(4) In the final stage of wear ([Fig. 4D]), the enamel fossette disappears as a result of continued attrition on the occlusal surface in the upper series. The fossette may vary somewhat in vertical depth in m1 and m2, but the amount of wear required for its effacement would be greater than in the upper teeth. Therefore, upon wear, the U-pattern would become characteristic of the final stage in M1 (and probably also M2), but the modified H-pattern described in [Fig. 4C] would prevail in m1 and m2. Perhaps, in extremely worn teeth, the labial fossette of m1 and m2 would disappear. If this advanced stage of effacement is obtained, then the two columns would be united across the entire surface of their protomeres from the center of the crown to its labial edge, and the occlusal pattern would be in the shape of a U.
The occlusal pattern, at least in M1 and M2, in the final stages of wear in Dikkomys resembles that in the subfamily Entoptychinae, but the U-pattern develops on only the first and probably the second molar in Dikkomys and not on all of the cheek teeth as it does in the entoptychines. Judging from the material that has been described, the U-pattern did not develop in the lower teeth of Dikkomys until the Hemingfordian (D. woodi), upper Rosebud, and specimens of D. matthewi from the earlier Arikareean, lower Harrison, suggest that the modified H-pattern, with secondary coalescence at the edge of the protomeres, persisted throughout life, without developing the U-pattern in the final stages of wear.
Essentially the same patterns of wear characterize the genus Pliosaccomys, except that the earlier stages were telescoped and the second stage was omitted while another (final) stage was added. The stages are reconstructed in sequence in figure 4, and all are based on preserved dentitions, as follows:
(1) The first phases of wear produced the pattern ([p. 4E] and I) described for Dikkomys in the previous account (Fig 4A).
(2) A small additional amount of wear produced the 2nd stage ([Fig. 4F and J]) characterized by a U-pattern, formed by union of the anterior and posterior columns at the edge of the protomeres of the first and second molars, both above and below, without first forming an H-shaped pattern. Union at the mid-points thus was omitted from the sequence of wear in these two teeth. In the premolars and third molars the primitive H-pattern did form, as in Dikkomys. The pattern of wear in the first two molars is the same as in the entoptychines of the early Miocene. The trend of evolution through which the Pliosaccomys lineage passed must have featured a progressively earlier union at the edge of the tooth until the lateral coalescence occurred simultaneously with the median union. At that stage, emphasis was shifted to the union at the edge of the tooth, and eventually the teeth failed to unite at their mid-points and the U-pattern developed directly. Therefore, the horizontally deep re-entrant fold that separates the two lophs of the U-pattern is equivalent to one fold plus the apex of the opposite fold.
(3) The horizontal re-entrant fold of the U-pattern was remarkably shallow vertically and disappeared with little additional wear. Thus the two parts of M1, and also of M2, are united into a single column except for a slight inflection on the labial side and this is true also of m1 and m2 except for a slight inflection on the lingual side ([Fig. 4G and K]). The inflection appears to have persisted in the upper teeth ([Fig. 4H]), but evidently with slight wear, disappeared in the lower teeth ([Fig. 4L]). The final monocolumnar pattern was attained early ontogenetically, evidently before the permanent premolar had fully erupted; hence, the earlier stages occurred only in transition, persisted for only a brief interval in the teeth of juveniles, and the final stage developed in the young animal and lasted throughout the rest of its life in Pliosaccomys. In Dikkomys the two columns never united into a single column, and a bilophodont occlusal pattern persisted throughout life.
The early phyletic development of the subfamily Geomyinae took place in the tribe Dikkomyini from the early Miocene into the early Pliocene. Compared with the rapid evolution of the specializations that distinguish the Entoptychinae, the structural changes in the early Geomyinae occurred at a remarkably slow rate. In fact the lineage changed but little from Dikkomys to Pliosaccomys, in parts of the animal that can be compared, as illustrated by the low-crowned and rooted cheek teeth, the continuous enamel bands, the lack of grooving of the upper incisor, the retention of the primitive H-pattern, both above and below, in the premolar and third lower molar, and the ridges and fossae of the mandible to which the muscles of mastication attach. The only major changes detected in the known fragments are in the pattern of wear and the final configuration of the first and second molars, as described above. The unification of the two lophs in each of these two teeth into a single column was a significant step in the evolution of the Geomyinae, and is a stage between the primitive bilophodont pattern of the early and middle Miocene geomyines having continuously bicolumnar teeth and the monolophodont pattern in the modern pocket gophers of both lineages in which these teeth consist of a single column in all but the initial stages of wear. The monocolumnar structure of the first and second molars in the final stages of wear, therefore, is closer to that in the lineage of Thomomys than it is to that of Dikkomys. Other specializations in the dentition of Pliosaccomys, especially in m1 and m2 where the H-pattern has been completely eliminated from the sequence of wear, are too far advanced for Pliosaccomys to have given rise to the tribe Geomyini. The teeth in the immediate ancestor of the Geomyini must have been less specialized in m1 and m2, perhaps about as in Dikkomys. In the m1 and m2 of the tribe Geomyini, the H-pattern is formed in the initial stages of wear; therefore, in the early Pliocene ancestor, presently unknown in the fossil record, the H-pattern probably was present. Even so, the ancestor of the Geomyini and that of Pliosaccomys probably were closely allied otherwise, and both probably had attained the highly specialized fossorial adaptations characterizing all modern pocket gophers, before the divergence of Pliosaccomys and the Geomyini took place.
The evidence points to a major divergence of the geomyines that lived in the latest Miocene or the early Pliocene (probably the latter) and that gave rise to the two modern lineages, Thomomyini and Geomyini (see [Fig. 3]). One, the most primitive of the two, gave rise to the Thomomyini lineage that eventually evolved into Thomomys. Pliosaccomys is closely allied to the ancestry of this lineage, although it is probably not the actual ancestor, as mentioned previously. Aside from the aforementioned specializations of the first and second molars, the features of the Thomomyini are less advanced than in the other specialized lineage (tribe Geomyini). Primitive traits retained in the tribe Thomomyini (and also characteristic of the ancestral tribe Dikkomyini) are: (1) Small size, in general no larger than the ancestral morphotype; (2) lack of grooving on the upper incisor (although a slight rudimentary groove is developed rarely in some living species); (3) retention of anterior and posterior enamel plates in lower and upper cheek teeth; (4) premolars having widely open re-entrant folds; (5) smooth and generalized skull lacking marked angularity, regosity or cresting (neither the sagittal nor the lambdoidal crest are ordinarily well developed except in Thomomys bulbivorus); (6) forefoot small, less modified for digging than in the Geomyini.