MS-A0211 - Differential and integral calculus 2, period III, 2020 David Radnell

1. Consider the curve with parametric equations x(t) = cos(t), y(t) = sin(t), z(t) = t2 for ≤ t0 ≤
6π.

(a) Sketch the curve and the tangent vector to the curve when t = π/4.
In the xy-plane the motion is 3 revolutions around the circle of radius 1. Note that
z(0) = 0, z(2π) = 4π 2 ≈ 39, z(4π) = 16π 2 ≈ 158, and z(6π) = 36π 2 ≈ 355. For a better
visualization, the plotted curve is x(t) = cos(t), y(t) = sin(t), z(t) = t2 /100 .

(b) Compute the tangent vector at t = π/4. Does your sketch match the computation?
The position vector is r(t) = hcos(t), sin(t), t2 i. The tangent vector is
√ √
r0 (π/4) = h− sin(π/4), cos(π/4), 2(π/4)i = h−1/ 2, 1/ 2, π/2i

It looks correct. At least it points in the right direction in the sense that it is h−, +, +i.

1
(c) Compute the arc length of the curve.
p √
First we compute the √ magnitude: ||r0 (t)|| = sin2 (t) + cos2 (t) + (2t)2 = 1 + 4t2 . The
R 6π
arc length is thus 0 1 + 4t2 dt. This integral is a bit annoying to compute. First let
R 12π √
x = 2t. Then dx = 2dt and the integral becomes (1/2) 0 1 + x2 dx. This is a very
standard integral you can look up how to solve. Here are the details.
First use the substitution x = tan θ. Then dx = sec2 θdθ, so we get
R√ R√ R√ R
1 + x2 dx = 1 + tan2 θ(sec2 θdθ) = sec2 θ(sec2 θdθ) = sec3 θdθ. Note that we
have assumed here that sec θ is positive. The final answer we obtain does not need this
restriction. You can check for yourself that the negative case leads to the same result.
Now integrate by parts R with u = sec θ and dv R= sec2 θ. Then du = sec θ tan θ and
v = tan θ. This gives sec3 θdθ = (sec θ)(tan θ) − sec θ tan2 θdθ
Applying a trigonometric identity and doing some algebra gives
Z Z
sec θdθ = (sec θ)(tan θ) − sec θ(sec2 θ − 1)dθ
3

Z Z
=⇒ sec θdθ = (sec θ)(tan θ) − (sec3 θ − sec θ)dθ
3

Z Z Z
3 3
=⇒ sec θdθ = (sec θ)(tan θ) − sec θdθ + sec θdθ

R R R
Adding sec3 θdθR to each side gives 2 sec3 θdθ = (sec θ)(tan θ) + sec θdθ Integrat-
Ring sec θ gives 2 sec3 θdθ = (sec θ)(tan θ) + ln | sec θ + tan θ| + C Dividing by 2 gives
sec3 θdθ = (1/2)(sec θ)(tan θ) + (1/2) ln | sec θ + tan θ| + C.√Now, since x = tan θ we have
x2 = tan2 θ, so 1 + x2 = 1 + tan2 θ, 1 + x2 = sec2 θ, and so 1 + x2 = sec θ. Thus, we can
plug in for sec θ and tan θ to obtain
Z √ √ √
1 + x2 dx = (1/2) 1 + x2 · x + (1/2) ln | 1 + x2 + x| + C.

And so
Z 6π √ h √ √ i6π
1 + 4t2 dt = (1/4) 1 + 4t2 · 2t + (1/4) ln | 1 + 4t2 + 2t| 

0 0
√ √
= 3π 1 + 144π 2 + ln( 1 + 144π 2 + 12π)

2
2. Consider the curve with parametric equation x(t) = cos(t), y(t) = cos2 (t) for −∞ < t < ∞.
(a) Sketch the curve and carefully describe the motion. Hint: think carefully about the range
of x(t) and y(t).
Oscillation on the part of the parabola with −1 ≤ x ≤ 1.
1.0

0.8

0.6

0.4

0.2

-1.0 -0.5 0.5 1.0

(b) Find the tangent vectors at the point (1/2, 1/4). Makes a sketch and relate your answers
to the direction of motion.
x = cos(t) = 1/2 implies t = π/3 or t = −π/3 plus multiples of 2π. Since x(0) = 1 and
x(π) = −1, the point is moving left when t = π/3 and right when t = −π/3. Let the
position vector be r(t) = hx(t), y(t)i. The tangent vector is
r0 (t) = −hsin(t), 2 sin(t) cos(t)i
√ √
Then r0 (π/3) = −h1/2, 3/2i and r0 (−π/3) = h1/2, 3/2i
(c) Find the tangent vector at the point (1, 1). Does your answer make sense. Is the curve
smooth at this point?
Directly substituting gives r0 (t) = h0, 0i. This zero vector gives no information about
the slope of the curve. Physically we can understand what is going on. The particle
moves along the curve, coming to a stop at the endpoint before moving back in the
opposite direction. So the velocity is zero at the endpoint. The curve is not smooth at the
endpoint and does not have a tangent vector at this point. Note however the velocity and
acceleration etc are smooth function so this is physically very reasonable motion. This is
the difference between a curve being smooth and the parametric coordinates being smooth
functions.
(d) Find the length of the curve.
Since the curve is being retraced infinitely often we need to be careful about the domain
of integration. The curve is traced out once as t goes from 0 to π. The arc length is thus
Z πq Z π p
2 2
sin (t) + 4 sin (t) cos2 (t) dt = | sin(t)| 1 + 4 cos2 (t) dt
0
Z0 π p
= sin(t) 1 + 4 cos2 (t) dt since sin(t) ≥ 0 on [0, π]
0
Z −2 √
= 1 + x2 (−1/2) dx letting x = 2 cos(t)
2
Z 2√
= (1/2) 1 + x2 dx
−2

And this integral we know how to compute because it is the same as in question 1(c) just
with different limits. Note that you reach the same integral by reparametrizing the curve
as ~r(t) = (t, t2 ), t ∈ [−1, 1].

3
3. Consider the function z = f (x, y) = x2 + 2y 2 .

(a) Sketch the graph of f (x, y). That is, the surface determined by z = f (x, y).
This is a paraboloid with elliptical cross-section determined by the equation x2 + 2y 2 = c
for any positive constant c.

(b) Find and sketch the level curves f = −1, f = 0, f = 1, f = 2 and f = 10.
For f = −1 the level curve is empty. For f = 0 the level curve is just the single
√ point
(0, 0).p For f = c, c > 0 we have the ellipse x2 + 2y 2 = c with intercepts ± c, 0) and
(0, ± c/2)
4

-2

-4
-4 -2 0 2 4

4. Do exercise 14.1.7 in Guichard’s Calculus text.

https://www.whitman.edu/mathematics/calculus online/section14.01.html
p
A cone can be defined by z = x2 + y 2 .
A paraboloid can be defined by z = x2 + y 2 .
For the cone the level surfaces are equally spaced. Each unit increase in z corresponds to an
equal increase in radius, no matter the height. For the parabola the level curves get closer
and closer together. For
√ example if we increase z by 1 from 1 to 2, the radius increases by
about 0.41 from 1 to 2 ≈ 1.41. But if√we increase z by 1 from say 9 to 10 then the radius
only increases by about 0.16 from 3 to 10 ≈ 3.16. This is because the slope of the parabola
increases.

4
5. Let f (x, y) be a function defined for all real numbers x, y. Is it possible for two different level
curves to intersect? That is, if a 6= b, is it possible that the level curves f = a and f = b
intersect?
No. We can prove this by contradiction. Say the curves f (x, y) = a and f (x, y) = b intersect.
This means there is a point (x0 , y0 ) such that f (x0 , y0 ) = a and f (x0 , y0 ) = b. But this is
impossible as f is a function, and so for a given input (x0 , y0 ) there can be only one output.
That is, we must have a = b. Contrast this to the case of the function we saw in class for which
the two-variable limit does not exist. In this case the computer pictures looked like the level
curves are intersecting. But the ”point of intersection” is not in the domain of the function.
There is a hole in the domain at that point.