It will be observed that these expressions contain non-axisymmetric
terms in sin 26 and cos 28 and therefore it is not generally possible to satisfy
the prescribed tangential displacement conditions using a series of the form
of eqn. (1). However, it would be possible to do so if these terms could be
neglected. This procedure can be justified on the following grounds.
(i) The terms in question are small. For example, in the first of eqns.
(11) the term in cos 28 adds at the worst points only + 9% to the correspond-
ing terms in r2, for v = 0.3.
(ii) In certain circumstances the offending terms are self-cancelling. For
example, if the incremental problem involves a central stuck region sur-
rounded by an annulus in microslip (as in all the cases treated by Mindlin
and Deresiewicz [2] and most of those treated here), the process of neutral-
ising the radius-dependent axisymmetric terms in the stick region conve-
niently neutralises the non-axisymmetric terms as well.
(iii) The additional stress distribution required for an exact solution
would be self-equilibrating and would therefore not affect the trajectory of
the sphere except in so far as the change in tangential stress caused a change
in the division of stick and slip regions.
(iv) The changes envisaged in (iii) would make the stick-slip boundary
non-circular and the problem would become intractable.
We therefore propose to neglect the terms in cos 20 and sin 28 in eqns.
(11) and (12).