154 Scientific Proceedings, Royal Dublin Society. 
Itis thus proved that, in titrating an alkali or an alkaline solu- 
tion with phenol-phthalein as indicator, the indications of the- 
latter are altogether interfered with, and the sharpness of the- 
reaction destroyed, if soluble silicates are present. As baryta. 
water acts on glass with the formation of such compounds, our 
employment of vessels coated with paraffin wax in Pettenkofer’s. 
process is amply justified, for that reason alone, if not for others ; 
for it is almost unnecessary to say that when such vessels are used 
no disturbance of the colour reaction occurs. 
Further we have, in the preceding experiments a clue to the 
explanation of the increased alkalinity of the baryta solution after 
it has been in contact with glass. Whatever the exact mechanism 
of the reaction may be, it seems probable that the dissolved con- 
stituents of the glass function on the whole as alkalies, and after 
our experiment with metasilicate of soda, such an egpamwbon 
does not appear improbable. 
What then, it may be asked, was the reason for our finding 
diminished alkalinity in our first experiments? We believe that. 
the explanation is a simple one, viz. that we used ordinary dis- 
tilled water for washing out the contents of the absorbing vessel, 
and that the carbonic anhydride of the former accounts for the 
weakening of the baryta solution. That this is probable the 
following consideration will show. 
The decrease in alkalinity observed correspond with about 1 c.c. 
of the solution, which was equivalent to 0:1 c.c. of carbonic anhy- 
dride. According to our determinations in ordinary distilled 
water, 0'l cc. of that gas would be contained in from 22-40 e.c. 
of the water, and in our later determinations made with air, 50 c.c.. 
were employed for washing out the receivers. 
Action of weak baryta solution on silicaa—We thought it of 
interest to pursue this matter, with the object, if possible, of ascer- 
taining what substance or substances are formed when silica and 
baryta water remain in contact. Therefore, at the close of the 
titration experiment described on p. 153, we filtered a quantity 
of the silicated solution, and determined in the filtrate Si0O.,, 
1 For at this time, the fact that distilled water contains carbonic anhydride had 
escaped our attention. 
