Electric Origin of Rigidity and Consequences. 439 
is 23°7-7°4, and so on. There is evidence (see “ Ionization 
&e.”) that a gramme atom of Li ions retain the volume 2 
characteristic of the combined state, of Na ions the volume 
74 and soon. Hence it appears that the contiguity of $ to 
Na causes a large reduction in its volume. On the other 
hand, any change of volume in the Cl and Br atoms on 
passing from Cl, and Br, into compounds is relatively small, 
and in the case of I the change seems to be an expansion 
instead of a contraction. It would appear that when $ and 
5 attract one another in a metallic atom, the atom acquires 
such a volume as makes centrifugal force equal to the 
attraction. When the attraction is neutralized the atom 
collapses to a much smaller volume. The simplest assumption 
we can make as to the mutual potential energy of a metal 
atom and $ and 6 is that Na and Z have minimum potential 
energy in contact, Na and ) when apart, that is Na and J 
attract, Na and Si repel. In metallic Na the 6 would be 
expelled but for the direct attraction of the J. The simplest 
assumption we can add as to the relation between the 
attraction and repulsion is that they are equal, or that the 
mutual potential energies of Na and $ and of Na and are 
equal in magnitude but of opposite sig 2 If there is a strong 
enough external electric field to counteract sufficiently the 
attraction between = and b, the repulsion of the metal atom 
for b will drive it out, giving the phenomena of the cathode 
rays. Ina similar way the Becquerel rays may originate. 
‘Returning to the energy changes to go with (24) we get 
for the heat of reaction 
Zee ane) =2h. . . «nk ee) 
For two metals we have 
(R,) —(R,) =: (R,S) — (RLS) =hy—hy ° . (26) 
Now, at a junction of these same two metals we may 
consider the passage of a current to be the exchange across 
the junction of % “for 5 bp. By the loss of % the first. metal 
gains potential energy (R,) and by the gain of it gains 
energy (R,}, and so for the sey ‘of 2e across the junction 
the gain of energy is 2(R,)—2(R,). Consequently there 
must be an E.M.F. V,—V, across the junction such that 
the total energy change of dike 2e crossing the junction is nil ; 
grate rel Vag Wea) iy hg ue eta a LT) 
Comparing this with (26) we see that if (R,S) —(R,S) is 
ot subordinate importance to (R,)—(R,), Lodge’s principle 
will be nearly true, namely, e(V,—V Ricerca e Now, in 
