SEDIMENTARY BELT OF THE COAST OF BRAZIL 229 
and in one lens they were associated with plant remains, in the char- 
acteristic form of jet, which is of frequent occurrence in the layers 
containing fossil vertebrates. In short, the re-examination of the 
section with this point especially in view convinced me that it repre- 
sents a geological unit. At the other locality, near Pojuca station, 
where mollusks occur, Mr. Mawson, a very competent and pains- 
taking observer, reports (10) vertebrate and molluscan fossil from 
the same layer of only a few inches’ thickness. 
In view of these observations, the hypothesis of an admixture, 
through careless ‘collecting, of fossils from two distinct formations 
may be safely put aside, and as the molluscan fauna is a compara- 
tively characterless one, but little weight should be given to it as against 
the well-characterized Mesozoic forms of the vertebrate fauna. 
The strictly fresh-water faunal elements of the Bahia basin are 
not sufficiently widespread, either in horizontal or vertical range, 
to characterize the formation as a whole, and it seems probable 
that, instead of being in a strict sense a fresh-water basin, it was rather 
an estuary or semi-detached arm of the sea in which deposits character- 
istic of fresh, brackish, and salt water might be laid down in different 
parts (or at different epochs in the same parts) and in close connec- 
tion, or rapid alternation, with each other. A widespread and 
characteristic feature is the terrestrial elements represented by 
plant and dinosaurian remains. The former are represented by 
scattered fragments of carbonized wood, usually in the form of more 
or less pulverulent lignite in the sandstones and of jet in the shales. 
Throughout the district the beds with recognizable fossils are 
overlain by soft sandstones, which, when they show plant remains 
or are distinctly inclined, have been referred to the Cretaceous, and 
in, the contrary case to the Tertiary. As Branner has pointed out 
these criteria are unsafe ones, and doubtless some errors have been 
committed. My own impression is that in most cases these errors 
will prove to have been in the sense of placing the limit between 
Cretaceous and Tertiary too low down, rather than, as Branner seems 
to think, not low enough. The plant-bearing beds are often hori- 
zontal, or apparently so, over considerable areas, and if, as may 
readily happen, no plant remains are observed in such areas, they may 
erroneously be referred to the Tertiary. Indeed, it is by no means 
