Place one gram of the fine asphalt in the bottom of a large nickel crucible. Cover with a layer of about four grams of mixed carbonates and about one gram of solid sodium hydroxide. Cover the crucible and heat gently until the gases are driven off. Then raise the heat and add small amounts of sodium peroxide, from time to time, until no further action is noticed on the addition of fresh Na,O,. Then, with the blast-lamp, heat to perfect fusion. This takes but a few minutes. Pour as much of the fusion as possible upon the cover of the crucible and cool slowly. Place the fusion, crucible, and cover in a casserole containing fifty cc. hot water, and after thorough digestion on the waterbath, filter off the residue, and wash well with hot water. The filtrate is made acid with hydrochloric acid, and evaporated down to dryness, dehydrated at 110°-115° C., the silica filtered off, and the sulphuric acid determined in the filtrate as already described. Results by this method:

IGNITION WITH CALCINED MAGNESIA-ESCHKA'S METHOD.

One gram of finely ground asphalt is intimately mixed with one gram of calcined magnesia and about one-half gram of dry mixed carbonates is added. Mix well and place in a platinum or porcelain crucible. With the crucible in an inclined position, heat at a dull red heat, stirring every few minutes until the ash is a dull yellow. Cool and add about one gram of finely ground ammonium nitrate and mix thoroughly. Heat slowly to a dull red, and continue the heating until the ammonium nitrate is entirely decomposed. Cool, extract the mass thoroughly with water (hot water preferred), filter, and wash well. Acidulate the filtrate with hydrochloric acid, and precipitate the sulphuric acid as before described. The results are as follows:

The relative accuracy of the different methods is best shown by tabulating the results, as follows:

It would appear, therefore, that the "sealed tube" gives the most accurate results, but on account of the explosion of the tubes and the time required for a determination, it would probably be barred out.

The deflagration method apparently gives low results. It is much shorter than the sealed tube, but longer than either the peroxide or Eschka's method.

Of the latter two, the Eschka's method is perhaps the better, as it requires less time and attention to carry out, and gives about as good results.

The peroxide method is shorter than the deflagration method and about as accurate.

The reagents employed were first tested for sulphur.

The sodium and potassium carbonate contained..
The sodium peroxide contained..................

......

Sulphur.
Per cent.

0.005

0.003

The rest of the reagents were sulphur free or contained but the merest trace.

NOTE.

Testing for a Yellow Azo-Color in Fats, Etc.-In this Journal, 20, 110, Joseph F. Geissler describes a delicate test for the detection of a yellow azo-dye used in coloring fats. The test has proved of the greatest service to me, and I make use of it constantly. While experimenting upon the subject, I have noted another simple test for this azo-color that seems fully as delicate as Geissler's. To a few cubic centimeters of the pure filtered fat in a large test-tube are added an equal volume of a mixture of one part strong sulphuric acid, and four parts glacial acetic acid. The contents of the tube are then heated almost to boiling, and thoroughly mixed by violently agitating the bottom of the tube. When now allowed to stand and separate, the lower layer of mixed acids will be strongly colored wine-red if the azocolor be present. Pure butter-fat imparts no color, or at most only a very faint brownish tinge to the acids. Strong hydrochloric acid may replace the sulphuric in the above mixture, or a mixture of one part strong sulphuric acid, and three parts water may be used, but I have obtained the best results in the manner described. ALBERT H. Low.

DENVER, COLORADO, August 27, 1898.

NEW BOOKS.

A BRIEF COURSE IN QUALITATIVE ANALYSIS. BY ERNEST A. CONGDON, PH.B., F.C.S. New York: Henry Holt & Co. 1898. iv + 62 pp. Price 60 cents.

Professor Congdon has given us a very satisfactory little textbook. While brief, the course, so far as it goes, is quite thorough.

The arrangement of the book is excellent and the selection of matter carefully made. Blank leaves for notes are inserted between the pages. The classification of the metals into groups for the purpose of analysis is first shown. Then the reactions. of the metals are given. Following the single reactions of each. group is given the method of analysis for the group. cases an alternative scheme is also printed. A notable exception is in the case of the barium group where the writer gives only the separation of strontium from barium, depending on the different solubilities of their nitrates in absolute alcohol. The reactions of the acids follow those of the metals. The order is alphabetical and not according to groups. The scheme for the detection of iodides, chlorides, and bromides, where more than one is present, by separate tests for each, does not commend itself to the reviewer, and might well be substituted by Professor Hart's easier and simpler method. The section on the treatment of solids is clearer and more concise than that given in most small books. Throughout the book the aim of the writer seems to have been to make the student think. This is borne out by the notes which follow the analytical schemes and by the table of solubilities and list of questions which close the book. RICHARD K. Meade.

ALKALOIDAL ESTIMATION; A BIBLIOGRAPHICAL INDEX OF CHEMICAL RESEARCH PREPARED FROM ORIGINAL LITERATURE FOR THE COMMITTEE OF REVISION. BY PAUL I. MURRILL, under the direction of ALBERT B. PRESCOTT. Ann Arbor, 1898. "Published by the Committee of Revision" of the Pharmacopoeia of the United States of America, 1890-1900. Not for sale.

This is a pamphlet of 58 pages "in boards," and embraces the easily found work upon its title subject from 1861 to 1898. It consists entirely of clerical labor very well done and its arrangement is convenient for reference.

The contents are first a list of the more important of the series of publications examined, and then an alphabetical list of other periodicals to which references are given. Then the body of the work is in three parts.

Part I is the chronological index where, in giving the references, original papers are first given. Then republications and abstracts; and the abbreviations used are, as far as practicable, those of Bolton's "Bibliography of Chemistry."