142 
MESSRS. W. R. BOUSFIELD AND C. ELSPETH BOUSFIELD 
This relation would indicate that the specific heat lowering of the solution is 
proportional to the specific contraction of the water. We may compare it with the 
before obtained relation 
hAs, 0 = — 69‘30 + 2'074 x , 
which may be written 
As 20 = 37-37 ( x —33-40)//T, 
whilst the new relation may be written 
As 30 = 38’2 ( x —33‘5l)/i7. 
These are obviously the same relations, though the different paths by which they 
are obtained bring out slightly different values for the constants. In Table XIV. are 
o-iven the calculated values for A s 20 from the above relation, with the differences. 
Though these are within the range of experimental errors yet the course of their 
variation indicates that the straight line may still be only a close approximation, and 
one is therefore led to look still further. 
