Crossability of corn and Tripsacum and the evolutionary history of the American Maydeae.


The cross incompatibility of Mexican and Guatemalan Tripsacum and corn has been demonstrated by three years of field experiments in regions of these countries where both Tripsacum and teosinte occur. The failure to obtain hybrid progeny from crosses involving the native corn of these regions supports the theory that corn, Tripsacum and teosinte originated from a common ancestor a long time ago. These experiments invalidate the alternative theory of Mangelsdorf and Reeves that important characteristics of modern varieties of corn originated in recent times from the natural hybridization of Tripsacum and corn in Central America and that teosinte is a product of such hybridization.


The very great diversity of existing types of corn, teosinte and Tripsacum in southern Mexico suggests that these plants have had a very long evolutionary history in this region and that the center of origin of corn and its near relatives was here rather than elsewhere in Central or South America. This simple interpretation is fully in accord with the Vavilov hypothesis that the centor of origin of cultivated plants is to be sought in the regions of greatest diversity of the existing cultivated forms and their wild relatives.


The first hybrids of maize and Tripsacum were obtained by Mangelsdorf and Reeves in 1929. They crossed both the diploid and the tetraploid forms of T. dactyloides with diploid corn. These crosses, which they made in Texas presumably with southern varieties of corn, have been repeated at Ithaca.


In using vigorous commercial corn hybrids, chiefly from the northern corn belt, little difficulty was experienced in obtaining large numbers of hybrids. The cross between 2n corn and 4n T. dactyloides ordinarily was much easier to make than the 2n corn x 2n T. dactyloides cross; from some of the hybrid ears of the former cross seedlings were obtained with the embryo culture technique from more than 50 per cent of the corn ovules to which Tripsacum pollen was applied. Tetraploid corn crosses readily with tetraploid T. dactyloides but the hybrids obtained thus far were highly sterile.


Repeated attempts to obtain hybrids between the tetraploid Tripsacums of Guatemala and Mexico and the native diploid corn of the regions where the Tripsacum occurs have failed completely. The Guatemala trials were reported in an earlier News Letter, over 200 corn ear shoots having been pollinated with tetraploid Tripsacum pollen without obtaining any hybrids. Similar experiments conducted in Mexico with 50 or more ear shoots being pollinated at each of three different locations in 1947 failed to yield any hybrids. In 1948 additional tests were conducted in Mexico at the Jolostoc field station of the Rockefeller Foundation where Dr. E. J. Wellhausen very kindly made available stocks of 6 different types of Mexican corn. These were crossed with pollen of the Acahuizotla diploid Tripsacum, and a few crosses also were made with pollen of a broad‑leafed 4n Tripsacum of the T. pilosum type. A total of 95 ears were pollinated, sufficient to test from 25,000 to 30,000 corn gametes for cross compatibility. Unfortunately, a very severe infestation of ear worms destroyed many of the ears completely and limited the test to approximately 2,000 gametes. No hybrids were obtained from any of the crosses made at Jolostoc in 1948.


Admittedly these tests of crossability of Tripsacum and corn in Mexico and Guatemala are not sufficiently extensive to warrant the conclusion that hybrids cannot be obtained from such crosses. Certain favorable combinations of genotypes not included in these tests might be compatible. But it can be concluded at present that there is a high degree of cross incompatibility among the combinations tested, and that Mexican and Guatemalan Tripsacum and corn cannot be crossed as readily as can the T. dactyloides and corn of the United States. Mangelsdorf and Reeves based their theory of the origin of corn on the crosses with T. dactyloides in Texas and did not subject their hypotheses to direct experimental verification in the region where natural hybridization of Tripsacum and primitive corn was assumed to have occurred.


The crossing experiments which I have conducted in Mexico during the past two years and the cytological and field studies of the various forms of Tripsacum and teosinte native in Mexico and Guatemala have convinced me that corn and Tripsacum and teosinte have existed as separate entities for long periods of time in these countries.


Tetraploid Tripsacum occurs over wide areas in both Mexico and Guatemala and is extremely variable. Species limits are poorly defined in many localities. There is clear evidence of hybridization and recombination of characters which have been used in defining the species. Intermediates between the broad‑leafed T. pilosum type and the T. lanceolatum type with narrow leaves are very prevalent.


Cytological examination of the tetraploids revealed the presence of multivalent association of chromosomes, indicating that they were of relatively recent origin and suggesting that the diploid ancestral species might still exist. It was postulated that the variability of the existing tetraploids was due to hybridization of diploid species similar to T. lanceolatum and T. pilosum, followed by chromosome doubling and segregation in the new tetraploid populations as they spread rapidly into the areas where they are found at the present time. Or, independent chromosome doubling in each of these species followed by hybridization of the autoploids derivatives could have accomplished the same result.


The problem then was to find the ancestral diploids. They were looked for, and both were found within 60 kilometers of each other in southern Mexico. As described elsewhere, the diploid of the T. pilosum type is a very tall, broad‑leafed plant with a thick stalk and many spikes in the terminal inflorescence. The diploid of the T. lanceolatum type is a low grass‑like plant, very leafy at the base and having a relatively short, slender stalk with only one or two spikes in the terminal inflorescence.


These new 36‑chromosome forms are referred to as diploids although it is conceivable that they are themselves tetraploids of remote ancestral species having 18 or 20 chromosomes like Manisuris or corn. Sporocytes of the Acahuizotla diploid have been examined and 18 pairs of chromosomes were seen. There was no evidence of multivalents, and until evidence either of their autoploid or alloploid nature is forthcoming they will continue to be referred to as diploids. However, it should be pointed out in this connection that longstanding differences in chromosome number between the progenitors of maize and Tripsacum must have prevailed and must have constituted a formidable barrier to the establishment in nature of hybrids between them.


The evidence from the known distribution of diploid and tetraploid forms of Tripsacum as well as the cytological evidence indicates that tetraploidy in the genus became established in relatively recent times. Diploidy is found at the limits of the geographic range of the species: to the southward in Brazil, where the diploid T. austale occurs and northward in the United States where the diploid form of T. dactyloides occurs as far north as Kansas. T. floridanum of southern Florida is also diploid. The tetraploid form of T. dactyloides in the United States may have originated as an autoploid independently of the alloploids of Mexico and Guatemala. Further study of the Tripsacum of northern Mexico is needed to clarify this situation.


Comparison of the existing forms of diploid Tripsacun, re­veals two distinct evolutionary trends. At one extreme are the grass­-like types of T. lanceolatum and allied species, of which the new diploid from Ca–ada del Zopilote represents the highest degree of specialization in this direction. These types have a slender stem, narrow leaves and few tassel branches. They are adapted to survival in dry arid climates as well as in moist situations. Such forms thrive on steep, rocky banks in full sunlight as well as on moist, shaded, rocky ledges and occasionally in roadside ditches. At the other extreme are the more corn‑like forns of the T. pilosum type which have thick leafy stalks, very broad leaves and are adapted to moist rich soil. The

Acahuizotla diploid is of this type but there are tetraploids of very similar appearance and growth habit.


It is idle to speculate about the derivation of these forms at present; but it is obvious that fewer mutational steps would have been required for the evolution of corn from a prototype resembling the latter than from a prototype resembling the former species of Tripsacum.


In this connection it is perhaps significant that existing types of native corn in Mexico and Guatemla exhibit morphological variations similar to those described above for Tripsacum. There are narrow‑leafed types with slender stalks and few tassel branches. There are also very broad‑leafed types with thick stalks and many tassel branches, as well as innumerable combinations of these and other traits which are sometimes referred to as "tripsacoid". Resemblances of this sort indicate that Tripsacum and corn have many genes in common. This is to be expected in species so closely related that they can be hybridized. In fact it has been estimated that 85 per cent or more of the genes in cross compatible species are identical. In view of the cross incompatibility of existing forms of Mexican Tripsacum and corn, it seems highly improbable that the characteristics, which they have in common, were due to hybridization in the recent past. Most of these similarities probably were due to inheritance from a common ancestor. A limited number may have arisen more recently by independent mutation.


The persistence of the two extreme types of diploid Tripsacum within a short distance of each other in southern Mexico in a region where teosinte and tetraploid Tripsacum are abundant over wide areas, suggests that this general region might have been the primary center in which the gradual differentiation of Tripsacum, Euchlaena and Zea from a common, but very remote, ancestral form took place. In this connection it may be significant that, at the present time, this general region is near the center of the geographic range of both Tripsacum and Euchlaena. Furthermore, it is generally believed that the cultivation of corn originated among the prehistoric Indian civilizations that migrated into and through this area over a period of many centuries.


If Zea, Tripsacum and Euchlaena originated from a common ancestor in southern Mexico or elsewhere in this general region, the cross incompatibility of Tripsaam and corn in this area at the present time is understandable on the assumption that genic cross incompatibility was the isolating mechanism responsible for their delimitation and segregation. As differentiation proceeded and Tripsacum spread northward into the United States, geographical, ecological, or other isolating mechanisms may have operated to maintain the autonomy of new species such as T. dactyloides. Genic incompatibility might have disappeared or was never present in these northern populations. This is one possible explanation of the fact that dactyloides crosses readily with corn under certain conditions and the Tripsacum of southern Mexico and Guatemala does not cross readily when similar techniques are employed.


It is questiomble whether natural hybridization of Tripsacum and Zea in recent times would have resulted in the origination of new species or a significant admixture of Tripsacum and corn germplasm. The experimental hybrids of Zea and Tripsacum are functionally male sterile and only occasional unreduced female gametes have been known to function in backcrosses to either parent. Thus the first generation hybrids cannot produce segregating populations. The plants of the first backcross to corn have 38 chromosomes including two sets of corn chromsomes and one set of 18 Tripsacum chromosomes; in these plants the Tripsacum chromsomes assort at random as univalents in meiosis and tend to disappear rapidly in later generations due to elimination at meiosis and selection against gametes carrying extra chromosones. If the Tripsacum chromsomes are unable to substitute for corn chromsomes or exchange segments with them in the formation of functional gametes, their association with the corn chromsomes in these few hybrid generations before they are lost is devoid of evolutionary significance.


Corn plants lacking Tripsacum chromosomes have been recovered in the second generation backcross progenies, which appear to be phenotypically pure corn, and no evidence of segments of Tripsacum chromsomes has yet appeared in the cytological examination of these and other plants in the same cultures.


A study of the cytology and genetics of each of the 18 Tripsacum chromosomes comprising the haploid set in combination with two sets of corn chromosomes is now in progress. These "trisomics" should be favorable material for a critical analysis of the degree of homology existing among Tripsacum and corn chromosomes.


L. P. Randolph