154 
ANNALS NEW YORK ACADEMY OF SCIENCES 
centages of extinct genera. The personal equation of the describe!’ and 
the imperfection of the types enter so much into the generic reference of 
extinct species, that correlations on this basis are not very reliable when 
taken at their face value. Comparisons should also be made with North 
American rodent genera of the Tertiary epochs and their modern survivors. 
A further consideration that should be taken into account is that the exten¬ 
sive immigration of northern forms into South America at the end of the 
Tertiary would naturally have caused a rapid extinction or modification of 
the native rodent genera with which they came into competition. These 
considerations have led the reviewer to agree with Scott in ascribing a 
Miocene and probably late Miocene age to the Santa Cruz rodentia. 
IV. The Patagonian Sandstone Formation. 
This thick and generally barren formation includes the Teliuelche of 
Gaudry and the Cape Fairweather Beds of Hatcher. It contains, according 
to Roth, an admixture of Santa Cruzian and Pampean genera (at Lago 
Fontana and Lago Blanco). Four subdivisions may be recognized,— Rio 
Frias, Nahuel Huapi, Santa Rosa ( = Cape Fairweather) and Rio Negro 
beds. The first three are regarded by the author as Miocene, the last as 
Pliocene. 
V. The Pampean Formation. 
Roth uses this term in a rather wide or comprehensive sense, including 
the Parana beds as well as the typical Pampean. He regards the lower 
part of the Parana formation (= Monte Hermoso beds of Ameghino), as 
not later than Miocene, the criteria being chiefly the ratio of extinct to living 
genera. The earliest precursors of the great faunal invasion from North 
America appear at this point (Pararctotherium etc.) but the great mass of the 
northern invaders ( Canis , Felis, Equus, Mastodon etc.) first appear in the 
upper strata of the Middle Pampean, although Cervidse and Ursidte appear 
somewhat earlier. Roth takes exception to the generally held view of the 
North American origin of the South American species of Equidae as follows: 
The occurrence of Equus in the Middle Pampean is no evidence for the Pleisto¬ 
cene age of these beds. We did not receive the Equidse from North America, as is 
often asserted. There occur here contemporaneously three well separated genera, 
of which two are not present in North America, a proof that they did not reach us 
from there. The genus Equus occurs as early as the Siwalik beds of India [which 
Roth accepts as Lower Pliocene]. Among all the immigrant mammals which occur 
in the uppermost beds of the Middle Pampean there is no single genus which does not 
occur in the Northern hemisphere as early as the Miocene. 
