136 L Remsen and M S. Southworth oji the action of 



faces of the insoluble crystals, was deposited in even layere 

 upon them, forming thus a complete envelope, and giving a 

 genuine pseudomorph. I was at first deceived by this strange 

 pseudomorph, believing it to be the product of the oxidation. 

 It was insoluble in water, and appeared to be insoluble in 

 alcohol. I found, however, afterward, that the alcohol dissolved 

 the central portions of the pseudomorphs leaving the envelopes 

 unchanged in form. 



5, The substance was heated loith water at 150° in a seaied 

 tube. At this temperature decomposition took place, bat not 

 at a lower temperature. The products of the reaction were 

 alcohol, and a solid, white crystalline substance which conducted 

 itself iu some respects like succinic acid. The alcohol was 

 detected by placing the whole product in a flask and distilling 



The experiments which have thus been described do not 

 suffice to enable us to judge positively in regard to the structure 

 of the substance under investigation. I have stated above the 

 view held by Wislicenus, and also my objections to this view. 

 It remains yet to be decided whether my objections are well 

 founded, and this can be done only by the aid of new experi- 



One of the most remarkable examples of so-called non- 

 saturated compounds is carbon monoxide. If we accept the 

 hypothesis of constant valence, the compound CO must possess 

 free affinities, or, as some chemists believe, the two affinities of 

 the carbon-atom, which are not saturated by the oxygen atom, 

 must exercise an influence upon each other. We can not ex- 

 plain this case by assuming that two carbon-atoms are joined 

 together by two affinities each, for we know that the formula 

 of carbon monoxide is CO, and not C^O^ or a higher multiple, 

 and, accepting this formula, it is plain that we cannot assume a 

 double union of carbon atoms in the compound. 



If, on the other hand, we accept the hypothesis of variable 

 valence, believing that the valence of an element depends 

 upon circumstances, we shall look in vain for circumstances 

 which, in the one case, can cause the bivalence, in the other the 

 ^>eratur( 

 linly does not < 



