Memoirs of the Horticultural Society of New York (1902) 117-124.
L. H. Bailey

Horticulturalist Cornell University, Ithaca, N. Y.

Professor Bailey's manuscript was so long that he did not read it, but gave the general results of his work in extemporaneous remarks; and he then made a running comment on the significance of the work with pumpkin-like plants and the general meaning and tendencies of the new theories that are now occupying the attention of plant breeders.

His work with cucurbitaceous plants was begun in 1887, and was continued for ten consecutive years. Its original purpose was to determine whether there is an immediate influence of pollen on the fruit, a question then under general discussion; but the work soon grew into a general line of crossing and experimenting for the purpose of producing new types of fruits that might have value to the horticulturist. More than one thousand hand-crosses were made. Notes and photographs were made of the results. In one season eight acres of land were required on which to grow the progeny of the crosses. Altogether, some twenty-five or thirty acres were employed in the work. Many more than one thousand kinds of fruit, undescribed in any literature were produced. Nearly all of these forms  are yet shown in photographs. The very magnitude of the results has prevented their publication. To show the work to advantage one hundred or more illustrations should be made. However, it is doubtful whether it is worth while to publish the results in detail, because no underlying principles were discovered. The results were very remarkable, however, because of the great number of strange forms that were produced. Some of the results are published in the author's "Plant-Breeding."

Most of the experiments were made with the races of Cucurbita Pepo. Crookneck, Bush Scallop, Bergen squashes, the Field pumpkin and various ornamental gourds were oftenest used as parents. There was the greatest possible diversity in the progeny, in most cases no two plants bearing the same kind of fruit. In the second and third generations part of the progeny was grown from plants again hand-crossed and part from plants that were left to themselves. In some cases the plants were inbredthat is, the flowers were fertilized with pollen from another flower on the same plant (Cucurbita is monoecious). There were no essential likenesses or unlikenesses between these various categories. Even the progeny of inbred fruits was as various as that from cross-bred fruits. In no case was there any immediate influence of pollen, or xenia. In very many cases the progeny showed marked [118] characters that were wholly lacking in either parent. These new characters were unusual colors, shapes, wartiness of the fruit and attributes of vine.

Hybrids of two species.—In all the work with Cucurbits, numbers of attempts were made to combine the three species, C. Pepo, C. maxima, C. moschata. It is a common notion amongst gardeners that these three intercross interminably. All efforts, however, to combine the three species have failed, and the speaker is convinced that under common garden conditions none of these species habitually hybridize.

He became convinced, however, that it is possible to amalgamate C. moschata with C. Pepo, and a definite result was secured in this direction in 1892. The result of many pollinations was seven fruits of the following progeny:

In al these crosses there was no immediate effect of pollen. In four of these fruits, although the fruits themselves were well grown, there were no perfect seeds. In some cases the seeds were full grown and plump, but they were empty. Only three fruits gave seeds. These were the gourd crossed by the Japanese field pumpkin and two fruits of the Field Pumpkin crossed by the Japanese Crookneck. All these were crosses between C. Pepo and C. moschata. From the Field Pumpkin crossed by the Japanese crookneck fruits, eighty-eight plants were grown. These fell into about eight types, although there were only four or five well marked forms. Most of them were like a small orange pumpkin. Some were small green pumpkins. None of them showed any influence of the staminate parent, the Japanese, except that in a few the scar of the blossom end was very large, which is usually not the case in the varieties of the pure Cucurbita Pepo. One of the forms simulated a bush scallop squash of light lemon color. One was striped. All of these forms had the fruit stems of Cucurbita Pepo.

The details of these progeny (of Field Pumpkin by Japanese Crookneck) are given in the following notes:

Crosses were made in 1893 between some of these crosses themselves, care being taken to choose the pistils and pollen from plants that bore very similar fruits. Of all these crosses only one fruit matured.

An interesting result of these experiments was the fact that squash and pumpkin flowers are nearly always infertile with pollen borne by the same vine. Over two hundred careful tests were made on this subject with more than fifty varieties of pumpkins and squashes. Out of the whole number, only seven fruits were obtained that had good seeds. In most cases the ovary failed to develop. In some cases the ovary remained alive for some days, and it enlarged to two or three times its size at anthesis; but in most cases it finally perished, beginning to die away from the blossom or pistil end. In some cases, however, the fruits matured, being to all appearances normal, but they were usually empty or produced hollow seeds. In one experiment with five varieties of Cucurbita Pepo, representing both summer squashes and gourds, one hundred and eighty-five flowers were hand-pollinated with pollen from the same plant. All but twenty-two of these flowers failed to develop their ovaries. These twenty-two fruits grew to full maturity and appeared to be normal squashes in every way. Some of them, however, were wholly seedless, the seeds being represented by very small, undeveloped seed-coats. In a few others the seeds appeared to be good, but when they were opened it was found that they had no embryos. Of the twenty-two fruits that came to maturity only seven bore good seeds, and even in some of these the seeds were very few. All the seeds of these seven fruits were sown for the purpose of determining what the effects of inbreeding would be. It was found, however, that the progeny was just as variable as that grown from crossed seeds. The record of the progeny of these seven fruits is as follows:

  • Fruit No. 1. Four vines were obtained from seeds of this fruit, with four different types, two of them being white, one yellow and one black.
  • Fruit No. 2. Twenty-three vines. Fifteen types very unlike, twelve being white and three yellow.
  • Fruit No. 3. Two vines. One type of fruit which was almost like one of the original parents.
  • Fruit No. 4. Thirty-two vines. Six types, differing chiefly in size and shape.
  • Fruit No. 5. Twenty vines. Nineteen types, of which ten were white, eight orange, one striped, and all very unlike.
  • Fruit No. 6. Thirteen vines. Eleven types, eight yellow, two black, one white.
  • Fruit No. 7. One vine.
  • Very unusual crossings sometimes resulted in the production of apparently good fruit. For example, a bush scallop squash crossed with the pollen of a cucumber produced a fruit to all appearances normal, but it was empty. In some of the hybridizations between the different species, as between Cucurbita Pepo, C. moschata and C. maxima, the same result was secured (as already noted). He was not positive whether these pericarps were made to grow to their normal size through the influence of the foreign pollen, or whether there may not have been other influences at work, as there is in the case of the hothouse cucumbers, fruits of which will develop to large size without any pollen whatever. However, many tests were made by withholding pollen from the flowers, but in no case did the ovary develop to any size.

    It is a common notion amongst gardeners that nearly all kinds of cucurbitaceous plants mix interminably. It is a common opinion, for example, that muskmelons are rendered insipid and worthless when cucumbers are growing in their vicinity. Close observation in the field will convince any person of the fallacy of this idea, but experiments were undertaken for the purpose of testing the matter. Muskmelon flowers were pollinated from cucumber flowers, both in the house and in the open. In one case ninety-seven muskmelon flowers of various kinds were pollinated from cucumber flowers of various kinds, but no fruits developed. Twenty-five cucumber flowers at one time were pollinated by muskmelon pollen, but only one fruit developed, and that was seedless. These experiments and others coincide with those made by other investigators, that cucumbers do not spoil melons.


    After giving the general results of crossing the cucurbits at some length Professor Bailey spoke of the bearing of this work on the recent discussions concerning hybridization, and also of the trend of recent evolution literature. On the surface, all the experiments with pumpkins and squashes seem to run counter to the results secured by Mendel with peas and other plants. As a matter of fact, however, the work with the squashes is not comparable with that of Mendel, since different objects were in view and different methods were employed. Mendel's work was conducted with specific differentiating characters, whereas this work with the pumpkins was concerned with the gross behavior of the plants and the gross characters of the fruits. It is possible that if the work were to be done over again, with Mendel's methods and results in view, the same laws would be found to hold with cucurbitaceous plants. However, it would be a very difficult matter to determine, because of the instability of the cucurbits, the fact that they are monoecious and that constant crossing therefore is necessary, and the fact that so many variants would need to be contrasted. The subject is far too complicated for Mendelian methods until one has thoroughly mastered the simpler forms of hybridization experiments.

    The work of Mendel, so recently revived, has two very important general bearings. In the first place, it is bound to set going a new discussion in respect to hybridity; second, it will be likely to revolutionize our methods of performing hybridization experiments, and of casting up the results of them. Whether or no Mendel's rules will hold good for all plants and for all characters is not yet known. The probability is that it will not. The very fact that Mendel chose his stock plants with such great care, selecting species which are relatively invariable, that do not intercross, and that he eliminated the weak and abnormal plants, would tend to give uniformity in the results. We are in danger of becoming partisans. Professor Bailey the remarked that he neither believed nor disbelieved in Mendel's laws. He desired only to know what the truth is. He thought that future experiments should be carried on along the lines suggested by Mendel, and not for the purpose of proving or disproving his conclusions.

    It has recently been said that the time is rapidly coming when we can predict the results of hybridization with certainty, and can produce new varieties of plants with almost no element of chance. This hope is far too sanguine. Mendel's laws come from a contrast and comparison of specific differentiating characters. It is not so much a contrast of plants as a contrast of single characters of those plants. In ordinary crossing it will often be impossible to secure plants that have differentiating characters. What the plant breeder wants is a plant in its entirety rather than a plant with  specific attributes alone; that is to say, it may be possible to secure some character that is wanted, but with this desired character undesirable ones of other kinds may be associated. Furthermore, Mendel's results show that the offspring of hybrids are not intermediates or new kinds, but that they are controlled by the characters of one or the other of the parents, so new forms may not arise as a result of crossing. Every plant has unknown and unrecognizable characters, attributes that we refer in a loose way to the "constitution" of the plant. Moreover, one does not know in advance what characters will become dominant and which will be recessive. In other words, Mendel's law must be applied and discovered for each kind of plant; and the probabilities are that the results will be considerably modified by the conditions under which the plants grow. Again, the uniformity in Mendel's results was secured by the average totals of a great number of plants. The individual plants often varied widely in the very characters which in the average totals were relatively invariable in behavior. Now, the starting point of a new variety must be one individual plant, and not the average of a hundred or a thousand. The general results may be predicted with some degree of certainty, but how the individual plants will stand with reference to that result will be unknown. The practical value of Mendel's work to the actual plant breeder is yet in doubt. but the value of these remarkable experiments in elucidating our notions of hybridity, and in sytematizing experiments, may be beyond calculation.

    Professor Bailey also spoke of the recent philosophy of De Vries and his associates. Heretofore our thought has been dominated very largely by the Darwinian principle; that is, it is supposed that great differences may come about because small differences are enlarged by means of natural or artificial selectiona variety may become more of a variety. The new notion is that the important and permanent forms of plants come about as sudden sport or jumps, and that the small individual variations are incapable of growing into large and permanent varieties by means of natural selection. De Vries does not deny the power of natural selection, but he believes that its range is limited, that it cannot give rise to species, and that it does not contribute to permanency; as soon as selection is discontinued the form again breaks up or reverts. De Vries' theory of mutations is, in a way, a rephrasing of the old idea of sports. It differs in some essential points, however. One is in the supposition that plants mutate or sport in periods, and that in the intermediate epochs they are only making ready for another mutation period; that is to say, there are non-mutation periods and mutation periods. In the pre-mutation periods, be they long or short, the plant produces incidental individual fluctuations or variations, but the great progress in variation is made in the mutation periods. This body of belief is bound to challenge our accepted notions and our way of looking at the organic creation. This, together with Mendel's suggestions in respect to heredity, promise to awaken the liveliest discussion during the next few years. The speaker thought it probable that when these discussions shall have passed their first stage of enthusiasm we shall return to the Darwinian hypothesis, although he doubted whether we should ever hold to it so completely and so strenuously as we have in the past. We are bound to make distinctions between the kinds of varieties, classifying them either as individual fluctuations and mutations, or, from another point of view, as quantitative and qualitative. In other words, it is probable that there are varieties and varieties, and that not all of them are destined to have the same influence on the phylogeny of the race.

    The general trend of the discussions at the meeting, he said, seemed to be too exclusively along the line of hybridization, as if there were no other means of breeding and improving plants.

    *See page 115.

    Mr. O'Mara's remarks* were heartily secondedthe fact that good care on the part of the grower is often more important than the variety merely. Often a good variety may become a poor one, or a poor one a good one, by the exercise of skill in the growing of it. Plant breeding alone cannot improve our cultivated plants. It must be combined with all good care.

    O. F. Cook: I wish to raise one question, because I think that we should give our predecessors credit for standing where they stood, in order that we may not accuse them later of holding things which they didn't hold. We had, I think, a very conspicuous instance in the case of Darwin. Darwin thought a great many things and was not nearly as sure of a great many of them as many of his successors have been and he is now frequently accused of having made mistakes which he never made, but which he is accused of having made because they were made by other persons who have taken the responsibility of representing him. I fear that this will be to a considerable extent the case with Mendel. He took the precaution, I find, for which he deserves all good credit, of saying when he announced his so-called laws, that these were things that happened with the peas which he cultivated and in his garden. He did not say that they applied to all creation, or any other part of creation. He raised the question why in two or three hundred experiments they did not all work out the same way; he freely admits, and he leaves the matter entirely open. It seems to me that it is hardly fair even to talk about Mendel's law until we have reason to believe that it is a law and that it is at least of wide application. It may turn out to be very much like many discoveries in physiology and other new sciences, which are made to apply to the cases where the original investigation was made, but that may not have any very wide application.

    W. Bateson: It gives me great pleasure to listen to the paper of Prof. BaiIey, which I am sure we all feel was most stimulating and enjoyable. I should like to say few words about the application of Mendel's law to these more complicated cases such as those of the squashes and pumpkins which were made the subject of his paper. I am sorry if any one who had heard that paper were to go away with the impressionI am sure Professor Bailey would not wish to give that impressionthat because of the great complexity of the results given by crossing the squashes and pumpkins and allowing their offspring to cross and the difficulty of classifying the offspring so produced, that therefore such a case was contradictory or in any way beyond the scope of Mendel's law. There is no reason so far as I can see to suppose that. If I had time, I could give you a number of cases that we do know are not included in the scope of Mendel's law, but such a case as this, on account of the great diversity of the offspring, is no evidence whatever that Mendel's law does not apply, for the following reason: Mendel's law in its original form is dealing with a statement of the results obtained with hybridization of simple characters. For instance, you cross together the green pea and the yellow pea; the germ cells of the hybrid will form themselves purely yellow or purely green. We are dealing with single characters that were put in with the parents. But Mendel's law deals with a more complicated, and, to the practical man, far more important group of cases than that, where the parental characters that are put in are not simple, and cases in which the hybrid when it comes to form its germ cells does not form the parental characters simply, but divides those parental characters into what, for the want of a better term, we call their components. For example, the color of these squashes and pumpkins. There is not the slightest doubt that these in other cases would not follow the simple rules of Mendel's creation. For example, these colors may consist not of one simple character, but six or eight or ten or more component characters. The shapes of the squashes again in all probability consist of at leas six or eight component characters. When you come to observe that each plant that you obtain can only take one of each of those components from one parent and one from another, you may have combinations of an immense number of different entities taken two together, so that the complexity reaches a degree that is always beyond the reach of experiment. We cannot infer from those facts that Mendel's law will not apply. It is simply that the enormous areas which must be under cultivation when we are dealing with such an immense number of characters make it practically impossible to draw any conclusions. One word with regard to the point of the cytological investigations that were told of previously. I am afraid there is a little difficulty there for this very reason in regard to the complexity of character: there is a little difficulty in the way of ever hoping to analyze ultimately by the microscope the characters in the way that Mendel's law teaches us to believe they night he analyzed. Because it is very true that in Ascaris and in a number of other forms referred to we have reason to believe that the chromosomes of the father plant and mother plant side by side represent blocks of parental characters, that is not enough to help us to trace out ultimately the different parental forms of gametes. To do that you would have to have particles representing each parental character, not merely the whole block of chromosomes representing the father plant and the mother plant; you would have to have fragments representing each of the constituents of the father and each of the constituents of the mother, and they would again combine in the various combinations that we must expect. Complexity itself is no bar at all to the application of Mendel's law.

    With regard to one point that the last speaker made: he said that Mendel's law is possibly a thing of small range of application, or, at all events, cannot be asserted to be of universal application, and consequently it may apply to comparatively a few things. It may he of interest to those who are not perfectly acquainted with those investigations if I just briefly run over the kind of characters that have been shown to follow Mendel's law. For instance, there were Mendel's seven original characters, shape of plants and characters of seeds and pods, carried into several details; then there are the animal cases; we know that it applies to the shape of the combs of fowls, to the extra toe in several races of fowls, we know that it applies to the colors of fowls; we know that in mice it applies to the curious waltzing habit of the Japanese mouse, that character involving, we may say, almost mental attributes. We know it applies to the whole series of colors into which the whole series of rats has been broken, and to, I may say, twelve or fifteen different colors in plants. I think it is not too much to say that it applies to almost every case where the test has been possible of application.

    N. L. Britton: The fruit of Cucurbita was the subject of experimentation. Even in the wild species of Cucurbita there is a very great diversity of form and size of fruits, as is well known, and I think it is possible that that might have lent difficulty.

    The Chair: Professor Bailey was perhaps somewhat unfortunate in the subject of his experiments.

    L. H. Bailey: I believed, and I became convinced before I came through, that I had got hold of the wrong topic. It was too large for me, and I believe that one to take up the discussion of the Cucurbits and variation through hybridization has got first to be well grounded in many simple things. And from the point of view that I occupy now I believe it is one of the last things for a man to take up to work with. I quite agree with what Mr. Bateson has said, that if I could work the thing over I might be able to discover some kind of law governing these facts, but looking it over now, I can't do so. All I know now is that I got lots of things, I don't know how.

    T. V. Munson: I don't know whether I can say anything that would be of further value on this discussion, but it does occur to me that we are apt to spend a great deal of time in discussing theories without arriving at any solution. It has come to me in my own work that the matter of hybridization is entirely too extended for us to begin to establish a general law. There are some intimations of existing laws in the work, but we find that whenever we begin to discuss those so-called laws, as Mendel's, we end with a great deal of pro and con discussion without solution, without a satisfactory conclusion. It appears to me like this: that everything in the entire organic worldI might make it universalwith reference to form, is the result of environment, and in that I include the subject itself as a part, a small part, sometimes a very large part, of the environment, and what we are looking at is what has produced this result. 1t is a result that has partly come out of the individual under view and the effects of the surroundings upon that. Now, in working upon plants, I think we should not confine our views entirely to a biological standpoint, but that we are all the time tracing chemical influences. There is a chemical laboratory in every plant and chemical changes taking place. Each variety of fertilizer given the plant produces its effect in taking up and elaborating the substances, and in carrying on those chemical changes it evidently brings about its own result. So that we might say there is one general law. It seems to me  that in every direction I have observed in my work the result has come out of a of surroundings of the environment. And when we study the environment, the soils we are using, the moisture in the atmosphere, the temperature, all those conditions have some influence. If we undertake to make a general application of such a law or so-called law as that of Mendel, which is, I think, a very small lawlaws are of different capacities; some reach very far and others but a very little distanceMendel's law applies to pure seedlings through several generations. Suppose, instead of using pure seedlings, we continue to hybridize each generation, and we have many other ways of producing varieties, applying other influences. This, then, has a very narrow limit to the hybridizer wishing to employ all the different influences that he may see fit to employ. So that, in spending time upon discussing this one small law, or a very short reaching law, we may overlook more general laws, and we shall be more likely to reach results if we collect facts, put them where the experimenter can use them, and show the extent to which we have absolutely proven. Then, like Kepler, after a time (and I don't think that time has yet arrived to establish any great number of laws in hybridization), we can probably draw some more general laws. Then I think our most practical direction in which to work is to collect facts, strive always after something that is useful, something that is practical, and make notes of every influence, everything obtained, and on these facts will grow laws that are exceedingly valuable. It is true we want laws; we want something by which we can guide ourselves in this work; but if you try to make a law out of a mere theory which is only set up for experimentation you are wasting time.