the premaxillae, a small part of the left maxilla, and the anterior parts of the palatines. The nasals are missing, but both incisors are in place including most of the roots. The single median sulcus on the anterior face of each incisor is typical of the genus Cratogeomys. The rostrum is long (25.8), as great in length as in the largest subspecies of the subsimus group of Cratogeomys castanops (see previous account for explanation) and as long as the rostrum of Cratogeomys perotensis which is now known only from Veracruz, México. The length of the rostrum was measured from the most anterior median projection of the premaxillary to the most posterior dorsal projection of the same bone. Actually, and especially in relation to its length, the rostrum of the fossil is remarkably narrow. The breadth of the rostrum measures 10.4, which is comparable to that in the subnubilus group of small subspecies, and less than that (11.4 in the smaller adult females to 13.7 in the larger adult males) in the subsimus group of large subspecies. The breadth of rostrum in the fossil is 40.3 per cent of the length of the rostrum. In living Cratogeomys castanops (both the large and small subspecies groups, and including both females and males) the breadth of rostrum amounts to between 44.0 and 51.4 per cent of its length. The rostrum in Cratogeomys perotensis (and in other species of the merriami group) is relatively much broader than in Cratogeomys castanops. Even though the rostrum of the fossil is narrower than in Recent species of Cratogeomys, the ventral border in the area of the palatine slits is more heavily constructed than in any of the living species, and it is nearly parallel-sided rather than tapered toward the midline anteriorly. At the lateral edge of the enamel plate of the incisors there is a distinct shelf, a characteristic of the merriami group of species and a feature not well developed in Cratogeomys castanops.

I hestitate to refer this fragment to any of the living species, although I would judge it to represent a form closer to the species castanops than to the merriami group (C. perotensis). The rostrum may represent, and probably does, an undescribed and extinct species of Cratogeomys, but in my opinion it should not be given formal taxonomic status until more adequate material is available.

If the fossil is actually Cratogeomys castanops, and if the fragment is from an earlier deposit in the cave than is the material here assigned to Cratogeomys castanops, the fossil stock could be ancestral to the group of small subspecies provided there had been a trend in evolution toward smaller size. Another possibility is that a shift in geographic range of the kinds of Cratogeomys that lived in the vicinity of the cave has occurred, and that the fossil represents an evolutionary line with no close relationship to Recent species and now is extinct. Additional material is needed before the history of these species can be reconstructed with validity.

Description of Holotype.—Differs from any known living species of Heterogeomys, by the significantly heavier and deeper ramus (see