Journal of Heredity 22(11) :328-343. (Nov. 1931)

Texas Agriculture and Experiment Station, College Station, Texas

Pistillate spikelets of Tripsacum, a wild relative of maize. The silks are less than all inch long and are divided. The seed is embedded in an indurated shell consisting of the rachis and the outer glume. Enlarged 4.7 times.

Figure 1
The Texas form (left) has narrower, stiffer, and more erect leaves than the Connecticut form (right). The somatic chromosome numbers are 36 and 72, respectively.
Figure 2
The husks are removed and the long silks are cut off to an inch or less in length. After pollination, an artificial shuck is provided to prevent contamination.

THE tribe Tripsaceae (Maydeae) is represented in America by three genera, Zea, Euchlaena, and Tripsacum. The first includes but one species, Zea mays, the well known and widely grown Indian corn.

*Quoted by Lamson-Scribner,
U.S.D.A. Farmers' Bull.,
No. 102.

Euchlaena, commonly known as teosinte, comprises at least two species. Annual teosinte, Euchlaena mexicana, is a plant which, when first discovered, was thought to be the prototype of maize, later considered as one of the parents of the hypothetical hybrid from which maize was supposed to have originated, and is now regarded by some as merely a closely related species, perhaps descending with maize from a common ancestor. It was formerly grown, and still is to some slight extent, as a soiling crop in the Southern states, though even for this purpose it has been largely replaced by other crops. Asa Gray's hope that teosinte is a plant "possibly affording an opportunity for one to make millions of blades of grass grow where none of any account grew before"* has not been realized.

Perennial teosinte, Euchlaena perennis, like the annual form, is native to Mexico, from where it was introduced into this country some years ago by Collins.3 As a crop plant it has even less value than annual teosinte but it has proven very valuable for certain genetic and cytological studies.

Tripsacum, the third American genus of this tribe, is represented by six or seven species. Three of these, T. dactyloides, T. fioridanum, and T. lemmoni, are indigenous to the United States while the others are found in tropical America. T. dactyloides is the most common species in the United States where it occurs naturally in moist places throughout the eastern half of the country, extending as far west as Kansas and Texas, and north to Connecticut. It has never been grown as a crop plant though several botanists, in the past, have suggested its possible value as a forage crop.

Previous Crossing Experiments

Both species of Euchlaena are fairly closely related to maize. They can be readily crossed with the latter and produce fertile or partially fertile hybrids. Their chromosomes are at least partially homologous with those of maize, as shown by the genetic studies of Emerson7 and the cytological observations of Longley11 and Randolph.†

†Unpublished, cited by Emerson.7

Tripsacum, on the other hand, is generally regarded as being much less closely related to maize. Weatherwax,13 for example, states: "Except for monoecism this genus resembles some of the Andropogoneae, such as Manisuris, much more than the other Maydeae." Collins,3a in a recent paper which has just come to hand, expresses a similar opinion in different language. He states: "The relationship between Tripsacum and Zea is remote and the chief reason for introducing Tripsacum into a discussion of the phylogeny of maize is that Tripsacum is the only American representative of the group of grasses to which Euchlaena and Zea belong, that is found growing outside of cultivated areas."

Tripsacum does not cross naturally with maize. Neither natural nor artificial hybrids of these genera have ever been previously recorded, with the possible exception of some "false" hybrids reported by Collins and Kempton.4 These writers attempted to hybridize Tripsacum with both Zea and Euchlaena. Their cross of Tripsacum X Zea produced only matroclinous plants, presumably the result of parthenogenetic development of the Tripsacum ovule. The cross Tripsacum X Euchlaena was even more surprising for it gave rise to a single plant which proved to be pure Euchlaena like the pollen parent, a condition for which they suggested the term "patrogenesis."

Parthenogenesis as a Means of Improvement

These results are important from several standpoints. The patroclinous plant from the cross Tripsacum X Euchlaena presumably originated through the development of a sporophyte from a male gamete and is one of the few recorded instances of this phenomenon. The matroclinous plants from the cross Tripsacum x Zea indicate that the female gametes also are capable of parthenogenetic development. The occurrence of these two types of parthenogenesis in two genera related to maize, one quite closely related, suggests some slight possibility of developing homozygous strains of maize through parthenogenesis as a substitute for prolonged inbreeding. This would be very important from the standpoint of crop improvement, both in saving time and labor, as East6 has recently pointed out, and in greatly increasing the opportunities of isolating extremely favorable combinations of genes because selection involves the gametes rather than the zygotes.

The relative efficiency of gametic selection and zygotic selection can be illustrated by a simple example. If ten allelomorphic pairs of genes are involved, one member of each pair being favorable to development, the other not, then one gamete, or one sporophyte originating from a gamete, in 1024 would receive all favorable genes, while with sexual reproduction only one sporophyte in 1,048,576 would be expected to receive all favorable genes. Planted a foot apart in three-foot rows it would require only about one-fourteenth acre to grow the former number of plants and more than 72 acres to grow the latter. Even if unlimited land were available it would be virtually impossible to identify the desired plant when sexual reproduction is involved.

Though the differences between gametic and zygotic selection would not, in actual practice, be as simple as this, there is no question that the possibility of gametic selection through induced parthenogenesis is one which has not been sufficiently considered and one which would justify continued effort in attempting to induce parthenogenesis in maize, either by hybridization or by artificial, chemical, or physical stimuli, such as X-rays. It offers the only reasonable hope of obtaining homozygous strains of maize equal or superior in productiveness to the best first generation hybrids, a goal which is scarcely worth attempting by present methods of selection, as Jones8 has so clearly pointed out.

With these considerations in mind we have made extensive efforts to induce parthenogenesis in maize by hybridization with Tripsacum and, though our results to date are negative, we have succeeded in obtaining true hybrids of Zea and Tripsacum. It is to report the occurrence of these hybrids that this account is primarily written.

Crossing Maize and Tripsacum

Our experiments with Tripsacum have involved mainly only two forms. One of these, hereafter known as Texas Tripsacum, was collected near Angleton, Texas, where it is common in the prairies of the Gulf Coast Region. This form has stiff, erect, narrow leaves and resembles T. floridanum in some respects, though it has been identified as T. dactyloides by Dr. A. S. Hitchcock. Its somatic chromosome number is 36. Practically all of our pollinations have been made with this form.

The second form, hereafter known as Connecticut Tripsacum, was collected in a marsh near New Haven, Connecticut. It is typically T. dactyloides. The somatic chromosome number is 72 and there is some indication that it is a tetraploid, the chromosomes during meiosis showing some tendency to assort in tetrasomes instead of disomes. The two strains are illustrated in Figure 1.

In 1929 pollen from the Texas Tripsacum was applied to the silks of maize. Within 24 hours the silks had wilted, as they normally do when corn pollen is applied, and microscopic examination showed that the pollen grains of Tripsacum had germinated and many of the pollen tubes had entered the styles.

Figure 3

Figure 4
The small watery "seeds" are not seeds, but merely swelled nucelli resulting from the stimulus of pollination. The true seeds on this ear, which are larger and more opaque, are hybrids and contain both embryo and endosperm. Some ears set seed almost 100 per cent. The hybrid seeds, however, are very small compared to maize seeds of the same age. A few maize seed near the middle resulted from pollinating with maize 24 hours after pollinating with Tripsacum. Note that some of the hybrid seeds adjacent to the maize seeds are larger than their sibs.

No seed development occurred, however, and fertilization presumably had not been accomplished. Since the silks of the Texas Tripsacum are usually less than an inch in length, while those of maize are generally several inches to more than 12 inches in length, the distance to be traversed by the Tripsacum pollen tubes in reaching the micropyle in maize is many times as great as that normally required of the tubes in fertilizing their own species. The obvious solution of this difficulty was to shorten the styles of maize. This was first tried in a group of 10 plants. In half of these plants the pollinations were made on silks of normal length. In the other five plants an incision was made through the husks on one side of the ear and the silks from a narrow region of the ear were cut back to less than an inch in length before the

Tripsacum pollen was applied. The results were very decisive. In the five ears with normal silks no development of any kind occurred. Of the five ears with shortened silks every one produced some seeds and these seeds were confined to the narrow region of the ear in which the silks had been shortened.

The season, by this time, was considerably advanced and few maize plants were available. Twenty-seven additional ears were pollinated, however, and from these a few very small, shriveled seeds were finally matured. Several typical maize seeds were also produced which later proved to be heterozygous and were undoubtedly due to contamination.

Figure 5
Most of the hybrid seeds are very abnormal in development. These photomicrographs show some of the peculiarities which occur. (A) Embryo and endosperm normal in structure though reduced in size. (B) Embryo severely distorted. (C) Embryo distorted and endosperm lobed. (D) Embryo at right angles to axis of seed.

In 1930 these experiments were repeated on a more extensive scale with much greater precautions against contamination. Seven ordinary commercial varieties of maize and several genetic stocks were grown in an isolated plat surrounded on two sides by woods and on the other two by buildings with no other maize nearer than 300 yards, and this to the north, directly opposite from the direction of prevailing summer winds. Tassels were removed each morning as soon as they had emerged and before pollen was shed. In addition, the usual precautions during pollination were observed. All ear shoots were bagged before the silks had emerged and ears were immediately covered after pollination.

In pollinating these plants with Tripsacum the entire shuck was removed by a circumcision at the base of the ear. The silks were then cut off with an ordinary pair of shears. After pollination, the ear was covered with an artificial shuck to prevent possible contamination by wind blown pollen and also to forestall insect damage and excessive drying. Ordinary crepe paper wrapped over a strip of cotton, wound around the base of the ear, furnished a very satisfactory shuck and one that expanded readily to accommodate the growth of the ear. Figure 2 illustrates the steps in this procedure.

Development of the Crossed Seeds

The results of pollinating 382 ears with Tripsacum are shown in Table I. The number of silks pollinated is an estimation arrived at by multiplying the number of rows on each ear by the number of silks exposed to pollination on one row. The number of seed set was determined by actual count during the third week after pollination, except in two lots, indicated by asterisks, where seeds on only part of the ears were counted, and the total numbers are estimates based on this sample.

In some varieties almost every pollinated ear developed at least a few seeds, some ears setting 90 per cent or more.

There is a great deal of variation in the development which occurs when maize is pollinated with Tripsacum. In addition to the true seeds which contain both endosperm and embryo, and have a white, opaque appearance, there are many parthenocarpic ovaries which possess neither embryo nor endosperm and are watery and translucent in appearance. Both types are illustrated in Figure 3. The development of the parthenocarpic ovaries is the result of a marked swelling of the nucellus which sometimes grows rapidly enough to burst through the pericarp. There is no evidence that fertilization has occurred and it is probable that the pollen tubes in the styles are capable of stimulating nucellar swelling without entering the micropyle.

In addition to the swollen nucelli, thousands of true seeds were formed and only these are included in the counts set forth in Table 1.

Even among the true seeds there is marked variation. Practically all of them, however, begin to shrivel about three weeks after pollination and at maturity only the empty pericarp remains. Anatomical studies of the developing seeds show that many kinds of abnormalities may occur. The embryo and endosperm may be normal in structure, though greatly reduced in size (Figure 5A). More frequently, however, the embryo is severely distorted (Figure 513) and this distortion may occur in a variety of ways. Apparently there is no definite growth pattern and cell division occurs more or less at random. This is also evident in the endosperm which is frequently lobed (Figure 5C). In some cases the embryo seems to have lost its polarity and its axis may lie at various angles to the axis of the seed. Figure 5D shows a section through an ovary in which the longitudinal axis of the embryo lies at right angles to the axis of the ovary. When the embryo and endosperm reach maturity they usually occupy only a small portion of the space within the ovary wall.

These figures illustrate but a few of the peculiarities which occur in the hybrid seeds, and merely serve to emphasize the fact that development, on the whole, is decidedly abnormal. The combination of germplasm from these two distantly related genera is apparently not conducive to normal development, at least in the early stages.

All of the crossed seeds are very much smaller than normal seeds of corn of the same age. Figure 4 illustrates an ear pollinated with Tripsacum and pollinated 24 hours later with maize.

Figure 6
(A) Texas Tripsacum: 36. (B) Endosperm of maize x Texas Tripsacum: 38. (C) Embryo of maize x Texas Tripsacum: 28.
(D) Connecticut Tripsacum: 72. (E) Root tip of maize x Connecticut Tripsacum: 46.

Effect of Maize Seeds on Adjacent Hybrids

Of possible interest in connection with the physiological aspects of nutrition of hybrid seeds is the fact, that hybrids adjacent to pure maize seeds are frequently larger and survive longer than hybrid seeds growing alone. This is illustrated by several seeds in Figure 4 and is further shown by the fact that mature hybrid seeds on ears which were later pollinated with maize were 39.8 per cent heavier than hybrid seeds on ears, of the same varieties, which had received only Tripsacum pollen. (Table II.) This may indicate that the growth of normal seeds stimulates a flow of nutritive materials into the region which they occupy and the hybrid seeds obtain some benefit of this increased nutrition, though they are incapable of inducing it. Perhaps the failure, reported by Laibach,9 of hybrid seeds of flax to develop in the ovaries of the maternal parent, though they could be matured in artificial media, is due to the fact that hybrid seeds are often incapable of stimulating a flow of nutrient materials and not because the maternal tissues are toxic to the developing hybrid, as Laibach seems to imply.

In any event the increased size of hybrid seeds adjacent to normal seeds is an observed fact; one which has proved useful in facilitating the hybridization of Zea and Tripsacum, and may possibly be utilized in the hybridization of other genera or species.

Chromosome Numbers in Embryo and Endosperm

Since one of our chief considerations in making these pollinations was that of inducing parthenogenesis, we have made chromosome counts in the developing seeds to determine whether any of them are haploid or diploid maize or Tripsacum. These counts were made both in the embryo and endosperm since parthenogenesis might conceivably occur in one structure and not in the other. Naturally only a relatively few seeds could be studied cytologically, but since the chromosomal condition was found to he the same in small and large seeds, there is no indication that part of the seeds are parthenogenetic and others true hybrids. The conditions found in all seeds studied are illustrated in Figure 6. Texas Tripsacum, as already indicated, has a haploid chromosome number of 18. The endosperm of the hybrid seeds usually has 38 chromosomes, which is the number expected if triple fusion takes place normally and two haploid sets of 10 chromosomes from corn combine with one set of 18 from Tripsacum. The embryos at these seeds usually have 28 chromosomes, the sum of the haploid numbers of the two parents.

Similar crosses with the Connecticut Tripsacum, (n=36), show 46 chromosomes in the cells of the root tips, the number expected if 36 chromosomes from Tripsacum combine with 10 from corn.

No indication whatever of induced parthenogenesis, either of the female or male gametes, has been observed.

Few Seeds Mature

Most of the hybrid seeds begin to shrivel about three weeks after pollination and at maturity practically all of them consist of little more than the empty pericarp, the endosperm and embryo having been partially or completely resorbed. A few, however, reach maturity, though these, too, are very small, the largest of them being about the size of a grain of pop corn. Table III shows the weight of some of these seeds in comparison with the weight of maize seeds of the varieties on which these hybrid seeds were borne.

Figure 7
The crossed seeds of maize X Tripsacum are shriveled and abortive and germinate only under optimum conditions. Sterile agar has proved the best medium for germination.

It is noted that the average weight of the crossed seeds is 38 mg. as compared to an average weight of 268 mg. for normal maize seeds of the same varieties. The crossed seeds weigh only 14.2 per cent as much, on the average, as normal maize seeds of the same varieties and are about the same size at maturity as normal maize seeds harvested in the third or fourth week after pollination.

How rarely even these small seeds mature is shown by Table I. Out of approximately 185,000 silks exposed to pollination, approximately 35,000 set seed and of these only 84 matured; an average of 4.54 seeds matured per 10,000 silks pollinated.

Apparently there are differences in maize varieties in ability to cross with Tripsacum. From the Horton variety we have never obtained a mature crossed seed, either in 1930 or the previous year. A first generation hybrid of Yellow Creole X Surcropper is intermediate between its two parents, indicating that differences in "crossability" are hereditary, a condition similar to that observed in wheat-rye crosses by Backhouse,1 Leighty and Sando,10 and others.

In addition to the 84 mature crossed seeds obtained by pollinating 382 ears with Texas Tripsacum, we obtained 15 seeds from crossing only a few ears with Connecticut Tripsacum. Apparently the Connecticut form crosses more readily with maize than the Texas form, though the divergence in chromosome number of the parents is much greater in the former cross than in the latter.

Germinating Shriveled Seeds

In both crosses the mature seeds are so small and shriveled that, planted in soil, they would probably never emerge. Even between moist blotters in a germinator the percentage of germination is low. By experimenting with partially developed, shriveled seeds of corn which showed no germination whatever under standard germinator conditions, we were able to develop a technique modified from that of Chen,2 Pearl and Allen,12 and others, which in some cases resulted in 100 per cent germination of seeds that showed no germination between moist blotters. This method has been used with such success in germinating hybrid seed of maize x Tripsacum, maize X perennial teosinte, and other poorly developed seed which would ordinarily be discarded, that it may not be amiss to include a brief description here.

The seeds are soaked about six hours in tap water at room temperature to induce softening or germination of seed borne spores, rendering them more susceptible to the action of fungicides. The pericarp is then removed completely. The seeds are soaked about 8 minutes in a one per cent solution of Semesan after which they are rinsed in boiled, distilled water and scattered in sterile, two per cent, agar in Petri dishes. The Petri dishes are kept in a glass-walled germinating chamber at a temperature of approximately 33° C. Some sprouting usually occurs within 18 hours. As the seedlings are exposed to light at all times, chlorophyll development begins immediately upon germination and this, no doubt, enables the young seedling to overcome partially the handicaps brought about by its deficiency in endosperm material and by its weak root system. Within six days the seedlings usually have two well developed leaves and can be potted off. After potting they are left in the saturated atmosphere of the germinator for several days. (See Figure 7.)

Of the 84 crossed seeds of maize x Texas Tripsacum, 45, or more than 50 per cent, germinated and 29 have survived transplanting in the field. Of the 15 seeds of the cross with Connecticut Tripsacum, only four germinated and only one still survives. Chromosome counts have been made from the root tips of a number of the hybrid seedlings. The chromosome number in the cross maize x Connecticut Tripsacum is 46; in the cross maize x Texas Tripsacum, 28.

After developing an adequate temporary root system, practically all seedlings have grown normally except that several of them became chlorotic in the third or fourth week of development. Some of the leaves were completely albinotic but new leaves emerging later were normal green.

Hybrids Resemble the Pollen Parent

The hybrid seedlings can be distinguished from those of either parent though they resemble Tripsacum, the pollen parent, more closely than maize, the seed parent. In rate of growth, width of leaves, color of leaves, tillering habits, and other characteristics, the resemblance to Tripsacum is very marked. In the cross maize X Connecticut Tripsacum, in which the Tripsacum parent contributed 36 chromosomes and maize but ten, the resemblance to Tripsacum is more pronounced than in the maize x Texas Tripsacum hybrids which received only 18 Tripsacum chromosomes. This is probably to be expected since the proportion of Tripsacum chromosomes to maize chromosomes is twice as great in the first case as in the second.

Figure 8

Figure 9
The hybrid resembles Tripsacum more closely than maize. It tillers profusely, has a perennial habit of growth, and an inflorescence very similar to that of Tripsacum. Most taxonomic botanists would probably call this an inflorescence of Tripsacum, so close is the resemblance to its pollen parent. It does, however, have some characteristics of maize. Note that the silks are fused most of their length but divided at the ends.

The hybrid of maize X Connecticut Tripsacum was planted six months earlier than the others and is the only one which has flowered when this is written. The inflorescence, though showing some characteristics of its maize parent, resembles Tripsacum very closely and would probably be classified as a species of Tripsacum by a taxonomic botanist. Like Tripsacum, the staminate and pistillate spikelets are borne on the same rachis, the former above, the latter below. The pistillate spikelets are enclosed in an indurated shell consisting of the rachis and the outer glume. The terminal inflorescence has five erect branches, two more than ordinarily occurs in Tripsacum. The silks are longer than those of Tripsacum, averaging 38 mm. as compared to 22 mm. for those of the Tripsacum parent. Most of them are divided only at the ends while Tripsacum silks are divided to the base and have the appearance of occurring in pairs. The silk color is pink, being intermediate between the red silks of the Tripsacum parent and the green silks of the maize parent. In time of flowering of the pistillate and staminate spikelets the hybrid resembles Tripsacum, which is always protogynous while maize is usually protandrous. As a matter of fact, the anthers of the hybrid do not dehisce at all because they contain only aborted pollen, but they are still green when the silks emerge and do not turn yellow until several days later.

Preliminary Studies of Meiosis

The hybrid seems to be completely sterile. No pollen has been shed and so far no seeds have set, either when Tripsacum or maize pollen was applied. Preliminary cytological studies of meiosis in the pollen mother cells suggest the probable causes of this sterility, though the details must be verified by further studies now in progress.

At the first division there are usually 18 bivalents and ten lagging univalents. The Tripsacum parent of this hybrid, as has already been suggested, is probably a tetraploid. Thus the hybrid receives a complete assortment of Tripsacum chromosomes and these are capable of pairing among themselves. Presumably the 18 bivalents represent the 36 Tripsacum chromosomes, and the ten lagging univalents constitute the ten maize chromosomes. This situation is similar to that reported by Collins and Mann5 in certain Crepis hybrids.

As the bivalents disjoin and proceed to the poles, the lagging univalents remain on the plate and undergo an equational division.

In the second division the split univalents lag again and many, if not all, of them fail to reach the daughter nuclei. These form many small nuclei and at least some of them give rise to supernumerary microspores. Others remain in the same microspore with the large nucleus and may be one of the chief causes of aborted pollen.

Whether or not any of the maize chromosomes pair with those of Tripsacum or pair among themselves remains to be determined by further studies.

Other Crosses With Tripsacum

The reciprocal cross, Tripsacum X maize, has also been attempted but, so far, without success. Though the pollen of maize germinates readily on the styles of Tripsacum and many pollen tubes enter the stylar tissue, no mature seeds have ever been obtained, perhaps mainly because it is difficult to make the enormous number of pollinations necessary to obtain viable seeds.

The cross of Tripsacum and perennial teosinte has never been successful, no matter which of the two genera was used as the maternal parent. Annual teosinte, however, can be crossed with both forms of Tripsacum if teosinte is used as the seed parent and its silks shortened. Seeds are set less frequently than in the cross, maize x Tripsacum, and all of them, so far obtained, have shriveled before maturity. Perhaps mature seeds could be obtained if enough crosses were made, but our observations indicate that teosinte crosses with Tripsacum less readily than does maize.

Chromosome counts in the cells of young seeds of teosinte x Tripsacum show that both embryo and endosperm are hybrids and neither are due to induced parthenogenesis.

In the light of our experiments it is difficult to account for the peculiar results of Collins and Kempton.4 It is surprising that they obtained seeds at all since we have been able to do so only by the use of a tedious technique and by dealing with enormous numbers. It is even more surprising that all of their crosses resulted in matroclinous or patroclinous plants, presumably resulting from parthenogenesis, while all of the seeds that we have so far obtained have proved to be true hybrids.

Literature Cited

TABLE 1.—Results of pollinating different varieties of maize by Texas Tripsacum.

Variety No. pollinated No.
No. matured
per 10,000
Ears Silks
Horton 21 11,848 469 4.0 0 0
Su su 24 14,762 995 6.7 4 2.71
Wx wx 16 6.736 916 13.6 2 2.97
Yellow Creole 46 24,886* 5,622* 22.5 10 4.02
Thomas 43 20,606 2,017 9.8 9 4.37
Creole X Surcropper 47 22,556 6,715 29.8 10 4.43
Mexican June 41 18,580 5,343 28.8 9 4.84
Misc. stocks 82 39,184* 7,827* 26.8 23 5.87
Surcropper 62 25,766 4,650 18.0 17 6.60
Total 382 184,924 34,554 18.7 84 4.54
*Estimates based on sample.

TABLE II.—Weights of hybrid seeds on ears also bearing maize seeds, compared to hybrid seeds occurring alone.

Variety Weight in mg. Gain or loss
in per cent
With maize Alone
Surcropper 71.9 43.4 65.7
Hastings' Prolific 50.0 19.8 152.5
Thomas 47.3 26.1 81.2
Mexican June 46.6 31.6 47.5
Creole X Surcropper 36.5 42.0 -13.1
Average 50.9 36.4 39.8

TABLE III.—Normal weights of maize seeds compared to weights of crossed seeds on ears of same variety.

Variety Weight in mg. Per cent
Maize Hybrids
Queen's Golden Pop 130 35 26.9
Waxy 204 40 19.6
Thomas 262 49 18.7
Yellow Creole 262 30 11.5
Hastings' Prolific 264 34 12.9
Mexican June 341 20 5.9
Surcropper 411 57 13.9
Average 268 38 14.2