338 
Manganese deposits of india : geology. [Part II : 
the photo-micrograpli (Plate 11, fig. 2). there is also a small amount 
of rutile present in the rock. The garnets are usually isotropic, but 
not infrequently, owing no doubt to strain effects, they exhibit 
anomalous double refraction. 
The relative proportions of the two constituents of gondite vary be- 
tween very wide limits, so that there is every gradation between quartz- 
rock free from spessartite and spessartite-rock tree from quartz. The 
typical rock, however, may be regarded as that which contains about 
equal proportions of the two constituents. The microscope shows that 
those varieties that show purphsh or reddish tints in the hand-speci- 
men do not as a rule owe their coloiir to the garnet being purple or 
red ; but to the fact that the separate grains of garnet each contain a 
small cloud of finely -di\aded red dust, probably oxide of iron, some- 
times uniformly distributed throughout the grain, but more often 
collected in the central parts of the grain so as to leave the periphery 
clear. The other variations of colour seem, however, to be due to vari- 
ations in the colour of the spessartite, which may range from almost 
colourless through pale yellow, bright yellow, and orange, to orange-red, 
or even deep blood-red. This variation in coloiir is no doubt due to vari- 
ations in the amounts of manganese and iron in the garnet. The fine- 
grained variety described above is perhaps the commonest variety of 
gondite and is to be regarded as the typical rock. From this degree of 
fine grain there is every gradation up to varieties of gondite of very coarse 
grain in which the garnets may be as much as 1 or even 2 inches in 
diameter. In these more coarsely crystalline varieties the garnets are 
often well developed so as to show their crystalline form. This is 
almost invariably seen to be the simple trapezohedron or icosatetrahe- 
dron, although other faces, for an account of which see the mineralogi- 
cal description of this mineral (page 172), are not infrequently found. 
Dana's ' System of Mineralogy ' does not indicate that particular 
varieties of garnet favour particular mineral habits. My experience, 
however, of the Indian garnets points to the fact that if a garnet exhibit 
predominant trapezohedral faces it is almost certain to give a strong re- 
action for manganese when fused with nitre and fusion mixture. In fact, 
in the Central Provinces one would almost invariably be correct if one 
designated all garnets exhibiting predominant trapezohedral faces as 
manganese -garnets, and all garnets in which other faces predominate as 
garnets practically free from manganese and far removed from spessartite 
in composition. There are exceptions to this rule, as is evident 
from the fact that a single example of an octahedral spessartite-garnet 
