91 
where it must be remembered that terms are included which 
only contribute to the integral, terms dependent upon the 
boundary ; and have only been introduced partly to add to 
the symmetry and partly for future reference, and comparison 
with results previously obtained by other writers. 
Finally, in order to produce the integral effect of the wave 
^2 = 0 7}2 = hcosp{at -x) 4 = ccosp(a^ - x 
The components of the displacement in the secondary wave 
must be such that 
where the terms omitted, will not affect the integral 
displacement, except by terms which depend on the form 
and size of the boundaries. And it appears that the terms 
so entering into the integral expression would probably 
be not capable of being determined by any experimental 
investigation. Possibly some considerations as to the nature 
of the strain or stress at an edge might enable us to dismiss 
some of these terms. 
IX. 
On comparing these results with those of Professor Stokes 
(between the methods of arriving at them, there is no 
possibility of comparison) we see that the only point of 
difference lies in the terms which are not according to my 
theory essential to the solution. Of the terms in the expression 
just given, the parts essential are 
that is to say they satisfy at once the equations of vibration. 
