AT NIGHT, FROM THE EARTH, ETC. 
of the scale ; I therefore determined to use this as a standard with which to 
compare every thermometer. 
I also found that it was absolutely necessary to have the division marked on 
the thermometer stems themselves ; for those to which scales were affixed, even 
those furnished with hinges, so that a part of the stem and the bulb were free, 
and those whose bulbs projected beyond the scale could not be laid so nicely on 
different substances as those without scales ; for that and other reasons all the 
thermometers used in the following experiments were without scales affixed to 
them, the divisions being engraved and coloured upon the stems themselves. 
In the course of the experiments I found that circular bulbs were the more 
sensible in proportion to their smallness, but with a bulb sufficiently small to 
have the desired sensibility, the column of mercury was so fine that it would 
have been impossible to observe accurately with it at night in the position in 
which the instruments were necessarily placed in these observations. 
The length of the thermometer finally used was thirteen inches, including the 
bulb, whose diameter was a quarter of an inch, and length three quarters of an 
inch (as shown in the figure) ; therefore, as the instruments were divided from 
0°to 130°, each degree was about 0T inch in length ; these were used during the 
night observations, and they were left on their respective substances till about 
9 o’clock in the morning, until it was found that many of them were broken ; the 
absorptive power of grass and the filamentous substances being such that before 
this time a temperature of more than 130° had taken place even in the month of 
April; other instruments were afterwards constructed both of the same length 
and of the same form, and graduated as far as 160° or 170°. These were occa- 
sionally used during the night observations, and always during those of the day 
in experiments upon the absorptive powers of different substances. 
These instruments were so delicate that on taking them from air of the tempera- 
ture of 60° to that of 37°, the latter temperature was indicated in about two mi- 
nutes ; and therefore if at any time a lull took place in a gale of wind of two 
minutes’ duration, or even less, the amount of heat lost by radiation under the 
then state of the sky would be correctly registered by these instruments. 
On September 13, 1843, I received twenty-five mercurial thermometers of the 
above form, and as many self-registering minimum thermometers, with circular 
bulbs, whose divisions were also on their own glass stems ; and at this time I com- 
menced the regular series of observations with the mercurial thermometers ; that 
with the minimum thermometers had been begun long before. A year was con- 
sumed in these preliminary experiments, and in ascertaining the precautions 
necessary to obtain correct determinations ; in consequence of several of these 
being neglected before this time, the previous observations must be regarded as 
undeserving of confidence ; the results from them therefore have been omitted in 
the following Tables. 
I now proceed to speak of the comparison of the thermometers with the 
r 2 
121 
