rawn on the general method of
procedure which constitutes his great discovery.

Einstein showed how to express the characters of the assemblage of
elements of impetus of the field surrounding an event-particle E in
terms of ten quantities which I will call J_{11}, J_{12}
(=J_{21}), J_{22}, J_{23}(=J_{32}), etc. It will be noted that
there are four spatio-temporal measurements relating E to its
neighbour P, and that there are ten pairs of such measurements if we
are allowed to take any one measurement twice over to make one such
pair. The ten J's depend merely on the position of E in the
four-dimensional manifold, and the element of impetus between E and
P can be expressed in terms of the ten J's and the ten pairs of the
four spatio-temporal measurements relating E and P. The numerical
values of the J's will depend on the system of measurement adopted,
but are so adjusted to each particular system that the same value is
obtained for the element of impetus between E and P, whatever be the
system of measurement adopted. This fact is expressed by saying that the
ten J's form a 'tensor.' It is not going too far to say that the
announcement that physicists would have in future to study the theory of
tensors created a veritable panic among them when the verification of
Einstein's predictions was first announced.

The ten J's at any event-particle E can be expressed in terms of two
functions which I call the potential and the 'associate-potential' at
E. The potential is practically what is meant by the ordinary
gravitation potential, when we express ourselves in terms of the
Euclidean space in reference to which the attracting mass is at rest.
The associate-potential is defined by the modification of substituting
the direct distance for the inverse distance in the definition of the
potential, and its calculation can easily be made to depend on that of
the old-fashioned potential. Thus the calculation of the J's--the
coefficients of impetus, as I will call them--does not involve anything
very revolutionary in the mathematical knowledge of physicists. We now
return to the path of the attracted particle. We add up all the elements
of impetus in the whole path, and obtain thereby what I call the
'integral impetus.' The characteristic of the actual path as compared
with neighbouring alternative paths is that in the actual paths the
integral impetus would neither gain nor lose, if the particle wobbled
out of it into a small extreme