We are apt to regard selenite or transparent gypsum as a useless mineral, apart from the very small quantity used in plates for microscope work. It is the object of this paper to show that from the first, to at any rate the seventeenth century in Europe and to the first half of the nineteenth in South America, considerable use was made of it. For the classical period the argument is based upon the identification of lapis specularis as mainly, if not entirely, selenite. Most translators render the name as talc or mica, the first word probably a vulgar error for the second, while some admit gypsum to have been included; Bailey in his notes on Pliny considers that two passages suggest gypsum but gives the preference to mica, while Zeitler claims nearly all the passages referred to below for this mineral. Pliny is, of course, our chief authority: I give the relevant passages in Philemon Holland's translation as more in keeping with the greater part of this paper, adding notes and the original Latin where necessary.

In 1911 an interesting find of barium-felspars was made in the Benallt manganese mine near Rhiw in the Lleyn peninsula of Carnarvonshire. A brief preliminary note on the occurrence was given by Mr. Arthur Russell in the same year. At the same time he very generously handed over to me for crystallographic and optical description a selection of the material he had collected, while he himself made a start on the chemical analysis of the material. Chemical analyses have now, at last, been made by Miss Hilda Bennett in the Mineral Department of the British Museum, and it therefore seems time to collect together the large volume of notes on the work done in 1911.

Although the Mendip Hills, as a district, have from time to time received the attention of geologists, details of the mineral occurrences are, on the whole, remarkably scarce. Mining was formerly carried on on the Hills for many centuries, but mineralogy as a science had not been sufficiently developed at the time when the mines were most active, and little interest was taken in the minerals other than those suitable or required for commercial and industrial purposes. Even in later years, when, in the course of treating old refuse and tailings left by earlier miners and of various attempts to resuscitate the mining industry, much material must have been available for examination, little attention seems to have been paid to it, or to the numerous quarries that were opened in the Carboniferous Limestone.

During the last few years large collections of rock specimens from the Toro-Ankole volcanic fields, east and south-east of Ruwenzori, have been sent to Durham by the Geological Survey of Uganda for petrological investigation. The specimens recently studied include representatives of a suite of highly potassic ultrabasic lavas. Some of these resemble melanocratic olivine-nephelinites or olivine-rich ankaratrites, but the nepheline-like felspathoid which they contain, instead of being nepheline, is a polymorph of KAlSiO4. For chemical reasons this mineral was at first thought to be kaliophilite. Other types, closely related, contain leucite, biotite, or melilite, in addition to the supposed kaliophilite.

It is well known that in the Schroeder Van der Kolk method of comparing the refractive index of a crystal grain with that of the material in which it is mounted, the effect can be reversed by moving the microscope sub-stage condenser, and that the position of this condenser must therefore be carefully controlled if ambiguities in the interpretation of the observations are to be avoided. This fact is mentioned in the majority of text-books and is referred to in a general manner by Wright and by Winchell. No clear explanation of the reversal is offered, however, and students are consequently inclined to forget this important limitation in the method or, alternatively, to forget in which sense the movement of the condenser influences the effect: unless a control experiment is invariably performed errors may therefore result.

The presence of much native iron in Great Namaqualand was heard of in 1836 by the traveller Sir James E. Alexander, but he himself did not see any of the masses, and the material he acquired consisted only of small fragments that had been detached by the natives for pointing their weapons. An approximate chemical analysis of this material, made by Sir John Herschel, proved it to be meteoritic. A fragment weighing 2 grams which Alexander presented to the Geological Society of London in 1838 was transferred to the British Museum collection of meteorites in 1911. This appears to be all of his material that has been preserved.

Various other vague reports and travellers' tales have mentioned the presence of large masses of iron beyond the old mission station of Bethany, between there and Beersheba, and on the east side of the Great Fish River. A critical summary of these accounts was given by Sir Lazarus Fletcher in this magazine.

In 1908, at the close of the Franco-British Exhibition at Shepherd's Bush, London, a series of rock phosphates from various localities in South Australia was transferred to the Mineral Department of the British Museum as a present from the Government of South Australia. One large block from Wait's quarry at Noarlunga, 20 miles south of Adelaide, showed cavities lined with crusts of minute crystals. These crystals were determined goniometrically and optically in 1908 to be orthorhombic, and qualitative chemical tests showed the presence of Al, Mg, P, F, and H2O (K was missed). This evidently represented a new mineral, and it is only now that the work has been successfully completed by the three junior authors. In the meantime, however, the same mineral was discovered in Western Australia by the late Dr. E. S. Simpson in 1932, and named minyulite, from Minyulo Well. This he determined optically to be orthorhombie, and the formula was given as KAl2(OH,F)(PO4)2.3½H2O or 2K(OH, F).4A1PO4.7H2O. Crystallographic and X-ray data are now given for this species.

Augelite was originally described from Sweden by Blomstrand (1868) who obtained the composition 2Al2O3.P2O5.3H2O from a massive mineral with cleavage in three directions. No further description of this inadequately defined mineral appeared until Prior and Spencer (1895) gave their complete and exact characterization of well-crystallized material from Machacamarca, department Potosi, Bolivia, supplemented by observations by Spencer (1898, 1907) on further finds at Tatasi, department Potosi, and at Oruro, Bolivia.

Lemmon (1935) first recognized augelite as cleavage masses in the commercial deposit of andalusite at White Mountain, Mono County, California, and gave determinative observations which established the identity of the mineral. In this locality Kerr (1932) had noted corundum, diaspore, pyrophyllite, alunite, lazulite, and rutile, associated with the andalusite. In addition to these minerals Professor Joseph Murdoch of the University of California at Los Angeles, has found small, white, tabular crystals of apatite in a specimen carrying augelite. Crystals of augelite from this deposit, with a number of new forms, were mentioned by Pough (1936) in an inconspicuous notice which we did not see until our work was nearly done.

Nevil Story-Maskelyne described in 1870 the mineral constituents of the Bustee meteorite, which was observed to fall in 1852 near Bustee about 45 miles west of Goruckpur, India. The stone, weighing 1429 grams, now preserved in the British Museum collections, was found to consist chiefly of enstatite and diopside, and two constituents proved to be new minerals. Pale, chestnut-brown spherules of calcium sulphide, CaS, Story-Maskelyne named oldhamite and golden-yellow octahedra found embedded chiefly in the oldhamite he named osbornite. One side of the meteorite has been ground to reveal the nodules of oldhamite and a coloured lithograph in W. Flight's book ‘A chapter in the history of meteorites’ reproduces the appearance of the chief mineral constituents very well. The minute octahedra of osbornite are easily visible with a lens in the meteorite itself.

If α, β, and γ are the three angles between a line passing through the origin and the three rectangular co-ordinate axes, then the equation connecting them is as follows:

cos2α+cos2β+cos2γ = 1.

This is the law of direction-cosines of a line.

Substituting (cos 2α+1)/2 for cos2α, &c., we get:

cos 2α+cos 2β+cos 2γ = −1.

During the course of a general investigation by one of us (K. L.) of diamonds from various sources, using X-ray analytical methods, the inquiry arose as to whether so-called artificial diamonds had ever been examined in this way. It was found that there was, in the Mineral Department of the British Museum, a glass slide bearing 12 minute specimens, labelled as being diamond, artificially prepared and presented by Mr. J. B. Hannay in 1880, presumably the remainder of those investigated by Prof. N. Story-Maskelyne, and referred to in a letter to the Editor of The Times which appeared in that paper on February 20, 1880, as follows:

Sir,—A few weeks since I had to proclaim the failure of one attempt to produce the diamond in a chemical laboratory. To-day I ask a little space in one of your columns in order to announce the entire success of such an attempt by another Glasgow gentleman.

The Skaergaard intrusion of Kangerdlugssuak, East Greenland (68° 10′ N., 31° 40′ W.) contains a series of gabbros and related rocks—the layered series—which has been interpreted as the result of fractional crystallization of a fairly normal gabbro magma (Wager and Deer, 1939). Analysis of these rocks and of some of the constituent minerals by Dr. W. A. Deer has shown a considerable and systematic variation in the major constituents (1939, table XXXVI and pp. 67–84) and it was considered therefore that it would be of value to determine the variation in some of the rarer elements.

The work of Goldschmidt and his collaborators on the distribution of trace elements in minerals (summarized in Goldschmidt, 1937) combined with our recently acquired knowledge of many crystal structures has made clearer the factors controlling the distribution of the trace constituents in minerals.

The Widmanstetter figures on a general plane section of a meteoritic iron of the octahedrite class consist of kamacite bands running in four directions, the traces of the four planes of the octahedron, while tessellated octahedrites show three more sets of bands, the traces of the cube planes. The angles which these bands make with one another give data which are in principle sufficient to define the crystallographic axes of the iron, but in practice there does not appear to be any simple direct method of effecting this definition. In the following, two simpler cases are first discussed, followed by a method for the general case somewhat shorter than any published method I am aware of, and by methods for orientation from the Neumann lines and from inclusions.

In a previous paper describing the geological and petrographical details of the adinoles of Dinas Head, short references were made to the common occurrence among the adinoles and associated rock-types of the magnesia-rich tourmaline, dravite. It is proposed now to describe these tourmaline-rich rocks in more detail and to discuss the question of their origin.

Adinoles with accessory tourmaline occur wherever true adinoles are developed at Dinas Head, although tourmaline is characteristically absent from types 2 a, b, c, and 3 [1, pp. 318–323]. Bands with up to 30 % tourmaline occur interbanded with normal adinoles of type 1 and with certain spilosites.

The determination of ferrous iron in minerals, such as staurolite and kornerupine, which cannot be effectively dissolved by dilute sulphuric and hydrofluoric acids has been made possible on the macro-scale by the work of H. P. Rowledge; the mineral is fused with sodium fluoride and boric oxide at 900°C. in a sealed pyrex glass tube, the fusion cake dissolved in hot dilute sulphuric acid in a current of carbondioxide, and the solution titrated with potassium permanganate. We have met with no difficulty in applying the method on the macro-scale, but when only small amounts of a mineral are available several difficulties arise, and modifications become necessary.

The air present in the sealed tube of a Rowledge fusion, using the quantities and dimensions he specifies, is already sufficient to oxidize about 3 to 5 mg. of ferrous oxide. But when the weight of the sample is reduced to 10 mg., it is not possible to reduce the dead-space in the tube in proportion ; it can hardly be reduced below about ¼c .c. which is sufficient to oxidize 0·3 mg. or 3% FeO and is likely to be more effective in this direction than on the larger scale, since the surface of the fusion is proportionally greater.