relationship there are two differences. P4 of H. tenuiceps is relatively larger than the P4 of H. gregoryi, and relatively smaller than the P4 of H. hatcheri. The width of the molars is relatively greater in H. tenuiceps and H. gregoryi than in H. hatcheri.

Measurements

(In millimeters)

[Note A: This and the following measurements at occlusal surface.]

Discussion.—Heliscomys tenuiceps shows beyond any doubt that the heteromyid pattern of skull was developed by mid-Oligocene times, and in this species was already undergoing lateral compression. The major change later made in heteromyid skulls is broadening of the dorsal surface of the skull in the interorbital area.

The complete confirmation of Wood's (1939) statement that the "sciuromorph" zygomasseteric structure had been developed by this time in the heteromyid rodents as it had been in the early Eomyids is demonstrated in this specimen. Further, it is to be noted that the infraorbital canal is not sciuridlike, but has been forced forward on the rostrum, as in the Geomyoidea.

In some ways this skull shows similarities to Florentiamys loomisi Wood, of the early Miocene, which aid in determining the relationship of that unusual genus to Heliscomys and to the heteromyids in general. When Wood described Florentiamys the peculiar combination of characters found in this animal prompted him to speculate that: (1) It was a typical heteromyid which had secondarily developed cingula; (2) its cheek teeth were nearer the primitive pattern than were those of any other known fossil heteromyid, and that Heliscomys represented a simplification in the reduction of the cingula; or (3) it was not a heteromyid, but a parallel development from the "Paramys" stock. Wood favored the second possibility. Now that a part of the skull of one species of Heliscomys is known, the undivided internal cingulum that is confluent with the hypocone, the lateral compression of the deep rostrum,