THE rodents are the most numerous order of mammals, both in terms of genera and species, and of individuals, including over half of the described living species of mammals. They are much less well represented among the known fossils, but this is, to a very large extent, the result of their small size and hence lesser conspicuousness, rather than of a former lesser diversification, at least since the latter part of the Eocene. The basic characteristic of the order-the development of gnawing incisors-has restricted the number of ecologic niches that were open to rodents. Since rodents have become very successful in this limited direction, many lines have developed, tending to evolve in the same general direction, due to similar selective pressures acting on a similar genotype. This has produced a great deal of parallelism and convergence within the order, as is now acknowledged by all students of the group. Due to this parallelism and convergence, one is faced with the problem of differentiating between similarities indicative of close relationships and those that are not. Hence, there has been, and still is, considerable difference of opinion as to the proper classification of the order. The classical arrangement, involving a subdivision into the Sciuromorpha, Myomorpha and Hystricomorpha, fundamentally on the basis of the structure of the masseter muscle, dates from Brandt (1855), although similar but less formal arrangements had been proposed before. Simpson (1945) is the latest major review of classification to adopt this three-fold arrangement. Although this has been the standard classification, many investigators felt (as, indeed, did Simpson) that its use involved forcing some rodents into unnatural relationships for purposes of convenience. Therefore, greater or lesser variants were proposed (Tullberg, 1899; Miller and Gidley, 1918; Winge, 1924). Tullberg's scheme divided the rodents into two major groups, based on the structure of the angle of the jaw, and then divided each of these on the basis of the masseter. For his secondary divisions, he used Brandt's three groups, erecting a fourth group for the South African Bathyergidae. Neither Winge nor Miller and Gidley use Brandt's terminology, but, essentially, accepted his groups and added a fourth for the forms which did not fit well into this scheme. These fourth, or wastebasket, groups have never met with wide acceptance. Even in cases where the suborders seemed morphologically valid, parts of the Brandtian classification seemed unreasonable to some authors. For example, the Hystricomorpha included certain Old World (particularly African) families, and a large array of South American ones, which were assumed to be of ultimate Old World origin. For many years, this was interpreted as indicating a South Atlantic land bridge. With increasing knowledge of the character and distribution of fossil mammals, no such land bridge seemed possible as an invasion route for the South American rodents (Simpson, 1950), and the problem of the origin of the South American forms became a major stumbling block in accepting the unity of the Hystricomorpha. Wood (1950), in view of the known extensive parallelism among rodents, suggested that the similarities between the New and Old World forms could best be explained as parallelisms, and that the source of the South American forms should be sought in North America. Within the past few years, there has
Read full abstract