As topsquark says, basically I want to use small angle approximations for the
trig function, and this substitution makes it easier to see what I am doing.
Also If I wanted to do a power series expansion, then if the limiting variable
is going to zero and I have expanded about 0, then I can discard the higher
order terms (but this is essentially the same reason as above).
Unless the limit is obvious it is the first move I normally make.
RonL