248 
ME. J. J. MANLEY ON THE APPAEENT 
some of the errors associated with the operations of refined weighing, we now proceed 
to give as brief an account as may be of the results obtained during the final 
experiments connected with this present research. 
(7) Of the Final Experiments Conducted with Silver Nitrate and Ferrous 
Sidphate Sohitions. —During the earlier part of this research use was made, as already 
stated, of inverted U-shaped vessels; but according to our experience such vessels 
are not so free from sundry inconveniences as we could wish for. It was therefore 
decided to substitute for those vessels others of a more or less globular form; and in 
order that the surfaces, .volumes, and figures of the two vessels might be as nearly as 
possible identical, recently purchased Kjeldahl flasks of Jena glass were chosen. The 
two flasks were then repeatedly softened in a blowpipe flame and re-blown until by 
trial it was found that they were in all essential respects very similar; it is believed 
that the repeated softening in the flame also ensured, as no other treatment could 
have done, a very close similarity in the nature of the surfaces of the flasks. 
With the object of separating the two reacting bodies, another flask of convenient 
volume was blown wflthin each of the two Jena glass vessels. The silver nitrate 
solutions were introduced into the small inner flasks, and the equivalent weights of 
ferrous sulphate solutions into the outer and larger vessels ; the necks of the outer 
vessels were then softened, thickened, and drawn out in the form of capillary tubes ; 
these when cold were re-heated and bent hookwise and their tips sealed in the flame. 
The volumes of the vessels were next measured hydrostatically, and to the one possessing 
the smaller volume was added the necessary spherical compensator having V = 1'59 c.c. 
The sealed tip of the lightest vessel was now removed and distilled water introduced 
until the weights of the two vessels, together with their contents, were almost exactly 
equal; the tip was then re-sealed. After subjecting the vessels to the usual 
preliminary and extended cleaning with {a) nitric acid, (5) potash, and (c) water, and 
finally wiping, the weighings were commenced and conducted under the newer con¬ 
ditions which have been outlined and discussed in the immediately preceding sections. 
In planning the final experiments it was decided that the operations of weighing 
should be simplified as far as might be consistent with the conditions demanded for 
the high degree of accuracy to which we hoped to attain ; the method of double or 
reversed weighing was therefore, in our opinion, with advantage discarded, for in this 
present enquiry a knowledge of the absolute weights of the reacting bodies is 
inconsequent. The question which we were endeavouring to answer was an 
apparently remarkably simple one, namely. Does the total mass of two or more 
interacting bodies undergo any appreciable change during chemical reaction? Now 
Landolt had already shown that if such changes take place they must be so small 
that there would result 1)ut a very slight displacement in the R.P. of a balance 
adjusted to a higli degree of sensibility ; in fact, so slight was any possible displacement 
likely to be that for the final comparison it would be quite unnecessary even to change 
the position of the rider upon the beam. 
