252 



has concluded that the laevogyratory rhodium-salts and the dextro- 

 gyratory cobalti-salts which are set free from the less soluble haiogeno- 

 d-tartrates, must possess the same configuration in space. Therefore, 

 according to this author, if in the triethylenediamine-complex of 

 the cobaltt-salts the central cobalt-atom be replaced by the rhodium- 

 atom, not only the size of the specific rotation would be altered, 

 but even its direction would be reversed. However, Werner does not 

 give any plausible argument to support his starting-point regarding 

 a direct connection between the configuration of the molecules and 

 their solubility. Moreover, this view seems to be quite fortuitous, 

 as the solubility of compounds is so highly complicated a constitutive 

 property of matter, that even where rules for homologous compounds 

 apparently universally valid seem to be established, the most unex- 

 pected and surprising facts often appear to make them illusory after- 

 wards. The crystal-form of the substances, on the contrary, is a pro- 

 perty much more directly connected with their geometrical character, 

 i. e. with their atomic configuration in space, than is the case with 

 their solubility. If in the chloro-d-tartrate of an optically-active tri- 

 ethylenediamine-cobalti-ion of definite configuration, the central 

 Co-atom, under complete preservation of the existent arrangement 

 of the radicals, be replaced by a Rho-atom, which, according to our 

 investigations, will replace it quite isomorphously, then it must 

 be obvious that the resultant rhodium-salt will appear perfectly 

 isomorphous with the original chloro-d-tartrate. Experience, however, 

 now proves beyond any doubt, that both the less soluble chloro-d- 

 tartrates of the Rho-, and Co-salt possess quite different crystal-forms, 

 the one being triclinic-pedial, the other monoclinic, while their 

 parameters do not show the least analogy. Moreover, direct analysis 

 proves that the chloro-d-tartrate of the Co-salt crystallises with five, 

 that of the Rho-salt with four molecules of water, a fact which 

 is also in full agreement with the absence of direct isomorphism 

 between both derivatives. Both chloro-d-tartrates , therefore, cannot 

 have an analogous configuration, but they must necessarily possess 

 antilogous configurations. This is the simple and true explanation 

 of the fact that the cobalti-, and rhodium-salts, set free from the 

 less soluble chloro-d-tartrates, also show optical rotations of opposite 

 algebraic signs. In contradiction to Werner's, our conclusion, 

 therefore, is that the cobalti- and rhodium-salts with the same direc- 

 tion of rotation possess also the same arrangement of the constituent 

 radicals in space. The influence of replacing a central cobalt-atom 



