Besides these forms, remains belonging to bats, various birds, snakes, and batrachians have been recognized. Of the fossils identified generically or specifically those belonging to Megalonyx, Tapirus, Mylohyus, Cervalces, Mammut, and Arctotherium are certainly extinct. Probably, too, the bison and the species of Felis are extinct. There are, therefore, pretty certainly close to 50 per cent of the species which are no longer living. This percentage and the history of some of the genera make it improbable that the assemblage belongs to the Late Wisconsin stage. Some of them could hardly have been living during the Wisconsin, when the foot of the glacier was within 100 miles toward the northeast and northwest. On the other hand, there are no species or genera present which make it necessary to refer the collection to the first interglacial. The assemblage probably belongs to the middle Pleistocene.

Coming now to the very southwestern corner of the State, we find that. Elephas columbi has been met with in the bed of Hargus Creek, 3 miles above Rogersville, in Greene County (p. [150]), and E. primigenius on Gray’s Fork of Ten Mile Creek, near Graysville (p. [133]). In the Rogersville Folio (No. 146, U. S. Geol. Surv.), Dr. F. G. Clapp described the geology of this quadrangle. On his page 10 he briefly discussed the meager Quaternary deposits of the area. These he referred to the Carmichaels formation, and indicated his opinion that it belonged to very early Pleistocene. On the geological map it is represented as occurring along Ten Mile Creek at and just below Rogersville. The occurrence of a tooth of Elephas columbi just above this town and of E. primigenius just above Harveys (p. [133]) renders it probable that other patches of the formation exist further up the stream and along some of its branches, and that the fossils were derived from that formation. It is, of course, possible that small patches of a later deposit exist there.

Reference has been made to the Carmichaels formation. The type locality is found at Carmichaels, on Muddy Creek in Washington County. The geological description of the locality has been presented by Marius R. Campbell in the Masontown-Uniontown Folio (No. 82, U. S. Geological Survey). The formation occurs extensively along Monongahela River and other streams of western Pennsylvania. For information the reader should consult the Geological Survey Folios Nos. 144, 146, 121, 82, and 177. The deposits occur at levels considerably above the present streams and are regarded as having been laid down in old and now abandoned river channels and in tributaries of these. The time when this occurred is believed by many, if not most geologists to belong to the early Pleistocene, the Kansas stage, or possibly the Nebraskan. In the opinion of some geologists the glacial ice dammed the streams and caused their valleys to be filled with detritus. More recent Pleistocene deposits, possibly of Wisconsin age, occur at lower levels in some places south of the Wisconsin moraine; and perhaps the age of some of them has not yet been recognized. When remains of vertebrate animals are discovered, it is of great importance to determine, if possible, the exact levels of their origin.

On another page mention is made of the finding of a tooth of Elephas primigenius at Lone Pine (p. [133]), 7.25 miles south of southeast of Washington. This village is on Little Ten Mile Creek. No details of the discovery have been received. From Folio 144 of the U. S. Geological Survey it is learned that patches of the Carmichaels formation are found for several miles along Ten Mile Creek, near the southern boundary of the quadrangle. It seems probable that there may be patches of the same deposit along Little Ten Mile Creek, in the neighborhood of Lone Pine.

As detailed on page [70], a mastodon tooth was found many years ago about 1.5 miles south of the village of Hickory, Washington County, about twenty miles southwest of Pittsburgh. Westland Run empties into Chartiers Creek, and this into the Ohio at Pittsburgh. The geology of Burgettstown and Carnegie Quadrangles has been described by E. W. Shaw and M. J. Munn (Folio 177, U. S. Geol. Surv. 1911). No Pleistocene deposits are mapped on the stream mentioned; but just a little lower down, on Chartiers Creek, is a patch of the Carmichaels formation. Below Hickory somewhere there must be a Pleistocene deposit of some kind, and it is more probably early than late Pleistocene.

From the vicinity of Pittsburgh there have been reported remains of the mastodon (p. [69]), of Elephas columbi (p. [150]), and of an undetermined species of elephant (p. [168]). Neither of the elephants is certainly determinable. The mastodon, represented by fragments of bones and teeth, is said to have been found in the river bank, at the junction of Monongahela and Allegheny Rivers. It is impossible to determine the Pleistocene stage to which any of these proboscidean remains belong. As shown on the geological map of the Carnegie Quadrangle (Folio 177, U. S. Geol. Surv.) there are indicated here Pleistocene deposits of early, intermediate, and late stages.

Little information is furnished by a mastodon reported found on Dicks Creek in Butler County. The statements regarding the finding of elephant remains on French Creek near Meadville are vague and valueless (p. [168]). Some remains of Elephas columbi have been found at Tryonville, at a depth of 7 feet (p. [150]). The town is on the Wisconsin moraine and the elephant probably belongs to the Late Wisconsin.

Nearly a hundred years ago a tooth of Elephas primigenius was reported from a place in Erie County, called Beaverdam (p. [133]). From Mr. Clyde C. Hill, civil engineer, Northeast, Erie County, the information is received that Beaverdam is a cross-roads hamlet about 23 miles south of the lake, near the prolongation of the western New York boundary line. This is within the area covered by Wisconsin drift, and it is pretty certain that the animal lived there after or near the close of the Wisconsin stage.

Just west of Erie a mastodon tooth has been found along Chase Creek (p. [70]). Unless there are some unrecognized pre-Wisconsin deposits along this creek, the animal must have lived there at a time after the lake had retired to about its present limits. This would be near the very close of the Pleistocene epoch. The same conclusion must be arrived at from a study of the proboscidean remains (supposed to be those of an elephant) found at Girard.