1 84 
THE ROMAN COMAGMATIC REGION. 
In the first place, as regards the amounts of silica alone some very decided 
breaks in the continuity are visible, as has been mentioned on a previous page. 
These are especially marked and wide between 48 and 50 per cent, and between 
52.5 and 54 . 6. There are narrower ones between 56 . 2 and 57 . 3 (with one analysis 
in the gap), and between 54.7 and 59.2, though probably the series of silica per- 
centages between 54.7 and 61 .9 may be regarded as continuous. As the number 
of analyses is rather large, it would be reasonable to suppose that the two large 
breaks are real and that they point to some definite differentiation of the general 
magma into at least three submagmas. It would be of very great interest to discover 
if similar clusterings, or indications of the presence of submagmas, are to be found 
in other regions. There are some indications that such is the case elsewhere, but 
30 
f+S;0 2 46 4S 50 52 4 56 59 6O 6* 
* A'o/7 leucitic rocks = Leucitic rocks 
FIG. 2. 
this is a matter the investigation of which must be deferred for the present, as it is 
too general in scope for the purposes of this paper. 
Returning to the immediate question at issue, it will be seen that the rocks 
below 52.5 of silica are uniformly leucitic, while those above 56.3 are as uniformly 
non-leucitic, and those with silica between 54.6 and 56.3 are of both leucitic and 
non-leucitic types. 
Confining our attention to the rocks with silica above 54 . 6, it is clear that the 
silica-potash ratios of the non-leucitic rocks are uniformly above those of the leucitic 
ones in the group between 54.6 and 56.3, which may be called the critical group, 
and which embraces most of the types of vulsinose and ciminose. To the right 
of this the silica-potash ratios tend to decrease and even reach the level of or fall 
below the ratios of the leucitic rocks of the critical group. The division line between 
them may therefore be regarded as a curve which inclines downward gently to the 
right, that is, as silica increases, slightly above, but parallel to, the line A-B of 
