In the former, garnet develops directly from pyroxenes with deposition of felspar as a by-product. In the latter, the garnet is the product of interaction between the labradorite and the augite of a dolerite.

IV. THE CHARNOCKITE SERIES IN THE EASTERN GHATS.

No attempt was made to map accurately the different types of charnockite in the main range, as this was outside the scope of my work. Several traverses into and across the hills were, however, undertaken. As a result, it is known that the commonest rock in the heart of the range is a charnockite of intermediate composition. Basic, ultrabasic, and acid types tend to be concentrated along the western edge of the main mass.

The chief difference between the charnockites of this area and those described by Holland, Walker, F. H. Smith and others is that these rocks have in most cases been crushed. They therefore show all sorts of strain phenomena extremely well.

Slide 23203 from near Bodapalle (18° 28′ 82° 19′) is typical of the intermediate charnockites in the the main range, except that it is rather too basic.

Crush phenomena.

It will be seen in linear streaks.

(Pl. 20, fig. 1) that the hypersthene is arranged It has been granulated, and altered to green chlorite and iron-ore both at the margin and along numerous cracks. A pale green pyroxene is also present, crystals of which generally show contorted cleavages.

Felspar (about andesine) is very abundant, and is cracked, strained, and granulated in the same way as is the hypersthene (Pl. 20, fig. 2). Albite is fairly common, but there is no quartz nor recognizable potash felspar.

Slide 23201 from the same locality is much more acid. It contains only a few streaks of hypersthene, but microperthite and quartz are relatively abundant. It shows all the cataclastic features of 23203. In the field it would be impossible to draw a boundary between these two rocks.

The cataclastic phenomena reach their maximum in the south of the area mapped. Here the rocks are rather more acid than usual. Pl. 20, fig. 3 illustrates a somewhat more advanced stage of cataclasis, in which the hypersthene is drawn out into elongated streamers, and the quartz and felspar occur in flattened bands.

Idiomorphic masses of garnet in association with iron-ore are often present at this stage (23226).

Further crushing leads to almost complete destruction of hypersthene, and an increased development of garnet and iron-ore. Granular hornblende is also abundant. The ultimate stage is a finely granular rock (23241, 23243) consisting of bands of quartz, hornblende, garnet, and iron-ore. Felspar may also be present, but it cannot be distinguished owing to the absence of twinning and the fineness of the grain.

In the more basic charnockites the pyroxenes and felspars are drawn out into parallel streaks at an early stage (23225). Later the pyroxene breaks up into hornblende, garnet, and iron-ore. Granulation is intense, and the final result is a banded rock resembling the crushed granitic charnockites, except that hornblende, garnet, and iron-ore are more abundant, while quartz is scarcer (Pl. 20, fig. 4). In all the slides which I examined there were residual crystals of pyroxene which had survived the granulation.

In many areas (see map) finely banded rocks composed of ironore, quartz, and garnet with a little hypersthene and pyroxene were found. These are believed to be due to the crushing of the more basic types of charnockite. In some cases they are sufficiently rich in iron to be used as ores, and are quarried to a small extent by the local iron-makers.

Uncrushed charnock

ites.

It is a remarkable fact that though most of the charnockite in the main range shows signs of crushing, there are numerous masses, both acid and basic, along the outer border of the range, which appear in the hand specimen to be uncrushed. Microscopic examination of these generally shows mortar structures, distorted twins, and similar phenomena, but some of the more basic ones show no more metamorphism than the rocks of similar composition described by Sir T. H. Holland (1900, pp. 133-170) in Madras. The phenomenon is all the more striking because the main mass in contact with these lesser intrusions is frequently intensely crushed. It is surmised that the uncrushed masses have been driven into their present position at the close of the charnockite period of igneous activity, and that the stress which ground up the earlier intrusions had waned before they were consolidated.

Garnet is locally abundant in the main charnockite mass, but it is by no means universally present. It is commonest where the

Garnet in the charnockites.

granulation is greatest, and is therefore correlated more with stress than with high temperature. This mineral is a marked feature of the hybrid gneisses in Bastar State. No evidence of its hybrid origin is here available; nevertheless it is quite likely that assimilation of older rocks by the charnockite magma has brought about conditions favourable to its development. A local increase in temperature where crushing has been greatest has probably also favoured its growth.

METAMORPHISM OF THE CHARNOCKITES IN THE MAIN RANGE.

The metamorphism of the main charnockite mass is characterized by the abundance of stress phenomena and by the presence of garnet. It has probably been brought about at quite a moderate temperature, but with the maximum stress possible at that temperature.

Admittedly some geologists (Stillwell, 1918, pp. 190-192) have claimed that the stress phenomena are only evidence of a second stage of metamorphism imposed on rocks which had already been subjected to a high grade of metamorphism, and they would quote the presence of hypersthene as evidence of this earlier high grade stage.

Hypersthene does frequently occur in high grade metamorphic rocks, and so this view cannot lightly be dismissed, but I propose to show that among the small intrusions of charnockite, which are a common feature among the older gneisses to the west of the main range (see map), hypersthene is an original pyromorphic mineral (pp. 423-425). It is difficult to believe that the hypersthene in the charnockite of the main range could have a different origin from that which occurs in its minor apophyses.

V.-CHARNOCKITE INTRUSIONS IN MALAKANAGIRI.

Norites.

Numerous small bosses of a rock closely resembling charnockite are intruded into the older gneisses all along the margin of the Eastern Ghats. Most of these lie within ten miles of the main range, but a few have been found even further afield. Basic, intermediate, and acid types are all present, but basic ones are relatively rare. They were seen, however, on hills south of Champakari (18° 22′: 81° 56′) and Dongorokolli (18° 31' 82° 6'), and in the Putparvatam and Guppakonda

reserved forests. In appearance and minerallo gy they resemble. the norites described by Holland 1900, pp. 133-170 in Madras.

They rarely show any signs of the crushing so commonly seen in the more acid rocks in this region. In most cases (24,426, and 24,431) the rocks are quite unaltered, but sometimes (24,434) the hypersthene has coronas of green biotite, and the pyroxenes tend to be altered to hornblende (23,289). In a specimen of this kind, albitization was also noted. The albite occurred on the outside edge of a biotite corona, and seemed to be replacing andesine, but as there was no development of myrmekite it was impossible to be sure of this.

INTERMEDIATE AND ACID HYPERSTHENE-GNEISSES.

The largest and most important occurrence of these rocks lies in the region between Malakanagiri and Balimela. Here a number of hills of a dark-coloured granitoid rock rise abruptly from the surrounding plains. In hand specimens this rock is coarse-grained, dark green or grey in colour, with porphyritic crystals of felspar inch long by inch wide, and fine interstitial material in which crystals of quartz can just be distinguished. No ferromagnesian minerals can be recognised with the naked eye.

Under the microscope the porphyritic features of the rock are less conspicuous than in the hand specimen. The great size of the porphyritic crystals is very readily seen, but their rhombic shape is generally disguised by the corrosion of their faces. It is clear enough, however, that the original margins of the porphyritic crystals were in many cases rectilinear.

This rock varies in mineral composition between wide limits, but speaking generally it is a mosaic of quartz and felspar with minor quantities of ferromagnesian minerals, none of which show distinct crystal faces. Mortar structures, cracked crystals, bent felspar twins, and brush polarization are all rather common, and show conclusively that the rock has been somewhat crushed. As the result of this crushing, and probably also due to subsequent albitization of the felspars, many of the specimens examined show marked granulitic features. As these igneous rocks are crushed and somewhat metamorposed, and because it is a convenient term, I propose to refer to them in the sequel as hypersthene-gneisses. I wish to make it clear, however, that they are not streaky nor banded as is the case with so many of the Indian gneisses,

Ferromagnesian minerals make up a variable but usually small part of these rocks. Foremost among them, and almost always It present, is hypersthene.

Ferromagnesian occurs in small minerals in the hypers- rounded or elongated grains, and never shows thene- gneisses. distinct crystal faces. Its extinction is straight, or nearly so, and its pleochroism is from pale pink to pale green, a well-known characteristic of hypersthene in the charnockite

series.

Rather large irregular masses of pale green pyroxene are also common. About Malakanagiri this mineral occurs along with hypersthene, but in the the crushed rocks of similar composition along the margin of the charnockite range it is sometimes the only pyroxene present.

Other femic minerals present are hornblende, biotite, and chlorite. The hornblende (24,448, 24,425, 24,430) is due to the alteration of pyroxene in most cases. It is then an olive-green variety. There are also a few grains of blue hornblende which are believed to be due to reactions between some of the ferromagnesian minerals and the soda solutions that have invaded the rock in many areas after its consolidation (see pp. 426-429).

In the fresher hypersthene gneisses (23,196, Pl. 21, fig. 1) there has been little reaction at the margin of the hypersthene. At most there is a faint greenish rim, probably due to the incipient development of biotite or chlorite. More altered varieties show much more striking phenomena. An example of this kind from the western face of Munas (18° 20′ 82° 57') contains rather a lot of hypersthene (24,425). This is all highly altered and has a dark ferruginous border outside which is a narrow zone consisting of minute flakes of hornblende or biotite in a matrix of quartz, possibly with some albite. This passes outwards into a comb of greenish yellow biotite which extends irregular fingers (Pl. 22, fig. 2) into the surrounding perthite and albite-oligoclase. In some cases these fingers put off irregular shoots along the cleavage directions of the felspar (Pl. 22, fig. 1).

In slide 24,430 from hill 1,449 north-east of Munusakonda (18° 21′ 81° 58′) the hypersthene is even more decomposed, and is largely replaced by granules of iron-ore. It can, however, just be recognized by its straight extinction and typical pleochroism. The reaction phenomena in this slide are very varied. case there is a central core of cryptocrystalline material representing

In one