METHODS OF ANALYSIS. 25 



To test the limits of this method a solution of potassium iodid con- 

 taining 0.1 mg of potassium iodid per cubic centimeter and a solution 

 of potassium bromid containing 1 mg of potassium broniid per cubic 

 centimeter were prepared. Measured quantities of the potassium iodid 

 solution were first taken and the iodin set free, extracted with carbon 

 bisulphid, and made up to 12 cc. It was found that the difference 

 between the following standard tubes could be easily read: 0.01, 0.02, 

 0.04, 0.06, 0.08, 0.1, 0.2, 0.3, 0.4, 0.5, 0.6, 0.7, 0.8, 0.9, 1 mg of potassium 

 iodid. Below this limit the amount of iodin was so small as to be con- 

 sidered only a trace, and above this limit the solution became too dark 

 colored for a good reading. It will be seen from this that when 1,000- 

 gram samples of water are used waters containing between 0.01 and 1 

 part per million of potassium iodid can be easily read by this method; 

 where 500-gram samples are used, waters containing between 0.02 and 

 2 parts per million of potassium iodid can be read, and where 200-cc 

 samples are used waters containing between 0.05 and 5 parts per mil- 

 lion of potassium iodid can be read. It was thought best not to take 

 a smaller sample than 200 cc since the error in parts per million would 

 be greatly increased. It is therefore evident that the limit of this 

 method for potassium iodid is 0.01 to 5 parts per million, and for iodin 

 itself 0.008 to 3.8 parts per million, according to the amount of sample 

 used. 



Working on potassium bromid in the same way it was found that 

 the difference between the following standards could be easily read: 

 0.1, 0.2, 0.4, 0.6, 0.8, 1, 2, 3, 4, 5, 6, 7, 8 mg of potassium bromid. 

 It will be seen from this that the limits for 1,000-gram samples of 

 water were 0.1 to 8 parts of potassium bromid per million; for 500- 

 gram samples, 0.2 to 16 parts per million of potassium bromid, and for 

 200-gram samples, 0.5 to 40 parts per million. This would make the 

 total limits of the method, never using less than 200 grams or more 

 than 1,000 grams of water, 0.1 to 40 parts per million for potassium 

 bromid, or 0.07 to 26.9 parts per million for bromin itself. 



To test the accuracy of this method the following experiments were 

 carried out: 



I. Two grams of sodium chlorid, 18 mg of potassium bromid, and 

 1.6 mg of potassium iodid were dissolved in a liter of water. The 

 water was evaporated with sodium carbonate, the residue dissolved in 

 water, filtered, and made up to a known volume. Aliquot portions 

 representing 500 cc of the original sample were taken for analysis. 

 The analysis was run as above described (extracting with alcohol, 

 determining iodin and bromin) and 9 mg of potassium broniid and 

 0.8 mg of potassium iodid were recovered, or 18 mg of potassium 

 bromid and 1.6 mg of potassium iodid per liter. 



