* 
139 
series of other curves relating to these continuously varying phases. 
b. If we imagine the conversion of « into y to come about, 
because primarily the a@ particles dissolve and y crystallizes from 
this solution, then it is possible that these two processes take place 
with very unequal rapidity. If the crystallization came about very 
rapidly in proportion to the solution of a, the values situated between 
curve I and curve [Il would be reached. 
As we said before however, the crystallization takes place extremely 
slowly, whereas the solution of the particles of product a happens 
more rapidly. Therefore explanation 6 does not hold true and as 
most probably it will not be easy to find a more plausible explanation, 
I believe I may best interpret the experiment by accepting a continuous 
transition between the solid phases that coexist on curves I and III 
with the solutions. For curve II this solid phase is the erystallized 
aluminiumhydroxide Al,O,.3H,O as microscopical examination as 
well as analysis of the rests prove. If we consider the rests of the 
points situated on curve I (Nos. 1—11) it then appears that the lines 
that unite solutions and accessory rests, do not display a mutual 
point of intersection inside the triangle. The gelatinous phase that 
coexists along curve [ by the side of the solutions, must therefore be a 
product of varying composition, rich in water and which besides 
(see analysis) has absorbed a certain quantity of alkali. Of course 
curve I is quite metastable with regard to curve ILI, but can be 
determined very well owing to the slight rapidity of transition. 
If we now survey the determinations performed with product g, 
it appears that they form curve II, which is situated between I and 
UI (fig. II, curve BE); besides the particles swell strongly in the 
more concentrated solutions, not appreciably in the diluted solutions. 
In any case we must therefore consider the form @ as metastable 
with regard to y; even though, in spite of numerous attempts I have 
not succeeded in experimentally realizing the transition By. Funda- 
mentally product 8 has nothing remarkable, it is but one of the 
many forms between I and III, which forms are all metastable as 
regards the crystallized hydroxide y. As to aspect, gis already much 
more like product y than a; this too pleads in favour of the 
continuous transition of the forms into each other. 
The equilibriums of Al, O, with NaOH-solutions of various 
concentrations. 
The determinations performed with the products d, ¢ and & dis- 
tinctly indicate that af 30° A/,O, is metastable as regards Al,O, .3H,0. 
The solutions in which product £ was added as a solid phase, give 
