EXPERIMENT STATION BULLETINS. C21 



however, tlie dicalciiim phosphate was not completely soluble. With 

 natural phosphates they found that mineral phosphates dissolved in 

 ammoniacal ammonium citrate to an inappreciable extent and they 

 go so far as to say that with superphosphate from such a source the 

 reverted phosphoric acid may be practically completely separated from 

 the original material by such a solution. As a result of their work they 

 recommend the use of Peterman's solution which is an alkaline one, 

 viz., a neutral solution having a density of 1.09 to which 50 c.c. of 

 NH4OH (S. G. 0.91) per liter has been added. 



Gibson (17) compared the action of three solutions on calcium phos- 

 phate. One of these was prepared by the A. O. A. C. ofificial method, 

 i. e., was neutral, while the other two were alkaline. The last two gave 

 concordant and constant results while Avith the first the figures were 

 variable. He ascribed this to a tendency on the part of the neutral solu- 

 tion to become acid. 



Gladding (18) found both di- and tricalcium salts completely soluble 

 in neutral ammonium citrate solution and says, "A slightly ammoniacal 

 citrate of ammonia solution alone of all the solutions which have been 

 proposed, is a perfect solvent for all the forms of reverted phosphate 

 while at the same time not unduly dissolving the raw or insoluble phos- 

 phate present." 



In a later article (19), however,- from the results of the analysis of 25 

 samples of natural phosphates using four citrate solutions of different 

 reactions, corresponding in a qualitative way to those described on page 

 IG of this paper, he shows that an acid solution dissolves a greatly in- 

 creased quantity of tricalcium phosphate and that an increasing alkalin- 

 ity causes a decrease in the solubility of all forms of phosphate of lime. 

 He therefore reverts to the neutral solution as the solvent of choice. 



In general, his acid solution gave the lowest results for citrate-insol- 

 uble P0O5 (the highest for citrate-soluble which was the way they were 

 reported) while an increase in alkalinity gave a corresponding increase 

 in the citrate-insoluble fraction. 



Bosworth (7) found that tricalcium phosphate was soluble in ammon- 

 ium citrate and also emphasizes the change in reaction of the solution 

 resulting from digestion at G5°. His results are, however, open to critic- 

 ism as he removed the ammonia with a current of air. 



Grupe and Tollens (20) found that Avhile dieak-ium phosphate was 

 more easily soluble than the tricalcium salt, the latter was, nevertheless 

 appreciably soluble in ammonium citrate. 



Huston (28) claimed that generally there was but little difference 

 in the solvent action of citrate solutions of varying reactions but that 

 in some cases an alkaline solution dissolved much more. He found that 

 the tricalcium phosphate of bones was soluble in ammonium citrate 

 solution. 



Joulie (31), shortly after the publication of the original description 

 of the ammonium citrate method, proposed a strongly alkaline reagent 

 instead of the neutral one recommended by Fresenius. Luck (36) in 

 replying to him voiced the objection that sucli a solution would cause 

 the conversion of some of the dicalcium salt into the insoluble tricalcium 

 compound which would then not be removed and would thus cause 

 erroneous results. This contention was subsequently supported by 

 Olsen (40). 



