|
Fishback, W.T. (1962), Projective and Euclidean Geometry (first edition), Wiley, pp. 164 9.5 Polarities
Definition 9.51. Two points are conjugate with respect to the conic with equation $∑ a_{i j} x_i x_j=0$ if their coordinates satisfy the condition $∑ a_{i j} x_i y_j=0$.
As an immediate consequence of this definition we have
Theorem 9.51. A point is self-conjugate with respect to a conic if and only if it lies on the conic.
Let us now fix a point $\left(x_1, x_2, x_3\right)$ and seek the locus of points conjugate to it. If the point were $(2,3,7)$ and the conic were
\[
x_1^2+x_1 x_3+x_3 x_1-2 x_2^2+4 x_3{ }^2=0,
\]
a point $\left(y_1, y_2, y_3\right)$ would have to satisfy
\[
2 y_1+2 y_3+7 y_1-6 y_2+28 y_3=0
\]
or
\[
9 y_1-6 y_2+30 y_3=0 \text {. }
\]
Thus the locus of points conjugate to $(2,3,7)$ would be a line. In general the locus of points conjugate to $\left(x_1, x_2, x_3\right)$ with respect to the conic with equation $∑ a_{i j} x_i x_j=0$ would be a line since the resulting bilinear form $∑ a_{i j} x_i y_j$ is linear in the $y$ 's. It is possible that all of the coefficients of the $y$ 's could be zero. In this case the locus of conjugate points is the entire plane. This would be the case for the point $(0,0,7)$ with respect to the conic
\[
x_1 x_2+x_2 x_1=0
\]
since the corresponding bilinear condition leads to
\[
0 y_1+0 y_2+0 y_3=0 .
\]
We summarize our results in the following theorem:
Theorem 9.52. The locus of all points conjugate to a given point with respect to a conic is a line or the entire plane.
In view of this result we make the following definition:
Definition 9.52. If the locus of all points conjugate to a fixed point with respect to a conic is a line, the line is the polar line of the fixed point with respect to the conic, and the fixed point is the pole of the line with respect to the conic. The relation between lines and points so determined is a polarity.
A polarity need not be defined everywhere since we have seen that the set of conjugate points of a given point may be the entire plane, in which case the point has no polar line. In the example above this occurred when the conic was degenerate. That this was no coincidence is shown in
Theorem 9.53. The polarity defined by a nondegenerate conic is a one-to-one correspondence between all of the points and all of the lines of the plane of the conic. [5]
Proof. Let the conic have equation $∑ a_{i j} x_i x_j=0$. Given a point $\left(x_1, x_2, x_3\right)$, the line coordinates $\left[u_1, u_2, u_3\right]$ of its polar line, if it exists, must be given by [6]
\begin{aligned}
&u_1=a_{11} x_1+a_{21} x_2+a_{31} x_3 \\
&u_2=a_{12} x_1+a_{22} x_2+a_{32} x_3 \\
&u_3=a_{13} x_1+a_{23} x_2+a_{33} x_3 .
\end{aligned}
[5] The result is valid even if the conic is imaginary.
[6] These equations establish a one-to-one correspondence between the points and the lines of a plane with coordinates linearly related. Thus a polarity is a projective transformation in the sense of Definition 8.72. |
|