lecture03
lecture03
MATH 136
Unit 3: Curves
It is also known as the velocity and just abbreviated r′ (t) . The first fundamental
form is
g = drT dr = |r′ (t)|2 .
It is the square of the speed. The arc length of the curve is defined as
Z b
L= |r′ (t)| dt .
a
Related to arc length is the action
Z b
I= |r′ (t)|2 dt .
a
which has the advantage that it can be computed better and produces equivalent
variational problems. Minimizing the arc-length is equivalent to minimize the action
and leads to geodesics. Here we are in flat Euclidean space and geodesics are straight
lines. We will say more about this in class. You show in the homework:
Theorem 1 (Archimedes). The straight line is the shortest path connecting A, B ∈ Rn .
1We will often write also just [cos(t), sin(t), t]T or simply [cos(t), sin(t), t] without the transpose for
typographic reasons.
Differential Geometry
3.3. A curve is called simple if r is it does not have self intersections. It is called
regular if the first fundamental form is nowhere zero. Equivalently, this means that
the velocity is nowhere zero. A simple closed curve in space is called a knot. An
example is the figure 8 knot
r(t) = [(2 + cos(2t)) cos(3t), (2 + cos(2t)) sin(3t), sin(4t)]T
parametrized on [0, 2π]. We talk more about this in class like that it lives on a torus
and why you can not tie knots in Rn for n > 3. r is simple can be rephrased that the
map r : [0, 2π) → R3 is injective.
3.4. A curve is parametrized by arc length if |⃗r ′ (t)| = 1 for all t. You will prove
in homework the following important result:
Theorem 2. Every smooth regular curve in Rn can be parametrized by arc-length.
3.5. It is custom to write r(s) to indicate that we have an arc length
√ parametrization.
√ √
For the helix above, the arc length parametrization is r(s) = [cos(s/ 2), sin(s/ 2, 1/ 2].
In general we do not bother to actually compute the arc length parametrization. Al-
ready in simple cases like the ellipse it would get nasty. We can use the theorem
however to build theory and prove stuff about curves.
Oliver Knill, knill@math.harvard.edu, Math 136, Fall, 2024