Andrews and Farr — Determination of Arsenic. 319 



the theoretical.* The precipitate, with the asbestos filter, is 

 now transferred quantitatively to the flask and shaken with the 

 iodine solution. f Enough of a five per cent solution of sodium 

 bicarbonate or of sodium phosphate should now be added to 

 maintain neutrality throughout the reaction, but not an unduly 

 large excess, and the shaking be continued till the asbestos is 

 thoroughly disseminated through the liquid and till it is cer- 

 tain that every particle of arsenic is dissolved. Fresh starch 

 paste is added and the excess of iodine is titrated back by 

 hundredth or thousandth normal arsenite solution. 



For quantities of arsenic smaller than 0*5 mg., thousandth 

 normal solutions may be employed, but it must not be forgot- 

 ten that in such high dilutions a correction must be employed 

 for the amount of iodine required to produce the end-reaction, 

 if satisfactorily exact results are required. This correction 

 usually amounts to about 0*6 cc of N/1000 for each 50 cc of solu- 

 tion, but it should be determined under the actual conditions 

 of the titration. A main factor influencing the sensitiveness 

 (aside from temperature) is the concentration of iodide in the 

 liquid. For larger quantities than 10 mg. of arsenic up to 

 100 mg. tenth normal iodine solution is preferably employed. 



The degree of precision to be expected in the results 

 obtained by this method may be seen from the following test 

 analyses. A hundredth normal arsenic solution (containing 

 0'375 mg. As per cc.) was made by dissolving 990 - mg. of 

 As 2 O s in water to make two liters, and also a fifth normal 



Table 





Arsenic 



Volume 



Vol. 













taken 



of 



of HC1 



Vol. of 



Vol. of 



Weight 







Milli- : 



solution 



used 



iodine sol. 



As sol. 



of As 



Error 





grams 



cc. 



cc. 



added 



req. 



found 



mg. 



1 



3-75 



10 



25 



50-02N/100 



25-02X100 



3*75 



o-oo 



2 



3-75 



10 



25 



41-05 



16-03 



3-753 



+ 0-003 



31 



3-75 



10 



50 



30-03 



5-23 



3-72 



— 0-03 



4- 



7-519 



20 



60 



60-06 



io-oo 



7-510 



— 001 



5 



0-375 



o 



10 



5-01 



2-46 



0-382 



+ 0007 



'3 



71-02 



10 



30 



51-OlN/lO 



35-2N/100 



71-23 



+ 0-21 



n 



75-00 



10 



50 



55-04N10 



49-4N/I00 



75-15 



+ 0-15 



^Xote. In analyses Xos. 3 and 7, abont 0*1 gram of crystallized copper 

 sulphate was added, without, as may be seen, influencing the results. In 

 another experiment, not recorded in the table, similar to No. 3 except that 

 the copper was omitted, a solution of titanous chloride was substituted for 

 the stannous chloride. In this case there was a negative error of 0'2 mg. in 

 the arsenic. Twenty hours was required for the reaction, which is slower 

 than when stannous chloride is used. 



* One cubic centimeter of centinormal iodine solution = - 15ing. As. 

 + It is ordinarily easy to see when the reaction is complete. When the 

 amount of arsenic is less than 5 mg., it is practically instantaneous. 



