Origin of Gravitation and Terrestrial Magnetism. 691 
Moreover, if we revise the previous working, using the 
law of attraction (1+ )e,2./r? between $ and 4 in place of 
€1¢2/7*, the only effect would be to replace (7), which is 
equiv: alent to 28? =32 x 10-*, by the relation 46? = 32 x 10~*, 
which makes the values of 8 deduced from gravitation and 
from terrestrial magnetism agree more closely than before, 
but at present we are concerned rather with agreement in 
the order of magnitude of the two than with strict con- 
cordance in absolute values. As 8 is 2°4x10-?, the value 
ie is) (2x 10—-”, 
An important point to be considered now is the non- 
existence of a field of electric force at the surface of the 
Karth comparable-in importance with that of its magnetic 
field. Let E be the amount of positive or negative electricity 
forming the matter of the Harth, then, according to the pre- 
ceding principles, if tree electricity obeyed the laws of elec- 
tricity in doublets, the attraction on a positive charge Q at 
the surface of the Earth would be 
EQ{—(1—26 — 6°) +14 6}/r?=2(8 + B°BQ/r*.- (19) 
With the data supplied for finding p in (11) we can 
estimate EH to be of the order oui and so the above force 
becomes of the order 10°Q dynes. This is quite incompatible 
with experience. Hence we infer that when a charge of 
electricity like @ becomes separated from its complement == 
the electrons forming it must lose their power of giving out 
and taking in ether with any velocity comparable with 
72x 10— found above for v. In the case of free electricity 
then B=0. And 8 vanishes not only as regards Q, but 
also as regards E and —E in their relation to Q. In 
breaking up doublets to form the two free charges Q and —Q, 
we apparently destroy the circulation of the sether which 
these doublets maintained, and so we withdraw their electrons 
from participation in the effects of the circulation maintained 
by other doublets. For, if we assume that n 2—(1+ 3)? 
the factor 1+8 for Q reduces to 1, while for E the factor 
1+ is retained, and similarly for —-E the factor 1 —£, the 
force given by (19) becomes 
EQ{—(—8) +14 B}/r?=26EQ/’, 
which is still incompatible with experience. 
Thus to account for the absence of a powerful electric 
field accompanying the Harth’s magnetic field, we must con- 
sider that the force between a free electron and any other 
9 
electron, whether free or bound, at distance r is é/7? 
