CHEMISTRY. 15 



is made up to 50 cc. by the addition of the sulphuric-nitric acid mixture descriljed 

 above. In case of soil solutions no further addition of acid is made. 



The 50 cc. of phosphoric acid solution is heated until bubbles of <jas ])egin to form, 

 s\vun<r a few seconds so that the sides of the beaker are not overlieated, and 50 cc. of 

 the snlphate-molybdic solution is poured into the middle of the solution ami the 

 l)eaker covered. As soon as the larger part of the precipitiite has settled, which 

 usually re(]uires not more than 5 minutes, the .solution is vigorously stirred for half a 

 minute with a glass rod. After 2 to 18 hours, the longer period being required for 

 soil .'-solutions containing less than )! mg. of phosphoric acid, the solution is rapidly 

 filtered through a jdatinum Gooch crucible according to the Kilgore method, using 

 ]iai)er disks, ami the precipitate washed about 4 times with 2 percent ammonium 

 nitrate solution. The crucible is then filled once full and twice about lialf full with 

 alcohol, care being taken each time not to allow the alcohol to be entirely drawn out 

 of the filter before a fresh supply is added. The precipitate is then treated in the 

 same way with ether, the precaution noted in connection with the alcohol washing 

 being especially important in this case, as the precipitate dries so quickly and is in 

 form of such fine powder that it may be drawn through the filter. The precipitate is 

 dried for 30 minutes in a vacuum desiccator containing no absorbents and weighed. 

 The ammonium phosphomolybdate obtained by this method contains 3.295 per cent 

 of phosphoric acid. 



The results obtained by the method were found to agree very closely with those 

 given by Wagner's molyl)dic method, but were uniformly somewhat lower than those 

 yielded liy lieitniair's citrate method. 



The determination of clay in soils, A. Pagnol'l {Auik Arjron., 27 {1901), No. 

 2, pp. 94-96; Bui. Sfa. Agron. Pas de Calais, 1900, pp. 1-4). — The author describes a 

 modification of Schloesing's method, which he has used with success, as follows: To 

 an amount of the sifted soil, corresponding to 5 gm. of dry soil, in a small dish add a 

 small portion of a solution of 1 cc. of ammonia in 100 cc. of distilled water, rub up the 

 soil against the sides of the dish with the finger, and pour off the liquid after standing 

 10 to 15 seconds. Repeat this operation three times and collect the soil and the 

 washings in a second dish and boil for 2 or 3 minutes. After cooling, add 1 cc. more 

 of annnonia, and wash the soil and solution into a 100 cc. cylinder having a side tube 

 with stopcock at the 20 ci;. mark (from the base), complete the volume to 100 cc. of 

 distilled water, shake vigorously, and allow to stand for 5 minutes. Then draw off by 

 means of the stopcock 80 cc. of the muddy liquid free from coarse sand and containing 

 the clay in suspension. Add 20 drops of acetic acid and allow to stand for 10 to 15 

 minutes for the clay precipitate to settle. Collect on a filter and wash until all lime is 

 removed. Then dry and calcine the filter. In this method the coarse sand is elimi- 

 nate l)y decantation, the lime by filtration, and the organic matter by calcination. 

 The weight obtained is therefore considered to represent the clay in 'suspension in 

 80 cc. of the decanted liquid. From this data the percentage of clay in the original 

 soil may be readily <'alculated. 



On the Adie and ■Wood method for the determination of potassium, V. B.\s- 

 KERVii.i.Kand I. F. IIakkis {.Tour. EU><ha Milchcll Sri. >SVir., 1900, No. 1, p. 7.?).— The 

 authors rei)ort comparisons of this method (E. 8. R., 12, p. 18) with the oHicial 

 method with very satisfactory results. 



The testing of drinking water with the Erdmann reagent, A. Ficrmax 

 {Oi'.^tcrr.-Ungar. ZLschr. Zuckcrind. u. Lundw., 29 {1900), pp. SGl-S6.i; abs. in Clwni. 

 C'entbl., 1901, I, No. 8, p. 475). — Attention is called to the unreliabilty of Erdmann's 

 reagent {Yj. S. R., 12, p. 18), due to tlie fact that very small amounts of organic 

 matter and other sub.stances will give the nitrite reaction. The author recommends 

 the use of m-phenylendiamin, for, while this reagent is less sensitive than Enhnann's, 

 it is less likely to give misleading results on account of reaction with other substances 

 besides nitrites. 



