328 
[Assembly 
observations, particularly in reference to the temperature of the mer- 
cury, which is hable, without extreme care, to a false indication by the 
attached thermometer, of a number of degrees. I have ascertained that 
the condition to be satisfied in order to be assured of accuracy in this 
respect, for any syphon barometer, is contained in the equation 
in which a, b, are respectively the upper and lower readings, T the 
temperature of the mercury as indicated by the attached thermometer, 
V the distance of the superior mercurial surface, and A, B, C, D, co- 
efficients which differ in different barometers, but are constant in the 
same. The appropriate conditions for the barometers No. 275 and 
No. 366, the former of which was used at Burlington, I have found to 
be, respectively, a — b = — 2,17 + 0,107 T, (1) 
and a' — b' = 35,14 + 0,107 — 0,004 (402 — a) (2) 
These formulae have been employed in rejecting some of the faulty ob- 
servations referred to above, and, assuming the correctness of T, in 
correcting the elevations of Lake Golden and Mt. Marcy, where the 
conditions expressed in (2) were not satisfactorily answered. 
As is not uncommon, even with good instruments, the column of 
No. 275 exceeded that of No. 366 by 2.50 millemetres, which I conse- 
quently added to the sum of the upper and lower readings of the supe- 
rior barometer. This difference between the columns is a mean de- 
rived from a comparison of more than 100 sets of observations, in 
which care was taken to secure as great a degree of uniformity in the 
temperatures of the atmosphere and mercury as possible, and to exclude 
all causes of change in the columns which were not equally operative 
in each, except those depending upon peculiarities in the constructions 
of the instruments themselves. 
Various experiments, which it is needless at present to detail, sug- 
gested the possibility that a part at least of this difference of columns 
might arise from a small portion of air in the summit of the tube of 
No. 366 ; and that consequently the correction above, instead of being 
constant, would depend upon the temperature and volume of the in- 
cluded air. On this hypothesis, which however I was prevented from 
verifying to the extent desired, by the loss of one of the barometers, I 
made the correction (451 — a) x 0,021 x (16 — T), in metres, which 
is additive or substractive according as T is less or greater than ]6. 
But to whatever extent this correction might differ from the truth, the 
