R. N, Maxson — lodometric Determination of Gold. 155 



Art. XYI. — The lodometric Determination of Gold in 

 Dilute Solution ; by Kalph 1^. Maxson. 



[Contributions from the Kent Chemical Laboratory of Yale University — CXYII.] 



In a former paper from this laboratory"^ it was shown that 

 under suitable conditions potassium iodide acts normally upon 

 auric chloride according to the equation 



3KI + AuCl3=: 3KC1 + Aul + I, 

 and it appeared that this reaction may serve as the basis of a 

 trustworthy method for the determination of small quantities 

 of gold, — the freed iodine being estimated by careful bleach- 

 ing with sodium thiosnlphate and the final addition of standard 

 iodine to the incipient coloration of the starch indicator, rose 

 in most cases, or only blue at the outset when the starch has 

 not undergone hydrolytic change. f Three series of experi- 

 ments were made upon solutions of auric chloride standardized 

 by the ferrous sulphate method and by the method of Yanino,:^ 

 and one series upon auric chloride made from pure gold foil. 



In the first series the solutions of auric chloride (0*8710 grm. 



to the liter) sodium thiosulphate nearly ( r— 77 ), and iodine 



near 



liter) sodium thiosulphate nearly (r-—j, 



ly(r— -) were of such strength that an inaccuracy of 



Q.Qj^cms -j^ measurement would correspond to about 0*000009 

 grm. for the gold solution and to a little more than 0*00001 

 grm. in terms of gold for each of the other solutions. It is 

 hardly to be supposed that individual readings can uniformly 

 approximate the truth within 0*01^=°^^, and as there are six 

 readings to be made in every complete determination, the 

 chances for inaccuracy amounting to several hundredths of a 

 milligram of gold, dependent to a great extent as to direction 

 and amount upon the personal equation of the operator, are 

 considerable. Should all these reasonable errors chance to lie 

 in the same direction, the cumulative errors are likely to reach 

 at least rb 0*00005 grm., and may be even more. 



The actual error, due to all causes, found in the first series of 

 twelve determinations, excepting a single determination mani- 

 festly not in line with the others, and tending to make the 

 average error smaller, amounted in the mean to - 0*00005 grm. 

 of gold between extremes of -1-0*00003 and —0*00010 grm. 



In the second series of twenty experiments, in which the 

 solution of gold was ten times as dilute as in the experiments 



*Gooch and Morley, this Journal, viii, 261 (1899). 

 f Hale, thiis Journal, xiii, 379 (1902). 

 jBer. Dtsch. Chem. Ges., xxxi, 1763. 



Am. Jour. Sci.— Fourth Series, Vol. XVI, No. 92.— August, 1903. 

 11 



