3

1

the surfaces of W and Q, are slightly convex between centimeters 15 and 45, while Q is convex to a less degree between centimeters 70 and 85. The surface of Q, is nearly plane when the cathetometer is supported at its neutral points.

RELATIVE ERRORS OF THE SUBDIVISIONS.

A plus sign indicates that the measured space is too short.

The relative errors of the 10 centimeters of the first decimeters of

Q, and Q, were found to be as follows.

3

of W and Q, were not investigated.

The centimeter subdivisions

The errors of the one-tenth millimeter subdivisions seem to be inappreciable, indicating that the screw employed has no sensible periodic error which is a function of a single revolution.

With regard to the decimeter subdivisions, it should be said that the graduation was done in three operations, the second zero being at about the thirty-third centimeter, and the third at about the sixty-sixth centimeter.

XVI.

A METHOD OF FILTRATION BY MEANS OF EASILY SOLUBLE AND EASILY VOLATILE FILTERS.

BY F. A. GOOCH.

Communicated March 11, 1885.

THE processes of analysis, in which it is desirable to redissolve precipitates from the filter after washing, or to separate a mixed precipitate into parts by the action of appropriate solvents, are many.

When a complete solution is the object, and the precipitate yields easily to solvents which do not affect paper injuriously, the use of the ordinary filter offers no difficulty. When, however, precipitates are to be treated with reagents which disintegrate paper filters, the case is otherwise; and the attempt to remove, by solvents, any individual part of a mixed heterogeneous mass upon a filter, is always an uncertain matter. As examples of cases of this sort, difficult to deal with, we may take the solution of acid sodic titanate in strong hydrochloric acid; or, the purification of baric sulphate from included salts, by digestion in strong hydrochloric acid; or, the separation of sulphides which are soluble from those which are insoluble in alkaline sulphides; or, the washing out of free sulphur from precipitated sulphides by means of carbon disulphide; or, the separation of calcic and baric sulphates by the action of sodium hyposulphite. In cases of this nature it is often convenient to make use of the asbestos filter which I have previously described; but this sometimes has its disadvantages. Thus, to recur to the examples just cited, acid sodium titanate may be filtered and washed upon an asbestos filter, and felt and precipitate treated together with hydrochloric acid, but it will be impossible to determine when solution is effected because of the floating asbestos; and in separating the sulphides it would be necessary to know the weight of the asbestos felt, since it must be weighed finally with the insoluble sulphides, unless removed by a special treatment which involves the solution, filtration, and reprecipitation of the latter.

These Proceedings, Vol. XIII. p. 342.

It is to meet cases like these that I have sought a filter which, in the reversal of the ordinary mode of separating filter and precipitate, should dissolve easily in solvents which do not affect the ordinary precipitates met with in analysis. The material which seems best suited to the case - light and fluffy, capable of making secure filters of any desirable degree of porosity, sufficiently insoluble in water and aqueous solutions of salts, alkalies, and acids (excepting strong sulphuric, strong nitric, and glacial acetic acids), easily soluble in naphtha, benzol, carbon disulphide, ether, boiling alcohol, and essential oils, and not too costly—is anthracene.

The mode of preparing and using the filter is simple. Anthracene is slightly moistened with alcohol to make it miscible with water, diluted to the right consistency, and applied to the same apparatus, and in the same way, as the emulsion of asbestos which is employed in making asbestos felts. That is to say, enough of the emulsion in water to form a layer of the proper thickness is poured into a perforated crucible which is held tightly in a packing of rubber tubing stretched over a funnel fitted in the usual manner to a vacuum-flask or receiver. After washing with water the filter is ready for use. If the felt happens to be too coarse for the use of the moment, it may be made as close as need be by coating the felt first deposited with a finer emulsion, made by dissolving anthracene in hot alcohol and precipitating with water. When voluminous precipitates are to be filtered, the large perforated cone described in the former paper, to which I have referred, may be substituted with advantage for the crucible; or Cooke's improved form of Carmichael's process of reverse filtration may prove most useful. In using the cone it is well to apply the anthracene in a thick layer.

To remove the anthracene filter from a precipitate, it is only necessary to act with the proper solvent. It is usually convenient to stand the crucible containing precipitate and felt in a small beaker, add enough of the solvent, and gently warm until the anthracene dissolves. On the addition of water, or the reagent to work upon the precipitate, the solution of anthracene floats, and nothing remains to obstruct or obscure the action. If the precipitate dissolves entirely, the solution of anthracene may be separated from the aqueous solution by simply pouring the fluid upon a filter previously moistened with water, when the solution in water runs through, and the anthracene and its solvent remain and may be washed indefinitely with water.

*These Proceedings, Vol. XII. p. 124.

of

If, on the other hand, the case is one of the division of precipitates, the anthracene and its solvent may be made to pass the filter, after the water has run through, by adding a little alcohol to overcome the repulsion between the solution and the water which fills the pores the filter, the precipitate which stays behind being washed first with a solvent of anthracene, and then, if necessary, with alcohol followed by water; or, if the vacuum filter be used (either paper or asbestos, according to the circumstances of the case), both liquids leave the precipitate and traverse the filter together.

In general, I prefer benzol as the solvent for anthracene, but some advantage may be gained in special cases by a proper choice of solvents. Thus, in removing intermixed sulphur from precipitated sulphides, both the anthracene and the sulphur may be dissolved in carbon disulphide in a single operation.

The ready volatility of anthracene, at a temperature very near its melting point, 213° C., makes it easily separable in cases when to remove it by a solvent is not advisable. The treatment of a solution of anthracene, for example, with strong sulphuric or nitric acid, is apt to produce carbonaceous or gummy residues. In such cases it is well either to heat precipitate and filter directly, or to first remove them from the crucible by means of a solvent for anthracene, then evaporate this and raise the heat gently until the anthracene has vanished. The purification of precipitated baric sulphate, by dissolving it in hot, strong sulphuric acid, and reprecipitating by dilution, is a case in point; and one, too, in which the reversed filter may be used with great advantage. It may be remarked, in passing, that, if one does not happen to possess a platinum rose, and does happen to have at disposal a perforated crucible, a very fair reversed filter may be improvised of the crucible, a piece of glass tubing, and a rubber stopper, the last being fitted to the crucible, and the tube passed through nearly to the perforated bottom.

At every point in the preparation and use of the anthracene filter, I have found the manipulation peculiarly easy.

XVII.

OBSERVATIONS OF VARIABLE STARS IN 1884.

1

BY EDWARD C. PICKERING.

Communicated March 11, 1885.

IN the communication entitled "Recent Observations of Variable Stars," it was stated that a similar circular would be published early in 1885. The friendly co-operation of several astronomers interested in the subject makes it practicable to present on this occasion a much fuller view of the progress of observation, in Europe as well as in America, than could be given last year. The various observers are named below in alphabetical order, with the abbreviations employed to designate them in the subsequent tabular statements.

B. These observations were made by Mr. T. W. Backhouse, at Sunderland, England. The instruments employed were a refracting telescope by Cooke, aperture 4 inches, with magnifying powers 38 and 75; the finder of this telescope, power 9; a field-glass and an opera-glass, with powers 4 and 2.2 respectively; other observations were made with the naked eye. The comparisons were made either in grades, in fractions of the interval between two comparison stars, or by approximate differences. A copy of the observations for 1884 has been received at the Harvard College Observatory.

C. These observations were made by Mr. S. C. Chandler, Jr., at the Harvard College Observatory. The telescope is by Clacey; aperture 6 inches, magnifying power generally 45, sometimes 125 or 200. The observations were made by Argelander's method. Most of them were made before April 28, and they were discontinued after June 30, owing to the requirements of other researches. They are not likely to be resumed at present.

D. These observations were made by Dr. N. C. Dunér, at the Observatory of Lund, Sweden, according to the method of Argelander.

E. These observations were made by Mr. John H. Eadie, at Bayonne, New Jersey. The telescope employed was made by John Byrne; its aperture is 31 inches, and the lowest magnifying power

1 Proc. Amer. Acad. of Arts and Sciences, XIX. 296.