( 1099 ) 
instance, however, the favourable circumstance arises that only a very 
small quantity of gas has to flow over, seeing that the dead space 
is extremely small. Against the widening of the capillary it may be 
urged that then the quantity of gas contained in it would lead to 
inaccuracy owing to the uncertainty existing regarding the distribution 
of temperature along it. After full consideration of the change of 
viscosity and density with temperature, and also of constructional 
difficulties, the low temperature portion of the capillary was made 
of 37 em. steel capillary of 0.5 mm. bore, and the part that is at 
practically room temperature was made of 50 em. copper ‘capillary 
of 1.0 mm. bore. The resulting uncertainty is, then, at the most, 
1°/,, while the viscosity is not yet excessive, seeing that it is 
possible to adjust to 0.01 mm. within a period of 2 minutes. 
§ 3. Densimetric Apparatus. The part of the dilatometer that was 
immersed in the helium bath consisted of a reservoir A, with a 
stem A,, a narrow glass capillary A, continued by a steel capillary. 
The mass of helium here present was determined volumetrically in 
the bulb V, with a graduated stem both above and below, whose 
temperature was determined by that of the surrounding water bath; 
the pressure was read on a seale by using the branch Vs, of the 
mercury reservoir }7,. The dimensions of 4, and WV, are so chosen 
that the position of the mercury for the desired pressure can be 
read on the lower part of V‚’s graduated stem before the dilatometer 
has been filled, and on the upper part after the filling has taken 
place. Moreover, the cross-section of the graduated stem has been 
chosen of such a size that when the dilatometer has been cooled 
again with Ay, and Ag, closed, after filling it at the boiling point 
to above the mark, the meniscus still remains in the stem even at 
the greatest densities employed. 
Although the capillary is very narrow at the part where its 
temperature is uncertain, the correction for the gas condensed from 
it when the dilatometer is cooled (keeping Ka, and 4, closed), 
which operates so as to cause a rise of the liquid meniscus in the 
stem, is of great importance when the question arises as to whether 
a maximum density can be shown to exist for helium or not. A 
second apparatus is therefore introduced having a very short stem 
and a similar capillary, but without a reservoir; to this the volu- 
menometer JV belongs. When the correction for the capillary of this 
control apparatus is calculated in exactly the same way as for the 
capillary of the dilatometer, it shows a rise of the level of the liquid 
that appears in-the tube, and by comparing this with the observed 
