170 



Messrs. Nicholson and Price on the 



residue with water, and after filtration to separate the silica, the 

 addition of chloride of barium produced in no instance more 

 than* the faintest indications of the presence of sulphuric acid. 



The per-centages indicated in column 2 may be regarded as 

 correct, as we have verified them by the employment of perfectly 

 pure nitric acid, and by adopting the precautions to which we 

 shall immediately allude. 



The cause of the error by the former process we found to be 

 owing to the presence of sulphuric acid in the acids employed, 

 which we failed to detect by the ordinary method of testing for 

 this impurity, namely by dilution of the acids with a large volume 

 of water before adding a solution of a baryta salt. 



By evaporating the acids nearly to dryness in a platinum dish 

 and then adding water, the sulphuric acid or sulphate with which 

 they were contaminated was recognized. The fact of the solu- 

 bility of sulphate of baryta in mineral acids was not known at 

 the time when we made the above examinations, which was in the 

 autumn of 1853*. It is not our intention now to dwell upon the 

 subject of the amount of sulphur in difi"erent varieties of pig-irou, 

 and the manner in which the property of the iron is afi"ected by 

 it. This we reserve for a future communication. 



From what has already been stated, it will be seen that errors 

 of two kinds may arise when nitric acid is employed : one where 

 evaporation to dryness is necessary, when, if the acid contain 

 sulphuric acid, an excess of sulphur will be obtained ; the other 

 where precipitation is proceeded with in very acid solutions, in 

 which case a loss will result from the solubility of sulphate of 

 baiyta in the acid. To prevent this latter som-ce of error, it is 

 necessary either to expel the excess of acid by evaporation, or to 

 neutralize it with an alkali previously to precipitating with a 

 baryta solution. With these precautions sulphur can be cor- 

 rectly estimated by oxidation with nitric acid. 



The following Table contains the results of some experiments 



* By an abstract of a paper in No. 16 of the Proceedings of the Royal 

 Societ}', Professor Calvert has, we see, been engaged in experiments on 

 the solubility of the sulphate of baryta. 



