508 Dr. E. J. Mills on Elective Attraction. 



pared, even with hydric nitrate distilled without ebuUition, they 

 were all found to contain iron, a circumstance which points to 

 the existence of a ferric oxide volatile with the vapours of that 

 nitrate. Hence it was necessary that they should all be fused, 

 dissolved, filtered, and evaporated to dryness. When ordinary 

 " pure nitric acid '^ was enjployed in their preparation, some 

 sulphate was invariably found, doubtless owing to the known 

 fact that hydric nitrate in large excess prevents the precipitation 

 of baric sulphate, thereby rendering the impurity less easy of 

 detection. The phosphoric oxychloride was prepared, according 

 to Gerhardt^s recommendation, by the action of hydric oxalate, 

 dried at 90°, on phosphoric chloride. By carrying out the pro- 

 cess in a large flask having a long neck closed with a watch- 

 glass, and adding oxalate in slight excess, a highly satisfactory 

 yield was produced. The oxychloride had next to be purified 

 from hydric chloride (with which it was saturated) by three dis- 

 tillations ; on again distilling and collecting the last fourth of 

 the distillate apart, a pure product was generally attained. This 

 was secured in stoppered bottles holding each about 30 cubic 

 centims., which were preserved under a desiccator containing 

 lime and oil of vitriol ; when about 24 cubic centims. had been 

 removed from any particular bottle, it was judged expedient to 

 reject the remainder, on account of the accumulated error due to 

 repeated contact with common air. A chlorine determination 

 was made in each portion to be actually employed, the adoption 

 or rejection of which was decided on the basis of the evidence 

 thus procured. The product was colourless and did not fume 

 in the air. 



In the first experiments, not only the nitrate, but the oxy- 

 chloride was weighed. It was soon evident, however, that the 

 eye can easily form an adequate estimate of the sufficiency of 

 such an excess of oxychloride as was here required ; and as, 

 owing to the occasional commencement of action in the tube at 

 the ordinary temperature, the weighing could not be generally 

 performed with accuracy, it was afterwards dispensed with 

 altogether. The amount of oxychloride taken was always more 

 than enough to cover the nitrate. 



The course of the actual performance of the reaction will be 

 apparent from the details above given. The following experi- 

 ments, however, to which all that succeed bear a substantial re- 

 semblance, are inserted as a proof of the efficiency of the general 

 arrangements. 



(1) 1'9300 grm. plumbic chloride mixed with 51351 grras. 

 phosphoric oxychloride and gradually heated to 127°'5 in the 

 dry air-current for I5 hour, increased in weight by '0027 grm. 

 = 00*19 per cent. 



