American Anthropologist 47(2): 235-243. (1945)
THE ORIGIN OF MAIZE: PRESENT STATUS OF THE PROBLEM
P. C. Mangelsdorf and R. G. Reeves

1 Whiting, 1944. (See bibliography at end of article.)
2 Although Zea Mays is universally known as "corn" in the United States, the term "maize" or "Indian corn" is usually preferred in international usage since the word "corn" in many countries is virtually synonymous with "grain."

IN A recent paper in this journal, Whiting,1 concluding that anthropologists have failed to understand the botanical evidence which bears upon the origin of maize,2 has undertaken to interpret and evaluate for their benefit the essential facts and theories involved. The results are not wholly fortunate. If there was confusion before, it is now, we fear, confounded.

It is with some reluctance that we take issue with Whiting's conclusions, for we fully recognize the right of any student to his own interpretations and evaluations. We are convinced, too, that time and space devoted to research and the publication of research are, by and large, more fruitful than time and space utilized in controversial discussions. Furthermore, we are aware that discussions on the origin of maize have on occasion in the past become acrimonious and have solved nothing. Such an outcome we earnestly desire to avoid. Nevertheless when there appears to be, as in this case, a misconception of the real nature of the problem, combined with an obvious misinterpretation of some of the facts and an omission of some of the pertinent literature, we feel that the conclusions can no longer go unchallenged.

Whiting apparently feels that the problem of the origin of maize is insoluble by botanical methods; at least he chides the anthropologists for continuing to believe that it is soluble by botanical analysis. He does not suggest how the problem is to be solved, beyond stating that "Any solution must rest upon a series of generalizations of a very high order of abstraction." With this conception of the problem we disagree perhaps more strongly than we do with his evaluation and interpretations.

It is true that the problem of the origin of maize, as of most cultivated plants, is exceedingly complex, and it is obvious that all of the evidence, botanical, archaeological, linguistic and historical, which can be brought to bear upon it should be utilized as completely as possible. But we cannot admit that the problem is insoluble by botanical methods, if we include among these methods those of genetics, cytology and physiology as well as those of comparative morphology. We venture to predict that the solution of the problem, if it is solved, will rest less upon abstract speculations and generalizations than upon new evidence slowly and systematically accumulated by acute observation and painstaking experimentation, accompanied by intelligent and testable working hypotheses which provide new avenues of attack. Such an approach to the problem Whiting apparently does not understand, or he dismisses as unpromising.

3 Mangelsdorf and Reeves, 1939.
4 Beadle, 1939.
5 Langham, 1940.
6 Weatherwax, 1942.

To come now to more specific points of difference: Whiting's statement that the theory that corn is a domesticated form of teosinte has long since been disproved is much more dogmatic than the circumstances warrant. It is true that the teosinte hypothesis has never been completely satisfactory. With the recent evidence3 that teosinte is a hybrid of maize and Tripsacum and hence the progeny rather than the progenitor of maize (evidence which Whiting apparently does not accept) the hypothesis has become even less satisfactory. But to say that it was "long since" disproved is straying from the facts. We did not regard it as disproved when we began our genetic and cytological investigations of maize-teosinte hybrids some fifteen years ago; many competent authorities on maize do not regard it as disproved today. Beadle4 and Langham5 (in papers which Whiting does not cite) have put forward new evidence in support of the teosinte theory. Weatherwax,6 whom Whiting credits with having contributed more sound thinking to this problem than any other writer, has recently (in another paper not cited by Whiting) pointed out that several objections to the teosinte hypothesis are less formidable than they were a decade ago. In view of these facts it can scarcely be truly said that the teosinte theory has long since been disproved.

In arriving at his conclusions regarding teosinte Whiting states that the proposition that interfertility is a better criterion of relationship than comparative morphology has now been "definitely discarded as obviously invalid." It is true that interfertility as a criterion of relationship does have serious shortcomings but this can also be said about comparative morphology. Certainly it is folly to rely completely upon any single approach to the problem of natural relationships, and especially is this true when the problem is as baffling as it is in the case of maize and teosinte. Recognizing this fact we have made a serious effort to focus all pertinent evidence upon its solution. It is disturbing, to say the least, to have this evidence casually brushed aside as of no importance.

It may be appropriate at this point to take account of some of the factors which have apparently led Whiting to reject the hypothesis that teosinte is a hybrid of maize and Tripsacum. We realize that a discussion of this matter involves technical botanical details in which anthropologists must, at best, he only moderately interested. Nevertheless it would be unfair to make the accusation of misinterpretation of facts without mentioning the evidence upon which the accusation is based.

7 Mangelsdorf and Reeves, 1939, pp. 203 and 306.

Whiting regards as an inconsistency in the hypothesis of Mangelsdorf and Reeves the conclusions (a) that teosinte differs from maize primarily by four segments of chromatin and (b) that the chromosome knobs of teosinte and maize are ultimately derived from Tripsacum. The inconsistency exists only if our statements are not considered in full, for we have repeatedly stated7 that other smaller segments, less readily detected by genetic techniques, are probably also involved. More recent experiments indicate that this is indeed the case. In some of the more Tripsacoid forms of teosinte, the so-called "pure" teosintes, these smaller segments may, in the aggregate, account for a considerable proportion of the genetic differences between teosinte and maize. This evidence also agrees with the fact that the more Tripsacoid forms of teosinte have more numerous chromosome knobs than the maize-like forms. The precise mechanism by which knobs are transferred first from Tripsacum to teosinte, subsequently from teosinte to maize, is admittedly still obscure. There is no doubt, however, of the strong, though by no means complete, correlation between knob number and various external morphological characteristics which some varieties of maize have in common with Tripsacum. Hence the suggestion that chromosome knobs of maize have been derived originally from Tripsacum has already proved to be a useful one, even if it should later, on the basis of additional evidence, become modified or be discarded. At present however, it is not inconsistent either with the facts or with our other statements.

Another apparent inconsistency to which Whiting calls attention is our conclusion that the cob of maize is toughened by Tripsacum influence in spite of the fact that Tripsacum is described by botanists as having a fragile rachis. This statement, perhaps more than any other in his paper, reveals the pitfalls which confront a writer who prefers abstractions to prosaic facts. It is true that Tripsacum has a fragile rachis in the technical botanical sense that the rachis segments disarticulate readily. But the rachis segments themselves, as well as the glumes and other structures are extremely tough or horny. In hybrids of maize and Tripsacum (as well as in maize-teosinte hybrids) many of the segregates exhibit the solid or non-fragile rachis of maize plus some of the horniness of rachis tissue characteristic of Tripsacum. Such a situation is not unusual in hybrids; indeed it is to be expected.

8 Mangelsdorf and Cameron, 1942.
9 Mangelsdorf and Reeves, 1939, pp. 160-170.

A third inconsistency which Whiting mentions is the statement of Mangelsdorf and Cameron8 that a high knob number is associated with (among other things) long ears, even rows of grains and strongly flattened kernels; yet the maize of the northern Plains Indians which exhibits these characters has low chromosome knob numbers. Whiting's discussion on this point is somewhat confused by the fact that he has inadvertently used the word low instead of high in citing the statement of Mangelsdorf and Cameron, but the nature of the inconsistency to which he refers is reasonably clear. The answer to his contention is that although the maize of the northern Plains Indians does resemble the many-knobbed varieties of Guatemala in having straight rows and flattened kernels it differs from them quite decidedly in many other respects, in which one of the most important is the type of endosperm. The Plains varieties are predominantly flint and flour corns. The many-knobbed maize varieties of Guatemala are predominantly dent. The differences between dent corns and flint-flour corns are genetically complex. If the straight rows and flattened kernels of the Plains maize are the product of Tripsacum admixture it certainly represents much less admixture than the Guatemalan dents have received. We have shown by actual experiments9 that segments of Tripsacum chromatin lacking knobs and almost too small to be detected cytologically can be transferred from Tripsacum to maize to produce visible effects upon the external morphology of the plant.

10 Cf. Mangelsdorf and Cameron, 1942, p. 217.

Whiting speaks of the "elaborate series of hypotheses" presented by Mangelsdorf and Reeves. We have regarded our hypothesis as a single tripartite hypothesis, the three parts of which, although to some extent independent of each other, combine to present an integrated and plausible picture of the origin of maize.10 If the hypothesis, or series of hypotheses, is elaborate it is only because the problem is, as Whiting has himself emphasized, extremely complex. Complex problems, whether mathematical or botanical, seldom have simple solutions. Actually we shall probably find, when all the facts are in, that our hypothesis, or any contemporary hypothesis, greatly over-simplifies the true situation. In any case the important consideration is not whether the hypothesis is elaborate but whether it accounts for the available facts, is testable, and is useful in providing new points of attack.

11 Cf. Weatherwax, 1942, p. 16.

More seriously in error, however, is Whiting's statement that our hypotheses rest upon the "basic assumption" that a wild corn without knobs on its chromosomes was native to the lowlands of Paraguay and Bolivia. If there is a "basic" assumption which underlies our hypotheses it is, as we have attempted to show, and as others have had no difficulty in recognizing,11 that teosinte is a hybrid of maize and Tripsacum. This conclusion obviously tells us nothing about the origin of maize but, if accepted, it clears the way for a more critical evaluation of the remaining evidence. Neither botanists nor anthropologists have been willing, until recently at least, to ignore the importance of teosinte as a putative maize ancestor. If, however, on the basis of recent evidence we can remove teosinte from consideration except as a source of admixture with maize the remaining evidence points strongly to South America as the center of origin of maize.

To the anthropologist the details of these hypotheses and the steps by which they were derived are perhaps of less interest than the concepts to which they lead. The hypotheses of Harshberger, Collins, Kempton, Weatherwax and Longley have all visualized maize as arising in Mexico or Central America, although several of these writers have admitted the possibility of a secondary center of origin in South America. The conclusions of Mangelsdorf and Reeves, in sharp contrast, point strongly to South America as the primary center of origin. Which of these conflicting concepts is more nearly correct only future evidence can finally decide.

Whiting concludes his evaluation of facts and theories with a table of probabilities purporting to show the "betting odds" that wild maize, if it is ever discovered, will be found in any one of ten possible areas which he considers. Perhaps this table possesses some merit as an original, simple and easily understandable method of presenting the writer's conclusions to his readers. Perhaps it also reflects Whiting's apparent conception of the problem of the origin of maize as a glorified guessing game. In either case we do not agree with the probabilities which he presents. But since it is a mathematical axiom that a series of guesses is still no better than a single guess, there is little to be gained by revising the table. Perhaps the greatest value of the table lies in its implications; it inevitably suggests when one carries the analogy a bit further that those who bet on races, except as an amusement, should know their horses.

Since we have repeatedly emphasized in this paper the importance of working hypotheses which are capable of being tested perhaps we may utilize this opportunity to make a brief and preliminary report upon some of the tests to which our hypotheses have been subjected and the new evidence (for the most part still unpublished) which has accrued from these tests.

To determine whether wild corn still exists in the lowlands of South America, Dr. Hugh C. Cutler, an experienced botanical explorer has, at our instigation, travelled extensively in Brazil, Paraguay and Bolivia since 1941 in search of wild or primitive maize. He has not discovered wild maize. He has, however, collected in southwestern Brazil, Paraguay and eastern Bolivia a group of varieties with unusual and primitive characteristics. One of these, a variety obtained from the Guarany Indians of Paraguay, but one which is rather widely distributed in the Matto Grosso and other parts of the lowlands, is the most extraordinary type of maize which we have ever studied. It possesses almost all of the known dominant genes of maize, including several which had previously been encountered only rarely. It has a low chromosome knob number and the few knobs which it possesses are small. The rachis or cob is slender and flexible; the spikelets are easily removed from it intact. The ears, in some environments, exhibit a partially indeterminate habit of growth. Once the ear grows beyond the end of the shucks, as it frequently does, it may elongate far beyond its normal length to produce a lax spike in which the true structure of the maize ear is revealed more clearly than has ever been done by anatomical studies. Cutler's collections also show that pod corn which previously has been reported as being unknown in Peru is sufficiently common in Bolivia to be recognized by the natives who call it paca sara or "hidden maize." His collections also show extraordinary diversity on the eastern slopes of the Andes in Bolivia from Sacaba and Cliza on the south to Chulumani and Coroico on the north. Two large tributaries of the Amazon, the Rio Beni and the Rio Mamoré, have their headwaters in this region, and it now appears that the basins of these rivers and the land which lies between may be a promising region in which to search for wild maize. Anthropologists familiar with South America will recognize this region as one of the most difficult in America to explore. Nevertheless Dr. Cutler plans to resume his explorations when he is released from the rubber work in which he is engaged for the duration of the war. In the meantime, however, the hypothesis that maize once had its wild habitat in the lowlands of South America has received considerable indirect support, and the hypothesis has been very fruitful in uncovering new and previously unsuspected types of maize.

12 The question has frequently been asked: If teosinte is a hybrid of maize and Tripsacum why should it not be possible to synthesize it by crossing maize and Tripsacum? The answer is that in species hybrids, as in games of chance, the odds against the fortuitous repetition of a particular very complex combination are almost overwhelming. Among the segregates of our hybrids of maize and teosinte we have found a number which approached teosinte in several but not in all characteristics. Perhaps a still closer approach to teosinte might be made by crossing Guatemalan species of Tripsacum with non-Tripsacoid Guatemalan maize. Unfortunately both are difficult to grow in our latitude.
13 Reeves and Mangelsdorf, 1942.

The hypothesis that teosinte is a hybrid of maize and Tripsacum has been further tested by transferring the principal segments of chromatin, which differentiate the two species, by repeated hackcrossing, from teosinte to a uniform inbred strain of maize so that the effects of each segment may be studied separately and the segments precisely located upon the chromosomes. This work is still far from completion; genetic studies of this nature require many years to consummate. But it has already become increasingly clear, almost certain in fact, that teosinte is an anomalous species. It is apparently nothing more than maize with segments of chromatin of another species, presumably Tripsacum, incorporated into its germ plasm.12 It is, in fact, so strongly maize-like in its genic and chromosomal composition, and it intergrades so strongly with maize that Reeves and Mangelsdorf13 have proposed that it be no longer classified as a separate genus, Euchlaena, but as two species of Zea. More recent data indicate that this change in classification is scarcely drastic enough; genetically, teosinte is really little more than a variety of maize.

Unpublished work of Reeves and Mangelsdorf also shows that in 54 of 55 morphological characters which were studied in maize, teosinte, and Tripsacum, teosinte either resembles one of the other plants or is intermediate between them. This study indicates that teosinte has but few morphological characters that it could not have inherited from one of its putative parents, maize and Tripsacum, and there is reason to suppose that these few characters are the result of recent mutations. In fact, the one exceptional character mentioned above can be explained even without the assumption of recent mutations. Such a situation is difficult to interpret without recourse to the hypothesis that teosinte originated as a hybrid between maize and Tripsacum, although it does not necessarily invalidate other hypotheses.

The third hypothesis that new types of maize originating directly or indirectly from the hybridization of maize and Tripsacum and exhibiting admixture with Tripsacum comprise the majority of Central and North American varieties has received considerable additional support from the studies, already referred to, of Mangelsdorf and Cameron on the maize of western Guatemala. These studies showed that a small area in the Department of Huehuetenango, an area in which Tripsacum grows in profusion and teosinte occurs as the dominant species on thousands of acres, embraces a greater diversity of maize than occurs in the entire United States. Here are found not only the most Tripsacoid varieties in this hemisphere but also remnants of the "pure" maize presumably once involved in the hybridization with Tripsacum. Here is the probable center of diversity of the Tripsacum-contaminated varieties, and in this area, significantly, chromosome knob-number is strongly correlated with the morphology of the plant. Reeves14 in another study has shown that Tripsacoid varieties are the predominating type also in Mexico and in countries of northern and eastern South America: Colombia, Venezuela, Dutch Guiana, Brazil and Uruguay. Mangelsdorf and Cameron (unpublished) find that this is true also of Panama, Nicaragua, Costa Rica and Cuba. Maize varieties of Paraguay like those of Peru, Bolivia and Ecuador have low knob numbers.

14 Reeves, 1944.

Reeves14 has also demonstrated a statistically significant relationship between knob number and distance from Central America. This agrees with the conclusion that maize was distributed over parts of North Central and South America before the advent of teosinte in Central America. The relationship between knob numbers and distance from Central America becomes even stronger when the Andean maize varieties are omitted from the computations on the assumption that maize was first highly domesticated in that region and became so well established there that encroachment of Tripsacum-contaminated varieties from the North has been a negligible factor.

Supplementing these studies Mangelsdorf (unpublished) has investigated the interchange of genes between maize and teosinte in the Valley of Mexico where teosinte occurs as a weed in the maize fields. It has been found that the incidence of hybridization is high and that (except for certain regions of the chromosomes) there is a free movement of genes between maize and teosinte. There is, in other words, no doubt that maize is being modified by teosinte admixture today; there is every reason to suspect that this process has been going on for centuries.

15 Anderson, 1944.

Thus each part of the tripartite hypothesis on the origin of maize has received substantial new support since it was first proposed. This is not to say that the problem of the origin of maize has been solved; indeed problems of this nature are seldom susceptible to a complete and final solution. The most that can he expected is that certain important facts will become established beyond a reasonable doubt. Nor is it to say that these hypotheses will not be modified as new evidence appears. As a possible example we may mention the fact that Anderson15 on the basis of a close resemblance between a variety of maize collected from a locality on the Rio Loa in northern Chile and a Chinese variety has revived the suggestion that maize may have been introduced into America from the Orient. The fact that there are several relatives of maize in the Orient, that pop corn is used as a food in parts of the East Inches and that the cultivated cottons of America are unquestionably hybrids of Asiatic cottons and wild American species, lends some degree of botanical plausibility to the hypothesis. There is, to be sure, much evidence against it; nevertheless it merits further investigation. But should such an hypothesis prove to he valid it would probably still leave South America as the primary American center of maize diversification. The search for wild maize in South America would he abandoned but the quest for primitive types would continue unabated.

To the extent that the origin of maize has a hearing upon the origin of prehistoric cultures in America the present available evidence, in our opinion, flow points more strongly than it has ever before to western South America as the region in which American maize agriculture, and American cultures based upon maize agriculture, had their beginnings.

16 Cutler, 1944.
17 Brieger, 1943.

NOTE: Since the above was written two additional papers bearing upon the subject have come to hand. Dr. Cutler's observations on pod corn in Bolivia have been published16 with an interesting suggestion as to how this relict character may have been maintained after selection began to operate against it. And Dr. Cutler has furnished us with a translation of a paper published in Portuguese by Dr. F. G. Brieger who is engaged in genetic research on maize and its relatives in Brazil. Dr. Brieger's studies17 tend to support one hypothesis that maize originated in the lowlands of South America. He does not, however, agree with us that teosinte differs from maize by only a few chromatin segments.

BOTANICAL MUSEUM, HARVARD UNIVERSITY
CAMBRIDGE, MASSACHUSETTS

TEXAS AGRICULTURAL EXPERIMENT STATION
COLLEGE STATION, TEXAS

BIBLIOGRAPHY

ANDERSON, EDGAR
  1944 A Variety of Maize from the Rio Loa. Annals of the Missouri Botanical Garden 30: 469-476.
BEADLE, G. W.
  1939 Teosinte and the Origin of Maize. Journal of Heredity 30: 245-247.
BRIEGER, F. G.
  1943 Origen do Milho. Revista de Agricultura, 18: 409-418, Piricicaba, Sao Paulo, Brazil.
CUTLER, HUGH C.
  1944 Medicine Men and the Preservation of a Relict Gene in Maize. Journal of Heredity 35: 291-294.
LANGHAM, D. G.
  1940 The Inheritance of Intergeneric Differences in Zea-Euchlaena hybrids. Genetics 25: 88-107.
MANGELSDORF, P. C., AND CAMERON, J. W.
  1942 Western Guatemala, a Secondary Center of Origin of Cultivated Maize Varieties. Botanical Museum Leaflets, Harvard University, 10: 217-252.
AND REEVES, R. G.
  1939 The Origin of Indian Corn and Its Relatives. Texas Agricultural Experimental Station Bulletin 574.
REEVES, R. G.
  1944 Chromosome Knobs in Relation to the Origin of Maize. Genetics 29: 141-147.
AND MANGELSDORF, P. C.
  1942 A Proposed Taxonomic Change in the Tribe Maydeae (Family Gramineae). American Journal of Botany 29: 815-817.
WEATHERWAX, PAUL
  1942 The Indian as a Corn Breeder. Proceedings of the Indiana Academy of Science 51: 13-21.
WHITING, ALFRED F.
  1944 The Origin of Corn: An Evaluation of Fact and Theory. American Anthropologist, 46: 500-515.