276 Mr. Walker's Observations on the 
repeat this, until the addition of snow or powdered ice will not 
make the thermometer rise : to be very accurate, it should be 
reduced in this manner to the proper strength, at the tempera- 
ture, whatever it be, at which the nitrous acid and snow, or 
powdered ice, are to be mixed together when cooled. 
In the course of my experiments I have endeavoured to as- 
certain the comparative powers of ice to produce cold with 
nitrous acid, in the different forms I have had occasion to use it. 
The result is, that fresh snow sunk a thermometer to — 32 0 ; 
ground ice to — 34, 0 ; and the most rare frozen vapour to be- 
low — 35 0 ; the vessel and materials each time being -f 30°. 
The vessels for these mixtures, particularly that in which 
the quicksilver is to be frozen, should be thin, and made of the 
best conductors of heat ; first, because thin vessels rob the mix- 
ture of less cold at mixing, i. e. if two mixtures of the same 
kind are made, one in a thin, the other in a thick vessel, the 
former will be coldest; secondly, because the air is a suffi- 
ciently bad conductor ; and thirdly, for the very obvious rea- 
son, that the cold is transmitted through them quicker. 
For these reasons, and from the difficulty I have found in 
procuring vessels of glass, which are undoubtedly fittest for 
experiments of this kind, I have used tin ; which is readily had 
in any form, and if coated with wax, is sufficiently secured 
against the action of the acids. 
I give the inside such a coating, by pouring melted white 
wax into the vessel, previously clean and dry, and turning it 
about by hand, so as to leave no point of the metal uncovered 
for the acid to act on, pouring the surplus away. 
In the experiment above described, I used a single vessel 
for cooling the nitrous acid ; a cupping-glass (represented by 
