1 68 Mr . Cavendish’s Account of 
and be again expofed to the cold, as the liquor is then in con- 
tact with the congealed matter, it begins to freeze as foon as it 
arrives at the freezing point, and the ice becomes much more 
folid and compact. 
The intent of decanting the fluid part, and fending both 
parts back, that their flrength might be determined, was 
partly to examine the truth of the fuppofltion laid down in my 
former Paper, that the flrength of the frozen part approaches 
nearer to ,411 than that of the unfrozen; but it is alfo a ne- 
ceflary Aep towards determining the freezing point anfwering 
to a given flrength of the acid ; for as the frozen part is com- 
monly of a different flrength from the unfrozen, the flrength 
of the fluid part, and the cold neceflary to make it freeze, is 
continually altering during the progrefs of the congelation. 
In confequence of this, the temperature of the liquor is not 
that with which the frozen part congealed ; but it is that ne- 
ceflary to make the remainder, or the fluid part, begin to 
freeze, or, in other words, it is the freezing point of the fluid 
part. This is the reafon that a thermometer, placed in fpirit 
of nitre, continually finks during the progrefs of congelation ; 
which is contrary to what is obferved in pure water, and other 
fluids in which no feparation of parts is produced by freezing. 
Moreover, from the above-mentioned experiments of Mr. 
M c Nab it appeared, that oil of vitriol, as well as fpirit of 
nitre, is fubjedl to the fpirituous congelation ; but it feemed 
uncertain, whether, like the latter, it had any point of eafieft 
freezing, or whether it did not uniformly freeze with lefs cold 
as the flrength increafed. For this reafon, fome bottles of oil 
of vitriol, of different ftrengths, were fent, which he was 
defired to try in the fame manner as the former. This point, 
indeed, has fince been determined by Mr. Keir, who has 
fhewn 
