March 18, 1904.] 



SCIENCE. 



447 



Following out these three concepts, we 

 are able, for the first time, to express such 

 compounds as CoCl3(NH3)5 satisfactorily. 

 Thus, around a central cobalt atom are 

 arranged the six ammonia groups attached 

 to the cobalt atom by secondary valences. 

 They are in the first sphere of influence, 

 and hence the whole complex Co(NH3)c 

 acts as a single ion. These six groups are 

 probably regularly distributed around 

 cobalt as the central atom and may, there- 

 fore, be represented by an octahedron with 

 an ammonia group joined to each of the six 

 vertices. In the second sphere, and not 

 directly connected with the cobalt atom, 

 lie the chlorine atoms. Being necessarily 

 farther removed from the cobalt atom, we 

 should expect more freedom of action for 

 them than for the ammonia, that is, they 

 would act as ions when the salt is in solu- 

 tion. All this may be briefly represented 

 by the formula Co(NH3)5Cl3. It has been 

 experimentally proved that such complex 

 ions as Co(NH3)e do actually exist in solu- 

 tion and that, in this particular salt, all 

 three chlorine atoms do act as ioils. 



Werner's theory also explains many 

 hitherto inexplicable i^henomena of the 

 simpler compounds. Wliy does ammonium 

 chloride, NH^Cl, dissociate while the cor- 

 responding compound methyl chloride, 

 CH3CI, its left-hand neighbor in the 

 periodic system, does not? According to 

 the idea given above, the hydrogen in the 

 ammonium chloride would be in the first 

 sphere of the nitrogen, the group NH^ act- 

 ing as an ion, while the chlorine would act 

 in the second sphere. The compound 

 should dissociate. It does. In the case of 

 the methyl chloride, there is no dissocia- 

 tion because both the hydrogen and the 

 chlorine are in the first sphere of influence 

 and joined directly to the carbon atom. 

 This same explanation also applies to the 

 oxonium, sulphonium, phosphonium, iodon- 

 ium and diazonium salts. 



It is impossible to explain these molecular 

 compounds by the old theory. The very 

 term 'molecular compound' is proof that 

 the old valence is insulficient. 



Various attempts have been made to dis- 

 prove Werner's theory, but in most cases 

 experiments have proved rather than dis- 

 proved the theory. This is especially true 

 with regard to coordinate number, which 

 gives to each element a fixed number and 

 secondary valence which has a definite 

 limit. It is not understood that this 

 theory was designed to replace the old 

 valence theory in cases of simpler com- 

 pounds like sodium chloride or in any case 

 where the present theory is satisfactory. 

 It Avas only intended as an extension of the 

 present theory. 



Solubility of Gold in Certain Oxidizing 

 Agents: Victor Lenher, University of 

 Wisconsin, Madison, Wis. 

 Metallic gold is soluble in such inert 

 acids as sulphuric and phosphoric when 

 heated in the presence of such oxidizing 

 agents as selenic, telluric, nitric and 

 chromic acids, red lead, lead dioxide, 

 nickelic oxide, manganese dioxide and the 

 higher oxides of manganese. Anode 

 oxygen will also readily cause solution of 

 a gold electrode with electrolytes of either 

 acids or alkali, most of the metal subse- 

 quently depositing on the cathode. In 

 case of such salts as sodium sulphate or 

 sodium nitrate, very little of the gold 

 passes through or enters the electrolyte, 

 and the gold anode is completely trans- 

 formed into gold oxide. 



On a Method for Preparing Salts with a 

 Definite Number of Molecules of Water 

 of Crystallization: Launcelot W. 

 Andrews, University of Iowa, Iowa 

 City, Iowa. 

 Salts containing a maximum amount of 



water of crystallization when enclosed in a 



