molecules in solid solution. Quartz veins are commonly found in sphalerite (Pl. 9, fig. 1). In one specimen a vein containing bismuthinite was observed in it and in another bismuthinite slightly replaces sphalerite.

Stannite, besides occurring as ex-solution droplets in sphalerite, also forms occasional coarser grains. This coarser stannite may, like sphalerite, contain minute ex-solution droplets and also veinlets of chalcopyrite. For the most part stannite preceded galena (Pl. 9, fig. 2).

Chalcopyrite also occurs as coarser grains; it replaces and forms veins in arsenopyrite (Pl. 8, fig. 2) and pyrite and it also forms veins in galena (Pl. 10, fig. 2).

Bismuthinite sometimes occurs as coarse crystals, but usually it is associated with, and is veined along its cleavage, by galena; intergrowths are common (Pl. 9, fig. 4). The association of these two minerals suggests that they were deposited close together in time but that bismuthinite was the first to form. In one specimen bismuthinite accompanied calcite as a veinlet in sphalerite, but in other sections carbonate veins occur in bismuthinite. Apparent veinlets of bismuthinite are seen in quartz, and it occasionally replaces the latter (Pl. 9, fig. 3).

Galena was not found to vein sphalerite. It is, however, veined by chalcopyrite, the veinlets following the cubic cleavage in galena. Remembering the close association of chalcopyrite with sphalerite, this might, perhaps, suggest that galena was earlier than sphalerite, but as chalcopyrite separated from sphalerite only on the fall of temperature, it is probable that sphalerite (with some chalcopyrite and stannite in solid solution) was deposited at a higher temperature than galena. Galena and bismuthinite replace both pyrite and arsenopyrite.

Carbonate commenced deposition late in the sequence, as it veins all the preceding minerals. However, one case of a vein of chalcopyrite and galena in platy carbonate was seen. The carbonate is apparently a normal calcite, but where associated with chlorite it is a brown ferrous variety. Occasionally the gangue consists almost entirely of calcite with only a very little tourmaline or micaceous material. It has often replaced quartz in a remarkably thorough manner (Pl. 12, fig. 3).

Scheelite replaces wolfram (Pl. 6, fig. 4; Pl. 7, fig. 1; Pl. 12, fig. 1), and the resulting pseudomorph may retain both the form

and the cleavage of the prismatic wolfram. As a rule unreplaced relicts of the original wolfram prisms are retained. Scheelite which is interstitial is white, whereas that which has replaced wolfram is yellow, but the two are optically continuous and there is no reason to suspect that they are of different ages. Scheelite is commonly interstitial to pyrite and later minerals. It has been observed to replace quartz along the border of wolfram and is occasionally interstitial to quartz. Instances of its replacement by lepidolite and muscovite were noted. Scheelite was also seen to replace calcite along the latter's cleavage (Pl. 11, fig. 4), but the platy variety of calcite is commonly euhedral to it.

Muscovite (very fine-grained gilbertite) and lepidolite occur together; the latter is, perhaps, the earlier. Muscovite forms veins in cassiterite (Pl. 12, fig. 1) and also replaces and forms veins in wolfram, scheelite, tourmaline (Pl. 10, fig. 4), phenacite and the sulphides. Lepidolite occurs as small radiating groups and shows a typical colourless to pale pink pleochroism. Some of the muscovite is altered to a pale brownish variety. In one section veins of gilbertite in cassiterite stop abruptly at the border of scheelite after wolfram and continue on from the further side of the replaced wolfram, its position in the latter being occupied by a line of white scheelite (Pl. 12, fig. 1). This might suggest that the replacement of wolfram by scheelite was closely associated with the introduction of the micas.

Chlorite is sometimes seen to have been derived from muscovite, but much of it is truly hypogene. It may replace any of the earlier minerals. In one section a few thin veins of calcite were seen to vein chlorite.

Supergene changes.

Supergene changes in these minerals as a result of surface alteration are surprisingly few. Very small amounts of chalcocite, covellite, cerussite and tungstic oxide have been found at all horizons down to No. 4.

Distribution of the minerals.

Within the limits of the lodes from which specimens were obtained, there is no discernible arrangement of the minerals. From the specimens examined the minerals appear to occur independently

of any particular horizon and of the country rock, that is, of course, apart from the local minerals beryl, phenacite, orthoclase, garnet and zoisite.

Mr. Hobson has observed, however, that relative to wolfram, cassiterite definitely does increase in depth, and also that the sulphides increase in depth relative to the oxides; the same relation holds also laterally towards the more central part of the granite mass. Indeed, in the lower barren sections of the veins, sulphides occur to the exclusion of oxides.

PROCESS OF MINERALISATION.

Crystallisation and mineralisation within the magma. Mineralisation at Mawchi has taken place in two ways-impregnation of cassiterite throughout the granite, and formation of the mineral veins. Both forms of mineralisation did not necessarily take place at the same time.

The granite is usually medium-grained, but is penetrated by both aplite and pegmatite dykes. It is apparent that, in granite, late magmatic or deuteric changes were widespread, and have arisen by the introduction mainly of H2O and B203, with a certain amount of F and Li2O; it may be presumed that the discrete cassiterite scattered throughout these granites was introduced at the same time that the deuteric reactions were taking place.

The granite is a massive rock through which, after crystallisation, both solutions and gases would have found it extremely difficult to penetrate, except along joint-planes or other fractures. Apart from local wall-rock alteration there does not appear to be any increase in deuteric change within the granite close to the veins. These changes are too widespread to have any relation to the veins, and there is no justification for attributing the post-magmatic changes in the granite to the same solutions as gave rise to the lode minerals.

The deuteric changes in the granite are apparently so evenly distributed that they immediately suggest a source within the body of the granite itself. The process of crystallisation of the granite from oligoclase through quartz and orthoclase and eventually to biotite, left a final liquid high in volatiles distributed in the remaining

pore-spaces of the rock. This liquid consisted of H2O, B2Og, F,

Sn, LigO, CaO, CO2 and probably also a little quartz-the combinations in which these molecules may have occurred within the liquid need not concern us. So long as the total pressure of the surrounding rock exceeded the vapour pressure of this liquid at any particular temperature during cooling, these liquids would remain in contact with adjacent minerals, reacting with them, removing alkalies from the felspars and giving rise to tourmaline, muscovite, fluorite and calcite. The formation of joint-planes, etc., on further cooling of the granite, would permit the ready dispersion of any liquid that may eventually have remained.

It is a curious fact that although disseminated SnO2 is so widespread throughout these granitic rocks, wolfram has not been similarly detected. Wolfram does, however, occur in certain segregations which are commonly found in the granite. These segregations consist of cassiterite, wolfram (usually considerably replaced by scheelite), tourmaline and completely kaolinised felspars-one such segregation contained 22-6 per cent. Sn and 6-23 per cent. WO,. Other segregations consist of tourmaline and quartz with a little cassiterite. These segregations may be regarded as of the nature of local pockets of final magmatic liquid, and in which sometimes WO, was concentrated along with the other constituents-its replacement by scheelite in these segregations is comparable with the relation between these minerals in the lodes and is presumably a matter of stability at lower temperature.

In view of the abundance of wolfram in the lodes, its apparent absence in the granite compared with cassiterite requires explanation. It is assumed that the Sn-bearing molecule in the magma would be much more volatile than the WO, molecule, in other words its vapour pressure is higher. Towards the end of crystallisation of the magma the greater part of the remaining liquid would be concentrated in the lower part of the magma reservoir but, with any tendency to pressure relief, the more volatile constituents would rise through the crystallising granite, in which there is sufficient liquid-filled pore-space to permit the ready movement of gases. Thus, the final liquid in the pore-spaces of the crystallised granite may acquire a certain proportion of Sn, whilst the less volatile WO, is relatively absent from them, although concentrated in the deep-seated residual liquid.

It is not supposed that these deuteric changes took place in the granite only after most of it had crystallised. As the magma

crystallised downwards and inwards from its roof and sides, these changes would be taking place in the crystallised rock. It is possible that some of the cassiterite in the upper granite had formed before the granite at greater depths had crystallised.

The formation of aplite and pegmatite dykes may have some connection with these changes. After a considerable proportion of the upper part of the magma had crystallised out, but whilst the remaining lower magma had not changed in composition to any considerable extent, apart from becoming somewhat more acid, relief to pressure along fissures in certain zones would permit this magma to rise into the overlying granite, and, with rapid loss of volatiles consequent upon this relief to pressure, such intruded magma could locally become "dry" and rapidly crystallise out as aplite. These aplites are usually rather less kaolinised than is the granite.

The pegmatites belong to a later phase than the aplites, and are more closely connected with the mineral veins themselves. They have been formed from a very late wet" magma, in which crystallisation would be long-delayed, permitting a coarser-grained texture and the accompaniment of much muscovite. Local pockets of such magma may be left isolated in the consolidated rock and may eventually give rise to pegmatites in the granite; such pegmatites need have no roots and even small tin veins may be formed in this way. Other pegmatites may be derived from magma which has come from depth. It is probable that some of the constituents (such as felspar and quartz) of these pegmatites were picked up by the liquid in consequence of its reaction on the granite during migration or the liquid may actually recrystallise the granite in situ. There is a gradation between some pegmatites and the quartz lodes proper.

Finally the stage of formation of the mineral veins was reached. There appear to be all stages between the pegmatites and lode quartz, even within some of the veins themselves. Hence the veins must be regarded as representing a final phase immediately following and, in fact, a continuation of the pegmatite intrusions. The residual liquid was obviously extremely high in volatiles and its vapour pressure must have increased enormously in the final stages. Once fissures in the overlying granite tapped this liquid its injection must have been almost instantaneous. The extreme pressure of these liquids would assist in widening the vein walls, even apart from any replacement which may have taken place.