the difference of the weights made up by a piece of thin platinum wire, to be attached to the hanger of the left-hand beam. The same operation is repeated with the third cylinder, after the removal of the cylinder on the left-hand side.

If this is satisfactorily accomplished, the beakers with water are removed and, when the temperature of the room is nearly o°, two tumblers are procured, with ground rims and ground covers, the latter each containing in the center a hole of about five-tenths cm. diameter, and near the rim a hole to admit a thermometer, which has a cylindrical vessel, and is graduated on the stem. The one tumbler is filled partly with fused calcium chloride and caustic lime or potash; the other is nearly filled with cold distilled water. The first tumbler is put under the hanger of the left beam, and the other under that of the righthand beam. After a time, sufficient for the cylinders and the water to acquire the temperature of the surrounding air (°), the weight is read off. It is:

x= the weight of each aluminum cylinder in grams.

y = the volume of each aluminum cylinder in cc.

z = difference of weights, if one is immersed in pure air, the other in pure water.

the same, reduced to the vacuum.

a = specific gravity of pure water at (air thermometer) and h barometric pressure.

=

b specific gravity of dry air at t° (air thermometer), and h barometric pressure.

= =

b, specific gravity of the air at t° (air thermometer) in the balance case.

c = specific gravity of the gilded weights at t° (platinum weights are to be preferred).

If this value of z, is substituted in equation 1, we get:

y=

2(c+b1)
c(a-b)

The same operation is repeated at about 25°, 50°, and 75°. In the latter two instances the temperatures of the water and air will, of course, differ considerably. From the values ob tained, a table is calculated of the volumes of the aluminum cylinders for each degree of the air thermometer from o° to 100°.

To ascertain the specific gravity of the air in the balance case at the time, when it is intended to determine the specific gravity of a liquid, the tumbler with pure water and thermometer is put under the left hanger of the balance, the cylinders hooked to the hangers of the beam, the one immersed in the water, the other in the air. After a time, sufficient for the water to have acquired the temperature of the air, the weight is read off. We have

then :

a

z (c+b1)
cy

= b1.

x + ay―z1 = x + b,y; or a Substituting for b, on the left side repeatedly the value

a

су

za

y су

1

i. e., the specific gravity of the air, if that of the water is assumed to be 1.00. If t° is not 4.1° C., b, has to be divided by a and the formula becomes :

The aluminum cylinder on the right hand is thence removed and the tumblers with the liquid to be examined, into which the third cylinder and a thermometer had already been immersed, is put in its place, the cylinder hooked to the hanger, and the weight ascertained.

If the specific gravity of a liquid at a high temperature is to be determined (i. e., the expansion coefficient), a hole directly underneath the hanger of the right-hand beam passes through the upper and lower bottoms of the balance case and also through the table. Through this hole passes a platinum wire

connected with the one cylinder hanging in the beaker, containing the liquid, on a water-bath over a Bunsen burner.

When the liquid in the beaker has about the desired temperature, the lamp is extinguished and the weight of the cylinder ascertained.

Now the thermometer is read off again and the weight again ascertained. If the temperature has risen in the meantime the operation is repeated, till the weight is less than at the first reading. The averages of all temperatures and of all weights are taken and used for the computation.

A SHORT STUDY OF METHODS FOR THE ESTIMATION OF SULPHUR IN COAL.'

BY G. L. HEATH.
Received June 15, 1898.

S manufacturer's requirements with regard to certain metallurgical processes become more exacting from year to year, the determination of sulphur in the fuel must become more important. The interest taken in this subject is evidenced by the appointment of the committee from the society who have just made their preliminary report. The writer has been led, as a matter of interest, to communicate some experimental work upon the determination of sulphur.

There has been a little controversy, or doubt, as to the relative accuracy of the two "sintering" or "ignition" methods in general use as compared with each other, or with the old method of fusion with sodium carbonate and potassium nitrate. It must be assumed and understood that the following work was done rather to study methods and show precautions necessary, than to obtain close check results. This remark explains a discrepancy in results on one coal of very high sulphur content, since the methods which will now be described as used by the writer, were purposely not modified to suit that special case.

I. THE FUSION" METHOD.

This is so well given in the text-books with so little variation that full description is unnecessary. Blair's modifications used.

1 Read at the meeting of the New York Section. June 3, 1898.

2 Blair's Analysis of Iron, 1888, p. 245.

In our experience, much care must be taken to dehydrate, by evaporation with acid, the silicic acid which dissolves in the solution of the fused alkali. This renders the process too slow for technical work.

2. ESCHKA'S METHOD.'

a. One gram of powdered coal is intimately mixed with one gram of pure magnesium oxide and one-half gram sodium carbonate, and heated with the flame of a large (eight-ounce) alcohol lamp in a platinum dish of about 100 cc. capacity. A dish is much to be preferred to a narrow crucible, since it offers a better exposure of the coal to the air, and prevents excessive heat and dry distillation in the interior of the mass.

The mixture is frequently stirred with a platinum wire and the heat is raised very slowly, especially so with very soft coals. The flame, which is kept in motion and barely touches the dish at first until strong glowing has ceased, is increased gradually until, in fifteen minutes, the bottom of the dish is at a low red heat. When the carbon has burned away the mass is transferred to a No. 2 Griffin beaker and boiled five minutes with 100 cc. of distilled water and some oxidizing agent.

Fifteen cc. of saturated bromine water, as recommended by Mack and by Handy, is the purest and most efficient reagent to use, safer than the plan of adding ammonium nitrate during ignition. The hot liquid is then passed through a washed filter and the residue washed with hot water until the total volume is approximately 200 cc. The filtrate is then acidified strongly with hydrochloric acid, boiled until the free bromine is driven off, and sufficient hot solution of barium chloride added to insure the precipitation of all the sulphuric acid as barium sulphate, which is settled, filtered, and ignited.

b. Fresenius recommends that the ignited mixture be boiled directly in bromohydrochloric acid, which will dissolve all the sulphur in high coals, but loads the solution with salts and possibly silica.

C. The modification, recently published, involving the use of silver oxide, does not seem to offer any advantage whatever 1 Chem. News, 21, 261 ; J. Anal. Appl. Chem., 6, 611.

over the other methods using inexpensive material, and is not included in the table.

3. HUNDESHAGEN'S METHOD.'

It is a familiar fact that this modification differs from Eschka's only in the substitution of potassium carbonate for sodium carbonate. The advocate of the first salt claimed that sodium carbonate did not retain all the organic sulphur from certain Bohemian brown coals.

J. O. Handy,' of Pittsburg, presented tests from which he deduced a conclusion in opposition to that of Hundeshagen, but he does not appear to have tested the particular class of coals specified by the latter chemist. The writer has, accordingly, experimented with a variety of samples, including Bohemian brown coal, and has experimented also with a few variations in the details of procedure. Part of the foreign samples were procured from the Michigan College of Mines, and a fine specimen of the Bohemian article from Prof. Lattimore, of the University of Rochester.

The proximate analyses were carried out as usual with the author, according to Blair's' modification of Heinrich's method. Blank analyses for the sulphur in reagents were also carried through. This is the more necessary, as I have sometimes met with goods labelled by importers "strictly C. P.," which were rather impure.

An inquiry, in one case, elicited the statement from a clerk that his house labelled articles C. P. to fill orders for the highest grade, even if they were only a common grade on the market at the time. The American firm of Baker & Adamson, it is but just to state, has furnished the purest chemicals for fuel analysis that we could obtain anywhere.

4 AND 5. WILEY'S AND CARIUS' METHODS.

I have had no experience with the Wiley and Carius methods, which are not so much used, as such, in America.

6. WET METHOD OF CALVERT.

The wet method" of Calvert' is useful for the scientific inves

1 Chem. Ztg., 16, 1070; also J. Anal. Appl. Chem., 6, 385.

2 J. Anal. Appl. Chem., 6, 116.

3 Blair's Analysis of Iron, 1888, p. 243.

4 Chem. News, 24, 26; also Watts' Dictionary of Chemistry.