892 Scientific Intelligence. 
was sealed, and at the same time the height of the barometer H 
was noted. The bulb was then taken to a room of uniform tem- 
perature provided in the laboratory for gas analysis, and after 
being mounted on a convenient support, the end of the stem was 
broken off under mercury, and the apparatus left to itself for a 
time to secure a perfect equilibrium of temperature. This tem- 
perature (T’°) was then observed by means of a standard ther- 
mometer hanging near the bulb; also the height (h) to which 
the mercury had risen in the bulb was measured by a cathetom- 
eter, and in addition the height H’ of the barometer (hanging in 
the same room) was again noted. Closing now the open stem 
with the finger, the bulb was quickly inverted and the pistes | 
mercury drawn out into a tared vessel and weighed (nipping o 
the end of the shorter stem in order to admit the air). Thi 
T?+273°2=(I"°-4-273°2) alr —T')°k}. 
It will be noted that as the mercury columns, including the 
heights of the barometer, were all measured at the same constant 
temperature, and, as we are dealing with relative values only, no 
reductions are necessary. Moreover, an error of one-tenth of a 
millimeter in the value of Hy Would make in determining the 
boiling point of sulphur (448°) a difference of only one-eighth of 
a degree, so that measurements of these heights are sufficiently 
close if accurate to one-half a millimeter, and might even be 
ith a common rule. The most uncertain element in the 
formula is the expansion of glass, but if the bulbs are made of 
flint glass (lead glass) tubing, such as is used for ornamental 
glass 8 to soften. e rate of expansiou of flint glass is not 
only less than that of crown, but it is also more constant and 
pared with the results of Regnault reduced to the corresponding 
pressures : 
