least in some quarters, to a perceived need for more formal systematics. In 1934, Winifred and Harold Gladwin published a short monograph entitled A Method for the Designation of Cultures and their Variations (Gladwin and Gladwin 1934). Their approach was strictly chronological and culture-historical, and it was based on an analogy with biological classification. The metaphor they used, however, was that of describing a tree. In the beginning, according to the Gladwin scheme, there were a few widespread and generalized regional roots. These in the course of time had given rise to stems, stems had subdivided into branches, and branches had subdivided into phases. In the U.S. Southwest (where the scheme was developed) there was an original southwestern root, which gave rise to stems that we now call Anasazi, Hohokam, and Mogollon (the Gladwins used different terms); the Anasazi stem in its turn had in time split into the Chaco, Mesa Verde, and Kayenta branches, and the Kayenta branch (to take one example) had split into Tsegi, Jeddito, and various other phases.

At nearly the same time, archaeologists in the Middle West were developing a fundamentally different classificatory system that ignored history altogether and was based purely on typological resemblances between artifacts and artifact groups. Formally designated the Midwestern Taxonomic Method, it came to be known popularly as the McKern System because it was first described in print by W.C. McKern (McKern 1939), although it had actually been formulated at a series of archaeological conferences several years previously.

The Gladwins’ system was essentially a splitting system, but the Midwestern System was conceived by its authors as a lumping system. The latter began at the lowest level by recognizing foci, which were made up of groups of sites in a localized area, that shared a very large number of traits in common. Foci were grouped into aspects, which shared some but not as many traits in common; aspects were in turn lumped into phases, and finally phases were grouped into patterns, representing the highest and most generalized level in the system. The original scheme comprehended only two patterns for the whole eastern United States, the Woodland and the Mississippian. The Midwestern System was devised to a considerable extent for the study and classification of museum and private collections, most of which were poorly dated, and for that reason it did not have a specifically chronological dimension. Phases, aspects, and foci might be either temporal or geographical variants of the parent pattern.

The Gladwin classificatory system involved too much speculation about historical connections to be congenial to most archaeologists, and it was never adopted, except by the Gladwins and their close associates. The McKern System, on the other hand, has been widely, though not very systematically, employed in many parts of North America besides the Midwest, but it has never been accepted as providing a fully satisfactory overall schema for the classification of North American prehistory. Like nearly all classificatory devices, it has been found to work better in some places and at some periods than at others.

Meanwhile, the discovery of so-called early man (now usually called Paleo-Indian) remains in the 1930s added a new and unexpected chronological dimension to American prehistory. Paleolithic-type remains were found that must date back at least several thousand years, and they could not be definitely related to the later Indian cultures. In classifying these early remains, American prehistorians followed much more closely the model of de Mortillet than their own methodology in dealing with the later culture. That is, the various Paleo-Indian cultures, like Clovis and Folsom, were treated strictly as chronological subdivisions in a single linear progression. This approach seems to be supported both by distributional evidence and by subsequent radiocarbon dating, but it probably owes something to the influence of the Old World Paleolithic canon as well.

In 1958, gordon willey and Philip Phillips proposed a comprehensive chronological schema for all of the native cultures of the New World (Willey and Phillips 1958) in which the prehistoric cultures were assigned to five developmental stages: lithic, archaic, formative, classic, and postclassic. This involved no formal systematics; it was a simple evolutionary chronology somewhat