10.2 Polar curves, tangents, and lengths 543
enable us to convert (7.26) into (10.281). For the special case in which
0(t) = t and ¢'(t) = 1, it is standard practice to put (10.281) in the form
(10.285) L = fa p jp2+
(Tdp^2
do
in which the variable of integration is (k
and//the
limits of integration are
called a and 0 (instead of a and b) because a and 0 "look more like angles."
It is worthwhile to know a little trick by which the above formulas
involving polar coordinates can be remembered. We can look at Figure
10.286 which shows, among other things, an arc of length As joining two
Figure 10.286
points P and Q which have polar coordinates p, 0 and p + Op, 95 + 04.
We can feel that the outer part of the figure resembles a rectangle enough
to enable us to write the approximate formulas
(10.287) As = p2 0¢2 + Opt, cos ¢ =
OP
P2 0O2 + Qp2
tan 4' =
and expect that correct results should be obtained by dividing by At
or by A¢ and taking limits. We can know that this optimism does not
prove formulas, but it can help us to recall theformulas when we have
forgotten them. When we wish to calculate the length L of the curve
having the polar equation p = f(4') with a < 0 < 0, we can, whenf is
continuous, sketch a figure more or less like Figure 10.286and use the
optimistic calculation
(10.288) L = lim As = lim J p2 XO2 -+ Apt
= lim p2 + fi() d¢
to lead us to the correct formula (10.285).