578 
PROFESSOR FORBES ON THE TEMPERATURES 
fixed, notwithstanding the celebrity of the makers, especially as I knew that the 
freezing point had undergone that permanent change which I have observed in most 
of the thermometers which I have examined. 
I first fixed the standard points, and then determined the intermediate points of 
the scale by a simple and, I believe, a new method. 
The freezing point was fixed by supporting the thermometer vertically in a tall 
glass vessel partially filled with pounded ice, freshly made by Leslie’s process. This 
vessel was placed in a saucer filled with ice-cold water, in order to keep it externally 
at the same temperature. By a mean of seven observations, none differing 0 o, 2 from 
the mean, and continued during an hour and a half, the freezing point was fixed at 
32°\33 on the scale. 
The boiling point was fixed with great care on two different days. A tall tin vessel 
was provided, with a steam orifice and in which eight ounces of water were placed, 
(spring water the first day, distilled water the second,) the ball of the thermometer 
being placed in the steam a few inches above the surface of the water. It is assumed 
that 212° ought to coincide with the boiling point of water under a pressure of thirty 
inches. 
Date. 
Barometer. 
Boiling point 
by Trough ton. 
Reduced to 
pressure 30 in. 
Error. 
1835. 
Nov. 7. 
Nov. 19- 
29-582 1 
Att.Th. 51°-0 J 
29-636 
Att, 
29-636 \ 
.Th. 50°*4 / 
210-9 
210-98 
211-61 
211-59 
Error at 212°. . . . Mean 
And we had before. Error at 32 3 .... Mean 
0-39 
0-41 
— 0-40 
+ 0-33 
The principle of the determination of errors at the intermediate points is the same 
as that of Bessel % viz. causing a detached column of mercury to traverse the tube, 
but is simpler in practice. Instead of employing columns of mercury quite arbitrary 
in point of length, and deducing by a complex and tentative process portions of the 
tube of equal capacity, I proceed at once in the following manner. 
I detach a column of mercury from the rest, (by a known method -f',) of such a 
length as to be nearly an aliquot part of 180°, which may be done with great accu- 
racy. I then cause it to step along the tube, the lower extremity of the column being 
brought successively to the exact points which the upper extremity had occupied, 
noting carefully these points. At length (having started from 32°) the upper end of 
the column coincides nearly with 212° if its length has been properly chosen. 
* See the Philosophical Magazine, vol. lxiii. p. 307. 
f A column of any length being detached from the body of mercury in a tube of moderately wide bore, may 
have its length adjusted with great accuracy by bringing the divided part just into contact with the remainder 
of the mercury, and at the same moment heating or cooling the ball. If heated, the detached column will be 
enlarged ; if cooled, some mercury will be abstracted from it. 
