CHEMISTRY. 429 



Over tliese two sets of'deteninnatioiis, wliieli ditteriii mean by nearly 

 0.2, some controversy has arisen. For details, see Tli()r[)e and Laurie 

 (Bericbte, xx, 3030, and Jonrn. Cliem. Soc, December, 18S7), and Kriiss, 

 (Bericlite, xx, 2305, and xxi, 128). Meanwhile Mallet has imblished a 

 l>reliminary note, without particulars, concernin.n- a determination which 

 he has made. He gets a direct ratio between gold and hydrogen, giv- 

 ing for gold a value somewhat higher than that obtained by Thorpe and 

 Laurie. (Chem. News, No. 1152.) 



Thv (Irfenninafion of molecular wcUjliis. — Early in 18S8 Victor Mc.yer 

 called attention to Kaoult's method for ascertaining molecular weights, 

 l)ronouncing it to be the most important advance in that tield of work 

 since the announcement of Dulong and Tetit's law. The niethod, in 

 l)riuciple, is as follows: When any substance is dissolved, the freezing 

 ])oint of the solvent is lowered ; and that depression of freezing point is 

 directly related to the molecular weight of the body in solution. If we 

 divide the observed depression of freezing point for each substance ex- 

 airiined by the weight of the substance dissolved in 100 grammes of 

 water, we get a coefficient of depression which gives, when multiplied 

 by the molecular weight, a constant quantity. The latter constant is 

 independent of concentration, and has approximately uniform value for 

 all similarly constituted bodies in any given solvent. It is called the 

 '))ioIccHlar depression ; and differs for different solvents; but in glacial 

 acetic acid it seems to have one value for all substances. Water, acetic 

 acid, and benzene seem to be the only solvents of general applicability. 

 If now it is desired to know the molecular weight of a soluble bod^' to 

 which vapor-density methods are inapplicable, we have only to deter- 

 mine its coefficient of depression and compute from that the sought for 

 value. The method, as a physical process, is not sharply accurate; but 

 it is close enough to decide between the lowest empirical formula assign- 

 able to a substance and the various multiples thereof. Although the 

 original investigation was published by Eaoult in 1883, it attracted 

 little attention until Meyer's paper appeared; but since then it has 

 been widely noted, and the method extensively applied. (See Meyer, 

 Bericlite, xxr, 530, and Aiiwers, ibid., xxi, 701.) Auwers also describes 

 a convenient form of apparatus, and so too does Beckmann. (Zeit. 

 Phys. Chem., ii, 638.) 



In illustration of the use of Kaoult's method, the following researches 

 out of many may be cited : 



Ramsay, working with an acetic acid solution, finds the molecular 

 weight of liijuid nitric peroxide to be 92, at a temi)eratiire of about 

 10°. Hence its formula is N2O.1, and it seems, iiirthermore, to undergo 

 no dissociation upon dilution. (Journ. Chem. Soc, Liii, 021.) 



Brown and Morris, using aqueous solutions, get dei)ression valines for 

 dextrose, cane sugar, maltose, lactose, and arabinose corresponding I0 

 their accepted molecular weights. For ratlinose the simplest empiri- 

 cal formula, CiitH320,G . SHaO, corresponds to the results obtained by 

 Kaoult's method. (Journ. Chem. Soc, Llii, 610.) 



