View source for Breeding of plants

{{Inc|
Breeding of plants. The definite producing of kinds of plants adapted to given uses and conditions is known now as plant-breeding. The existing varieties are of course the result of the action of natural tendencies and laws, but the producing of them has not been, for the most part, a conscious, or at least not a regulated, act on the part of man. The laws of variation and inheritance are now beginning to be understood, and the application of this knowledge is to produce orderly and more or less predictable results.

In beginning the artificial cultivation of plants, our early ancestors, even with their crude understanding and methods, probably selected seed for planting from the best wild individuals of any plant. The selection of seed from the best individuals has thus been continuing for thousands of years, ever since the dawn of civilization. While this is a crude method of breeding, if long continued on an extensive scale, it could not, as is now recognized, fail to have results. The greatness of the changes produced is shown by the fact that some of the most extensively cultivated crops, such as wheat and maize, have been so modified that the wild types from which they sprang cannot now be recognized, although the original wild ancestors probably still exist.

Breeding did not become established as an art until comparatively recently. The sexuality of plants was not established until it was proved experimentally by Camerarius in 1691; and the first hybrid of which there is record was made in 1719 by Thomas Fairchild, an English gardener, who crossed the carnation with the sweet william. The first exact knowledge of hybridization dates from about 1761 when Koelreuter began publishing the results of his observations, but even his work had little bearing on practical plant-breeding. The systematic breeding of plants may be said to have begun with the work of Knight and Von Mons about the beginning of the nineteenth century.

Knight worked mainly in hybridization and in 1806 said: "New varieties of every species of fruit will generally be better produced by introducing the farina of one variety of pollen into the blossoms of another than by propagating from a single bud." Von Mons worked mainly in selection and it is interesting to note that his experiments were made primarily with pears. He emphasized continuous selection and produced very many valuable varieties. It is probable that a large part of the success of Von Mons work was due to the fact that pears are normally sterile to their own pollen, requiring cross-fertilization, and, therefore, many of his new varieties were probably hybrids. He was not aware of this fact, however, and it made no great difference in the establishment of the principle which has since proved to be so important.

A most important stimulus to the development of plant-breeding was given by the publication of Darwin's famous works, particularly his "Animals and Plants under Domestication," in 1868. His extensive researches, masterful compilation and systematization of the existing knowledge may be said to have established breeding on a systematic basis.

Following Darwin, little advance was made in the knowledge of the principles of breeding until in 1900, when Mendel's papers on plant hybridization, describing his now famous principles or laws of inheritance, were rediscovered independently and brought to attention by DeVries, Correns and Tschermak. The discovery of these laws and the publication of DeVries' "Mutation Theory" in the same year, marked the beginning of a new era in plant- breeding. No matter what the final conclusions may be regarding Mendel's principles and the mutation theory, the stimuja- tion which these two theories have given to breeding has already served greatly to modify and extend knowledge, both in scientific and practical directions.

The great advance that has been made in the discovery of the underlying principles of breeding puts experimentation in this field on a much surer basis and the breeder can now approach his subject with definite understanding.

Classification of varieties.

To understand clearly the character of organisms with which breeding deals, careful definitions of the different groups of cultivated plants which are ordinarily known as varieties are needed. One speaks of varieties of wheat, corn, apples and pears, yet it is known that these varieties differ from each other as natural groups. In order to distinguish clearly these differences, the following classification of varieties into races, strains and clons has been proposed:

Races are groups of cultivated plants that have well-marked differentiating characters, and propagate true to seed except for aimple fluctuating variations. The different groups of beans, peas, wheat, oats, com, cotton, and the like, referred to commonly as varieties, are thus in a more restricted sense races. Boone County White, Learning, Golden Bantam, and ?? on, would be recognized ?? races of corn, and Turkey Red, Fulcaster, Fultz, as races of wheat, and Early Paris, Dwarf Erfurt and Snowball as races of cauliflower.

Strains are groups of cultivated plants derived from a race, which do not differ from the original of the race in visible taxonomic characters. When the breeder, by a careful selection of Blue Stem wheat, produces a sort of Blue Stem that differs from the original race only in the quality of yielding heavily, it would be called a strain of Blue Stem.

Clons are groups of cultivated plants the different individuals of which are simply transplanted parts of the same individual, the reproduction being by the use of vegetative parts such as bulbs, tubers, buds, grafts, cuttings, runners, and the like. The various sorts of apples, potatoes, strawberries, chrysanthemums, and so on. commonly denominated varieties, in a more restricted eense would be clons. Clons of apples, pears, strawberries, potatoes, arid the jike, do not propagate true to seed, while this is one of the most important characters of races and strains of wheat, corn, and others. The term variety would thus be used in a general sense, and would include races, strains and clons.

Heredity.

The laws of heredity are of primary importance to the breeder. It is a general principle that like begets like, but it is also true that like frequently gives rise to unlike. In general, by heredity is meant the tendency which an organism manifests to develop in the form and likeness of its progenitors, and the study of heredity includes thus the inheritance of characters. It is of the utmost importance that organisms in general reproduce their kind, as otherwise the breeder would be confronted with confusion, but it is of equal importance that the offspring does not always reproduce the parental characters. There are thus apparently two conflicting principles in plant-breeding. On the one hand, the breeder seeks to produce variations in order to get new types as the foundations for improvement. On the other hand, when such a variation from or improvement on the normal type is secured, he reverses the process and tries to establish heredity and reduce the amount of variation, so that the aphorism, "like begets like,'' will hold true.

In pedigree- or grade-breeding, and in breeding to produce new varieties, the importance of hereditary strength cannot be overestimated, as it is only by rendering this power very great that any new form can be brought to what is called a fixed type.

In recent years, the ideas of fixity of type have been greatly modified, and it is now held that fixity of type is secured by purifying a race from all admixtures so that any character represented in a race will be pure.

Unit-characters.

The modern studies of heredity have led to a new conception of organic characters that should be clearly understood by the breeder. A careful study of species or varieties of plants or animals focuses the attention not on the generality of the differences existing but rather on the differences in certain characters; one observée whether a plant is smooth or hairy, cut-leaved or entire- leaved, much branched or simply branched, erect or procumbent, tall or short, and the like. This leads to the conception that a plant is not of simple organization but is comprised of a combination of characters. These characters or the physiological units which cause them are nc>w thought of as in considerable measure independent of each other and as representing distinct organic units. The classical studies of Gregor Mendel on the hybridization of races of peas that exhibited different characters established the fact that at least certain characters are inherited separately and may form permanent new combinations.

A unit-character, then, may be defined as any characteristic quality or set of qualities or expression of character in an organism that is inherited as a whole and independent of any other quality or set of qualities. They are the organic units of inheritance. The units that are considered in hybridization are not the species or varieties themselves, but the unit-characters of which they are composed. The origin of a new variety would then consist in the acquirement of a new character by the organism or the loss of an old character or of the production through hybridization of new combinations of characters that already existed but in different combinations.

Nature of variation.

While, as indicated in the discussion of heredity, organisms are usually reproduced in the likeness of their parents, nevertheless it is well known that all plants vary. Individual plants differ from one another just as do men. The fact that plants can be improved by selection depends upon the occurrence of these so-called variations. One is accustomed to think of plants as very stable and uniform. Casually looking over a field of ox-eye daisies and admiring their beauty, one distinguishes no apparent variability; all seem to be alike. Nevertheless, if the plants are examined carefully and the different individuals studied, it is found that each one possesses certain peculiarities. Some have large flower-heads, others small flower-heads; some have very many rays or petals, others comparatively few; some have broad rays, others narrow rays. Some plants are tall, others short. No two plants can be found which do not differ from each other in some noticeable character. They present different facial expressions, the same as do people or cattle, so that different individuals may be recognized after one has studied them and made their acquaintance. This is one of the interesting studies which the breeder pursues. Careful gardeners learn to recognize the individual plants that they handle day after day as the shepherd recognizes the different members of his flock.

The inheritance of a character ordinarily does not mean its exact expression in the offspring as in the parent. In considering variations from the standpoint of the conception of unit-characters, it must be remembered that only the determiners of a character are inherited and the expression of the character in the new individual is influenced by the environment under which the individual develops. It must also be remembered that in the higher plants and animals with which the breeder ordinarily has to deal, an individual results from a fertilized egg-cell which contains the heritage determiners of two parents and, as there are a very large number of characters making up any individual and as different individuals possess different determiners which are brought together in fertilization, rarely or never can one individual be conceived to be an exact counterpart of another.

Variations are of very great difference in magnitude and kind; and while many different names have been given to the different types of variation, the most generally accepted usage at present is to classify all variations either as fluctuations or mutations.

Fluctuations are those variations that are supposed to be due to the direct action of environment and that are not inherited. The variation in size as a result of richness of soil, is such a fluctuating variation and, as well recognized, is not a heritabje character. A similar illustration of such a variation is the difference in size of oat or wheat plants due to crowding in the field (Fig. 637). It is known that if a pole bean be transferred to the North, it tends to produce a bush type, and if a cowpea be transferred to the North, it tends to shorten up its vine and assume a bush habit. An interesting illustration of such modifications is shown in the ordinary red cedar, Juniperus virginiana (Fig. 638). In the rich, moist soils of Pennsylvania, Maryland and Virginia, this tree forms a beautiful tall columnar top with dense foliage (Fig. 638 o). On the dry, sterile, limestone hills of Kansas, Nebraska, and Kentucky, and in the sandy soil of Florida, the same tree produces a spreading, scraggly top of entirely different character (Fig. 6386). If one of these trees is transplanted while young, from sterile barren soil to moist rich land, it assumes the tall columnar habit as a result of the environment.

Plant-breeders have sometimes assumed that such modifications, which are the result of environment (Fig. 639), are of great importance to them. This matter, however, is in grave doubt. The information at command indicates that these characters, which are physiological adaptations, are not hereditary, and are lost as soon as the plant is transferred again to its normal environment. If, for example, it is desired to produce a bush cowpea and the selection is undertaken in the South with a viny variety, a search should be made among the plants for the individual that approaches most nearly to the bush type, and it is probable that this plant would be as likely to transmit this character to its progeny as a similar bushy type selected under northern conditions. As a matter of fact, it may be that this tendency could be recognized much more clearly in a southern location, where the plants normally produce vines, than in a northern location.

Mutations, on the other hand, are changes that are more profound and effect the germinal cells of the organism in such a way that the changes are inherited. The most typical illustrations of mutations are the striking large type-variations that are known to gardeners as sports, and which ordinarily reproduce true to seed. It must not be understood, however, that all mutations are large type-variations. This, it is true, was in large measure the meaning given to mutations by DeVries in his development of the mutation theory of evolution, but the moie general interpretation of biologists at present is to consider any type of variation that is inherited as a mutation. Many small variations, such as a slight difference in height of ear in corn, may be regularly inherited, and in some instances differences that are so slight as to be distinguished only by careful biometrical analysis are regularly inherited, generation after generation, even under very different conditions. Recent scientific studies have emphasized the great importance of such variations in the production of cultivated varieties and the evolution of species. As is well known to gardeners, these sports or mutations appear suddenly without warning or reason, so far as is known. They cannot be produced, and one must simply wait until they appear and then be prepared to recognize and propagate them. Mutations usually reproduce their characters without much reversion to the parental type except such as is caused by cross-pollination. Mutations of self-fertilized plants thus usually come true to type, while in cross-fertilized plants the mutation must usually be cultivated in an isolated place and carefully selected to weed out the effect of such crossing as has occurred. Many seedsmen examine their trial-grounds regularly for sports or mutations, and many of the best varieties have resulted from the selection of such sports. Livingston, of Ohio, who during his life was famous for the number of new varieties of tomatoes which he produced, made it a practice to search regularly the fields of tomatoes, which he grew for seed purposes, for such sports, and almost all of his numerous varieties were produced by the discovery of such striking variations.

A very interesting case of a variety that originated as a seedling sport or mutation is the now familiar case of the Cupid sweet pea. Until about fifteen years ago the only sweet peas known were the ordinary tall twining sorts which grow to a height of 3 to 6 feet, depending upon the richness of the soil. At this time there was found in California, a small dwarf sweet pea plant only about 6 or 8 inches high. This was growing in a row of the Emily Henderson variety, one of the ordinary tall sorts from which it evidently had sprung. Seed of this dwarf plant was saved and grown, and it was found to reproduce plants of the same dwarf character. The variety was designated the Cupid, under which name it was introduced to the seed trade and distributed over the world. The Cupid differed from other sweet peas not only in height but in its closely set leaves and general habit of growth. Indeed it is as distinct from other sweet peas as are distinct species of plants in nature. From the original Cupid, there have sprung many different sorts, until now there are varieties of Cupids representing almost all variations of color and shape of flower known in the sweet pea family.

Causes of variation.

Understanding of the causes of variation is as yet very imperfect. Fluctuations are in general interpreted as the direct physiological action of environment on the plant, or. in other words, environmental reactions. There would seem to be no doubt of the correctness of this view for the cause of ordinary fluctuations, and it may be accepted as the cause of such fluctuating variations as the breeder will commonly meet. Such reactions as the changes in structure and form of the entire air- leaves and finely divided water-leaves of certain buttercups (Ranunculus) and the floating and submerged leaves of pondweeds or Potamogeton (Fig. 640), and the loss of knees on the bald cypress when cultivated on high land where the soil is well aerated, may be interpreted merely as extreme environmental reactions. Even these extreme changes are not inherited other than that the ability to react in this way under different environments is inherited.

To account for mutations is, however, a much more difficult matter and no definite conclusion as to their cause has yet been reached. Lamarck and his followers have strongly maintained the hypothesis that changed environment would stimulate the production of variations that would permanently effect the organism and its progeny in the direction of better adapting them to their environment. Many scientists, even today, believe in the effectiveness of environment in developing adaptive changes. Weisman and his followers, however, appear to have shown that characters acquired through external influences, the so-called acquired characters, do not affect the germ-cells, which are early differentiated in the development of the organism, and are thus not inherited. While, in general, it is certain that the ordinary environmental reactions are not inherited, it is known that plants long grown under a certain environment become modified to suit that environment, and that such adaptive changes have in some way so modified the organism that the adaptive changes are rendered heritable. Thus the conclusion follows that in some way environment by its stimulation does occasionally affect the germ-cells and produce changes that are inherited. Plants that have long been cultivated under widely varying conditions almost invariably develop numerous heritable variations that would be classed as mutations. The older breeders strongly held to the belief that such conditions as change of food-supply, change of altitude, artificial cultivation, budding, and grafting, indeed the ordinary manipulation of agricultural cultivation, lead plants to vary in directions of importance to the breeder. Clearly, no problem is of more importance to the breeder than to be able to produce or cause such new characters to appear.

It is only very recently that the idea has developed that one can go farther than possibly to change the environment. With the publication of Mac Dougal's researches in 1906, describing mutations that were apparently caused by injecting the capsules of plants with certain solutions, such as zinc sulfate and magnesium chloride, a possible new method of forcing variations was introduced. Mac- Dougal apparently obtained marked variations as a result of his treatment, that were inherited in succeeding generations.

Tower, by subjecting potato beetles during the formation of the germ-cells to extremely hot and dry or hot and humid conditions with changes of atmospheric pressure, was able to cause the development of marked changes or mutations that were found to transmit their characters true through several generations and which segregated as unit-characters following hybridization. He concludes from his experiments "that heritable variations are produced as the direct response to external stimuli."
Gager has produced similar changes in plants by subjecting the developing ovaries to the action of radium rays, and a number of similar studies by Hertwig and others indicate that radium emanations have a very active effect on both plants and animals.

While the evidence favoring the value of such external stimuli as the above in producing new heritable characters is apparently definite and positive, the extent to which the method can be used in practical breeding has Dot been determined, and indeed further experience must be awaited before the evidence, or the interpretation of the evidence presented in these very valuable ana suggestive researches, can finally be accepted. Humbert has made experiments in which the capsules of a pure line of a wild plant (Silens noctiflora) were injected with the solutions used by MacDougal, and although the number of plants handled (about 15,000) was apparently as great or greater than was used in Mac- Dougal's experiments, no mutations were found in the treated plants that were not also found in the untreated or check plants.

Some observations and experiments are recorded in literature which indicate that mutilations or severe injury may induce the development of mutations. Most noteworthy among such observations are those of Blaringhem who by mutilating corn plants in various ways, such as splitting or twisting the stalks, apparently produced variations that bred true without regression and which he described as mutations. Observations on the great frequency of striking bud-variations on recovering trunks of old citrus trees in Florida, following the severe freeze of 1894-5, also furnished evidence in support of this theory.

While the evidence at command regarding the artificial production of mutations is not yet sufficiently exact and trustworthy to enable one to draw definite conclusions and formulate recommendations for practical breeders, it may be stated that this is apparently one of the most profitable lines of experimentation for the immediate future.

Principles of selection.

Selection is the principal factor of breeding, both in the improvement of races and in the production of new races and varieties (Fig. 641). The keynote of selection is the choice of the best, and a factor of the highest importance in finding the best is the examination of very large numbers.

In evolutionary studies, it has long been recognized that variation is the foundation of evolution and that no evolution is possible without variation; but, to selection has been assigned an all-important part as guiding and even stimulating the variation in a certain direction. Darwin, and particularly some of his more radical followers, have assigned to selection a creative force, in that it has been assumed that when nature by a slight variation gave the hint of a possible change in a certain direction, natural or artificial selection, by choosing this variation and selecting from among its progeny the most markedly similar variants, could force the advance in the direction indicated. Since Darwin's time, this cumulative action of selection had been emphasized so forcibly that selection had come to be recognized as an active force in creation rather than merely as a determinative agency.

It is certain, of course, that, by careful observation and selection from any particular race, ultimately a new race may be produced. The question is whether the individual or individuals selected in producing the new race have not varied by mutation or seed-sporting rather than being merely representative of the cumulative result of the selection of slight individual variations. The sugar-beet furnishes an interesting illustration in this direction. It will be remembered that Louis Vilmorin started the selection of sugar-beets for richness in sugar between 1830 and 1840, selecting first by means of specific gravity, the method being to throw the beets into solutions of brine strong enough so that the great majority of them would float, the few that sank being of greater specific gravity and presumably of greater sugar-content. Considerable improvement was produced by this method. About 1851, the method of chemical analysis was introduced to determine the exact sugar-content. At this time, the sugar-content was found to vary from 7 to 14 per cent, and in the second generation of selection individuals with 21 per cent of sugar were found. The selection based on percentage of sugar, using the beets highest in sugar as mothers, has been continued regularly since that time, and the industry has come to rely entirely on careful selection for high sugar-content. It would be expected that under these conditions, the percentage of sugar would have increased sufficiently so that the selected plants could be considered a different race or strain. Yet, after fifty years of selection, the highest sugar-content found is only about 26 per cent, and this in a very few instances, seldom over'21 per cent being found. At the present time, many thousand analyses are made every year, so that abundant opportunity is afforded to find individuals producing a high sugar-content. On the contrary, when Vilmorin's work was started, the determination of sugar- content was made by very laborious methods, and was limited to comparatively few individuals. It is not improbable that if Vilmorin had been able to make analyses of the sugar-content in many thousands of roots, he would have found certain individuals producing as high as 26 per cent. The inference from this illustration would be that the limitations of the variation within the race have not been surpassed as a result of selection.
Of recent studies favoring the active influence of selection in creating or strengthening characters, the most noteworthy are those of Castle and Smith.

Castle and his assistants made an extensive series of experiments with hooded rats to increase the black-colored dorsal band on the one hand and to decrease or obliterate it, on the other. He appears to have obtained very positive evidence favoring the gradual cumulative action of the selection, as he succeeded in markedly increasing the amount of black in one strain until the rats were almost wholly black and in the other strain almost wholly obliterating the black. Castle has also obtained similar results in producing a four-toed race, and a change of coloring in guinea-pigs. His view may be summarized in the following quotation: "In Johannsen's view, selection can do nothing but sort out variations already existing in a race. I prefer to think with Darwin that selection can do more, than this; that it can heap up quantitative variations until they reach a sum total otherwise unattainable, and that it thus becomes creative."

The experiments conducted by Smith and others at the Illinois Experiment Station on selecting high and low strains of corn with reference to oil- and protein-content, have resulted in markedly distinct strains possessing these qualities. Experiments have also been made in cultivating these varieties without selection and the new characters have been maintained for several years without marked regression.

The standard researches of DeVries, now familiar to all, challenged the correctness of the selection theory and sought to show that species originated by sudden jumps or mutations. It may be admitted that DeVries proved that species or new characters were formed suddenly as mutations, but this would not prove that they might not also be formed or actually induced to mutate by a continuous process of selection. Indeed, in his experiments on the production of a double-flowered variety of Chrysanthemum segetum ("Mutationstheorie," Vol. I, p. 523), a few generations of selection led to increasing markedly the number of ray-florets before the ligulate corollas appeared among the dink-florets, the change that he interpreted as the mutation that gave him the double variety.

Tower's experiments with the potato beetle in attempting to create by selection large and small races, albinic and melanic races, and races with changed color-pattern, although conducted carefully from ten to twelve generations, failed to give any evidence of producing permanently changed types. While strains of plus and minus varieties gave populations with a range of variation apparently markedly restricted to their respective sides of the normal variation range, still these selected strains did not greatly exceed the normal range of variation in either direction, and when the selection was discontinued, in two or three generations, again populations exhibiting the normal range of variation were produced. Jennings, in a series of selection experiments wjth paramecium extending over twenty generations, and Pearl, in an extensive experiment in the selection of chickens in an attempt to produce a breed of high egg-laying capacity, failed to secure any evidence favoring an accumulative effect of selection.

No series of experiments have had a more profound influence on the conception 01 selection than those of Johannsen, the Danish investigator. In studying commercial varieties of beans, he found that such characters as weight and size of seed fluctuated around a certain average, and when large seed or small seed was chosen the progeny showed the influence of the selection, being smaller or larger in accordance with the direction of the selection. The progeny, however, did not exhibit the extreme sizes of the selected parents, there being a certain regression toward mediocrity. In investigating this matter. Johannsen was led to use the ordinary pedigree method of cultivating the progeny of different individuals separately and inbreeding or selfing all seed used to prevent the crossing of different strains. By this method, he found that the progeny of each individual fluctuated around an average or typical size, as had the commercial varieties, but that while some strains were exactly the same in average size as the commercial variety, others fluctuated around a larger mean or smaller mean than the commercial variety. He tried the experiment of selecting from these large and small strains extreme variants, and found that no advance was made as a result of the selection. He was thus led to conclude that in a pure self-fertilized strain from a single plant— what he called a pure line—no advance could be made by selection and that the commercial variety with which he first experimented was a mixed race. In the course of his experiments with pure lines, several variations were obtained which reproduced true to type, but these were interpreted by him as changes of type by mutation. While, before the publication of Johannsen's results, breeders clearly recognized the importance of determining individual performance and using pedigree methods, still his pure-line conception was a distinct advance and forcibly brought to attention the fact that most commercial varieties and races consist of a number of distinct types—biotypes, as he called them,—and that much of our work of selection consists merely in isolating and purifying these types.

Is one, then, to conclude that the practice of breeders in continually selecting from the best for propagation is useless, and must one advise practical breeders to discontinue their selection? There can be no doubt that the practical breeders have made advances by selecting from the best individuals. No scientific breeder will deny this. It is simply the question of the interpretation of how the results were secured that is in doubt and whether these results can be considered as permanent new unit-characters.

It appears that one is dealing in breeding with two markedly distinct types of selection, based on different principles and arriving at different results, both correct m principle and productive of equally valuable practical results, but of very different value when considered from a strictly evolutionary standpoint. The first of these types would be that in which mutations are selected and new races established, while the second would be illustrated by that type of selection which is intended merely to maintain a maximum strain of the race.

It would seem that such cases of improvement as are illustrated by the sugar-beet indicate that the continuous selection, generation after generation, of maximum fluctuations shown by a character, will result in maintaining a strain at nearly the maximum of efficiency; and that within a pure race the progeny of a maximum variate which would probably be classed as a fluctuation, does not regress entirely to the mean of the race in the first generation succeeding the selection, but that there is only a certain percentage of regression similar to the regression determined by Galton.
These races or selected strains maintain themselves as long as the selection is continued, and when the selection is discontinued rapidly regress to the mean of the species.

The practical breeder should clearly recognize that the act of selection, the choice of the best, remains just as important whether it has a cumulative effect, thereby augmenting the character, or whether he is merely purifying an already existing superior race. The final result remains the same.
Methods of selection, or pedigree breeding.

By methods of selection is meant those practices that the breeder uses to find promising variations, determine their value, and purify or develop them into fixed races coming true to seed.
Choosing superior plants.

The first concern of the breeder is to find the valuable variations. How he had best do this will depend largely upon the plant with which he is working. In all cases, it is of the greatest importance to find the best possible plants and this is likely to require the examination of a very large number of individuals. This factor cannot be too strongly emphasized. If, for example, one attempted to find a man 7 feet high, one would probably have to examine, or pass over, a million individuals to find him. The superior individuals fitted to be the progenitors of a new or improved race are very few. Certain individuals far above the average may be found by examining a comparatively limited number, but the very best possible individual is but rarely produced.

The plants from which selections are to be made should be grown under as uniform conditions as possible, so that the experimenter may have opportunity to examine and select the best. Two methods of growing plants for selection are in general use, and may be termed the nursery method and the field method.

The nursery method, which was first used by Hallett about 1868, consists in cultivating each plant under the most favorable conditions possible for its best development. By this method with wheat, for example, Hallett pursued the policy of planting the indidivuals in squares a foot apart, which would give each plant abundant opportunity for stooling, and also the investigator an opportunity clearly to distinguish each individual plant and determine its characteristics, total yield, and so on. In recent years, this method of growing the individual plants at a standard distance from each other, in order to test their yielding capacities and the like, has been used very extensively.
The field method was used by Rimpau about 1867, and probably by many others before that time. By this method, the selections are made from plants grown under normal field conditions. The advantages of this method are that it can be judged only what a plant will do in the field under ordinary conditions of field culture, by growing and selecting it under these conditions. In the large majority of cases, the first selections are probably made from plants grown in the field in the regular course of crop-production, which thus were not specially grown for the purpose.
If one is to use the nursery method, the plants must be especially planted. While the nursery method certainly allows the breeder to distinguish the individual plants more clearly, in wheat, oats, and other crops that are sown broadcast or drilled, it entails very much extra work and is probably to be recommended only for the use of experimenters who are giving their entire time to the work. In the greater number of horticultural crops, the individuals are normally cultivated one in a place, as in the case of tomatoes, cabbages, strawberries, currants and the like, and the examination of individuals in the field thus satisfies the requirements of both the above methods.
The breeder may have in mind either of two purposes in his work: (1) On the one hand, he may desire to secure an improved strain of a certain race, that is, by selection to keep his seed up to the maximum of efficiency. This may be called strain breeding. (2) On the other hand, he may desire to produce an entirely new race with different characters, and this may be called race-breeding.
He should clearly recognize which of these types of breeding he is following. As an illustration, suppose that the breeder is growing the Stone tomato and desires to maintain the best-yielding strain possible of this race. He would then attempt to choose from a very large number of plants of the Stone variety, the best- yielding plants having the largest number of perfect fruits and typical of the variety in habit of growth, quality, character of fruit, and the like, and would hope by a process of continuous selection to maintain his selected strain in a state of high productivity. This is the type of selection pursued by the sugar-beet breeders described earlier in this article.

On the other hand, if he desires to produce an improved new race, he would search among large numbers of tomato plants of any or all varieties for the appearance of mutations or sports, or plants of new type differing from any known variety. As a matter of experience, it should be stated that it is very easy to find types of plants differing from the varieties or races ordinarily grown, but far the larger part of such variations are worthless types. Good new types, the superior or even the equal of the known varieties, are of very rare occurrence.

If the general improvement of a variety is the breeder's purpose, he should choose a considerable number of apparently superior plants of good type, which will form the basis of his selection work. Breeders who are conducting careful experiments will find it necessary and desirable to use careful methods of judging their plants. While one is breeding possibly for one primary improvement, as, for example, increased yield, it is necessary, at the same time, that one should keep the product up to the standard in other characteristics, namely, quality, disease-resistance, drought-resistance, and the like, and that one sees that all of the good qualities of the variety are retained. To do this properly necessitates the use of a score-card, on which each character of the plant that is important is given its relative weight or grade. By the use of such a score-card, the breeder can judge each character separately, and by the adding up of the score-card get the rank of different plants in a comparative way.

Inheritance test.—When a number of plants have been chosen, the next important factor is to test each individual as to its inheritance. It must be continuously remembered that a plant is valuable only as it produces good progeny. To determine the inheritance, the usual method is to plant the seed from each individual selected in a row by itself, or in a marked part of a row. This is the so-called "plant-to-row" method, and brings the offspring of a single individual together so that they may be readily compared with each other and their qualities carefully judged. These progeny rows should be grown in a special breeding-patch in which the soil is as uniform as can be secured.
It is frequently found that two select plants that are equally good so far as their yield is concerned will give progeny that, as a whole, differ greatly in this respect. In the progeny of one, almost every plant may have inherited the desired quality, while in the progeny of the other only a few of the plants may show, in any noticeable degree, the inheritance of the quality. To determine the degree of inheritance, it is necessary to grade carefully the progeny of each individual.

Finally, with the use of his best judgment, the breeder determines the superior progenies, and these would be the ones which have most nearly given the ideal type and produced the best yield of the highest quality. This would end the work of the first generation of the selection as the breeder now has the data which shows him which of the original plants selected was the superior one. It will be seen that this is a method of judging the individual by its progeny.

Continuation of the selection the second year.

Having determined the superior progeny or progenies at the end of the first year, the breeder then makes his selections of seed-plants from these best progenies for continuing the breeding. While one progeny may be and usually is superior to all others, this may be due to the season or other accidental conditions and for a few generations it is usually the best policy to make selections from several of the best progenies. Select from each of the superior progenies several of the best plants, using the same care in selecting these plants as was used in choosing the first plants. Preserve the seed from each of these plants separately and keep it carefully labeled so that its origin may be known.

The further work with these plants consists in planting each individual by the plant-to-row method, testing the inheritance as described in the first generation, and finally selecting again the best progenies. This would be followed by again selecting from the best progenies a number of superior individuals to continue the selections in the third year.

The third and succeeding years of the selection would be conducted in the same way as long as it was thought necessary or desirable to continue the work.

Securing general stock seed of the improved strain.

In carrying out selection work as outlined in the preceding section, it is ordinarily the object of the breeder to secure an improved strain of the race with which he is working, and usually he desires to utilize such improvements as he can make at the earliest possible time. With ordinary annual crops such as beans, peas, tomatoes, corn, and cotton, it will be found a good policy at the end of the second year of the selection, after taking the seed from the few special plants used in continuing the pedigree breeding, to harvest the seed from a number of the best plants remaining in the chosen progenies and using this seed to plant a multiplication plat from which stock seed may be secured to plant a fairly large crop. Each year following this, seed may be taken in the same way from the best progenies in the breeding patch to plant a multiplication plat. By this method, seed of a gradually improving grade may be secured for planting a general crop.

Control of parentage.

In plant-breeding, as in animal-breeding, the isolation of the parents is a very important consideration. It is necessary that the character of both parents should be known whenever this is possible. In breeding plants, more attention is given ordinarily to the mother parent and in very many instances the characters of the father parent are entirely neglected. Animal- breeders, on the contrary, give more attention to the characters of the male parent, and much improvement in ordinary herds has been accomplished by the introduction of improved heritage through the male. In plant-breeding, it is desirable that the seed of the select individuals be planted in a field by themselves. This insures that only progeny of carefully selected individuals will be planted near together, and thus no ordinary stock will enter as a contamination. One can be certain that each plant of the progeny is fertilized with pollen from another similarly good plant, or at least from a plant derived from good parentage. One difficulty, however, has been experienced by plant-breeders in planting continuously their selected stock in such isolated plats. If this method is continued year after year, it results in fairly close inbreeding, which, in the case of plants, frequently results in loss of vitality and vigor. In animals there is frequently no noticeable effect from close inbreeding, and many of the most famous animals have been produced as a result of the closest in-and-in- breeding. In plants, however, it is possible to secure much closer inbreeding than in animals, as in many cases a plant can be fertilized with its own pollen. Within recent years, much activity has been shown in the careful breeding and improvement of corn. The corn plant has been shown, as a result of experiments made by various investigators, as, for example, the Illinois Experiment Station and the United States Department of Agriculture, to lose vitality very rapidly when self-fertilized. Within three or four generations, by the most careful inbreeding, it is possible to reduce corn almost to total sterility. The general practice of corn- breeders who have been giving attention to the production of pedigree strains is to plant the rows of corn from different select ears side by side, giving a row to each select ear, and each year selecting, from the progeny of those rows that give the largest yield, plants to continue further the selection. Planting these select ears together ever}' year, therefore, means that they are more or less inbred, as the closest relatives are planted together in the same row. While in following this practice at first no effect was visible, corn- breeders are now finding in some cases an apparent decrease in yield, which seems to be traceable to the effect of inbreeding. It seems necessary, therefore, in corn and in other plants that are affected by inbreeding, to use methods that will avoid close inbreeding. The detrimental effect of inbreeding is largely limited to those plants that are normally cross-fertilized, this fact being strikingly brought out in Darwin's "Effects of Cross and Self Fertilization in the Vegetable Kingdom." Tobacco, wheat, and some other plants that are normally self- fertilized do not show this decrease in vigor as a result of inbreeding.

Considerable effort has been made within recent years, in the selection of certain crops, particularly corn, to follow both parents, choosing good males and good females of known parentage and crossing these by artificial means. There can be no doubt but that by the use of such methods more rapid progress could be made, but in the greater number of cases the methods thus far devised require so much work as to be almost prohibitive.

A method of breeding has recently been devised by J. B. Norton in the improvement of asparagus, which is worthy of careful consideration and may be applicable, at least in modified form, to use in the improvement of various crops. Asparagus is dioecious, the male and female flowers being borne on different plants. The first part of the process consists in selecting a number of superior plants of both sexes, attention being given to all important characters such as yield, quality, rust- resistance, and the like. This having been done, the next process consists in crossing each female with each selected male plant and testing the progeny produced by the cross. If, for example, ten superior females and ten superior males were chosen, a sufficient number of flowers on female No. 1 would be crossed with pollen of each of the ten males to obtain sufficient seed to test the comparative value of the progeny of female No. 1 with each of the ten males. The ten lots of seed from the crosses on female No. 1 would be grown separately and the comparative value of the different progenies determined by careful observations on vigor, rust-resistance, quality of product, yield, and the like. In this way, it would ultimately be determined which of the ten male plants was the superior one to use in crossing with female No. 1. In like manner, female No. 2 would be crossed with each male and the progenies tested to determine the superior male in this combination. Finally the combination of each female with each male can be compared and if the work has been conducted with sufficient care and for a long enough period, it can be determined which combination has uniformly given the best results.

Asparagus is perennial and is easily propagated vegetatively by separation of the roots, so that when once the superior male and female combination has been determined, these may be propagated vegetatively as clons, in alternate rows in an isolated place so that all seeds developed will be of the desired combination. Asparagus produces numerous seeds and by such a method an indefinite quantity of seed of the desired combination can be produced. It would doubtless be possible in a few years, if desired, to have all of the seed used commercially grown from a single superior combination.

If experiments of this nature could be made on an extensive scale so that the males and females of the highest or maximum grade could be discovered, they would be of almost fabulous value. The importance of this method of breeding may be better appreciated by imagining the value of the best bull and the best cow in the world if they would live indefinitely and if it were possible for them to reproduce rapidly enough to supply all of the individuals desired.

The direct application of this method is possible only with perennial dioecious plants that can be propagated as clons and that develop numerous seeds. The hop is another plant to which this method of improvement could be applied.

Hybridization.

Aside from selection, hybridization has played the most important role in the formation of the varieties and races of our cultivated plants; but the results obtained are in many cases closely connected with selection. Ever since the time of Knight, hybridization has been used extensively in plant-breeding, and it seems that this is the only sure means that the breeder can use in producing new and desirable combinations of characters. In -hybridization, as generally used, the breeder does not expect to cause or produce new unit- characters, although such changes may occasionally occur under the stimulus of hybridization. What he can do with certainty is to secure hybrids combining the different characters of two distinct sorts. The extent to which such recombination of characters can be carried is almost unlimited. In many cases, two or three or even four distinct species and the characters represented in their various varieties have been recombined in hybrids.
In older literature, the term hybrid was restricted to crosses resulting from the combination of distinct species, while combinations of different races of the same species were known as crosses or mongrels. The term hybrid is here used as designating any product of a cross when the parents were noticeably distinct from each other, whether the parents belonged to different clons, races or species. This broader use of the term hybrid has become almost universal in recent years. If, in discussion, it is necessary to refer to the degrees or grades of difference in the parents, the hybrids may be characterized as species hybrids, racial hybrids, clonal hybrids and the like.

Choosing varieties to hybridize.

In starting any work in hybridization, the first important step is the choice of the varieties to be hybridized. It is interesting to make crosses of any two plants with distinct characters and observe the recombinations of characters which result, but this haphazard work takes too much time and is not to be recommended. The breeder, in general, should hybridize with some definite aim in view and use systematic methods in attempting to accomplish that aim. One cannot ordinarily expect to obtain in hybrids of any two varieties any characters which are not present in the parents. The unit-character conception explained in the beginning of this article is of fundamental importance in understanding hybrids. The breeder, by a careful study of varieties, determines the good characters and the poor or weak characters of each variety. He may, as an illustration, if working with tomatoes, find all of the varieties with yellow pear-shaped fruits to be large bushy plants, the so-called standards, and he may desire a dwarf type of plant and red fruits. By examining the different races of tomatoes, he would soon find a variety, such as the Quarter Century, which possesses the characters of dwarf plant and red fruit. By crossing these two varieties, he would obtain new combinations of the characters of the two sorts, and if he grew a sufficient number of the second generation of these hybrids, he would be certain to find some plants in which the pear-shape had been combined with the red color and dwarf habit of the Quarter Century variety. The study of the varieties of any crop thus gives the breeder an idea of the characters available, and he must then use his originality and judgment in determining what combinations of these characters would form the best commercial variety. If this combination does not already exist, he may start out with considerable confidence that it is possible for him to obtain such a combination and thus a valuable new variety. Plants, however, are not simple in their organization and the reaction of different characters on one another in different combinations may not always be what one expects. Again, in no plant has a complete analysis been made of all characters, and it may be impossible for us ever to reduce all the characters of a plant to a unit-character basis; thus there is always an element of doubt as to the value of any new combination of characters until this combination has been produced and tested.

Methods of crossing plants.

Plants, like animals, bear male and female organs, and an act of fecundation is necessary in all ordinary cases to insure the development of seeds. In probably the larger number of plants, the male and female organs or the stamens and pistils are borne in the same flowers on one plant. In some cases, as in the castor bean, corn, and the like, both sexes are borne on the same plant but in different flowers. In still other cases, as in the date palm, asparagus, hop and hemp, the sexes arc on different plants.

In hybridizing plants, it is necessary to insure that the plants are not fertilized with their own pollen or with pollen from any other source than that desired. If, therefore, the plant to be operated on has the stamens and pistils in the same flower, the stamens must be removed from the buds before they burst and discharge the pollen. This act of removing the stamens, or emasculation, as the process is called, is necessary in order to prevent self-fertilization. In some plants, it is necessary to emasculate the buds very early, as the pollen develops considerably in advance of the pistils. In other cases, the pistils reach maturity or a receptive condition before the pollen is shed. In this latter case, the emasculation may be delayed until a time just previous to the normal opening of the flower.

The process of emasculation may be illustrated by the columbine. Here large-sized buds are chosen just before they open normally (Fig. 642). The tips of the petals can then be easily pried apart so that the stamens may be pulled off with small forceps. This process should be performed carefully to avoid crushing or injuring the pistil. The bud should then be inclosed in a small light paper bag in order to prevent pollen from any foreign source being brought to the pistil by insects or wind (Fig. 643). The bud should remain covered until sufficient time has elapsed to allow the pistil to reach normal maturity, when the bag should be removed and the pollen from the desired variety dusted over the pistil. After this act of pollination, the bud should again be covered with the paper bag, which should not be finally removed until several days later, after fecundation has taken place. As soon as a flower is pollinated, it should be labeled with a small tag of some sort which may remain attached to the flower-stem until the fruit is ripe. In some cases, the pollen may be placed on the immature pistil without injury, when the flower is emasculated, and this is a great saving of time when it can be done. However, in most cases, premature pollination is liable to injure the pistil and prevent the setting of seed. One should ordinarily attempt to pollinate the pistil at as nearly the normal time as possible. Many plants are difficult to hybridize and every process must be as natural as possible to insure results.

Many handy methods have been devised to use in pollination work and are described in breeding literature. In all work fine copper wire is better to attach bags and labels than is string. In emasculation work also, it will often be found convenient when some pollen has accidentally fallen on the pistil to wash it off with water by means of a small dental syringe. In many cases, such as apples, pears and cotton, the best means of emasculation is to remove the outer floral envelopes by cutting them off, using a sharp scalpel. With a little practice this can be done quickly and with minimum injury to the essential organs (Fig. 644)

Difficulty is frequently experienced when hybridizing different varieties, in getting plants of each variety to bloom at the same time. This difficulty may be overcome in many cases by keeping the pollen, which can be done for a limited period by slightly drying the pollen without allowing it to become desiccated, and preserving it in a tightly corked bottle.

After the pollen has been placed on the stigma of the pistil by the act of pollination, each pollen-grain develops a small tube which grows down through the pistil to the ovary. Through this tube, the male germ-cells pass down and finally a male germ-cell comes in contact with each egg-cell of the different ovules in the ovary (in most plants there are several ovules in each ovary) and fuses with them. This constitutes the act of fecundation or fertilization. This fecundated egg-cell is then the beginning of the hybrid and from the seed containing it, when grown, there develops the hybrid plant. The plant developed directly from this hybrid egg-cell is known as the first-generation hybrid (F ,). Seeds from this first-generation hybrid, when grown, give second-generation hybrids (F,). The expressions F 1, F2, and F3, meaning first, second and third filial generations, are used very commonly to designate the first, second and third generations of hybrids.

Laws of inheritance in hybrids.

When plants of different pure races are crossed, as, for example, different races of wheat, corn or cotton, the hybrids are usually all very similar to each other in the first generation, exhibiting in general the same characters. And this is the case also when different fixed species are crossed. If, however, individuals belonging to unfixed races are crossed, there is usually a considerable variation in the first generation. This is well illustrated by the crossing of different clons of apples, pears, oranges, and the like, when the different so-called varieties are merely transplanted parts of the same individual seedlings which have not been bred to a purity of type. It is well known that if seeds of an apple variety be planted, the resulting plants exhibit many different variations in the first generation. The parents, themselves, therefore, not being of pure type, when they are hybridized produce progeny which in the first generation is variable. In the crossing of races which nave been bred true to type, whether of the same or of different species, the first-generation hybrids, however, are nearly uniform in the characters presented, and in such instances it is necessary to secure a second generation of the hybrids in order to accomplish the segregation of the characters and the production of a large number of variations. Ordinarily, therefore, desirable variations are looked for in the second generation. This, as has been explained above, is true only in the case of hybrids of species and races that are fixed in type.

Mendel's law of hybrids.

The preceding discussion represents fairly well the general understanding of hybrids until about 1900, when DeVries and Correns rediscovered what is now termed "Mendel's law of hybrids." These laws or principles are of great value from an economic standpoint, and are, furthermore, of the greatest scientific interest. They should thus be thoroughly understood by every practical breeder of plants. It has been known for many years that a splitting-up and redistribution of parental characters occur in hybrids, and it is on this fact largely that the practical application of hybridization in plant-breeding depended. Until Mendel's law was discovered, however, there was no understanding of why or how such a recombination could be made, and it was necessary to experiment extensively in order to determine what could be accomplished.

If one carefully studies a number of first-generation hybrids with special reference to the characters of the parents exhibited in the hybrids, it will be found that certain characters possessed by the male parent are plainly represented in the hybrid, while other characters possessed by the female parent are also represented in the hybrid. Many characters of the parents are thus plainly represented in the hybrid, but it is probable that other characters will be blends of the similar parental characters, or possibly differ from any definite characters distinguishable in the parents. Attention has already been called to the complexity of organisms in general and the difficulty of recognizing all of the unit-characters. Thus far it has been possible only to follow carefully certain plainly marked characters. This commingling of the different characters of each parent gives the hybrid a mosaic appearance, as if certain characters had been taken from each parent and thrown together to make up a hybrid individual.

Character-pairs.—To understand this commingling of characters in the first-generation hybrids, it is necessary to know that the parents used in the hybridization differed from each other in certain characters. One parent may have had red fruits, hairy stems, and dwarf habit, while the other may have had yellow fruits, smooth steins, and tall habit. Such characters are opposed to each other, and such opposed qualities or characters are termed "character-pairs." A plant may have red fruits and smooth stems, but it could not have red fruits and yellow fruits at the same time. As an illustration of such character-pairs, may be cited, scarlet and yellow fruits of peppers, reversed or erect fruits of peppers (Fig. 645), starchy and sweet kernels of corn, standard and dwarf size in tomatoes, stringy and stringless pods of beans, and the like. Such pairs of characters have been termed by Bateson "allelomorphic pairs of characters," and this terminology is commonly used in the literature on hybrids. When parents possessing opposed or contrasted characters are crossed, the hybrid egg-cell receives, through the male and female germ-cells uniting in the fecundation, the determiners which represent the different contrasted pairs of characters, and all cells making up the first-generation hybrid will contain in like manner the determiners representing these characters, and are thus hybrid in nature. This being the case, it might be expected that all characters in the hybrid would show as blends of the parental characters or exhibit some stage of intermediacy between the characters of the parents. This is indeed frequently the case, but more commonly one of the characters is very strong, or "dominant," as Mendel expressed it, and only this character will show in the first-generation hybrid, the other character remaining recessive or masked, although present. As an illustration, in the character-pairs mentioned above scarlet fruits of pepper, reversed fruits of pepper (this is true only in certain varieties), starchy kernels of corn and standard size of tomato plants, are dominant over their corresponding contrasted characters. Illustrations of blended or intermediate characters are found, for example, in first-generation hybrids of round with pear- shaped tomatoes, and large with small fruits of tomatoes or peppers.

The law of segregation and purity of the germ-cells.—The second important principle of Menders law is what is termed the law of segregation and purity of the germ-cells. It seems certain from the researches that nave been conducted that, when the germ-cells of the first-generation hybrids are formed, the determiners which represent the two different characters under consideration, and which were united by the hybridization, ordinarily segregate again in the cell-divisions, which lead to the formation of the germ- cells, so that certain germ-cells include the determiner of one only of the two characters. There are thus two kinds of germ-cell formed with respect to this one character-pair. Choosing as an illustration a hybrid of a pepper having scarlet fruits with one having yellow fruits (Fig. 645), when the germ-cells were formed a segregation of the determiners representing the two opposed characters would take place and there would be germ-cells of one kind, both male and female, containing the scarlet fruit determiners and of a second kind, both male and female, containing the yellow fruit determiners. This segregation takes place in the formation of both the egg-cells and the sperm-cells or pollen-grains. It is thus seen that the first-generation hybrid, when two such allelomorphic characters are combined, forms two kinds of egg- cells and two kinds of sperm-cells, so far as this one character-pair is concerned. This segregation of characters, which has been termed the law of segregation, is one of the most important facts of inheritance and, in enabling us to get recombinations of characters, is of the highest importance in breeding.

The law of probability in recombination of characters.—The third important principle of Mendel's law is what is termed the law of probability, and explains what may be expected in plants of the second generation of such a hybrid. Remembering that there are formed in the first-generation hybrid, as explained above, two kinds of egg-cells and two kinds of sperm-cells with reference to the opposed characters, what would happen if the hybrid were bred with its own pollen; or, in the case of an animal, if it were bred with another hybrid of the same parentage? For the purpose of illustration, suppose that a hybrid of a scarlet-fruited pepper with a yellow-fruited pepper be fertilized with its own pollen, and that 100 egg-cells be fertilized with 100 pollen-grains of the same hybrid. There are two kinds of egg-cells produced, some carrying determiners of the scarlet fruit, and others determiners of the yellow fruit, and the same is true of the pollen-grains. Taking the egg-cells and pollen-grains without choice, as equal numbers are produced of each kind, one would expect to have of the egg-cells fifty with scarlet determiners and fifty with yellow determiners. In the pollen-grains, also, one would expect to have fifty with scarlet determiners and fifty with yellow determiners. If, then, the 100 egg-cells and 100 pollen-grains are brought together in fertilization by chance, as would occur in nature, according to the law of probability, there would be twenty-five scarlet uniting with twenty- five scarlet; twenty-five scarlet uniting with twenty-five yellow; twenty-five yellow uniting with twenty-five scarlet; and twenty- five yellow uniting with twenty-five yellow. Representing scarlet determiners by the capital letter S because scarlet is the dominant character, and the yellow determiners by the small letter y, as yellow is recessive, the unions may be represented as follows:

One Hundred Ego-cells Bt 100 Sperm-cells.

Female       Male        Composition

Cells           Cells            of hybrids

25S     X      25S      -       25SS            These do not contain determiners of y and will reproduce true.

25 S   X       25 y    -        25 Sy      

25 y    X   25 S   -          25 yS     These are hybrids so far as this character-pair is concerned exactly                                              
the same as in the first generation and contain determiners of both S and y. These will not reproduce true to type and will break up like second-generation hybrids.

25y   X   25y      -      25 yy  These do not contain the determiners of S, and will reproduce I true.

The above illustration explains the law of segregation, and the probable ratio of recombination when hybrids are inbred with their own pollen, and when only one pair of characters is considered. When an egg-cell with scarlet determiners unites with a sperm- cell with scarlet determiners, this gives rise to a pure germ-cell, or zygote containing only scarlet determiners, and the progeny in subsequent generations will breed true Mo far as this character is concerned. Also, when an egg-cell with yellow determiners unites with a sperm-cell with yellow determiners the result is a pure germ-cell, containing only yellow determiners and the progeny would reproduce true, so far a this character is concerned, in subsequent generations. In the other two cases, when in fecundation gametes with scarlet determiners unite with gametes with yellow determiners giving the combinations Sy and yS, which amount to the same thing, there result in reality, hybrids exactly the same as in the first generation and the progeny from these in the next generation behave exactly the same as did the first-generation hybrids in the second generation.

In such a case as the one under consideration, in which the scarlet is a strong dominant character, all combinations that contain the determiners of this character, whether pure or of hybrid nature, show i In - character only. Thus in the above 100 combinations the twenty-five yy would come with yellow fruits while the seventy five other combinations would have scarlet fruits, although fifty of these would be of hybrid nature. To determine which of these seventy-five scarlet-fruited plants are the combination Sy, that is, scarlet with yellow, and which are SS, that is, scarlet with scarlet, requires the growing of self-fertilized progeny from them to determine which are reproduced true to type, as these would be the pure scarlet. The progenies of any of these plants that produced both scarlet- and yellow-fruited plants would show that the parent of such progeny was a hybrid.

In the hundred combinations there is thus produced a ratio of one pure scarlet to two hybrid scarlet and yellow to one pure yellow, 1 SS:2 Sy:l yy. or three scarlets to one yellow and this is the famous 3:1 Mendelian formula.

This process of union of an allelomorphic pair of characters in hybridisation, the formation of four kinds of germ-cells, both male and female, by the hybrid, and their four different unions, is graphically illustrated in Fig. 646.

While in certain hybrids of parents possessing two opposed parental characters, this ratio of probabilities is not produced, if large numbers are used the ratio will be found in many cases with little deviation. A sufficiently large number of cases have now been studied with various plants and animals to place this conclusion beyond question. It is not known, however, how many characters follow Mendel's law, nor is it yet entirely certain whether those character-pairs that sometimes follow the law of segregation always follow it.

The individuals of the second generation which contain the determiners of both characters of the pair, if self-fertilized or bred with similar individuals containing the determiners of both characters, exhibit in the third generation exactly the same nature that first-generation hybrids exhibit in the second generation. The two determiners are commingled in their cells, and to all intents and purposes they are exactly the same as first-gene rat ion hybrids. When such self-fertilized hybrids are grown they give again, in the third generation, the regular Mendelian proportion of 1 S8:2 Sy: 1 yy. Here the individuals containing only determiners of one character, that is, SS and yy, would come true to these characters in succeeding generations, while those individuals containing the determiners of both characters, S and y. would be expected to segregate again in the fourth generation in similar proportions.

When dealing with more than one character-pair, ratios of segregation become complicated but are 
easily understood. If the character of reversed fruits (R) and erect fruits (e), two plainly marked characters of ordinary garden peppers, caused by the pedicel of the fruit curving backward in one case and remaining straight in the other, are combined with the above allelomorphic characters, it can be foretold exactly what combinations will occur and the relative number of each. This is a second allelomorphic pair of characters that behaves in inheritance the same way as did the two colors of fruit. In this case, the reversed pedicel is the dominant character, as in the hybrids of reversed with erect sorts the peai eel 8 are always or very generally recurved. These characters would thus be represented by R for the recurved or dominant character and e for the erect or recessive character. In this character-pair one would expect a splitting and segregation to have occurred in the formation of the germ-cells of the first generation so that the hybrid plants of the second generation would exhibit these characters in Mendelian proportions as in the character-pair first described. The progeny in the second generation would thus exhibit these characters in the following combinations; and proportions: 1 RR:2 Rc:l ee. This theoretical proportion should hold rather constantly, either in small or large numbers of hybrids, though in large numbers it would be more nearly realized. The determiners of the four characters, or two character-pairs, are commingled in the cells of the first-generation hybrid. When the egg-cells and pollen-grains are formed, however, a segregation of the determiners of the two character-pairs occurs, but independent of each other. Each egg-cell or pollen-grain will receive only the determiner of one character of a certain character-pair but will, at the same time, receive determiners of other characters belonging to other character-pairs. Considering the two character-pairs described in poppers, an egg-cell receiving the determiner of the scarlet color of fruit S, might also receive the determiners for either R or e representing the characters of recurved or erect fruits. These two character-pairs would thus give egg-cells of four combinations, SR, Se. yR, and ye. .

In the formation of the pollen-grains, the same combination occurs, so that with reference to the two character-pairs described. the pollen-grains that would be formed have the same combinations of determiners as the egg-cells, namely, SR, Se, yR, and ye. There would thus be four kinds of egg-cells and four kinds of pollen grains so far as these two character-pairs are concerned. If these are brought together, sixteen combinations are possible as follows:

Examining these combinations carefully, and placing together those combinations that contain the same character-determiners as indicated by the letters, and this can properly be done as it does not matter in the fecundated egg whether a certain determiner is furnished by the egg-cell or the pollen-grain, there result the following nine combinations, all of which are different in germinal constitution with reference to these two character-pairs:

Table Showing Number Of Germinal Combinations And Chab- Acteb Of F, Pepper Hybrids With Two Allelomorphs.

No. of combina- tions	Germinal constitu- tion	Visual characters of hybrid	Nature of hybrid

1				SRSR		Scarlet recurved	Pure scarlet and recurved

1				SeSe		Scarlet erect	Pure scarlet and erect

1				yRyR		Yellow recurved  Pure yellow and recurved

1				yeye		Yellow erect 	Pure yellow and erect 

2				SRSe		Scarlet recurved	 Pure scarlet and hybrid reurved X erect

2				SRyR		Scarlet recurved	Hybrid scarlet x yellow and
pure recurved

2				Seye		Scarlet erect		Hybrid scarlet X yellow and
									pure erect

2				yRye		Yellow recurved 	Pure yellow and hybrid recurved X erect

4				SRye	Scarlet recurved Hybrid scarlet x yellow and 	hybrid recurved X erect

An examination of the preceding table, in which are grouped the sixteen possible combinations when two allelomorphic pairs are concerned in the hybridization will show that among these sixteen there are nine groups with different germinal constitutions. The visual character of the hybrid plants of these nine different groups is given in the third column and is easily understood by examining the germinal constitution and remembering that scarlet S, and reversed R. are the dominant characters in the two allelomorphs and that the presence of one determiner of either of these characters will cause the appearance of that character in the hybrid plant. It will be observed that by grouping the hybrid plants according to the characters they show, there will be nine scarlet and reversed, three scarlet and erect, three yellow and reversed, and one yellow and erect. This is the Mendelian formula: 9:3:3:1. The nature of the nine different groups of hybrid plants with different germinal constitution is given in the fourth column of the table. When a character is pure, it may be expected to reproduce true in succeeding generations but in those cases in which both determiners of a character-pair are present, the character is of hybrid nature and will segregate in succeeding generations.

In the illustration of the character-pair, scarlet and yellow fruits and the probable ratio of number of unions in F, hybrids, it was shown that out of 100 unions one should expect 25 SS:50 Sy:25 yy. If now the second character-pair recurved and erect fruits is coneidered in connection with these same 100 unions, there would occur the following combinations, according to the law of chance:

These nine combinations are the same as given above, but the percentage of each combination out of the 100 unions is shown.

If a third character were considered, the proportions of the combinations can be determined in exactly the same way. Each one of the above nine possible combinations would be again divided into three different unions in the same way as the three combinations of the one character-pair gave nine different combinations with the second character-pair. In the consideration of the three character-pairs, there would thus be twenty-seven different combinations of parental characters. And again in each ovary fecundated, when only one determiner of each character-pair occurred, the opposing character-determiner being in each case eliminated, such a cell should give a plant that would reproduce its character true to type. It is well known that almost any two different races or species that may be chosen for hybridisation will ordinarily differ from each other in numerous characters. When there are a number of these opposing characters which form Mendelian character-pairs, the determination of the possible combinations by Mendel's formulae becomes very complex and difficult to under- etand. It is only by taking a few well-marked character-pairs and carefully studying them that the segregation and new combinations according to Mendelian proportions can be followed and understood.

Any character-pairs following Mendel's law would segregate as indicated above, in the case of scarlet and yellow fruits and reversed or erect fruits of the pepper. A very large number of characters of various plants and animals are now known to be Mendelian and while many modifications of the principles have been necessary to harmonize them with special cases, still it may be said that there is no other general law of heredity and Mendel's law has thus furnished us with a working basis of great value.

The study of hybrids has been resolved into a study of unit- characters and their relation to each other. By hybridizing related types having opposed characters and observing the segregations which occur in the later generations, the characters of each type arc family  and it is determined when a character-pair occurs. The researches on this subject by Mendel, Batewon, Davenport, Castle, Punnett, Shull, Hurst, Correns, Tschermak, East and dozens of other now well-known investigators, have developed a science of heredity of which there was no conception a few years ago.

The characters presented by the different varieties of a plant or of different species, which can be crowd with it, can now be studied, and one can definitely plan the combination of characters desired in an ideal type, and can with considerable confidence estimate the number of plant it will be necessary to grow to get this combination. It is now known in general how characters behave in segregation and inheritance, ?? that one can go about the fixation of a desired type, when one is secured, in an orderly and intelligent way.

The further the study of characters is carried, the more it is coming to be realized that the appearance of apparently new types following hybridization is due to recombinations of different units which in their reactions give apparently new characters. As an illustration, in a study of pepper hybrids, which has been conducted during the past four years, it has become evident that the form of plant and branching is due to three pairs of characters or allelomorphs; namely, first, erect or horizontal branches; second, large or small branches; and third, many or few branches. In crossing two medium-sized races, one with large horizontal and few branches, and the other with small erect and numerous branches, there result many new combinations of characters, among which appear some with small horizontal and few branches, which gives a dwarf plant, and others will have a combination of large erect and numerous branches, which gives a giant plant (Fig. 647). These dwarfs on the one hand and giants on the other appear as distinct, new creations, though they are very evidently merely the recombinations of already existing unit characters, and dwarfness and giantness are the results of the reaction of the different units combined.

When the large number of distinct characters that are presented by the very numerous varieties of any of our cultivated plants is remembered, an understanding is secured of the possibilities of improvement which the field of hybridization affords.

The development of hybrids into pure races.

When hybrids have been produced between species or varieties possessing certain characters that it is desired to unite in a variety, the recombinations of characters as explained in the preceding section become visible in the second generation, and it is thus among the plants of this generation of the hybrid that one should expect to find the combination of characters desired. The breeder would thus very carefully examine a large number of second-generation plants and choose for further experimentation those plants that were found to have inherited the characters which he desired to combine. The entire batch of F2 plants should be carefully examined to determine what characters behave as character- pairs and also the dominant or recessive nature of each character. This knowledge is necessary in order to determine the practice to be pursued in choosing plants in which the characters desired will be pure with reference to these characters. If, for example, the breeder is working to get a combination of two characters only, such for instance as a yellow- and erect-fruited pepper, from the combination of character-pairs discussed above in explaining Mendel's law he would discover that both of these characters are recessive, and thus when a hybrid was found in which these two characters were united, he could be sure that by self-fertilizing such an individual it would reproduce true with reference to both of these characters in the next and succeeding generations. He would know furthermore in dealing with only two pairs of characters that he should, according to the law of chance, secure on an average about one such combination in sixteen hybrids.

If, however, the combination desired was a scarlet reversed fruit, both dominant characters, the process would be much more difficult. As shown in the preceding section describing the segregation and recombination of characters, nine plants out of the sixteen possible combinations would have red, reversed fruits, while only one of the nine would be pure with reference to both of these characters. The breeder would thus be compelled to self-fertilize a number of the plants having red and reversed fruits and grow a number of plants from each in order to determine which one, if any, was pure with reference to both characters. If, then, the progeny from any one of the plants chosen and self- fertilized came true to type with reference to both characters, he would be certain of its purity and would again self-fertilize some of the best plants of this progeny, which should give him a pure type.

If a combination of a dominant and recessive character is desired, the examination of the Fs 
hybrids would enable the breeder to choose a pure plant so far as the recessive character is concerned, but he could not determine the purity of the dominant character and would be compelled to self a number of plants exhibiting the two characters and grow progenies in the third generation, when be should be able to select a pure type with reference to both characters.
If, as frequently occurs, neither character of an allelomorphic pair is dominant, but gives in the hybrid an intermediate form, the fixation becomes simple, as in such cases those hybrids in which either character is pure can be recognized.

While these methods appear very complex at first, they will be easily understood with careful study, and are far simpler than the methods breeders were compelled to employ in fixing hybrids before they had an understanding of Mendel s law.

When more than two characters are concerned in the recombination, the process becomes more difficult, and indeed one cannot limit one's consideration to two characters in practical breeding unless one is combining standard varieties where all characters are good. As in simple selection work, one must necessarily consider all important characters that go to make up a good variety, and usually one will be able to recognize Mendelian segregation only in a few prominent differential characters. The breeder should use the knowledge of inheritance that he possessess with all characters which he can recognize, but at the same time the plants which he inbreeds to secure purity of type should be perfect plants of all-round good type, and in every generation of the hybrids grown he should exercise his best judgment in selecting the best plants for seed-bearers.

In the fixation of cotton hybrids, the policy was pursued of selecting for inbreeding the most fruitful and best-shaped plants of those hybrids having the desired characters, using very large numbers of hybrids from which to choose. The self- fertilized seed of a certain type was then planted by the plant-to-row selection method in an isolated plat, in order to give an opportunity to select not only the pure combination of the desired characters but the best all-round plants. As soon as the plants in such an isolated plat were sufficiently developed to show their characters and it could be recognized that certain ones had inherited the desired qualities, the fields were carefully searched and all plants not true to type were pulled up, leaving only a few good plants of the correct type. This process of roguing, as the seedsmen call it, insures that at least the greater part of the seed developed would be fertilized with pollen of similar plants of good type. 
This sort of selection and purification of type will probably in most cases be found necessary even after such Mendelian characters as can be recognized have been secured in a pure state.

The inheritance of many fundamental characters will doubtless remain obscure for many years.
The use of impure first-generation hybrids.

In the case of very many of the most important horticultural crops, fortunately, it is possible to use hybrids without the necessity of purifying or fixing them as described in the last section. Plants such as apples, pears, oranges, grapes, roses and strawberries, which are grown as clonal varieties, being propagated by buds, grafts or slips, are merely parts of one individual and it does not matter whether they are germinally pure, as seeds are not needed. This makes it possible to use Ft hybrids and, as hybrids are notoriously vigorous, this is a factor of very great importance. Again, characters which blend and give intermediates in the Fj generation may, in such cases, prove very valuable.

The work that has been carried out by the Department of Agriculture in the breeding of citrus fruits very clearly indicates that valuable intermediates may sometimes be secured. The writer, in conjunction with Walter T. Swingle, hybridized the hardy cold-resistant trifoliate orange (Poncirus trifoliata) with several varieties of the tender sweet orange, and as a result at least five different varieties of hardy oranges or citranges have been produced. These hybrids are nearly intermediate between the two parents, having the characters in the first generation nearly blended. The leaves are trifoliolate, but are much larger than the leaves of the ordinary trifoliate orange tree, and show a tendency to drop off, the lateral leaflets producing an unifoliolate leaf. The trifoliate orange is deciduous, while the sweet orange is evergreen. The hybrids are semi-deciduous, holding a large share of their leaves through the winter In hardiness they also seem to be intermediate, being much more cold-resistant than the ordinary orange, but not so hardy as the trifoliate orange. They are sufficiently hardy so that they doubtless may be grown with safety as far north as South Carolina, or 300 to 400 miles north of the present orange region. Some of the fruits produced are as large as the ordinary orange, but most of them are very nearly intermediate in size. They are very variable, however, in the first generation. At least five of the fruits that have been produced are juicy and valuable. It is not probable that they would be reproduced true to seed, but orange varieties are clons, and the different types will, of course, be normally reproduced by buds or grafts, so that from a practical standpoint it does not matter whether or not they would reproduce true through the seed. In the second generation it is probable that these different characters would split up, possibly according to Mendel's law, and it is likely that still more valuable varieties will be secured when a second generation has been grown. See Citrange.

Similar groups of valuable intermediate types of fruits have been produced by Wm. Saunders, until recently the Director of the Canadian Experimental Farms, by crossing varieties of the ordinary apple, such as the Pewaukee and Wealthy, with a very hardy cold-resistant crab (Pyrus baccata). Saunders has produced already numerous hardy intermediate types which bid fair to be of very great economic value, particularly in the cold regions of Manitoba and Saskatchewan (Fig. 648). Second generation seedlings of these valuable types may be expected to yield still more important improvements. The reproduction of such unfixed hybrids may be said to form the basis of fruit-culture, as all of the apple, peach, pear, plum, orange, lemon and grape varieties, as well as the varieties of small fruits, are of mixed parentage and do not reproduce true to seed. Most of the varieties ot these fruits are either known to be hybrids or are superior seedlings that have been selected and propagated. These latter, doubtless, in the greater number of cases were of hybrid nature as all of these fruits are normally cross-fertilized and natural hybridization is exceedingly common.

The same may be said of most flowers, such as carnations and roses, that are cultivated extensively for the cut-flower trade. Practically all of the varieties are unfixed hybrids.

The selection of bud-variations.

No consideration of the methods of plant-breeding would be complete without a mention of the improvements that can be produced by what may be termed the selection of bud-variations. While, in general, all buds of a plant are practically the same, as is shown by the fact that buds taken from the Baldwin apple almost uniformly produce Baldwin apples, yet there is considerable variation frequently in the product from different buds, and it is evident that bud-variations may be classified like seedling-variations, into fluctuations and mutations or the so-called bud- sports (Fig. 649). Hybrid plants also frequently, for some cause, show segregations of characters in different buds similar to the segregations shown in F2 hybrid seedlings. It would thus seem natural to suppose that these variations could be utilized in producing new varieties much as the similar types of seedling-variations are used.

In violets, for example, the propagation is normally by slips that are developed from different buds. These slips when grown into plants frequently show considerable difference, and B. T. Galloway and P. H. Dorsett, of the national Department of Agriculture, have demonstrated that by the selection of slips from plants which are very productive the yield in the number of flowers to the plant can be increased considerably. In the case of the orange, seedling trees are almost always very thorny, yet certain branches may show a tendency to be more nearly thornless, and by the selection of buds from such branches the thorny character of almost all the standard varieties has been reduced. By the systematic selection of vegetative parts, such as buds, slips, suckers, and the like, in many cases very important improvements could doubtless be secured, and the plant-breeder should have a thorough understanding of this method of improvement. In hybrids of mixed parentage, frequently a bud on one side of a plant will sport, showing different tendencies, and many of our new varieties of roses, chrysanthemums and carnations have been produced by the selection of such bud-sports. Many standard varieties of carnations have produced bud-variations that have proved valuable; the Lawson has given rise to the Red Lawson and White Lawson; the Enchantress has produced the Pink Enchantress and White Enchantress. The practice of exercising care in choice of chrysanthemum or carnation cuttings and of cions for fruit trees is, therefore, seen to rest on rational reasons.

Variations in the character of the seed from different bolls, in the case of hybrid cottons, are frequently found and may be of value to the breeder even in cotton that is propagated by seed. In the study of cotton, similar bud-variations have been found, showing in the lint characters of hybrids. In a number of instances, certain bolls have been found which produced much longer lint than other bolls on the same plant, and similar variations in strength and uniformity of length have been observed. Experiments indicate that such variations, which are doubtless to be classed as bud-variations, are inherited in considerable degree. This being the case even in seed-propagated plants, it becomes desirable to observe and search for bud-variations.

The importance of bud-selection in oranges and jemons has recently been called to attention by the investigations of A. D. Shame], of the United States Department of Agriculture. It has been found that groves planted with the Bahia or Washington Navel, which is grown extensively in California, frequently show a number of different types with reference to productiveness and form of fruit and that these conditions remain the same from year to year. The same has been found to be the case also in lemon groves, several distinct types not infrequently being produced on the same tree (Fig. 650). These barren trees, and trees producing poor fruit, greatly reduce the production of the grove and in many cases are a serious handicap. Evidence has been collected showing that when buds are taken from productive trees of good type they may ordinarily be expected to produce good types.
In experiments which have been conducted during the last six years in the selection of potatoes, it has been clearly demonstrated that, in a family of potatoes developed from a single tuber and thus positively known to be pure, low- and high-yielding strains can be produced by selecting from low- and high-yielding hills (Fig. 651). Such low- and high-yielding strains have now maintained themselves for three years in over thirty different cases representing work with eighteen different varieties.

The importance of bud-selection is only beginning to be realized and further data is necessary before it can be determined how important this is in different cases. The evidence now at hand, however, clearly indicates that this method of improving plants should be given careful consideration.
}}