CLOVER IMPROVEMENT

[ABSTRACT.]  IN 1928 the United States Department of Agriculture, through the cooperation of the State agricultural experiment stations, began a thorough search for red-clover stocks that had been grown continuously on the same farms or in the same communities for periods of 10 years or more without the introduction of seed from outside sources. These were planted in small observation plots and used as a basis for determining regional needs in red-clover-breeding programs for the humid Eastern States. It turned out that in the southern region red clover must be especially resistant to anthracnose; in the central region it must be both winter-hardy and disease-resistant; in the northern region it must be able to withstand a long period of winter dormancy; and in all regions resistance to powdery mildew is of varying importance. Breeding work based on these needs is now in progress, and some superior strains have already been introduced or are well advanced.

THROUGHOUT the world clover has been repeatedly referred to as the keystone of a permanent agriculture. Its introduction into agricultural use in the sixteenth century profoundly affected the entire system on which the old agriculture was based. As early as 1663, Yarranton wrote, in The Great Improvement of Lands by Clover, “for I perceive the land doth receive wonderful advantage by these leaves and branches; and as the leaves and branches, so the root doth very much contribute towards the enriching of the land.”  The improvement noted by Yarranton has been observed in every generation since that time, and to this has been added an increasing appreciation of clover as forage. Red clover, especially, is today the most widely adapted dual-purpose crop available to the farmer of the humid parts of the United States. Even older is the belief in clover as a lucky plant.  Systematic attempts at improvement, however, are of recent origin.

Many plants have been called clovers, probably because of a similarity in usefulness and appearance to the true clover; but the true clovers and the sweetclovers only will be considered herein.

The true clovers belong to the genus Trifolium, of which there are some 250 described species of annual, biennial, and perennial forms that are widely distributed. In nearly every continent there are found indigenous species that occur only in that continent. In general they inhabit cool, moist regions or their growth is confined to the season of the year when cool climatic conditions prevail. Only four species—red (Trifolium pratense L.), alsike (T. hybridum L.), white (T. repens L.), and crimson (T. incarnatum 1.)—are of primary importance and widely used, although several others are minor agricultural plants, in some cases of great importance locally.

Though recently considered a weed, sweetclover has had a phenomenal rise to the position of a forage crop of major importance in the United States. As an immigrant, it established itself along roadsides and railroad beds before its value was recognized. Sweetclover is now widely distributed over the world, but its native habitat appears to be in Asia Minor. Twenty species of sweetclover are recognized by Engler and Prantl. Three species—Melilotus alba Desr., M. officinalis Lam., and M. indica All.—are of importance in agriculture.

OF THE true clovers, red clover (Trifolium pratense) is by far the most important. In the wild state it ranges over most of Europe and far into Siberia. The plant was known to be generally cultivated in the Netherlands 370 years ago, and it is possible that there may have been a much older clover culture among the ancient Letts.

In the wild, red clover is an extremely variable plant. There are known to be early, late, smooth, hairy, prostrate, erect, and semierect forms. These forms can be found today, and it seems probable that one of them was the ancestor of the clover first used in agriculture, which was substantially like the double-cut clover of Europe today.  The many varieties that have developed since the introduction of red clover into England in 1645 have resulted from the action of local conditions rather than conscious selection by man. Many such more or less local forms still exist, but they differ physiologically rather than morphologically; that is, they differ in resistance to cold and disease or in yielding ability rather than in characters that can readily be distinguished by the eye. In the Netherlands, however, a variety occurs, known as Maas, that is characterized by the almost total absence of the crescent-shaped white spot on the leaflets; and the character of rough hairiness readily distinguishes the American clover from its European progenitor.

Nothing is definitely known regarding the origin of red clover now common in North America. It is known that red clover was grown in Rhode Island in 1663 and that in later years there was a constant importation of seed from England. The American red clover is therefore certainly descended from the cultivated red clover of England, which in turn was introduced from Flanders. Just how the American form developed its characteristic hairiness is not known, but it has been surmised that it may have been because of the continued attacks of leafhoppers, which are common on clover and which prevent seed production in the smooth more than in the hairy plants. The English clover contains a certain percentage of hairy plants that, on this theory, would persist and hence gradually develop a hairy form (fig. 1).

Figure 1.—Stems of Italian red clover (A), American (B), and English broad red (C), showing the hairiness of American, believed to have been derived from the English by natural selection.

CLASSES OF CULTIVATED RED CLOVER

In both Europe and North America the cultivated red clovers fall into two classes—(1) early, or double-cut, these giving two hay crops in a season; and (2) late, or single-cut, these giving but one hay crop in a season; and in most cases two or more forms are known in each class. Such forms are more numerous in the double-cuts than in the single-cuts and with a few exceptions are local or regional, rather than true varieties. Thus, in England, among double-cut clovers are the English Broad Red, Dorsett Marl, and Vale of Clwyd Early; among the single-cut clovers, Montgomery, Cornish Marl, English Late, and Vale of Clwyd Late. In the countries of continental Europe also, several regional varieties occur, distinguished by the name of the country or local area where the seed is produced, as Silesian, Hungarian, French, Italian, Swedish. Claims of superiority are made for all of these regional strains, and the evidence of comparative trials shows that in most cases each such regional variety is superior to others in the environment where it was developed. Some varieties, as the Mattenklee of Switzerland, are reported by investigators to be especially distinguished by persistence.

These strains have not arisen as a result of conscious efforts toward improvement of clover, but are rather the products of different environments acting on a highly variable plant. It seems probable also that in some cases, as in England, natural crossing with the native wild red clover may have affected the progeny of the cultivated form first introduced from the low countries.

In North America regional strains are present as they are in Europe and differ in productivity, winter hardiness, and disease resistance.  In the main, again, such characteristics have developed as a result of the action of local environments. In many States these strains have been grown continuously for 15 to 40 years without being mixed with other seed and have in many cases proved themselves decidedly superior. It must not be thought, however, that just because seed was grown in a certain State it has taken on a special character.  A given lot of seed may represent one generation or many, and the trade names “Ohio”, “Michigan”, etc., have little significance.

   Definite attempts at the improvement of clovers are of relatively recent origin. In the United States studies looking toward the selection of a superior red clover have, at one time or another, been started at several State agricultural experiment stations, but, with the exception of the Tennessee anthracnose-resistant strain, such work has never been brought to fruition. The Kentucky station has found a superior strain that is called Kentucky 101, but it did not result from conscious selection but rather is an old stock grown for many years under the same environmental conditions. Similar valuable stocks exist in other States and can doubtless be made the basis for future breeding work.

   In 1906 the Tennessee Agricultural Experiment Station selected a red clover resistant to the attacks of the fungus Colletotrichum trifolii S. M. Bain (fig. 2). Here again the form arose through the elimination of susceptible plants—man aided by seizing the opportunity to preserve a character brought out by the environment.  The same process of selection, with perhaps less conscious human aid, has probably taken place elsewhere, since lines of long standing, highly resistant to this disease, are found in Kentucky and Virginia.

Figure 2.—Strains of red clover resistant and susceptible to southern anthracnose, caused by Colletotrichum trifolii: A, Tennessee anthracnose-resistant strain; B, Italian red clover. Arlington Experiment Farm, Arlington, Va.

In 1928 the United States Department of Agriculture, through the cooperation of the State agricultural experiment stations, began a thorough search for red clover stocks that had been grown continuously on the same farms or in the same communities for periods of 10 years or more without the introduction of seed from outside sources. Approximately 75 strains or lots were secured, many of which had been grown for a continuous period of 40 years. These were planted in small observation plots, in cooperation with the Kentucky, Ohio, and Iowa Agricultural Experiment Stations. As a result of these studies and previous observations on the adaptation of foreign and regional strains of red clover, three regions of adaptation have been recognized in the humid eastern part of the United States, namely, the southern, central, and northern.

The chief limiting factor to which the plants must become adapted in the southern region is disease, such as that caused by southern anthracnose (Colletotrichum trifolii). In addition the plants must be able to survive frequent great changes of temperature during the winter months, and high summer temperatures. This southern region is composed of the States of Tennessee, Kentucky, and Maryland; the piedmont area of North Carolina and Virginia; parts of New Jersey; the southern parts of Ohio, Indiana, and Illinois; and a section of Missouri.

The central region includes the territory north of the southern region to approximately the latitude of the Illinois-Wisconsin boundary. Red clover adapted to this region must be disease-resistant and winter-hardy, since low temperatures with little snow protection are often met.

The northern region extends from the Wisconsin-Illinois line to the Canadian border. In this region no serious diseases are encountered, but the plants must be capable of tolerating a long period of dormancy, which at times may be accompanied by very low temperatures.

Throughout the three regions attacks of the powdery mildew fungus (Erysiphe polygoni DC.) and the potato leafhopper (Empoasca fabae Harr.) occur regularly with fluctuating intensity. In general the injury produced by the clover root borer (Hylastinus obscurus Marsham) is one of the principal causes of red clover killing-out during the second winter. In each of these regions the severity of the adverse factors varies from season to season and according to location.

After several years of study of the old strains in the central region of the red clover belt, superior strains were selected as a basis for breeding studies. Several hundred plants of each of 10 strains were studied, and it is of interest that the percentage of superior plants in these strains ranged from 22 to 58, indicating the need of improvement.

In cooperation with the Kentucky station, seed of Tennessee anthracnose-resistant and Kentucky 101, two superior strains, was planted under controlled conditions, and the plants were artificially inoculated with spores of southern anthracnose (Colletotrichum trifolis).  In these tests the percentage of resistant plants of Kentucky 101 and of Tennessee anthracnose-resistant were, in 1925, 57 and 41, respectively, while only 10 percent of the plants of an old strain from Wisconsin remained alive.

In cooperation with the Wisconsin station, studies are in progress to develop lines resistant to powdery mildew (Erysiphe polygoni). Several lines have been developed that are resistant to two physiologic forms of the fungus (fig. 3). Both self-fertile and self-sterile mildew- resistant lines have been developed, but all of these are lacking in vigor and require further selection and hybridization.

Figure 3.—Resistance of red clover to the powdery mildew disease, caused by Erysiphe polygoni: a, Seedling plants from commercial seed, covered with mildew; b, seedling plants from mildew-resistant lines. University of Wisconsin, Madison, Wis., 1935.

Investigations on the improvement of red clover are in progress at the Minnesota, Pennsylvania, Tennessee, Indiana, New Jersey, and Idaho stations, and at the Kentucky, Wisconsin, and Illinois stations in cooperation with the United States Department of Agriculture.

Attempts to develop improved clovers are being made in nearly every country in Europe and also in Australia and New Zealand.

In Canada two double-cut forms and two single-cuts are the result, of deliberate selection. The Ottawa selection, the result of a mass selection process begun by the late M. O. Malte and carried forward for many years by the late R. I. Hamilton, is claimed to be of superior winter hardiness and is now coming into general use in Ottawa. An early or double-cut strain, called the Dollard, of superior winter hardiness, has been produced by J. N. Bird at MacDonald College, Quebec.

The Altaswede red clover, selected by G. H. Cutler from a Swedish strain, is said to be widely used in Alberta, but being a smooth clover it is not satisfactory for use in the eastern United States. The Manhardy, another smooth strain of single-cut, was selected at the Manitoba Agricultural College by W. Southworth out of various escaped and cultivated lines and is believed by Canadian authorities to be the hardiest strain of red clover in Canada.

In England the firm of Sutton & Sons selected a clover the seed of which was pure yellow instead of purple or yellow and purple. The stock came from yellow seed picked out of Chilean red clover. The effort to place the variety on the market was abandoned, however, since it was found impossible to prevent crossing in the field, and the purple color again appeared in the seed.

The firm of Gartons, Ltd., in England, began development work as early as 1890, but nothing of value has resulted from these efforts nor from the efforts of other private workers. Since the establishment of the Welsh Plant Breeding Station, systematic work has been undertaken on the study of red clover. Breeding of red clover under the direction of R. D. Williams has been important. So far one improved variety, S-123, has been offered to the public. This was released in 1936 and is believed to be useful for 2- and 3-year meadows.  The following strains, built up through crossing various lines, are being tested but have not been released: S-87, S-106, S-110, and S-141.

Work of a similar kind has been undertaken in the Irish Free State, but results are not yet available. In New South Wales, Australia, an effort is being made to select a more persistent type out of a local strain known as Santhia. In New Zealand a strain known as Runciman’s red clover is said to be truly perennial and of great value in pastures.

Martinet, and later Lindhard, tried to produce a variety with a short tube or corolla. Lindhard (1921) claimed that honeybees visited this variety freely and that it produced seed abundantly.

In Denmark various selected strains have been introduced as being more valuable than the local parent stocks. These are Øtofte early, semilate, and late. The Øtofte late was selected from a local strain of red clover. The strain called @Øtofte semilate was bred from Hersnap, another improved late clover used in Denmark. There is also Tystofte 40, a late form, and others.

In Germany various attempts have been made at improvement, notably at Bonn-Poppelsdorf, but no record has been found regarding results. Among private breeders was A. Dreger, who commenced his work in 1908 and produced violet-seeded and yellow-seeded varieties. According to Nessler (1931) the only bred clover in use in Germany is Original Lembke’s red clover. This is an early red clover developed by Lembke in Wismar.

In Sweden such local strains as Norrum, for dry conditions; Karaby, produced by L. Karaby; and Mardal, said to be resistant to Sclerotinia, have become locally established. The experiment stations have also introduced Svalof Purebred red clover, produced by H. Witte, and Mercur red clover, produced by N. Sylven, both of these strains being valued for yield and resistance to Sclerotinia and Tylenchus. Mercur was developed by mass selection from a local strain called Spannard. Weibwell’s Smaragd, out of a Finnish local strain, and Géta Red, out of middle Swedish strains, are also in use. The improvement work in Sweden has been carried on by mass selection following intercrossing of various Swedish strains, or by mass selection out of a single strain.

F. Chmelar, of Brno, in correspondence, advises that original Dregers, out of Bohemian stock, and Mattenklee Rekord, selected by K. Holy out of Swiss Mattenklee, are bred varieties grown in Czechoslovakia.

In the Union of Soviet Socialist Republics special attention has been paid to native strains of wild red clover. These wild strains are said to vary in earliness, winter hardiness, production of aftermath, leafiness, and resistance to disease. The opinion is held by Lissitzyn and certain other Russian workers that these wild red clover strains rather than selections from existing cultivated forms should be made the basis of improvement.

Red clover heads are composed of from 50 to 150 florets, which in themselves are complete units of reproduction. The florets develop and open in an ascending order from the base to the top of the head.  The pistil is usually curved, the stigma extending beyond the anthers, though florets have been found in which the style was greatly shortened, thus placing the stigma deep in the corolla tube (fig. 4). The ovary of red clover has two ovules, one normally developing upon fertilization and the other aborting. Plants have been found with a high percentage of pods having two seeds.

Figure 4.—Longitudinal section of a red clover flower, showing how anthers and style lie in the keel. (From Siebler and Schroter, The Best Forage Plants, pl. 24, fig. 2.)

The anthers shed their pollen in the bud stage shortly before the floret or individual flower of the head opens. At this stage the pollen does not scatter freely. After the floret is open the stamens and pistil remain in the keel, and the pistil becomes exposed only after the floret is tripped. When the pressure causing the tripping is released, the stamens and pistil return to the keel unless the floret is ruptured. Fertilization occurs from 18 to 50 hours after pollination, the time depending on the temperature. When the atmospheric temperature is high, fertilization occurs much earlier than when it is cool. The degree of success of pollination can be approximately determined 2 to 3 days after fertilization.

Self-sterility in red clover has generally been recognized since Darwin caged plants to exclude bees. The more recent studies of Fergus of Kentucky, Kirk in Minnesota, and Williams in Wales have clearly established the fact that nearly all the plants are self-sterile. However, lines have been established that are homozygous or pure for self-fertility, and Johnson in Minnesota has inbred one line for nine generations without materially decreasing the self-fertility. One of Fergus’ self-fertile lines maintained its self-fertility for six generations, after which the line was lost.

Self-sterility in red clover is due to the slow growth of the pollen tube in the style, the ovules disintegrating before the generative nucleus reaches the egg. Recent investigations by Williams (fig. 5) in Wales indicate that this phenomenon is controlled by a series of many sterility alternative genes or allelomorphs. Whenever the same sterility factor is present in the pollen and in the female plant, growth of the pollen tube is inhibited. He has also shown that in self-fertile lines of red clover the presence of a noninhibitory self-fertility factor or gene permits the pollen tube to grow at the same rate as it does in the case of pollen from an unrelated plant.

Figure 5.—R. D. Williams, Welsh Plant Breeding Station, Aberystwyth, Wales, who has devoted much attention to the subject of fertility and sterility of red clover.


SELF FERTILITY

It is true that upon self-fertilization many red clover plants do produce a few seeds, but there is a distinct incompatibility in the succeeding progeny and the line cannot be maintained beyond the second or third generation. This phenomenon may be called pseudo- self-fertility as distinguished from true self-fertility.

Self-fertilization in red clover is accompanied by a decided decrease in vigor of the plants. This decrease is most pronounced in the first and second generation and differs in different lines. Apparently Johnson’s inbred line in Minnesota has maintained more vegetative and reproductive vigor that any other.

The number of true self-fertile plants in red clover is relatively small. Out of many hundred self-pollinations only a few self-fertile lines have been secured, though a large number of pseudo-self-fertile lines have been found. In connection with red clover breeding studies in the Department of Agriculture, from a spaced population of approximately 4,000 plants, 500 superior open-pollinated plants were selected; each of these 500 plants was self-pollinated, but not a single true self-fertile progeny was obtained.

The structure of the red clover floret prevents cross-pollination by wind, since the anthers and pistil remain enclosed in the keel unless artificially tripped (fig. 4). Under natural conditions pollination is done by insects, nearly all of which are bees (order Hymenoptera).  The bumblebees (Bombus) and the honeybees (Apis) are the principal pollinators, although at certain times and places bees of other genera, such as the ground bees (Tetralonia and Melissodes) and the leaf-cutter bees (Megachile), contribute to cross-pollination.

Bees visit the red clover floret for nectar and pollen or both, tripping the florets and transferring pollen from plant to plant, thus constantly maintaining the condition of mixed inheritance in the species. Other insects, such as moths, are constantly seen on red clover heads, but they do not come in contact with the pollen and therefore do not effect cross-pollination.

There has been considerable controversy as to the extent to which pollination can be accomplished by honeybees. Discussion has centered upon the fact that the tongue of the honeybee is not long enough to reach the nectar. The literature on this subject is voluminous and cannot be reviewed here. More recent investigations clearly indicate that honeybees visit red clover principally for pollen and seldom obtain nectar, but regardless of what is obtained, pollen is transferred and cross-pollination is effected.

In the breeding of red clover inherent difficulties exist that have greatly retarded advancement in the improvement of this crop. These factors are chiefly the small size of the red clover floret and the fact that the plants are self-sterile. In general the method followed in the past in breeding for an improved variety of red clover has been to start with a valuable local stock and attempt to secure a better strain by mass selection of the most desirable types. Because of the necessity for cross-pollination, red clover is in a hybrid condition and the progeny of selected plants will continue to produce a certain percentage of undesirables. These may be eliminated in some cases by continued selection and in others by unfavorable factors in the environment. The latter is the case with the anthracnose-resistant strains introduced in the southern region of the red clover belt and the winter-hardy strains of the northern region, since they are naturally subjected to disease or to severe winter conditions. All varieties successfully established in the United States and Canada to date have been developed by the aid of this natural selection.

Special inoculation beds with favorable conditions for development of the disease organisms have proved to be advantageous in studies on disease resistance. Experience has shown that field inoculations are not, always successful, and there is always the possibility that surviving plants may have escaped inoculation rather than that they are resistant. The worker must also consider the susceptibility of the host as related to age and environment.

In details, the method of making crosses varies with the individual investigator. Toothpicks, camel’s-hair brushes, cards, and crooked needles have all been successfully used for the transfer of pollen, Since most of the red clover plants are self-sterile, reciprocal crosses can readily be made without the emasculation or removal of anthers necessary with many plants.

Bumblebees from which pollen has been washed have been successfully used by Williams for cross-pollination, and this method offers the distinct advantage of securing large numbers of seeds in any specific cross. Lindhard caught queen bumblebees in the spring and developed the broods in special boxes that could be moved to cages in which the plants were enclosed. Natural nests have been transferred to artificial domiciles and moved about at will in studies by the Bureau of Plant Industry. Honeybees have been successfully used when many plants were to be cross-pollinated.

Williams (17), of the Welsh Plant Breeding Station, follows three methods: (1) Strain building—selecting superior plants of old stocks and combining them by controlled crossing; (2) brother-and-sister matings—combining desirable sibs of the same families for several generations, eliminating plants with undesirable recessive characters after each crossing, and later outcrossing to unrelated families that have been sib-crossed; (3) dialed crossing—intercrossing several lines with one another to determine the combinations producing the best progenies, followed by intercrossing among such progenies to form a new strain.

The development of mildew-resistant lines in Wisconsin has resulted from selections made from artificially inoculated beds. The progeny of these selections were sib-crossed and selfed to secure lines homozygous or pure for mildew resistance. The crossing of unrelated lines resistant to mildew is in progress to restore vigor.

The results of studies by Williams and by the Bureau of Plant Industry clearly show the effects of sib-crossing. Loss of vigor with succeeding sib crosses for several generations is usually evident.  Whether full vigor can be restored and maintained by combining a limited number of unrelated lines remains to be proved. Experience has shown that the growing of large populations of spaced plants is essential if progress is to be made. ~While controlled crossing may be accomplished in the field, a greenhouse is a valuable asset. The necessary manipulation can be more easily performed and a higher percentage of seed may be expected in the greenhouse than in the field.

The isolation and use of self-fertile lines in the breeding of red clover, while discredited by some investigators as offering little promise, has the distinct advantage of developing plants that are pure (homozygous). The loss of vigor by making sib crosses for several generations is practically as great as that resulting from selfing, and self-fertile lines are more readily maintained. The selfing of one head per plant is sufficient to determine whether it is self-fertile.

In Europe the problem of isolation of stocks in the early stages of development has been met by planting in an isolated field among other plants. Lang planted his red clover in the midst of a field of other plants blooming at the same time so that the bees might work on these plants first and thus have any foreign red clover pollen cleaned off before they arrived at the red clover plants. Dreger isolated his breeding stock in a field of Vicia villosa. By this means he claims to have produced a strain of red clover with uniformly violet seed.

Williams and Evans found that contamination by extraneous pollen depended on two factors—(1) the distance separating the bred strain from other red clover plants and (2) the profusion with which the strain flowers. In their experiments it was found that when there was little bloom on the pure strain the percentage of contamination was great, no matter how well isolated the plot.  On the other hand, with abundant bloom relatively little contamination occurred even when the plot was not well isolated.

WHITE clover, a common inhabitant of lawns, pastures, and roadsides, is widely distributed in every continent of the world, with a natural distribution probably as great as that of any other plant of the legume family. Moist and cool situations are its most favorable habitat and under such conditions growth is continuous. Though it is believed not to be indigenous to North America, it was brought over early and the year of its introduction is not known.

The white clovers of agricultural value have been grouped as wild, cultivated or Dutch, and Giant or Ladino. Five other varieties are mentioned but are only of botanical interest.

In form these groups differ chiefly in plant size, the English wild white clover being smaller in all vegetative parts than the Dutch, and this again smaller than the Ladino. In England the wild white clover differs also in being more persistent than the Dutch white clover, which is often short-lived. As a group, the native or wild white clover is distinguished from the Dutch white clover and the ladino white clover by the presence of a cyanophoric glucoside.  This is an organic substance present in many herbage plants and which in the course of digestion forms hydrocyanic acid. The quantity present in white clover is small and harmless. The more persistent plants have a high glucoside content, while the short-lived Dutch and the Ladino have but little or none.

The development of varieties and strains of white clover has resulted from the action of environment on this variable species. Thus in the United States it is possible to distinguish the strain produced from Louisiana seed from that produced from Wisconsin seed, the Louisiana strain being more persistent under summer conditions but slightly less winter-hardy than the Wisconsin strain.

White clover was introduced into New Zealand from England, but the best white clover produced as wild white in New Zealand today differs from the wild white of England in its larger growth.

In New Zealand, where intensive work has been done with white clover, Bruce Levy has distinguished four strains of New Zealand white clover, differing in productivity and persistence, the best strain being like English wild white clover and the least satisfactory form like Dutch white clover. In Denmark two distinct varieties of white clover are grown, Morsg and Stryng. Of these the Morse appears to be the more permanent and the better seed producer, but the Stryng is the better yielder of herbage. In Sweden a very persistent strain called Svea has been produced at Svalsf and is now on the market. In Finland the Tammisto strain, which has proved hardy and a good yielder, has been developed from native Finnish stocks.

Though no critical studies have yet been made, observations indicate that the white clover growing in old pastures in the northeastern United States has the same growth habit and persistency as the English wild white clover. A common descent may be presumed, since white clover was brought to this country by the early settlers.

Ladino white clover was first discovered in northern Italy, but its exact origin is unknown. It does well under irrigation but has not been successfully established in the eastern United States.

The difficulties experienced in the breeding of white clover are similar to those in the breeding of red clover, and the same procedure is applicable. White clover, like red clover, is extremely variable in leaf size, color, and markings; flower and seed color; size of runners; and persistence. Combinations of one or more characters in individual plants have been observed from time to time. In many cases these have been perpetuated by vegetative propagation, and in a few cases strains have been developed that are homozygous for certain characters. All the strains to which reference has been made are the result of natural selection with only incidental help from man.

At present special breeding work is being conducted at the Weraroa station in New Zealand, where the Weraroa strain was isolated as no. B-95. It is said to be truly perennial and a good yielder. At the Welsh Plant Breeding Station a variety, S-100, was released in 1936 and is described as being more productive than commercial white clover and as lasting longer. It was produced from Dutch, New Zealand, and wild white clover and is intended for use in 1- to 3- year meadows. Another variety, S-99, has not yet been released.  Breeding work is also being carried on at various stations in Germany, Denmark, Czechoslovakia, and Sweden.

Experiments by Williams (17), Erith (3), and others have clearly shown that self-sterility in white clover is not so marked as in red clover but is the result of a similar phenomenon. The possibility of developing self-fertile lines is, therefore, slightly greater. A pronounced reduction in vigor results from inbreeding, with the appearance of dwarf and chlorophyll-deficient seedlings and other recessive characters.

CRIMSON clover—the most important winter annual of the true clovers—was first introduced in the United States in the early part of the nineteenth century, and the Patent Office made a distribution of the seed in 1855. This species is self-fertile and is less variable than the clovers previously discussed. The florets are not self-tripping and for maximum seed crops require insect visitation.

Several varieties of crimson clover are recognized, differing in date of bloom and in color of the flowers. Vilmorin-Andrieux & Cie., of Paris, refer to Extra Early Crimson, Early Crimson, Late Crimson, Very Late White Flowering, and Extra Late Crimson Flowering. A difference of 4 to 5 days in time of maturity is shown between each of these varieties in the order named. An early white variety is also said to exist.

In the United States a local strain, developed in northern Georgia and known locally as Pitt’s clover, is said to have given better results in Georgia than commercial seed.

The first recorded introduction of alsike clover into the United States occurred about 1839, when the Genesee Farmer, an agricultural journal, made a small distribution. No recognized varieties exist, but regional strains give varying results. InSweden the native strain yielded more than that from Silesia or that from North America. In preliminary studies in the United States, differences of this sort have been noted.  The plants are generally self-sterile, though Wilson in Minnesota has shown that self-fertility exists; loss of vigor follows inbreeding.

The clovers known as subterranean, berseem, and Persian are all annuals, the first two distinctly winter annuals, and are used for hay grazing. Subterranean clover is little used in the United States.  In Australia several strains of subterranean (Trifolium subterraneum L.) differing chiefly in date of maturity, are known, and in Egypt and elsewhere three or four distinct strains of berseem (T. alexandrinum L..) are recognized. Strains of Persian clover (T. resupinatum L.), differing in maturity, have been selected by workers in Turkistan. In none of the other true clovers are strains recognized, and no special improvement work, so far as known, is being carried on with these.

ALTHOUGH a relatively new crop, sweetclover offers one of the most promising possibilities for forage improvement. The different species appear on the whole to be self-fertile, although some are not self-pollinating. Large numbers of honeybees visit the florets, effecting a certain amount of cross-pollination and thus increasing variation among the plants.

In general, the sections where sweetclover is principally grown are the north-central region and the Great Plains States. This territory may be divided into two areas of adaptation—(1) the area east of the Nebraska-Iowa boundary and (2) the Great Plains States. In the eastern region resistance to a disease complex is of first importance in connection with the development of more palatable late-maturing strains for pasture. In the Great Plains States diseases as yet are of minor importance, but drought-resistant palatable strains for hay and pasture are needed. A coumarin-free strain would be desirable everywhere, since the presence of coumarin affects the palatibility of the herbage and, as shown later, is probably related to the development of a toxic substance in spoiled hay.

The sweetclovers grown in the United States belong to the genus Melilotus. They infiude white sweetclover (M. alba), yellow sweetclover (M. officinalis), and sourclover (M. indica). Redfield yellow, an early introduction of the United States Department of Agriculture, appears to be M. suaveolens Ledeb., though in many respects it resembles M. officinalis. The first species mentioned includes both biennial and annual forms, while M. indica is an annual, and of M. suaveolens and M. officinalis the biennial form only is known.

Of the white sweetclover group, the Grundy County, Hubam, Arctic, Alpha, Iowa Late White, Ohio Evergreen, and Madrid White varieties have characteristic qualities and are in use to a greater or less extent. Grundy County White was first noted in Grundy County, Ill., in 1917, but is of unknown origin. It is early, of erect habit, with slender stems, and it is not so tall as the common biennial white sweetclover. Arctic was named by J. Bracken at the University of Saskatchewan, Saskatoon, Canada, and was derived from a Siberian importation brought in by N. E. Hansen. It is especially noted for winter hardiness and behaves as a dwarf variety in the central part of the United States, although making a much greater growth in Canada.

Madrid White, introduced by the United States Department of Agriculture through seed received from the Botanic Garden in Madrid, Spain, is a large-growing variety characterized by early seedling vigor and tolerance of the first year’s growth to frost. Hubam is an annual mutation from Melilotus alba and was first brought to general attention by H. D. Hughes, of Iowa, in 1917.  This variety resembles common biennial white sweetclover, except that it blooms and seeds freely the season of planting and does not survive a second year.

Alpha sweetclover is a variety developed by L. E. Kirk at the University of Saskatchewan. The first plants were found in a field of Arctic sweetclover in 1924, and Kirk developed the variety by a process of inbreeding and selection. The plants of the Alpha variety branch profusely at the crown, the slender stems being leafy and resembling alfalfa. It is a dwarf variety in the central part of the United States and is not adapted to that region.

Iowa Late White is a large-growing, late-maturing selection made by the late F. S. Wilkins, in 1935, at the Iowa Agricultural Experiment Station, from a lot of seed secured from an Illinois farmer. Ohio Evergreen is similar in growth habits and maturity to the Iowa Late White. Selections of this variety were made by J. B. Park, of the Ohio Agricultural Experiment Station, as a result of several years of mass selection of desirable types secured from roadside plants. A selection made by the Illinois Agricultural Experiment Station and called Illinois No. 8 is another late-maturing uniform strain selected from wild plants.  Selections made by Brink at the University of Wisconsin were developed in the effort to improve disease resistance. Hayes, Johnson, and Doxtater, of the Minnesota station, have made selections from Alpha and commercial white, resulting in a strain called Minnesota No. 1 and a strain of Alpha that is more disease-resistant than Alpha from Canada. In the Pacific Northwest the growing of sweetclover west of the Cascade Mountains was unsuccessful until the development of a strain resistant to what is believed to be stem rot. This variety was developed by H. A. Schoth, at Oregon Agricultural Experiment Station, over a period of years, by growing successive generations in the same soil, allowing the disease organism to eliminate the susceptible plants.

Less attention has been given to the selection of varieties from yellow sweetclover (Melilotus officinalis) than from white sweetclover.  In general, the varieties of yellow sweetclover are less productive and mature earlier than those of white sweetclover. Many varieties, such as Albotrea, Madison County or Switzer, and St. Louis Valley, are very much like the commercial yellow. Zouave, developed at the University of Saskatchewan, is more erect the first year than the commercial yellow and has seed that is densely mottled with purple.  Daghestan Yellow, introduced from the Province of Daghestan, Transcaucasia, and having large and rounded leaflets, appears to be a form of M. suaveolens.

Redfield Yellow, a variety received by the Department of Agriculture from Manchuria in 1915 as Melilotus alba, but which later proved to be a yellow form, was developed and named at the Department’s forage-crop field station at Redfield, S. Dak. The fact that Redfield Yellow crosses readily with M. alba while M. officinalis does not cross with alba and the further fact that Redfield Yellow does not cross with known varieties of M. officinalis appear to confirm the classification as M. suaveolens. Madrid Yellow, introduced by the Department from seed received from the Madrid Botanical Garden, has vegetative characteristics similar to those of Madrid White, previously described. The character of early seedling vigor makes both the Madrid Yellow and the Madrid White superior for Great Plains conditions, where early establishment is of the utmost importance. In addition to the above-mentioned varieties of both white and yellow sweetclover, other local strains are in existence, developed largely by the continued growing of strains over a period of years.

Definite breeding work has progressed more in Canada under the direction of L. E. Kirk than elsewhere (fig. 6). Kirk has made extensive studies on pollination, selfing, and crossing. Improvement work in Canada has been directed toward "the production of winter-hardy, fine stemmed, leafy, disease resistant and more palatable sorts.”  Three methods are used—(1) isolation of desirable plants or of plant groups so as to provide for immediate increase of a superior form; (2) isolation of desirable forms through selection within inbred lines; and (3) hybridization, to combine the desirable features of the best inbred lines (fig. 7). Studies of variations in coumarin content of individual plants have given promising results. Lines have been selected that appear to be approaching homozygosity for low and for high coumarin content.

Figure 6.—L. E. Kirk, Dominion agrostologist, Dominion of Canada Experimental Farm, Ottawa, Canada, who has pioneered in the genetics of sweetclover.



Figure 7.—Segregating sweetclover line, showing branched dwarf character in center, compared to normal growth on each side. Arlington Experiment Farm, Arlington, Va., 1936.

The Wisconsin station, in cooperation with the Department of Agriculture, has been studying the relationship between coumarin and palatability and toxic properties of spoiled sweetclover hay.  Results indicate that coumarin is intimately connected with the toxic principle and is one of the important factors in the palatability of sweetclover for pasturage.  Hybridization experiments are in progress within and between lines of Melilotus alba, M. suaveolens, M. officinalis, M. dentata Pers., and other species.

At the Minnesota station various lines have been inbred for 4 years. In agreement with the work in Canada, selfing has not resulted in any material reduction in vigor in either common biennial white clover or Alpha. The development of a variety of low coumarin content and high yield is the objective of this work.

In Kansas, Washington, and Texas, and at Guelph, Canada, selection of desirable sweetclover lines is in progress. The Nebraska and Illinois stations, in cooperation with the Department, have established nurseries for selection and breeding looking toward the development of varieties having superior value. In West Virginia various selfed lines are being grown with a view to securing one that will thrive at a higher acidity level than that at which sweetclover now does well. The Idaho station reports the development of a sweetclover having purple seeds, a crown similar to alfalfa, and nonshattering pods.  A report for 1934 stated that none of the purple-seeded types proved homozygous for this character and that the plants were segregating for numerous other characteristics.

The Russians have been interested in the breeding of sweetclover, especially at the Institute of Plant Breeding, at the Black Soil Regional Plant Breeding Center, at the Maikop and the Detsko-Selo stations in the north Caucasus, and at Saratov. Both commercial varieties and native wild strains are said to be used, various forms having been isolated to serve as breeding stock.  Attention is being paid to variations in coumarin content.

At the Kaiser Wilhelm Institut für Züchtingsforschung, Müncheberg, Germany, under the direction of M. Ufer, many thousands of plants from different sources have been grown and studied for relative freedom from coumarin. Of these, 51 individuals were selected.  Among the species studied were Melilotus alba, M. officinalis, M. wolgica Poir., M. dentata, and M. indica. This work was reported in 1934, but no later information is available.

Pollination of sweetclover under natural conditions is effected principally by honeybees, except insofar as the species, varieties, or individuals are spontaneously self-fertilized. Various workers have studied this problem, but only a brief summary of the information as given by Kirk and by Kirk and Stevenson can be given here. While Melilotus alba is generally highly self-fertile, three groups have been recognized: (1) Plants in which nearly all florets produce seed without manipulation—spontaneously self-pollinated and self-fertile; (2) plants that do not, produce seed without manipulation—self-fertile but, not, self-pollinating; and (3) self-sterile plants. The varieties of M. officinalis are not spontaneously self-pollinated, but are not completely self-sterile as is commonly thought. When the stigmatic surfaces are scratched in selfing, some seed is formed, though the proportion is not so great as when the florets are open-pollinated. The Redfield Yellow variety is spontaneously self-pollinated and self-fertile, and this is also the case with many lines of M. dentata.

With the advent of more extensive genetic studies of plants, an occasional investigator became interested in the inheritance of particular characters of red clover. For the most part, however, the inherent difficulties occasioned by the self-sterility of the species and the manipulation of the florets in making crosses discouraged the initial efforts. Until recent years the published reports of such investigations were principally confined to general statements of what appeared to be evident.

Flower and seed-coat color of red clover were the basis of most of the early observations. De Vries (14) described the selection of a line that in the fifth generation had a large proportion of leaves composed of seven leaflets. He also reported that white flower color was recessive to red, being inherited in a simple ratio.

In 1903 Fisher likewise developed lines in which 85 to 100 percent of the progeny had yellow seed coats and another line in which 86 to 99 percent had dark-violet seed coats. In 1912 Kajanus (6) studied a blue-flowered red clover plant and reported that blue was recessive, being inherited in approximately a 15:1 ratio. He also stated that red flower color was dominant over white. In 1921 Witte (20) reported finding a white seeded red clover in which the white seed was associated with white flower. When this plant was crossed with those having yellow seed, the F₂ generation segregated in a 3:1 ratio.

In any large population of red clover plants, many have a crescent-shaped leaf marking while in others no markings are evident.  Gmelin’s (5) studies in 1916 showed that the presence of leaf markings was dominant. His work also confirmed the report of Kajanus (6) that white flower color was recessive to red. Fruwirth also studied flower color, counting the number of plants producing white and red flowers. The more recent work of Wexelsen (16) on the presence and size of leaf spot, leaf color, anthocyan pigment and its presence in stems and stipules, flower color, and resistance to mildew indicates an increasing interest in the inheritance of characters of red clover.  Nijdam (11) also investigated incompatibility, seed-coat color, flower color, chlorophyll difference, sterile stamens, and dwarfs. These studies include the occurrence, variation, and segregations of crosses.  Segregation ratios were reported.

With the establishment of the Welsh Plant Breeding Station an orderly investigation of red clover breeding began. Experimental proof by Williams and Silow (19) of sterility and fertility relationships of red clover, mentioned in a preceding section, affords a basis on which genetic studies may proceed soundly. The comprehensive studies of Williams (18) on flower color clearly show the presence of additional epistatic recessive factors with linkage relationships to the sterility genes, a condition not unexpected when the characters of large populations are critically examined. Besides the characters mentioned above, many investigators have observed the occurrence of dwarf forms, chlorophyll-deficient seedlings, abnormal flower-head development, and other variations.

With the increased interest in cytological study as related to genetics, and its apparent usefulness in interspecific hybridization, studies on chromosome number and morphology have been undertaken by a few investigators, notably Bleier (1) Karpechenko (Karpetschenko) (7), and Wexelsen (15). In the majority of cases these investigators agree on the number of chromosomes, particularly in the most important species. Differences of opinion exist, however, on some points, and additional studies, with improved methods, are needed for a clearer understanding. The basic numbers of chromosomes are seven and eight, with multiples thereof, but species belonging to a given subsection of the genus Trifolium do not always have the same number of chromosomes. Ascherson and Graebner record various alleged hybrids as having been found in Europe, but experimental interspecific crossing has invariably been unsuccessful. A table of chromosome numbers of all species of Trifolium for which data are available is presented in the appendix.

Genetic studies on the inheritance of different characters of white clover are in the developmental stage. In a preliminary report on a few crosses made between several varieties, Erith (3) states that leaf color and flower color are inherited as a single-factor difference. Normal green leaves are dominant to purple, and pink flower color is dominant to white. Cytological studies by Wexelsen (15) on two varieties—English wild and Ladino—show no difference in the chromosome numbers. Those of Ladino are, however, much larger, though considerable variation in this regard exists between plants of the English wild white clover. In crosses between varieties, the chromosome size of the F₁ generation was found to be intermediate between those of thé parents.

As indicated in the preceding section, improvement of sweetclover has been principally concerned with the introduction and trial of various sweetclovers from the Old World. The occurrence of natural hybrids and mutations has stimulated interest in characters and their inheritance.

In breeding sweetclover, emasculation is necessary unless the female parents carry a recessive character that may readily be determined in the F₁, generation. Kirk’s water-suction method for greenhouse studies and Savage’s automobile vacuum method for field emasculation, as described in the article on Alfalfa Improvement in this Yearbook, are ingenious devices that have been successfully used. In view of the numerous hazards in breeding work, the need of greenhouse facilities cannot be overemphasized. A plant generation can be grown in the greenhouse in the winter either by placing the young seedlings under continuous light or by freezing seedlings a few weeks old and then placing them under a long-day exposure. These methods have been successfully used by Kirk in Canada and by Johnson in Minnesota.

The appearance of Hubam and of the Alpha form undoubtedly stimulated inheritance studies. Smith concluded from his studies that the annual character was dominant and counts of segregating progeny indicated a close agreement of a 3: 1 ratio. Elders, Kirk (9), and Clarke (2) have shown that the Alpha type is recessive to normal and is inherited in a ratio of 3 : 1. Clarke also reports the presence of a recessive spreading dwarf character and of two factors for pale-green seedling. Linkage between pale green and spreading dwarf is evident.

Using the F₂ and F₃ lines developed from a natural hybrid between Melilotus alba and M. suaveolens (Redfield Yellow), Kirk (8) reports that the inheritance of flower color is in close agreement with a three-factor hypothesis, two factors for cream, varying in intensity, together giving yellow, and a third factor that inhibits the action of factors for cream and yellow giving white. Segregations in several of the above lines, however, indicate a deficiency of yellow plants. Reference to seed-coat color has been the basis of determining whether seed is from M. alba or M. officinalis. Various workers have considered the clear seed coats of M. alba as a distinguishing characteristic differing from those of M. officinalis in which a varying percentage of seeds are flecked with purple. Kirk and Stevenson have reported the finding of speckled seed coats in a line of Alpha (M. alba). Previous crosses indicated that color markings are dominant over clear seed-coat color.  Kirk and Armstrong (10) also describe a mutation characterized by a reduction of the size of leaf with a pattern dissimilar to normal. This abnormality is also expressed in size and pattern of corolla. The plants are dwarfed, self-sterile, and cross-sterile to pollen from normal types.

The characters in sweetclover are as diverse as in other plants; purple seed coats, green cotyledons, dwarfs, and chlorophyll-deficient seedlings, germinable seed not needing scarification, and floret abnormalities have already been observed.

The fact that all species of Melilotus have the same basic number of chromosomes, supported by the occurrence of natural and artificial hybrids between species, has encouraged the belief that interspecific hybridization may be successful. The reports of early botanists on the occurrence of natural hybrids have been summarized by Schulz (12) and later by Ascherson and Graebner in their systematic treatise on the genus Melilotus. These authors mention hybrids between M. alba and M. altissima Thuill, M. alba and M. officinalis, M. altissima and M. officinalis, and M. officinalis and M. wolgica. The authors describe these alleged hybrids as having characters of both supposed parents, but genetic data on the supposed hybrids are wanting. Kirk and Stevenson failed to obtain a single hybrid between crosses of M. alba and M. officinalis. In an experiment designed to favor natural crossing between a line of M. alba and its varieties and species of yellow-flowered sweetclovers, Kirk secured one hybrid with cream-colored flowers. This is similar to hybrids secured later from artificial crosses between M. alba and Redfield Yellow. Furthermore, reciprocal crosses between Redfield Yellow and Albotrea and between Zouave and Redfield Yellow failed to produce a single hybrid. Natural hybrids of the same nature, involving crosses between Redfield Yellow and M. alba, have been selected by Garver at Redfield, S. Dak., and at Arlington Experiment Farm, Arlington, Va. Sylven, in Sweden, reports a natural hybrid between M. alba and M. officinalis, selected from seed secured from Canada. It is possible that this hybrid may have arisen from the cross between M. alba and Redfield Yellow or the reciprocal.  More recent work of Stevenson and Kirk (13) shows that hybrids can readily be secured by artificial cross-pollination between M. alba (Alpha) and M. suaveolens (Redfield Yellow). The evidence tends to the conclusion that, while M. alba and M. officinalis and M. officinalis and M. suaveolens are reciprocally sterile, M. alba and M. suaveolens are reciprocally fertile. Compatibility varies, however, between individuals, depending upon which species is used as the male or female parent. In several cases crosses between M. alba and M. officinalis and between M. alba and M. dentata have resulted in the development of pods and shrunken nongerminating seed. Efforts to dissect out hybrid embryos at an early stage of development and grow these to maturity in nutrient cultures have failed. Further attempts at interspecific hybridization are in progress at several places.

Intergeneric crosses between Melilotus, Medicago, and Trigonella have been attempted by Fryer (4) and by Stevenson and Kirk (13), but without success.