Journal of Heredity 15 (3): 132-144. March, 1924.
REMARKABLE VARIATIONS IN TARWEEDS
Many Abnormalities Found in Plants Long Domesticated Appear
in First Generation Raised in Garden

ERNEST B. BABCOCK
Professor of Genetics, University of California

*The taxonomic status of these two and related forms will be discussed in another place. Both should be included in the collective species, Hemizonia congesta, although one has received the name, H. citrina Greene and the other, H. rudis Benth. For the sake of brevity the two forms treated in this paper will be referred to as citrina and rudis.

DURING the course of a genetical investigation of two of the Hayfield Tarweeds* a number of interesting variations have appeared among populations only one, two, three, or four years removed from the wild state. Some of the variations which were found among about 400 plants during the first season of garden culture are the following: Size of plant, habit (form) of plant, color of flowers (corolla and stamen tube), structure of flowers (petallody), size of leaves (width and length), pubescence on leaves (long, appressed or short, spreading), and other less striking differences. From the first year garden cultures numerous pairs of plants were chosen and crossed because the individual tarweed, like the individual sunflower, is self-sterile. From their progeny similar pairs were again chosen and crossed, and this was repeated, so that inbreeding by brother and sister mating was practiced in some strains for two or three years, when these strains became so weakened that very few progeny were obtained. It was in these inbred strains that most of the other variations herein described were found.

Variations in Size of Plants

There is considerable variation in the size of wild plants of these forms, but probably such extreme differences as appeared in the garden would seldom occur in nature. This would be expected because in the garden nearly every seedling is preserved and each is given an opportunity to develop to the limit permitted by its hereditary determiner (genotype). Many different genotypes as regards size of plant would occur among the wild plants, if there were several Mendelian factors which condition growth and a number of other factors which modify growth and vigor. It must be remembered that a single plant is self-sterile, so that every fertile seed is the result of crossing two plants, a situation which provides for the maximum amount of heterozygosity within the population.

The differences in size of plant found in the first garden culture of citrina were greater than is indicated in Figure 16, A, B, but these plants are illustrated because of the size differences among their progeny. Much more striking differences in size appeared the second year in the garden as a result of segregation of size factors. This is shown in Figure 16, C, D, which were progeny of the same parents, Figure 16, A, B. Furthermore the general result of inbreeding was to cause reduction of size and vigor and loss of reproductive power. Thus Figure 16, E and F, are members of the same family after two generations of inbreeding, and they represent about the average of the eight plants in the family. One inbred strain, however, did not follow the usual course, but instead displayed increased vigor during three generations of inbreeding and developed certain remarkable abnormalities in structure of the flowers (see Figure 21). Several of the inbred strains also produced some seedlings with retarded development, some of which failed to reach maturity, while others produced tiny plants one or two inches high bearing two or three flower heads, and others grew somewhat larger. Similar variations in size of plants were found in the other subspecies, rudis. Thus it appears that there are many genetic factors which determine or influence size and vigor of the plant in these tarweeds, and that, in a wild population, any two plants which happen to look quite similar may possess either similar or different size factors with resulting similarities or differences among their progenies when tested in the garden or subjected to inbreeding.

Variations in Habit or Form of Plants

In wild plants of these species the single erect stem is usually branched above, somewhat as in Figure 17, A, so that the flowers are lifted above the grass or stubble, among which the plants usually grow. The diversity of habit found among the plants of the first garden generation was truly remarkable. Some of the most striking types are shown in Figure 17, A-E, all of which appeared in the first garden culture of citrina. Many of these first season's plants had the slender, wiry stem and branches seen in Figure 17, A and C, but others had the robust, succulent appearance, the many branches, and the pronounced leafiness of Figure 17, B, and it was from two such plants as the latter that the robust inbred strain mentioned above was derived. The procumbent habit exhibited by the plant shown in Figure 17, C, was inherited as a Mendelian recessive character. The next plant, Figure 17, D, was a great-grandparent of the one just below, Figure 17, F, which is of the stricta type. This type of habit breeds true and behaves as a Mendelian character. In Figure 17, E, is shown the short-stemmed spreading habit which proved to be' unstable and is the expression of a heterozygous genotype. Thus we find that the Hayheld Tarweeds, when removed from their usually crowded and often semi-arid environment, are capable of producing a striking array of plant forms. Some of these forms, particularly the procumbent habit (Figure 17, C,), would tend to be eliminated by natural selection in the wild state, but the recessive factor conditioning it would occur in heterozygous combination among a certain number of the wild plants, so that plants of this habit would be expected to appear from time to time. Other forms, such as stricta, would tend to be preserved and to segregate out from time to time, although often partially masked by the influence of the environment or of other factors as in Figure 17, D. The masking of the potential genetic qualities of these plants both by their rigorous natural environment and by the heterozygous condition is of considerable interest to taxonomists as well as to plant breeders.

It is noteworthy, too, that, although the first garden cultures of rudis were fully as extensive as those of citrina, and although all the citrina plants were grown from seed collected from a few plants in one spot, while the rudis cultures came from seed collected in several widely separated localities, yet the rudis cultures were more uniform as regards habit of plant than were the citrina. There were only two definite habit types among the rudis plants with minor variations, while the citrina population contained the five types shown in Figure 17 and others less striking. This difference may have some phylogenetic significance.

Variations in Color of Flowers in Citrina

The conspicuous portion of the flower head in these plants is the outer circle of ligules or rays, which in the wild plants are bright yellow of the degree called lemon chrome in Ridgway's Color Standards and Nomenclature. Slight variations appear among wild plants, and by sufficient searching one would probably discover the more extreme variations which appeared in the garden cultures. The next lighter color than lemon chrome (illustrated in Ridgway) is lemon yellow, and this was noted in two or three plants in the first garden generation. One of these (Figure 16, B) was crossed with another (Figure 16, A) whose flowers were lighter than lemon chrome (probably Ridgway No. 22, which comes between lemon chrome. No. 21; and lemon yellow, No. 23, but is not illustrated), and among the seven offspring secured four had lemon chrome, two had lemon yellow, and one, pale lemon yellow flowers (23b). The progress made during three generations in attempting to establish a strain of citrina with very pale yellow flowers is summarized in Table 1. In the same table the reduction in size of plant as a result of only two generations of inbreeding is also. shown. It will be noted that in F3 three of the eight plants had flowers that were lighter than the pale lemon yellow of Ridgway, but that all eight plants were so small and few-flowered that inbreeding could not be longer continued. Accordingly, this strain was outcrossed to a strain with lemon chrome flowers. As a result of this experiment it became evident that there are several factors for the yellow color of the rays in citrina. This is further indicated by the results of crossing citrina with rudis which has white rays, except for a reddish purple tinge on the outside of the ray in some plants. The flowers of the F1 generation plants were intermediate in color, and in the F2 generation nearly all of the 181 plants. bore flowers of some intermediate shade — only one being recorded as having lemon chrome rays and none with pure white ligules, although three or four plants had flowers of a very pale greenish yellow tint.

Another color variation was discovered in one plant of the first garden generation, viz., cream stamen tubes. Usually the stamen tube is a deep purple color, but in this one plant there was no purple color present in the stamen tube. This character was inherited as a simple Mendelian recessive.

Variations in Structure of Flowers

Petallody. In the first garden culture of rudis one plant was found with "double ligules" clue to the development of petal-like structures within and near the base of the corolla-tube (Figure 19, B). Among twenty-five seedlings of this plant (from unguarded seed) one was found which exhibited petallody in part of its ray flowers. It was crossed with a normal sib, and among the eighteen seedlings obtained, there was one which displayed an extreme form of petallody. From such slender evidence no conclusion can be reached as to the mode of inheritance of this interesting variation but that it is inherited seems fairly certain. Later this teratological form appeared occasionally in cit rina (Figure 19, C-F). It is probable that in a species more amenable to inbreeding it would have been possible to secure a strain with "double ligules."

Collarette. (Figure 18). This striking type of flower head, so named because of its similarity to the well-known Collarette Dahlia, first appeared after two generations of inbreeding in a strain selected for the stricta type of habit. Nothing unusual had been noted about the flower heads, except that the rays were in two rows, a not uncommon condition. This sudden appearance of a character new to the species simulates mutation, but a little consideration of the circumstances leads to the conclusion that it was due to a fortuitous combination of genetic factors and is reversionary in nature. This view is strengthened by the fact that one of the prototypes of the Compositae is believed to have had a five-lobed, bilabiate corolla, in this connection it is of interest to note that of the three individual florets shown in Figure 18, the right-hand one has its posterior lip notched at the top, thus producing a five-lobed corolla. This, so far as the writer is aware, is the first time the bilabiate type of corolla has been reported in the Tarweed Subtribe (Madinae) of the Compositae.

Corolla Lobes. Variation in the number of lobes in rays which were otherwise normal was also worthy of note. As is shown in Figure 19, A, C, the normal number is three, but in several of the strains, including those showing petallody, rays with four and even five lobes occur. Figure 19, E, F. Reduction in the number of lobes was characteristic of certain other inbred strains of citrina, particularly those having lighter flower color. This tendency became so marked after one generation of inbreeding that the descriptive name biloba was added. In every case reduction was accomplished by omission of the middle lobe.

Corolla-Tube. In the normal ray flowers the corolla is tubular only at the base (Figures 18 and 19), but in one of the first stricta plants showing the collarette type in some of the rays, other rays were found in which the tubular portion was half the length of the corolla. This type of variation occurs frequently in the common Gaillardia and other cultivated species of the Compositae.

Proliferation and Virescence

After one generation of inbreeding in one of the robust strains of citrina a family was produced in which all of the plants were exceptionally late in maturing but remarkably vigorous. When these plants began to flower, it was found that, in some heads at least, only the ray flowers were normal. Instead of disk flowers foliaceous bracts developed, as shown in Figure 21. Later a whorl of branchlets appeared around the margin of the disk and inside the whorl of rays (Figure 20). Similar cases of proliferation have been found in the common marigold.

*BISSET, P. Prolification in a double-flowered form of Calendula officinalisJour. Hered. 9: 323-325. 1918.

In Bisset's report of such a case* he attributes the abnormality to environmental conditions but states that it appeared on several of the plants. As the plants seem to have been grown under uniform greenhouse conditions, it may fairly be questioned why it did not appear on all the plants, if it were due merely to over-stimulation, as Bisset assumes. In view of the present evidence it seems more probable that the plants which exhibited prolification were different genetically from those which did not. It is possible, of course, that the prolificated plants would not have developed the abnormality under ordinary garden conditions, and, if so, the expression of their genotype was conditioned by their environment. But the writer has seen the same abnormality in the same plant when grown in the garden without special care, and the same type of abnormality is found occasionally even in wild plants.

For example, I have in my possession prolificated specimens of Erodium moschatum, a common native forage plant of California. These were collected by Mr. Joseph P. Tracy of Eureka, who wrote as follows: "This is a "freak" form, its peculiarity being in a tendency of the umbels to be compound, i.e., to produce secondary umbels or small branches, instead of the usual, few, short-pedicelled flowers. I first noted this form in a single plant in the spring of 1911 on G Street, where I pass daily going to and from work. This spring, 1912, there were many plants scattered for half a block." There can be no question as to a genetic cause for this case of prolification, any more than there can be for the tarweed case described above. It may be inferred. therefore, that such abnormalities always depend upon a germinal change or difference of some sort, although the manifestation of that germinal condition may in turn depend upon the environment.

Two plants with virescent flowers have been found during six years' observations on hundreds of plants, one being citrina, the other, rudis. The former is shown in Figure 22, as this plant had the additional peculiarity of being chimerical. All of the branches on one side of the plant were several inches longer than the remainder. All of the flowers were virescent and proliferous, both ray and disk flowers being much elongated (Figure 23). The ovary when present was about one-fourth inch long and greenish. The ray flowers were tubular, sometimes being partially expanded into a greenish yellow ligule. The style and stigma were usually present and apparently normal. The disk flowers were two to four times the normal length and virescent; the anthers were brownish instead of purple and apparently devoid of pollen.

TABLE I—Pedigree of Pale Yellow Citrina Strain.

P1 C 2-17 (Fig. 16, A) 19"
Ridgway No. 22 ?
X C: 9-16 (Fig. 16, B) 15"
lemon yellow (R,23)

16.8 — 7 plants - 4 lemon chrome: 2 lemon yellow: 1 pale lemon yellow.
 
F1 P2 (Fig. 16, C) 15"
lemon yellow (R,23)
X P6 (Fig. 16, D) 6"
pale lemon yellow (R,23b)

17.44 — 3 plants lived - 2 lemon yellow: 1 pale lemon yellow
       
F2 P4 (not illus.) 13"
pale lemon yellow (R,23b)
X P2 (not illus.) 12"
lemon yellow (R,23)

20.88 — 8 plants - 1 R, 22: 14 R,23: 3:R,23d.
     
P8 (Fig. 16, E) 8"
lemon yellow. (R,23)
—— P7 (Fig. 16, F) 6"
picric yellow (R,23d)
Height of plant is stated in inches after number of illustration. Ridgway's No. 22 is intermediate between lemon chrome and lemon yellow; lighter tints than lemon yellow are pale lemon yellow (23b), picric yellow (23d), and martius yellow (23f), with 23c and 23e unnamed.

Conclusion

It has long been known that introduction of a wild plant species into cultivation induces variation. But the reason why some of the variations so induced were easily "fixed" while others were not remained unexplained until the discovery of Mendelian inheritance and the development of the science of genetics provided the necessary knowledge of heredity. It seems to have been the idea of many of the older writers on the subject of plant improvement that the horticulturist's art could, in some mysterious way, "break the type" or cause variation. In the case of these wild tarweeds it is very clear that the cause of the striking variations which appeared in the very first garden cultures are fully explained by the simple fact that the individual plant is self-sterile. It is probable that many problems in practical plant improvement would be simplified by a preliminary study of the plant itself, especially its provision for sexual reproduction.

During the first four years that these tarweeds were under cultivation many of the types of abnormality common among species long domesticated have been discovered. In the case of ornamental plants some of these abnormalities would be considered to be improvements, and they might be thought to be modifications induced by cultivation, but their presence in the first garden cultures or appearance after only one, two or three years of inbreeding shows that they are merely recessive characters which are represented in the germ cells but are seldom displayed by the wild plants which are cross-pollinated and hence heterozygous. The occasional appearance of such abnormal types even among wild plants would be due to the chance meeting in fertilization of two germ cells bearing the recessive factors for the abnormality. This may also explain the occurrence of such forms among the seedlings of cross-pollinated garden plants.

The demonstration that there is a genetic basis for such characters as plant size, plant habit and time of maturity is of some significance in taxonomy where the tendency may be strong to assign all such variations to ecological factors.


VARIATION IN SIZE OF CITRINA TARWEED PLANTS

Figure 16. Two plants of the first generation raised in the garden are shown at A and B. They are nineteen and fifteen inches high respectively. C and D are the largest and the smallest plants in the progeny of A and B, and are fifteen and six inches high. The next generation is not illustrated, but the three plants averaged less than fourteen inches in height. Two of them were crossed and produced a family of eight plants, averaging less than eight inches in height. Two of these are shown at E and F.


VARIATION IN HABIT OF GROWTH IN CITRINA TARWEED PLANTS

Figure 17. A, B, C, D, and E are types found in the first generation grown in the garden, from seed of wild plants. A resembled the usual habit of wild plants. B resembled the progenitors of the most robust inbred strain. C was unique, and transmitted its habit of growth as a recessive character. D transmitted the stricta habit of growth to its great-grandchild, F. The type shown in E proved unstable.


THE "COLLARETTE" TARWEED

Figure 18. Below is shown one head from a plant of the collarette type of citrina, with all but three of the ray flowers removed. Note the posterior erect lip in each corolla. In the upper left-hand corner is shown a similar corolla attached to its achene (seed). Opposite it in the right hand corner is shown another two-lipped corolla seen from above, showing the two lobes of the posterior lip. Between these is shown another ray flower from the same head, lacking the posterior lip, but with a small extra lobe on the left side of ray.


DOUBLE RAY FLOWER IN TARWEEDS

Figure 19. A, normal flower head of rudis; B, head from rudis plant in first garden culture, showing two extra petal-like organs within each ray; D, a normal three-lobed ray; E, a four-lobed ray; F, a five-lobed ray, each containing one petal-like organ. A and B are enlarged about one and one-half times; C about three times; and the others about two and one-half times.


LEAF-LIKE BRACTS IN THE FLOWER-HEAD

Figure 20. Note the foliaceous bracts developing inside the circle of ray-flowers. Such a head is shown in Figure 21. after branches have developed from the region of these bracts. The occurrence of so many variations in the early generations raised in the garden shows that many of the abnormal plant forms developed by man are present in the wild stock, only awaiting selective breeding to bring them to light.


PROLIFERATION OF INFLORESCENCE

Figure 21. This shows a head similar to that shown in Figure 20, after it has been covered with a paper bag for some weeks. The whorl of branches has developed from the outer region of the flower disc, but within the circle of ray flowers, which are normal and fertile. Natural size.


A CHIMERICAL, VIRESCENT TARWEED PLANT

Figure 22. The five branches forming the right-hand portion of this citrina plant were several inches longer than the remaining ones. All of the flowers were abnormal and sterile, with virescent (green) flower parts. See Figure 23.


NORMAL AND VIRESCENT TARWEED FLOWERS

Figure 23. a, Normal flower head and bud of normal citrina; a', normal ray flower with achene surrounded by involucral bract; a", normal disc flowers. x, flowers from plants shown in Figure 22; x', ray flowers from this plant, the left one with involucral bract; and x", the disc flowers. Instead of normal ray flowers, green foliaceous bracts have developed.