( 947 ) 
fig. 2) which may be seen in use in the apparatus used for the 
experiments of. Comm. N°. 88 (Jan. ’04) and N°. 92 (Sept. ’04) was 
needed neither for our present experiments nor for those on the 
compressibility of hydrogen vapour. The reservoir 4’ (F of N°. 84, 
Pl. Il) has a separate mercury holder Q, (Pl. I) and the capillary 
connecting this reservoir with the volumenometer is so constructed 
that the gas contained in it may be driven completely out by the 
mercury. A second tap #, was put in the fork of the tube connect- 
ing the volumenometer with the reservoir /”; when gas bas been 
admitted into the volumenometer from the reservoir J” the mercury 
is allowed to rise above the tap 4, of the reservoir /” and in the 
volumenometrie determinations the volume of the mercury that has 
risen in /—k, is allowed for by reading the position of the mercury 
from the graduations of the calibrated capillary. 
The dead space consists of the portion of the volumenometer that 
projects above the level (§ in Pl. 1) of the liquid in the bath in 
which it is placed, and of the spaces between the volumenometer 
and the taps of the apparatus that are connected with it (for instance, 
hk, and #, Pl. D (and usually too, the dilatometer containing that 
portion of the gas that remained after the greater part had been 
transferred to the volumenometer), but the great advantage of con- 
structing the reservoir /” in the above way lay in the fact that in 
the calculation of the masses that are measured, “this dead space 
occurs only at low pressures, so that it is not necessary to determine 
its temperature so accurately. In fact, nearly all the gas to be 
measured can be transferred to the reservoir /’ which is evacuated 
beforehand and then filled with mercury, and if the volumenometer 
is used as if it were a pump there remains to be measured in’ the 
dead space only gas that is at a very low pressure. Then, keeping 
k, closed, the volumenometer is evacuated through #%, and &,,, and 
the gas that has been stored in the reservoir /” is transferred to it: 
this large quantity is then measured at a temperature that is known 
With great accuracy. 
The connections between the manometer tube J/ and the large 
reservoir # in which a pressure is sustained equal to, or if one so 
desires, slightly different from atmospheric had also undergone some 
modification for the research on hydrogen vapour compressibilities. 
Amongst other alterations we may. mention that the connections 
which are to withstand a vacuum are either ground or fused together 
so that it is possible to measure volumes over the whole volumeno- 
meter at all pressures between O and somewhat more than one at- 
mosphere. Moreover, a connection may now be made between the 
