and Attached Water. 367 



urinations for the solubilities above 0° C. by adding 0*1 to his 

 percentages. 



Table XXVI. 



AgN0 3 



H 2 



Temperature 

 at which soli- 



Nature of 



per cent, by 



per cent, by 



dification 



solid formed. 



weight. 



weight. 



begins. 





10 



90 



o 

 - 0-8 



Ice. 



20 



80 



- 2-7 



n 



30 



70 



- 4-7 



>i 



40 



60 



- 60 



>> 



48-3 



51-7 



- 6-5 



Cryohydrate. 



50 



50 



- 5-5 



AgN0 3 . 



53 



47 



- 2 2 



it 



55 



45 



00 



tt 



(Kr.) 69-4 



304 



+ 19-5 



» 



§ 138. Thoughts suggested by fig. 1. — As the whole field 

 of inquiry contained below the zero-line of fig. 1 has, I be- 

 lieve, remained hitherto unexplored, I am tempted to submit 

 a few remarks upon it. First, of course, the tracings should 

 be continuous curves between 0° and the cryohydrates, and 

 again between the latter and the boiling-point of the saturated 

 solutions, and doubtless, under pressure, beyond. Further, it 

 seems that the curves are all varieties of the same kind. The 

 figure 2 shows what we may liken to the backbone of the 

 curve, with a typical curve about it. There is no absolute 

 measure of what may be called the Fig. 2. 



eccentricity of the curve, because 

 there is no common measure be- 

 tween percentage composition and 

 temperature ; both are measured in 

 arbitrary units. But this does not 

 affect the type or order of the curve's 

 equation. On the left-hand branches 

 there appears to be a point of con- 

 trariflexure. The only apparent 

 exceptions to this are KI, alcohol, 

 and HC1. Concerning the two lat- 

 ter, doubtless we only see a fragment of their curves before 

 their final cryohydrates. If there be such a point with KI it 

 must lie between 40 per cent, and the cryohydrate, which is 

 well possible. There seems to be a point of contrariflexure 

 also on the right-hand branch, though sometimes the curvature 

 is so small that this point is practically obliterated. The ave- 

 rage slope of descent to the cryohydrate is invariably less than 



