WOODWORTH: GEOLOGICAL EXPEDITION TO BRAZIL AND CHILE. Sl 
result of deposition from floating ice and such distribution can not be 
depended upon as a basis for discussing advances and recessions of 
glaciers. It can only be said here that the evidence seen points to the 
existence of two horizons on which ice-action directly or indtrectly 
arising from glaciation appears to be demanded. 
It would seem from a comparison of the partial sections visited in 
the course of this Expedition that the boulder-beds are not persistently 
parallel formations, that the more typical tillite of one district may 
pass by gradation or intercalation of deposits into contemporaneous 
waterworn gravels or sands, now conglomerates and sandstones in 
another district along the strike of the beds. Thus in the sections 
along the Jaguaricatu in northeastern Parana, the tillite with large 
angular blocks surmounts with few intervening feet of beds the white 
fine sandstone which appears to be a fairly persistent basal and pre- 
glacial member of the Permian series. In the latitude of Ponta 
Grossa the tillite appears at a much higher horizon, the apparent 
place of the Jaguaricatu tillite formation being taken by beds of 
waterworn pebbles. Again in the gorge of the Iguassti at Serrinha 
Station, if the beds be truly of Permo-Carbonic age at this horizon, a 
conglomerate with striated pebbles occupies an inferior position near 
the base of the series. ‘At Orleans, yet further south, the tillite is 
replaced in the section by waterworn conglomerates possibly, though 
not certainly, laid down in the presence of ice in the manner of eskers. 
This alternation from point to point along the present roughly 
meridional] line of exposures, thus described of deposits approximately 
at the same level in the series, and probably more or less contem- 
poraneous, finds a parallel in the existing deposits of glacial origin 
within the glaciated areas of Europe and North America. In travers- 
ing the glaciated region of the latter area we pass from north to south 
over belts of till with alternating strips of waterworn gravels and 
coarse sands and clays. In this particular case the deposits are 
successively newer in the direction in which the ice retreated, 7. e. 
towards the north. We encounter another mode of deposition of 
alternating accumulations of ice-laid and water-laid drift, however, in 
which the likeness to the Brazilian distribution, so far as it is at 
present known, is equally close. That is where glaciers coming down 
either as distinct valley glaciers or as outflowing tongues from a 
central ice-cap reach the coastal plain or sea-floor at the base of a 
highland region so as to deposit till in the vicinity of the paths by 
which they reach the low grounds while the intervening areas receive 
only the waterworn débris. In south Brazil, what seems to be evi- 
