MATH 140 Calculus 1 • McGill University, Montreal

Corrected exercises: continuity and the Intermediate Value Theorem (MATH 140)

This is the corrected exercise set for continuity and the Intermediate Value Theorem in MATH 140, Calculus 1, at McGill University, section 2.5 of Stewart. The chapter looks light after the limit laws, and it is where the first proofs of the course are marked: a continuity question is graded on the three conditions written out, and an IVT question on the hypotheses checked before the theorem is named. Every number is exact and chosen to be done by hand.

The thread running through the whole set: continuity is CHECKED, never assumed. At a point, lim⁡x→af(x)=f(a)\lim_{x\to a} f(x) = f(a) hides three checks made in order. At a seam, both sides are computed as limits, never by plugging into a formula that gives 00\frac{0}{0}. The type of a discontinuity is read from the one-sided limits. And the IVT is used only after continuity on the whole CLOSED interval is written, and then gives existence only.

The traps named in the solutions: classifying a discontinuity by the first condition that fails, plugging a seam value into a 00\frac{0}{0} piece, removing an absolute value without cases, calling every zero of a denominator an asymptote, keeping one root of a quadratic condition on a parameter, forcing a value where the one-sided limits disagree, applying the IVT across a point where the function is undefined, testing f(0)f(0) and f(4)f(4) only and missing two roots, reading uniqueness or bounds into the IVT, and alternating signs blindly in a sign table.

10 corrected exercises • 100 points • 150 minutes

Revision sheet for this chapter → Every MATH 140 chapter →

Self-checking set Type your answers under each question: the page tells you right or wrong before you open the solution. With an account, every answer right on the first try earns points.

Course recap

  • • ff is continuous at aa when (1) f(a)f(a) is defined, (2) lim⁡x→af(x)\lim_{x\to a} f(x) exists, (3) lim⁡x→af(x)=f(a)\lim_{x\to a} f(x) = f(a). From the left or the right: the same with a one-sided limit.
  • • Removable: the limit exists but differs from f(a)f(a) or f(a)f(a) is undefined. Jump: finite one-sided limits that differ. Infinite: a one-sided limit is ±∞\pm\infty.
  • • Polynomials, rational functions, roots, exponentials, logarithms, trigonometric functions and their inverses are continuous on their domains; so are sums, products, and quotients where the denominator is non-zero.
  • • If lim⁡x→ag(x)=b\lim_{x\to a} g(x) = b and ff is continuous at bb, then lim⁡x→af(g(x))=f(b)\lim_{x\to a} f(g(x)) = f(b).
  • • Piecewise function: continuous on each open piece, then one equation per seam, left limit == right limit =f(a)= f(a).
  • • IVT: ff continuous on [a,b][a, b], NN strictly between f(a)f(a) and f(b)f(b), then f(c)=Nf(c) = N for some cc in (a,b)(a, b). Existence only.

Part A: the basics (/50)

Exercise 1: Reading continuity on a graph: three conditions, three kinds of discontinuity

A function ff is continuous at aa when lim⁡x→af(x)=f(a)\lim_{x\to a} f(x) = f(a). That single equation hides three checks, made in this order: (1) f(a)f(a) is defined; (2) lim⁡x→af(x)\lim_{x\to a} f(x) exists, which means that both one-sided limits exist, are finite and are equal; (3) that limit equals f(a)f(a). A discontinuity is REMOVABLE when the limit exists but (1) or (3) fails, a JUMP when the two one-sided limits are finite and different, and INFINITE when at least one one-sided limit is infinite.

The figure shows the graph of a function ff defined on [−4,5][-4, 5] except at x=−3x = -3 and x=3x = 3. A filled dot belongs to the graph, an open dot does not, and the dashed line x=3x = 3 is a vertical asymptote.

-4-3-2-112345-4-3-2-1123456x = 3y = f(x)x
  • a) For each of a=−3a = -3, −1-1, 11 and 33, read f(a)f(a), lim⁡x→a−f(x)\lim_{x\to a^-} f(x) and lim⁡x→a+f(x)\lim_{x\to a^+} f(x) (write undefined, ∞\infty or −∞-\infty where needed), then name the FIRST of the three conditions that fails.
  • b) Classify each of the four discontinuities as removable, jump or infinite.
  • c) Is ff continuous from the left at −1-1? From the right? At the endpoint x=−4x = -4, which kind of continuity is the only one that makes sense, and does ff have it?
  • d) At which points can ff be made continuous by defining or redefining ONE value? Give that value, and explain why no value works at the other points.
  • e) List the largest intervals on which ff is continuous. Is ff continuous on [−1,1)[-1, 1)? On [−1,1][-1, 1]?

Type your answers, the page tells you right or wrong 0/16

a)
b)
c)
d)
e)
Show the solution

Answers

  • a) a=−3a=-3: undefined, −1-1, −1-1, condition 1 fails. a=−1a=-1: f(−1)=3f(-1)=3, 11, 33, condition 2 fails. a=1a=1: f(1)=4f(1)=4, 22, 22, condition 3 fails. a=3a=3: undefined, −∞-\infty, ∞\infty, condition 1 fails.
  • b) Removable at −3-3 and at 11; jump at −1-1; infinite at 33.
  • c) At −1-1: continuous from the right, not from the left. At −4-4: continuity from the right only, and ff has it (f(−4)=−2f(-4) = -2).
  • d) Define f(−3)=−1f(-3) = -1 and redefine f(1)=2f(1) = 2; nothing works at −1-1 (the sides disagree) or at 33 (no finite limit).
  • e) [−4,−3)[-4, -3), (−3,−1)(-3, -1), [−1,1)[-1, 1), (1,3)(1, 3), (3,5](3, 5]. Yes on [−1,1)[-1, 1), no on [−1,1][-1, 1].

a) At a=−3a = -3 there is an open dot and no filled dot above or below it, so f(−3)f(-3) is undefined; on both sides the line climbs to the open dot, so lim⁡x→−3−f(x)=lim⁡x→−3+f(x)=−1\lim_{x\to -3^-} f(x) = \lim_{x\to -3^+} f(x) = -1. The first condition fails. At a=−1a = -1 the filled dot gives f(−1)=3f(-1) = 3; from the left the line arrives at the open dot (−1,1)(-1, 1), from the right the curve starts at the filled dot: lim⁡x→−1−f(x)=1\lim_{x\to -1^-} f(x) = 1 and lim⁡x→−1+f(x)=3\lim_{x\to -1^+} f(x) = 3. Condition 1 holds, condition 2 fails because the one-sided limits differ. At a=1a = 1 the filled dot is at height 44 while the curve passes through the open dot (1,2)(1, 2) from both sides: f(1)=4f(1) = 4 and both one-sided limits equal 22. Conditions 1 and 2 hold and condition 3 fails, since 2≠42 \ne 4. At a=3a = 3, f(3)f(3) is undefined, the curve plunges to −∞-\infty on the left and comes down from +∞+\infty on the right. Condition 1 fails. Read the value from the DOTS and the limits from the CURVE: the filled dot at (1,4)(1, 4) says nothing about where the curve is heading.

b) At −3-3 the limit exists (it is −1-1) and only the value is missing: removable. At 11 the limit exists (it is 22) but the value is wrong: removable as well. At −1-1 the one-sided limits are finite and different: jump. At 33 both one-sided limits are infinite: infinite discontinuity. The trap is to classify by the condition that fails first: at −3-3 and at 33 it is the same condition, the value is undefined, and yet one discontinuity is removable and the other is infinite. The TYPE is decided by the one-sided limits, never by the list of failed conditions.

c) Continuity from the left at −1-1 asks lim⁡x→−1−f(x)=f(−1)\lim_{x\to -1^-} f(x) = f(-1), that is 1=31 = 3: false. Continuity from the right asks lim⁡x→−1+f(x)=f(−1)\lim_{x\to -1^+} f(x) = f(-1), that is 3=33 = 3: true. So ff is continuous from the right at −1-1 and not from the left; the filled dot sits on the right-hand branch. At x=−4x = -4 the function is not defined to the left, so the only question that makes sense is continuity from the right: lim⁡x→−4+f(x)=−2=f(−4)\lim_{x\to -4^+} f(x) = -2 = f(-4), and ff is continuous from the right at −4-4. This is exactly what continuity on a closed interval asks at its endpoints.

d) A removable discontinuity is repaired by one value, the limit. At −3-3, defining f(−3)=−1f(-3) = -1 makes all three conditions hold. At 11, redefining f(1)=2f(1) = 2 instead of 44 does the same. At −1-1 no value can work: whatever f(−1)f(-1) is, it cannot equal both 11 and 33, so condition 2 still fails; changing one value never closes a gap between two different one-sided limits. At 33 there is no finite limit, so condition 2 fails whatever value is chosen. The question is always answered by the limit, and only a limit that exists and is finite can be used.

e) ff is continuous on [−4,−3)[-4, -3), (−3,−1)(-3, -1), [−1,1)[-1, 1), (1,3)(1, 3) and (3,5](3, 5]: each discontinuity is excluded, and an endpoint is kept exactly when ff has the one-sided continuity it needs there (from the right at −4-4 and at −1-1, from the left at 55, where f(5)=2f(5) = 2 is the end of the curve). On [−1,1)[-1, 1), ff is continuous at every interior point and from the right at −1-1: yes. On [−1,1][-1, 1], continuity at the right endpoint 11 would require lim⁡x→1−f(x)=f(1)\lim_{x\to 1^-} f(x) = f(1), that is 2=42 = 4: no. A closed interval makes a demand at each end, and those demands are one-sided.

Tick the exercises you have done or want to review: a free account, no password, keeps your ticks from one visit to the next and tells you which chapter to tackle next. Create your space, an email is enough.

Exercise 2: Locating and classifying discontinuities from a formula

Every function built from polynomials, roots, ∣x∣|x|, exponentials, logarithms and trigonometric functions is continuous at every point of its domain. The only candidates for a discontinuity are therefore the points EXCLUDED from the domain and the points where the formula CHANGES. At each candidate, compute the two one-sided limits: they alone decide the type.

Notation: ⌊x⌋\lfloor x \rfloor is the greatest integer less than or equal to xx, so ⌊2.7⌋=2\lfloor 2.7 \rfloor = 2 and ⌊−0.5⌋=−1\lfloor -0.5 \rfloor = -1.

  • a) Find and classify the discontinuities of f(x)=x2−4x2−3x+2f(x) = \frac{x^2 - 4}{x^2 - 3x + 2}.
  • b) Same question for g(x)=x2−1∣x−1∣g(x) = \frac{x^2 - 1}{|x - 1|}.
  • c) Same question for h(x)=x−⌊x⌋h(x) = x - \lfloor x \rfloor. At an integer nn, is hh continuous from the left or from the right?
  • d) At x=0x = 0, study k(x)=sin⁡(1x)k(x) = \sin\left(\frac{1}{x}\right) and m(x)=xsin⁡(1x)m(x) = x\sin\left(\frac{1}{x}\right). Does either discontinuity fit one of the three types? Can a value at 00 make the function continuous?
  • e) A student writes: the denominator of ff in a) vanishes at 11 and at 22, so the graph of ff has two vertical asymptotes. Correct the statement.

Type your answers, the page tells you right or wrong 0/12

a)
b)
c)
d)
e)
Show the solution

Answers

  • a) Removable at x=2x = 2 (limit 44), infinite at x=1x = 1.
  • b) Jump at x=1x = 1: left limit −2-2, right limit 22.
  • c) A jump at every integer nn: left limit 11, right limit 0=h(n)0 = h(n); continuous from the right.
  • d) kk: no limit and not infinite, none of the three types, no value works. mm: removable, m(0)=0m(0) = 0 works.
  • e) Only x=1x = 1 is an asymptote; at x=2x = 2 the factor x−2x - 2 cancels and the graph has a hole at (2,4)(2, 4).

a) Factor first: f(x)=(x−2)(x+2)(x−1)(x−2)f(x) = \frac{(x - 2)(x + 2)}{(x - 1)(x - 2)}, with domain R\mathbb{R} minus {1,2}\{1, 2\}. As a rational function, ff is continuous at every other point, so only 11 and 22 are candidates. At 22: for x≠2x \ne 2, f(x)=x+2x−1f(x) = \frac{x + 2}{x - 1}, which is continuous at 22, so lim⁡x→2f(x)=41=4\lim_{x\to 2} f(x) = \frac{4}{1} = 4. The limit exists and f(2)f(2) is undefined: removable discontinuity. At 11: in the simplified form, the numerator x+2x + 2 tends to 33 while x−1→0x - 1 \to 0; for x→1+x \to 1^+, x+2x−1\frac{x+2}{x-1} is a number close to 33 divided by a small positive number, so lim⁡x→1+f(x)=∞\lim_{x\to 1^+} f(x) = \infty. One infinite one-sided limit is enough: infinite discontinuity.

b) The domain excludes 11 only, and gg is continuous elsewhere. An absolute value is removed by cases. For x>1x > 1, ∣x−1∣=x−1|x - 1| = x - 1 and g(x)=(x−1)(x+1)x−1=x+1→2g(x) = \frac{(x - 1)(x + 1)}{x - 1} = x + 1 \to 2. For x<1x < 1, ∣x−1∣=−(x−1)|x - 1| = -(x - 1) and g(x)=−(x+1)→−2g(x) = -(x + 1) \to -2. The one-sided limits are finite and different: jump discontinuity at 11, and no value g(1)g(1) can repair it. Writing x2−1∣x−1∣=x+1\frac{x^2 - 1}{|x - 1|} = x + 1 without the cases gives the limit 22 and a false removable discontinuity: the absolute value changes the sign on one side only, and that is the whole point.

c) Between two integers, ⌊x⌋\lfloor x \rfloor is constant, so h(x)=x−nh(x) = x - n on (n,n+1)(n, n + 1) is a polynomial there and hh is continuous at every non-integer. At an integer nn: for n≤x<n+1n \le x < n + 1, h(x)=x−n→0h(x) = x - n \to 0 as x→n+x \to n^+; for n−1≤x<nn - 1 \le x < n, ⌊x⌋=n−1\lfloor x \rfloor = n - 1 and h(x)=x−n+1→1h(x) = x - n + 1 \to 1 as x→n−x \to n^-. And h(n)=n−n=0h(n) = n - n = 0. The one-sided limits 11 and 00 differ: a jump at every integer. Since lim⁡x→n+h(x)=0=h(n)\lim_{x\to n^+} h(x) = 0 = h(n), hh is continuous from the right at nn and not from the left. The trap is to compute ⌊x⌋\lfloor x \rfloor to the left of nn as nn: just below 33, ⌊2.99⌋=2\lfloor 2.99 \rfloor = 2.

d) For kk: sin⁡1x=1\sin\frac{1}{x} = 1 when 1x=π2+2jπ\frac{1}{x} = \frac{\pi}{2} + 2j\pi, that is at x=2(4j+1)πx = \frac{2}{(4j + 1)\pi}, and sin⁡1x=−1\sin\frac{1}{x} = -1 at x=2(4j+3)πx = \frac{2}{(4j + 3)\pi}, for j=1,2,3,…j = 1, 2, 3, \dots: both families of points come as close to 00 as we like, on the right (and, by oddness, on the left). So kk has no limit from either side, and since ∣k∣≤1|k| \le 1 neither one-sided limit is infinite. This discontinuity is none of the three types: the classification of Stewart is not a complete list. No value k(0)k(0) helps, because condition 2 fails. For mm: −∣x∣≤xsin⁡1x≤∣x∣-|x| \le x\sin\frac{1}{x} \le |x| for x≠0x \ne 0, since ∣sin⁡∣≤1|\sin| \le 1, and both bounds tend to 00, so by the squeeze theorem lim⁡x→0m(x)=0\lim_{x\to 0} m(x) = 0. The limit exists and m(0)m(0) is undefined: removable, and m(0)=0m(0) = 0 makes mm continuous, as the figure of the solution shows.

e) A vertical asymptote needs an infinite one-sided limit, and a zero of the denominator does not guarantee one. At 22 the numerator vanishes too, the factor x−2x - 2 cancels, and the limit is the finite number 44: the graph has a HOLE at (2,4)(2, 4), not an asymptote. At 11 only the denominator vanishes, and the asymptote is real. Correct statement: ff is discontinuous at 11 and at 22, with an infinite discontinuity (a vertical asymptote) at 11 and a removable one at 22. Simplify the common factors BEFORE naming any asymptote.

-1-0.75-0.5-0.250.250.50.751-1-0.75-0.5-0.250.250.50.751y = |x|y = -|x|y = x sin(1/x)x

Exercise 3: Continuity from the theorems: sums, quotients and compositions

The theorems of section 2.5. If ff and gg are continuous at aa, so are f+gf + g, f−gf - g, cfcf, fgfg, and fg\frac{f}{g} provided g(a)≠0g(a) \ne 0. Polynomials, rational functions, root functions, exponentials, logarithms, the trigonometric functions and their inverses are continuous at every point of their domains; at an endpoint of a domain, such as 00 for x\sqrt x, the continuity is one-sided.

Composition: if lim⁡x→ag(x)=b\lim_{x\to a} g(x) = b and ff is continuous at bb, then lim⁡x→af(g(x))=f(b)\lim_{x\to a} f(g(x)) = f(b). In particular, if gg is continuous at aa and ff is continuous at g(a)g(a), then f∘gf \circ g is continuous at aa. The hypothesis is on ff at the point bb, and it must be checked.

  • a) On which intervals is F(x)=4−x2x−1F(x) = \frac{\sqrt{4 - x^2}}{x - 1} continuous? Say what happens at the endpoints.
  • b) Same question for G(x)=ln⁡(x2−5x+6)G(x) = \ln(x^2 - 5x + 6) and for H(x)=ex1+cos⁡xH(x) = \frac{e^{\sqrt x}}{1 + \cos x}.
  • c) Compute lim⁡x→πsin⁡(x+sin⁡x)\lim_{x\to\pi} \sin(x + \sin x) and lim⁡x→2arctan⁡(x2−44x−8)\lim_{x\to 2} \arctan\left(\frac{x^2 - 4}{4x - 8}\right), naming the continuity used at each step.
  • d) The function u↦⌊u⌋u \mapsto \lfloor u \rfloor is discontinuous at u=0u = 0. Decide whether ⌊x2⌋\lfloor x^2 \rfloor and ⌊−x2⌋\lfloor -x^2 \rfloor are continuous at x=0x = 0.

Type your answers, the page tells you right or wrong 0/9

a)
First interval ,
Second interval ,
b)
GG: first interval ,
GG: second interval ,
c)
d)
Show the solution

Answers

  • a) On [−2,1)[-2, 1) and (1,2](1, 2], one-sided at −2-2 and 22.
  • b) GG: on (−∞,2)(-\infty, 2) and (3,∞)(3, \infty). HH: on [0,π)[0, \pi), (π,3π)(\pi, 3\pi), (3π,5π)(3\pi, 5\pi), and so on.
  • c) 00 and π4\frac{\pi}{4}
  • d) ⌊x2⌋\lfloor x^2 \rfloor is continuous at 00; ⌊−x2⌋\lfloor -x^2 \rfloor has a removable discontinuity there (limit −1-1, value 00).

a) The root requires 4−x2≥04 - x^2 \ge 0, that is −2≤x≤2-2 \le x \le 2, and the quotient requires x≠1x \ne 1. On this domain, 4−x24 - x^2 is a polynomial, so continuous, and  \sqrt{\ } is continuous on [0,∞)[0, \infty), so 4−x2\sqrt{4 - x^2} is continuous on [−2,2][-2, 2] by composition; x−1x - 1 is continuous and non-zero for x≠1x \ne 1, so the quotient is continuous on [−2,1)[-2, 1) and (1,2](1, 2]. At −2-2 and 22 the function is defined on one side only, and the continuity there is one-sided: from the right at −2-2, from the left at 22, with F(±2)=0F(\pm 2) = 0. At 11, FF is undefined. Answering R\mathbb{R} minus {1}\{1\} forgets the root; answering (−2,1)∪(1,2)(-2, 1) \cup (1, 2) throws away two endpoints where FF is perfectly continuous.

b) GG: the logarithm is continuous on (0,∞)(0, \infty), so GG is continuous wherever x2−5x+6=(x−2)(x−3)>0x^2 - 5x + 6 = (x - 2)(x - 3) > 0, that is for x<2x < 2 or x>3x > 3 (the product of two factors is positive outside the roots). GG is continuous on (−∞,2)(-\infty, 2) and (3,∞)(3, \infty), and undefined on [2,3][2, 3]. HH: x\sqrt x needs x≥0x \ge 0, eue^u is continuous everywhere, so the numerator is continuous on [0,∞)[0, \infty); the denominator is continuous and vanishes exactly when cos⁡x=−1\cos x = -1, that is at x=π+2jπx = \pi + 2j\pi. So HH is continuous on [0,π)[0, \pi), (π,3π)(\pi, 3\pi), (3π,5π)(3\pi, 5\pi), and so on: every odd multiple of π\pi is excluded, and 00 is kept with continuity from the right.

c) Let g(x)=x+sin⁡xg(x) = x + \sin x: a sum of continuous functions, so lim⁡x→πg(x)=g(π)=π+0=π\lim_{x\to\pi} g(x) = g(\pi) = \pi + 0 = \pi. The sine is continuous at π\pi, so lim⁡x→πsin⁡(g(x))=sin⁡π=0\lim_{x\to\pi} \sin(g(x)) = \sin \pi = 0. For the second limit, the inner function q(x)=x2−44x−8q(x) = \frac{x^2 - 4}{4x - 8} is NOT defined at 22, so direct substitution is impossible, but its limit exists: q(x)=(x−2)(x+2)4(x−2)=x+24→1q(x) = \frac{(x - 2)(x + 2)}{4(x - 2)} = \frac{x + 2}{4} \to 1. The arctangent is continuous at 11, so the limit is arctan⁡1=π4\arctan 1 = \frac{\pi}{4}. This is the limit form of the composition theorem: the inner function needs only a limit, the OUTER function needs continuity at that limit.

d) For ∣x∣<1|x| < 1 we have 0≤x2<10 \le x^2 < 1, so ⌊x2⌋=0\lfloor x^2 \rfloor = 0, which is also its value at 00: ⌊x2⌋\lfloor x^2 \rfloor is constant near 00, hence continuous there, although ⌊ ⌋\lfloor\ \rfloor is discontinuous at 00. For 0<∣x∣<10 < |x| < 1, −1<−x2<0-1 < -x^2 < 0, so ⌊−x2⌋=−1\lfloor -x^2 \rfloor = -1, while ⌊−0⌋=0\lfloor -0 \rfloor = 0: the limit is −1-1 and the value is 00, a removable discontinuity. The composition theorem gives a SUFFICIENT condition only. When the outer function is discontinuous at bb, anything can happen, and here the answer depends on the side from which the inner function reaches 00: x2x^2 arrives from above, where ⌊ ⌋\lfloor\ \rfloor is continuous from the right, and −x2-x^2 from below, where it jumps.

Exercise 4: Gluing pieces: one parameter, two parameters, and no parameter at all

A function defined by pieces is continuous on each OPEN interval where one formula applies, as soon as that formula is continuous there. Only the seams remain, and at a seam aa continuity means: left limit == right limit =f(a)= f(a). Each seam gives one equation, and the limits at a seam are LIMITS: a piece that takes the form 00\frac{0}{0} at the seam is simplified, never plugged into.

In each part, the unknown numbers are constants, and the function must be continuous on its whole domain.

  • a) Find kk such that f(x)=kx2+3xf(x) = kx^2 + 3x for x<1x < 1 and f(x)=2x3−kxf(x) = 2x^3 - kx for x≥1x \ge 1 is continuous on R\mathbb{R}.
  • b) Find aa and bb such that gg is continuous on R\mathbb{R}, where g(x)=x2−1x−1g(x) = \frac{x^2 - 1}{x - 1} if x<1x < 1, g(x)=ax+bg(x) = ax + b if 1≤x≤31 \le x \le 3, and g(x)=x2−x−6x−3g(x) = \frac{x^2 - x - 6}{x - 3} if x>3x > 3.
  • c) Find every cc such that h(x)=c2x2h(x) = c^2x^2 for x<1x < 1 and h(x)=3cx−2h(x) = 3cx - 2 for x≥1x \ge 1 is continuous on R\mathbb{R}.
  • d) Find cc such that p(x)=x2+3x∣x∣p(x) = \frac{x^2 + 3x}{|x|} for x≠0x \ne 0 and p(0)=cp(0) = c is continuous at 00.

Type your answers, the page tells you right or wrong 0/8

a)
b)
c)
d)
Show the solution

Answers

  • a) k=−12k = -\frac{1}{2} (both sides give 52\frac{5}{2})
  • b) a=32a = \frac{3}{2}, b=12b = \frac{1}{2} (g(1)=2g(1) = 2, g(3)=5g(3) = 5)
  • c) c=1c = 1 or c=2c = 2
  • d) No value of cc: the one-sided limits are −3-3 and 33 (jump).

a) For x<1x < 1, ff is a polynomial, and for x>1x > 1 as well, so ff is continuous at every x≠1x \ne 1. At the seam: lim⁡x→1−f(x)=k+3\lim_{x\to 1^-} f(x) = k + 3 (the left polynomial is continuous at 11), and lim⁡x→1+f(x)=f(1)=2−k\lim_{x\to 1^+} f(x) = f(1) = 2 - k. Continuity at 11 requires k+3=2−kk + 3 = 2 - k, so 2k=−12k = -1 and k=−12k = -\frac{1}{2}. Check: left −12+3=52-\frac{1}{2} + 3 = \frac{5}{2}, right 2+12=522 + \frac{1}{2} = \frac{5}{2}. The value f(1)f(1) is given by the piece that contains the equality sign, here x≥1x \ge 1, and it must match the OTHER side's limit.

b) The left piece is continuous on (−∞,1)(-\infty, 1) (its denominator vanishes only at 11, which is excluded), the middle one is a polynomial, and the right piece is continuous on (3,∞)(3, \infty). Seam at 11: g(1)=a+bg(1) = a + b and lim⁡x→1+g(x)=a+b\lim_{x\to 1^+} g(x) = a + b; on the left, plugging in gives 00\frac{0}{0}, so simplify: (x−1)(x+1)x−1=x+1→2\frac{(x - 1)(x + 1)}{x - 1} = x + 1 \to 2. Condition: a+b=2a + b = 2. Seam at 33: g(3)=3a+bg(3) = 3a + b and x2−x−6x−3=(x−3)(x+2)x−3=x+2→5\frac{x^2 - x - 6}{x - 3} = \frac{(x - 3)(x + 2)}{x - 3} = x + 2 \to 5 as x→3+x \to 3^+. Condition: 3a+b=53a + b = 5. Subtracting, 2a=32a = 3, so a=32a = \frac{3}{2} and b=12b = \frac{1}{2}. Check: g(1)=2g(1) = 2, g(3)=92+12=5g(3) = \frac{9}{2} + \frac{1}{2} = 5; the figure shows the three pieces forming one unbroken line. A student who plugs x=1x = 1 into the left piece and writes that gg cannot be continuous because of the 00\frac{0}{0} has confused the value of a formula at a point with the limit of the function there.

c) Both pieces are polynomials, so only x=1x = 1 matters: lim⁡x→1−h(x)=c2\lim_{x\to 1^-} h(x) = c^2 and lim⁡x→1+h(x)=h(1)=3c−2\lim_{x\to 1^+} h(x) = h(1) = 3c - 2. Condition: c2=3c−2c^2 = 3c - 2, that is c2−3c+2=(c−1)(c−2)=0c^2 - 3c + 2 = (c - 1)(c - 2) = 0, so c=1c = 1 or c=2c = 2. Both work: with c=1c = 1 the two sides give 11 and 11; with c=2c = 2 they give 44 and 44. Dividing the equation by cc, or keeping only the first root found, loses a mark: when the parameter enters squared, the answer is a SET of values, and each one is checked.

d) For x>0x > 0, ∣x∣=x|x| = x and p(x)=x(x+3)x=x+3→3p(x) = \frac{x(x + 3)}{x} = x + 3 \to 3. For x<0x < 0, ∣x∣=−x|x| = -x and p(x)=−(x+3)→−3p(x) = -(x + 3) \to -3. The one-sided limits are finite and different, so lim⁡x→0p(x)\lim_{x\to 0} p(x) does not exist and condition 2 fails whatever cc is: there is NO value of cc. This is a jump discontinuity, and the value at one point cannot close a gap between two sides. The expected answer is a sentence, not a number: a student who solves c=3c = 3 from the right side alone has made pp continuous from the right only.

-2-112345-2-112345678x + 1(3/2)x + 1/2x + 2x

Exercise 5: The Intermediate Value Theorem: hypotheses first, then a root

Intermediate Value Theorem. Suppose that ff is continuous on the CLOSED interval [a,b][a, b], that f(a)≠f(b)f(a) \ne f(b), and let NN be any number strictly between f(a)f(a) and f(b)f(b). Then there is a number cc in (a,b)(a, b) such that f(c)=Nf(c) = N. For a root, take N=0N = 0: a continuous function that changes sign on [a,b][a, b] vanishes somewhere in (a,b)(a, b).

The marks are earned by three written sentences: the function and the interval, WHY the function is continuous on the whole closed interval, and the two values with their signs. No calculator: every value used must be exact.

  • a) Prove that x5−3x+1=0x^5 - 3x + 1 = 0 has at least three real solutions, locating each between two consecutive integers.
  • b) Prove that 2x=3x2^x = 3x has at least two real solutions.
  • c) Prove that ln⁡x=2−x\ln x = 2 - x has a solution in (1,2)(1, 2).
  • d) tan⁡π4=1\tan\frac{\pi}{4} = 1 and tan⁡3π4=−1\tan\frac{3\pi}{4} = -1. Does tan⁡x=0\tan x = 0 have a solution in (π4,3π4)\left(\frac{\pi}{4}, \frac{3\pi}{4}\right)? What does this say about the theorem?
  • e) The equation x3+2x−4=0x^3 + 2x - 4 = 0 has a root in (1,2)(1, 2). With two steps of the bisection method, in exact fractions, find an interval of length 14\frac{1}{4} that contains a root.

Type your answers, the page tells you right or wrong 0/13

a)
b)
c)
d)
e)
Show the solution

Answers

  • a) f(−2)=−25f(-2) = -25, f(−1)=3f(-1) = 3, f(0)=1f(0) = 1, f(1)=−1f(1) = -1, f(2)=27f(2) = 27: roots in (−2,−1)(-2, -1), (0,1)(0, 1), (1,2)(1, 2).
  • b) g(x)=2x−3xg(x) = 2^x - 3x: g(0)=1g(0) = 1, g(1)=−1g(1) = -1, g(3)=−1g(3) = -1, g(4)=4g(4) = 4; roots in (0,1)(0, 1) and (3,4)(3, 4).
  • c) h(x)=ln⁡x+x−2h(x) = \ln x + x - 2: h(1)=−1<0<h(2)=ln⁡2h(1) = -1 < 0 < h(2) = \ln 2.
  • d) No solution; the theorem does not apply, since tan⁡\tan is undefined at π2\frac{\pi}{2}.
  • e) f(32)=198>0f\left(\frac{3}{2}\right) = \frac{19}{8} > 0, f(54)=2964>0f\left(\frac{5}{4}\right) = \frac{29}{64} > 0: a root in (1,54)\left(1, \frac{5}{4}\right).

a) Let f(x)=x5−3x+1f(x) = x^5 - 3x + 1, a polynomial, hence continuous on R\mathbb{R} and on every closed interval. Exact values: f(−2)=−32+6+1=−25f(-2) = -32 + 6 + 1 = -25, f(−1)=−1+3+1=3f(-1) = -1 + 3 + 1 = 3, f(0)=1f(0) = 1, f(1)=1−3+1=−1f(1) = 1 - 3 + 1 = -1, f(2)=32−6+1=27f(2) = 32 - 6 + 1 = 27. The sign changes on [−2,−1][-2, -1], on [0,1][0, 1] and on [1,2][1, 2], so by the IVT with N=0N = 0 there is a root in each of (−2,−1)(-2, -1), (0,1)(0, 1) and (1,2)(1, 2). These open intervals are disjoint, so the three roots are distinct: at least three real solutions. Note the words AT LEAST: the IVT counts sign changes, not roots, and a degree-5 polynomial could have up to five real roots. The figure of the solution shows the three crossings; ff has no fourth one, but proving it is not an IVT question.

b) The theorem is about ONE function, so move everything to one side: g(x)=2x−3xg(x) = 2^x - 3x, the difference of an exponential and a polynomial, continuous on R\mathbb{R}. g(0)=1−0=1>0g(0) = 1 - 0 = 1 > 0 and g(1)=2−3=−1<0g(1) = 2 - 3 = -1 < 0, so there is a root in (0,1)(0, 1). Then g(3)=8−9=−1<0g(3) = 8 - 9 = -1 < 0 and g(4)=16−12=4>0g(4) = 16 - 12 = 4 > 0, so there is another in (3,4)(3, 4). Two disjoint intervals, at least two solutions. Finding the second one is a matter of looking further: g(2)=4−6=−2g(2) = 4 - 6 = -2 stays negative, and the sign returns only at 44. Testing only 00 and 44 would show g(0)>0g(0) > 0 and g(4)>0g(4) > 0, no sign change, and hide both roots.

c) Let h(x)=ln⁡x+x−2h(x) = \ln x + x - 2. The logarithm is continuous on (0,∞)(0, \infty) and x−2x - 2 is a polynomial, so hh is continuous on (0,∞)(0, \infty), and in particular on [1,2][1, 2]. h(1)=0+1−2=−1<0h(1) = 0 + 1 - 2 = -1 < 0 and h(2)=ln⁡2+0=ln⁡2>0h(2) = \ln 2 + 0 = \ln 2 > 0, because 2>12 > 1 and ln⁡\ln is increasing with ln⁡1=0\ln 1 = 0. By the IVT there is cc in (1,2)(1, 2) with h(c)=0h(c) = 0, that is ln⁡c=2−c\ln c = 2 - c. The sign of ln⁡2\ln 2 is decided exactly, without any decimal: that is what no calculator means.

d) tan⁡x=sin⁡xcos⁡x\tan x = \frac{\sin x}{\cos x} is zero exactly when sin⁡x=0\sin x = 0, that is at the multiples of π\pi, and none lies in (π4,3π4)\left(\frac{\pi}{4}, \frac{3\pi}{4}\right). So there is no solution, although tan⁡\tan goes from 11 to −1-1. The theorem is not contradicted, because its hypothesis fails: tan⁡\tan is undefined at π2\frac{\pi}{2}, which lies inside [π4,3π4]\left[\frac{\pi}{4}, \frac{3\pi}{4}\right], so it is not continuous on that closed interval. It jumps from +∞+\infty to −∞-\infty at π2\frac{\pi}{2} instead of crossing 00. A sign change is evidence of a root ONLY for a function continuous on the whole closed interval, and that has to be written.

e) f(x)=x3+2x−4f(x) = x^3 + 2x - 4 is a polynomial, continuous on every closed interval. f(1)=−1<0f(1) = -1 < 0 and f(2)=8>0f(2) = 8 > 0. First step, the midpoint 32\frac{3}{2}: f(32)=278+3−4=278−88=198>0f\left(\frac{3}{2}\right) = \frac{27}{8} + 3 - 4 = \frac{27}{8} - \frac{8}{8} = \frac{19}{8} > 0. The sign change is now on [1,32]\left[1, \frac{3}{2}\right]. Second step, the midpoint 54\frac{5}{4}: f(54)=12564+104−4=12564+16064−25664=2964>0f\left(\frac{5}{4}\right) = \frac{125}{64} + \frac{10}{4} - 4 = \frac{125}{64} + \frac{160}{64} - \frac{256}{64} = \frac{29}{64} > 0. The sign change is on [1,54]\left[1, \frac{5}{4}\right], and by the IVT there is a root in (1,54)\left(1, \frac{5}{4}\right), an interval of length 14\frac{1}{4}. Each step keeps the half where the sign changes, and each step is one more application of the IVT, on a smaller closed interval where ff is still continuous.

-2-1.5-1-0.50.511.52-5-4-3-2-112345(-1, 3)(0, 1)(1, -1)f(-2) = -25f(2) = 27x

Part B: problems and reasoning (/50)

Exercise 6: What the IVT guarantees, and what it does not

The IVT is an EXISTENCE theorem: it guarantees at least one cc, and says nothing about how many there are, where exactly they lie, or what happens when its hypotheses fail. Its most useful consequence sits silently behind every sign table: a function continuous on an interval and never zero there keeps the same sign on the whole interval.

The figure shows ONE function ff continuous on [0,6][0, 6] with f(0)=1f(0) = 1 and f(6)=5f(6) = 5, together with the line y=3y = 3.

123456123456y = 3y = f(x)x
  • a) Let ff be ANY function continuous on [0,6][0, 6] with f(0)=1f(0) = 1 and f(6)=5f(6) = 5. Which statements must be true? (i) f(c)=3f(c) = 3 for some cc in (0,6)(0, 6); (ii) f(c)=3f(c) = 3 for exactly one cc; (iii) 1≤f(x)≤51 \le f(x) \le 5 for every xx in [0,6][0, 6]; (iv) f(c)=0f(c) = 0 for some cc; (v) ff takes every value between 11 and 55.
  • b) The converse of the IVT is false. Give a function on [0,2][0, 2] that takes every value between f(0)f(0) and f(2)f(2) and is not continuous.
  • c) Prove that if gg is continuous on an interval II and g(x)≠0g(x) \ne 0 for every xx in II, then gg has the same sign on all of II. Use it to solve x2−x−2x−3>0\frac{x^2 - x - 2}{x - 3} > 0.
  • d) A function gg is continuous on [1,5][1, 5], its only zeros are x=2x = 2 and x=4x = 4, and g(1)=3g(1) = 3, g(3)=−1g(3) = -1. Give the sign of gg on [1,2)[1, 2) and on (2,4)(2, 4). What can be said about its sign on (4,5](4, 5]?

Type your answers, the page tells you right or wrong 0/10

a)
c)
Solution: first interval ,
Solution: second interval ,
d)
Show the solution

Answers

  • a) (i) and (v) must be true; (ii), (iii) and (iv) need not be.
  • b) f(x)=xf(x) = x on [0,1][0, 1], f(x)=x−1f(x) = x - 1 on (1,2](1, 2]: every value of [0,1][0, 1] is taken, and ff jumps at 11.
  • c) A sign change would force a zero by the IVT. Solution: (−1,2)∪(3,∞)(-1, 2) \cup (3, \infty).
  • d) Positive on [1,2)[1, 2), negative on (2,4)(2, 4); one constant sign on (4,5](4, 5], which the data do not decide.

a) (i) TRUE: ff is continuous on [0,6][0, 6] and 33 lies strictly between f(0)=1f(0) = 1 and f(6)=5f(6) = 5, so the IVT gives such a cc. (ii) FALSE: the function of the figure meets the line y=3y = 3 five times; the IVT gives existence, never uniqueness. (iii) FALSE: the same function dips below 11 (to about 0.70.7) and rises above 55 (to about 5.35.3); the IVT says which values ARE taken, not which values are NOT. (iv) FALSE in general: 00 is not between 11 and 55, and the function of the figure, like 1+2x31 + \frac{2x}{3}, never vanishes; but it is not forbidden either, some continuous function with these end values does reach 00. (v) TRUE: each NN strictly between 11 and 55 is reached by the IVT, and 11 and 55 are reached at the endpoints.

b) Take f(x)=xf(x) = x for 0≤x≤10 \le x \le 1 and f(x)=x−1f(x) = x - 1 for 1<x≤21 < x \le 2. Then f(0)=0f(0) = 0 and f(2)=1f(2) = 1, and every value NN in [0,1][0, 1] is already taken on [0,1][0, 1], at x=Nx = N. Yet lim⁡x→1−f(x)=1\lim_{x\to 1^-} f(x) = 1 and lim⁡x→1+f(x)=0\lim_{x\to 1^+} f(x) = 0: a jump at 11. Taking every intermediate value does not make a function continuous. The theorem reads in one direction only: continuous on [a,b][a, b] implies every intermediate value is taken.

c) Suppose gg took two opposite signs on II, say g(p)<0<g(q)g(p) < 0 < g(q) with pp and qq in II. Since II is an interval, it contains every point between pp and qq, so gg is continuous on the closed interval with endpoints pp and qq, and the IVT with N=0N = 0 gives a zero of gg between them, contradicting g(x)≠0g(x) \ne 0 on II. So gg has one sign on II. Application: r(x)=x2−x−2x−3=(x+1)(x−2)x−3r(x) = \frac{x^2 - x - 2}{x - 3} = \frac{(x + 1)(x - 2)}{x - 3} is continuous wherever it is defined, vanishes at −1-1 and 22, and is undefined at 33. These three points cut the line into (−∞,−1)(-\infty, -1), (−1,2)(-1, 2), (2,3)(2, 3) and (3,∞)(3, \infty), on each of which rr is continuous and non-zero, so ONE test value per interval gives its sign: r(−2)=−45r(-2) = -\frac{4}{5}, r(0)=23r(0) = \frac{2}{3}, r(52)=7/4−1/2=−72r\left(\frac{5}{2}\right) = \frac{7/4}{-1/2} = -\frac{7}{2}, r(4)=10r(4) = 10. The sign line of the solution sums it up, and r(x)>0r(x) > 0 exactly on (−1,2)∪(3,∞)(-1, 2) \cup (3, \infty). Forgetting the point 33 merges (2,3)(2, 3) and (3,∞)(3, \infty) into one interval, on which rr is not continuous: a single test value would then give the wrong sign on half of it.

d) By c), gg has one sign on [1,2)[1, 2), which contains no zero, and g(1)=3>0g(1) = 3 > 0: gg is positive there. On (2,4)(2, 4), again one sign, and g(3)=−1<0g(3) = -1 < 0: negative. On (4,5](4, 5] the sign is constant, but the data do not decide it: a zero is not necessarily a change of sign. Two functions fit every datum: g1(x)=(x−2)(x−4)g_1(x) = (x - 2)(x - 4), with g1(1)=3g_1(1) = 3, g1(3)=−1g_1(3) = -1 and g1(5)=3>0g_1(5) = 3 > 0; and g2(x)=(x−2)(x−4)g_2(x) = (x - 2)(x - 4) for x≤4x \le 4, g2(x)=4−xg_2(x) = 4 - x for x>4x > 4, continuous at 44 (both pieces give 00), with g2(5)=−1<0g_2(5) = -1 < 0. A sign table built by alternating ++ and −- at each zero would have guessed wrong for g2g_2.

-12300undef.−+−+

Exercise 7: Continuous extension: filling a hole with the only value that works

If ff is not defined at aa but lim⁡x→af(x)=L\lim_{x\to a} f(x) = L exists and is finite, the function equal to f(x)f(x) for x≠ax \ne a and to LL at aa is continuous at aa: it is the continuous extension of ff at aa, and LL is the ONLY value that works. If the limit does not exist, or is infinite, no extension exists.

The limits needed here use only algebra: factor, multiply by a conjugate, remove an absolute value by cases.

  • a) Extend f(x)=x3−8x−2f(x) = \frac{x^3 - 8}{x - 2} continuously at x=2x = 2.
  • b) Extend g(x)=x−3x−9g(x) = \frac{\sqrt x - 3}{x - 9} continuously at x=9x = 9.
  • c) Extend F(x)=e2x−1ex−1F(x) = \frac{e^{2x} - 1}{e^x - 1} continuously at x=0x = 0.
  • d) Find the domain of u(x)=x−1∣x∣−1u(x) = \frac{x - 1}{|x| - 1}, then decide at each excluded point whether uu can be extended continuously.
  • e) Find aa and bb such that q(x)=x2+ax+bx−2q(x) = \frac{x^2 + ax + b}{x - 2} can be extended continuously at x=2x = 2 with the value 55.

Type your answers, the page tells you right or wrong 0/9

a)
b)
c)
d)
e)
Show the solution

Answers

  • a) f(2)=12f(2) = 12
  • b) g(9)=16g(9) = \frac{1}{6}
  • c) F(0)=2F(0) = 2
  • d) Domain R\mathbb{R} minus {−1,1}\{-1, 1\}; u(1)=1u(1) = 1 extends; no extension at −1-1 (infinite discontinuity).
  • e) a=1a = 1, b=−6b = -6

a) ff is a rational function, continuous on its domain R\mathbb{R} minus {2}\{2\}. At 22 the form is 00\frac{0}{0}, so factor the difference of cubes: x3−8=(x−2)(x2+2x+4)x^3 - 8 = (x - 2)(x^2 + 2x + 4), and for x≠2x \ne 2, f(x)=x2+2x+4f(x) = x^2 + 2x + 4, a polynomial continuous at 22. Hence lim⁡x→2f(x)=4+4+4=12\lim_{x\to 2} f(x) = 4 + 4 + 4 = 12, and setting f(2)=12f(2) = 12 gives the continuous extension, which is simply the polynomial x2+2x+4x^2 + 2x + 4 on all of R\mathbb{R}. The simplification by x−2x - 2 is allowed because a limit never looks at x=2x = 2 itself.

b) The domain is x≥0x \ge 0, x≠9x \ne 9. At 99 the form is 00\frac{0}{0}. For x≥0x \ge 0, x−9=(x−3)(x+3)x - 9 = (\sqrt x - 3)(\sqrt x + 3), a difference of squares in disguise (equivalently, multiply by the conjugate x+3\sqrt x + 3). So for x≠9x \ne 9, g(x)=1x+3g(x) = \frac{1}{\sqrt x + 3}, continuous at 99, and lim⁡x→9g(x)=13+3=16\lim_{x\to 9} g(x) = \frac{1}{3 + 3} = \frac{1}{6}. Setting g(9)=16g(9) = \frac{1}{6} extends gg continuously. A frequent slip is to write 19+3=112\frac{1}{\sqrt 9 + 3} = \frac{1}{12}, adding 99 instead of 9\sqrt 9: evaluate the simplified formula slowly.

c) The denominator vanishes only at x=0x = 0, where the form is 00\frac{0}{0}. Put w=exw = e^x: the numerator is w2−1=(w−1)(w+1)w^2 - 1 = (w - 1)(w + 1), so for x≠0x \ne 0, F(x)=ex+1F(x) = e^x + 1. The exponential is continuous at 00, so lim⁡x→0F(x)=1+1=2\lim_{x\to 0} F(x) = 1 + 1 = 2, and F(0)=2F(0) = 2 is the continuous extension. No rule beyond algebra is needed: an exponential expression often factors once it is read as a polynomial in exe^x.

d) ∣x∣−1=0|x| - 1 = 0 exactly when x=1x = 1 or x=−1x = -1, so the domain is R\mathbb{R} minus {−1,1}\{-1, 1\}, and uu is continuous on it. Remove the absolute value by cases. Near 11, x>0x > 0, so ∣x∣=x|x| = x and u(x)=x−1x−1=1u(x) = \frac{x - 1}{x - 1} = 1: lim⁡x→1u(x)=1\lim_{x\to 1} u(x) = 1, and u(1)=1u(1) = 1 extends uu continuously at 11. Near −1-1, x<0x < 0, so ∣x∣=−x|x| = -x and u(x)=x−1−x−1=1−x1+xu(x) = \frac{x - 1}{-x - 1} = \frac{1 - x}{1 + x}: the numerator tends to 22 while the denominator tends to 00, so ∣u(x)∣|u(x)| grows without bound and no finite limit exists. At −1-1 the discontinuity is infinite and no extension exists. The solution figure shows both behaviours on one graph: a hole at (1,1)(1, 1) on a horizontal line, and an asymptote at x=−1x = -1. The same two symbols, a zero denominator and a zero numerator, meet at 11; only the denominator vanishes at −1-1.

e) Suppose the limit at 22 is a finite number LL. The numerator N(x)=x2+ax+bN(x) = x^2 + ax + b equals (x−2)q(x)(x - 2)q(x) for x≠2x \ne 2, so lim⁡x→2N(x)=0⋅L=0\lim_{x\to 2} N(x) = 0 \cdot L = 0; since NN is a polynomial, continuous at 22, this limit is N(2)N(2). So 4+2a+b=04 + 2a + b = 0 is NECESSARY. With b=−4−2ab = -4 - 2a, N(x)=x2+ax−4−2a=(x−2)(x+2+a)N(x) = x^2 + ax - 4 - 2a = (x - 2)(x + 2 + a), so q(x)=x+2+aq(x) = x + 2 + a for x≠2x \ne 2 and lim⁡x→2q(x)=4+a\lim_{x\to 2} q(x) = 4 + a. The required value is 55, so a=1a = 1 and b=−6b = -6. Check: x2+x−6x−2=(x−2)(x+3)x−2=x+3→5\frac{x^2 + x - 6}{x - 2} = \frac{(x - 2)(x + 3)}{x - 2} = x + 3 \to 5. The first condition is the key sentence: a quotient whose denominator tends to 00 can have a finite limit only if its numerator tends to 00 as well.

-4-3-2-11234-6-4-2246hole at (1, 1)x = -1x

Exercise 8: Five statements to correct

Each statement below was written by a student on a MATH 140 assignment, and each is false. Say what is wrong, give a counterexample, and write the correct statement.

  • a) If lim⁡x→af(x)\lim_{x\to a} f(x) exists, then ff is continuous at aa.
  • b) If f(a)f(a) and f(b)f(b) have the same sign, then ff has no zero in [a,b][a, b].
  • c) If ff and gg are both discontinuous at aa, then f+gf + g is discontinuous at aa.
  • d) If ∣f∣|f| is continuous at aa, then ff is continuous at aa.
  • e) If ff is continuous on (a,b)(a, b) and f(a)<0<f(b)f(a) < 0 < f(b), then ff has a zero in (a,b)(a, b).

Type your answers, the page tells you right or wrong 0/5

a)
b)
c)
d)
e)
Show the solution

Answers

  • a) False: x2−4x−2\frac{x^2 - 4}{x - 2} has limit 44 at 22 and is undefined there. Continuity needs the limit to EQUAL f(a)f(a).
  • b) False: x2−1x^2 - 1 on [−2,2][-2, 2]. A sign change is sufficient for a zero, not necessary.
  • c) False: HH and 1−H1 - H jump at 00, their sum is 11. True if one is continuous and the other not.
  • d) False: s(x)=1s(x) = 1 for x≥0x \ge 0, −1-1 for x<0x < 0. The converse holds.
  • e) False: f(0)=−1f(0) = -1, f=1f = 1 on (0,1](0, 1]. Continuity is needed on the CLOSED interval [a,b][a, b].

a) FALSE. f(x)=x2−4x−2f(x) = \frac{x^2 - 4}{x - 2} has lim⁡x→2f(x)=lim⁡x→2(x+2)=4\lim_{x\to 2} f(x) = \lim_{x\to 2}(x + 2) = 4, yet f(2)f(2) is undefined; and if we set f(2)=0f(2) = 0, the limit still exists and ff is still discontinuous. The existence of the limit is only condition 2 of the three. Correct statement: ff is continuous at aa if and only if lim⁡x→af(x)\lim_{x\to a} f(x) exists AND equals f(a)f(a), which requires f(a)f(a) to be defined.

b) FALSE. f(x)=x2−1f(x) = x^2 - 1 has f(−2)=f(2)=3>0f(-2) = f(2) = 3 > 0, and yet f(−1)=f(1)=0f(-1) = f(1) = 0: two zeros in [−2,2][-2, 2]. The IVT is an implication in one direction: continuity plus a sign change gives a zero; the absence of a sign change gives nothing. Correct statement: if ff is continuous on [a,b][a, b] and f(a)f(a), f(b)f(b) have opposite signs, then ff has a zero in (a,b)(a, b). With the same sign at both ends, ff may have no zero, one zero (a touching zero, like x2x^2 on [−1,1][-1, 1]), or several.

c) FALSE. Let H(x)=0H(x) = 0 for x<0x < 0 and H(x)=1H(x) = 1 for x≥0x \ge 0. Then HH and 1−H1 - H both jump at 00, and H+(1−H)=1H + (1 - H) = 1 is constant, hence continuous. The jumps cancel. Correct statement: if ff is continuous at aa and gg is discontinuous at aa, then f+gf + g is discontinuous at aa, because otherwise g=(f+g)−fg = (f + g) - f would be a difference of two functions continuous at aa, hence continuous at aa. Two discontinuities can cancel, one cannot be cancelled by a continuous function.

d) FALSE. Let s(x)=1s(x) = 1 for x≥0x \ge 0 and s(x)=−1s(x) = -1 for x<0x < 0. Then ∣s(x)∣=1|s(x)| = 1 for every xx, continuous everywhere, while ss jumps from −1-1 to 11 at 00. The absolute value erases exactly the information (the sign) that makes ss jump. Correct statement: the converse is true, if ff is continuous at aa then ∣f∣|f| is continuous at aa, as the composition of ff with the continuous function ∣ ∣|\ |. One useful special case survives in the direction of the statement: ∣f(x)∣→0|f(x)| \to 0 implies f(x)→0f(x) \to 0, so if f(a)=0f(a) = 0 and ∣f∣|f| is continuous at aa, then ff is continuous at aa.

e) FALSE. On [0,1][0, 1] let f(0)=−1f(0) = -1 and f(x)=1f(x) = 1 for 0<x≤10 < x \le 1. Then ff is continuous on (0,1)(0, 1), f(0)<0<f(1)f(0) < 0 < f(1), and ff never vanishes: the only jump is AT the endpoint 00, where lim⁡x→0+f(x)=1≠f(0)\lim_{x\to 0^+} f(x) = 1 \ne f(0). Continuity on the open interval says nothing about how ff joins its endpoint values. Correct statement: if ff is continuous on the CLOSED interval [a,b][a, b], which includes continuity from the right at aa and from the left at bb, and f(a)<0<f(b)f(a) < 0 < f(b), then ff has a zero in (a,b)(a, b).

Exercise 9: The hiker who climbs on Saturday and comes down on Sunday

A hiker leaves the base of a trail at 6:00 on Saturday and reaches the summit, 1212 km further along the trail, at 14:00. She sleeps at the summit, leaves at 6:00 on Sunday by the SAME trail, is back at the base at 9:00 and waits there for her ride until 14:00. Measure time tt in hours after 6:00, so that 0≤t≤80 \le t \le 8 on both days, and let u(t)u(t) and v(t)v(t) be her distances from the base along the trail, in km, on Saturday and on Sunday.

The figure shows one possible pair of walks: on Saturday she climbs at 22 km/h for four hours, rests for an hour at the viewpoint 88 km up, then covers the last 44 km in three hours; on Sunday she comes down at 44 km/h. The proof asked in b) must not use these particular speeds.

1234567824681012Saturday: u(t)Sunday: v(t)t (hours after 6:00)km from the base
  • a) Give u(0)u(0), u(8)u(8), v(0)v(0) and v(8)v(8). Why are uu and vv continuous on [0,8][0, 8], including during the rest and the wait?
  • b) Prove that some point of the trail is passed at exactly the same time of day on Saturday and on Sunday, whatever the speeds, stops and even backtracking of either walk. Name the function, the interval and each hypothesis.
  • c) For the walks of the figure, find that time of day and that point exactly.
  • d) Is such a point unique in general? Describe a pair of walks, on the same trail, with three such moments.
  • e) Two changes to the story. (1) On Sunday she comes down by a DIFFERENT trail, which meets the first only at the base and at the summit. (2) On Sunday she leaves the summit at 10:00 instead of 6:00, and still descends at 44 km/h. Does the conclusion of b) survive each change?

Type your answers, the page tells you right or wrong 0/11

a)
c)
d)
e)
Show the solution

Answers

  • a) u(0)=0u(0) = 0, u(8)=12u(8) = 12, v(0)=12v(0) = 12, v(8)=0v(8) = 0; a position cannot jump, so both are continuous, constant pieces included.
  • b) d(t)=u(t)−v(t)d(t) = u(t) - v(t) is continuous on [0,8][0, 8] with d(0)=−12<0<d(8)=12d(0) = -12 < 0 < d(8) = 12: by the IVT, u(c)=v(c)u(c) = v(c) for some cc in (0,8)(0, 8).
  • c) At 8:00 (t=2t = 2), 44 km from the base.
  • d) Not unique. Sunday at a steady 1.51.5 km/h; Saturday up to 99 km at 9:00, back down to 55 km at 10:00, up to the summit at 14:00: three meetings.
  • e) (1) The argument collapses: positions on two trails cannot be compared. (2) It survives: d(0)=−12d(0) = -12, d(8)=12d(8) = 12; here they meet at 11:00, 88 km up.

a) On Saturday she starts at the base and ends at the summit: u(0)=0u(0) = 0 and u(8)=12u(8) = 12. On Sunday she starts at the summit and is at the base from t=3t = 3 on: v(0)=12v(0) = 12 and v(8)=0v(8) = 0. A distance along a trail cannot jump from one value to another without passing through the values in between: a walker has no teleporter, so uu and vv are continuous functions of time. A rest or a wait is a CONSTANT piece of the graph, which is continuous; what matters at the junctions, such as t=3t = 3 on Sunday, is that the constant value is the one she has just reached, v(3)=0v(3) = 0, so the two pieces meet.

b) Let d(t)=u(t)−v(t)d(t) = u(t) - v(t) on [0,8][0, 8]. As a difference of two continuous functions, dd is continuous on the closed interval [0,8][0, 8]. Its end values are d(0)=0−12=−12<0d(0) = 0 - 12 = -12 < 0 and d(8)=12−0=12>0d(8) = 12 - 0 = 12 > 0. By the Intermediate Value Theorem with N=0N = 0, there is cc in (0,8)(0, 8) with d(c)=0d(c) = 0, that is u(c)=v(c)u(c) = v(c): at the time cc hours after 6:00, she is at the same point of the trail on both days. The speeds never entered: only the continuity of the two walks and their four end values did, which is why the conclusion holds for any walks, backtracking included. The trick is to compare the two days on the SAME clock, as if two hikers left at 6:00 on the same morning, one from each end: they must meet.

c) For 0≤t≤30 \le t \le 3, u(t)=2tu(t) = 2t (she is still in her first four hours) and v(t)=12−4tv(t) = 12 - 4t. Setting 2t=12−4t2t = 12 - 4t gives 6t=126t = 12, so t=2t = 2, which lies in [0,3][0, 3]. For 3<t≤83 < t \le 8, v(t)=0v(t) = 0 while u(t)>0u(t) > 0, so no other solution. The time is 22 hours after 6:00, that is 8:00, and the point is u(2)=4u(2) = 4 km from the base; check on Sunday: v(2)=12−8=4v(2) = 12 - 8 = 4. On the figure it is where the two graphs cross, and the IVT had promised that they would, before any computation.

d) In general the point is not unique: the IVT guarantees at least one zero of dd, never exactly one. Each change of sign of the continuous function dd gives a meeting, so it is enough to make dd change sign three times. Take a slow Sunday: she comes down at a steady 1.51.5 km/h, v(t)=12−1.5tv(t) = 12 - 1.5t, reaching the base at 14:00. On Saturday she climbs at 33 km/h to 99 km at 9:00 (t=3t = 3), where v(3)=7.5v(3) = 7.5: she is now ABOVE Sunday's position, so dd went from −12-12 to 1.5>01.5 > 0. Then she walks back down to 55 km by 10:00 to fetch a forgotten jacket, where v(4)=6v(4) = 6: she is BELOW, d(4)=−1<0d(4) = -1 < 0. Finally she climbs to the summit by 14:00, d(8)=12>0d(8) = 12 > 0. Three sign changes on [0,3][0, 3], [3,4][3, 4] and [4,8][4, 8], hence three meetings by the IVT, at t=83t = \frac{8}{3}, t=185t = \frac{18}{5} and t=5613t = \frac{56}{13}. With the fast Sunday descent of the figure this is much harder to arrange: she would have to run down faster than 44 km/h to fall below Sunday's position again.

e) (1) The function d=u−vd = u - v no longer means anything: uu is a position on one trail and vv a position on another, and the equality u(c)=v(c)u(c) = v(c) would say only that two different points are at the same distance from the base. She is at a common point only at the base or at the summit, and on Sunday she is at the summit only at 6:00, when on Saturday she is at the base. The conclusion fails. (2) Now v(t)=12v(t) = 12 for 0≤t≤40 \le t \le 4, then v(t)=12−4(t−4)v(t) = 12 - 4(t - 4) until she reaches the base at t=7t = 7, and v(t)=0v(t) = 0 afterwards. The end values are unchanged, d(0)=−12d(0) = -12 and d(8)=12d(8) = 12, and the IVT applies as before. With the Saturday walk of the figure: on [4,5][4, 5], u=8u = 8 and v=12−4(t−4)=28−4tv = 12 - 4(t - 4) = 28 - 4t, equal when t=5t = 5; that is v(5)=8=u(5)v(5) = 8 = u(5), at 11:00, 88 km up. The conclusion survives, and only the meeting moves.

Exercise 10: A final exam question: fixed points and the temperature on a ring

A number cc is a FIXED POINT of ff when f(c)=cf(c) = c: on a graph, the curve y=f(x)y = f(x) meets the diagonal y=xy = x. The figure shows a function continuous on [0,1][0, 1] whose values all lie in [0,1][0, 1], and the diagonal of the unit square.

In the second half, a thin circular ring is heated unevenly. The position on the ring is an angle θ\theta in radians, and the temperature T(θ)T(\theta) is continuous with T(θ+2π)=T(θ)T(\theta + 2\pi) = T(\theta). The points at angles θ\theta and θ+π\theta + \pi are diametrically opposite.

0.250.50.7510.250.50.751y = xy = f(x)x
  • a) Let ff be continuous on [0,1][0, 1] with 0≤f(x)≤10 \le f(x) \le 1 for every xx in [0,1][0, 1]. Prove that ff has a fixed point in [0,1][0, 1].
  • b) Show that each hypothesis is needed, with three counterexamples: a discontinuous ff; a continuous ff on [0,1][0, 1] whose values leave [0,1][0, 1]; a continuous ff from (0,1)(0, 1) into (0,1)(0, 1).
  • c) Use a) to prove that cos⁡x=x\cos x = x has a solution in [0,1][0, 1].
  • d) Prove that at every moment there are two diametrically opposite points of the ring at the same temperature.
  • e) For T(θ)=20+4cos⁡θ+2sin⁡2θT(\theta) = 20 + 4\cos\theta + 2\sin 2\theta (degrees Celsius), find exactly all the pairs of opposite points with equal temperatures, and their common temperature.

Type your answers, the page tells you right or wrong 0/2

e)
Show the solution

Answers

  • a) g(x)=f(x)−xg(x) = f(x) - x is continuous with g(0)≥0≥g(1)g(0) \ge 0 \ge g(1): a zero in [0,1][0, 1] (at an endpoint if g(0)=0g(0) = 0 or g(1)=0g(1) = 0, by the IVT otherwise).
  • b) f=1f = 1 on [0,12]\left[0, \frac{1}{2}\right], 00 on (12,1]\left(\frac{1}{2}, 1\right]; f(x)=x+12f(x) = x + \frac{1}{2}; f(x)=x2f(x) = x^2 on (0,1)(0, 1).
  • c) cos⁡\cos maps [0,1][0, 1] into [cos⁡1,1]⊂[0,1][\cos 1, 1] \subset [0, 1] since 0<1<π20 < 1 < \frac{\pi}{2}.
  • d) g(θ)=T(θ+π)−T(θ)g(\theta) = T(\theta + \pi) - T(\theta) has g(π)=−g(0)g(\pi) = -g(0), so gg vanishes on [0,π][0, \pi].
  • e) Only the pair θ=π2\theta = \frac{\pi}{2}, 3π2\frac{3\pi}{2}, both at 2020 degrees.

a) Let g(x)=f(x)−xg(x) = f(x) - x, continuous on [0,1][0, 1] as a difference of continuous functions. Since f(0)≥0f(0) \ge 0, g(0)=f(0)≥0g(0) = f(0) \ge 0; since f(1)≤1f(1) \le 1, g(1)=f(1)−1≤0g(1) = f(1) - 1 \le 0. If g(0)=0g(0) = 0, then c=0c = 0 is a fixed point; if g(1)=0g(1) = 0, then c=1c = 1 is. Otherwise g(0)>0>g(1)g(0) > 0 > g(1), and the IVT with N=0N = 0 gives cc in (0,1)(0, 1) with g(c)=0g(c) = 0, that is f(c)=cf(c) = c. The case split is not pedantry: the IVT needs NN STRICTLY between the end values, and a solution that ignores the equality cases uses the theorem outside its hypotheses. On the figure, the curve starts on or above the diagonal and ends on or below it, so it must meet it.

b) Discontinuous: f(x)=1f(x) = 1 for 0≤x≤120 \le x \le \frac{1}{2} and f(x)=0f(x) = 0 for 12<x≤1\frac{1}{2} < x \le 1 has values in [0,1][0, 1], but f(x)=1≠xf(x) = 1 \ne x on the first piece and f(x)=0≠xf(x) = 0 \ne x on the second: no fixed point, because ff jumps over the diagonal at 12\frac{1}{2}. Values outside: f(x)=x+12f(x) = x + \frac{1}{2} is continuous on [0,1][0, 1], and f(x)=xf(x) = x would mean 12=0\frac{1}{2} = 0; here g(1)=12>0g(1) = \frac{1}{2} > 0, the sign change is lost. Open interval: f(x)=x2f(x) = x^2 maps (0,1)(0, 1) into (0,1)(0, 1) and is continuous, but x2=xx^2 = x only for x=0x = 0 or x=1x = 1, both excluded; the fixed points escaped through the missing endpoints. Each hypothesis of a) is used exactly once in its proof, and each counterexample breaks exactly one.

c) The cosine is continuous. For 0≤x≤10 \le x \le 1 we have 0≤x≤1<π20 \le x \le 1 < \frac{\pi}{2}, and on [0,π2]\left[0, \frac{\pi}{2}\right] the cosine decreases from 11 to 00, so cos⁡1≤cos⁡x≤1\cos 1 \le \cos x \le 1 with cos⁡1>0\cos 1 > 0. Thus cos⁡\cos maps [0,1][0, 1] into [0,1][0, 1], and by a) it has a fixed point: some cc in [0,1][0, 1] with cos⁡c=c\cos c = c. Neither endpoint works (cos⁡0=1≠0\cos 0 = 1 \ne 0 and cos⁡1<1\cos 1 < 1), so cc is in fact in (0,1)(0, 1). The whole proof rests on the inequality 1<π21 < \frac{\pi}{2}, that is π>2\pi > 2: exact reasoning, no decimal of cos⁡1\cos 1 needed.

d) Let g(θ)=T(θ+π)−T(θ)g(\theta) = T(\theta + \pi) - T(\theta), the temperature difference between a point and the opposite point. gg is continuous, since TT is and θ↦θ+π\theta \mapsto \theta + \pi is. Then g(0)=T(π)−T(0)g(0) = T(\pi) - T(0) and g(π)=T(2π)−T(π)=T(0)−T(π)=−g(0)g(\pi) = T(2\pi) - T(\pi) = T(0) - T(\pi) = -g(0), by periodicity. If g(0)=0g(0) = 0, the points 00 and π\pi answer the question. Otherwise g(0)g(0) and g(π)g(\pi) are non-zero and of opposite signs, and the IVT on [0,π][0, \pi] gives cc in (0,π)(0, \pi) with g(c)=0g(c) = 0: the points cc and c+πc + \pi have the same temperature. Nothing is assumed about how the ring is heated; only continuity and the identity g(π)=−g(0)g(\pi) = -g(0) are used, the same comparison trick as for the hiker.

e) cos⁡(θ+π)=−cos⁡θ\cos(\theta + \pi) = -\cos\theta and sin⁡(2θ+2π)=sin⁡2θ\sin(2\theta + 2\pi) = \sin 2\theta, so T(θ+π)=20−4cos⁡θ+2sin⁡2θT(\theta + \pi) = 20 - 4\cos\theta + 2\sin 2\theta and g(θ)=−8cos⁡θg(\theta) = -8\cos\theta. Equal temperatures at opposite points means cos⁡θ=0\cos\theta = 0, that is θ=π2\theta = \frac{\pi}{2} or θ=3π2\theta = \frac{3\pi}{2} in [0,2π)[0, 2\pi), and these two angles are opposite each other: ONE pair of points. Common temperature: T(π2)=20+0+2sin⁡π=20T\left(\frac{\pi}{2}\right) = 20 + 0 + 2\sin\pi = 20 and T(3π2)=20+0+2sin⁡3π=20T\left(\frac{3\pi}{2}\right) = 20 + 0 + 2\sin 3\pi = 20 degrees. The solution figure shows T(θ)T(\theta) and T(θ+π)T(\theta + \pi) crossing exactly at these two angles. As d) predicted, g(0)=−8g(0) = -8 and g(π)=8g(\pi) = 8 have opposite signs, and here the computation also finds WHERE, which the IVT alone never does.

123456714161820222426T(θ)T(θ + π)θ = π/2θ = 3π/2θ

© Ahmed Squalli Houssaini. Exercise set published at www.letuteurscientifique.ca/en/exercices/math140-continuity. Free for personal and classroom use; republishing it elsewhere requires written permission (legal notice).

See also

Struggling with MATH 140?

I tutor first-year calculus at McGill and Concordia, in English or in French, in Montreal or online. Get in touch for a first session.

Site by Studio Squalli