The evolute of an ellipse may be defined in terms of the curvature at a point on the ellipse. Suppose that the ellipse is parameterized by [(r)\vec](t)=áacos(t),bsin(t)ñ, 0 £ t < 2p. The curvature k(t) at [(r)\vec](t) can be evaluated using the standard calculus formula in parametric form
k(t) = |
|(-acos(t))(bcos(t))-(-bsin(t))(-asin(t))|
[a2sin2(t)+b2cos2(t)]3/2
|
|
|
and from it one obtains the radius of curvature,
R(
t)=1/
k(
t). By definition the center of the osculating circle is located a distance
R(
t) along the inward pointing unit normal
N(
t) from its point of origin on the ellipse, so the position vector of this center is simply
C(
t)=
r(
t)+
R(
t)
N(
t). By direct calculation and simplification one finds
|
®
C
|
(t) = |
|
a2-b2
a
|
cos3(t), |
b2-a2
b
|
sin3(t) |
|
. |
|
(4) |
If one adopts the definition that the evolute is the locus of the centers of the osculating circles, then for the ellipse it is this parametrized curve C(t) as t varies from 0 to 2p. Since the focal distance from the center along the major axis is c = Ö{a2-b2}, the two cusps of the evolute at a distance c2/a=c(c/a) < c < a along that axis must fall short of the foci inside the ellipse, while the other two cusps exit the ellipse along the minor axis when c2/b > b or a > Ö2 b.
FIGURE 3: The center of the osculating circle traces out the evolute of the ellipse.