176 METHODS OF ANALYSIS. 



The accuracy of this method is indicated by the agreement of the percentages 

 of potassium bitartrate in cream of tartar powders containing no free tartaric 

 acid, obtained by calculation from the tartaric acid, with those obtained by 

 calculation from the potassium oxid.o 



8. Starch, 

 (a) DIBECT INVERSION METHOD. 



(For all baking chemicals free from lime.) 



Weigh 5 grams of the powder into a graduated 500-cc flask. Convert into 

 dextrose by the modified Sachsse method ("VI. General Methods," page 53), 

 and determine the reducing power by the Allihn method (page 49). 



(b) INDIRECT METHOD. & 

 (For phosphate, alum phosphate, and all other baking powders containing lime.) 



Mix 5 grams of the powder in a graduated 500-cc flask with 200 cc of 3 per 

 cent hydrochloric acid, and allow the mixture to stand for 1 hour, with fre- 

 quent shaking. Filter on an S. & S. No. 575, 11 cm, hardened filter, taking care 

 that a clear filtrate is obtained. Rinse the flask once without attempting to 

 remove all the starch, and wash the paper twice with cold water. Carefully 

 wash the starch from the paper back into the flask, with 200 cc of water, 

 using a small wash bottle. Add 20 cc of 25 per cent hydrochloric acid and pro- 

 ceed according to the Sachsse method. Determine reducing power by Allihn's 

 method. 



The treatment with 3 per cent hydrochloric acid, without dissolving the 

 starch, effectually removes the lime, which otherwise would precipitate as 

 tartrate in the alkaline copper solution. 



(c) MODIFIED McGnx METHOD. 



Digest 1 gram of the powder with 150 cc of 3 per cent hydrochloric ;i<-k 

 for 24 hours at room temperature, with occasional shaking. Filter on a good 

 crucible, wash thoroughly with cold water and, finally, once with alcohol 

 and once with ether. Dry at 110 C. (4 hours is usually sufficient), cool 

 and weigh. Burn off the starch, weigh again, and determine by difference. 



The results by this method on cream of tartar powders and tnrtaric aci( 

 powders agree closely with those obtained by copper reduction. On phosphat 

 alum, and alum-phosphate powders the results are usually satisfactory, but ii 

 some instances they may be over 2 per cent too high. 



9. Potassium Bitartrate. 



If, as is usually the case, no other potassium snlt but the bitartrate 



present, multiply the pen-entire of total pol.-ish determined MS directed nn< 

 section 12 (d) on page 17*. by : I. '.n :;;. 



Conn. Ap;r. Kxper. Stat. Rept.. 1no. p. 180. 

 MVinton, Conn. Am-. Kxpn-. Stal. U.-pt- I'-'""- '" 



