Volume Number: 1  2  3  4  5  6  7  8  9  10  11  12 




In farming districts of the Mississippi Valley they have a curious custom in selling cattle at auction. They drive a herd of cows together and the auctioneer asks his audience to bid for first choice, no individual animal being specified. The highest bidder makes his choice, and the cow he selects is taken from the herd. Then the auctioneer starts over, receiving bids for "first choice" among the remaining animals. This process is repeated again and again until all the exceptional animals have been selected. A curious result of the method is that it very commonly happens that different bidders have their eyes on different animals. Farmer A, who bid highest at the outset, did not have in mind the animal for which farmer B was bidding. And so it often happens that after six or eight selections have been made a cow still remains that was regarded by some of the bidders as the very best one of the entire herd. A man who bid unsuccessfully again and again may thus, in some cases, finally have his choice precisely as if he had made the highest bid at the outset. The obvious explanation both of the method and of its somewhat anomalous results is found in the fact that individuals differ in their judgment as to what constitute the superior qualities of a cow. Each bidder has noted an animal that particularly appeals to him, and each is backing his own judgment in making selection. The result is a process of elimination that may or may not select from the herd the best animals at the very outset. "But what have cows and their selection to do with the development of new varieties of plants?" you ask. Nothing direct and obvious to be sure. But it has often occurred to me that the process of selection at the Iowa auctions is closely comparable to that which is employed by the plant experimenter in the course of his every-day work. In lieu of a herd of cattle, he deals with a group of seedlings. But his task is precisely like that of the auction bidder in that he must select from among scores of plants of the same kind, and often of closely similar appearance, the one that seems to him the choice of the entire lot; and then in succession the second and third and fourth best, until he has chosen possibly six or eight individuals out of a group of hundreds or thousands. These six or eight individuals will be preserved for use in further experiments. They are the ones with which the attempt to improve the variety to which they belong will be carried out. And the ultimate success of the entire experiment in plant breeding will very largely be determined by the perspicacity with which the selection of these few individuals was made. Nor can we doubt that it must often happen, in the case of the seedlings as in that of the cattle, that after the final selection has been made there remain, unknown to the experimenter and in contravention of his judgment, better plants among those rejected than any one that he has chosen. It could not be otherwise when we consider the large numbers involved, the variety of plant characteristics, and the great diversity of traits represented in a single generation of hybridized seedlings. Yet, on the other hand, experience should enable the experimenter to choose with a relative degree of certainty, and it is possible to acquire a degree of skill, based on careful observation of the minute details of plant structure, that will give full assurance of a capacity to select with at least a large measure of success.


It is usually a surprise to any visitor who comes to my orchard at a time when I am making selections among seedlings of many kinds to observe my method. Many people have expressed astonishment when they have seen me walk rapidly along a row of plum trees saying: "Kill this one, and that one, and that; save this one, and that one yonder"; indicating the choice between plants to be saved and those to be destroyed so rapidly that the men following me can scarcely tie strings to the selected ones as fast as they are chosen. In this way I may test from five to ten thousand young trees as I walk along the row, scarcely pausing for more than what seems the most casual glance. But my eye takes in the important thing. I know just what I am looking for. And if my judgment in the matter had not proved in the main good, the output from my orchard would have been quite different from what it has been. I may recall by way of illustration an experience in which my selective judgment was put to a practical test-no different a test, to be sure, from thousands that I myself have made, but having added interest because it was made by another. It chanced that a well known judge, who is also a horticultural enthusiast, who had been very much interested in my work, was visiting me at a time when I was sorting out plum trees from among a lot of several thousand seedlings about a foot high. I had a man carrying them away as fast as selected. They were thrown in three piles, the first containing those I had declared to be the best ones for continuing the test; the second pile containing those I thought possibly worth trying; and the third pile those that seemed to me no good at all. The judge watched me for a few minutes and then said: "You are picking them altogether too fast. You cannot possibly tell like that which are good and which are not." I replied: "Wait and see, or test the matter for yourself if you wish." "Very well," said my visitor, "I will do so." And therewith he selected a few seedlings from each of the piles and took them home with him to graft on trees of his own. Of course it was necessary to wait two or three years for results. But when the time came, the judge very cheerfully admitted that I had been quite right all along the line. The cions from my discarded pile bore fruit that was almost worthless; those from the intermediate pile gave fairly good fruit; and from the pile of my first choice seedlings he secured a fruit of such quality that he named it the Klondike, declaring that it gave him more good plums than he had ever had before from a similar tree. I cite the incident as showing the possibility of gauging fruiting qualities of a seedling at a time when the plant itself is a mere sapling a few inches in height. The capacity to make such selection has sometimes been spoken of as intuition; but it is really a matter of observation and practice. One learns through long experience to judge what characteristics of the seedlings are suggestive of possibilities of fruit-bearing. And after all this is no more than judging the man of the future by observation of the child of today.


If we were to state the matter a little more technically we might say that such selective judgment as I have just illustrated is based on a knowledge of the correlation between the different parts or members of a plant's organization. It was first prominently brought out, I believe, by the French naturalist Cuvier something over a hundred years ago that there is always a correlation between the different structures of a given animal, to accord with its habits of life. For instance the teeth and claws of a cat are associated with its carnivorous habits and are linked with a certain structure of legs and muscles adapting the creature to spring forward with great celerity upon its prey. A somewhat different structure of body and limb is associated with the talonless feet of the dog tribe which are adapted to rapid running for prolonged periods rather than to sudden leaping and clutching. It was by careful study of the correlation, of parts, of which these are only crude and familiar examples, that Cuvier was enabled to gain an insight into the characteristics of fossil animals of which only small fragments of skeletons were preserved in the rocks. The science of comparative anatomy was the outgrowth of his observations. Now it is at once obvious to anyone who studies plants attentively that their structure also shows a corresponding and no less invariable correlation of parts. The more conspicuous illustrations of this are obvious to the most casual observer-various adaptations of form of tree and shrub and vine to their natural surroundings are so patent that they cannot escape attention. But of course the plant experimenter must deal with correlations of a very delicate order. He is called upon to make nice distinctions between individual seedlings of the same variety. All will have the same general formation of stem and leaf. He must look, then, for details of variation that would altogether escape the notice of the untrained observer. But that such differences exist, and that they are signs that to the practiced eye are of the utmost importance, any successful plant experimenter can testify. It would obviously be futile to attempt a detailed description of the nice shades of distinction between various seedlings of the same race upon which the plant experimenter depends in forming his selective judgments. That, clearly, must be matter for practical observation. It can be learned nowhere but in the field. But perhaps two or three illustrations may be given that will at least serve in a general way to suggest what manner of traits are taken into consideration when the plant experimenter is choosing the individuals with which he is to continue his experiment.


In selecting raspberry or blackberry plants for color of fruit, for example, there is almost always a correlation of the plant and fruit that will foretell the future crop. I have observed in thousands of instances that vines that have purple spines and canes will in future produce berries that are dark purple or dark red in color. Pinkish leaves, on the other hand, foretell fruit of light pink or red color; plants with yellowish vines and foliage may be expected to produce berries of a yellowish color. Very pale foliage and canes usually indicate that the crop will be of a whitish or amber color. A knowledge of this correlation between vine and fruit was of great service to me in my later experiments for the development of the race of white blackberries. It enabled me to select for transplantation and particular care vines that would produce the type of berry I was seeking. It was not necessary to await the time of fruiting in order to gauge progress. The correlation of characters between the vine and the fruit of the grape is not always quite so clearly established, yet it is often observable. Grape tentacles may give clear indication of the size and flavor of the future bunches of fruit. Long before a grape vine has come to the age of fruiting, the taste of the tendrils may give a fair idea of the flavor of the grapes it will ultimately bear. Moreover the seedling vines that produce bushy stems that are small and much branched, and have small leaves, will almost invariably produce meager clusters of small fruit of poor quality. So the wise experimenter will root out such vines without letting them come to maturity. Among plums and peaches the correlation of characters is exceedingly valuable. The case of the plum seedlings already cited suggests the possibility of pre-judgment of fruit from observation of small seedlings. There are a good many characters of leaf and twig that are almost too intangible for description, like the changing expressions of the human face, or like delicately graded colors, yet which to the practiced eye are full of meaning.


A broad general distinction that is fairly obvious to any observer is found in the color of the foliage. It may be expected that a plum or peach seedling having foliage of a reddish purple color will produce fruit dark-colored not only in skin but in flesh. And of course the selection made from any given lot of seedlings will depend largely upon the particular qualities that one desires to develop. But, as repeatedly pointed out, in practical work one is usually looking for a combination of qualities; and, by the same token, one usually inspects his seedlings for the combination of characteristics of stem and leaf and color. He seldom has his choice determined by a single characteristic, obvious or otherwise.


Yet there are cases where an experimenter is working with a single plant-characteristic in view, as, for example, when I successfully attempted to develop scented callas and dahlias and verbenas. Here, obviously, the task of selection is comparatively simple. We are dealing in each case with a flower that has certain desired qualities of color that are firmly fixed in its heredity. The one conspicuous point of variation among thousands of specimens is the presence or absence of a pleasing aroma. It is necessary, then, merely to select the individual plants that have the most pleasing perfume and to use these only for carrying on the experiment. By making such selection generation after generation, choosing always the sweet-scented and rejecting the others, it proves possible to accentuate and fix the quality of perfume-production without altering the other characteristics of the respective flowers in question. Again the quality sought may be a particular color of blossom, and it may be desirable to pay attention to this only, practically disregarding all other qualities. Such, for example, was the case with my experiments with the crimson Eschscholtzia, commonly known as the California poppy. The blossoms of the plant from which my new type of poppy was developed, had a narrow strip of crimson on the inner side of one petal. This was an anomaly that appeared "spontaneously." Doubtless it was due to some crossing of ancestral strains that brought out a latent character that had long been suppressed. But as to this we can only surmise. The simple fact of the matter was that a blossom did appear that had this narrow strip of crimson on one petal. I seized on this individual blossom as offering material for an experiment in color variation. Seeds from this plant produced the next year several plants that had a trifle more crimson on their blossoms. The following year there was still further improvement, as plants appeared that showed a much larger invasion of the flower petals by the crimson coloring. And by selecting year after year blossoms that showed this increasing tendency to adopt the new color, I produced presently a plant that bore blossoms of a beautiful uniform clear crimson. No trace of the original color remained. This furnishes a very good illustration of selection for color where the material consisted of a small strip of an unusual color appearing on blossoms otherwise of a fixed hue. But the same method of selection may sometimes be applied to the improvement of the shade of color, or even to the development of a new color, from a flower that shows only a faint departure in shade from the normal. And the same principle of selection, followed out in precisely the same manner, applies to the development of flowers or fruits of varying size, of larger or smaller stem, abundance of blossom, profusion of leaf or flower or fruit, and the like. It is equally possible to alter the proportions of the chemical constituents of a plant in certain instances. The case of my sugar prunes, which were developed to have a sugar content of more than 23 petcent, as against the 15 percent, of their ancestral type, will be recalled. In a similar way the sugar beet has by mere selection been developed until the races now cultivated contain several times the proportion of sugar of the ancestral beets even of twenty years ago. An interesting experiment in causing the progeny of a certain plant to vary in opposite directions through selection, has been made at the Illinois Agricultural Station. Here the quality under consideration was the protein content-that is to say the amount of nitrogeneous matter-in the kernels of a given variety of corn. The specimen with which the experiment started showed on analysis 10.92 percent of protein. Selection was made, among the ears of corn grown from this seed, of the individual specimens having the highest protein content on one hand, and those having the lowest protein content on the other. By continuing this double selection for ten generations, two races of corn were developed, one of which produced seed having an average protein content of 14.26 percent, while the other, grown in the same field, showed a decrease to 8.64 percent. This experiment illustrates the possibility of selecting out and fixing new races varying widely as to a single important quality of grain among the descendants of a parent plant of relatively fixed strain. In point of fact no plant is so fixed that its individual members do not show variation; none so fixed that it does not supply material with which the experimenter may work in producing new varieties. Another illustration of the same thing was given by an allied series of experiments at the Illinois Station at which selection was made with reference to the height of the ear on the corn stalk. Seed from the same cob was planted in two fields and grown always under closely similar conditions. But in one field selection was made for breeding purposes from stalks having the ears higher from the ground than the average; and in the other field from ears that were lower than the average. At the end of five years the two fields were so widely diversified that the average height of the ear from the ground in one of them was less than three feet (33.2 inches), whereas in the other field the average height of the ears was fully six feet (72.4 inches). One could not well ask a more striking illustration than this of the possibility of developing new races, differing as to some conspicuous character, by simple selection from a given stock. The case of my winter rhubarb, which came to have a relatively gigantic stalk, will be recalled as of similar import; although in that case the experiment was complicated by having to bear in mind various other qualities in addition to mere size of stalk. My giant corn and the corn with the rainbow-striped leaves are other examples.


But, as repeatedly pointed out, the experiment usually is complicated by the necessity for considering more qualities than one whenever selection is made with an eye to the production of a commercially valuable variety of flower or fruit or vegetable, and not merely for the purpose of scientific record. We have seen this illustrated again and again; and we have seen also how great are the complications which result when we are called upon to make a selection that will give us not merely one quality-merely size or a given color or sugar content-but a combination of six or eight or ten qualities, all presented in superlative measure. We have seen that the chance of securing any given combination of qualities decreases at a startling geometrical ratio in proportion as the number of qualities increase. The precise formula, as calculated by the biometricians, runs something like this. In case a single pair of qualities is in question-say high protein content versus low protein content in corn-the chances are, if the two strains are crossed, that there will appear in the second generation of their progeny one offspring in four that closely resembles each parent. But when we are considering two qualities-say protein content and height of ear on the stalk-in combination, the chance that there will be an individual of the offspring like each parent in the progeny of the second generation, is only one in 16. And when three qualities are in question the ratio jumps to one in 64; with four qualities it advances to one in 256; with five qualities, to one in 1026. When eight qualities are in question, the chance of producing one offspring showing precisely the combination of qualities of each parent is only one in 98,496. And when we deal with ten qualities we encounter the altogether disconcerting ratio of one to 1,575,936! All of which makes it very clear that the wise plant experimenter does not depend upon mere chance to give him the combination of desired qualities in the production of a new form of flower or fruit. He mut make his selection, in any given generation, with reference to one or two pre-eminently desirable qualities, and must be content to accept for the moment such other qualities, however undesirable, as are associated with the desired ones.


For example, in developing a stoneless plum, my earliest selections were made with an eye to stonelessness alone. Then as I gradually developed a race of plums in which I was certain of finding a fairly large proportion of individuals growing stoneless fruit, I could select among these the ones that combined with stonelessness the largest proportion of other good qualities, such as size and color and flavor and abundant bearing. When presently I had, through selection, developed a somewhat fixed strain that combined the qualities of stonelessness with fair size and good flavor, I could then select among the many individuals showing these qualities the particular ones that showed them in fullest measure; and at the same time I could now have in mind one or two other qualities-say color of fruit and keeping quality-and be guided in my selection by a consideration of these traits in addition to the others that had already been fairly fixed. Thus the matter of selection, even when many qualities are to be combined in the ultimate product, is not quite so hopelessly complex as the calculations of the biomnetricians might lead one to suppose. Yet it is assuredly complex enough to test the patience and the ingenuity of the experimenter to the last degree. So the amateur who enters this fascinating field will do well to begin with simple cases, paying heed to a single quality of any flower or fruit with which he experiments; endeavoring to advance along one line till he acquires skill enough through practice to attempt more complex experiments. Let him, for example, increase the perfume of some familiar garden plant, or develop a race having large blossoms, or one having peculiar brilliancy of color. Any flower bed will show him, among different specimens of the same species, enough of variation to furnish material for his first selection. And he is almost sure to find encouragement through discovery, among the plants grown from this seed, of some that will show the particular quality he has in mind in a more pronounced degree than did the parent plant. So here he will have material for further selection, and step by step he can progress in successive seasons, often more rapidly than he had dared to hope, toward the production of the new variety at which he aims. Of course the time will presently come when the amateur who thus begins with what may be called the alphabet of plant experimentation, will wish to advance to more complicated projects. He will wish to urge his plants along a little more rapidly on the path of variation by means of hybridization. But even here, as will be obvious on a momnent's reflection, the experimenter is still dealing with selection. For of course he will not make his hybridizing tests at random, but will select for his parent stock individuals that manifest in pronounced degree the qualities that he wishes to combine in his projected new race. So when we pass from the stage of simple selection of "spontaneous" variations, to the stage of inducing variations along given lines by cross-breeding, we are not abandoning selection but are only dealing with selection in its more complicated aspects. Rightly understood, then, it is not too much to say that the entire task of the plant developer is a matter of selection. First, he may select varieties as nature presents them to him. Second, he may through selective breeding improve these varieties. Next he selects among these and makes combinations for further variations; and then he is ready for a new series of selections. So from first to last it is only the same story presented in different aspects. How important a part does selection play in the life about us! Whether it be in animal or human life, whether it be the selection of materials for a nest, an appropriate club or stone to lay low an enemy, the selection of materials for a dynamo or a pyramid, or of words to convey certain thoughts, aspirations or emotions; but selection alone from among all the materials supplied by nature no matter how skillfully carried out can never produce the artist's ideal in pigments or marble, the architect's vision of a great structure for the shelter of thousands-universal standards of excellence-unless their production is accomplished by means other than metrical and statistical! The beginning is selection and the end is selection.

This text is from: Luther Burbank: his methods and discoveries and their practical application. Volume 3 Chapter 9