It will be noted that for the metals in which the natural con- 

 stants have actually been determined there is a very fair agree- 

 ment in the three values of (L). Marked divergencies are, however, 

 to be noted in the case of the polyvalent elements Hg, Sn and Tl, 

 and in those metals of high boihng points. In this latter case the 

 fact that thermionic emission is strongly marked is an indication 

 that the Haber relationship does not give the correct periodicity 

 of the binding electron, but one of the electrons liberated by 

 thermal agitation. Closer agreement is scarcely to be expected 

 since, in manv cases, the whole basis of calculation is the Linde- 

 mann melting point and the Haber vibrational relationship, both 

 onlv approximations. 



There is, however, a sufficiently good agreement to afford evi- 

 dence for the validity of the quantum relationship in this physical 

 molecular energy interchange. It would be of interest to deter- 

 mine whether the extension of this hypothesis, involving new 

 assumptions, were also vahd, i.e. whether the vapour pressure of a 

 metal, e.g. potassium, could be raised by illumination with either 

 monochromatic radiation of A - 1516-6 /^/z or with a dichromatic 



