The two exceedingly interesting specimens which form the subject of the present notes came under my observation in the following manner. In 1905 there was sold by public auction in London an old and very fine collection of minerals, mostly Cornish, which had been formed by John Hawkins, F.R.S. (b. 1758, d. 1841), of Trewithen, Probus, Cornwall; a gentleman well known in his day both as a scientist and traveller. The major part of the collection unfortunately left this country, having been purchased by the dealer F. Krantz of Bonn. A few lots were, however, bought by myself, and one of these contained the first specimen to be here described.

The separation and estimation of small quantities of manganese in the course of exact analysis of igneous and metamorphic rocks has long presented a difficulty. The element is almost universally present, in amounts varying from a few hundredths of one per cent. to over one per cent., the higher figure being reached in the basic rocks, especially in those rich in ferromagnesian minerals. In the ordinary course of analysis, if no special method be adopted to separate and remove manganese, this element tends to be distributed between the 'ammonia precipitate' and the lime and magnesia. It has been shown by G. Steiger that when the usual methods are employed, the manganese tends to be distributed in a quite haphazard way among the three precipitates; no definite proportion is retained by any one of them, although as a general rule most of the manganese is found with the magnesia and least with the lime.

Experience with the usual types of glass pyknometers long ago convinced the writer that these instruments as ordinarily used are subject to considerable errors and may be quite unreliable when dealing with small quantities of material, It is true there are forms intended to be used in conjunction with a very precisely regulated water-bath which are capable of giving results of high accuracy, but this method of determination at constant temperature involves too much time and trouble for routine mineralogical work. But no matter what the design of glass pyknometers, and their variety is legion, in actual practice they suffer in common from tile defects of the material of which they are constructed.

The heavy-solution method is inconvenient ; it cannot be used for the heavier materials and there is often danger of contaminating the material which may be intended for analysis.

The general outline and weight of the body of the instrument are devised to ensure perfect rigidity and steadiness during use, all necessary adjustments, such as the focusing of the rotating eyepiece, movement of the revolving stage, reflector, or screen, being quite independent of the optical system; which, being contained in the main body in a dust-proof receptacle, entirely prevents any vital part of the refractometer getting out of adjustment, unless deliberately tampered with.

The eyepiece, which by rotation is capable of giving a wide variation of focus, is set at the most convenient angle to ensure easy and comfortable observation.

The forward projecting portion attached to the main body is rectangular in shape, and hollow, the upper surface forming a platform in which revolves a circular plated stage with a milled edge capable of being turned about a vertical axis through 360°, with an engraved scale marked at every 45° to enable the observer to read with accuracy the least and greatest refractive indices of all doubly refracting stones.

A very fine and unusual crystal of sperrylite, recently presented by Mr. G. H. Beatty of Johannesburg to the mineral collection of the British Museum, came from a new adit on the Tweefontein farm (no. 1083), about 10 miles NNW. of Potgietersrust, Waterberg district, Transvaal. It had been found embedded in copper-stained limonite in crush-zones of banded ironstone underlying norite. The limonite matrix is presumably an alteration-product of sulphides (pyrrhotine, pentlandite, and chalcopyrite). The banded ironstone is an altered sediment belonging to the dolomite series, and is penetrated by apophyses from the basic margin of the Bushveld igneous complex.

In the metamorphic aureole of the Falmouth. granite, the greenstones I which are largely referable to epidiorites form an important and interesting assemblage of rocks. These are lower Palaeozoic in age and are pre-granite and have suffered from metamorphism as all effect of the granite-intrusion in post-Carboniferous times.

The chief mineral constituents, viz. hornblende and plagioclase, are quite normal for a rock of this nature except that the former is clearly a derivative of augite, many relics of which are still preserved ; among the most important accessory minerals are biotite, chlorite, epidote, iron-ores, quartz, augite, axinite, and garnet. One may pass over some of the accessories as being quite usual, being partly primary and partly secondary in origin, but biotite is one of those the presence of which requires to be specially explained.

The spots of a Laue crystal-photograph may be regarded as due to the reflection of X-radiation by the planes of the crystal and the interpretation of these photographs involves the determination of the index of the plane by which any given spot is produced. This 'interpretation can be readily accomplished by two distinct though related methods. We may, in the first place, utilize the stereographic projection and determine the indices of the planes by means of intersecting zone- circles. This method, the earliest in point of time, was explained by W. L. Bragg in 1913. In recent years, however, it has become usual to employ for this purpoee the gnomonic projection which affords a solution in some respects impler and more direct than the stereographic method used by Bragg, and detailed illustrations of its application in this field have been given by R. W. G. Wyckoff, and E. Schiebold.

The crystal used in the first part of this investigation, an untwinned specimen from Ceylon in the British Museum collection, was measured goniometrically and described in detail by Dr. G. F. Herbert Smith, who gave as its crystallographic constants :

Density (p) 5.82 (grs./c.c.). Monoelinic prismatic,

Owing to the comparatively largc size of the crystal and its high absorptive power it was unsuited either to the 'rotating crystal' or Laue photographic methods of X-ray examination, Reflections from the faces were observed on the Bragg ionization spectrometer and spacings measured with an accuracy of about 2 per cent.

In a proper recently published by Prof. Vegard is contained an analysis of the structure of the zircon group of crystals, which belong to the tetragonal holohedral class, the axial ratio being approximately a : c = 1 : 0.656.

As the present work was completed when the above paper appeared, it is interesting to find that the principal deductions concerning the structure of zircon, ZrSiO4, are confirmatory. The parametral values are, however, slightly different. This difference, though small in itself, is found to introduce larger discrepancies between the observed and calculated intensities of reflection.

Lamellar twinning is of frequent occurrence in corundum, as evidenced by rulings on the surfaces and by the well-marked pseudo-cleavages (or partings) parallel to the faces of the primary rhombohedron and also parallel to the basal plane. Distinctly developed juxtaposition and interpenetration twins would, however, appear to be rare, for only a few isolated examples have hitherto been described.

There are two twin-laws for corundum crystals, the twin-plane being either r (101̄l) = (100) or c (0001) = (111). The former of these is the one considered in the present note.

Gold and silver have been detected not only in the peculiar facies of the granite occurring on Bittleford Down, but also in the normal grey biotite-granite of numerous tor- and quarry-localities; several hand-specimens of granite showing minute flakes of visible gold have been collected.

The modes of occurrence of gold on Dartmoor, the environment of the granite from which specimens showing visible gold have been taken, and the gold-silver content of the granite generally have been investigated in some detail, with the results described below.

Small portion (B.M. 1921, 274) of the Vaalbult meteoric iron, weighing about 184 grams, was kindly placed at the disposal of the writer for examination by the late Dr. L. Peringuey, Director of the South African Museum, Cape Town. According to information received from him, this meteorite is known as the 'Vaalbult' siderite, and was presented to the South Afican Museum on April 22, 1918, by Mr. S. Grove, Member of the Provincial Council of the Cape for the Prieska Division, who obtained it from Mr. Japie Greeff, the owner of the farm on which it was found, probably early in the same year. The original mass weighed about 26 lb. and was very deeply and broadly pitted. In external appearance it is somewhat similar to the large (1,210 lb.) Ratteldraai iron from the Kenhardt Division, more than 60 miles distant, found in 1909. A cast of the iron sent by Dr. Peringuey shows that it is of very irregular shape.