Planar Reflections:

Reflecting a Point (or Curve) Across a Dynamic Tangent Line

by

Frank J. Attanucci

[email protected]
Department of Mathematics
Scottsdale Community College
9000 E. Chaparral Road
Scottsdale, AZ 85256-2626

ABSTRACT: Let C be a smooth curve in the plane and let P be any (fixed) point in the same plane as
C. Let L be a line in the plane that is tangent to C at the (dynamic) point Q. Let R be the reflection of
P across L. Of course, as Q moves along C (thus changing the orientation of the tangent line), R also
moves in the plane. In this paper, I first find general equations that describe this dynamic reflection
of P across L; then by considering some specific examples (based on choices for C and P), one
discovers new(?) ways in which familiar curves can be generated. I also discover some surprising
results when P is a well-chosen parameterized curve and not merely a point. I finish by indicating
possible directions in which the investigations contained this paper could be continued.
1

The mathematical foundation of this entire paper is contained in our first theorem (Theorem 1)for
every other result is a mere consequence.
A note on terminology: In the parametric equations that will appear, some will have only one
parameter, while others will have two. In those that have only one parameter, the parameter will be
either the object parameter (for, say, a curve) or the parameter of animation for the changing
position of a point. (From the context it should be clear what role the parameter is playing.) For those
equations having two parameters, one will be an object parameter for a curve and the other will be the
parameter of animation. (Again, from the context, it should be clear which is which.)

Theorem 1: Let C be a smooth curve in the plane having parametric equations:

x f (t ) and y g (t ) ,

(1)

where f '(t ) and g '(t ) do not simultaneously vanish for all t I , and let P x0 , y0 be any (fixed)
point in the same plane as C. Let L be a line in the plane that is tangent to C at the (dynamic) point Q.
Let R be the reflection of P across L. Of course, as Q moves along C (thus changing the orientation of
the tangent line), R also moves in the plane. In fact, parametric equations (X, Y) for the trajectory of R
are
2 f (t ) g '(t ) x0 f '(t ) 2 y0 f '(t ) g '(t ) 2 g (t ) f '(t ) g '(t ) x0 g '(t )
2

f '(t ) g '(t )
2

and

(2)

2 g (t ) f '(t ) y0 g '(t ) 2 x0 f '(t ) g '(t ) 2 f (t ) f '(t ) g '(t ) y0 f '(t )

2

f '(t ) g '(t )
2

where t I .
Proof: We proceed in steps (using column matrices). For readability, the details of the calculations
have been omitted. Figure 1 below will help reveal the proofs overall coherence.

P-Q

Q
S

x
S
Figure 1
x
f (t )
Step 1: Let Q be the dynamic point Q
on the curve C and let P be a fixed point P 0 in

g (t )
y0
x f (t )
the same plane as C. Then the vector from Q to P is P Q 0
.
y0 g (t )

Step 2: Let U be the unit tangent vector to the curve C at the point Q. Then

f '(t ) g '(t )
.
g '(t )

2
2
f '(t ) g '(t )

f '(t )
2

Step 3: Let S be the vector projection of P Q onto U. Then S is parallel to the tangent line L and is

given by
S P Q U U
3

 x0 f '(t ) f (t ) f '(t ) y0 g '(t ) g (t ) g ' t ) f '(t )

2
2
f '(t ) g '(t )

.
x
f
t
f
t
f
t
y
g
t
g
t
g
t
g
t

'(
)
(
)
'(
)
'(
)
(
)
'
)
'(
)

0
0

2
2

f '(t ) g '(t )

Step 4: If N is the vector difference: N = (P Q) S, then R, the reflection of the point P about the line

L, is given by
R QS N

Q S P Q S

2Q 2 S P
2 f (t ) g '(t ) 2 x0 f '(t ) 2 2 y0 f '(t ) g '(t ) 2 g (t ) f '(t ) g '(t ) x0 g '(t ) 2

2
2

f '(t ) g '(t )

.
2
2
2
g
t
f
t
y
g
t
x
f
t
g
t
f
t
f
t
g
t
y
f
t
2
(
)
'(
)

'(
)

2
'(
)
'(
)

2
(
)
'(
)
'(
)

'(
)

0
0
2
2

f '(t ) g '(t )

(3)

The parametric equations (X, Y) for R are given by the rows of the matrix in (3); these are as stated in
(2). This completes the proof of Theorem 1.

Before considering some illustrative examples, I note that the fixed point P x0 , y0 in Theorem 1
could have been just one point lying on another parameterized curve (also called P), having parametric
equations

x0 x0 ( s) and y0 y0 ( s ) , where s I *.

(4)

In this case, R is no longer merely the reflection of a point across the dynamic line L; it is the reflection
of the (fixed) curve P across L. Parametric equations for the dynamic reflected curve R would be
2 f (t ) g '(t ) x0 ( s) f '(t ) 2 y0 ( s) f '(t ) g '(t ) 2 g (t ) f '(t ) g '(t ) x0 ( s) g '(t )
2

X ( s, t )

f '(t ) g '(t )
2

and

(5)

2 g (t ) f '(t ) y0 ( s) g '(t ) 2 x0 ( s) f '(t ) g '(t ) 2 f (t ) f '(t ) g '(t ) y0 ( s) f '(t )

2

Y ( s, t )

f '(t ) g '(t )
2

where s I * and t I . Note: In this context, s would be the object parameter for the reflected
curve R and t would be the parameter of animation.

Some illustrative examples of Theorem 1

Theorem 2: Let C be an ellipse and let P be one of its foci. If L is a tangent line to the ellipse at the

point Q and R is the reflection of Q across L, then as Q moves once around the ellipse, R moves in a
circle whose radius has a length equal to that of the major axis of C and whose center is at the other
focus.

Proof: Let C have parametric equations

x f (t ) a cos(t ) and y g (t ) b sin(t ),

(6)

where 0 a b and t 0, 2 . Let P be the focus the point F1 0, c 0, b 2 a 2 . Let L be

the tangent line to the ellipse C at the point Q f (t ), g (t ) a cos(t ), b sin(t ) . Note: As t increases
from 0 to 2 , Q will move CCW around the ellipse. With these choices for C and P, we have
f '(t ) a sin(t ), g '(t ) b cos(t ), x0 0, and y0 b 2 a 2 .

(7)

Substituting (6) and (7) into equations (2) gives the parametric equations (X, Y) for R

2ab cos(t ) sin(t ) b 2 a 2 b

2a 2b sin(t ) b 2 a 2 (b 2 a 2 ) cos 2 (t ) a 2

and Y
X
.
(b 2 a 2 ) cos 2 (t ) a 2
(b 2 a 2 ) cos 2 (t ) a 2

(8)

The claim of Theorem 2 is that equations (8) are parametric equations for a circle of radius r = 2b that

is centered at the other focus F2 (0, c) 0, b 2 a 2 . Letting E X h Y k and

2

simplifying, one finds that

E X h Y k
2

X 2 Y b2 a 2

4b 2 .

Thus, the parametric equations in (8) define a circle of radius r = 2b (which equals the length of the

major axis, since 0 < a < b) and having center at the focus: F2 (0, c) 0, b 2 a 2 . This completes

the proof of Theorem 2. The following illustrates the result (with a = 4, b = 5).

Before continuing on to a second example, we first provide some other illustrations of Theorem 1, for,
remember, the point P x0 , y0 can be any point in the plane of the curve C. Taking C to be the
same ellipse as above, we have:

From the illustrations above (especially the one on the lower leftwhich looks an awful lot like a
cardioid), we are led to our next theorem. Although it appears here as Theorem 3, it was actually the

first in my order of discovery (and was followed by Theorem 2).

7

Theorem 3: Let C be a circle and let P any point in the plane of C. If L is a tangent line to the circle at

the point Q and R is the reflection of Q across L, then as Q moves once around the circle, R moves
around the graph of a limaon whose graph (i) will have a loop if P is outside C; (ii) will be a cardioid
if P is on C; and (iii) will not have a loop if P is inside C (if P is at the center of C, R will be a circle
concentric with C and with twice the radius). The limaon will have its axis of symmetry on the line
passing through the center of C and the point P.
Proof: Let C have parametric equations
x f (t ) a cos(t ) and y g (t ) a sin(t ),

(9)

where t 0, 2 . Let P be the point P b, 0 , where b > 0. Let L be the tangent line to the circle C
at the point Q f (t ), g (t ) a cos(t ), a sin(t ) . Note: As t increases from 0 to 2 , Q will move
CCW around the circle. With these choices for C and P, we have
f '(t ) a sin(t ), g '(t ) a cos(t ), x0 b, and y0 0.

(10)

Substituting (9) and (10) into equations (2) gives the parametric equations (X, Y) for R

X b 2a 2b cos(t ) cos(t ) and Y 2a 2b cos(t ) sin(t ), where t 0, 2 .

(11)

But equations (11) can be written in the form X b r cos(t ) and Y r sin(t ), where r is the polar
equation

r 2a 2b cos(t ), where t 0, 2 .

(12)

Equation (12), however, is the polar equation of a limaon. Since replacing t with t leads to the same
equation for r, the graph of (12) is symmetric with respect to the polar axis (and, hence, with respect to
the line through the center of C and the point P). By hypothesis, 0 < a, b. Therefore,

(i)

If b > a > 0, P lies outside C; moreover, since 2b 2a 0 , the graph of (12), and hence, (11) has
a loop.

(ii)

If b = a > 0, P lies on C; moreover, since 2b 2a 0 , the graph of (12), and hence, (11) is a
cardioid.

(iii) If a > b > 0, P lies inside C; moreover, since 2a 2b 0 , the graph of (12), and hence, (11) does

not have a loop. Note: If b = 0, then r 2a and the parametric equations in (11) reduce to

X 2a cos(t ) and Y 2a sin(t ), where t 0, 2 .

These are, or course, parametric equations for a circle that is concentric with C and a radius whose
length is twice that of C. This completes the proof of Theorem 3.

Theorem 4: Let C be a parabola and let P be its focus. If L is a tangent line to the parabola at the point

Q and R is the reflection of P across L, then as Q moves along the parabola, R moves along the graph
of its directrix.
Proof: Suppose the equation of the parabola C was x 2 4 py , where p 0 is a constant. Then C has
parametric equations
t2
x f (t ) t and y g (t )
,
4p

(13)

where t , . Since P is the focus of the parabola, P 0, p . Let L be the tangent line to the
t2
circle C at the point Q f (t ), g (t ) t ,
. Note: As t increases from to Q will move
4p

along the parabola from left to right. With these choices for C and P, we have
f '(t ) 1, g '(t )

t
, x0 0, and y0 p.
2p

(14)

Substituting (13) and (14) into equations (2) gives the parametric equations (X, Y) for R

X t and Y p, where t , .

(15)

These are, of course, parametric equations for the directrix of the parabola x 2 4 py. This completes
the proof of Theorem 4.

What if P is a (fixed) parameterized curve?

Earlier in this paper, I mentioned that the fixed point P x0 , y0 in Theorem 1 could have been just
one point lying on another parameterized curve (also called P), having parametric equations
x0 x0 ( s ) and y0 y0 ( s ) , where s I *.

(4)

And that, in this case, R is no longer merely the reflection of a point across the dynamic line L; it is the
reflection of the (fixed) curve P across L. Parametric equations for the dynamic reflected curve R
would be

2 f (t ) g '(t ) x0 ( s) f '(t ) 2 y0 ( s) f '(t ) g '(t ) 2 g (t ) f '(t ) g '(t ) x0 ( s) g '(t )

2

X ( s, t )

f '(t ) g '(t )
2

and

(5)

2 g (t ) f '(t ) y0 ( s ) g '(t ) 2 x0 ( s ) f '(t ) g '(t ) 2 f (t ) f '(t ) g '(t ) y0 ( s) f '(t )

2

Y ( s, t )

f '(t ) g '(t )
2

where s I * and t I . I further noted that, In this context, s would be the object parameter for the
reflected curve R and t would be the parameter of animation. We now want to consider some
examples of (5). The first example is my favoritein fact a theorem in its own right!
10

Theorem 5 (Rolling a curve along itself): Let C be a smooth oriented curve in the plane having

parametric equations:
x f (t ) and y g (t ) , where t I .

(1)

and let P x0 , y0 f ( s ), g ( s ) be a (fixed) copy of the curve C. Let L be a line in the plane that is
tangent to C at the (dynamic) point Q. Let R be the reflection of P across L. As Q moves along C
(thus changing the orientation of the tangent line), R will roll along C.
Proof: Its immediate; substituting x0 ( s ) f ( s ) and y0 ( s ) g ( s ) into equations (5), we see that
parametric equations (X, Y) for animating the rolling of R along C are

2 f (t ) g '(t ) f ( s) f '(t ) 2 g ( s ) f '(t ) g '(t ) 2 g (t ) f '(t ) g '(t ) f ( s) g '(t )

2

X ( s, t )

f '(t ) g '(t )
2

and

(16)

2 g (t ) f '(t ) g ( s ) g '(t ) 2 f ( s ) f '(t ) g '(t ) 2 f (t ) f '(t ) g '(t ) g ( s) f '(t )

2

Y ( s, t )

f '(t ) g '(t )
2

where s I and t I . This completes the proof of Theorem 5.

Note: For the mathematical rolling of R to have the appearance of real rolling, one would require

that the smooth oriented curve C always turns in the same direction (i.e., always to the left or always to
the right). This aesthetic requirement will be satisfied if f '(t ) g ''(t ) g '(t ) f ''(t ) 0 for all t I . We
now consider some examples of Theorem 5.

Example 1: Let C be the ellipse used in the above illustration of Theorem 2. C had parametric

equations x f (t ) 4 cos(t ) and y g (t ) 5sin(t ). Then f '(t ) g ''(t ) g '(t ) f ''(t ) 20, for all

t [0, 2 ]. Using equations (16), we obtain

11

4 16 cos( s ) 41cos( s ) cos 2 (t ) 50sin(t ) cos(t ) sin( s ) 50 cos(t )

X ( s, t )
9 cos 2 (t ) 16

and
Y ( s, t )

5 32sin(t ) 32 cos(t ) sin(t ) cos( s ) 41cos 2 (t ) sin( s ) 16sin( s )

9 cos 2 (t ) 16

where s [0, 2 ] and t [0, 2 ]. Note: In this context, s is the object parameter and t the parameter
of animation of the rolling ellipse. Below is an illustration of the resultwhich includes the circular

trajectories of its foci (cf. Theorem 2).

t2
Example 2: Let C be the parabola having parametric equations x f (t ) t and y g (t ) . Then
8

1
f '(t ) g ''(t ) g '(t ) f ''(t ) , for all t (, ). Using equations (16), we obtain
4
t 3 16s ts 2 st 2
1 32t 2 64 st t 2 s 2 16s 2
X ( s, t )
and Y ( s, t )
,
16 t 2
8
16 t 2
12

where s , and t , . Below is an illustration of the resultwhich shows that the focus
of the parabola [here, (0, 2)] travels along the directrix y 2 as t increases (cf. Theorem 4).

With parabolas, we saw (cf. Theorem 4) that the reflection of its focus across a dynamic tangent line
moves along the directrix of the parabola. Our next theorem shows that, in a sense, the directrix
returns the favor.

Theorem 6: Let C be a parabola and let P be its directrix. If L is a tangent line to the parabola at the

point Q and R is the reflection of the directrix P across L, then as Q moves along the parabola, every
(line) R passes through the focus of the parabola.
Proof: Suppose the equation of the parabola C is x 2 4 py , where p 0 is a constant. Then C has
parametric equations
x f (t ) t and y g (t )

t2
,
4p

(17)

where t , . Since P is the directrix of the parabola, P has parametric equations

13

x0 s and y0 p , where s , . Let L be the tangent line to the circle C at the point
t2
Q f (t ), g (t ) t ,
. Note: As t increases from to Q will move along the parabola from
4p

left to right. With these choices for C and P, we have

f '(t ) 1, g '(t )

t
, x0 s, and y0 p.
2p

(18)

Substituting (17) and (18) into equations (5) gives the parametric equations (X, Y) for R

t 3 4 sp 2 4tp 2 st 2
4tsp 3t 2 p 4 p 3
and
,
Y

4 p2 t 2
4 p2 t 2

(19)

where s , and t , . Note: In this context, s is the object parameter and t is the
parameter of animation. Therefore, in spite of appearances, equations (19) are parametric equations

for a lineas they must be, since the reflection of a line (the directrix P) across a line (the dynamic
tangent L) must be another line (here R). What we must show is that, for each value of t, there is a
value for s such that R passes through the focus F 0, p of the parabola. To this end, consider the
linear system: X 0, Y p, which, from (19), is
t 3 4 sp 2 4tp 2 st 2

4 p2 t 2

2
3
4tsp 3t p 4 p

4 p2 t 2

0
.

(20)

Solving (20) for s gives: s = t. Therefore, every reflection R of the directrix of the parabola passes
through the focus of the parabola. This completes the proof of Theorem 6. The figures below
illustrate the result (here, with p = 2).

14

15

Some observations on symmetry

A topic studied in most elementary algebra courses are even and odd functions and the symmetries of
their graphs with respect to the y-axis and origin, respectively. In my opinion, a nice problem for
college algebra (or calculus) students is to generalize a bit and to formulate definitions for (i) a
function f whose graph is symmetric with respect to a point P x0 , y0 , or (ii) a function f whose
graph is symmetric with respect to the vertical line x x0 . One arrives at:
Definition 1: The graph of a function f is said to be symmetric with respect to the point P x0 , y0 if

f ( x0 x) y0 y0 f ( x0 x)
or, equivalently,

(21)

f ( x0 x) f ( x0 x) 2 y0
for all x such that x x0 domain of f. On the other hand,
Definition 2: The graph of a function f is said to be symmetric with respect to the vertical line x x0 if

f ( x0 x) f ( x0 x)

(22)

for all x such that x x0 domain of f.

Note that, in both definitions, it is not necessarily the case that x0 domain of f. In addition, from
(21) we see that the graph of a function f is symmetric with respect to the point P x0 , y0 if and only

if the function g defined by g ( x) f ( x x0 ) y0 is an odd function, i.e., g ( x) g ( x) . Similarly,

from (22) the graph of a function f is symmetric with respect to the vertical line x x0 if and only if
the function g defined by g ( x) f ( x x0 ) is an even function, i.e., g ( x) g ( x). With these
preliminaries behind usand, more importantly, the accumulated experience of having tried out
various combinations of curves and pointswe are now ready to state and prove our final theorem
(which has two parts).

16

Theorem 7: Let g be a function whose derivative exists on its domain.

(i) If, in addition, g is an odd function, let P = (0, 0) and let L be the tangent line to the graph of
y g ( x) at the point Q. Let R be the reflection of the point P across L. Then, the trajectory of
R (as Q moves along the graph of g) is symmetric with respect to the origin.
(ii) If g is an even function, let P (0, y0 ) be any (fixed) point on the y-axis and let L be the
tangent line to the graph of y g ( x) at the point Q. Let R be the reflection of the point P
across L. Then, the trajectory of R (as Q moves along the graph of g) is symmetric with respect
to the y-axis.
Proof: Let C be the graph of y g ( x). Then C has parametric equations (x, y), where
x f (t ) t and y g (t ) ,
for all t domain of g. Note that f '(t ) 1.
(i) Suppose that g is an odd function. Let P = (0, 0) and let L be the tangent line to the graph of
y g ( x) at the point Q. Let R be the reflection of the point P across L. Then, from Theorem 1,
parametric equations for the trajectory of R (as Q moves along the graph of g) are given by

2t g '(t ) 2 g (t ) g '(t )
2

and Y

1 g '(t )

2 g (t ) 2tg '(t )
1 g '(t )

(23)

where t domain of g. To show that the graph of equations (23) is symmetric with respect to the
origin, it suffices to show that X (t ) X (t ) and Y (t ) Y (t ). From (23), we have
2t g '(t ) 2 g (t ) g '(t )
2

X (t )

1 g '(t )

2t g '(t ) 2 g (t ) g '(t )
2

1 g '(t )

(since g odd g ' even )

17

2t g '(t ) 2 g (t ) g '(t )
2

1 g '(t )

X (t ).
Again, from (23), we have
Y (t )

2 g (t ) 2tg '(t )
1 g '(t )

2 g (t ) 2tg '(t )

1 g '(t )

2 g (t ) 2tg '(t )
1 g '(t )

Y (t ),
for all t domain of g. This completes the proof of part (i).
(ii) Suppose that g is an even function. Let P (0, y0 ) be any (fixed) point on the y-axis and let L be
the tangent line to the graph of y g ( x) at the point Q. Let R be the reflection of the point P
across L. Then, from Theorem 1, parametric equations for the trajectory of R (as Q moves along
the graph of g) are given by

2t g '(t ) 2 y0 g '(t ) 2 g (t ) g '(t )

2

1 g '(t )

2 g (t ) y0 g '(t ) 2tg '(t ) y0

2

and Y

1 g '(t )

(24)

where t domain of g. To show that the graph of equations (24) is symmetric with respect to the yaxis, it suffices to show that X (t ) X (t ) and Y (t ) Y (t ). From (24), we have
2t g '(t ) 2 y0 g '(t ) 2 g (t ) g '(t )
2

X (t )

1 g '(t )

18

2t g '(t ) 2 y0 g '(t ) 2 g (t ) g '(t )

2

1 g '(t )

(since g even g ' odd )

2t g '(t ) 2 y0 g '(t ) 2 g (t ) g '(t )

2

1 g '(t )

X (t ).
Again, from (24), we have
2 g (t ) y0 g '(t ) 2tg '(t ) y0
2

Y (t )

1 g '(t )

2 g (t ) y0 g '(t ) 2tg '(t ) y0

2

1 g '(t )

2 g (t ) y0 g '(t ) 2tg '(t ) y0

2

1 g '(t )

Y (t ),
for all t domain of g. This completes the proof of part (ii) and, hence, the proof of Theorem 7.
Below are three illustrations of the result. First, with g (t ) t 3 t , we have:

19

Then, with g (t ) cos(t ) 2, and t 2 , 2 , we have:

Finally, to illustrate the fact that the continuity of g is nowhere needed, with g (t )

2t 2
, we have:
t2 4

20

Concluding remarks
This paper grew out of a problem taken up by my Linear Algebra class last fall (November 2010).
Having reminded the class that they had studied the reflections of y f ( x) across the x- and y-axes, I
asked: But what if we wanted to reflect the graph of some function across any y mx b . How
would we proceed? After solving this problem, our discussion continued along the following lines:

What if we wanted to reflect any parameterized curve C [e.g., x f (t ) and y g (t ) ] across

the line y mx b ?

We have interpreted C to be the graph of some parametric equations. But we could just as
correctly interpret C as giving the time-dependent position of a moving particle (which I
represented as a small circle centered on C). The class was intrigued by an animation that
showed a particle moving along C while its reflection moved along the reflection of C across
the line.

I then stated how other animations (e.g., spinning C, or having C morph from one shape into
another) could also be reflected across the line y mx b .

Then, as class time was running out, I remarked, in passing, that we could also consider animating the
line of reflection ( y mx b ) in some waye.g., changing its slope, m, or its y-intercept, b.
When I returned to my office after class I thought about how the details of such changes would be
worked out. Since changing the vertical intercept seemed to lack pizzazz, the first change I considered
(not discussed in this paper) was spinning the line of reflection L about the origin. But after that, how
else could one animate the line of reflection? It was then that I considered allowing L to be the tangent
line to a circle whose point of tangency moves around the circle. Even if the object I reflected across L
was merely a (fixed) point P x0 , y0 , I knew that as L changed, the reflection of P across L would
change. It was this that led me to Theorem 3 (as a first application of Theorem 1).

I close this paper by indicatingin question formthree directions in which these investigations
could be continued:

21

What if the object P reflected across the dynamic tangent line L was a region in the plane
(rather than a point or a parameterized curve)? The only challenge, as I see it, would be finding
parametric equations for the region P x0 , y0 , i.e., some
x0 x0 (u , v) and y0 y0 (u, v).

(25)

In fact, if P is a region defined by (25) over some domain, parametric equations (X, Y) for the
(dynamic) reflection of the region P across the (dynamic) tangent line L are simply found by
substituting (25) into equations (2) above:

2 f (t ) g '(t ) x0 (u , v) f '(t ) 2 y0 (u, v) f '(t ) g '(t ) 2 g (t ) f '(t ) g '(t ) x0 (u, v) g '(t )

2

X (u, v, t )

f '(t ) g '(t )
2

and

(26)

2 g (t ) f '(t ) y0 (u , v) g '(t ) 2 x0 (u, v) f '(t ) g '(t ) 2 f (t ) f '(t ) g '(t ) y0 (u, v) f '(t )

2

Y (u, v, t )

f '(t ) g '(t )
2

where t I and the values of (the object parameters) u and v are determined by the domains of
x0 and y0 .

What if the dynamic line L was normal to the curve C at the point Q (rather than being tangent
to C)? (Clearly, equations similar to (2) would be obtained by using a unit normal vector to C
at the point Q, rather than a unit tangent vector U.)

How might one extend this analysis to 3-space where, for starters, instead of reflecting a fixed
point P x0 , y0 , z0 across a (dynamic) tangent line L to a parameterized curve C, you reflect
the point across a (dynamic) tangent plane to a parameterized surface S?

22