C. F. Poorer, Cytologist, Division of Fruit and Vegetable Crops and Diseases, Bureau of Plant Industry

[ABSTRACT.]  THE old method of plant improvement is inadequate for our present needs. The numerous kinds of insects and fungus diseases attacking such root vegetables as the turnip, rutabaga, and radish necessitate the production of new resistant or immune strains by the more recent methods, including inbreeding, crossing, and the use of wild ancestral forms in the search for superior qualities.  In all the cruciferous root vegetables the deliberate attempt to breed for resistance to diseases or to produce special kinds of vegetables has been neglected thus far.  But if practically nothing has been done in the way of improvement by modern breeding methods, there has been enough genetic research and working out of adequate techniques to lay the foundations on which to build a practical program.

Did the ancient civilizations arise in the regions where our common cultivated food plants originated and were naturally abundant? Or did man plants with him, so that the early centers of civilization only seem to be the centers of origin of the plants? This must remain an interesting subject of speculation, but few students doubt that civilization was dependent on the natural locations of the plants.  At any rate, the regions that are now believed to be the natural centers of origin of the root vegetables (40, 41), which were used as food long before recorded history, include practically all of the centers of the oldest civilizations. The present belief is that in the Old World there were six of these, with five of which we are here concerned, and in the New World two major and two minor centers, all of which produced valuable root vegetables:
          (1) Central and western China—radish, turnip, taro (dasheen).
          (2) India (except northwestern India)—taro.
          (3) Middle Asia (Punjab and Kashmir)—turnip, rutabaga, radish, carrot.
          (4) Near Asia—turnip (secondary center), beet, carrot.
          (5) The Mediterranean—turnip, rutabaga, beet, parsnip, salsify.
          (6) Ethiopia—no root vegetables.
          (7) Mexico—sweetpotato.
          (8) South America (major)—taro, potato.
          (8a) Chile—potato.
          (8b) Brazil-Paraguay—cassava.

The theory is that the place where a plant exhibits the greatest diversity of subspecies and varieties in its natural state must have been a center of origin of that plant.

The value of the root vegetables is due to the fact that they are biennials, storing food2 in their roots during the first season to support the second season’s growth. Among the cultivated root vegetables, man has been able to select fairly true breeding varieties, which differ from one another chiefly by the shorter or longer time intervals they require to reach maturity. In some regions an ordinarily biennial crop may be changed into an annual, or vice versa, by planting seed early enough so that the plant will bolt, or produce seed the first season, instead of enlarging its root for winter dormancy. In all the biennial crop plants there are numerous varieties reputed to be non-bolters, meaning that they take longer to mature than varieties not so designated. Genetically, the hereditary factors governing time to maturity in most of our root crops must be numerous, and the terms annual and biennial are not strictly accurate because of the many gradations.

From the beginnings of agriculture almost to the present there has been no conscious effort to improve the root vegetables beyond selecting seed from particular plants whose roots struck the grower’s fancy.  Within the last century, however, there has been more or less continuous activity in improving plants by crossing diverse types and selecting new combinations of characters. With the development of Mendelism into the science of genetics it became apparent that cross-breeding was essential for rapid plant improvement, and furthermore that an intelligent hybridization technique required a knowledge of the breeding behavior of chosen parents. In the case of cross-pollinated species, like the root crops, the breeding behavior or the real genetic make-up of the parents can be determined only by inbreeding to produce lines homozygous or "pure” for their own characteristics. This must be combined with the keeping of pedigree records. With all our root vegetables, however, the inbreeding program is sometimes exceedingly difficult, for two reasons to be discussed later.


The old method of plant improvement is inadequate for our present needs. The numerous kinds of insects and fungus diseases attacking cruciferous root vegetables (turnip, rutabaga, radish) necessitate the production of new resistant or immune strains by the more recent methods, including inbreeding, crossing, and the use of wild ancestral forms in the search for superior qualities. When established commercial varieties are planted in regions where insects are most numerous or in soils or regions where the fungus diseases are most damaging, seed can be produced only by plants that are entirely immune or partially resistant to attack. If no natural immunity or resistance is found in commercial stocks, it can usually be found by the. introduction of wild ancestral forms of these cultivated vegetables from their original home sites. In all the cruciferous root vegetables the deliberate attempt to breed for resistance to diseases or to produce special kinds of vegetables has been neglected thus far. But if practically nothing has been done in the way of improvement by modern breeding methods, there has been some genetic research and the working out of adequate technique, as will be brought out later.  The ground work has been laid on which to build a practical program.


The name turnip is commonly applied to vegetables that, in their present cultivated condition, are botanically classified in three separate species of the genus Brassica. The point should be stressed, however, that cultivated forms of plants represent complexes of characters that make them quite different from their wild prototypes.  This has resulted from many generations of selection for characters, usually Mendelian recessives, which accumulate in combinations not found in nature.

In the United States the name turnip is applied to plants of the species Brassica rapa L,3 a species wherein all plants examined cytologically have 10 pairs of chromosomes. In Europe the name turnip is frequently applied to the group of plants that we call rutabagas (bagas, or sometimes Swedish turnips), known as B. napus var. sativa rapifera Hort. by some botanists, and as B. campestris var. napobrassica DC in Bailey’s Standard Cyclopedia of Horticulture.  The rutabaga has shown 18 pairs of chromosomes in two cytological examinations by Karpechenko and Frandsen and 19 pairs in one examination by Nagai and Sasaoka (29). It is important to know the number of pairs of chromosomes in a species, or at least in the particular stocks used for parents, because crosses between species having unlike numbers of pairs are less easily made and are less regular in breeding and fertility than crosses between those with the same number. A discussion of chromosome numbers in the genus Brassica is given by Pearson (30).

In general, the chief distinguishing characteristics of the turnip are that the roots are mostly disklike or decidedly flattened, though ranging from spherical to elongated conical; the leaves are hairy, usually not fleshy, and greatly varied in outline; and the plants reach maturity in from 42 to 80 days. The commonly grown varieties show a wide variation in time to maturity, as the following indicates:

White Milan42
Snow Ball43
Purple Top Strap Leaved46
Purple Top White Globe55-60
Golden Ball60-65
Yellow Aberdeen70-80
White Norfolk76

The principal turnip varieties reach maturity more quickly than the principal varieties of rutabagas. Since the turnip and the rutabaga will readily cross, however, it is probable that these turnips reaching market size around 65 to 80 days after planting are the results of natural turnip-rutabaga crosses.

The chief characteristics of the rutabaga, or Swedish turnip, as it is sometimes known, are a root that is tankard-shaped or elongated, although sometimes globular; a fleshier and larger leaf than the turnip; leaves not hairy; and a longer period of time (from 85 to 90 days) required for reaching the best pulling stage. Among the most commonly available yellow-fleshed rutabaga varieties on the American market are American Purple Top, Early Neckless, and Bangholm.  Two white-fleshed rutabaga varieties are Sweet Russian and White Rock.

Disease Resistance in the Turnip

Although no deliberate attempt has been made to breed disease-resistant varieties of turnips or rutabagas, a new variety of turnip known as The Bruce, that is highly resistant to the slime mold disease called clubroot, has recently been introduced in Great Britain, New Zealand, and Australia. Clubroot attacks many species of the crucifer family, including cabbage, radish, mustard, cauliflower, sweet alyssum, and many others. It is now present in every country where the common cruciferous vegetables are cultivated. The Bruce is supposed to be a natural hybrid between the turnip and the rutabaga, and it first appeared in Scotland about 1820, some 40 years after the time that clubroot first appeared in Great Britain in 1780. The seed of The Bruce was carefully guarded in Aberdeenshire, and as it was disseminated throughout the district it was known to local farmers under different names, until within the last 10 years all agreed on its present name. The variety is still in a highly heterozygous state; that is, it appears in a great range of shapes, colors, firmness of flesh, size of seed, and degrees of resistance to clubroot. Tradition says that the rutabaga parent had a purple top and white flesh and that the turnip parent had a green top and yellow flesh, but these differences are not enough to account for the persistence of a high degree of variability.

Evidently much of the variability is due to the fact that the two parents had different chromosome numbers—10 pairs in the turnip and 18 pairs in the rutabaga—and the present descendants are still segregating for differences in number of chromosomes. From a plant-breeding standpoint it is of interest to note that in tests conducted recently in Scotland (7) the stocks of The Bruce that are less resistant to clubroot proved to be the stocks that have grown on soils less heavily infected with the slime mold.

Many different forms of clubroot exist, and the studies show that different varieties of turnips or rutabagas may be resistant to one or several strains of the slime mold but susceptible to others. Evidently The Bruce combines the largest number of resistance factors, and some strains are therefore nearly immune to this disease.

In addition to the fact that some strains have practical immunity to clubroot, whether grown on limed or unlimed soil, the best strains of the variety also have a higher percentage of dry matter than other turnip varieties included in the same test.


The radish, Raphanus sativus L., is a native of China and India, and its cultivation was practiced in ancient Egyptian and Grecian gardens. Because of its low-food value it has always been more popular as a home-garden vegetable than as a truck crop. It is a good subject for genetic study, since it exhibits a wide range of varietal forms in shape, size, and color of roots, time required from planting to market maturity, and keeping quality. A number of species crosses involving the radish have proved of considerable value to the plant breeder as well as the cytogeneticist. The condition known as self-incompatibility, which means that plants cannot normally be fertilized by their own pollen, is widespread in commercial varieties, but this can be removed by the selection of self-fertile inbred lines.

The leading radish varieties are most frequently classified according to the seasons in which they are grown. Spring varieties, such as French Breakfast, White Tipped, and Scarlet Turnip, are quick-maturing and require from 24 to 30 days’ average time to marketing.  The leading summer varieties—White Icicle, White Strasburg, and Giant Strasburg—are larger and better keepers and require from about 30 to 42 days to market maturity. The winter varieties, such as Long Black Spanish, Round Black Spanish, White Chinese Winter, and Mammoth White, have the largest roots, require 50 to 60 days to reach market maturity, and keep for several months when well stored.

Our varieties were all developed by crude selection without the aid of artificial cross-breeding to produce special types. A considerable number of genetic analyses have been made with radish, however, and with these as a beginning it is quite likely that plant breeders will take advantage of cross-pollination to produce new forms better adapted for special purposes or localities.

Of great interest from the standpoint of plant breeding is the fact that the radish has been successfully used in generic crosses with the cabbage. According to reliable records, the first successful cross between the radish and the cabbage was made in 1826 by Sageret, a gifted French plant breeder. This wide cross between two genera has subsequently been duplicated by several other investigators, with especially interesting results in the two instances to be reported later.


ALTHOUGH the origin center of the beet, Beta vulgaris L., is well known to be western Asia and the Mediterranean region, our commercial varieties have resulted from mass selections, based on the appearance and quality of roots at the end of the first season, to meet western European tastes. The beet flower is largely wind-pollinated, and cross-fertilization can be effected in areas separated a dozen or more miles. The practice of gathering seed for commercial stocks of beets from open-pollinated plants is responsible for the maintenance of self- incompatibility in beet varieties. Recently, however, seedsmen and geneticists are using specially produced inbred lines to obtain improved stocks free from rogues and from self-incompatibility (fig. 1).

Figure 1.—Effect of inbreeding or line breeding in developing high uniformity and market quality in beets: A, Stock of Ohio Canner developed by selection and line breeding from Detroit Dark Red; B, good average commercial stock of Detroit Dark Red. Note superior uniformity in shape of beets in A as compared with those in B.

The pollination technique in the beet, described later, demands unusual care, because even the protection given by airtight bags will be undone upon removal to effect controlled cross-pollination unless the breeding plot is removed from possible contamination. Self-pollination for inbreeding requires isolation, either geographic or physical, and it is necessary to start with strains that are already self-fertile in some degree.

The leading beet varieties differ considerably in length of growing season, and the time to market maturity of the most popular varieties extends from about 57 to 78 days. Beets may be grown so as to be available on our markets all year round by storing the slower-maturing ones, like Long Smooth Blood, under relatively humid conditions.  The leading varieties are all of the red beet type, and they include Asgrow Wonder, Crosby Egyptian, Ohio Canner, Detroit Dark Red, and Long Smooth Blood.

New beet species from Anatolia, recently imported into Germany and examined cytologically by Scheibe (1, 33), have chromosome numbers running in an arithmetical series, with 9, 18, and 27 pairs. A species with 27 pairs, Beta trigyna Waldst. and Kit., exhibited a sugar content almost twice as great as that of the best commercial varieties of sugar beets. From the plant breeder’s standpoint this information indicates that there are still untried species in the wild origin centers that offer new characters for hybridization and selection. Moreover, the occurrence of chromosome numbers in arithmetical series suggests the operation of chromosome doubling in the evolutionary history of the genus and gives hope of producing new constant- breeding hybrids made to order.


ALTHOUGH the home of the carrot, Daucus carota L., is western Asia, all of our present varieties represent western European selections, many of which were made by Vilmorin, a French plant breeder and seedsman of the middle nineteenth century. Up to the present, carrot breeding has been almost entirely a matter of crude mass selection, based on the appearance of roots at the end of the first season’s growth.

This method has produced a surprisingly good range of varietal types in length and shape of roots, relative thickness of central core, color of roots, and length of time from planting to market maturity.  Because of the increasing demand for carrots in the dietary, however, much attention is being given today to the more modern methods of inbreeding for the purification of commercial stocks and for genetic analysis, followed by cross-pollination.

Some notion of the uniformity developed in inbred carrot lines may be gained from the fact that the coefficient of variability in carotene content (carotene is a yellow coloring matter and the source of vitamin A) between 28 roots of a commercial strain of Danvers Half Long was about 33 percent, while the variability between 28 roots of a line inbred for four generations was only about 7 percent.

The leading carrot varieties over the last 20 years have been Chantenay, Danvers Half Long, Oxheart, and Nantes, all of which are medium long, require from 70 to 75 days from planting to marketing condition, and vary from a large core in Chantenay to a minimum of core in Nantes. Although the present trend is to develop narrower cores, there is a limit, because an increasing brittleness of the tops accompanies an increasing narrowness of core. The carrot is the last of the root vegetables to receive the attention of geneticists, owing largely to misconceptions of the difficulties involved. Selected plants were allowed to set seed by open pollination because it was believed that carrots were self-incompatible, and this practice, of course, has largely nullified the effects of the root selection.

For many years carrot growers and seedsmen did not understand that the failure of isolated umbels to set seed was due not to self- incompatibility but to the fact that the stamens of a carrot flower ripen and shed their pollen some hours or days before the pistil is receptive to pollen. Consequently, self-pollination is effective only if the stigmas of the older flowers can be served by pollen from the anthers of younger ones. This service is performed by blowflies introduced into cages in which the entire plant or a major branch is encased; and by using this technique which is described in more detail later, Borthwick and Emsweller (2) demonstrated that all carrot varieties tested were entirely self-fertile.

Miller (28), in Louisiana, in working with the problem of poor color in carrots grown on certain soils, found that merely selecting the best colored roots on soil that was adversely affecting color produced progenies (whether self-pollinated or open pollinated) having higher percentages of well-colored roots on unfavorable soils than check commercial strains.

Recent studies indicate that hybrid vigor in the carrot acts as it does in corn. When two inbred lines producing roots averaging 12.8 and 24.9 g, respectively, were crossed, the first-generation progeny produced roots that weighed 80.5 g on the average.

Genetic analysis in the carrot has included studies of the inheritance of such characteristics as branched roots, cracked roots, root shape, and root color.


Among the root vegetables taro, parsnip, and salsify are relatively unimportant in the United States, and the number of contrasting characters available for study is so small that no attention has thus far been paid them by breeders or geneticists.


The taro, or dasheen, Colocasia esculenta (L.) Schott, seems to have three centers of origin, China, India, and South America, and up to very recent times its cultivation was practically confined to the Hawaiian and other Pacific Islands, where its fleshy root is the staple food of the natives, and also to the Mediterranean region. Recently, however, several introduced varieties of the taro, Trinidad, Ventura, and Sacramento, were thought sufficiently promising to serve as potato substitutes in the Southeastern States (43, 44). As early as the seventeenth century a few varieties were successfully grown in South Carolina as a cheap food source for the plantation hands, but the industry failed to meet competition with importations from the West Indies. The starch of the root is regarded in Hawaii as superior to that of any other root vegetable in the ease with which it is digested and assimilated.  Young, (43) gives many ways in which the taro root may be prepared as a food, including practically all the ways used with the potato and the sweetpotato.

No breeding or genetic work has been done with taros, and only in Hawaii is there any record of seed production. Moreover, many varieties of taro fail to produce an inflorescence, even in Hawaii. The United States Department of Agriculture has recently begun a project on taro breeding with the object of introducing a large number of species and varieties from Hawaii and elsewhere with a view to investigating their genetics and cytology, and also to provide new varieties with appeal to the growers and consumers of the Southeastern States.


The parsnip, Pastinaca sativa L., is a relatively unimportant root vegetable cultivated in the United States as a market garden crop.  Records show that its center of origin was the Mediterranean region and that it was introduced into Virginia in 1609 and Massachusetts in 1629, but thus far there has been no interest in the improvement of varieties or in the analysis of character inheritance. Only one variety of importance, Hollow Crown, is grown in this country, and the lack of varieties with character contrasts is one reason for the failure of geneticists to attempt any character analysis.

The parsnip is an umbelliferous plant in the same family as the carrot, and like the carrot it exhibits the phenomenon of protandry, the pollen being shed long before the stigmas are receptive. The pollination technique of producing pure strains for variety improvement or genetic analysis is therefore identical with that for the carrot.


Salsify (Tragopogon porrifolius L.) or vegetable-oyster—so called because of a mildly oysterlike flavor—is grown in the United States on a very small scale and is represented mainly by a single variety, Mammoth Sandwich Island. Salsify originated in the Mediterranean area, and its introduction into the United States first occurred about 1806. Practically all the salsify seed produced in this country comes from a narrow strip along the coast of central California. Its limited use, as well as the meager representation of character differences, accounts for the failure of plant breeders or geneticists to take any interest thus far in its improvement.

Salsify is one of the Compositae, belonging to the same family as the sunflower. Its fleshy taproot, which resembles a small parsnip, is handled for curing and storing very much like the dahlia root.


In spite of the limited amount of practical breeding work with the root vegetables, there are certain aspects of the breeding technique and of the theoretical knowledge attained that are exceptionally interesting from the standpoint of plant breeding in general. An effort will be made here to discuss these aspects as simply as possible for the nontechnical reader. (A technical discussion of root-crop genetics follows later in this article.)

Inbreeding in the root crops is often difficult for two reasons. (1) It brings out recessive factors for small size and lack of vigor; in other words, it is just as likely to make undesirable factors “pure” as desirable ones. Inbred lines frequently are so enfeebled that they cannot live. In carrots, for example, this is so general as to make it necessary to proceed with caution in inbreeding and to inbreed several generations, then outcross to secure enough vigor to go on, then continue inbreeding until the line is as uniform and homozygous or “pure” as desired. With beets, on the other hand, many vigorous upstanding inbred lines have been selected without recourse to outcrossing. (2) Inbreeding may also uncover recessive hereditary factors for self-incompatibility.

How do these factors for self-incompatibility act? In the cruciferous vegetables—including turnip, rutabaga, radish—and probably beets, the rate of pollen-tube growth of all pollen grains containing such a factor is retarded if that factor is also present in the tissues of the style (the stalk of the pistil) through which the pollen tube has to grow before it can reach the ovary; in other words, if the factor is present in both male and female cells (fig. 2). When the rate of pollen-tube growth is retarded the male cell cannot reach the female cell in time to effect fertilization. Self-pollination is thus impossible in plants that lack the normal alternative (allelomorph) of these self-incompatible genes. The difficulty can be overcome, however, by bud pollination; that is, applying pollen to the stigmas of unopened buds. This permits many extra hours of pollen-tube growth before the flower would normally shed its own pollen.

Commercial varieties that contain genes for self-incompatibility cannot be pure or homozygous for this characteristic or they could not reproduce by crossing with other plants of the same variety.  They must in some way be cross-compatible with other plants of the variety; that is, in a highly heterozygous condition so far as the genes for self-compatibility are concerned.  The best way to remedy this situation is to remove the factors for self-incompatibility that prevent reproduction, by making appropriate crosses to stocks containing the "normal" alternative genes or allelomorphs; or, more simply, all lines that are not self-fertile may be eliminated. The situation is illustrated in figure 2.

The last method has been shown to be entirely feasible even in such characteristically self-incompatible species as cabbage and beets.  In these species numerous varieties are available that are completely self-fertile.

All our common root vegetables are biennials, and this somewhat reduces the speed with which breeding operations may be conducted.  In each case there are also special impediments to breeding operations, such as self-incompatibility in the cruciferous species, self-incompatibility and wind-pollination in beets, difference in the time of maturity of the male and female organs in carrots, and failure of taro to produce flowers when grown in most regions. These handicaps, however, merely tax the ingenuity of the vegetable breeder and make it necessary for him to develop a special pollination technique for almost every species. To appreciate some of the problems confronting the vegetable breeder, it will be well to review the pollination technique now most favored for cross-pollination or self-pollination in the root vegetables.

The cruciferous root vegetables—turnips, rutabagas, and radishes— have relatively large flowers, which are insect-pollinated. This makes it necessary, in order to effect controlled pollination, to isolate all flowers from insect contamination by encasing them in cloth or paper bags or in cages made of netting. Crawling insects may be excluded by gluing a cotton lining on the inside of the neck of the bag.

To cross-pollinate, the six anthers from the flower of the female parent must be removed a day before the pollen is shed, thus preventing any chance of self-pollination. The following morning the pollen from the intended male parent, which likewise has been bagged, is applied, either by a camel’s-hair brush or by transferring an entire flower that is shedding its pollen, to the stigma of the emasculated flower of the seed parent.

Figure 2.—Diagram showing possible combinations in crosses involving self-incompatibility. S represents the normal self-compatible genes, dominant to all the members of the multiple allelomorphic series. s1, s2, and s3 represent self-incompatible genes, recessive multiple allelomorphs of S.  No plant can have more than two (any two) members of this series.

   To self-pollinate, it is necessary, as before, toisolate all flowers from insect contamination.  If the plant is self-compatible, it will automatically pollinate itself.  If it is self-incompatible, however, it will be necessary to bud-pollinate; that is, to apply pollen from a flower opened this incompatible combinations morning to an as yet unopened bud of the same plant. The pollen will thus have extra time in which to reach the ovules. It is even possible that the retarding principle does not become effective until after the pollen of a given flower is shed normally.

   The beet is an example of a wind-pollinated species that is also insect-pollinated to some extent. Beet pollen is so small and dust- like that samples of it have been found by airplanes at elevations of 1,000 feet. Consequently when air currents are in motion the removal of a bag in order to effect pollination, even for an instant, may result in contamination of the stigmas with undesired foreign pollen. The best practice is to pollinate in still air; for example, within a greenhouse.  The floral organs are so small that emasculation is impracticable. When cross-pollination is desired, an excellent procedure is to make a cross with a male parent that has some easily identified, dominant characteristic. All progeny that result from self-pollination will then show the recessive characteristic and can be discarded in favor of the hybrids, which can be identified by the dominant characteristic; or if it is not completely dominant, they will be intermediate in character.

The technique developed for sugar beets by Brewbaker (3) is to isolate the female parent flowers in a hand-made bag 5½ inches by 17 inches made of 30-pound vegetable parchment. When cross-pollination is desired, the danger of uncontrolled pollination may be practically eliminated by not removing the bag at all, and blowing pollen into it with a pollen gun, as described by Jenkins (13). As already explained, self-incompatibility in beets and crucifers prevents successful self-pollination unless a normal alternative gene (allelomorph) is present.  A strain with such a normal gene might be called a low self-fertile one, and, as indicated in figure 2 (the fourth ovary from the left), a low self-fertile strain may be converted by inbreeding and selection into a high self-fertile strain containing only the dominant normal gene; that is, a strain homozygous for self-fertility.

As a result of inbreeding experiments with sugar beets, Brewbaker (3) says: "The characters of high and low self-fertility appear to be heritable, and by selection and continued self-pollination highly self-fertile lines would be obtained.” It should be said that no genetic analysis of the inheritance of what is called self-sterility in beets has been made, but it is assumed to bé of the same nature as that outlined in figure 2, which in general illustrates the interaction of genes for self-incompatibility in the cases thus far studied, tobacco and cabbage.

In carrots no self-incompatibility is encountered, despite previous beliefs (2). Self-fertility is complete provided one overcomes the barrier of protandry, which can be easily done by using the fly-pollination technique of Jones and Emsweller as applied to carrots by Borthwick and Emsweller (2), for either cross-pollination or self- pollination.

Since carrot flowers are as small and difficult to emasculate as beet flowers, cross-pollination is effected most readily by bagging single umbels in muslin or cheesecloth bags, waiting several days after the first flowers have come into full bloom, then introducing an umbel of the male parent with cut stem in a jar of water, along with a supply of freshly hatched clean blowflies, which will effectively transfer the pollen to the female umbels. In such cases self-pollination is impossible on a single umbel because of the existence of protandry. Self- pollination is readily performed by caging an entire plant in a muslin or cheesecloth bag, then periodically introducing clean freshly hatched flies. By caging the entire plant the flies are enabled to pollinate the older flowers with their own kind of pollen from the younger flowers,

The technique of pollination in taros is very simple. The taro flower is in the form of a spadix or fleshy axis enclosed by a modified leaf or spathe, as in the jack-in-the-pulpit or the calla. Each spadix contains staminate florets on its upper end and pistillate florets on its lower end. All that is required is to remove the spathe and enclose the spadix in a glassine or other transparent bag to exclude insects, and self- or cross-pollinate the female florets as desired with a camel’s-hair brush. This is sometimes impossible on a single spadix because in some species of taro the stigmas are past receptivity by the time pollen is ready. (This is known as protogyny, the opposite of protandry.)  As a precaution against self-pollination when crossing is desired, it is advisable to remove the male end of the spadix. The real difficulty with taro, however, is that most varieties will not produce an inflorescence even where they are extensively cultivated, as in Hawaii.

Hybridization between species or genera of the root vegetables has produced at least three artificial hybrids with all of the chromosomes doubled. Plants with doubled chromosome numbers of this kind are known as amphidiploids, and they are usually hybrids between remotely related parents. It would seem that on account of the remote relationship, the chromosomes of the parents are so dissimilar as to be unable to mate or pair properly.  The cell-division mechanism is consequently so fundamentally disturbed as to produce a doubling of the number of chromosomes in the sex cells.  The first-generation hybrid is likely to be either entirely sterile or almost so, but the doubling of the chromosome number provides all unmated chromosomes with mates for pairing, so that in future generations sex cell division can proceed in the regular manner. The result may be a true-breeding amphidiploid. In figure 3 all visibly unpaired chromosomes of the F1 or first hybrid generation can be seen to be doubled in the amphidiploid hybrid. Frequently, however, the chromosome doubling takes place in the cells of the plant body instead of in the sex cells.5

Figure 3—Somatic (diploid) chromosome sets of two parent species of hawks-beard: A, Crepis rubra; B, C. foetida; C, their F1 hybrid; D, their amphidiploid hybrid. Chromosome doubling in some F1 sex cells, followed by a union of two of them, has produced a complete diploid set, qualitatively as well as quantitatively, from both parents. X about 1,800.

It was only in 1925 that the first amphidiploid hybrid obtained under experimental conditions was recognized and described (4). Since then an almost bewildering number of such hybrids have been described.  Many of them, because of the wide separation of their two parents, are practically true-breeding and might truly be called new genera or new species and given new names (42). In other cases, however, the parent species were so closely related to begin with that many of the chromosomes were able to pair, and the subsequent chromosome doubling therefore only resulted in a more complex hybrid, but not a true-breeding one or one that could be expected to produce stable derivatives (31, 32).

Many plant genera contain species whose sex cell chromosome numbers occur either exclusively or mostly in multiples of some common number. In wheat (Triticum), for example, some species have 7 chromosomes, others 14, and still others 21, in an arithmetical series. Other examples of genera in which this occurs, with the sex cell chromosome numbers, are:
Beta (beet)9, 18, 27
Campanula17, 34, 51, 8, 16, 10, 13
Chrysanthemum9, 18, 27, 36, 45
Papaver (poppy)7,14, 21, 28, 11,22
Rosa7, 14, 21, 28

The occurrence of arithmetical series among the chromosome numbers of many genera prompted Winge in 1917 to offer the hypothesis that such chromosome numbers arose as a consequence of chromosome doubling following species crosses. At that time he knew of no experimental verification of his hypothesis, but now there are scores of artificially produced amphidiploids, as well as evidence of amphidiploid hybrids in the wild, and correlated analyses of characters and chromosome numbers for their parent species. Two species known to be natural hybrids are successfully growing in the wild—one a pentstemon, in the foothills of California, and the other a grass, in the harbors and on the beaches of the English Channel. The artificially produced constant breeding amphidiploids in the Cruciferae, with the number of pairs of chromosomes, are:

(1) Radish, Raphanus sativus, 9 pairs, X cabbage, Brassica oleracea L., 9 pairs,
produced "Raphanobrassica", 18 pairs.
(2) Radish, R. sativus, 9 pairs, X Chinese cabbage, B. chinensis L., 10 pairs,
produced “Brassicoraphanus", 19 pairs.
(3) Turnip, B. campestris L., 10 pairs, X rutabaga, B. napus, 18 pairs,
produced "Brassica napocampestris”, 28 pairs.

Although none of these constant breeding hybrids has commercial possibilities at present, Brassica napocampestris, combining features of the turnip and the rutabaga, may easily have value with further selection, and the others can be used as parents for further crosses to provide selection material for commercial stocks.

The fact that the chromosome numbers in Beta, from which our garden and sugar beets come, occur in an arithmetical series, 9, 18, 27, suggests that species formation in this genus was influenced to some extent by chromosome doubling following crosses between species.



According to Kajanus, turnip more commonly contains hereditary factors for self-incompatibility than rutabaga. This is expected if among the larger number of chromosomes of the rutabaga the chromosome containing the locus S versus s1, s2, etc. (see fig. 2), had been duplicated; hence segregations for compatibility versus incompatibility will show at least a tetraploid rather than a diploid ratio; for example, an expectation of 35:1 instead of 3:1. The production of self-fertile inbred lines in turnips will, therefore, be more difficult than in rutabagas.

Kajanus (15, 16, 17, 18, 19) has made the greatest number of genetic analyses in both these species, the main results of which are shown in tables 1 and 2. Among other geneticists who have worked with these vegetables are Hallqvist (11, 12), Malinowski (25, 26), and Sylven (34). Kajanus studied the progeny of hybrids between these two species, and Frandsen and Winge (8) produced an F1 hybrid between the turnip and the rutabaga that later doubled its chromosome number. The ensuing amphidiploid stock of plants had the potentialities of a constant new species, combining features from both parents, but most of them came from the rutabaga, because it contributed most of the chromosomes.

Table 1.—Genetic analyses of character inheritance in turnip
Color of barkColor of fleshColor of flower

M v. m, white v. yellow flesh color, also yellow v. orange flower color.
P v. p (anthocyanin), red v. cream-yellow upper root bark color.
V v. v (chlorophyll), green v. cream-yellow upper root bark color.
P is epistatic to V, thus producing the phenotypic combinations shown above.

Mendelian ratios obtained (Kajanus):

The heterozygotes of these genes exhibit an intermediate intensity of expression; consequently the classification of any hybrid population may be more complex than is indicated.

Table 2.—Genetic analyses of character inheritance in rutabagas
Color of barkColor of fleshShape of leafColor of flower
P1P2M1M2E1E2ABDeep redWhitePinnatifidYellow
P1P2M1m2e1e2aBEntirePale yellow
P1P2m1M2E1E2abPinnatifidPale orange
P1p2M1M2e1e2ABPale redEntireYellow
p1p2M1M2E1e2aBGreenPinnatifidPale yellow
p1p2m1m2e1e2abYellowEntirePale orange.

A v. a, orange v. pale orange flowers.
B v. b, yellow v. pale yellow flowers.
P1 v. p1, pale violet red v. green.
P2 v. p2, deep violet red v. green.
M1M2 v. m1m2, yellow v. white flesh color.
E1E2 v. e1e2, pinnatifid v. entire leaf outline.
P2 is epistatic over P1, giving the phenotypes shown above.

Mendelian ratios obtained:

The duplicate nature of factorial relations here is an additional reason for the suggestion that the genes concerned are included in the extra set of eight chromosomes by which the rutabaga is distinguished from the turnip. For example, the two 15:1 segregations for flesh color and leaf shape indicate duplicate dominant genes for the dominant allelomorphs in each case. Malinowski, who analyzed the cross biennial X annual, obtained an observed ratio of 349 biennials to 57 annuals, which gives a significant deviation from any assumed Mendelian ratio.

Turnip X Rutabaga Crosses

From 1912 to 1917 Kajanus reported studies from a number of crosses between the two species, turnip, Brassica rapa (10 pairs of chromosomes), and rutabaga, B. napus (18 pairs of chromosomes), with results that have been summarized by Matsuura (27). The work was done before the acceptance by geneticists of cytological aid in investigating crosses between parents with differing chromosome numbers. The studies summarized by Matsuura would have been more valuable had the investigators considered this feature of the problem. Most of the analyses made, however, deal only with simply inherited leaf characters that segregate in the F2 generation with 1 :2:1 monohybrid ratios, as, 1 pubescent : 2 intermediate : 1 smooth; or 1 bloom : 2 intermediate : 1 nonbloom.

The Amphidiploid Hybrid "Napocampestris”

In 1932 Frandsen and Winge (8) reported the production under experimental control of a hybrid between the turnip, Brassica campestris var. sativa rapifera (10 pairs of chromosomes) and the rutabaga, B. napus var. sativa rapifera (18 pairs of chromosomes). It is unfortunate that an unnecessary confusion should be introduced by calling the turnip B. campestris here and B. rapa at other times, but the authors’ names have been used. As expected, plants of the first filial generation were, as a rule, quite sterile compared with the parent plants. One of the F1 plants, however, proved to be almost as fertile as the parents. A cytological examination of the root tips of several F1 plants showed the expected chromosome number of 28 (10 plus 18), where half the chromosomes from rutabaga were added to half the number from the turnip parent. Consequently it is believed that all the F1 plants, with the exception of the one that was highly fertile, possessed 28 somatic chromosomes. The highly fertile plant is believed to have doubled its chromosomes from 28 to 56 immediately following fertilization, as in Nicotiana digluta (4), because in the progeny of this plant 21 plants that were examined cytologically were highly uniform both as to outward appearance and as to the chromosome number of 56.

The progeny of all 28-chromosome F1 plants exhibited a more or less high degree of sterility, because at the time of formation of sex cells the 18 chromosomes from the rutabaga parent had only 10 chromosomes from the turnip parent with which to mate, leaving 8 chromosomes to be distributed to the sex cells at random. Consequently, when these plants were self-pollinated, all sorts of fertilizations with odd chromosome numbers were effected, many of which proved to be nonviable, hence the high degree of sterility. Sterility of this kind is called "generational" sterility, to distinguish it from incompatibility, already considered. In the progeny of the 56-chromosome fertile plant, however, the distribution of chromosomes to the new sex cells was as regular as though the plant had 28 pairs of chromosomes, the members of each pair separating from each other at germ cell formation with the regularity of true-breeding natural species.

An idea of the uniformity of this fairly true-breeding new hybrid may be gained by comparing its F2 progeny with the parents and the F1 generation.

Table 3.—Distribution of root shapes of turnip-rutabaga hybrids
Nature of hybridGlobeOvalHalflongLong conicalBark color
Amphidiploid[none observed][none observed]8416Green

In outward appearance as well as in chromosome content most of the progeny, diploid or amphidiploid, resembled the rutabaga parent which contributed eight mote chromosomes than the turnip. This may be compared to the observation that in The Bruce turnip most of the characters usually favor the presumed rutabaga parent. From the foregoing discussion of Kajanus’ analysis of inheritance of root color in turnips and rutabagas it is seen that green bark is hypostatic to red bark, which indicates that the rutabaga parent of Frandsen’s amphidiploid hybrid must have been heterozygous for the red bark factor, and therefore of the genotype PpVV.


An outline of the chief genetic analyses in the radish, Raphanus sativus, together with the names of the investigators reporting them, is given in table 4. The list is compiled chiefly from the work of Frost (9), Uphof (39), and to a lesser extent Malinowski (25) and Karpechenko (21, 22). The character contrasts include all organs of the plant, and the segregations reported show that a fairly simple factorial situation usually exists.

Some question arises regarding the interpretation of results from the cross red-striped v. white (made from crossing the variety Triumph with either Early White or Icicle). In F1 red-striped was completely dominant to white, but in F2 instead of 3 red-striped:1 white, Uphof obtained 1 red-striped:1 white. No cause of the supposed disturbed ratio was learned, but Uphof suggested the action of a gamete lethal, although this ratio might actually be a 9:7 interaction, in which two recessive whites are involved. Frost’s crosses in some cases showed a pronounced hybrid vigor in F1, a phenomenon most frequent in open-pollinated species maintained in a highly heterozygous state.

Table 4.—Character inheritance in Raphanus sativus
P1 charactersF1F2Investigator
Yellow v. whiteYellow3 yellow:1 whiteUphof, Malinowski
Red v. whitePurple1 red:2 purple: 1 whiteUphof
Purple v. red3 purple:1 redFrost
Red primary cortex v. white15 red: 1 whiteUphof
Red striped v. whiteRed striped1 red striped:1 white
Long v. globeIntermediate1 long:2 intermediate:1 globe
Corky v. smoothCorky3 corky:1 smooth
Long leaf v. shortIntermediate1 long:2 intermediate:1 short
Early flowering v. lateEarly3 early:1 lateFrost
White v. yellowWhite3 white:1 yellowKarpechenko
Woody capsule v. paperyWoodyComplex inheritanceFrost
Linkage.—Purple v. red and early v. late are on the same chromosome.
Species Crosses in the Radish

Crosses are easily made between the wild radish, Raphanus raphanistrum L., yellow flowered, and the cultivated radish, R. sativus, red, purple, or white flowered. Both species have nine pairs of chromosomes, and the cross appears to segregate regularly both as to chromosome distribution and Mendelian factors. F1, interspecific hybrids usually have violet flower color, and, according to Trouard-Riolle (36, 37), the starchy root of the wild radish is dominant over the sugary root of the cultivated radish. Frost found indications of a linkage between the locus for purple-red root pigment and that for earliness v. lateness of flowering, with a cross-over of 4.78 percent. Frost also established the presence of self-incompatibility in the radish, but made no analysis of its genetics.

Karpechenko’s "Raphanobrassica”

Karpechenko (21) obtained a first-generation hybrid between the radish (2n=18) and cabbage (2n=18) which was highly sterile as a result of disparity in chromosome content and structure, although not in chromosome number, between the two parent species. In consequence of the cell-division disturbances, some of the functioning sex cells had double the expected number of chromosomes. The chance conjugation of two such unreduced sex cells resulted in the production of a plant having the combined diploid number of chromosomes of the two parents, known as an amphidiploid hybrid.  Karpechenko’s amphidiploid hybrid was in some respects like the amphidiploid hybrid that Frandsen obtained in the turnip-rutabaga cross already described. There is one important difference, however, between Karpechenko’s radish-cabbage amphidiploid and Frandsen’s turnip-rutabaga amphidiploid. Frandsen’s hybrid presumably resulted from a suspended cell division following the union of the turnip sperm with the rutabaga egg, whereas Karpechenko’s hybrid resulted from the chance meeting of two unreduced sex cells in the ovary of a first-generation hybrid. Thus we have illustrations of the two ways in which amphidiploids can be produced, (1) where the doubling occurs in the asexual or sporophyte generation, and (2) where the doubling occurs in the sexual or gametophyte generation.

Figure 4 presents a diagram of the types of seed capsules characteristic of the parents and hybrids occurring in the production of Karpechenko’s amphidiploid, which he called Raphanobrassica.  The hybrid plants were clearly intermediate between the two parents in the structure of capsules, as will be seen from figure 4, but also in possessing as many kinds of chromosomes from one parent as the other—radish, Raphanus sativus (2n=18); cabbage, Brassica oleracea var. capitata (2n=18); Raphanobrassica (2n=36).

Figure 4.—Diagrammatic representation of the relative capsule shapes and sizes, as well as chromosome numbers, in (A) Raphanus sativus, (B) Brassica oleracea, (C) their F1 hybrid, and (D) their amphidiploid hybrid. Redrawn from Karpechenko, but the relative chromosome sizes were somewhat enlarged in the copying.

   No effort has been made to utilize Raphanobrassica to improve either the radish or the cabbage, and Karpechenko (22) states that it will not cross back to either of its parents. It will cross readily, however, with a large number of other cruciferous species including turnips and rutabagas which in turn will not cross with either radish or cabbage. Karpechenko attempted to produce still other amphidiploids with an even higher chromosome number by adding the chromosomes of turnip, rutabaga, etc., to those of Raphanobrassica.  The first-generation hybrids have succeeded, but thus far the chromosome doubling to stabilize sex cell formation has not occurred.

From the plant-breeding standpoint the chief value of these studies is the demonstration of the principle that if a cross cannot be made directly between species A and B because of their wide evolutionary separation, then the gap may be bridged by first forming a hybrid between A and C, with which B will later successfully cross.

Terasawa’s "Brassico-raphanus"

In 1932 Terasawa (35) reported the appearance of an amphidiploid hybrid called Brassico-raphanus with 2n=38 chromosomes, which bred approximately true in the F4 generation from the cross pakchoi (Brassica chinensis, 2n=20) X radish (Raphanus sativus, 2n=18). He explained the tardy appearance of this amphidiploid in F4 instead of F2 as resulting from insect pollination of F2 plants, which somehow induced nonreduction of chromosomes in sex cells. Consequently, the origin of Brassico-raphanus from unreduced gametes was in the main similar to that of Raphanobrassica. The phenotypes of the two parents and their intermediate hybrid are shown in table 5.

Table 5.—Comparative morphology of Brassica chinensis, Raphanus. sativus, and their amphidiploid
SpeciesLeaf marginsRootsFlower colorCalyxFruit structure
Brassica chinensisEntireSlenderYellowOpen2 locules
Raphanus sativusLobedFleshyPurpleClosedNot dehiscent
Brassico-raphanusIntermediateQuite fleshyWhite with purple veinsIntermediateUpper half not dehiscent, lower 2 locules


Our knowledge concerning the genetics of Beta is obtained chiefly from the investigations of Kajanus (14, 16, 20) and Keller (23). Some years ago Kajanus, also Lindhard and Iversen (24), investigated the inheritance of root color in this biennial root crop, and all believed that the postulation of two pairs of factors was sufficient to account for color inheritance. Certain discrepancies in Kajanus’ segregation, which he ascribed to faulty pollination technique, were observed also by Lindhard and Iversen, who interpreted them as caused by about 37-percent linkage between the two postulated loci. More recently, however, Keller, with greater care against the contamination of wind-pollination, investigated the genetics of root color. Instead of only the two pairs of factors assumed by Kajanus et al., giving 4 segregating phenotypes in F2, he finds that each of the two loci carries a series of three multiple allelomorphs.

The genes postulated by Keller in the order of their dominance are:
Locus 1. R=red hypocotyl, Rt=red top white hypocotyl, r=yellow hypocotyl.
Locus 2. Y=yellow root, Yr=green top yellow root, y= white root.

From two different F2 populations the following series of nine phenotypes was obtained from combinations of these six genes:
Root phenotypes in garden-beet crosses
Red hypocotyl, white rootRy
Red hypocotyl, red rootRY
Pale red hypocotyl, pale red rootRY±
Striped red beetRtYr
Green top, red rootRYr
Red top, white rootRty
Yellow hypocotyl, white rootry
Green top, yellow rootrYr
Yellow beetrY

Moreover, instead of a loose linkage of 37 percent, he calculated an average linkage of 7.3 percent for four crosses made in the coupling phase and a linkage of 8.8 percent for a single cross made in the repulsion phase. Different members of the multiple allelomorph series were used in the two types of crosses, as shown in table 6.

Table 6.—Linkage values derived from beet crosses in the coupling and repulsion phases
CrossesParental combinationsRecombinationsTotal
Coupling phase crosses. Recombination percentage=7.3:
ry/ry X RtYr/ ry97114145230
RtYr/ry X ry/ry145137712301
ry/ry X RtYr/ry8916298268
RtYr/ry X ry/ry1331091211265
Repulsion phase cross. Recombination percentage=8.8:
Ry/ rYr X ry/ry3153072337682

Kajanus also investigated the genetics of leaf color, but found the ratios less satisfactory than for root color. In the red-fleshed beets the distribution of anthocyanin is sometimes general throughout the stems and leaves and sometimes confined to the stems, petioles, and larger veins, the rest of the leaf being green. Crosses between red- containing phenotypes and entirely green-leafed ones gave conflicting results. Sometimes green appeared to be dominant to red and at other times it was clearly recessive. He interpreted this paradox to indicate the presence of a dominant inhibitor for red leaves in some genotypes which is absent in others.

Finally, for root shape, Kajanus postulated four pairs of genes:

It has already been noted that the root crops carry several genes for self-incompatibility, and Beta vulgaris is no exception. Although no attempt has been made to study the inheritance of self-incompatibility in Beta, it is easy to keep these genes out of the populations by selecting self-fertile inbred lines. The Division of Sugar Plant Investigations of the Bureau of Plant Industry has selected several completely self-fertile strains of sugar beets, and Roy Magruder, of the Division of Fruit and Vegetable Crops and Diseases, has several promising inbred lines of garden beets.

Dudok van Heel (5) with sugar beets and Bateson (10) with sugar, garden, and stock beets, have investigated the inheritance of bolting.  Both investigators thought inbreeding in beets was impossible because of self-incompatibility, and thus made their selections from open-pollinated mother plants. Despite the poor pollen control, both were able in a short time to select lines that were practically or entirely nonbolters under any conditions. Bateson rapidly eliminated bolters by forcing under glass. Dudok crossed high-bolting and low-bolting strains and found evidence in F1 that low bolting or nonbolting was dominant, but he attempted no further analysis.

In view of the fact that in radishes Frost found high bolting to be completely dominant to low bolting in the ratio of 3:1 in F2, and that in rutabagas Malinowski found low bolting partially dominant in F1 and F2 with indication of multiple factors, it is regrettable that more complete information is not available for beets. It is probable that many genetic factors are in operation, from the fact that a wide range exists between beet varieties in length of time to market maturity.


The project of producing inbred lines in carrots as the initial step in genetic analysis was begun at Davis, Calif., for the purpose of general genetic analysis, variety improvement, and increase of carotene content. The attempt to increase carotene content in the carrot is important, because it is a rich source of vitamin A, so necessary in the correction and prevention of certain nutritional disorders. Betacarotene and vitamin A appear to produce identical physiological effects. The general genetic analysis includes the following studies: Branched roots, cracked roots, root shape; wild v. cultivated root type; three different leaf colors; two flower colors; purple central flower v. normal; red, white, yellow, purple, and orange root colors.

At the outset it was observed that inbreeding reduced plant vigor, but the extent of reduction was difficult to learn, because some carrot varieties are genetically fast growing while some are slow growing. A barrier to studying the effect of inbreeding on carotene content was encountered in the difficulty of securing representative carotene samples from individual roots. Carotene is not uniformly distributed throughout the root, but apparently is first deposited at the top of the phloem zone, whence it gradually diffuses to the bottom of the phloem, and then to the top and bottom of the xylem core. After many trials it was learned that the most reliable index of carotene content was a colorimetric reading of the total extracted pigment before separation of carotene and xanthophyll.  The correlation between the color reading and milligrams of carotene per 100 g of dry matter is very high. Since the pigment is fairly easily extracted, but the carotene determination is made only with much labor, the method is a valuable aid to carrot breeding for increase of carotene (6).

Studies on average root weight of inbred lines, and of hybrids among them, show that the numerous growth factors are segregated to the various lines in a manner corresponding to that for inbred maize lines. Furthermore, the crossing of enfeebled inbred carrot lines gives the same manifestation of hybrid vigor as in F1 hybrids of maize inbred lines. Presumably the hypotheses developed for size and weight inheritance in maize are applicable also to carrots, and we may expect important contributions from carrot genetics to the solution of the exceedingly difficult problem of size inheritance. An illustration of production of F1 generations with excessive hybrid vigor is shown in the following cross between two inbred carrot lines at Davis, Calif. (data unpublished):

Hybrid vigor in a carrot cross
[Average weight, in grams]

        Parent lines, A, 34.28; B, 36.42
        Fourth-generation inbred, A, 12.85; B, 24.87
        F1 hybrid, 80.5.

The variety was Danvers Half Long and parent lines A and B were from strains inbred one generation. A planting of the variety from noninbred seed grown and pulled at the same time had a mean weightof 75 g. Since carrots increase in weight as they grow, it is necessary to pull at the same time all roots that are to be compared.

Some work is being done in California on analysis of root color in carrots.7 In the cross White Belgian x Yellow Belgian the F1 generation showed plants with the following different phenotypes:
  1. White xylem, lemon phloem, white skin
  2. Lemon xylem, white phloem, white skin
  3. "Ringing” of lemon and white, white skin

In a cross between an Indian purple carrot (purple phloem) x Commercial Danvers Half Long (orange phloem), the F1 showed varying degrees of purple.

From the foregoing analysis of Yellow Belgian and White Belgian it is probable that the picture of a simple monohybrid contrast given by Tschermak (38) is inadequate.


 (2)  Borthwick, H. A., and Emsweller, S. L.  1934, CARROT BREEDING EXPERIMENTS. Amer. Soc. Hort. Sci. Proc. (1933)  30: 531-533.
 (3)  Brewbaker, H. E.  1934. SELF-FERTILIZATION IN SUGAR BEETS AS INFLUENCED BY TYPE OF ESOL ATOR AND OTHER FAcToRS. Jour. Agr. Research 48: 323-337, illus.
 (5)  Dudok van Heel, J. P.  1927. aenuoM OF BOLTING IN SUGAR BEET. Genetica 9: [217]-236, illus.
 (6)  Emsweller, S. L., Burrell, P. C., and Borthwick, H. A.  1936. STUDIES ON THE INHERITANCE OF CAROTENE IN CARROTS. Amer. Soc. Hort. Sei. Proc. (1935) 33: 508-511, illus.
 (7)  Findlay, W. M.  1931. A DISEASE-RESISTING TURNIP. Scot. Jour. Agr. 14: 173-183, illus.
 (8)  Frandsen, H. N., and Winge, Ö.  1932. BRASSICA NAPOCAMPESTRIS, A NEW CONSTANT AMPHIDIPLOID SPECIES HYBRID. Hereditas 16: {212}-218.
 (9)  Frost, H. B.  1923. HETEROSIS AND DOMINANCE OF SIZE FACTORS IN RAPHANUS. Genetics 8: 116-153, illus.
(10)  Hall, A. D.  1928. BATESON’S EXPERIMENTS ON BOLTING IN SUGAR BEET AND MANGOLDS. Jour. Genetics 20: 219-231.
(11)  Hallqvist, C.  1915. BRASSICAKREUZEN. Bot. Notiser 1915: 97-112.
(12)  ― 1916. EIN NEUER FALL VON DIMERIE BEI BRASSICA NAPUS. Bot. Notiser 1916: 39-42, illus.
(13)  Jenkins, M. T.  1936. CORN IMPROVEMENT. U.S. Dept. Agr. Yearbook 1936: 455-522, illus.
(14)  Kajanus, B.  1912. GENETISCHE STUDIEN AN BETA. Ztschr. Induktive Abstam. u. Vererbungslehre 6: [137]-179, illus.
(15)  ― 1912. GENETISCHE STUDIEN AN BRAssica. Ztschr. Induktive Abstam. u. Vererbungslehre 6: [217]-237, illus.
(16)  ― 1912. MENDELISTISCHE STUDIEN AN RÜBEN. Fibling’s Landw. Ztg. 61:142-149.
(17)  ― 1913. ÜBER DIE VERERBUNGSWEISE GEWISSER MERKMALE DER BETA- UND BRASSICA-RÜBEN. 1. BETA. Ztschr. Pflanzenzücht. 1: [125]}-186, illus.
(18)  ― 1917. ÜBER DIE VERERBUNGSWEISE GEWISSER MERKMALE DER BETA- UND BRASSICA-RÜBEN. II. BRASSsIcA. Ztschr. Pflanzenzücht. 1: [419]}-463, illus.
(19)  ― 1917. ÜBER BASTARDIERUNG ZWISCHEN BRASSICA NAPUS L. UND BRASSICA RAPA L. Ztschr. Pflanzenzücht. 5: 265-322.
(20)  ― 1917. ÜBER DIE FARBENVARIATION DER BETA-RÜBEN. Ztschr. Pflanzenzücht. 5: [357]}-372.
(21)  Karpechenko, G. D.  1928. FOLTELOW HYBRIDS OF RAPHANUS SATIVUS L. X BRASSICA OLERACEA... Ztschr. Induktive Abstam. u. Vererbungslehre 48:1-85, illus.
(22)  ― 1930. A CONTRIBUTION TO THE SYNTHESIS OF A CONSTANT HYBRID OF THREE sPEcIES. Vsesoiuz. Sezd Genetike, Selek., Semenov. i Plemenn. Zhivotnov. Trudy (U.S.S.R. Cong. Genetics, Plant and Anim. Breeding Proc.) 2: 277-294, illus. [In Russian. English summary, pp. 293-294.]
(23)  Keller, W.  1936. INHERITANCE OF SOME MAJOR COLOR TYPES IN BEETS. Jour. Agr. Research 52: 27-38.
(24)  Lindhard, E., and Iversen, K.  1919. VERERBUNG VON ROTEN UND GELBEN FARBENMERKMALEN BEI BETA-RÜBEN. Ztschr. Pflanzenzücht. 7: 1-18.
(25)  Malinowski, E.  1916. O DzIEDZICZENIU NIEKTORYCH CECH U RZODKIEWKI (ON THE INHERITANCE OF SOME CHARACTERS IN THE RADISHES). Spraw. Towarz. Nauk Warszawsk. (Compt. Rend. Soc. Sci. Varsovie) 9: 757-776, illus. [In Polish. English summary, pp. 770-776.]
(26)  ― 1924, KILKA OBSERWACJI NAD MIESZANCAMI RODZAJU BRASSICA (QUELQUES OBSERVATIONS SUR LES HYBRIDES DU GENRE BRASSICA). Pam. Zakl. Genetyez. Szkoly Glownej Gosp. Wiejsk. (Mém. Inst. Génétique Ecole Supér. Agr. Varsovie) 2: 145-162, illus. [In Polish. French summary, p. 162.] Reviewed in Ztschr. Pflanzenzücht. 11: 179.
(27)  Matsuura, H.  1933. A BIBLIOGRAPHICAL MONOGRAPH ON PLANT GENETICS (GENIC ANALYSIS) 1900-1929. Sapporo. Ed. 2, rev. and enl., 787 pp.
(28)  Miller, J. C., Cochran, F. D., and Garrison, O. B.  1935. SOME FACTORS AFFECTING COLOR IN CARROTS. Amer. Soc. Hort. Sci. Proc. (1934) 32: 583-586.
(29)  Nagai, K., and Sasaoka, F.  1930. CHROMOSOME NUMBER OF CROP VARIETIES BELONGING TO CRUCIFERAE. ADDITIONAL REPORT. Jour. Okitsu Hort. Soc. 25: 76-77, illus. [In Japanese.]
(30)  Pearson, O. H.  1929. A SUGGESTED CLASSIFICATION OF THE GENUS BRASSICA. Amer. Soc. Hort. Sci. Proc. (1928) 25: 105-110.
(32)  ― 1933. CONSTANT SPECIES HYBRIDS. Amer. Nat. 67: 188-190.
(34)  Sylven, N.  1927. KREUZUNGSSTUDIEN BEIM RAPS (BRASSICA NAPUS OLEIFERA). I. BLUTENFARBEN. Hereditas 9: 380-390. (English summary, p. 390.)
(35)  Terasawa, Y.  1932, KONSTANTE AMPHIDIPLOIDE BRASSICO-RAPHANUS BASTARDE. Proc. Imp. Acad. Tokyo 8: 312-814, illus.
(36)  Trouard-Riolle, Y.  1916. HYBRIDATION ENTRE UNE CRUCIFERE SAUVAGE ET UNE CRUCIFERE CULTIVEE A RACINE TÜBERISf&E. Compt. Rend. Acad. Sci. (Paris) 162: 511-513.
(37)  ― 1920. LES HYBRIDES DE RAPHANUS. Rey. Gén. Bot. 32: [438]-447, illus.
(38)  Tschermak, E. V.  1916. ÜBER DER GEGENWARTIGEN STAND DER GEMUsEziicutune. Ztschr. Pflanzenzücht. 4: 65-104.
(39)  Uphof, J. C.  1924. ON MENDELIAN FACTORS IN RADISHES. Genetics 9: [292]-304, illus.
(40)  Vavilov, N. I.  1926. STUDIES ON THE ORIGIN OF CULTIVATED PLANTS. Trudy Prikl. Bot., Genetike i Selek. (Bull. Appl. Bot., Genetics and Plant Breeding) 16 (2): 1-248, illus. [In Russian and English.]
(41)  Vavilov, N. I.  1935. [Botanical-geographic principles of selection.] 60 pp. Moskva and Leningrad. [In Russian. Translation in U.S. Dept. Agr., Bur. Plant Indus.]
(42)  Winge, Ö.  1932. ON THE ORIGIN OF CONSTANT SPECIES-HYBRIDS. Svensk Bot. Tidskr 26: [107]-122.
(43)  Young, R. A.  1924, TAROS AND YAUTIAS: PROMISING NEW FOOD PLANTS FOR THE SOUTH. U.S. Dept. Agr. Bull. 1247, 24 pp., illus.
(44)  ― 1936. THE DASHEEN; A SOUTHERN ROOT CROP FOR HOME USE AND MARKET.  U.S. Dept. Agr. Farmers’ Bull. 1396, 38 pp., illus.

1. Italic numbers in parentheses refer to Literature Cited, p. 322.
2. The stored food is most frequently in the form of starch and starchlike substances, important elements in the human diet.
3. Frandsen and Winge (8), however, in reporting on the cytology and geneties of the progeny from a cross between the turnip and rutabaga, call the turnip parent B. campestris var. sativa rapifera, and Vavilov (41) calls the turnip B. campestris var. rapifera Metzg.

4. Influence on carotene content of inbreeding 4 generations (after Emsweller, Burrell, and Borthwick, 6):

Statistics calculated[Unit measured]Commercial strainInbred line
Roots separately analyzednumber288
Mean carotene contentmg/ 100 g dry matter6656
Range in carotene content32-16436-80
Standard deviation23.004.06
Coefficient of variabilitypercent33.6±3.37.0±0.6
The average carotene content per line among 18 inbred lines descended from 2 original plants varied from 32 to 63 mg per 100g of dry matter. One may infer from the above that lines derived from more than 2 original plants would have given a greater spread and therefore offer even better inbred lines with higher carotene content.

5.  That is, amphidiploids may be formed in two ways—in the gametophyte or sexual generation, and in the sporophyte or asexual generation.
7.  Unpublished data communicated by P. C. Burrell.

[Previous Chapter: Cabbage & other crucifers used for green parts, etc.       Next Chapter: Salad Vegetables
1937 USDA Yearbook Table of Contents        Portal Homepage]