[PASEA] RELATIVE BULK OF WEAK SOLUTIONS 33 
The observations are not sufficiently accurate to say with definite- 
ness that there is a contraction, but indicate a ‘28 per cent solution as 
the most concentrated solution with a volume less than its constituent 
water would have in the free state. 
In the case of ferrous sulphate all the solutions examined were 
made by mixing the salt directly with water, the water being freshly 
boiled so as to free it as far as possible from air that it might contain 
dissolved. As this salt is found to undergo progressive decomposition 
after being dissolved in water, each solution was analysed as soon as 
possible after its preparation. The specific gravity of the solution was 
then determined, and another analysis made. In all cases the diver- 
gence between the results of the analyses was found to lie within the 
limits of the experimental error. 
The burette used for analysis was of the Mohr type, with glass 
stop-cock. It had a capacity of 50 c.c. and was graduated to lee. By 
means of an Erdmann fioat it could be read to 025 c.c. Corrections 
were applied when found necessary to the volumes between every two 
consecutive cubic centimetre marks. ‘They had been found by weighing 
the water delivered at a known temperature. 
For analysis a solution of potassium permanganate was made up 
and its strength carefully found a number of tinies by titrating with a 
weighed amount of oxalic acid at about 60°. Reducing the oxalic acid 
to its equivalent amount of ferrous sulphate, the results of the titration 
showed that 1 c.c. of the permanganate solution was equivalent to a 
solution of ferrous sulphate containing the following numbers of 
grams per cubic centimetre: ‘01553, ‘01550, ‘01548, ‘01554; giving 
a mean value of ‘01551, and indicating a possible error of about :2 
per cent. As a solution of potassium permanganate is liable to undergo 
progressive decomposition, the strength of the solution used for analysis 
was ascertained from time to time. 
To analyse a solution a known amount was taken out by means of 
a pipette, and placed in a beaker, to which a crystal of sodium carbonate 
and some sulphuric acid were added, so that the carbonic acid gas gen- 
erated might preserve the solution as far as possible from contact with 
the air. This was then titrated with the permanganate solution. In 
the titration of a solution of concentration ‘5, the amount of perman- 
ganate required did not differ from the mean value by more than ‘025 c.c., 
which would indicate a possible error of ‘1 per cent. Hence the total 
possible error in the concentration is "3 per cent. 
For solutions with a concentration less than ‘3, an error of ‘3 per 
cent in the concentration would cause an error in the sixth place of 
decimals in the volume of the water in one gram of the solution. For 
concentrations between ‘3 and ‘6, the error would be 1 in the fifth place. 
