The isomorphous replacement of one element by another, in a pair of crystalline salts, is usually accompanied by a substantial change in the molecular volume. This progression is exceptionally well illustrated in the 'eutropic' elements K, Rb, Cs, and Cl, Br, I, which formed the subject of the first part of this paper. It was there shown that the differences produced by interchanging these elements stood in the same ratio in all the salt-series, though the size of the differences varied with the acid radicle. The object of the present note is to draw attention to the existence of similarly constant (though different) ratios for the salts of O, S, Se, Te, and for the salts of Ca, Sr, Ba. The spacings in these compounds have recently been obtained by X-ray determinations, in many cases due to u M. Goldschlnidt and his co-workers, so that it is possible to verify the rule of proportional differences in these series, although the results do not possess an accuracy so high as that of densities measured by the pyknometer.

Baron A. De Schulten, a political refugee from Finland, worked during the latter part of his life in the chemical laboratories of the Sorbonne in Paris. There he prepared in a crystallized form number of minerals and compounds analogous to them. These were determined crystallographlcally and compared in isonlorphous series. For example, having obtained crystals of artificial monetite, CaHPO4, he then proceeded to prepare the corresponding compounds in which strontium, barimn, or lead take the place of calcium, and arsenic the place of phosphorus. In this series he determined the crystallographic constants and optical data for CaHPO4, SrHPO4, BaHPO4, PbHPO4, SrHAsO4, and PbHAsO4. Crystals of CaHAsO4 were also obtained, but they were too small tbr measurement. Any of these compounds might be expected to occur in nature, but hitherto only monetite has been found as a mineral.

The object of this note is not to give elaborate descriptions of rocksections, but to call attention to tile fact that the old volcano of Nimrud, situated on the west side of lake Van in Armenia, was the centre of eruption of alkali-lavas presenting striking similarities to some of the alkali-rocks of the tlift Valley in East Africa.

To Dr. Felix Oswald, in his book on the geology of Armenia, we owe a detailed description of this ancient volcano, of which the vast crater, now half occupied by a deep lake, is more than five miles across. An interesting series of the lavas from the volcano was collected by Dr. Oswald and was presented by him to the British Museum. On looking over thin slices of these rocks under the microscope, I was at once struck by the surprising resemblance of some of them, with their phenocrysts of anorthoclase and scattered tufts of aegirine and the soda-amphiboles, riebeckite and cossyrite, to rocks in Prof. J. W. Gregory's collection from the Rift Valley which I had described.

The range of temperature within which tourmaline can arise in nature has not hitherto been determined. No synthesis of this mineral is on record, and experimental data are therefore lacking. On the other hand, rutile, brookite, and anatase have all been prepared synthetically, and the experimental data suggest their rutile is formed at the highest temperatures, brookite at temperatures considerably lower, and anatase at a point still lower on the scale. It is generally conceded that for rutile a pyrogenic origin is possible, and that brookite and anatase appear only as secondary minerals. The status of tourmaline as a possible pyrogenic mineral is regarded as doubtful by Clarke, who states that' in igneous rocks tourmaline seems to have been produced by fumarole action, and not as a direct separation from the magma'.

Bauxite is of twofold interest due firstly to its economic importance, and secondly to the scope it affords for scientific investigation. When these two reasons are taken into consideration, one can easily account for the enormous and rapid accumulation of literature concerning this subject. Any attempt to give a detailed account of the literature will only mean unnecessary repetition of what has been already dealt with at great length by several writers, of whom Anderson and Fox may be specially mentioned. The relation between bauxite and laterite is so intimate that a detailed study of the one is impossible without any reference to the other, and as such the description of and frequent reference to the latter should not be considered as any digression from the subject.

The rare mineral nadorite has hitherto been recorded from one locality only, namely, Jebel Nador in the province of Constantine, Algeria, where at one time it was found in considerable quantity associated with bindheimite, smithsonite, and galena in a deposit lying between lacustrine limestone and brown marl.

In the present notes I am able to record the occurrence of this mineral at the small antimony mine of Bodannon, St. Endellion, Cornwall, where I found a single specimen in 1907. Immediately after the reading of this paper, Professor A. Hutchinson of Cambridge most kindly sent me a specimen for examination which he had himself collected from the small dump at Bodannon in 1919, and this mineral, which had attracted his attention, also proved to be nadorite.

The old mine Settlingstones, situated three miles north of Haydon Bridge in the county of Northumberland, was originally worked as a lead mine, but, during the htst fifty years, as the vein has been explored farther west, it has attained far greater importance as the largest and most consistent producer of witherite in the British Isles. The Settlingstones main vein has a direction approximately north-east and southwest, and has been proved for a distance of nearly 1½ miles. The vein has varied in width from 3 feet up to 40 feet, the most productive witherite-bearing portion averaging 10 to 12 feet. The enclosing rocks are sandstones ('whetstones') and shales ('plate') of Carboniferous age, and dolerite (Whin-sil) which here attains a thickness of a little over 21 fathoms.

Six analyses of berthierite from various localities are quoted by Dana and show a varying composition. The analyses correspond approximately to the formulae FeS.Sb2S3, 3FeS.2Sb2S3, and 3FeS.4Sb2S3. It is considered by Dana that the variation is probably due to the impurity of the analysed material and that the composition of berthierite is probably FeS. Sb2S3. This view is supported by the examination and analysis of a specimen of berthierite from Xullarnanna, 18 miles north of Inverell, Now South Wales.

The specimen was kindly lent by Mr. G. W. Card, Curator of the Mining Museum, Sydney, for comparison with certain minerals under investigation, and measured approximately l x l x O-7 cm. It was sliced and one portion was polished for examination under the reflecting microscope while the other portion was used for the chemical analysis.

By the publication of his paper on 'The source of the purple zircons in the sedimentary rocks of Scotland', Dr. Wm. Mackie has done his fellow-workers on the petrology of British sedimentary rocks a signal service, He has proved that this distinctive purple (or sometimes rosecoloured) variety of zircon is relatively abundant in the Lewisian gneiss. The variety occurs usually as rounded or ovoid grains, but occasionally with crystalline form ; it is often finely polished, as if by wind, and is sometimes zoned, with a darker centre, or displays colour of varying intensity; it is often definitely, although but slightly, pleochroic, and usually possesses high bireffingence. These characteristics at once bring to mind grains of zircon frequently met with in sedimentary rocks and recorded by numerous other observers.

Of the garnet group of minerals, two members only appear to have been synthesized from artificial melts, viz. melanite (L. Michel, 1892) and spessartine (L. Bourgeois, 1883, and A. Gorgeu, 1883); though grossular is once reported by E. S. Shepherd, G. A. Eankin, and F. E. Wright (1909) as being produced by reaction under pressure of calcium orthosilicate and aluminium chloride. In dry melts, pyiope is represented by a cordierite-forsterite-spinel assemblage, but no data are yet available to indicate the dry melt assemblages representing almandine. The home of almandine in sediments is the dynamically metamorphosed pelite, and its widespread regional occurrence in these crystalline schists implies a ready synthesis under the influence of shearing stress.

Some years ago the writer read a paper on the determination of the optic axial angle of a crystal by the Becke method. This proposed method was to record the trace of the optic axial plane when horizontal, the position of one axis, and the extinction-direction. From these data and the established constants of the microscope, the position of the second optic axis may be found, as the extinction bisects the angle between the centre of the pole of the section and the two optic axes. Hence, the optic axial angle may be determined.

Considerable errors arise in these determinations owing to defects in the optical system of the microscope, and also to the inability to determine the exact position when the trace of the optic axial plane is parallel to the E-W. lines in the microscope.

Bolivian minerals have been sent tbr examination and preservation in the British Museum oolleetiml at various times since 1894 by my old school-fellow Mr. Malcolm Robmts, A.R.S.M., now General ganager of the Compagnie Aramayo de Mines en Bolivie. Several of these have been described in the pages of this magazine (see vol. 14, p. 309). Although he has industriously collected Bolivian minerals for over thirty years he is only now rewarded by the discovery of a species new to science. The material which he has been good enough to entrust to me for description was recognized by him as being probably something new, and this was confirmed by a chemical analysis made by Mr. T. Burns MeGhie in the assay office at Quechisla in Bolivia.

Willemite is a mineral of comparatively rare occurrence, but it has long been mined in New Jersey as an important, ore of zinc (Zn2Sio4, containing Zn 58.6 %), and it would seem also to be abundant in Rhodesia. In the standard mineralogies of Dana (1892) and Hintze (1889) it is mentioned from only four districts. As there is some confusion in the statement of these few localities, and further as several others have been discovered during recent years, a complete list of localities for this mineral has been compiled from the literature; and some others represented by specimens in the British Museum collection are added, inchding the African occurrences now to be described.

The following notes are intended to bring to the notice of workers on heavy mineral grains the possible confusion between blue tourmaline, purple tourmaline, glaucophane, lepidolite, andalusite, pleochroic sillimanite, and dumortierite. Dumortierite (when blue) is very liable to be passed over as the more common mineral glaucophane, the optical properties of the cleavage-flakes of these minerals being similar in several respects.

Having seen a grain of dumortierite from the Wealden of Surrey, kindly lent me by Mr. H. A. Hayward, I was impressed by the remarkable nature of its pleochroism. A day or so later, while examining some concentrates of the Blackheath Beds of Shirley, Surrey, belonging to Mr. G. M. Davies, I came across a single grain which he had described as glaucophane. The pleochroism seemed too intense for a grain of glaucophane of that thickness, and it occurred to me that the tint observed, when the grain was in the position of maximum absorption, was exactly that of the dumortierite grain.

During an examination of zinc ores I had occasion recently to employ methods for the identification of smithsonite, and was surprised to find how scanty and indefinite are the data of its physical properties as stated in text-books. Most of the smithsonite known occurs either compact or as very small crystals many of which are of a composite nature. In this connexion Dr. L. J. Spencer recently acquired for the British Museum some very fine crystals from the Rhodesia Broken Hill mines, and he suggested that this material would be suitable for a detailed examination. These consist of two specimens obtained in 1921 from Mr. W. E. Harris [B.M. 1921, 769] and in 1923 from Dr. O. L. Brady [B.M. 1923, 619] respectively, and are portions of the same mass. Smithsonite crystals were first described from this locality by Dr. Spencer in 1908, but these crystals were minute and showed only the forms v (201̄) and s (223̄).

In ordinary circumstances crystal-angles are measured by means of light reflected from the crystal-faces. In the partial or complete absence of faces the determination of axial ratios and angles may be impossible. In a recent investigation of certain derivatives of glyoxaline, an example of this was found. It was, in fact, the substance glyoxaline itself, which always occurred as long prisms having four faces in the prism-zone; the ends of the prisms were jagged and rough, and no siglls of optically reflecting-planes could be found on these ends. In such a case, X-ray methods of investigation are now available, and the problem in question is the determination of the dimensions of the unit-cell, not the complete X-ray inveatigation of the structure.

An attempt was made to determine the true symmetry of the cleavage flakes of aramayoite by means of the Laue photographic method. Since X-ray diffraction takes place within the crystal, the lack of welldeveloped faces other than main cleavage faces was no disadvantage. One feature, however, which did prove a difficulty, was the pliability of thin clevage flakes referred to by Dr. Spencer. It was essential that the whole of the main cleavage face upon which the X-ray beam was incident should be strictly perpendicular to that beam. In order to ensure this, only those flakes were used which gave a single optical reflection, and two distinct methods of setting were resorted to. In the first case the slit (circular aperture) was used vertically ; the flat end of the slit system was ground truly perpendicular to the slit itself and was thus horizontal.

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.