CHANGE IN WEIGHT DURING CHEMICAL REACTION. 
231 
prepared and used together in every experiment. Each limb of a reaction vessel was 
charged, the one with a convenient quantity of concentrated silver nitrate solution, 
and the other with an equivalent amount of solution of pure re-crystallized ferrous 
sulphate. The charges were introduced, as in Landolt’s own experiments, through 
side tubes, ^2 (fig- I)- After the vessels had been charged they were weighed, and 
the weights of the two vessels and their contents were almost equalised by adding 
distilled water to the lighter one. Then the vessels were allowed 
to stand side by side for some hours, after which the side tubes 
were hermetically sealed by the blowpipe. The external volumes 
of the charged and sealed vessels were now separately measured 
by the hydrostatic method, and to the one having the smaller 
volume was added a sealed auxiliary bulb, the volume of which 
had been adjusted by trial, until it was as nearly as possible 
equal to the difference in the volumes of the two vessels. One 
of the side tubes of the vessel having the larger volume was 
then opened, and water introduced to compensate the weight of 
the auxiliary bulb of the other vessel; the tube was then re¬ 
sealed. The vessels were next immersed in faudy concentrated 
nitric acid for ten days ; after being removed from the acid, they were continuously 
washed with water for some hours ; and then finally wiped with a fine linen cloth and 
placed under a glass cover until required for use. Proceeding in the manner described, 
we obtained two charged reaction vessels having as nearly as possible a common 
volume and tlierefore an almost equal air displacement; and as their weights differed 
but slightly, the vessels when suspended from the arms of the balance almost exactly 
counterpoised each other. 
( 5 ) Of the Manner in which Weighings were Conducted. Series /.—-At the outset 
it was desired to discover the degree of approximation that might l)e attained when 
Landolt’s determinations were compared with my own ; the first series of experiments 
were therefore planned and conducted on the same general lines as those laid down 
by Landolt, but wibh the addition of the distinctly more favourable conditions which 
accrue from the use of a protected balance beam. Here it may also be stated that 
throughout this research all weighings were carried out in a darkened room, the one 
window of which faces N.E. The ivory scale of the balance was, at the time of weighing, 
brightly illuminated by a convergent beam of yellow light; for this purpose a large 
lens was placed in front of a 6-litre flask filled with a dilute solution of potassium 
chromate solution in which a carbon filament lamp was placed centrally. 
On July 11, 1910, the prepared and charged vessels which have been described in 
the preceding section were momentarily immersed in distilled water and then wiped 
with old fine linen, care being taken to avoid contact between the vessels and the 
hand ; they were then suspended from the Imlance by means of attached platinum wire 
loops. As vessel A was a little heavier than vessel B, the necessary weights were 
