782 PROF. W. H. MILLER ON THE CONSTRUCTION OF THE NEW STANDARD POUND. 
The thermometer R was used in some of the earlier observations. It has an ivory 
scale the dimensions of which vary with the quantity of moisture present in the 
atmosphere, and consequently the error is very sensibly different at different times. 
By a mean of twenty comparisons of R, B in October 1846, their corresponding 
readings were : — 
R. B. B-R. 
19-31 18-90 -0-41 
But in October 1846, at 19°, B= — 0’15. Therefore R= — 0°’56. 
Comparison of Barometers. 
Up to the end of August 1844 a siphon barometer by the late Mr. Robinson was 
employed. It resembles Bunten’s improved Gay-Lussac’s barometer in all respects 
except that it is graduated in English inches, and the attached thermometer in 
degrees of Fahrenheit’s scale. The observations were reduced, for the mercury to 
32° Fahr., and for the scale to 62° Fahr., by the tables in Schumacher’s Jahrbuch 
fiir 1837. 
From the beginning of September 1844 a cistern barometer by Ernst of Paris was 
used. It is graduated in millimetres, and the attached thermometer in centesimal 
degrees. In the following comparisons with the barometer of the Paris Observatory, 
made by one of the Assistants of the Observatory, O, T denote the readings of the 
Observatory barometer, and of its attached thermometer; F, E those of Ernst, and 
of its attached thermometer. 
0. 
T. 
F. 
E. 
0-F. 
mm. 
mm. 
mm. 
753-16 
21-4 
752-90 
21-0 
0-26 
753-64 
21-8 
753-30 
21-3 
0-34 
754-00 
22-3 
753-60 
21-7 
0-40 
754-14 
22-3 
753-70 
21-7 
0-44 
754-32 
22-5 
753-90 
21-9 
0-42 
757-40 
21-1 
757-00 
20-9 
0-40 
756-90 
22-4 
756-50 
21-7 
0-40 
758-64 
19-5 
758-25 
19-3 
0-39 
758-20 
21-7 
757-70 
21-1 
0-50 
Mean 0-393 
Ernst stands 0*393 millimetre lower than the Observatory barometer. The latter 
requires no correction. Ernst was suspended close to the Observatory barometer 
all night previous to the day on which the comparisons were made, their tempera- 
tures must, therefore, have been very nearly equal. Yet E is less than T, while, by a 
subsequent comparison with B, E was found to be 0°*45 too great. There is reason 
to believe that this discrepance is due to the lodgment of a small quantity of mercury 
in the upper end of the tube of the thermometer, which occurs sometimes after the 
instrument has been conveyed in a carriage over a rough pavement, as was done pre- 
vious to its comparison with the Observatory barometer. 
In March 1845 the thermometer B was suspended so that its bulb was in contact 
