44 The American Geologist. January, ]904 
The only conclusion is that the hypothetical river did not 
exist. The error is partly due to reliance on surficial maps and 
to ignoring the rock geology. It illustrates the truth that ge- 
ography must be based on geology. The complicated problem 
here raised can not be fully solved by any study of stream di- 
rections and present contours on topographic maps. The mod- 
ern topography and the present trend of valleys and drainage 
may suggest certain directions of ancient stream-flow, and may 
properly form a basis for a working hypothesis, if not contra- 
dicted by other facts. But the basal facts to which all theories 
of ancient stream-flow must yield are those relating to the hard- 
rock topography, or the old valley forms with the drift-filling 
ignored. Until borings in the valleys in question have been 
made in number sufficient to determine rock altitudes of critical 
points and the slopes of the rock valleys the true direction of 
the preglacial drainage must remain in question. However, the 
non-existence of an ancient valley at Ithaca, from the northeast, 
can not be in question. 
Another factor which should be considered, although it may 
not be of importance in this case, is the preglacial, differential, 
northward uplift. In the district of Cayuga and Seneca valleys 
this uplift is probably not less than two or three feet to the mile ; 
which is a far greater slope than the gradient of mature streams 
of large size. 
The remarkable rhomboidal forms of the gross topography 
between the meridians of Ithaca and Cazenovia present an in- 
teresting problem, but much more complicated than it might 
appear. There is a large element of stratigraphical geology 
involved, along with preglacial epeirogenic movements. By 
way of stimulating investigation the writer will venture the fol- 
lowing prediction, (i). That borings will show, in general, 
relatively shallow drift in the northeast-southwest valleys, 
which implies that these broad valleys were the product of 
weathering along the strike of the strata, with some early work 
of beheaded or diverted streams. In other words, these val- 
leys were cols between deeper valleys. (2). That the main 
drainage of the area, preceding the ice invasion, was northward 
by several parallel valleys, of which the present broad areas at 
Freeville and Cortland are open sections. These north-leading 
valleys were subsequent to the greater east-west Ontario valley, 
