CPIANGE IN WEIGHT DUEING CHEMICAL REACTION. 
‘237 
Granting, then, the existence of air streams which vary both in direction and 
intensity as proven, the question which now confronted us was : How can these air 
streams be best suppressed, or their effects neutralised ? As it appeared impossible 
to institute any simple or convenient scheme for neutralising the disturbances, an 
attempt was made to eliminate their cause. 
As a result of earlier experiments, it was known that a delicate bolometer placed 
within an ordinary balance case almost invariably reveals the existence of continuous 
fluctuations in the temperature of the contained air. In general, the fluctuations 
assume the form of minute and somewhat irregular oscillations about some mean 
temperature. Such disturbances occurring in the vicinity of the beam may be 
reduced to a vanishing point by the simple expedient of enclosing the whole beam 
with a small additional inner case. With these facts before us, it was concluded that 
the difficulties arising from the presence of air streams might be successfully met by 
completely enclosing the pans and stirrups by means of shells having a very high 
conductivity for heat. Theoretically, the following very distinct advantages are 
gained by the use of such shells, more particularly if the exterior surfaces are polished 
and the interior coated with carbon :— 
(1) The effective volume of the disturbed air is very greatly reduced ; 
(2) A shell having a high conductivity for heat may, when placed within a balance 
case, be regarded as possessing throughout a strictly uniform temperature ; 
in consequence of this, combined with the fact that the inner or carbon skin 
has the power of rapidly absorbing and radiating heat energy, the contained 
air will also acquire the same uniform temperature; 
( 3 ) Radiant energy falling upon the outer polished surface of the shell will be 
almost entirely reflected; the insignificant fraction that may enter will be 
incapable of producing any appreciable disturbing effects in the air enclosed 
by the shell. 
All these advantages were effectually secured in the following way:— 
Four half cylinders of stout sheet copper were prepared, the diameters of two 
being made a trifle less than the diameters of the other two. Using two half- 
cylinders, one of each size, a complete closed cylinder could at any time be formed Ijy 
so placing the two portions that the vertical edges of the wider half slightly over¬ 
lapped and just gripped the two corresponding edges of the narrower one. The 
height of the compound cylinders was as nearly as possible equal to the vertical distance 
between the upper surface of the base of the ordinary balance case, and the lower 
surface of the aluminium base-plate of the auxiliary beam case above. Before being 
used their inner surfaces were coated with Indian ink and their outer surfaces rough 
polished with fine sand-paper. The device was now tested. First, the reaction 
vessels were re-wiped and again suspended from the arms of the balance; then the 
