30 ROYAL SOCIETY OF CANADA 
delivered by a 50 c.c. pipette had the foliowing values: 04493, 0°4490, 
04487, 0:4487, 0:4493, 04495, giving as mean value 0-4491 grams and 
indicating a possible error of about 0°1 per cent. 
The pipettes used for analysis and in dilution were of the approxi- 
mate volumes 25, 50 and 100c.c. They were treated as nearly as pos- 
sible in the same way as when calibrated, Ostwald’s suggestions’ being 
carried out, as to time of delivery and mode of handling. They were 
calibrated by weighing the amount of water they delivered at 18°, the 
mean of at least three determinations being taken. 
Other solutions were then made up from those prepared directly, by 
taking out a certain quantity in a pipette and adding water from the 
same or another pipette, the temperature of the solution and water 
being 18°. 
The percentage concentration of the solutions 7.e., the number of 
grams of salt in 100 grams of solution, was calculated as follows :—Let 
M be the mass of salt found in volume P, of pipette used for analysis, 
P, the volume of the pipette used for measuring the solution which was 
to be mixed with water, P,, the volume of the pipette used for measur- 
ing the water which was to be added to the solution, D the density of 
the solution from which the mixture was made up, D,, the density of 
the water used for dilution, and p the percentage concentration. Then 
VIDE 100 X mass of salt 

P = “mass of solution 
Je. 
100 =e M 
X P X 

WHO L,) E Ce Py 
The second term in the denominator did not need to be calculated, 
as it had been found experimentally when the pipettes were calibrated, 
this being the mass of the water delivered, since the calibrations of the 
pipettes and the making up of the solutions were performed at the same 
temperature. Also in the solutions analysed directly, the above formula 
gives their concentration by putting P, equal to P, and the last term 
of the denominator equal to zero. 
The weights of water delivered by a pipette in no case differed from 
their mean value by more than ‘0157. Hence the possible error in the 
numerator is ‘12%, and in the denominator is ‘035%, giving a possible 
error in the result of about ‘15%. 
Knowing the concentration and the density of a solution, (1) the 
volume of 1 gram of the solution at that temperature, and (2) the 


1 Physico-Chemical Measurements, p. 85. 
