246 
MR. J. J. MANLEY ON THE APPARENT 
recommend a preliminary immersion of the vessel in distilled water. Various observers 
have found that the apparent weight of a vessel which has been wiped with silk is 
somewhat changed if re-wiped with linen, and vice versa. Certain experiments 
which we have made with tlie view of elucidating the cause of the admitted and 
observed difierences, suggest that the apparent change in weight is intimately bound 
up both with the manner in which the vessel is finally prepared for weighing and 
with the nature of the substance used for wiping. A typical case which is simpler 
than but equivalent to one of my own experiments will make this clear. The 
selected case is that of a glass sphere having a capacity of 150 c.c. and filled with 
water. The operation of wiping and the attendant rise in the temperature of the 
vessel and its contents will, in general, effect— 
(1) Slight variations in the weight of the water skin ; 
( 2 ) A very small alteration in the air-displacing power of the vessel; and 
( 3 ) A decrease in the buoyant properties of the surrounding air. 
Passing over the first-named variable which we have already fully discussed, and 
Ignoring yet another associated with possible minute changes in the weight of an air 
skin as distinguished from one of water, we observe that a glass sphere having a 
volume V = 150 c.c. at some temperature t° C., and a cubic coefficient of expansion 
= '00003, will at i + l° C. assume the approximate volume of 150'005 c.c. ; ’005 c.c. of 
air weighs something like l/l70 mgr. The great majority of workers would, by 
following the methods usually adopted for accurate weighing, quite fail to detect so 
small a change with any degree of certainty; and therefore, for our present purpose, 
the variation may be treated as an insignificant factor which calls for no further 
comment. 
In considering the effects which may be produced by the last of three variables 
named above, we would first note that when the buoyancy of the air is to be allowed 
for, a thermometer is placed in some convenient position witliin the balance case and 
read at the time the object is weighed ; the density of the air at the observed 
temperature is then taken into account and the proper correction applied. For very 
many cases this way of treating the problem in hand is doubtless sufficiently accurate 
and calls for little or no criticism. Yet it must be admitted that a close examination 
of the actual facts will show that such a plan must not infrequently lead to the 
adoption of an inaccurate correcting value for the buoyancy. To obtain a rigorously 
true correcting value, the temperature of the air should be taken, not as is often done 
anywhere within the balance case, but at a point infinitely near the object weighed ; 
for, unless the temperature of the air which is in close contact with the object upon 
the balance pan be accurately known, it will be impossible to assign to the correcting 
factor a precise value. The following remarks will, it is hoped, render evident the 
nature of the error that may be introduced by pursuing the usual instead of the ideal 
method. 
