International Conference on Plant Breeding and Hybridization (1902)
W. Bateson
Cambridge University, England.

Mr. President, Ladies and Gentlemen: It is impossible for me to begin the serious discussion of these subjects without expressing the pleasure that I feel in having this opportunity of addressing you. It is and must be always a great pleasure to a man who is engaged in a very special line of inquiry, as breeding experiments are, to meet others who are engaged in such work, whose thoughts are centred on the same problems as his own. Especially may I welcome this opportunity of speaking here in the United States, where what is being done in this line of inquiry is on a scale of comprehensiveness which I may truly say far exceeds anything that is being done in any other country in the world. We have only to glance at the publications of the agricultural experiment stations to know what progress is being made in this line. Here amid vast diversities of soil and climate the great resources of the States are being applied to the elucidation of these problems, with the result that the scope of the work carried on entirely surpasses that which is attempted by other nations. It is therefore with especial satisfaction that I welcome the opportunity of addressing those who in the United States are devoting themselves to the study of experimental breeding.

In these studies we have reached a critical moment. That crisis, as it is known to many of those present, has been brought about by the rediscovery and confirmation of Mendel's work on heredity. These discoveries intimately concern the art of the practical breeder, and I propose to use the present opportunity to indicate some of the ways in which we can employ them for his purposes.

The essential point which Mendel discovered in peas, and others are now discovering in various fields of inquiry, is that, though a plant or an animal may be made up of a great complex of characters, height, size, color, hairiness, form of fruits and organs, etc., yet in a very considerable number of cases, a number which increases almost every month, those characters may possess an individuality manifested in the formation of the germ cells. When two varieties differing in, say, color, or form, or hairiness, or whatever it may be, are crossed together a "hybrid" is formed. That hybrid when it comes to make its own germ cells, male cells, or female cells, makes them in a number of cases, indeed in all Mendelian cases, such that each germ cell represents one of the pure grandparental characters, and not both. That is the essential discovery of Mendel. The cases that are most familiar are those in peas, the subject on which he originally worked. If a pea with green cotyledons be crossed with one having yellow cotyledons a hybrid is produced. That hybrid grows up and bears peas in its turn. Those peas will be composed, each individual pea, of a union of two germs, each germ being a carrier of either one or the other of the pure parental characters. Therefore we may have two green germs uniting, or two yellow germs uniting, or a yellow germ uniting with a green. Each gamete in such a case is pure to one or the other of the two parental characters which you first put into the hybrid. In other words, we can recognize many different characters in animals or plants which are unit characters, and in the formation of gametes are treated as distinct entities or units.

If, instead of using pure parental forms differing from each other in respect of one pair of antagonistic — allelomorphic characters, as we call them — we use parental forms distinguished in respect of two, three or more pairs of allelomorphs, then each germ cell in Mendelian cases will contain or transmit one character only of each pair.

To use an illustration: In chemistry you may have a body, say, a simple salt, from which you can take out the base, or the acid radical, replacing the base by another base, or the acid radical by another acid radical. You can in that way decompose your substance into component parts, reforming them in various combinations. So we must imagine a plant which has one element of color, for example, another element of texture, etc;, and we must conceive that when two varieties are crossed together the unit characters can be combined and recombined in the gametes of the hybrid, alternating with and replacing each other by substitution. You can take out greenness and put in yellowness; you can take out hairiness and put in smoothness; you can take out tallness and put in dwarfness, etc. The characters have their fixed possibilities of union, and hence it may be possible for us to form some mental picture of the constitution of the organism.

Now when we come to the question of the significance of these things to the breeder or to the hybridist, it will be found that the significance is exceedingly great I am afraid of saying that we have already reached a point when the practical man who is doing these things with a definite, economic object or commercial object in view can take the facts and use them for his definite advantage. But we do for the first time get a clear sight of some of the fundamentals on which he will in future work, and it cannot be now very many years, if the investigations go on at the present rate, before the breeder will be in a position not so very different from that in which the chemist is: — when he will be able to do what he wants to do, instead of merely what happens to turn up. Hitherto I think it is not too much to say that the results of hybridization had given a hopeless entanglement of contradictory results. We did not know what to expect. We crossed two things; we saw the incomprehensible diversity that comes in the second generation; we did not know how to reason about it, how to appreciate it, or what it meant. We got contradictory results, and the thing looked hopeless. But with the discovery of the purity of the germ cells we have the first step, which, I think, is bound in a very short time to become a path through many of those wonderful mazes of heredity.

To the practical man, I take it, the importance of this discovery comes in, first, somewhat as follows: Seeing that the gametes are pure with respect to their characters, it follows that an individual which is produced as the offspring of a cross will be composed, in respect to any one pair of these characters, either of two similar gametes or of two dissimilar gametes. Take the case of the pea. Any one pea descended from an original cross between yellow and green will either be composed of two similar green gametes, or of two similar yellow gametes, or of a yellow and a green. Now, as it happens in the case of the pea, and in a great number of other cases, unfortunately for the breeder, there is no means of distinguishing outwardly by any test that we can apply whether the organism is a hybrid or pure to the dominant character. There is no way of distinguishing in the cases where yellow meets green whether the organism is a hybrid that is composed of yellow and green, or whether it is pure to the yellow character and is composed of two yellow gametes. There is no possibility of distinguishing, because the yellow is, as Mendel calls it, dominant, and the green is hidden, or, as he calls it, recessive. We have, therefore, in such a case as that, two classes of organisms, pure and hybrid, each showing the dominant character, and it is owing to the fact that the pure dominant cannot be distinguished from the dominant hybrid containing both dominant and recessive characters that an immense number of the contradictions which the practical breeder experiences have come about. For example, a breeder or seedsman introduces some strain of a new variety of his seed — peas, or whatever it may be. He finds a number of rogues which are not true to the character which he desires to put on the market — rogues which he is unable to eliminate. Formerly we said it was only a question of time; he must hoe out the rogues and go on, and he will gradually fix his type. But now we begin to see what the facts really mean. He hoes out the rogues, and again they come — in diminishing number, no doubt, but they are still there. We now suspect the nature of such rogues in a considerable number of examples. For example, the bearded wheats occurring among wheats which are intended to be beardless. Every year they grow, and every year the seedsman hoes them out, and again they come back. Those bearded wheats may come from the fact that the beardless wheats had a bearded ancestor, and some of them contain bearded germs. If you cross a bearded wheat with a beardless wheat, the first cross will be a beardless wheat. You allow it to fertilize itself, and you sow your crop; you begin to get beards and beardless. You take out the bearded, and again there will be a beardless crop with a certain number of bearded. To get your crop pure in a few generations you should make your selection from individual plants. Then you will begin to find that some plants carry only the bearded character and some will carry both. It is the coming out of these recessive characters, owing to the fortuitous union of recessive germs, which shows itself in the offspring you desire to get rid of.

Whenever, then, it is desired, in a crossbred strain, to fix a dominant character selections must always be made of single families containing no recessive members.

We reach, therefore, a fact of immediate interest to the practical breeder. We have lost forever, I think, the conception that fixity of character is solely or chiefly a function of the number of generations during which that character has been manifested, or of the number of successive selections of that particular variety which have been made. Purity of strain or fixity of character is, on the contrary, due primarily to the union of similar gametes in fertilisation. Such purity may therefore occur among the immediate offspring of crossbred organisms.

*The extra toe of the Dorking fowl, for instance, is uniformly dominant in some strains, but not in others.

Another question of considerable practical significance is that of the nature and causation of dominance, involving the further question whether the breeder has any means at his disposal by which dominance may be created, modified or controlled. Upon this point experimental results are still to seek, and though there are a few cases* where we know that the dominance of one character over another varies in intensity, we have no clear indication as to the causes governing these differences of intensity. We may naturally be disposed to consider whether continued pure-breeding, or, perhaps, in-breeding, may not be concerned in the creation of dominance, but the facts at present ascertained give no clear light on this question. We have, however, abundant evidence that pure breeding is not essential to the constitution of dominance; for in any simple Mendelian case the pure dominants, offspring of one crossbred and one dominant parent, or of two crossbred parents, may, and commonly do, show unimpaired dominance over recessives of pure lineage. But there is another class of facts which, to my thinking, is far more interesting than that, and is of more significance to the practical breeder, and that is this: I spoke in the case of the green and the yellow pea of the offspring resembling the dominant, the yellow. But in a great number of cases we find a phenomenon not nearly so simple as that. When similar germs meet they produce a pure bred organism, which in my terminology is railed a homozygote, a yoking together of like germs. When the germs are dissimilar they make a new form, a hybrid form, which in this terminology we may call a heterozygote, the yoking together of two dissimilar germs in the zygote form. Until we have seen the heterozygote, its form is not predicable in any specific case. You cannot say until you have made the specific cross what the character of that heterozygote will be. It may be that, through dominance, one character only prevails to the exclusion of the other, or the heterozygote may have some form totally distinct from that of either of the parents.

For example, in a case that I see a great deal of, in the sweet pea, you may by crossing two sweet peas produce the old purple sweet pea with chocolate colored standards and purple wings. That purple sweet pea so produced will not breed true. The old purple pea of the gardeners was a pure pea and would perpetuate itself truly from seed; but the purple pea produced as a heterozygote form will not breed pure, but will split up into the components which produced it. So that we recognize that there is a new form, a heterozygote form, which, though it may resemble some pure form, will not breed true. This is a case that may not interest the seedsman, because he does not want the old sweet pea. Nevertheless, in his fertilizations he may produce another new form which he does want, and after all his laborious selection he may find it is only a heterozygote which will never breed true.

It is a curious and unexplained fact — constituting one of the most fertile fields of inquiry — that when dissimilar gametes meet they should so often produce an ancient form. That is what we now recognize as the rationale of Darwin's "reversions on crossing." When Darwin crossed his pigeons he brought back an old form; and so in crossing many plants you can get back a reversionary form by uniting two dissimilar gametes.

In my own experience a most extraordinary case of this nature has occurred. When the Mendelian discoveries were first announced it was obviously desirable to cross two varieties differing in some visible character of the gametes (whether alike in other respects or not). By such means we might hope to make visible that mixture of dissimilar gametes which must certainly occur in Mendelian hybrids. Unfortunately the actual gametes of flowering plants are not adapted to this experiment, but the nearest things to them are the pollen grains. So I cast about for a case of visible variation in pollen. By good fortune I found them at once in a certain sweet pea.

Ordinary sweet peas have their pollen grains elongated, with three pores. The white variety known as Emily Henderson, an American sort about eight to ten years old, usually has pollen grains which, when treated with acids, etc., are seen to be roughly spherical, with only two pores. Various grains of intermediate types are found from time to time in the pollen of E. Henderson, and the round grains not very rarely have three pores. But the pollen of a round-pollened plant can generally be distinguished immediately from that of a long-pollened plant.

Proposing, then, to cross E. Henderson with sweet peas having typical pollen, I sowed a quantity of that variety. But when the plants flowered I discovered that, though a majority of the Hendersons had round pollen, a few, though otherwise indistinguishable from the others, had nevertheless long pollen, exactly like a common sweet pea. I then crossed the round pollened Henderson with the long, and vice versa. The same experiment was also made independently by Miss E. R. Saunders. Every seed then produced (from four capsules) has given a plant with chocolate-purple standards and blue-purple wings! There are many details respecting this remarkable case which I hope ere long to publish, but I mention it now as illustrating in a striking way how paradoxical are the phenomena empirically produced by the experimental breeder, and how puzzling are these heterozygous forms.

I may say that my experiment entirely failed to fulfil its original purpose, which was to see a mixture of true long grains and true round grains, for all the pollen of these plants was typically long, showing dominance of the long pollen as a plant character.

I know no example of the production of an atavistic heterozygote so curious as this. It should be stated that white color of flowers is, in general, a pure recessive character in sweet peas. Hence, had I happened to cross a long and a round pollened Henderson together without knowing of this pollen difference, I should not have been aware of the heterozygous nature of the purple mongrel, and should have been still more hopelessly unable to bring the facts into line. Even as it stands, we may feel fairly sure that something more than simple Mendelian phenomena are presented by this case, but pending the next generation we cannot analyze it any further.

The occurrence of these heterozygous forms concerns the practical breeder very closely. The breeder may breed a new variety of value, and he may be most anxious to obtain its seed pure. Year by year he selects it, but every year, if it be a heterozygote, it fails to come true; because, as we now see, its germ cells do not transmit or represent the heterozygous character, but merely the pure characters of its components.

In the garden of my friend, Mr. Sutton, of Reading, I have seen a case of this kind. It is a beautiful Chinese Primrose (Primula sinensis) of a curious lavender color. The seeds of the self-fertilized lavenders are sown each year, but of the total offspring only about half are lavenders, one-quarter being a tinged white and one-quarter magentas. We can scarcely doubt that the lavenders are formed as the heterozygote of that particular white and magenta, the whites being homozygotes formed by the union of two white gametes, while the magentas are similarly formed by the union of two magenta gametes. In such a case statistical study of the offspring will show the breeder with approximate certainty what he is dealing with, and will give him a good indication whether it is worth while for him to continue in his attempt to get the variety true.

A case almost certainly of the same nature occurs in poultry — the case of the Andalusian fowl. The Andalusian was at one time a favorite breed. Its plumage is of a peculiar blue-gray, mixed with black. You may go to the poultry shows and buy the winning Andalusians, thinking that they will breed true. But they will not. Andalusians have been bred for at least forty or fifty years, and there is no good reason for thinking that they breed any truer now than formerly. Every one is agreed that the breed possesses this drawback. The "impurity" manifests itself in the production of numerous black birds and numerous white birds irregularly splashed with blackish gray. From such evidence as I can obtain it seems almost certain that these two objectionable forms are produced in about equal numbers, and that the number of true Andalusians is about double the number of either. The Andalusian is almost unquestionably a heterozygous form made by the union of the black gamete with the white-splashed gamete. It is, moreover, on record that the two sport-forms crossed together produce only Andalusians, as they should do if the case is a simple Mendelian one. We may, therefore, predict that the Andalusian, like the lavender Primula, will never breed true, however well or long it be selected.

In these brief remarks I have indicated some of the lines along which the Mendelian discoveries will have a close bearing on the work of the practical breeder. We have for the first time a conception of the true nature of at least a part of the facts which underlie the outward and visible phenomena witnessed by the breeder. As I have attempted to show, we have at last a clear notion of the meaning of purity or fixity of type, of the consequences of dominance and of the nature of heterozygous forms — phenomena which go to make up the daily experience of those who are practically engaged in these pursuits. It is impossible on the present occasion to go into many other fascinating problems suggested by these simple facts. For example, we do not proceed far with the practice of experimental breeding before we meet the phenomenon of the decomposition or resolution of compound characters into simpler constituent characters (hypallelomorphs), themselves possessing a measure of individuality. Then again we are presented with a whole series of possibilities of the utmost consequence both to the naturalist and the practical breeder.

It is difficult to see this phenomenon of the decomposition or resolution of compound characters without feeling the conviction that we have here the key to a great part of the mystery of parallel variations. We are led to suspect that the series of colors, for instance, into which the original color of the Carnation has been split up may be a similar series to that into which, say, the sweetpea has been split up. We can in this way imagine that each series of component colors consists of a number of definite terms related to each other in a definite way such that, if we could ascertain the relation of yellow in the one series, we could predict somewhat similar relations for yellow in the other series. The colors of flowers give us many such series, and even classes of series, of which some have obviously distinct laws of their own. Nevertheless, it is in a high degree likely that if one such series of colors were studied statistically in such a way that what I have called the mutual relations of its terms could be stated, we should have a model which would enable us to reconstruct other similar series, to predict its terms, and possibly to set about producing them at will.

In this paper I have spoken only of the simpler deductions from Mendel's principles. To this audience I need scarcely say that we are well aware that those principles in their simple form cover only a part of the phenomena of heredity. In trying to extend them or to cast them into a general form many reservations must be made that cannot now be detailed, and a vast field must be covered by specific experiment before such generalizations can be successful. Chief, perhaps, of the difficulties we can at present foresee is that caused by the existence of numbers of specific heterozygotes, which may appear quite unexpectedly owing to the presence of unknown differentiations between parent-strains presumed to be identical. Such a case is that of the E. Henderson mentioned above. Phenomena of this kind will doubtless be found elsewhere, and will lead to great difficulties of interpretation. Against such cases the observer must be on his guard. The significance of such forms can only be studied by an analysis of their offspring.

In addition to the general development of the inquiry we may note three chief subjects that call for immediate investigation:

  1. The resolution of compound characters and a statistical study of their components.
  2. The nature of dominance and its possible limitations.
  3. The detection of differentiation among the gametes of cross-bred organisms.

As to the last two we are still in ignorance how to proceed, but the first is a question we can at once attack by Mendelian methods.

But apart from the profounder mysteries, the unravelling of the problems of heredity has now become a matter for simple statistical research. Owing to the scale on which they must be pursued, it is likely that for their further elucidation we must perhaps look rather to the practical breeder whose operations are of large extent than to the scientific investigator whose resources are generally of a more limited character. But if in the future some cooperation between these two groups of workers can be secured, we may confidently look forward to the time when the laws of heredity, hitherto a hopeless mystery, will, in their outward presentments, at least, be, as the laws of chemistry now are, a matter of every day knowledge. The period of confusion is passing away, and we have at length a basis from which to attack that mystery such as we could scarcely have hoped two years ago would be discovered in our time.

[For a fuller account, in English, of Mendelian facts and problems, the reader is referred to the Report to the Evolution Committee of the Royal Society No. 1, by W. Bateson and E. R. Saunders; also to Mendel's Principles of Heredity, by W. Bateson, containing a translation of Mendel's papers, together with a discussion; published by the Cambridge University Press (in America, the Macmillan Company). These papers give references to the chief writings on the subject, especially those of De Vries, Correns, and Tschermak, who almost simultaneously announced the rediscovery and confirmation of Mendel's work. In the latter publication, on p. 71, 4th line from bottom, Abab should be Ab, ab; and aBab should be aB, ab. The following corrections should be made in the Report referred to above: p. 24. The offspring of S. inermis x S. armata should stand in the column headed "S. ar." p. 105, 2nd line. For "agree precisely, being 3.0:1" read "are 2.7:1." p. 160. Note. For "talls" read "Cupids."]

The Chair: We feel very greatly indebted to Mr. Bateson for his admirable presentation of these principles underlying fertilization. I am sure he has brought to each one of us here who has had any practical experience in this work the explanation of some difficulties that we have run against, whatever our work may have been. And 1 want to commend Mr. Bateson for the admirable presentation of a subject so full of information for us.

L. H. Bailey: Mr. President, I should like to say one word in regard to this matter of the Mendelian hypothesis. I have tried to follow it myself in this last year or two. I wish to say to you that if you wish to follow this with the greatest degree of accuracy you should get Mr. Bateson's recent book, "Mendel's Principles of Heredity." I don't believe that we shall get ready for a long time to formulate laws by means of which we may predict what is coming, because our premises are as yet in a way unknown. But it seems to me that the resuscitation and revival of Mendel's theories are going to open a whole new field to speculation in regard to the principles of heredity. It seems to me that the next few years are going to see a discussion of the principles of heredity in regard to plants that is comparable to that which followed Darwin's discovery. It seems to me it is as important as that. I expect to use this book as a basis for all our work in plant breeding.

William Saunders: Mr. President, this paper has thrown light upon many subjects which have been somewhat dark in my mind. For instance, in the cross-fertilizing of wheats we have often found that the crossing of two beardless forms will produce a bearded form, or we have a beardless wheat as the result of the crossing of two bearded forms. This explanation that Professor Bateson has given us throws light on that point and on many similar points which have puzzled many of us who are practical workers in this very interesting field.

H. H. Groff: The principles referred to by Mr. Bateson are certainly of great interest as regards primary crosses, and those crosses are related to our comparatively early experience in work of this kind. But the great question of interest to us in the future (and even now to those workers of extended experience) will be in regard to those crosses which are multiple to a limitless degree. These will present the questions in the future. It is not so much what we expect to find between hybrids or crosses containing a limited number of characteristics, but when crosses contain many thousands the problem will be far greater.

W. J. Spillman: I have with me some specimens, or rather some figures, of the specimens of wheat illustrating this law. I place them on exhibition.

The Chair: I could have presented from my own fields this season ten acres of illustration of Mr. Bateson's statement in regard to growing wheat. I have been growing a hybrid wheat for a number of years in a practical way as a farmer, and the seedsmen have taken the crops, and every year I have had to fight these bearded specimens of plants that came up in this field. To me it has been one of my greatest puzzles, as I was making no progress whatever; and while I never allowed one of those plants to go into my field, yet year after year I had the same result. I can see that it has been a bottomless work that I have been trying.