14 RESEARCHES UPON ATOMIC WEIGHTS. 



The close agreement of the results in each series leaves little doubt of the 

 identity of the different samples, although they represent material from different 

 sources as well as different fractions of the same material. The slight discrep- 

 ancy between the results by the two methods is undoubtedly due, in part, to 

 the difficulty in determining silver chloride with accuracy, owing, in the first 

 place, to loss of chlorine by the silver chloride in the processes of manipulation 

 and drying, and in the second place to the sHght solubility of silver chloride 

 even in dilute silver nitrate solutions. In the Hght of these possibihties, it is 

 probable that the results of Series I are shghtly too high. On the other hand, it 

 is probable that the average of Series II is slightly too low, for the average of 

 experiments 7, 8 and 9, in which the experience gained in the previous analyses 

 was a very considerable aid, is 112.415, 0.006 of a unit higher than the average 

 of the whole series. 



ACTION OF HYDROCHLORIC-ACID GAS UPON PHOSPHORUS PENTOXIDE. 



Some time after the completion of the foregoing series of experiments it was 

 found that in fusing manganous chloride, in a current of hydrochloric-acid gas 

 which had been dried by concentrated sulphuric acid and finally by means of 

 phosphorus pentoxide, an insoluble residue of manganous phosphate was in- 

 variably obtained when the salt was dissolved in water. The quantity of this 

 residue varied with the amount of moisture contained by the salt when brought 

 in contact with the hydrochloric-acid gas, being extremely slight if the salt was 

 very nearly dry, but amounting to several milligrams if the salt still contained 

 much of its crystal water. Although it seemed certain that the phosphorus had 

 its origin in the phosphorus pentoxide, and was volatilized in the form of either 

 phosphorus pentachloride or oxychloride through the action of the hydro- 

 chloric acid upon the pentoxide, in order to obtain still more positive evidence 

 that this was really the case, the experiment was tried of passing hydrochloric- 

 acid gas which had been dried thoroughly by means of sulphuric acid, first 

 over phosphorus pentoxide which had been freshly sublimed in a current of 

 dry air, and then into water. The aqueous solution, upon evaporation and 

 testing with ammonium molybdate, gave a considerable amount of the charac- 

 teristic ammonium phosphomolybdate. This result confirms that of Bailey 

 and Fowler,^ who have found that both hydrochloric and hydrobromic acids 

 react with phosphorus pentoxide at ordinary temperatures to form the oxy- 

 chloride and bromide of phosphorus respectively. The manganous phosphate, 

 then, must have been produced by the action of the volatilized chloride of 

 phosphorus upon the moisture contained by the manganous chloride to form 

 phosphoric acid, with subsequent displacement of hydrochloric acid from the 

 salt by the phosphoric acid, 



' Chem. News, 58, 22 (1888). 



