348 MR. T. CARNELLEY ON A COLORIMETRIC 



permanent faint pink colour is obtained; the whole is 

 made up to i litre, when it forms what may be called the 

 "water-test solution." In each of the cylinders i cub. 

 cent, of the potassium ferrocyanide is added, and then a 

 measured quantity of the water-test solution put into one 

 of them (x) ; both are next filled with water up to the 

 mark, and I cub. cent, of the standard nitric acid added 

 to each. After x has been well stirred, the standard 

 iron solution is gradually run into y, the liquid being 

 stirred after each addition, and the colours in the two 

 cylinders compared by placing them side by side over a 

 sheet of white paper in front of a window ; this is repeated 

 till the colours in each of the cylinders appear to be equal, 

 which point completes the operation. 



Every cubic centimetre of iron solution used corresponds 

 to o*i milligrm. of iron, from which the amount of iron 

 added to cylinder y can be calculated. Then, assuming 

 that equal shades of colour are, ceteris paribus, produced 

 by equal weights of iron, the amount of the latter in 

 cylinder x is equal to that added to y ; and since the 

 volume of the original sample of water in the test solu- 

 tion is known, and also the volume of the latter put into 

 x, the amount of iron in a measured quantity of water 

 can therefore be calculated. The volume of the test solu- 

 tion put into cylinder x should be such as not to require 

 more than 5 cub. cent, of the iron solution to be added to 

 y to produce an equal shade; for if more be added the 

 colour obtained will be too dark to compare with ease 

 and accuracy. 



If the sample of water contains such a small amount of 

 iron as, after oxidation, not to give a coloration directly 

 with the ferrocyanide and nitric acid, a sufficient quantity 

 of it must be evaporated with half a cubic centimetre of 

 the sulphuric acid till it occupies from 100 to 200 cub. 

 centims. The liquid is then poured into a flask, together 



