62 oe REPORT—1857. . 
tube seemed to become irregular; and on solidifying, the mercury sank com- 
pletely into the bulb. 
A curious circumstance happened with this instrument on one occasion. 
Believing the quicksilver in it to be pure, I placed it beside one of Adie’s, ex- 
posed to the air at a temperature about the freezing-point of mercury, for the 
purpose of ascertaining the exact degree indicated by Adie’s at the moment 
of solidifying. Whilst attentively watching it, to my surprise the column of 
mercury suddenly shot up the stem to —4°, then slowly but steadily de- 
scended into the bulb. Though I heard no sound of the glass cracking, I 
thought the bulb had given way, and the entrance of air had forced the mer- 
cury up the tube; but in this I was mistaken, for having taken it on board 
and thawed it, nothing wrong could be detected, and it worked as well as 
before. ‘The explanation which offered itself to my mind was, that the sur- 
face of the mercury in the bulb becoming at once solidified, its contracting 
pressed the central and still fluid portion of the metal into the stem with a 
jerk, and thence again gradually absorbed it as the process of freezing ap- 
proached the centre. 
Whether these instruments by Adie were absolutely correct seems doubtful. 
In my Journal I find the following remarks regarding them :— 
“Feb. 2, 1854.—Temperature fell to —39° in the night, when I had a good 
comparison of the thermometers, those of Adie’s remaining within a degree 
of —39°, whilst a quantity of quicksilver in a teacup partially froze. The 
quicksilver remained out all night, and did not become completely fluid again 
until 9 a.m., when the temperature had been some hours at —36°, —35°, and 
—34°, A mercurial thermometer placed in it also stood at —34°, and the 
same one now blackened for exposure to the sun’s rays and enclosed in a 
glass case has fallen to —52°, i. e. become solid, whilst the one (Adie, 11) 
in constant use shows only —37°5.” 
“Feb. 3, 1854.—One of the new thermometers was kept in the vessel in 
which the quicksilver was exposed; and it remained all day at —36°, whilst its 
fellows showed —39° and —40°. On removing it a small portion of the soli- 
dified metal adhered to the bulb and still remains attached, although the tem- 
perature indicated by it and the others is —37°. The result of this is either 
that the mercury is impure, which I believe is not the case, or the instru- 
ments have an error of 3 or 4 degrees.” 
“ Feb. 4.—-The mercury adhering in the solid state to the bulb of the spirit 
thermometer remained in the same state until half-past two this morning, 
when it dropped off, that and the four other thermometers by Adie showing 
— 36°." 
From these experiments, I incline to the belief that an error of 3 or 4 
degrees will be found to exist in these instruments at the freezing-point of 
mercury. The quantity of metal in the teacup was several ounces, and there- 
fore too large, unless its indications be taken while partially solid either in 
freezing or melting. I have considered the dropping off of the small portion 
adhering to the bulb of the spirit thermometer as the best index. 
The mercurial thermometer alluded to as descending to —52°, was one at- 
tached to ascale apparently graduated regardless of accuracy ; but from some 
experiments made with it, I considered the tube was tolerably uniform in 
calibre; I therefore removed the scale, and attached another reaching down 
to within half an inch of its bulb. This scale was graduated by comparison 
with: Pastorelli (mercurial), as low as —32°, and thence the graduation was 
continued to the bottom of the scale in the same proportion, bringing it down to 
—50°, about two degrees below which the mercury always stood when solid. 
The spirit thermometer for use was placed in a tin cylinder 23 inches in 
3h we 
