Chap. XII.] kodurite series : composition. 259 
If the Kotakarra analysis be compared with the analysis of a specimen 
from Garb ham given on page 168 it will be seen that the two analyses are 
very similar, the only important difference being that in the Garbham 
garnet a much larger proportion of the alumina is replaced by ferric 
oxide than in the Kotakarra garnet. In either case these garnets can be 
designated spandite, a contraction of spessart-andradite, indicating 
their intermediate composition. The Boirani garnet, however, owdng 
to its small percentage of manganese protoxide and high percentage 
of lime, must be regarded as intermediate between grossularite and 
andradite rather than spessartite and andradite, and consequently might 
be called on the same system grossular-andradite, or for short grandite. 
The use of this term is not, however, as necessary as that of spandite, 
for only one locality for the rocks of this series has yet been discovered 
in the Ganjam district. 
Taking the analyses expressing the mineralogical composition given on 
page 258, it is possible by making fairly legitimate assumptions to arrive 
at the approximate original composition of the rock before alteration. 
Let us take first the Kotakarra rock. In this case we need only consider 
three minerals, apatite, spandite, and opal. Taking the specific gravity 
of the spandite to be 4' 0 (G. of the Garbham garnet was fomid to be 4* 02), 
that of apatite to be 3 • 20, and that of the rock analysed 2 * 85, it is easy to 
calculate the specific gravity of the opal. This works out as 1 ' 87. This 
figure is, of course, a little too low for opal, no doubt on account 
of the fact that the rock contains some closed spaces that were not filled 
by water, even after prolonged soaking, when the specific gravity was 
being taken. Hence the figure really represents the specific gravity 
of the opal and air spaces taken together. As has already been explained 
this opal has probably been formed by the replacement of the original 
felspar of the rock, which, judging from all the other occurrences of 
kodurite, must have been a potash variety probably containing a small 
percentage of soda in the form of microperthitic inclusions. By replacing 
the opal by orthoclase (G. = 2" 57) the following works out as the original 
mineralogical composition of the rock : — 
Apatite 3 '70 
Spandite „ . . . 60 "58 
Orthoclase ......... 45 '44 
Ti02 0-32 
CuO 0 '02 
110-06 
