310 Mr. J. F. Heyes on Valency, 



element and the reciprocally conditional " validity M of 

 negative or non-metallic atoms is well illustrated by the 

 hypothesis of Friedel or Wurtz*, where Friedel's compound is 

 given as either 



H 3<\ nIV /Cl H3C-O 



H 3 C H H 3 C— CI— H 



as contrasted with the formula now proposed : — 



)0=C1— H or >==C1— H, 



HsC 7 H.CT 



where oxygen can only be tetravalent if chlorine be trivalent, 

 and where, as it may be said, the mutual validity of the 

 oxygen and the chlorine is really one of the points at issue. 



Temperature is, then, a function of valency ; and Dr. 

 Crookes' address to the British Association served to bring 

 out very clearly that it is also probably a periodic functionf 

 of the atomic weight, or at least chemical reacting weight, of 

 an element with one alternative valency, as this paper has 

 suggested, rather than many " alternatives " such as the 

 hypotheses of valencies from 2 to 8 or more. The idea is 

 well illustrated by a pendulum arrangement of the periodic 

 series. 



Paramagnetic. Diamagnetic. 



vjqv// ivQ// mjv 

 Sc'" Ca" K' 



Na' Mg" Al"' + 

 111 CF "S" y F'" — 



The validity alternative valency is indicated on the left of the 

 symbol. It will be interesting to see how future researches 

 on the associations of vortex-rings bear upon this view. At 

 the least it may be urged that it would be very disastrous for 

 chemists to abandon these valence-marks, because the con- 

 nexions with chemical affinity are not measurable at this 

 stage of the inquiry. 



The alternatives of (C 3 H 5 )'" and (C 3 H 5 )', indicated in my 

 first paper, suggest analogies between the probable ordinary 

 monovalent and rarer trivalent behaviour of the halogens. In 

 some cases, also, an element may be trivalent but occasion- 

 ally univalid. But they suggest the further resolution of our 

 present chemical units. If, by way of illustration, (C 3 H 5 )'" be 

 represented by (E)'", we may say we have its "allotropic 



* See 'Atomic Theory,' p. 250. 



t It is curious that the oxygen-group 0, S, Se, Te have atomic weights 

 which are very close multiples of the oxygen atomic weight. 



