THE following few lines give a necessary elucidation to my critical note and to the answer given by Messrs. H. B. Baker and H. L. Riley (NATURE, Mar. 5, p. 348). My principal first theoretical argument is based on theinterdependenceof the atomic weights of the elements silver, nitrogen, and chlorine, resulting from the classical life-work of Richards and his school, in which I have the greatest confidence. If the atomic weight of silver = 107.876, then nitrogen = 14.006±0.0011, and chlorine = 35.456±0.002, most probably 35.458. If we accept Baker and Riley's value, silver = 107.864, then nitrogen would become 13.999, a value exceedingly improbable, especially having regard to the fact that Baxter found recently (Proc. Amer. Acad.,12, 12, p. 699, Dec. 1926) by an extremely careful physico-chemical research the value N = 14.006(7), which confirms the higher atomic weight of silver, namely, 107.876. This important argument was not referred to by Messrs. Baker and Riley.My second, no less important, practical argument was based on the assumption that Messrs. Baker and Riley have lost exceedingly small quantities of silver vapour on fusing the metal in hydrogen, so that the atomic weight found by them is slightly lower than the true one. They did their best to convince themselves that no visible condensation of metallic silver could be observed in their tubes, and they say that they have begun a new series of experiments to investigate the volatility and condensability of silver. I beg to remark that some experiments on a large scale in this directi6n were published by J. S. Stas so long ago as 1865 ("CEuvres completes," T. I, p. 457), who was my first " atomic weight teacher " in 1875 (but who would read such 'antiquated' papers to-day ?). He describes the distillation of 50 gm. of his purest silver in the flame of the oxyhydrogen blowpipe and says: " Je dois avouer toute fois que, dans les operations que je viens de d6crire, la moiti6 au moinas de I'argent employe a Ut6 perdue. En effet, il a 6t6 entrain4 a 1'6tat de vapeur bleue pale avec le courant de gaz tonnant, quoiqu'il fit cependant moder6, et sans excbs trop grand d'oxygene; il a 6te repandu dans l'air ambiant dont il a trouble la transparence, et auquel il a communique une saveur m6tallique tres sensible." From this important observation it follows that when silver once passes into the state of vapour it is not easily condensed in a solid state, but forms only a colloidal dispersion as a fog. Large quantities of silver heated in tubes give a condensation of the metal, but when a small quantity was heated and fused, the silver vapour-the weight of which was, in the said experiments, of the order of 0 0001 gm. and which would occupy in the solid state 0 00001 cm.3-may have passed out of the apparatus.Messrs. Baker and Riley say that they controlled the weight of the fused silver obtained by repeatedly melting and weighing it to constant weight. But the question arises: What was the weight of the silver obtained in a fine state of division immediately after decomposition of its oxide by heat and before fusion ? Such silver has a very great surface, and during fusion a small loss by evaporation may have taken place. After fusion, its surface has become very small and, last but not least, it was " coated with a very thin film of dross consisting of silica." To these circumstances the fact is very probably due that no appreciable loss of weight of the silver was observed after repeated fusio