Albite, oligoclase, and myrmekite are all present in abundance in these rocks, but they present no features which have not already been described in connection with the hypersthene-gneisses. The quartz in these rocks generally shows the rod-like inclusions seen. in charnockite.

Iron-ore, apatite, and zircon are all rather common, but rarer than in the hypersthene-gneisses. Their mode of occurrence is identical in both cases, and their relative scarcity in these soda-granites is probably due to the absence of ferromagnesians, round which these minerals tend to cluster.

Iron-ore, apatite,

and zircon.

VII.-ANALYSES OF THE MALAKANAGIRI GNEISSES. Analysis No. 1 is of a specimen collected from the high hills north of Korukonda (18° 17′ 81° 59′). No. 2 is of a rock collected from the hill one mile east of Potrelu (18° 16′ 82° 2′). Both rocks come from the igneous massif which extends eastwards from near Malakanagiri to within three miles of the main charnockite range of the Eastern Ghats. (Analyst-Mahadeo Ram.)

No. 1 is probably as basic as any of the rocks here. No. 2 is typical of the acid rocks in the outlying parts of this massif descibed elsewhere as gneissic soda-granites. Most of the rocks in this massif would have a chemical composition intermediate between these two

extremes.

As regards the norm of No. 1 the calculated percentage of lime felspar is probably too high, as some of the lime has been used up in the formation of hornblende, which mineral is rather abundant in the mode. After allowing for this the percentages of the quartz, felspar, ferromagnesian and other minerals shown in the norm is much the same as that estimated microscopically for the mode.

As regards the norm of No. 2 the mafic minerals shown are probably represented by biotite and iron-ore in the mode. The large percentage of albite shown probably occurs partly intergrown with a potash felspar in the perthite, partly as albite and myrmekite, and partly with anorthite to form oligoclase. The ratio of felspar to quartz in the norm differs little from that observed in the mode.

In the graphs on pl. 23 I have plotted the percentages of the various oxides in a series of charnockites against their silica contents. These analyses have been made of specimens collected from localities in Ceylon, from St. Thomas' Mount, from the Shevaroy hills, and from Pallavaram in Madras. The analyses are by H. S. Washington (1929, p. 481) R. J. C. Fabry (1929, p. 481) and J. T. Donald (1929, p. 481). Along with these I have also plotted the two analyses of the Malakanagiri rocks by Mahadeo Ram.

This procedure is open to all sorts of criticism. The most that can be hoped for it is to get a general idea of the differentiation of the charnockites. Even for this the number of analyses, especially towards the basic end, is inadequate.

The range of variation among the analyses, even from a single area like Ceylon, is very considerable. The graphs drawn are intended to be an average of the various analyses. In the Malakanagiri hypersthene-gneiss the soda and alumina are above average, iron is below, while potash, lime and magnesia are about normal. The high soda and alumina may be due to the introduction of albite

by soda solutions in the end stages of consolidation. On the other hand the difference from normal is well within the limits of variation shown by the other analyses, so that there is really no evidence to show that the high soda in this case is not a mere local abnormality.

In the analysis of the more acid of the Malakanagiri rocks the soda and alumina are far too high and the potash much too low. The differences greatly exceed the limits shown by the other analyses. If this rock can be regarded as a charnockite at all, and the field evidence strongly favours that view, soda and alumina must have been added to it and potash removed. This in my opinion is due to albitization in the final stages of consolidation.

VIII. THE CHARNOCKITES OF MALAKANAGIRI AND

ST. THOMAS' MOUNT.

A comparison of the hypersthene-gneiss of Malakanagiri with the charnockite of St. Thomas' Mount shows many similarities and some differences.

Though they occur hundreds of miles apart, both are intrusions in much older gneisses, form similar hill masses, and are indistinguishable in hand specimens. Moreover both are closely associated with norites and with acid granulites free from hypersthene.

Microscopic examination shows that the hypersthene, the microperthitic potash felspars, and the acicular inclusions in the quartz, which are considered characteristic of charnockite, also occur in the Malakanagiri rocks. The presence of quartz de corrosion' or myrmekite and of the accessory minerals apatite and zircon, are further points in common.

The only difference in field occurrence between these two rocks is that the Malakanagiri ones have been definitely crushed subsequent to their consolidation. whereas the Madras charnockites have not noticeably suffered in this way, though Sir T. H. Holland (1900, p. 154), expresses the opinion that their granulitic texture is due to movement in the last stages of crystallization.

Texturally they differ in that the hypersthene-gneiss is porphyritic, and shows distinct signs of cataclasis, whereas the charnockite is granulitic, and there is no obvious cataclasis.

Cataclastic phenomena are of little genetic importance, as they are secondary. The greatest difference between the two sets of

rocks lies in the porphyritic nature of the one as compared with the granulitic texture of the other. The granulitic texture of the Madras charnockite is according to Holland not an original feature, and therefore is not of any great significance. In any event accepted charnockites having a porphyritic texture occur in parts of Madras and in the Eastern Ghats, as well as in Malakanagiri. The s'ight textural differences may therefore be disregarded.

As far as general composition is concerned there is nothing in the hypersthene-gneiss analysis from Malakanagiri to prevent it being a charnockite. Microscopic examination confirmed by the chemical analyses shows that soda felspars and apatite are common in the rocks of this area than in the Madras ones, but I think this is adequately explained by the albitization which they have suffered during their consolidation.

more

The resemblance between the Malakanagiri gneiss and charnockite in occurrence, mineralogy, and composition is very striking. The points of difference are small, unimportant, and readily explained. I, therefore, regard these gneisses as true members of the charnockite scries. They occur within three miles of the vast charnockite boss of the Eastern Ghats, and are almost certainly apophyses from it. These views are not new nor revolutionary, but have been expressed by all field workers in the region west of the Ghats. They may be accepted without reserve.

IX.-IGNEOUS NATURE OF THE MALAKANAGIRI
CHARNOCKITES.

Sir T. H. Holland, who first described the charnockite series, regarded its rocks as igneous. He considered (1900, p. 242) that their textural and mineralogical characters were original features due to their igneous origin, which had been little affected by subsequent metamorphism. Most geologists who have worked on the charnockite series in India have agreed with Holland in a general way, though instances of charnockites which have been considerably metamorphosed, especially by crushing, have been described.

Mr. E. Vredenburg (1918, pp. 423-448) in the year 1918 put forward the theory that the charnockites in Mysore were the high grade metamorphic derivatives of the great basic igneous series of that region. About the same time Dr. F. L. Stillwell (1918, pp. 128-130, 138-139, 190-192) advanced somewhat similar ideas, about

More

the charnockite-like rocks found by him in Adelie Land. recently Mr. A. W. Groves (1933, pp. 150-208) has given an account of charnockite-like rocks in Uganda and has put forward a formidable array of chemical and petrological facts to prove that these rocks were formed by the plutonic metamorphism of a normal series of plutonic rocks.

The chemistry of the Malakanagiri gneisses and charnockites is not well enough known to permit a comparison with that of the similar Uganda rocks, but the field and petrological characteristics of the Malakanagiri rocks are known, and I think they suffice to prove conclusively that the charnockites of this region are slightly altered igneous rocks like those described by Holland in Madras, and not intensely metamorphosed igneous rocks like those described by Groves in Uganda.

One of Groves' main points is that there is a gradual passage from gneisses to charneckites. In Malakanagiri the charnockite hills often rise like igneous plugs 1,000 feet above the flat plain of felspathic gneiss. The actual junction is rarely seen, owing to the dense forest and debris, but its position can be fixed within a few yards. The change from one rock to another is, therefore, quite sharp in contrast to the very gradual change noted in Uganda.

Moreover, both acid and basic charnockites in Malakanagiri are surrounded by a uniform granitic gneiss. It is difficult to see how any metamorphic process could produce two rocks of totally different composition from a uniform granitic gneiss. The position in Uganda is quite different. There the gneisses are chemically almost identical with the charnockite-like rocks said to be their metamorphic equivalents.

All are agreed that metamorphic hypersthene is only found in rocks of the highest metamorphic grade. Such a grade may be due to regional or thermal causes. I have shown (pp. 407-408) that the regional metamorphism in Malakanagiri is always relatively low nor are there any important intrusions there which could have thermally metamorphosed the charnockites. It is difficult to see how they could ever have been raised to a temperature high enough to produce hypersthene.

The dolerite dykes are of some interest in this connection. Groves describes how similar rocks in passing from gneiss to charnockite exhibit all the stages of metamorphism from a dolerite to