over from an enemy to an ally. If a Fox Terrier breeder should by continued and careful selection raise the average of his own strain from the general race average of six inches to the ideal average of seven inches, he would not have to worry about length of heads so long as he exercised enough selection to hold the very great advantage he had gained.
Besides the pure line inheritance and the bisexual inheritance there is another, named after the man who discovered it, Mendelian inheritance. Mendel, an Austrian monk, studying the crossing of different varieties of garden peas, made important discoveries that were quite unappreciated for thirty-five years. In 1900 his work was rediscovered and confirmed by De Vries, Tschermak, and Correns, each working independently. The garden pea shows sharply differentiated characters in its different varieties. Mendel crossed these and observed the way that these different characters were inherited in the hybrids. Mendelian inheritance then is primarily the inheritance of hybrids, or cross-breds, but subsequent study has shown that many individual characters in straight bred animals follow this same law.
It is beyond our needs to go into all the technical details of Mendel's experiments, or to know how he succeeded in being sure that he was crossing certain plants. We will confine ourselves to the results obtained in one particular case. Crossing the tall variety, which is about six feet high, and the dwarf, which is about a foot and a half high, Mendel got a generation of plants, every one of which was just as tall as the tall parent. This is certainly not what one would naturally expect, for we generally look upon cross-breds as a combination of their parents, and we would think the offspring of the tall and short varieties would be about four feet tall. These tall cross-breds were allowed to fertilize themselves, which is the usual method of reproduction, and from the resulting seeds a second generation was raised next year.
This second hybrid generation behaved in a truly extraordinary manner. Many plants were just as tall as their tall parents and the tall half of their grandparents; others, however, were just as short as their dwarf grandparents. There were absolutely no plants of intermediate height. What is even more remarkable, the dwarfs bore a constant numerical proportion to the talls. There was one dwarf to three tall, or twenty-five per cent, of the second hybrid generation were dwarf. In the next and in all succeeding generations, these dwarfs continued to produce only dwarf plants. Here was a hybrid, breeding absolutely true, a perfect dwarf produced from tall parents, produced in turn from crossed tall and dwarf.
The seventy-five per cent. talls in the second hybrid generation behaved very differently. Self-fertilized, some produced both tall and dwarf plants, while others produced only talls. It was found that twenty-five per cent, of the apparent talls of the second hybrid generation were true talls and continued to produce talls indefinitely. Those seeming talls, half of the whole second generation, continued to produce both talls and shorts, in the ratio of twenty-five per cent. true dwarfs, twenty-five per cent. true talls, and fifty per cent. seeming talls, but in reality hybrids in inheritance.
To sum up the results of this important discovery: the first cross between tall and dwarf produced all talls. Mendel expressed this by saying that tallness is in this cross dominant and dwarfness recessive. This hybrid, tall-dwarf, but tall looking generation, produced twenty-five per cent. true dwarfs, twenty-five per cent. true talls, and fifty per cent. hybrids with tallness dominant. This is known as segregation, or the sifting out of the offspring in definite proportions of the characters employed in the cross. This proportion is 1 : 2 : 1. These same results can be expressed in a chart:
The first cross is represented by the letters T (tall) and D (dwarf). The fact that tallness is dominant and dwarfness is recessive is represented by the symbol Td, a tall looking plant in which dwarfness lies recessive or hidden. The true talls (TT) and the true dwarfs (DD), which are later segregated, continue, as is shown, to breed true.
This principle of segregation, or the splitting up of the off-spring into the Mendelian ratio of 1 : 2 : 1 is a fundamental part of this type of inheritance. Dominance of one character over another does not invariably occur. Sometimes there is a blending. In such cases the symbols in the chart would be changed from Td to TD, representing in this particular case a plant of intermediate height. The subsequent splitting off into true talls and true dwarfs would be the same in each case.
Beyond all further doubt Mendel's law has been demonstrated to hold in its mathematical relations. The breeder, however, must remember it applies only to one character or set of characters, not to the entire individual. Judged as individuals, the Mendelian nature of a hybrid cross might not be at all apparent, though each character be following strict Mendelian inheritance. Some characters would be dominant in one parent, others in the other parent, and still others might be a blend. In this way, the different characters, viewed as a whole, would seem a hopeless muddle. For this reason the true nature of such inheritance was so long obscured, only to be discovered by the careful isolation and study of each character by itself.
The dog breeder can make no use of Mendel's law until he establishes what characters in dogs, if any, follow it. That there are such characters is highly probable. Color in chickens, pigeons, rabbits, guinea pigs, and cattle; hair and eye colors in man; presence or absence of horns in cattle; the shape of the comb in chickens, and many other similar characters have been found to follow Mendelian inheritance.
A. L. Hagedoorn has done some work on color inheritance in Dachshundes, and C. C. Little has made a statistical study of coat colors in Pointers from data in the A. K. C. Stud Book. Their work, which supplements the rather scanty data of Professor A. Lang, indicates that black and brown (liver) follows the same Mendelian inheritance observed in these colors in mice, guinea pigs, and rabbits. Dr. C. G. Darling believes eye coloring in Airedale Terriers is Mendelian, the light color being dominant. He acknowledges that he has not sufficient data to either prove or discredit this hypothesis, but, as an eye specialist and a terrier breeder, his opinion bears weight. If he is correct, it is probable that all eye coloring in dogs follows Mendelian inheritance.
It is also probable that the smooth and broken coats in Fox Terriers, a form of cross breeding that is common, is Mendelian, the broken coat being, in this case, dominant. The red and black coloring in Chow Chows and self colored spaniels is also probably according to Mendelian inheritance. However, before a positive statement can be made in any of these cases, more evidence is required. Such evidence would be a valuable contribution to the equipment of breeders, and it is to be hoped that some day it will be collected. To be of practical value, it must be determined by a careful study of a great number of individuals from all possible combinations, for large numbers are necessary to establish the true ratio, and of course, the greater the number of cases the less the probable error.
In view of the great likelihood of different characters in dogs being subject to Mendelian inheritance there is a practical value in knowing what are the average numerical results to be obtained from crossing characters following this ratio. Let us take a simple case when black and red colors are crossed, the black being dominant. The symbols used are the same as before, i.e., BB, a true black; RR, a true red; and Br, a seeming black with red recessive. A hundred offspring will in every possible cross give the following approximate results:
Mendelian inheritance is particularly applied to crosses of certain sharply defined characters. But quite aside from this practical application of this type of inheritance, in a peculiar manner it throws a strong light on the nature of the germ plasm and the whole subject of heredity.
From the action of characters under Mendelian inheritance we can see that the units of heredity in the germ plasm remain, even when crossed, true and pure in respect to any given character. To return to Mendel's original experiment, the unit for tallness is carried by one plant and the unit for dwarfness by another. On combining the two germ plasms these two units remain distinct, or all offspring of the cross would forever afterwards be a blend, and there could never be separating of the offspring back to the original sizes. The tallness and the dwarfness remain distinct, though they may blend.
Each hybrid germ plasm contains heredity units represented by T and D. When crossed the T's of one plasm combine with the T's of the other plasm, giving TT or true tall, or they may combine with a D, resulting in TD which may be either a blend or one factor may dominate the other. The D's act in the same way, they may combine with other D's giving DD, or true dwarfs, or with T's giving DT, or TD, which is the same thing.
Each germ plasm of every individual has two determinants as they are called. These may be TT, or DD, or TD. On crossing these couples in each individual act independently, and one determinant of one parent will combine with one determinant of the other parent. Accordingly in crossing TT x DD the only possible result will be TD. This is exactly what happens in the first hybrid cross. But on crossing TD x TD we can get either TT, or DD, or TD, and it is a mathematical certainty that the chance of TD combining is just twice as great as TT or DD, hence the establishment of the Mendelian ratio of 1TT : 2TD : 1DD.
It is just as if you tossed two coins in the air. The only possible combinations for you to get would be two heads, one head and one tail, or two tails. If you did this five times, it might happen that you got two heads every time; but if you did it a thousand times and kept count, you would find that you would get very close to 250 two heads, 500 heads and tails, and 250 two tails — the Mendelian ratio of 1 : 2 : 1.
The practical application of this is the lesson it teaches that in Mendelian inheritance it is useless to try to establish in a strain a blend between two characters. Such a blend will never breed true. The characters will continually be splitting up into the two original forms.
Another very practical lesson is that very evidently the germ cells of both parents each contain a complete set of hereditary units. Every possible character is represented in both male and female, which applies to all inheritance whether Mendelian or otherwise. This upsets the idea that the sire is more important than the dam so far as the physical appearance of the offspring is concerned. This is a time honored belief that dies hard, but the sooner it is buried the better it will be for all breeders.
The question of whether or not acquired characteristics are inherited has been long debated by biologists. The tendency is to place less and less credence in this once popular belief. Practical breeders ought to be able to distinguish true acquired characteristics, so as to appreciate their relation to his operations.
First, such a character is only acquired during the lifetime of the individual. Those characters that have been acquired by the whole race are beyond the scope of this definition. The retrieving habit, which must be taught to a bird dog, is an acquired characteristic: the pointing habit, which they have inherited, is not.
Second, a factor outside the dog, something in his habits, training, or environment, must have brought about the change. Cutting off a terrier's tail is an acquired character: the tendency displayed by many terriers to go thick in skull, though this happens in the dog's lifetime, is not.
Third, and this is the most difficult point to establish in the individual case, the acquired character must affect only the body of the dog and not his germ plasm. Bad raising during puppyhood may result in rickets and other weaknesses. If these weaknesses go further and affect his fertility they cannot be strictly considered as acquired so far as that dog's heritage is concerned.
In this strict, scientific sense, acquired characteristics are obviously non-inheritable, else long ago our Terriers and Spaniels would have been born with short tails and no trainings would be necessary for bird dogs and hounds. Diseases, as such, are not inherited strictly, though, of course, communicable diseases may be transmitted by the dam to the pups. This is not inheritance but infection. The tendency to develop certain diseases is, however, passed on from one generation to another. Use and disuse of certain faculties or organs probably act much in this same way. The fact that Pointers and Setters have for generations been broken to the field makes Pointer and Setter puppies easier to train. Exercise of certain muscles develops them and makes them stronger. Effects from use and disuse must, however, be very slow in their action. They are felt more in transmission of the capability for further development than in a direct inheritance.
In practice the breeder need not worry over the inheritance of acquired characters, provided he is assured they are acquired in the strict meaning of this term. This is not so in the case of care and treatment of his breeding stock and puppies. Environment is a very different thing, and poorly housed, dirty, under-fed stock are not good breeding stock. Environment has a very direct action on development, and the breeder must maintain his kennels under favorable conditions that will insure strength and health among his dogs.
In our conceptions of heredity we dog breeders have made two mistakes. These are natural ones, and it is some consolation to know that other breeders, and even trained biologists, have fallen into the same errors. In the first place, we have paid too much attention to the exceptional individual, the dog that is a "stormer," way above the average of his race. Secondly, and this sounds somewhat paradoxical, we have not paid enough attention to the individual points that go to make up the whole dog.
In our almost fetish worship of the Champion of Record, we have been led astray in formulating any sound systems of breeding. We have overlooked the great average of the race and the drag that this average always exerts. This has been very strikingly demonstrated in the statistical studies of inheritance which were pointed out earlier in this chapter.
Although as breeders we are continually working for the development or effacement of certain points, we have overlooked the fact that these different characters behave differently in transmission. Some blend, others never do. Some are correlated, others are quite independent.
The fact that heredity is from the whole race more directly than from the individual is forcibly impressed on us, and the fact that heredity keeps all variations close to the race average, together with the fact that many characters combine in definite proportions, bring out the mathematical nature of all inheritance. We are working with tremendously complicated material. It is little wonder that this mathematical relation of variation and heredity should be obscured. But picking out individual characters and working with them in large numbers give new ideas and fresh inspiration to the careful breeder. We can now appreciate the real significance of scientific breeding, and understand that it is not merely fine spun theory.
The principles of variation and heredity in the light of modern biological knowledge enable us to make our selection in matings with a fuller understanding of the problem before us and with a more reasonable expectation of success. It is very much more effective than the old hit and miss methods.