"incipient tendency to form lophs," and H. hatcheri does the same when worn, but by union with the anterior cingulum. If cusps in H. tenuiceps do form lophs, the process is definitely not by union of the cusps with the anterior cingulum. The transverse median valley is deep and divides the tooth on the buccal side. The anteroposterior valleys are shallow and hanging, and can be said to exist only as indentations between the two sets of cusps. The paracone and metacone are much higher than the other two cusps, but much of this disparity in height may be the result of greater wear on the protocone and hypocone; H. gregoryi agrees with H. tenuiceps in these respects. Possibly the protocone and hypocone were much larger than the paracone and metacone. The internal cingulum of M1 exhibits only one large cusp opposite the medial end of the transverse valley, and shows no evidence of having been divided into two cusps. It is barely possible that there may have been two cusps and that wear makes it appear that there was only one. I doubt that there were two cusps because the cingulum is still so high (as high as the outer edges of the paracone and metacone) as to suggest that it is only slightly worn. Posteriorly this single cusp in the cingulum is united with the hypocone. Anteriorly the cusp is confluent with an anterior cingulum that is small, but, nevertheless, plainly visible as it crosses the occlusal face of the tooth to the paracone. There is some reason to believe that there was a posterior cingulum, but wear, which has obliterated even the posterior wall of the hypocone, prevents my being certain about this. This cingulum is absent in H. gregoryi and present in H. hatcheri.

M2 compares favorably with M1 except for the following differences: The protocone and hypocone are equal to the paracone and metacone in area, but not in height; although the internal cingulum is undivided, there is no evidence of a cusp as in M1. Here, too, the cingulum is as high as the paracone and metacone. Possibly the cingulum was confluent with the hypocone. The internal cingulum continues around the margin of the tooth to the paracone as an anterior cingulum which is sharper and plainer than the anterior cingulum on M1. There is no evidence of a posterior cingulum.

M3 shows a great amount of wear, and the occlusal pattern is not too clear. The median transverse valley is reduced almost to a pit, and the paracone and metacone are divided by a small notch. The protocone and paracone, the latter being much higher, are larger than the metacone which is reduced in size, and not all this difference in size can be the result of wear. The hypocone is absent. The internal cingulum is as high as the paracone and shows no evidence of division into two cusps, but in M3 this character is apparently variable for H. gregoryi does not have the internal cingulum divided and H. hatcheri has it markedly so. A slight anterior arm of the internal cingulum may have reached forward to the anterior face of the protocone. Wear prevents knowing whether a crest surrounds the tooth completely, or only on three sides.

In size the teeth of H. tenuiceps average twenty per cent larger than any of the upper teeth of H. gregoryi, H. hatcheri, or the Pipestone Springs specimen, and exceed any of the known lower teeth including those of H. vetus and H. senex by twenty-five per cent or more. Inasmuch as the upper teeth rarely exceed the lower in length in all the related genera of heteromyids, it is assumed that a similar relationship existed between the upper and lower molars of H. tenuiceps and, therefore, that this species can be distinguished by its large size. The relative size of the premolars and molars is the same in H. tenuiceps as in other species of Heliscomys. However, within the framework of this similar