24 
Remarks on Leslie's Differential Thermometer. [Jan. 
This subject tempts me to offer some remarks on the scale of Leslie s Differential 
Thermometer, which protean instrument, in its hygrpu.etnc form, you have discarded 
from your meteorological table in favour of the moistened bulb thermometer, -not. 
withstanding the extravagant anonymous commendations of it bv its inventor, it he 
may so be called, as the only difference between it and Van Helmont s Thermome- 
ter consists in an alteration which is no improvement to its ..ci wi.u y . 
Van Helmont's Thermometer was in the shape of the letter U, one end h ingopen 
to the atmosphere, while the other was furnished with a bulb, which bulb being 
filled with air, and Ilia tube under it with a coloured fluid, any increase of tempera- 
te 
filled — , ... 
ture in the bulb was accompanied by a rise of the coloured fluid in the opv»*dc tu >e. 
While the atmospheric pressure continued uniform, this instrument would form, 
bating the slight variations of hydrostatic pressure in the open tube, a perfect ther- 
mometer, each additional degree of heat acquired by the bulb being indicated by an 
equal rise of the fluid in the open tube. But as the volume of the air in the bulb 
varied not only with the temperature, but with the pressure also of the atmosphere, 
its value, as a thermometer for general use, was thereby greatly diminished. 
An obvious mode of excluding this variable pressure was to shut the open end of 
the tube; adapting to it, of course, a bulb full of air, to allow, by its compressibility, 
of the rise of the coloured liquid. This was Leslie’s alteration ; but as he .-.tiUcon- 
tinued to attach a scale of equal parts to the instrument, it v as any thing hut An 
improvement : in fact the instrument then ceased to be a thermometer, sinking 
into a mere thermoscope, which indicates a difference of nearly 6'0 centesimal de- 
grees, when the actual difference of temperature between the two bulbs is onlv 50° ; 
thus erring by 18° of Fahrenheit’s scale ! I have had no opportunity of experimen- 
tally verifying this conclusion; but its truth, though hitherto unnoticed, will appear 
from the following considerations. 
Two points, in Leslie’s scale, are necessarily connect, the zero being marked when 
the temperatures of the two bulbs is the same, and 1000 ° when the difference be- 
tween their temperatures is 100 ° of the Celsian scale, or 180° of Fahrenheit’s, the 
former decupled having been adopted by Leslie, probably to facilitate the introduc- 
tion of his instrument upon the continent. But he divides the space between ti° and 
1000 ° into equal parts, forgetting, that while the liquid rises iu the scale-tube, the 
increasing re-action of the compressed air in the bulb of that side diminishes the 
extent through which each successive degree of temperature, added to the warmer 
ball, is able to propel the fluid ; and that the scale ought therefore to he logarithmi- 
cally graduated, the lowest degrees being made about twice as large as the highest 
The ratio of diminution is thus ascertained. All gases in being heated through 
180° of Fahrenheit’s scale, acquire an addition of § to tliftir volume, when exposed to 
a uniform atmospheric pressure. Therefore, as 8 : 11 : : 1000 : 1375 . the index 
which rises through 1 000° would rise to 1375° of Leslie’s equafde- rees ’if the tube 
degrees up to 1000° !h* so 
Leslie s 6 P ° mt 1000 ° ° f the new sca,e sliaI1 coincide with the point 1000 ° in 
Geometrical 
Division in 
millesimal 
degrees. 
Arithmetical 
division in 
millesimal 1 
degrees. 
0 ° 
0 ° 
10 
13,7 
20 
27,3 
30 
40,8 
40 
54,2 
50 
67,5 
60 
80,7 
70 
93,8 
80__ 
106,8 
Difference 
of 
Geometrical 
division. 
80° 
100 
110 
120 
130 
140 
150 
1U0 
1000 
Arithme- 
tical divi- 
sion. 
Diff. 
1 10 °, 7 
5°, 3 
132 5 
5 ,8 
145 .9 
6 ,3 
157 ,9 
6 ,8 
I/O ,4 
7 ,8 
182 ,8 
7 ,7 
lbo ,1 
8 ,1 
207 ,3 
8 ,5 
1000 
0 i J 
The subjoined table, constructed on the above data, shews the respective indiea- 
ratter tl 5'“”* a " d *»>• P™P»sed geometrical 
H Ol,g l0U the ordinary hygrometric range ; stating also, in degrees of 
Fahrenheit, the amount of error in the former. aegrtes oi 
