_Proechimys guyannensis_ appears to be more plastic than any other species. In size of animal, width of aristiforms, color and number of counterfolds in the cheekteeth, it shows marked response to variations in geographic conditions. _Proechimys longicaudatus_ is apparently less plastic; only the number of counterfolds shows marked variation. _Proechimys semispinosus_ varies much within its range. _Proechimys goeldii_ seems to be relatively uniform. _Proechimys canicollis_ shows relatively little variation throughout its range but probably is divisible into two or more subspecies. The primitive _Proechimys_ probably was large with a short tail, narrow aristiforms, strongly built skull, and five counterfolds in each molariform tooth. Primitiveness here is inferred from characters which now are of general occurrence in the whole group as opposed to those restricted in geographic occurrence. It is a curious fact that in this genus, populations from small islands are more primitive than populations on the mainland. Apparently a small population restricted to a small island tends to revert to the primitive type. The homozygous condition will tend toward a generalized genotype and the disappearance of secondary biotypes. _P. i. iheringi_ on the Island of São Sebastião averages larger, has thinner aristiforms, and a stronger skull than the same subspecies on the mainland, and the cheekteeth usually have two and three counterfolds. The same subspecies on the mainland has no more than two counterfolds. _Proechimys semispinosus gorgonae_ and _Proechimys semispinosus ignotus_, living on Gorgona and San José islands, respectively, are both characterized by large size, short tails, strong and conspicuously ridged skulls, and cheekteeth frequently with four and five counterfolds. On the mainland, closely related subspecies, like _P. s. panamensis_, _chiriquinus_ and _gularis_, far less frequently have four counterfolds in more than one or two teeth. More striking still is the population-sample of _gularis_ from the island of Llunchi, in the Rio Napo, eastern Ecuador. In it there is a higher ratio of cheekteeth with four counterfolds than there is in the samples from the banks of the river. The two insular forms, _P. s. gorgonae_ and _P. s. ignotus_, referred to as primitive in the discussion above, have wide aristiforms, which is contrary to what would be expected in a primitive _Proechimys_. Supposing, however, as actually seems to be the fact, that narrowness of the aristiforms depends on an increased number of genes, we deduce that the population from the mainland, that gave rise to the populations of the islands, did not have all of the genes necessary to make the aristiforms narrow. In fact the subspecies known on the mainland, near the aforementioned islands, have wide aristiforms.