40 J. P. Cooke—Buoyancy of the Atmosphere. 
here assumed, we can easily correct for temperature by simply 
adding to the observed height of the barometer (in tenths of an 
inch) the difference between 27°C. and the temperature ob- 
served. Of course the correction becomes negative if the tem- 
perature exceeds 27°C. Having thus eliminated the effect of 
temperature we can (after taking a few weighings under as 
great a variation of temperature and pressure as we can com- 
mand) easily find the difference in weight which corresponds to 
a variation of one-tenth of an inch in the barometer, an 
thus obtain a constant for the vessel (or other object weighed) 
by means of which we can rapidly reduce the weights to the 
standard of thirty inches barometric pressure, having previ- 
ously reduced them to the standard of 27° C. for temperature. 
The weights, having now been corrected for buoyancy, can be 
compared, and although the standards may be as unusual in 
their association, as is one of them in its value, they are as 
legitimate as any others and will be found in practice more 
t 
-convenien 
To apply this method we simply leave the load equipoised 
on the balance, shifting the rider with the varying weight, and 
noting the corresponding temperatures and pressures, until a 
sufficient difference has been observed; and a difference corre- 
sponding to 20° C., or two inches of mercury, is adequate in 
most cases. ‘The process corresponds to calibrating a flask, and 
the constant, once obtained, can be afterwards used for the 
same vessel, unless the weight of its contents is materially 
altered. The following examples will show the application of 
the method. 
In each case the load was a closed absorption-tube of pecu- 
larly irregular construction, but not much larger in volume 
than those generally used in organic analysis. We give in the 
accompanying tables: first, the date; secondly, the observed 
weight; thirdly, the temperature of the balance-case; and, 
fourthly, the height of the barometer at the time of weighing. 
These are the observed data. In the fifth column we give the 
reduced heights of the barometer for 27° C., and these values 
are obtained by simply subtracting the observed temperatures 
from 27°, and adding the remainders to the observed baro- 
metric heights. Below the tables we print in each case the 
largest weight observed over the least weight observed, and on 
the same lines the corresponding reduced barometric heights. 
Dividing, now, the difference of weight in milligrams by the 
difference of height in tenths of an inch we obtain the value 
last given, which we then called the “constant.” With this 
constant we can readily reduce all the weights to the common 
standard of thirty inches, and this we do by multiplying the 
difference between 300 and the reduced barometric heights by 
