« 1B7 ) 
led to Hantzsch and Sebaldt’s attempt to render it plausible that 
this compotmd is not NH 4 OH, but a hydrate (NH,) m (H 2 0)n of 
unknown composition. 
Abegg and Riesenfeld l ) have felt able to support this idea by 
demonstrating that the NH,-pressure of an aqueous ammonia solution 
increases most on addition of salts which do not yield complex com¬ 
pounds with NH„ if these salts have themselves a great tendency 
to form hydrates. 
If we consider this and other arguments which have led to the 
conclusion that in aqueous solution of ammonia hydrates ot the same 
should be present, and not NH 4 OH, it appears that the argument is 
far from convincing and that the conclusion is, therefore, subject to 
great objections. 
When instead of NH, OH we write NH,. H,0, we only mean to 
say in fact that we are dealing with a compound which does not 
behave as a base. Now, thq solution of NH, in water does behave as 
a base, but this it is endeavoured to explain in the light of the theory 
of Werner, by assuming that a part of the free NH,-molecules will, 
in consequence of their secondary valencies, unite with the H -ions 
of the water to cations, the water equilibrium being shifted in such 
a manner that the OH-ions may appear in large excess. 
It is true that the phenomena observed may be explained, even 
when we deny the existence of NH,OH„ but for this denial there is 
no sufficient evidence, while there is not a single phenomenon opposed 
to the assumption that NH.OH occurs in the aqueous ammonia 
solution, which is, on one side, in equilibrium with NH, and H, 
and on the other side with the ions NH 4 * and OH . 
The research, the first results of which will be communicated here, 
did not relate to the investigations of previous workers but was simply 
undertaken to answer the question “what are, apart from the structure, 
the compounds which may be deposited on cooling from mixtures of 
NH, and H a O”. This question seemed to us an important one, also 
because from the separation of a compound in solid condition, it 
, may be concluded with a high degree of probability that such com¬ 
pound was previously present in the solution. 
The investigation of the mblting point line now gave the following 
result 
In this graphical representation the results of Pickering s investi¬ 
gations 2 ) are indicated by circlets, and thus we notice that, whereas 
!) Zeitschr. f. phys. Chem. 40 . 84, (1902) and 45 . 462 (1903). 
Journ. Chem. Soe. 63, 141 (1891). 
