748 DR CHARLES HUNTER STEWART ON THE ESTIMATION OF 



ten minutes. Should the acid quickly become decolorised, then probably too little 

 permanganate has been added, and the flask must again be cooled and more added until 

 the colour, after again gently heating, remains greenish or pinkish. (The author has 

 found a second addition of permanganate necessary only in such substances as narcotine 

 or naphthaline.) The flame is now removed and the flask cooled with the same pre- 

 cautions as previously. When cool, the tube is clipped at (a) and the clip at (b) opened 

 and the entire system swept out for half-an-hour. The operation is now entirely ended. 

 The tube at (D 5 ) is clipped, the pump turned off a little though not altogether, (Ej) is 

 disconnected from (D 5 ) and the pump immediately turned off. The contents of the 

 baryta tubes are emptied into bottles with well-ground glass stoppers, which had 

 previously been rinsed out with some baryta solution and drained. The stoppers are sealed 

 with paraffin. Eight to twelve hours are necessary for the precipitate to entirely settle. 

 [In a paper, " On the Variations of the Amount of Carbonic Acid in the Ground- Air,"* the 

 author has detailed experiments which he made, proving that, during such work as the 

 above, the glass of the tubes or bottles has no effect in altering the titre of the baryta 

 solution.] The contents of the combustion flask are ready for dilution, neutralisation, 

 and distillation as in the ordinary Kjeldahl process for the estimation of nitrogen. In 

 the estimation of the carbon the acid which I have used for titration is the same as 

 recommended by Pettenkofer for the estimation of carbonic acid in air. It is made by 

 dissolving 1"4107 grammes recrystallised and air-dried oxalic acid in 1 litre distilled 

 water. Each c.c. of this is equivalent in power of combining with baryta or other base 

 to 0'25 c.c. carbonic acid at 0° C. and 760 ms- Each tube, as we have seen, contains 

 100 c.c. baryta solution. Instead of titrating the whole 100 c.c.'s with an acid each c.c. 

 of which is equivalent to 1 c.c. of carbonic acid, 25 c.c. (one-quarter of the whole) is 

 titrated with an acid each c.c. of which is equivalent to one-quarter of a c.c. of carbonic 

 acid. In both cases the number of c.c.'s of oxalic acid solution used is the number of 

 cc's of carbonic acid which 100 c.c. of the baryta solution is equivalent to. Three 

 separate titrations can be made from the contents of each tube without disturbing the 

 precipitated carbonate of barium, giving thus an opportunity of checking the work. The 

 oxalic acid solution should be prepared at least weekly. Where few determinations 

 require to be made, a solution of sulphuric acid is more convenient on account of its 

 stability. The indicator used in the experiments was phenol-phthaleine. The strength 

 of baryta solution was such that each 100 c.c. was equivalent in combining power to 

 between 45 and 50 c.c.'s carbonic acid at 0° C. and 760 ms " As, in some of the 

 experiments with organic substances, very small quantities were used, every means known 

 to the author of making the method of titration accurate and consistent with itself was 

 used. This was especially necessary in applying the method to the analysis of potable 

 waters. At the risk of being tedious these may be shortly described. 



First. — About five-sixths of the amount of oxalic acid solution which would likely be 

 necessary for the titration was run into the titrating flask and the indicator added. The 



* Travis. Roy. Soc. Edin., vol. xxxvii. 



