Maize Evolution from 500 B. C. to the Present

Abstract

Variation in modern maize relates to recent and ancient cultural events. Large complexes of maize races are usually characteristic of certain cultural regions and help to define boundaries and histories. Most of the regions were established by 2000 years ago; clear differentiation of maize must have also occurred by then. Twenty-four racial complexes, numerous sub-complexes, and six regions have been defined here and elsewhere. Most of the complexes seem descended from seven ancestral types plus continual teosinte introgression. Spreading and interaction of these ancient types has had as important consequences as their within regions. Testing of the component hypotheses will be done through continued interdisciplinary study. SINCE THE EVOLUTION OF CULTIVARS is linked more to cultural factors than to natural ones, anthropological information is necessary to explain maize evolution. The formation of cultivar, like that of species, is a step-by-step process, through the accumulation of differences caused by mutation, recombination, selection and isolation (Davis and Heywood 1963:420); because each step is affected by man's activities and desires, many close parallels can be found between variation in maize and differences among and within cultures and their histories. Most of the genetic variation which was to be molded by numerous cultures into the present extraordinary array of races, ecotypes, and forms had been blended into maize by 500 B.C., date chosen just to predate the great spread of Intermediate Area maize races and cultural traits in Mesoamerica and South America (below). An explanatory model is being sought which will at once describe the present cultivars (races), any archaeological types, and the processes which led to them from the ancestral populations. Basic to this approach is multi-faceted systematic description reflecting observations made in many fields. In studying modern maize, so many races have been described using widely differing sets of variables (Goodman and Bird 1977, Hernandez X. 1973, Kato 1976, and the many race booklets listed therein), that it is often hard to perceive broad patterns. The voluminous data can be reanalyzed using computerized multivariate analyses; as new data are obtained and added to the model, the patterns will change, but the changes will diminish with each addition. As broad range of evidence as possible must be evaluated and used for forming hypotheses (Harlan and de Wet 1973, Hawkes 1970). Many contradictory hypotheses about maize evolution have been proposed in the past 50 years. Resolving the contradictions and adding more detail could greatly help both maize breeding and the reconstruction of New World prehistory. The early (pre-500 B.C.) evolution of maize and its relationships to other species has been reviewed recently using evidence from cytogenetics, archaeobotany, and archaeology (de Wet et al. 1971; Galinat 1974, 1977; Goodman 1976; Mangelsdorf 1974; Randolph 1976; Wilkes 1977), and the interbreeding of species and races of tripsacum, teosinte, and maize has answered some questions about these relationships. Chromosome morphology of these taxa is conservative and very useful in comparisons. There have been many changes in theories because of the location or morphology of very early archaeological cobs. Land shifts and diffusions of archaeological traits are accompaned by new maize types. In studying the last 2500 years of maize evolution, such evidence is also important. However, discussions of archaeological maize of this period are very limited in area and scope, and, with some exceptions (Cutler and Agogino 1960; Galinat, Reinhart and Frisbie 1970), it is difficult to relate archaeological maize specimens to one another or to modern races using published data. Other studies are possible-comparing maize geography to cultural geography, determining interregional influences, and describing hypothetical ancestral types and their in-