Lerts & Bhake—The Carbonic Anhydride of the Atmosphere. 148 
the action which a weak baryta solution exercises on glass, and 
especially on the glass of which the greenish-white Winchester 
quart bottles used in our experiments are made.!. We may, how- 
ever, first draw attention to the fact that the actual quantities in- 
volved both in the experiments we made in the past and in those 
we are about to describe, were very small. Thus— 
(A.) Strength of baryta water used, 1 ¢.c. = 0-1 c.c. CO, at 
IN. TP: 
(B.) Actual volume of carbonic anhydride to be determined in 
a Winchester quart of air, about 1 c.c. 
(C.) Theoretical residue left on evaporating 10 c.c. of the 
baryta water with hydrochloric acid, 0:0093 grm. 
We are, therefore, dealing with quantities which so far as 
weights are concerned involve single milligrammes, while as regards 
volumes, any disturbing influence affecting hundredths of a c.c. of 
carbonic anhydride would sensibly affect the determinations. 
Repetition of Previous Experiments.—Quantities of 50 c.c. of 
the baryta solution were exposed as already described in Win- 
chester quart bottles of light-green glass, which had been carefully 
cleaned, rinsed with distilled water, and well drained. The bottles 
contained ordinary air, as for the purpose we had in view we did 
1 This action cannot, of course, occur in our modification of Pettenkofer’s process, 
and we should not have gone any further into the matter but for the fact that our state- 
ment that glass is attacked by weak baryta solution was challenged. It was contended 
that Reiset and others had already investigated the subject, but had found no such 
‘extraordinary action’’ as that indicated by our experiments; and it was further 
contended that the method adopted in the first of these was unreliable, owing to the 
fact “that it was impossible to obtain accurate weighings of the barium chloride 
owing to the rapidity with which it rehydrates itself.’’ 
If this objection had any solid foundation in fact, then it would obviously be 
impossible to determine the water of crystallization in barium chloride. This was a 
simple point to test, and was easily decided experimentally. Two determnaitions were 
made by heating quantities of the crystallized salt in platinium crucibles to 150° C., 
covering the crucibles with their lids, and allowign them to cool in a dessicator, the 
heating, &c., being repeated until their weights were constant. With large crystals 
the loss was slightly too great (15°02 and 15-32 per cent.), but with small erystals— 
obtained by shaking a hot saturated solution in a corked vessel, cooled rapidly by im- 
mersion in a current of cold water—the correct amount was obtained, viz. 14°66 and 
14-67 per cent. in duplicate experiments, the theoretical quantity being 14-76. 
