612 



SCIENCE. 



[N. S. Vol. IX. No. 226. 



proposed for this simple acid. For the 

 familiar potassium chloroplatinate, K^FtOg, 

 there are four constitutional formulas seri- 

 ously advocated at present. It may be 

 Kj = PtCJg, with octavalent platinum ; 



KCl = CI p p, 

 KCl = CI -^ ^^^' 



with tetravalent platinum and trivalent 

 chlorine, as required by Eemsen's theory ; 

 (PtC]^)Kj in the sense of Werner's theory, 

 the two potassium atoms being combined 

 with the PtCl^ as a whole, or it may be a 

 molecular compound in which two mole- 

 cules KCl as wholes combine with PtCl^ as 

 a whole. The formulas suggested for most 

 minerals are pure guess work. The silicates 

 are usually written as if containing the 

 group Si;=0, by analogy with carbonyl, 

 C = O, yet there is not a single silicate in 

 which this assumption rests on any experi- 

 mental evidence, and the little we do 

 actually know of the chemical behavior of 

 silicon speaks against it. Such formulas, if 

 not purely speculative and devoid of all 

 basis and all value, as they frequently are, 

 at best do not represent structure in the 

 sense that the best established organic 

 formulas do; they are at most reaction 

 formulas only, or they represent partial 

 molecules, in the same way that CH may 

 stand for benzene (CgHJ or HPO3 for a 

 metaphosphoric acid. The attempt to in- 

 terpret the double salts and halides, the com- 

 pounds with water of crystallization or 

 hydration, the metal-ammonias, the peculiar 

 compounds of the zeolites described by 

 Friedel, and other so-called molecular com- 

 pounds, in the sense of the valence hypothe- 

 sis, seems almost hopeless without taking 

 such liberties with it as to render it nearly 

 useless, and without making assumptions 

 of very narrow and limited applicability. 

 One may well question whether this 

 hypothesis must not be very considerably 

 qualified before it can be taken as the basis 



of a general theory of the structure of in- 

 organic compounds. 



One of the most striking indications of a 

 revival of inorganic chemisti-y is the recent 

 attempt of "Werner to break away from the 

 bonds of the organic-structure theory as 

 applied to inorganic compounds and to es- 

 tablish a more general theory in which 

 valency plays a comparatively insignificant 

 role. The arguments on which Werner's 

 hypothesis is founded are too numerous 

 and elaborate to be presented here. Suf- 

 fice it to say that it was primarily based on 

 that peculiar class of bodies known as the 

 metal-ammonias, consisting of metallic 

 salts, combined with usually six or four 

 molecules of ammonia, and in which the 

 ammonia may be wholly or in part replaced 

 by pyridine, water, acid radicals and other 

 groups. These groups are supposed to be 

 arranged symmetrically about the metallic 

 atom, forming a radical, which, according to 

 its nature, can combine as a whole with met- 

 als, halogens or other positive or negative 

 groups. Thus, in the compound C0(ISrH3)g 

 CI3, cobalt forms with NHj a radical, which 

 combines as a whole with the three chlorine 

 atoms ; in (PtClJK, the two potassium 

 atoms are combined with the whole group 

 PtClg and not attached to any one part of 

 it; the same applies to (NHJCl, and to 

 K,(SOJ and K,(FeCN),. In the formation 

 of these radicals the bivalent NH3, the 

 neutral II^O and the univalent CI can re- 

 place each other indiscriminately ; the 

 valence theory is, therefore, practically 

 thrown overboard entirely and in place of 

 combination by bonds we have an exten- 

 sion of the old theory of molecular com- 

 pounds applicable alike to the metal-am- 

 monias, the ordinary oxj'gen salts, the 

 double halides and the compounds with 

 water of crystallization. It is yet too soon 

 to predict the future of this hypothesis, 

 which has already won numerous active 

 adherents. It is scarcely too much to hope 



