107 



phuric acid formed), and heat to boiling till nitric acid fumes are all expelled Cool 

 make up to mark with distilled water, mix thoroughly, filter through a dry filter and 

 take .-><> < "i tlu- iiltrate (0.5 gram) for the determination of arsenic oxid Transfer 

 this "><> portion to a 400 cc Erlenmeyer flask, dilute to about 100 cc with water add 

 grame . -i p< .tassium iodid and 5 cc of concentrated sulphuric acid, heat to boiling 

 and evaporate to about 40 cc. Cool, dilute to 150 cc to 200 cc and add approximately 

 tenth-normal sodium thioeulphate just to disappearance of color caused by the free 

 iodin. In case the solution is slightly colored from iron or incomplete oxidation of 

 the oiganic matter, add the thiosulphate until nearly colorless, then add a few drops 

 in-h paste ami continue adding the thiosulphate slowly until the blue color just 

 disappears. The exact end point can easily be obtained in this way. Neutralize 

 immediately with sodium carbonate, make slightly acid with dilute sulphuric acid, 

 and u hen all Lumps of sodium carbonate are dissolved add sodium bicarbonate in con- 

 siderable excess. Titrate with twentieth-normal iodin solution in the usual way 

 iisin- -tardi solution a- indicator. Subtracting from this the number of cubic centi- 

 meters <>f iodin solution corresponding to arsemous oxid as determined by Method I, 

 the number of cubic centimeters of iodin solution corresponding to the arsenic 

 oxid (As..,() ram of the sample. 



Totul arsenic oxid (As. 2 & ). 



COMMENTS OK AJTAIiYVH AM) KISCU88ION. 



/.'. J.Danihtnn: Then- i- n-.-^n-at dilliculty in working the London purple by Method 



I\ i rhaps a little more troublesome and you have to use 



another in th : ::iu' tin- ar.-enimis oxid. I believe Method III, provided it 



factory re-ult-. i- the simplest method, both determinations being made 



from the same weighed sample. 



/'. IT 'Hir modification (Method III) makes the end point some- 



whai ietermine. The modification suggested by the referee does not seem 



to mi- to off, T any advantage over the otlicial method, since the yellow color left after 

 oxidation with nitrate mixture ob-cun-s the end reaction as badly as does the original 

 anl ha- the di-ad vantage of requiring a separately weighed sample for the deter- 

 mination of th- ar-enic in arsenious form. 



l>. UW/.v; I would emphasize the fact and recommend that it be included in 

 din-et ion- for in-e, ticide work, that this kind of work is difficult for a beginner and 

 that -.-v.-nil preliminary determinations should be run before a man new to this work 

 attem; rt results. 



A - ha- been the case in previous years, the results on arsenic oxid are very unsatis- 

 ry, there being a difference of over 4 per cent between the highest and lowest 

 determinations by Methods I and III, and over 2 per cent difference by the same 

 method by different analysts. Methods II and III give lower results than Method I, 

 but there dot^ not appear to be any uniformity in the amount that these methods 

 fall short, the determinations made by different analysts and even those by the same 

 analy-t at different times sometimes agreeing with those made by Method I and at 

 other^ showin.i; a variation of several per cent. Why this is so has notas yet been 

 determined, but is under investigation. It is the referee's opinion that on precipi- 



