120 F. H. SCOTT. 



Pollacci' 4 ' using zinc chloride as the reducing agent found colours 

 varying from dark blue to grey which lie considered to indicate the 

 presence of phosphorus. Heine (5) , however, found that phosphorus- 

 free as well as phosphorus-holding bodies would, after treatment with 

 the nitromolybdate reagent, turn blue with zinc or stannous chloride. 



Macallum (6) , believing in the principle of the Lilienfeld and 

 Monti reaction, used phenyl-hydrazine as the reducing reagent since, 

 in the presence of acids, this will reduce the phosphomolybdate but 

 leave ammonium molybdate unaffected. This is a great improvement 

 on zinc or stannous chloride, as it is impossible to wash the sections 

 free from adsorbed ammonium molybdate. The reaction with phenyl- 

 hydrazine has been used more or less extensively as a micro-chemical 

 test for phosphorus, among others by the writer' 7 '. It has, however, 

 recently been pointed out by Bensley (8) that this reaction of Macallum, 

 when applied to tissues, often gives misleading results and shows a green 

 colour in the tissue where nucleiu compounds do not exist. On 

 investigation Bensley discovered that the blue colour produced by the 

 phenylhydrazine was due to the reduction of adsorbed molybdic acid 

 and had nothing to do with phosphomolybdate, as his results show there 

 could be no phosphomolybdate present in the sections. The reaction 

 with phenylhydrazine, if Bensley be correct, is therefore no test for 

 phosphorus. 



Observations. I thought it desirable to determine in the first place 

 whether the yellowing of the tissues, which is produced by the nitro- 

 molybdate reagent, might not be taken as sufficient evidence of the 

 presence of phosphorus. 



In the nitromolybdate reagent there are two factors tending to 

 cause a yellowing of the tissue, viz. the nitric acid producing the 

 xanthoproteic reaction, and the formation of ammonium phosphomo- 

 lybdate from any inorganic orthophosphate which may be present. 

 On investigation it was found that, if great excess of acid be avoided, 

 a solution of ammonium molybdate in hydrochloric acid is as good 

 a precipitant of orthophosphate as a solution in nitric acid. The reagent 

 I used had the following proportions : 



Ammonium molybdate 10 / solution 80 c.c. 



Hydrochloric acid sp. gr. 1-16 12 c.c. 



Ammonium chloride 20 grs. 



Potassium persulphate saturated solution 10 c.c. 



The molybdate is poured slowly into the acid with continued 

 shaking, then the chloride added, and when this has dissolved the 



