28 Mr. H. E. Armstrong on Valency, 



sulphines. The possible occurrence of isomeric change in 

 the formation of molecular compounds, however, is a subject 

 which certainly deserves careful study at the present time. 



In the case of phosphorus, the existence of the highly 

 stable gaseous pentafluoride PF 5 , discovered by Thorpe, is 

 undoubtedly regarded by many as final proof of the pentad- 

 icity of this element ; but the existence of compounds such 

 as H 2 F 2 , HFFR, &c, which clearly belong to the class of 

 molecular compounds, is an indication of so marked a 

 tendency on the part of fluorine to combine with itself, that 

 for this reason alone (as Naumann and others have asserted) 

 the pentafluoride is by no means necessarily regarded as 

 an atomic compound. And I would here add that stability 

 affords no criterion as between atomic and molecular com- 

 pounds, every degree of stability being met with even among 

 those of the former class. 



An argument in favour of the pentadicity of phosphorus, 

 which apparently cannot be disposed of by any explanation 

 based on conventional considerations, has, however, been 

 advanced by La Coste and Michaelis (Berichte, 1885, p. 2118), 

 who have shown that the compounds obtained from diphenyl- 

 chlorophosphine, PC1(C 6 H 5 ) 3 and phenol is not identical with 

 the triphenyl-phosphine oxide, OP(C 6 H 5 ) 3 , obtained by oxi- 

 dizing triphenyl-phosphine, as it should be if the latter were 

 a compound of the formula (C 6 H 5 ) 2 P . OC 6 H 5 ; this last cor- 

 responding to the formula CLP . OC1 which has been 

 suggested as that of phosphorus oxychloride, and which 

 appears to derive considerable support from Thorpe's obser- 

 vations on the specific volume of the oxychloride (Chem. Soc. 

 Trans. 1880, p. 388). It is, however, conceivable that the 

 oxygen and phosphorus are united by residual affinities, 

 thus : — 



C 6 H 5 



V.U-, 



tcfEg o^ 



Michaelis and Polis (Berichte, 1887, p. 52) have argued in 

 the case of bismuth, which also is a member of the nitrogen 

 group, that the pentadicity of this element is proved by the 

 existence of the triphenyl dibromide (C 6 H 5 ) 3 BiBr 2 . But the 

 mere production of such a compound proves nothing so long 

 as its constitution is undetermined; it at most serves to 

 strengthen the conviction gained from the general study of 

 the element, that bismuth is a member of the nitrogen- 

 phosphorus group. 



In other cases also it is possible that undue importance 



