380 PROCEEDINGS OF THE AMERICAN ACADEMY. 



in another otherwise similar case. The conditions attendant upon mixing 

 two liquids are so uncertain that it is unsafe to attempt the quantitative 

 interpretation of the relationship ; but qualitatively this result is wholly 

 in accord with the law of mass action, provided that the undissociated 

 group is supposed to be the one concerned. 



The nature of the electrically neutral group with which we have to do 

 in the present case is less easy to imagine. Kiister and Thiel assume 

 that it is Ba[Fe(S04)2]2 ; but this is evidently an unsafe assumption. 

 The analogy to the case of chromium which they quote is an excellent 

 point, but they do not seem to have made themselves conversant with 

 the literature upon this subject, and hence apply their useful analogy 

 unfortunately. 



They seem to consider that chromium exists in the green solution as 

 part of an acid ion, instead of in the basic semi-colloidal form ascribed to 

 it upon excellent grounds by both Recoura and Whitney.* One could 

 hardly expect a basic substance to form such a salt as Kiister and Thiel's 

 with barium. 



Some light upon the subject is to be had from the study of the quan- 

 titative results in the paper of Jannasch and myself before mentioned. 

 We found that the impure baric sulphate precipitated from a ferric solu- 

 tion (containing enough sulphuric acid to have yielded 1.1608 grams of 

 pure baric sulphate) as a matter of fact weighed on the average only 

 1.1170 grams, and contained 0.0144 gram of ferric oxide, hence the baric 

 sulphate actually present amounted to 1.1170 — 0.0144 = 1.1026 grams. 

 This is 0.0582 gram less than the amount which should have been 

 present ; hence, supposing that no sulphuric acid remained in the solu- 



SO 

 tion, there must have been 0.0582 -r. ,^t, == 0.0199 gram of sulphur 



trioxide in the iron complex. But 0.0144 gram of ferric oxide corre- 

 sponds to 0.0216 gram of sulphur trioxide in ferric sulphate; hence 

 there was a deficiency of nearly two milligrams. If sulj^huric acid 

 remained in solution, the deficiency in the complex must have been still 

 greater. It is true that the observed difference is only half so great as 

 that demanded by a formula analogous to Recoura's, Fe4(S04)5(OH)2, 

 and is perhaps not very much greater than a possible experimental error; 

 but, as far as the argument goes, it points toward the existence of a basic 

 (complex in the precipitate. 



* A convenient resume of the argument in favor of this point of view is to be 

 found in a recent article bv Whitney, Jl. Am. Chem. Soc, 21, 1075 (1899). 



