198 Sanitary Studie s of Baking Powders 



tion value (92.9 mg.). There was apparently a re-conversion of 



AI2O3 to AIPO4. 



From these findings it might be supposed that a correct result 

 could be obtained by adding ammonium phosphate to a precipitate 

 of AIPO4, before ignition and weighing, but experience has shown 

 that, unless repeated additions of very small amounts are made, this 

 cannot be accomplished ; f or, if a quantity of ammonium phosphate 

 sufficient to effect the change is added all at once, it is extremely 

 difficult to volatilize the excess of P2O5. In the cases cited above, 

 it required 12 and 15 hrs. of ignition, respectively, at nearly white 

 heat in the muffle, to remove the excess of P2O5, yet in neither case 

 was sufficient phosphate added the first time to convert all the oxid 

 to phosphate. In an attempt to make a single addition suffice, it 

 was found, after 17 hrs. of ignition, that at least 70 hrs. would be 

 required for complete removal of the excess of P2O5, obviously an 

 impossible condition for ordinary work. 



In two repetitions of the foregoing tests — 6 and 7 in the accom- 

 panying Table (i) — conducted some time after the earlier ones, it 

 was found that ignition with sulfur did not produce a significant 

 loss in the weights of the ignited precipitates. The addition, how- 

 ever, of ammonium phosphate always caused perceptible increases 

 in the weights of such precipitates. There was evidently production 

 of oxid, or a modified phosphate, as a result of the first ignition, 

 independently of any influence of the sulfur. Ignition with ammo- 

 nium phosphate re-converted the resultant product, wholly or in 

 part, into normal phosphate. 



It is impossible to say, under the circumstances, that the con- 

 version of phosphate into oxid was dependent, in any of the fore- 

 going tests, on the intermediate action of the associated sulfur. It 

 is possible that temperature differences, and other unknown factors, 

 were influential in effecting the observed variations in the results 

 with sulfur. The influence of impurities seems to have been 

 excluded.^ 



* AIPO4, made by the Schmidt-Hoagland method, and precipitated only once 

 (not re-precipitated as they direct), contained traces of potassium, chlorin and 

 Sulfate. When precipitated twice, it was free from these materials. 



