OP ARTS AND SCIENCES. 



271 



These determinations exhausted the material employed. The fol- 

 lowing diagram brings together all the results. 



111.3 



11395 



B, 



Left. 



RCl a 



1U0.U5 



Right. 



104.05 



102.75 



A, 



105.1 



104.65 

 A 2 B 3 



[10.46 



B, 



The diagram in the case of the oxides from cerite and monazite 

 shows very clearly that, as with the oxides from gadolinite, the atomic 

 masses B 1? B 2 , B 3 increase, while those of A u A 2 , A 8 diminish, but at 

 a different rate, perhaps, as already remarked, because the conditions 

 were not precisely the same. It seems very desirable that similar 

 experiments should be made with the basic nitrate process, which 

 has been so much used, so as to determine which method gives 

 results that converge most rapidly toward the atomic masses of 

 pure oxides. Only in this manner can the relative values of the 

 two methods be determined. I consider it probable that further ex- 

 perience with the oxychloride process will lead to a very material 

 shortening of the process. I have employed for the mosl pari porce- 

 lain crucibles holding about 130 c.c, but with larger muffles it would 

 be easy to work up a kilogram of oxides at each operation. Also 

 much is to be expected from a judicious mixture of the d fferenl pro- 

 ducts on the right and left having nearly the same atomic ma All 

 points fairly considered, I am I believe justified in offering the oxy- 

 chloride process as worthy of further trial. 



It appeared possible that basic bromides might be more advantageous 

 than basic chlorides as means of differentiation, but the experiments 

 made were not conclusive on this point. Observing the formation of 

 beautiful well defined crvstals when the oxides from gadolinite and 

 the cerite and monazite residues were dissolved m chlorhydric or 



