Considered by itself the variation in the tooth pattern can lead to erroneous conclusions as to differentiation of species, because the number of folds on the occlusal face of a tooth and the depth of certain folds may be subject to great individual variation as shown by examination of more than one large series of specimens of the same kind, age and sex from a single locality. Also there are geographic gradients or clines, in number of folds. Nevertheless the variation in number of folds, when measured at sufficient intervals along a cline, may provide quantitative characters useful in differentiating subspecies. [Illustration: FIGS. 2-17. Second left upper molar of the two subgenera _Proechimys_ and _Trinomys_. All × 8. Anterior border of tooth is at the top of each figure (Nos. 9 and 17 excepted). Note especially that main fold is short in _Proechimys_ and long in _Trinomys_. FIGS. 2-9. _Proechimys_ (_Proechimys_) _semispinosus liminalis_, female, MN no. 6243, Rio Quichito. Fig. 2, unworn crown. Figs. 3-8, cross sections at 0.5 mm. intervals, showing changes in the main fold and counterfolds at increasing depths as the tooth was ground down. Fig. 8 is 3 mm. below surface shown in fig. 2. Fig. 9, posterior view with proximal end of the tooth open showing basal ends of folds. Later in life the proximal end closes and three roots are formed. FIGS. 10-17. _Proechimys_ (_Trinomys_) _iheringi denigratus_, female, SEPFA no. 17060, Mata do Ribeirão da Fortuna. Figs. 10-16 corresponding to figs. 2-8. Fig. 17, posterioventral view with proximal end of the tooth open and part of walls cut away, showing basal ends of folds. Later in life, as in _Proechimys_, the proximal end closes and three roots are formed. ]