( 470 ) 
The line 5 is the solubility line of a metastable form of the solid 
compound; this line, however, can only be partially determined in the 
presence of an excess of pyridine. With a large excess the stable 
form was formed too readily. If the line could have been continued, 
it would have been continuous, in distinction from 1, as it is situated 
entirely below the mixing point M. 
It could not be determined at the side of the mixtures which are 
richer in CH,I as these erystallise very slowly and then we always 
obtain the stable form. 
The most noteworthy result of this research is, however, that two 
liquids which are miscible in all proportions, may yield two sets of 
coexisting liquids owing to the formation of a chemical compound. 
In the formation of less stable compounds such has never as yet 
been observed and the better known stable compounds have not as 
yet been studied from this point of view chiefly because the com- 
parison of combined and uncombined liquids is so often rendered 
difficult by the great differences in the melting points of the components. 
The sharp intersection of the melting point lines 1 and 3 at 117° 
and the strong elevation of the boiling point after the combination 
(pyridine 116°, CH,I 42°, combined liquid 270°) show that even in 
the liquid state, the compound is certainly for the greater part undis- 
sociated. . 
Probably the partial miscibility of this combined liquid with its 
components is connected with the fact that the chemical nature of 
the compound differs so greatly from those of the components. On 
this point also we possess but very little knowledge at present. 
Chemistry, — “The reaction of Frepen and Crarts’. By Dr. J. 
BörseKEN. (Communicated by Prof. A. F. HorLEMAN). 
As is well known, the reaction of FRisppr and Crarts does not 
always proceed uniformly. Sometimes traces only of the catalyzer 
seem to suffice for the preparation of large quantities of the desired 
product; in other cases equimolecular quantities of the products to 
to 1’,, mol. of the reagent. Inva 
be condensed require from !/, 5 
20 
great many condensations it has been shown that at least 1 mol. 
of AICI], is required in order to obtain the highest possible yield. 
The reaction is also dependent on a number of circumstances 
which are either connected with a secondary action of aluminium 
chloride (Ree. XXII p. 302) on one of the substances present during 
