One of the four species, P. setosus, subspecies elegans, was used by [Winge] (1941:80, 82) as representative of the genus Proechimys when he was estimating the relationships of that genus. Because Cercomys, with four crests in each of its cheekteeth, was, on other grounds, regarded by him (op. cit.:80) as "... the most primitive genus within the group.", and because he noted in P. s. elegans 4 crests in P4 and in some first molars, he concluded that Proechimys was "very closely related to Cercomys." His conclusion seems to be correct, but actually other species of Proechimys (subgenus Trinomys), for example, P. dimidiatus, have four or more crests in each cheektooth, and, therefore, may be considered as more closely related to Cercomys than is P. setosus. If a large number of crests indicates primitiveness, P. dimidiatus, always with four, is more primitive than any other species in the subgenus Trinomys. Also, the large skull, long hind foot, short tail and thin aristiforms of P. dimidiatus, in my opinion, are primitive characters.
SUBSPECIFIC VARIATION IN THE SUBGENUS
TRINOMYS
One of the species of Trinomys, Proechimys iheringi, is here subdivided into six subspecies which show a clinal variation. P. i. iheringi, in the southernmost part of the range of the species (Ilha de São Sebastião), has three counterfolds in the upper cheekteeth of almost every young specimen but one of these counterfolds, since it is small, very shallow, and disappears after little wear, is probably in the process of disappearance; all lower cheekteeth have two counterfolds or, rarely, m3 has only one. P. i. bonafidei is the next subspecies northward, where it was collected at 850 m altitude (Fazenda Boa Fé). This subspecies still has two counterfolds in all the upper cheekteeth; only 3 out of 16 specimens fail to have these counterfolds coalesced in one or more of the teeth. In the lower cheekteeth the coalescence is evident in 18 per cent of the specimens. P. i. gratiosus, from Floresta da Caixa Dagua (alt. 750 m), geographically is well removed from bonafidei (more than two degrees north), and no samples were obtained from the intervening area. It shows such great reduction in the counterfolds that the existence of intermediate populations is clearly suggested. Every upper cheektooth of this subspecies has the two counterfolds coalesced and in 40 per cent of the specimens M3 has only one counterfold; in the lower cheekteeth 60 per cent of the specimens have only one counterfold in m3. P. i. panema, occurring approximately 100 kilometers to the northward of P. i. gratiosus (lowland form), has one counterfold in M3 in only 20 per cent of the specimens but the lower third molar has only one counterfold in 80 per cent of the specimens. In P. i. denigratus, from about 3 degrees north of the range of P. i. panema, the reduction is proportionately greater: P4 now is the only upper cheektooth with two counterfolds in every specimen; all molars tend to have only one; p4 has also two counterfolds but all lower molars have only one.
The relative size of the tail also varies in a cline from south to north. Its length is approximately 87 per cent of the length of the head and body in P. iheringi; 88 per cent in bonafidei; 99 per cent in gratiosus; 100 per cent in panema; and 103 in denigratus.
One of the subspecies, P. i. paratus, however, seems to be completely out of the dental cline. It was collected in the near proximity of the type locality of P. i. gratiosus, at an elevation of 120 m lower. This subspecies has two counterfolds in all molariform teeth and only one of the two specimens known shows these counterfolds coalesced in P4 and M1. The sample, 2 specimens, is too small to be trustworthy; hence it is impossible satisfactorily to account for the break in the clinal variation. Conceivably two full species are involved, but I prefer at present to defer decision on this problem until such time as more evidence is accumulated.
P. setosus is poorly represented, both of the available skins being faded. Furthermore, no type locality is known for the subspecies P. s. setosus.
P. albispinus has only two known subspecies: P. a. albispinus, living in a region of higher humidity, is slightly the darker and has subapical zones of the setiforms on the sides Ochraceous-Tawny; P. a. sertonius, living in a much drier region, has the same subapical zone Ochraceous-Buff. The number of specimens of P. a. sertonius is so few that no gradient can be detected, even if one exists.