242 
MR. J. J. MANLEY ON THE APPARENT 
may be formed upon the several volume-compensators discussed in Cases 1 and 2 
(pp. 239 and 240) to be as follows ;— 
Case 1.—Splierical compensator. Water skin . 
Short cylindrical compensator. Water skin. 
Long 5, ,, J, 5, • 
Case 2.-—Spherical compensator. Water skin . 
Cylindrical ,, ,, . 
= ‘Old mgr. 
= -017 
= -029 
> 5 
?? 
= -006 
= -007 
It is almost superfluous to point out that the density of the water skin will be 
subject to fluctuations; it is, in fact, a variable dependent upon three others ; namely, 
the pressure, temperature, and hygrometric state of the adjacent air. 
Tlie difficulties and nncertainties associated with the variations in the weight of a 
water skin were, it is believed, successfully met and overcome by tlie simple device 
which we now describe. 
(6) /3:j. Of the Methods used for the Removal of Water Skins. —Some observers 
may Ije considerably surprised to learn that anyone should attempt to weigh accurately 
glass vessels without previously drying the air within the balance case in the usual 
manner. This departure from a time-honoured, orthodox, and, as we shall try to show, 
sometimes worse than useless custom, was, on our part, of deliberate design. 
Let us consider the actual facts. As ordinarily practised, the air within a balance 
case is dried by means of such substances as solid calcium chloride, sticks of caustic 
potash or soda, phosphorus pentoxide, or by concentrated sulphuric acid contained in 
shallow vessels placed upon the bottom of tlie case. Any of the substances named 
will lap up the moisture in tlie lowest layers of air with great avidity ; but as the 
density of air increases with the removal of moisture, and as an attempt is usually 
made to maintain the temperature of the balance both uniform and steady, it will be 
necessary to depend almost exclusively upon the inter-diffusion of the dry and damp 
air for completing the removal of the moisture. E,emembering the law governing 
the rate of the diffusion of one gas into another, and that usually for our present 
purpose we may take the ratio of the densities of ordinary and of dried air as l/r005, 
it will be at once apparent that such a process is both lengthy and inconvenient. 
Again, whenever the shutter of the balance case is lifted the greater portion of 
the dried air will be lost; this loss can only be replenished by the further expenditure 
of valuable time. The introduction of some form of stirrer for mixing the air durinir 
the Initial process of drying would frequently be oljjectionable ; for a stirrer in motion 
would create a suction in places and so dust, otherwise harmlessly reposing in out of the 
way crevices, would be drawn into the air and in part be subsequently deposited upon 
the balance pans and upon the objects placed upon them. But a still further objection 
to the use of the ordinary method of drying the air may be given. 
In the course of some preliminary experiments with desiccating reagents, a distinctive 
