AND GEOLOGICAL RELATIONS OF CERTAIN HOT SPRINGS. 
583 
If we now project these errors as before, and form an interpolating 1 curve, we shall 
obtain the following 1 Table of Corrections applicable to the scale of Crichton’s ther- 
mometer. The observations made with it are to be considered simply as checks upon 
the general accuracy of the other, and will not finally be employed, unless in the few 
cases where Troughton’s thermometer has not been used. 
Table of Corrections applicable to Crichton’s Thermometer. 
j Reading. 
Correction. 
Reading. 
Correction. 
o 
o 
o 
° !. 
32 
0-0 
80 
— 1-2 
40 
— 0-4 
90 
- 1-7 
50 
- 0-7 
100 
- 2-0 
60 
- 0-8 
110 
— 2-1 
70 
- 0-9 
120 
- 2-9 
There is but one other thermometer specifically referred to in the following paper, 
and that only in two instances. So long ago as the year 1826 I took the temperature 
of the hot spring in the cavern of Nero’s baths near Naples, with a thermometer made 
by Cary, and not belonging to myself. I have recently had an opportunity of ex- 
amining 1 this identical thermometer. I restricted my experiments to determining 
the error at the particular temperature at which it had been employed, viz. 183°‘5, by 
a comparison with my standard Troughton. To make the comparison I plunged 
both instruments in diluted alcohol in a state of ebullition, just enough of water being 
added to bring it up to the required temperature. I satisfied myself by many expe- 
riments that the indications were sufficiently constant, and fixed the error of the 
Cary’s thermometer at + l 0l 3 (at temp. 183°). I had stated when I first published 
the experiment in Dr. Brewster’s Journal*, that I conceived the indications of the 
thermometer to be about a degree too high, which was thus nearly confirmed. This 
thermometer was likewise employed (together with another now broken) to determine 
the temperature of the spring of La Pisciarella, also near Naples, which was 114° on 
its scale. By another special comparison with Troughton’s thermometer, by plunging 
them in hot water (intended only as an approximation), I found the error of Cary 
at 115° to be + 2 0l l by several concordant observations. To avoid an affectation of 
accuracy which the original observation would not warrant, 1 shall simply reduce 
the indication of Cary by 2°. 
§ 2. Springs of the Pyrenees. 
The following observations were made in the months of July and August 1835. I 
follow the natural order of the springs from west to east. 
* New Series, vol. ii. p. 90. 
