46 On definite Proportions. 
If, on the other hand, we attempt to pulverise the same salt 
equally well dried, but consisting of larger crystals, we obtain a 
damp ‘powder, sorhewhtit cohering, from containing a portion of 
the mother liquor, that oozes out of the pores, and causes the 
appearance of moisture. . If we dry this again, we may then re- 
duce it to a state of dust; except in the case of particular salts, 
which exhibit some degree of cohesion even when perfectly dry. 
The water mevhanieally included im crystals is generally ex- 
pelled by the effect of heat, together with the water of erystalli- 
zation ; ‘but in the salts which contain no water of crystallizatien, 
and which are exposed in large masses to a high temperature, 
the water is suddenly changed into vapour, and causes the salt 
to decrepitate. Consequently the water thus expelled cannot 
have been water of crystallization, since it must have been un- 
equally distributed through the crystal,—a condition incompatible 
with chemical combination. For this reason, the estimates of 
the quantity of the supposed water of crystallization in decrepi- 
tating salts have been extremely various, according to the mag- 
nitude of the crystals examined, the larger crystals affording the 
greater proportion. 
That decrepitating salts might contain water of crystallization 
is not altogether impossible, for the water of crystallization might 
be expelled at a higher temperature than that which produces 
decrepitation: but as far as we have hitherto examined, the com- 
binations which retain water the most obstinately are always 
fused at a moderate temperature, and the water which is only 
united mechanically to the crystals is expelled by boiling. 
I have premised these observations, in order to show how dif- 
ficult it may be to determine the quantity of the water of ery- 
stallization with perfect accuracy. But hereafter this determi- 
nation will be rendered very easy : for, when the law of the water 
of crystallization is once de termined, some experimental approxi- 
mations only will be required, in order to determine among the 
different suppositions which can possibly be admissible : and this 
law | imagine that I have established in the experiments which 
are now to be related. I have generally employed granular ery- 
stallizations ; and if the salt was not much disposed to crumble, 
} have rubbed it into a fine powder, suffered it to remain exposed 
for some hours to dry air, and then heated it in a crucible of 
platina, of which the weight, as usual, had been accurately deter- 
mined. I have also examined different forms of crystallization 
of the same salt, but have been unable to find any difference in 
the quantity of water contained in them. 
The principal subjects of the present division of my essay are 
four: jirst, Laws for the combinations of water with acids, 
bases, 
