he Courses of Instruction in these Departments are arranged for Students preparing to become Manufacturing or Analytical Chemists, Metallurgists, or Science Teachers. The Courses in Chemistry qualify for the Examinations of the Institute of Chemistry, and of the Universities of Glasgow, Edinburgh, and London.

SESSION 1910-11 begins on TUESDAY, SEPTEMBER 27, in the New Builings recently erected for the College The ENTRANCE EXAMINATION begins on MONDAY, SEPTEMBER 19.

CALENDAR (by post Is. 4d.), containing Regulations for the

DENTISTRY,

Leading to Degrees and Diplomas.

The Session 1910-11 commences October 3rd, 1910. ALL COURSES AND DEGREES ARE OPEN TO BOTH MEN AND WOMEN STUDENTS.

In the Medical School Courses of Instruction are arranged to meet the requirements of other Universities and of Licensing Bodies.

Graduates, or persons who have passed Degree Examinations of other Universities, may, after one year's study or research, take a Master's Degree.

SYLLABUSES with full information as to Lecture and Laboratory Courses. Fees, Regulations for Degrees, Diplomas, &c., Exhibitions and Scholarships, will be sent on application to the Secretary of the University.

UNIVERSITY OF DURHAM.

ARMSTRONG COLLEGE,

Newcastle-on-Tyne.

Principal-W. H. HADOW, M.A., D.Mus.
SESSION OF 1910-II.

MATRICULATION EXAMINATION,

SEPTEMBER 20th to 23rd.

Particulars of Curricula for University Degrees and College Diplomas in ENGINEERING, ELECTRICAL ENGINEERING, NAVAL ARCHITECTURE, MINING, METALLURGY, AGRICULTURE, PURE SCIENCE, ARTS and LETTERS, and of Fellowships, Scholarships, and Exhibitions, on application toF. H. PRUEN, M.A., Secretary, Armstrong College, Newcastle-upon-Tyne.

The Proprietors of the PATENT No. 2603 of

1898 for "IMPROVEMENTS IN THE PREPARATION OF EXTRACTS OF TANNIN," are desirous of entering into arrangements, by way of Licence and otherwise, on reasonable terms, for the purpose of exploiting the same and ensuring its full development and practical working in this country. All communications should be addressed, in the first instance, to HASELTINE, LAKE, and Co., Chartered Patent Agents and Consulting Engineers, 7 & 8, Southampton Buildings, Chancery Lane, London, W.C.

SULPHUROUS

ACID and SULPHITES.

Liquid SO2 in Syphons, for Lectures, &c.

PHOSPHORIC

ACID and PHOSPHATES.

obtained from the SECRETARY.

TO ADVERTISERS.-Particulars of Scale of Charges

and Contract terms for extended periods, obtainable on

for all purposes.

Manager, CHEMICAL NEWS Offices, 16, Newcastle Street, Farringdon Street, London, E.C.

Stratford, London, E.

CHEMICAL NEWS, Aug. 12, 1910

}

73

SINCE the publication of the first part of my paper on Scandium (Phil. Trans., A, ccix., 15, March 4, 1908), G. Eberhard, of the Astrophysical Observatory, Potsdam (Sitzungber. Kgl. Preuss. Akad. Wissensch., 1908. xxxviii., 851), reasoning from the fact that the strongest lines of the scandium spectrum are observed in the spectra of stars in diverse stages of development, has come to the conclusion that scandium must be universally distributed on the earth. Investigating the arc spectra of 366 minerals and rocks, he obtained the remarkable result that scandium in small quantities is actually one of the most widely distributed earth elements. He shows that it occurs most frequently in zirconium minerals-in beryls, titanates, columbites, and titanocolumbites of the rare earths-also in micas; and, finally, that specimens of wolframite and tinstone from Saxony and Bohemia contain scandium in sufficient quantity to make its extraction advantageous. These results of Eberhard were quickly verified by Prof. R. J. Meyer, Berlin University (Zeit. Anorg. Chem., Ix., 134, November 17, 1908; and CHEMICAL NEWS, xcix., 85, 97, February, 1909), who, in a preliminary paper, has described experiments on extracting scandia from the Zinnwald wolframite. Prof. Meyer concludes that this wolframite contains from 0.14 to 0.16 per cent of rare earths, and that these rare earths contain o 30 to 0:33 per cent of scandia. Thus the original wolframite would contain about o'15/100 x 0.315/100 = 0.04725/10000, or about 0.0005 per cent of scandia. Prof. Meyer has worked out two methods of separating scandia from the wolframite rare earths :-(a) Precipitation with hydrofluoric acid from an intermediate product in which the scandium has been concentrated by an oxalic acid precipitation, and (b) precipitation with hydrofluosilicic acid, or sodium silicofluoride in acid solution. The raw scandia prepared by the above methods still contains small quantities of yttria and ytterbia, particularly if method (b) is used. Prof. Meyer says that his scandia, when freed from cerium and yttrium earths, has the atomic weight 45 to 46, and still contains about 1 per cent of thorium. (A mixture of 99.25 parts of scandium with 0.75 part of thorium would have an atomic weight of about 45'5).

On reading Prof. Meyer's paper I examined the felspar and mica which were associated with the parcels of wiikite procured by me from Finland. The felspar contains a small percentage of rare earths, but no trace of scandia could I detect on examining their photographed spectra, the dominant line, wave-length 3613-984, being entirely absent.

The mica was found to contain about 1 per cent of rare earths, and in the photographed spectrum the dominant lines of scandium are plainly visible. From a comparison of the spectra with those taken with the rare earths from wiikite I should estimate the scandia at about half that in

the earths of the best black wiikite.

A specimen of wolframite from Zinnwald was found to contain less than 1 per cent of rare earths, and in the photographed spectrum of these earths, the dominant lines of scandium were strong, accompanied by only the faintest trace of yttrium or ytterbium lines.

* A Paper read before the Royal Society, January 20 and June 2, 1910. + The descriptions and analyses of the following salts were communicated to the Royal Society on December 14, 1909, as now printed: -Scandium aurochloride, platinocyanide, iodate, sulphite, malate, malonate, tartrate, racemate, lævotartrate, and mesotartrate.

Through the kindness of a friend, Dr. Wahl, of the obtained the following authentic specimens of minerals from the locality whence my wiikite was mined :

No. 1. Felspar from the pegmatite vein carrying monazite and the scandium-wiikite. Lokasaari.

No. 2. Monazite from Lokasaari.

No. 3. Black wii ite from Lokasaari, Impilaks.

No. 4. Brown wiikite from Lokasaari.

No. 5. Wiikite from Lokasaari, another specimen.
No. 6. Mica from Hunttila, Impilaks. (Probably from

the wiikite quarry).

No. 4, brown wiikite, contained 23'77 per cent of rare earths; the spectrograph showed only a small amount of scandium.

No. 5, wiikite, contained 24.85 per cent of rare earths; the photographed spectrum showed considerable quantities of scandium.

No. 6, mica, contained o 63 per cent of rare earths, which, on spectrographic examination, contained an amount of scandium about equal to that in the rare earths in Nos. 3 and 5.

I have also to thank Dr. A. Hiorth, of Kristiania, for a typical collection of 23 specimens of rare minerals from Norway; the results of examination of most of these specimens have already been given.

The scandia used in the preparation of some of the salts here described was not absolutely pure. Chemically, no other earth could be detected in it, but the spectrograph revealed traces of yttria and ytterbia. These traces could have been removed by one or more operations, but I thought it advisable to leave them in, for the following reasons:In each operation of purification some loss is unavoidably incurred, and when chemical reactions are insufficient to find the other earths it is not worth diminishing my lessening stock of scandia for the sake of academic purity. The chief reason, however, for leaving these traces in is that they might afford evidence of a difference of behaviour between one earth and another in the presence of some of the acids used. After each quantitative determination the scandia was dissolved in acid and a spectrogram taken to see if yttria or ytterbia were present. The residual earth was then collected from the mother-liquor and a photograph taken of its spectrum. A comparison of the pair of spectra shows at once if any separation has been effected between the earths present. When separation is apparent further experimentation on a larger scale is reserved to a

future occasion.

SCANDIUM IODATE, Sc(IO3)3 + 18H20.

Scandium iodate is best prepared by the metathesis of a soluble scandium salt with ammonium iodate. It forms a After being well washed and dried in the air it contains 18 white crystalline powder almost insoluble in water. molecules of water. The water is held loosely; long drying in a desiccator, or exposure to a temperature of 100 and slightly higher, drives some of it off; only by repeated trials with slight alterations of temperature have I obtained certain definite hydrates. There are indications of other hydrates, but I have not attempted to prepare them.

Analysis.

1. 15:483 gr. of 18-hydrate scandium iodate, dried in the air at the temperature of about 15°, were dissolved in hydrochloric acid and heated until chlorine ceased to be evolved. It was then evaporated to

5

1.63

0'77

THE CORRELATION OF ROCK AND RIVER-WATER ANALYSES.

By H. S. SHELTON.

THE analysis of rocks, and of the substances dissolved in river-waters, tends to become a mechanical process, and to be undertaken for diverse and special purposes, without due regard to the broader scientific aspects. One point in particular has not received the requisite attention. It is not sufficiently recognised among analysts that, between the analyses of rocks and of the dissolved matter in riverwaters there must be an intimate correlation. The solid matter in solution is, in the main, obviously derived from the erosion of the rocks. If therefore any soluble substance widely distributed in rocks fails to show a due equivalent in river-water analyses, the matter calls for investigation. If, on the other hand, any widespread constituent of riverwater, not to be accounted for in other ways, is found only in minute proportion in the rocks, there must be serious errors either in rock analyses, or in water analyses, or in both. Although some attempts have been made to correlate river-water analyses with the geologic formation of the districts in which the analyses have been taken, there appears, as yet, to have been no serious effort to follow the dissolved constituents of river-waters to their origin in the rocks (see Hanaman, "Analysis of Bohemian Waters," quoted by Clarke; "Data of Geochemistry," P. 79).

The present paper is not concerned with the correlation of particular formations with particular rivers, but with the broad average analyses of river-waters. It is a preliminary survey of the ground, in order to see whether every sub stance which appears in sensible proportion in water analyses can be identified with a corresponding widespread constituent in the rocks. A good statistical average, which will serve the purposes of this paper, has been compiled by Sir John Murray. His average of the substances in solution in the great rivers of the world, though it was made more than twenty years ago, and is open to criticism in detail, is the best extant, and any errors may contain are probably confined to minor details and proportions (Scottish Geogr. Mag., 1887, " Data of Geochemistry," p. 88). This estimate, reduced to ionic form by Prof. Clarke, is quoted Per cent of inorganic solid.

below.

MgO CaO Na2O K20

Water

CO2
S..
Combined oxygen
Other minor con-
stituents (b).

Total

(a) Expressed as 0·64 SO3.

(b) Containing a large proportion of titanium.

To show the relation of these tables to the problems of erosion and of r ck solution, it is necessary to state that sandstone, shale, and limestone are estimated to occur, in the area subject to erosion, in the relative proportions of 80, 15, and 5. The proportion of igneous rock in the part of the earth's crust subject to erosion is somewhat more uncertain, but it is usually estimated as about one-quarter of the sedimentary. The presence of metamorphic formations, which may arise both from igneous and from sedimentary rocks, makes exact calculation impossible. the same reason, a general average of all rocks subject to erosion cannot readily be made; but a study of the tables will show that, within the limits of statistical error, this differs only from the composition of igneous rock in a deficiency of sodium and an excess of carbon, both of which differences are readily explained.

For

In a preliminary rough survey, which is all that is possible in the present paper, we must first of all note the constituents of river-waters which can have any origin other than that of solution of the rocks. The proportion of carbonate is undoubtedly largely due to atmospheric carbon-dioxide, which is continually renewed volcanically, and to the special solubility of calcium carbonate in carbondioxide solution. The latter remark also applies to the remarkable percentage of calcium. The nitrate is prob ably over-estimated; but what proportion exists we can refer to decaying organic matter. The chlorine, a portion of the sodium, and a minute proportion of magnesium and of other salts, is due to the precipitation of sea-salt in the

rain.

Making these necessary allowances, a rough comparison of the two tables will show that nearly all the constituents found in sensible proportion in river-water-calcium and magnesium carbonates, iron and alumina, sodium potassium, and silica are found in large proportion in the igneous and sedimentary rocks. Moreover, no considerable constituent of the rocks is missing from the riverwater table.

One substance found in river-water analyses I have already discussed at some length in a previous paper (CHEMICAL NEWS, 1909, xcix., 253). To those remarks, and to the suggestions contained therein, I will merely add the expression of opinion that, in the ordinary water analysis the sodium is over-estimated. Theoretical reasons for that opinion are confirmed by the fact that the statistical abstracts which record only the work of well-known and careful analysts (cf. Dubois, Proc. Amsterdam Acad., 1904)