Any excess of bromine liberated in the first reaction above that necessary for the second, will exist in the free state, and from the amount of bromine which remains free the quantity of carbolic acid can be calculated, always provided the strength of the bromine solution is first known. The volumetric part of the analysis, therefore, merely amounts to the determination of free bromine, which is best found by causing it to react on potassium iodide, and ascertaining the amount of free iodine by titration with a standard solution of sodium thiosulphate. In other words, titrate in this way the standard alkaline bromine solution, using as an indicator starch paste until the blue colour disappears. Another method of indicating the end of the reaction is by the use of strips of paper first soaked in starch solution, and dried, and then the same papers moistened with zinc iodide, and again dried; the least excess of bromine sets free iodine, and strikes a blue colour. =Colorimetric Method of Estimation.=--A very simple and ever-ready way of approximately estimating minute quantities of the phenols consists in shaking up 10 grms. of the sample with water, allowing any tar or insoluble impurities to subside. Ten c.c. of the clear fluid are then taken, and half a c.c. of a 5 per cent. solution of ferric chloride added. The colour produced is imitated by a standard solution of carbolic acid, and a similar amount of the reagent, on the usual principles of colorimetric analysis. § 237. =Carbolic Acid Powders.=--Siliceous carbolic acid powders are placed in a retort and distilled. Towards the end the heat may be raised to approaching redness. The distillate separates into two portions--the one aqueous, the other consisting of the acids--and the volume may be read off, if the distillate be received in a graduated receiver. Carbolic acid powders, having lime as a basis, may be distilled in the same way, after first decomposing with sulphuric acid. The estimation of the neutral tar oils in the distillate is easily performed by shaking the distillate with caustic soda solution, which dissolves completely the tar acids. The volume of the oils may be directly read off if the receiver is a graduated tube. Allen[218] has suggested the addition of a known volume of petroleum to the distillate, which dissolves the tar oils, and easily separates, and thus the volume may be more accurately determined, a correction being of course made by subtracting the volume of petroleum first added. [218] _Op. cit._, i. p. 310.