MANSFIELD: ROXBURY CONGLOMERATE. 169 
toward the south. According to Walcott the nearest place where 
quartzite containing the same fauna is exposed is at Great Belle Island 
near Newfoundland (Walcott, b, p. 327). If this place is accepted 
as the probable source of these pebbles it is necessary to invoke the 
aid of some agent of transportation powerful enough to carry pebbles 
of large size for great distances without permitting them to be com- 
minuted during the process. Such a view would favor the idea of 
glacial action as advocated by Shaler (Shaler et al., p. 57-59). It is 
possible, however, to conceive the existence of a land mass, to the 
south and east of the present coast, large enough to supply the materials 
observed in the conglomerate. The marked increase in the coarse- 
ness of the conglomerate and in the number of fossiliferous pebbles 
in that direction tends to favor such a view. 
The distortion of the pebbles of the Purgatory Conglomerate has 
been such as. almost completely to obliterate any traces of fossils, so 
that it is not definitely known whether the great pebbles in that rock 
really belong to the fossiliferous quartzite. The writer saw some 
indications in the rock at the shoreward end of the Purgatory Penin- 
sula that strongly suggested the occurrence of Oboli in the large quart- 
zite pebbles of that locality. At the southeast and southwest of the 
Narragansett Basin quartzites occur in situ. Those to the southwest 
at least have furnished material to the conglomerates but they do not 
contain fossil Oboli (Foerste, b, p. 382). 
The quartzite of the Harvard Conglomerate is unlike that of neigh- 
boring localities in having small grains of uniform size, showing a 
considerable degree of attrition. Professor Emerson, in conversation 
with the writer, suggested that the quartzite represented an aeolian 
deposit. Daubree has shown that in water-laid sands the grains less 
than one-tenth of a millimeter in diameter are angular (Daubree, 
p. 256). In wind-blown deposits coarse and fine grains alike are 
more or less rounded. i 
Granite. 'The granites most frequently represented in the conglom- 
erates of the Boston and Norfolk Basins are a fine grained pinkish 
variety, containing little ferro-magnesian material, and a coarser type 
corresponding to the granites now exposed at Dedham, Randolph, 
and Cohasset, and consisting of quartz, pink and green feldspars, and 
biotite with hornblende. Similar, though not certainly identical, 
granite occurs in the northern highlands in considerable abundance- 
and is typically exposed near Saugus. No granite of the Quiney-Lynn ' 
variety (bluish gray hornblende granite) has been seen by the writer 
in any part of the conglomerate. The occurrence of granite both 
