MATH 203 Calculus I • Concordia University, Montreal

Corrected exercises: continuity (MATH 203)

This is the corrected exercise set for continuity in MATH 203, Differential and Integral Calculus I, at Concordia University, section 2.5 of Thomas' Calculus. The definition fits in one line, lim⁡x→cf(x)=f(c)\lim_{x\to c} f(x) = f(c), and yet this is where the course starts marking reasoning: three conditions checked in order, a type of discontinuity justified by the one-sided limits, a seam solved as an equation, a theorem quoted with its hypotheses. The scientific calculator is allowed, and it is used where it belongs: in a bisection table, after the theorem.

The thread running through the whole set: continuity is decided at a handful of suspect points, the excluded points and the seams, and at each of them the verdict is a LIMIT that only correct algebra can reach. That algebra is where MATH 203 marks are lost, not in the definition: a denominator factored halfway, a repeated factor that survives the cancellation, a complex fraction cleared with the wrong sign, an absolute value removed on one side only, a negative exponent read as a minus sign. Every correction names the algebraic gesture and what the slip costs.

The traps named in the solutions: reading a limit from an isolated dot, calling a seam a break, stopping a factorization by grouping halfway, cancelling one copy of (x−1)2(x - 1)^2 and calling the point removable, losing the sign of 3−xx−3\frac{3 - x}{x - 3}, forgetting the reversal in −2x≥−6-2x \ge -6, copying the bad point of the outer function of a composition, dropping the factor 22 in a conjugate, keeping one root of a quadratic condition, applying the IVT across a point where the function is undefined, running a bisection in degree mode, counting an endpoint twice, reversing a bracket through 1x\frac{1}{x}, and reading a failed hypothesis as a proof of the opposite.

10 corrected exercises • 100 points • 150 minutes

Revision sheet for this chapter → Every MATH 203 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 cc when (1) f(c)f(c) exists, (2) lim⁡x→cf(x)\lim_{x\to c} f(x) exists, (3) lim⁡x→cf(x)=f(c)\lim_{x\to c} f(x) = f(c). From the left or right: the same with a one-sided limit; at an endpoint of a closed interval, only the one-sided version applies.
  • • Removable: the limit exists. Jump: finite one-sided limits that differ. Infinite: a one-sided limit is ±∞\pm\infty. Only a removable discontinuity has a continuous extension, and its value is the limit.
  • • Polynomials, rational functions, roots, ∣x∣|x|, exe^x, ln⁡x\ln x and the trigonometric functions are continuous on their domains; so are sums, products, and quotients where the denominator is not 00.
  • • Composition: gg continuous at cc and ff continuous at g(c)g(c) give f∘gf \circ g continuous at cc; if lim⁡x→cg(x)=b\lim_{x\to c} g(x) = b and ff is continuous at bb, then lim⁡x→cf(g(x))=f(b)\lim_{x\to c} f(g(x)) = f(b).
  • • Piecewise function: continuous on each open piece, then one equation per seam, left limit == right limit =f(c)= f(c).
  • • IVT: ff continuous on [a,b][a, b] and y0y_0 between f(a)f(a) and f(b)f(b) give f(c)=y0f(c) = y_0 for some cc in [a,b][a, b]. Bisection: keep the half where the sign changes; after nn steps the interval has length b−a2n\frac{b - a}{2^n}.

Part A: the basics (/50)

Exercise 1: Reading continuity on a graph: three conditions at four points

Thomas 2.5 defines continuity at an interior point cc of the domain by one equation, lim⁡x→cf(x)=f(c)\lim_{x\to c} f(x) = f(c), which is checked as three conditions in this order: (1) f(c)f(c) exists; (2) lim⁡x→cf(x)\lim_{x\to c} f(x) exists, that is, both one-sided limits exist, are finite and are equal; (3) lim⁡x→cf(x)=f(c)\lim_{x\to c} f(x) = f(c). The TYPE of a discontinuity is read from the one-sided limits: removable when the limit exists, jump when the one-sided limits are finite and different, infinite when a one-sided limit is ±∞\pm\infty.

The figure shows the graph of a function ff defined on [−4,6][-4, 6]. A filled dot belongs to the graph, an open dot does not, and the dashed line x=4x = 4 is a vertical asymptote on the left side only.

-4-3-2-1123456-2-11234567x = 4y = f(x)x
  • a) At c=−2c = -2: read f(−2)f(-2), lim⁡x→−2−f(x)\lim_{x\to -2^-} f(x) and lim⁡x→−2+f(x)\lim_{x\to -2^+} f(x), then name the first of the three conditions that fails.
  • b) At c=1c = 1 and at c=3c = 3: read the value and the limit. Is ff continuous at 33, where the formula of the graph visibly changes?
  • c) At c=4c = 4: read the two one-sided limits and f(4)f(4). Is ff continuous from the left at 44? From the right?
  • d) Classify the discontinuities at −2-2, 11 and 44. Which one can be removed by defining or redefining a single value, and with which value?
  • e) Give the largest intervals on which ff is continuous, endpoints included whenever the one-sided continuity allows it.

Type your answers, the page tells you right or wrong 0/19

a)
b)
c)
d)
e)
First interval ,
Second interval ,
Third interval ,
Fourth interval ,
Show the solution

Answers

  • a) f(−2)=2f(-2) = 2, left limit 11, right limit 44: condition 2 fails.
  • b) f(1)f(1) undefined, lim⁡x→1f(x)=1\lim_{x\to 1} f(x) = 1: condition 1 fails. At 33: f(3)=3=lim⁡x→3f(x)f(3) = 3 = \lim_{x\to 3} f(x), continuous.
  • c) Left limit +∞+\infty, right limit 11, f(4)=1f(4) = 1: continuous from the right only.
  • d) Jump at −2-2, removable at 11 (define f(1)=1f(1) = 1), infinite at 44.
  • e) [−4,−2)[-4, -2), (−2,1)(-2, 1), (1,4)(1, 4), [4,6][4, 6].

a) The filled dot above x=−2x = -2 gives f(−2)=2f(-2) = 2. From the left, the line climbs to the open dot (−2,1)(-2, 1): lim⁡x→−2−f(x)=1\lim_{x\to -2^-} f(x) = 1. From the right, the V-shaped piece starts at the open dot (−2,4)(-2, 4): lim⁡x→−2+f(x)=4\lim_{x\to -2^+} f(x) = 4. Condition 1 holds, since the value exists; condition 2 fails, since the one-sided limits differ. Read the VALUE from the dots and the LIMITS from where the curve is heading: the filled dot sits between the two ends, on neither branch, and it says nothing about either limit.

b) At 11 there is only an open dot, at the bottom of the V, and no filled dot above or below it: f(1)f(1) is undefined and condition 1 fails. The curve reaches the height 11 from both sides, so lim⁡x→1f(x)=1\lim_{x\to 1} f(x) = 1 exists. At 33 the filled end of the V is (3,3)(3, 3), so f(3)=3f(3) = 3; from the right, the next piece starts at the same height, so both one-sided limits equal 33. The three conditions hold: ff IS continuous at 33. A change of formula, or a sharp turn of the graph, is not a break: continuity only asks that the pieces meet at the value of the function.

c) Approaching 44 from the left, the curve climbs along the asymptote without bound: lim⁡x→4−f(x)=+∞\lim_{x\to 4^-} f(x) = +\infty. From the right, the last segment starts at the filled dot (4,1)(4, 1): lim⁡x→4+f(x)=1=f(4)\lim_{x\to 4^+} f(x) = 1 = f(4). Continuity from the left would require lim⁡x→4−f(x)=f(4)\lim_{x\to 4^-} f(x) = f(4), impossible with an infinite limit: no. Continuity from the right requires lim⁡x→4+f(x)=f(4)\lim_{x\to 4^+} f(x) = f(4), that is 1=11 = 1: yes. An asymptote on ONE side does not prevent a value, nor continuity from the other side.

d) At −2-2 the one-sided limits are finite and different: a jump discontinuity. The isolated value f(−2)=2f(-2) = 2 does not change the type, and no value can repair it, since no number equals both 11 and 44. At 11 the limit exists and only the value is missing: removable, and defining f(1)=1f(1) = 1 makes the three conditions hold. At 44 a one-sided limit is infinite: infinite discontinuity, and no finite value repairs it. Only the removable discontinuity can be repaired, and only by the value of the limit.

e) Each discontinuity cuts the domain. At −4-4, ff is continuous from the right (f(−4)=−1f(-4) = -1 is where the line starts), so −4-4 is kept. At −2-2, ff is continuous from neither side (1≠21 \ne 2 and 4≠24 \ne 2), so −2-2 is excluded from both intervals around it. At 11 the function is undefined. At 44, ff is continuous from the right, so 44 opens the last interval, which ends at 66 with continuity from the left. Answer: [−4,−2)[-4, -2), (−2,1)(-2, 1), (1,4)(1, 4) and [4,6][4, 6]. The interval (1,4)(1, 4) contains the seam at 33, where b) showed continuity.

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: factor completely first

A rational function is continuous at every point of its domain, so the only candidates are the zeros of its denominator. At each candidate the verdict comes from the limit, and the limit comes from the algebra: factor the numerator and the denominator COMPLETELY, cancel the common factors (allowed, because a limit never looks at the point itself), then look at what is left.

The algebra is where the marks go. Factor by grouping, use a3−b3=(a−b)(a2+ab+b2)a^3 - b^3 = (a - b)(a^2 + ab + b^2), clear a complex fraction by multiplying top and bottom by the common denominator, and remove an absolute value by cases.

  • a) Find and classify the discontinuities of f(x)=x2+x−6x3+3x2−4x−12f(x) = \frac{x^2 + x - 6}{x^3 + 3x^2 - 4x - 12}.
  • b) Same question for g(x)=2x2−5x+3x3−1g(x) = \frac{2x^2 - 5x + 3}{x^3 - 1}.
  • c) Same question for h(x)=1x−13x−3h(x) = \frac{\frac{1}{x} - \frac{1}{3}}{x - 3}.
  • d) Same question for k(x)=∣x−1∣−2x2−2x−3k(x) = \frac{|x - 1| - 2}{x^2 - 2x - 3}. Is kk continuous at 11, where ∣x−1∣|x - 1| has its corner?
  • e) A student cancels x−2x - 2 in a) and writes: f(x)=1x+2f(x) = \frac{1}{x + 2}, which is continuous at 22, so ff is continuous at 22. 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 −3-3 (limit −1-1) and at 22 (limit 14\frac14); infinite at −2-2.
  • b) Only x=1x = 1, removable, limit −13-\frac{1}{3}.
  • c) Removable at 33 (limit −19-\frac{1}{9}), infinite at 00.
  • d) Removable at 33 and at −1-1, limit 14\frac14 at both; continuous at 11, k(1)=12k(1) = \frac12.
  • e) f(2)f(2) is undefined, so ff is NOT continuous at 22; its continuous extension is.

a) Group the denominator: x3+3x2−4x−12=x2(x+3)−4(x+3)=(x+3)(x2−4)=(x+3)(x−2)(x+2)x^3 + 3x^2 - 4x - 12 = x^2(x + 3) - 4(x + 3) = (x + 3)(x^2 - 4) = (x + 3)(x - 2)(x + 2). The numerator is (x+3)(x−2)(x + 3)(x - 2). The candidates are −3-3, 22 and −2-2, and for x≠−3,2x \ne -3, 2, f(x)=1x+2f(x) = \frac{1}{x + 2}. At −3-3: lim⁡x→−3f(x)=1−1=−1\lim_{x\to -3} f(x) = \frac{1}{-1} = -1, finite, f(−3)f(-3) undefined: removable. At 22: the limit is 14\frac{1}{4}: removable. At −2-2: the numerator tends to (1)(−4)=−4(1)(-4) = -4 while the denominator tends to 00, so ∣f(x)∣→∞|f(x)| \to \infty: infinite discontinuity. Stopping after x2(x+3)−4(x+3)x^2(x + 3) - 4(x + 3), which is still a SUM, finds no common factor at all and misses both removable points.

b) Factor the trinomial with its leading coefficient: 2x2−5x+3=(2x−3)(x−1)2x^2 - 5x + 3 = (2x - 3)(x - 1), checked by expanding, 2x2−2x−3x+32x^2 - 2x - 3x + 3. The denominator is a difference of cubes: x3−1=(x−1)(x2+x+1)x^3 - 1 = (x - 1)(x^2 + x + 1), and x2+x+1x^2 + x + 1 has discriminant 1−4=−3<01 - 4 = -3 < 0, so it never vanishes. The only candidate is 11, and for x≠1x \ne 1, g(x)=2x−3x2+x+1g(x) = \frac{2x - 3}{x^2 + x + 1}, continuous at 11: lim⁡x→1g(x)=−13=−13\lim_{x\to 1} g(x) = \frac{-1}{3} = -\frac{1}{3}. Removable discontinuity. Two classic slips: x3−1=(x−1)3x^3 - 1 = (x - 1)^3 (expand it: x3−3x2+3x−1x^3 - 3x^2 + 3x - 1), and a factorization (2x−1)(x−3)(2x - 1)(x - 3) that gives −7x-7x in the middle instead of −5x-5x.

c) hh is undefined at 00 (because of 1x\frac{1}{x}) and at 33. Clear the complex fraction by multiplying top and bottom by 3x3x: h(x)=3−x3x(x−3)h(x) = \frac{3 - x}{3x(x - 3)}. Now 3−x=−(x−3)3 - x = -(x - 3), so for x≠0,3x \ne 0, 3, h(x)=−13xh(x) = -\frac{1}{3x}. At 33: lim⁡x→3h(x)=−19\lim_{x\to 3} h(x) = -\frac{1}{9}, removable. At 00: −13x-\frac{1}{3x} is unbounded, infinite discontinuity. The sign is the whole point: writing 3−xx−3=1\frac{3 - x}{x - 3} = 1 gives +19+\frac{1}{9}, a wrong answer that looks perfectly reasonable. Check with the calculator: h(3.001)=1/3.001−1/30.001≈−0.1111h(3.001) = \frac{1/3.001 - 1/3}{0.001} \approx -0.1111.

d) Factor the denominator: x2−2x−3=(x−3)(x+1)x^2 - 2x - 3 = (x - 3)(x + 1), so the candidates are 33 and −1-1. Near 33, x−1>0x - 1 > 0, so ∣x−1∣=x−1|x - 1| = x - 1, the numerator is x−3x - 3, and k(x)=x−3(x−3)(x+1)=1x+1→14k(x) = \frac{x - 3}{(x - 3)(x + 1)} = \frac{1}{x + 1} \to \frac{1}{4}: removable. Near −1-1, x−1<0x - 1 < 0, so ∣x−1∣=1−x|x - 1| = 1 - x, the numerator is 1−x−2=−(x+1)1 - x - 2 = -(x + 1), and k(x)=−(x+1)(x−3)(x+1)=−1x−3→−1−4=14k(x) = \frac{-(x + 1)}{(x - 3)(x + 1)} = -\frac{1}{x - 3} \to -\frac{1}{-4} = \frac{1}{4}: removable as well. Writing ∣x−1∣=x−1|x - 1| = x - 1 everywhere gives 1x+1\frac{1}{x + 1} near −1-1 and a false infinite discontinuity. At 11, ∣x−1∣|x - 1| and x2−2x−3x^2 - 2x - 3 are continuous and the denominator equals −4≠0-4 \ne 0, so kk is continuous by the quotient rule, with k(1)=−2−4=12k(1) = \frac{-2}{-4} = \frac{1}{2}: the corner of ∣x−1∣|x - 1| is not a discontinuity.

e) The cancellation shows that ff and 1x+2\frac{1}{x + 2} agree for x≠−3,2x \ne -3, 2, which is exactly what a limit needs: lim⁡x→2f(x)=14\lim_{x\to 2} f(x) = \frac{1}{4}. It does not make f(2)f(2) exist: the original formula gives 00\frac{0}{0}, so condition 1 fails and ff is discontinuous at 22. Correct statement: ff has a removable discontinuity at 22, and its continuous extension, with the value 14\frac{1}{4} at 22 (and −1-1 at −3-3), is the function 1x+2\frac{1}{x + 2} on its own domain. The domain is decided on the ORIGINAL formula, before any cancellation.

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

The theorems of Thomas 2.5: sums, differences, products, constant multiples and quotients (where the denominator is not zero) of functions continuous at cc are continuous at cc; polynomials, rational functions, roots, ∣x∣|x|, exponentials, logarithms and the trigonometric functions are continuous on their domains; and if gg is continuous at cc and ff is continuous at g(c)g(c), then f∘gf \circ g is continuous at cc.

The limit form (Theorem 10 of the section): if lim⁡x→cg(x)=b\lim_{x\to c} g(x) = b and ff is continuous at bb, then lim⁡x→cf(g(x))=f(b)\lim_{x\to c} f(g(x)) = f(b). Angles are in radians.

  • a) On which intervals is F(x)=6−2xx2−x−2F(x) = \frac{\sqrt{6 - 2x}}{x^2 - x - 2} continuous? Say what happens at each endpoint.
  • b) On [0,2π][0, 2\pi], where is H(x)=cos⁡x2sin⁡x−1H(x) = \frac{\cos x}{2\sin x - 1} continuous?
  • c) Compute lim⁡x→0cos⁡(πx2+πx3x)\lim_{x\to 0} \cos\left(\frac{\pi x^2 + \pi x}{3x}\right) and lim⁡x→3ln⁡(x3−27x−3)\lim_{x\to 3} \ln\left(\frac{x^3 - 27}{x - 3}\right), naming the continuity used at each step.
  • d) Let f(u)=1u−1f(u) = \frac{1}{u - 1} and g(x)=x2g(x) = x^2. Where is f∘gf \circ g discontinuous? Where is g∘fg \circ f discontinuous?

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

a)
First interval ,
Second interval ,
Third interval ,
b)
c)
d)
Show the solution

Answers

  • a) On (−∞,−1)(-\infty, -1), (−1,2)(-1, 2) and (2,3](2, 3], from the left at 33.
  • b) On [0,π6)\left[0, \frac{\pi}{6}\right), (π6,5π6)\left(\frac{\pi}{6}, \frac{5\pi}{6}\right), (5π6,2π]\left(\frac{5\pi}{6}, 2\pi\right].
  • c) 12\frac{1}{2} and ln⁡27=3ln⁡3≈3.2958\ln 27 = 3\ln 3 \approx 3.2958
  • d) f∘gf \circ g at x=−1x = -1 and x=1x = 1; g∘fg \circ f at u=1u = 1 only.

a) The root needs 6−2x≥06 - 2x \ge 0. Solve it with care: −2x≥−6-2x \ge -6, and dividing by −2-2 REVERSES the inequality, x≤3x \le 3. The denominator factors as x2−x−2=(x−2)(x+1)x^2 - x - 2 = (x - 2)(x + 1), so x≠2x \ne 2 and x≠−1x \ne -1. On this domain, 6−2x6 - 2x is a polynomial and the square root is continuous on [0,∞)[0, \infty), so the numerator is continuous on (−∞,3](-\infty, 3] by composition; the denominator is a polynomial, non-zero away from −1-1 and 22; the quotient is continuous on (−∞,−1)(-\infty, -1), (−1,2)(-1, 2) and (2,3](2, 3]. At 33, FF is defined on the left only and F(3)=0F(3) = 0: the continuity there is from the left. At −1-1 and 22, FF is undefined. Writing x≥3x \ge 3 after forgetting the reversal puts the whole answer on the wrong side of 33.

b) The cosine and the sine are continuous everywhere, so HH is continuous wherever 2sin⁡x−1≠02\sin x - 1 \ne 0, that is sin⁡x≠12\sin x \ne \frac{1}{2}. On [0,2π][0, 2\pi] the unit circle gives sin⁡x=12\sin x = \frac{1}{2} at x=π6x = \frac{\pi}{6} (first quadrant) and x=π−π6=5π6x = \pi - \frac{\pi}{6} = \frac{5\pi}{6} (second quadrant), and nowhere else. So HH is continuous on [0,π6)\left[0, \frac{\pi}{6}\right), (π6,5π6)\left(\frac{\pi}{6}, \frac{5\pi}{6}\right) and (5π6,2π]\left(\frac{5\pi}{6}, 2\pi\right], one-sided at 00 and 2π2\pi. The second solution is the one that goes missing: sin⁡\sin is positive in two quadrants.

c) Inner function first: for x≠0x \ne 0, πx2+πx3x=πx(x+1)3x=π(x+1)3\frac{\pi x^2 + \pi x}{3x} = \frac{\pi x(x + 1)}{3x} = \frac{\pi(x + 1)}{3}, so its limit at 00 is π3\frac{\pi}{3}, although it is undefined at 00. The cosine is continuous at π3\frac{\pi}{3}, so by the limit form of the composition theorem the limit is cos⁡π3=12\cos\frac{\pi}{3} = \frac{1}{2}. Second limit: x3−27x−3=x2+3x+9\frac{x^3 - 27}{x - 3} = x^2 + 3x + 9 for x≠3x \ne 3 (difference of cubes), which tends to 9+9+9=279 + 9 + 9 = 27. The logarithm is continuous at 27>027 > 0, so the limit is ln⁡27=3ln⁡3≈3.2958\ln 27 = 3\ln 3 \approx 3.2958. The inner function only needs a LIMIT; the outer function needs CONTINUITY at that limit.

d) f∘g(x)=f(x2)=1x2−1f \circ g(x) = f(x^2) = \frac{1}{x^2 - 1}. The function ff fails at u=1u = 1, so f∘gf \circ g fails where g(x)=1g(x) = 1, that is x2=1x^2 = 1: at x=−1x = -1 AND x=1x = 1 (infinite discontinuities). The trap is to copy the bad point of ff, 11, as the bad point of f∘gf \circ g: the question is which xx are SENT to the bad point, and that is an equation to solve. In the other order, g∘f(u)=(1u−1)2=1(u−1)2g \circ f(u) = \left(\frac{1}{u - 1}\right)^2 = \frac{1}{(u - 1)^2} is discontinuous at u=1u = 1 only, where ff itself is undefined; gg is continuous everywhere and adds no bad point. Composition does not commute, and neither do its discontinuities.

Exercise 4: Gluing pieces: a conjugate, a complex fraction and a parameter that enters squared

A function defined by pieces is continuous on each OPEN interval where one continuous formula applies. Only the seams remain, and at a seam cc continuity means: left limit == right limit =f(c)= f(c), one equation per seam. The value f(c)f(c) comes from the piece that contains the equality sign.

A piece that gives 00\frac{0}{0} at its seam is simplified BEFORE the limit is taken: that simplification is the algebra of the question, and it is where the marks are.

  • a) Find kk such that f(x)=2x−1−3x−5f(x) = \frac{\sqrt{2x - 1} - 3}{x - 5} for 12≤x<5\frac{1}{2} \le x < 5 and f(x)=kx2−8f(x) = kx^2 - 8 for x≥5x \ge 5 is continuous on [12,∞)\left[\frac{1}{2}, \infty\right).
  • b) Find aa and bb such that gg is continuous on R\mathbb{R}, where g(x)=2x+2x+1g(x) = \frac{\frac{2}{x} + 2}{x + 1} if x<−1x < -1, g(x)=ax+bg(x) = ax + b if −1≤x≤2-1 \le x \le 2, and g(x)=x−2x+2−2g(x) = \frac{x - 2}{\sqrt{x + 2} - 2} if x>2x > 2.
  • c) Find every cc such that h(x)=x2−c2h(x) = x^2 - c^2 for x<3x < 3 and h(x)=cx−1h(x) = cx - 1 for x≥3x \ge 3 is continuous on R\mathbb{R}.
  • d) Find cc such that q(x)=x2−1x2−2x+1q(x) = \frac{x^2 - 1}{x^2 - 2x + 1} for x≠1x \ne 1 and q(1)=cq(1) = c is continuous at 11.

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

a)
b)
c)
d)
Show the solution

Answers

  • a) k=13k = \frac{1}{3} (both sides give 13\frac{1}{3})
  • b) a=2a = 2, b=0b = 0 (g(−1)=−2g(-1) = -2, g(2)=4g(2) = 4)
  • c) c=2c = 2 or c=−5c = -5
  • d) No value: q(x)=x+1x−1q(x) = \frac{x + 1}{x - 1} near 11, an infinite discontinuity.

a) For 12≤x<5\frac{1}{2} \le x < 5 the first piece is continuous (a root of a positive-or-zero polynomial over a non-zero polynomial), and the second is a polynomial. At the seam 55: f(5)=25k−8=lim⁡x→5+f(x)f(5) = 25k - 8 = \lim_{x\to 5^+} f(x). On the left, the form is 00\frac{0}{0}; multiply by the conjugate 2x−1+3\sqrt{2x - 1} + 3: the numerator becomes (2x−1)−9=2x−10=2(x−5)(2x - 1) - 9 = 2x - 10 = 2(x - 5), so f(x)=22x−1+3→23+3=13f(x) = \frac{2}{\sqrt{2x - 1} + 3} \to \frac{2}{3 + 3} = \frac{1}{3}. Condition: 25k−8=1325k - 8 = \frac{1}{3}, so 25k=25325k = \frac{25}{3} and k=13k = \frac{1}{3}. The factor 22 in 2x−102x - 10 is the mark: forgetting it gives 16\frac{1}{6} and k=49150k = \frac{49}{150}, a number that should alarm you.

b) The pieces are continuous on (−∞,−1)(-\infty, -1) (the only zeros of denominators, 00 and −1-1, are not in it), on (−1,2)(-1, 2) and on (2,∞)(2, \infty) (there x+2−2>0\sqrt{x + 2} - 2 > 0). Seam at −1-1: g(−1)=−a+bg(-1) = -a + b. The left piece is a complex fraction: multiply top and bottom by xx to get 2+2xx(x+1)=2(x+1)x(x+1)=2x\frac{2 + 2x}{x(x + 1)} = \frac{2(x + 1)}{x(x + 1)} = \frac{2}{x}, which tends to −2-2. Condition: −a+b=−2-a + b = -2. Seam at 22: g(2)=2a+bg(2) = 2a + b, and on the right multiply by the conjugate: (x−2)(x+2+2)(x+2)−4=x+2+2→4\frac{(x - 2)(\sqrt{x + 2} + 2)}{(x + 2) - 4} = \sqrt{x + 2} + 2 \to 4. Condition: 2a+b=42a + b = 4. Subtracting, 3a=63a = 6, so a=2a = 2 and b=0b = 0. Check: 2(−1)=−22(-1) = -2 and 2(2)=42(2) = 4; the figure shows three pieces forming one unbroken graph.

c) Both pieces are polynomials, so only 33 matters: lim⁡x→3−h(x)=9−c2\lim_{x\to 3^-} h(x) = 9 - c^2 and h(3)=lim⁡x→3+h(x)=3c−1h(3) = \lim_{x\to 3^+} h(x) = 3c - 1. Condition: 9−c2=3c−19 - c^2 = 3c - 1. Move everything to ONE side, keeping the signs: c2+3c−10=0c^2 + 3c - 10 = 0, that is (c+5)(c−2)=0(c + 5)(c - 2) = 0. So c=2c = 2 or c=−5c = -5. Check both: with c=2c = 2, 9−4=59 - 4 = 5 and 6−1=56 - 1 = 5; with c=−5c = -5, 9−25=−169 - 25 = -16 and −15−1=−16-15 - 1 = -16. The negative value looks strange and is perfectly correct; a student who writes c2−3c−10=0c^2 - 3c - 10 = 0 (a sign lost in the transfer) finds 55 and −2-2, and the check exposes it at once: 9−25≠15−19 - 25 \ne 15 - 1.

d) Factor both: x2−1=(x−1)(x+1)x^2 - 1 = (x - 1)(x + 1) and x2−2x+1=(x−1)2x^2 - 2x + 1 = (x - 1)^2. One factor x−1x - 1 cancels and ONE REMAINS: q(x)=x+1x−1q(x) = \frac{x + 1}{x - 1} for x≠1x \ne 1. As x→1x \to 1 the numerator tends to 22 and the denominator to 00, so q(x)→−∞q(x) \to -\infty from the left and +∞+\infty from the right. The limit does not exist as a finite number, condition 2 fails whatever cc is: there is NO value of cc. Seeing a common factor and concluding removable, then computing cc from the leftover formula, is the trap: count the multiplicity of the factor before cancelling.

-4-3-2-112345-3-2-11234562/x2x√(x + 2) + 2(-1, -2)(2, 4)x

Exercise 5: The Intermediate Value Theorem, then a bisection table in radians

Intermediate Value Theorem (Thomas 2.5). If ff is continuous on the CLOSED interval [a,b][a, b] and y0y_0 is any value between f(a)f(a) and f(b)f(b), then y0=f(c)y_0 = f(c) for some cc in [a,b][a, b]. For a root: a function continuous on [a,b][a, b] whose end values have opposite signs vanishes somewhere in (a,b)(a, b).

The theorem gives existence; the scientific calculator then LOCATES the root by bisection: evaluate at the midpoint, keep the half where the sign changes, repeat. Every value is computed in RADIAN mode and written with four decimals. The figure shows the curves y=cos⁡xy = \cos x and y=x2y = x^2.

0.250.50.7511.251.50.511.522.5y = cos xy = x²x (radians)
  • a) Prove that x3−4x2+2=0x^3 - 4x^2 + 2 = 0 has three real solutions, locating each between two consecutive integers.
  • b) Prove that cos⁡x=x2\cos x = x^2 has a solution in (0,1)(0, 1). Name the function, the interval and each hypothesis.
  • c) Four steps of bisection on [0,1][0, 1], in a table: give the final interval of length 116\frac{1}{16}. What goes wrong if the calculator is in degree mode?
  • d) A student writes: w(x)=x+1xw(x) = x + \frac{1}{x} has w(−1)=−2<0w(-1) = -2 < 0 and w(2)=52>0w(2) = \frac{5}{2} > 0, so ww has a zero in (−1,2)(-1, 2). Is this correct?
  • e) Prove that xx=3x^x = 3 for some xx in (1,2)(1, 2), then do one step of bisection.

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

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

Answers

  • a) p(−1)=−3p(-1) = -3, p(0)=2p(0) = 2, p(1)=−1p(1) = -1, p(3)=−7p(3) = -7, p(4)=2p(4) = 2: roots in (−1,0)(-1, 0), (0,1)(0, 1), (3,4)(3, 4).
  • b) g(x)=cos⁡x−x2g(x) = \cos x - x^2: g(0)=1>0g(0) = 1 > 0, g(1)=cos⁡1−1<0g(1) = \cos 1 - 1 < 0.
  • c) [0.8125,0.875][0.8125, 0.875]; in degrees g(0.875)g(0.875) comes out positive and the wrong half is kept.
  • d) No: ww is undefined at 00, and x+1x=x2+1xx + \frac{1}{x} = \frac{x^2 + 1}{x} is never 00.
  • e) 11=1<3<4=221^1 = 1 < 3 < 4 = 2^2; 1.51.5≈1.8371<31.5^{1.5} \approx 1.8371 < 3, so a solution in (1.5,2)(1.5, 2).

a) p(x)=x3−4x2+2p(x) = x^3 - 4x^2 + 2 is a polynomial, continuous on every closed interval. Values: p(−1)=−1−4+2=−3p(-1) = -1 - 4 + 2 = -3, p(0)=2p(0) = 2, p(1)=1−4+2=−1p(1) = 1 - 4 + 2 = -1, p(2)=8−16+2=−6p(2) = 8 - 16 + 2 = -6, p(3)=27−36+2=−7p(3) = 27 - 36 + 2 = -7, p(4)=64−64+2=2p(4) = 64 - 64 + 2 = 2. The sign changes on [−1,0][-1, 0], [0,1][0, 1] and [3,4][3, 4], so by the IVT there is a root in each of (−1,0)(-1, 0), (0,1)(0, 1) and (3,4)(3, 4): three distinct roots. A cubic has at most three real roots (an algebra fact, not an IVT fact), so these are all of them. Stopping at x=2x = 2 or 33, where the values stay negative, hides the third root: keep tabulating until the sign returns. Watch −4x2-4x^2 at x=−1x = -1: −4(−1)2=−4-4(-1)^2 = -4, not +4+4.

b) The theorem is about ONE function, so move everything to one side: g(x)=cos⁡x−x2g(x) = \cos x - x^2, the difference of the cosine and a polynomial, continuous on R\mathbb{R} and in particular on [0,1][0, 1]. g(0)=1−0=1>0g(0) = 1 - 0 = 1 > 0. g(1)=cos⁡1−1<0g(1) = \cos 1 - 1 < 0, because cos⁡1<1\cos 1 < 1 (the cosine equals 11 only at multiples of 2π2\pi); the calculator in radians gives g(1)≈−0.4597g(1) \approx -0.4597. By the IVT with y0=0y_0 = 0, there is cc in (0,1)(0, 1) with cos⁡c=c2\cos c = c^2. The figure shows where the curves cross; the figure suggests, the theorem proves.

c) Table in radians. Midpoint 0.50.5: g(0.5)=cos⁡0.5−0.25≈0.6276>0g(0.5) = \cos 0.5 - 0.25 \approx 0.6276 > 0, keep [0.5,1][0.5, 1]. Midpoint 0.750.75: g(0.75)≈0.7317−0.5625=0.1692>0g(0.75) \approx 0.7317 - 0.5625 = 0.1692 > 0, keep [0.75,1][0.75, 1]. Midpoint 0.8750.875: g(0.875)≈0.6410−0.7656=−0.1246<0g(0.875) \approx 0.6410 - 0.7656 = -0.1246 < 0, keep [0.75,0.875][0.75, 0.875]. Midpoint 0.81250.8125: g(0.8125)≈0.6877−0.6602=0.0275>0g(0.8125) \approx 0.6877 - 0.6602 = 0.0275 > 0, keep [0.8125,0.875][0.8125, 0.875], of length 116\frac{1}{16}. Each line is one more application of the IVT on a smaller closed interval. In degree mode the calculator computes cos⁡(0.875∘)≈0.9999\cos(0.875^\circ) \approx 0.9999, so g(0.875)g(0.875) comes out ≈0.2343>0\approx 0.2343 > 0: the student keeps [0.875,1][0.875, 1], which contains no root at all. A value of cos⁡\cos that stays near 11 for every input is the signature of degree mode.

d) Not correct. The IVT needs ww continuous on the whole closed interval [−1,2][-1, 2], and ww is undefined at 00, which lies inside. The conclusion is also false: put ww over one denominator, x+1x=x2+1xx + \frac{1}{x} = \frac{x^2 + 1}{x}, and x2+1≥1x^2 + 1 \ge 1 never vanishes, so ww has no zero anywhere. The sign changes by jumping from −∞-\infty to +∞+\infty at 00, not by crossing the axis. The single fraction is the algebraic gesture that settles the question in one line.

e) For x>0x > 0, xx=exln⁡xx^x = e^{x\ln x}: the product xln⁡xx\ln x is continuous on (0,∞)(0, \infty) and the exponential is continuous everywhere, so xxx^x is continuous on [1,2][1, 2] by composition. 11=11^1 = 1 and 22=42^2 = 4, and 33 lies between them, so by the IVT with y0=3y_0 = 3 there is cc in (1,2)(1, 2) with cc=3c^c = 3. Here y0=3y_0 = 3, not 00: the theorem works for any intermediate value, no need to rewrite. One bisection step: 1.51.5≈1.8371<31.5^{1.5} \approx 1.8371 < 3, so the value 33 is reached in (1.5,2)(1.5, 2). (Two more steps give 1.751.75≈2.66271.75^{1.75} \approx 2.6627 and 1.8751.875≈3.25001.875^{1.875} \approx 3.2500, so cc is in (1.75,1.875)(1.75, 1.875).)

Part B: problems and reasoning (/50)

Exercise 6: Continuous extension: the algebra that finds the only possible value

Thomas 2.5: if f(c)f(c) is undefined but lim⁡x→cf(x)=L\lim_{x\to c} f(x) = L exists and is finite, the function FF equal to f(x)f(x) for x≠cx \ne c and to LL at cc is continuous at cc. It is the continuous extension of ff to cc, and LL is the ONLY value that works. If the limit is infinite or does not exist, there is no continuous extension.

Each part below is decided by one algebraic gesture: factor, multiply by a conjugate, use sin⁡2x=1−cos⁡2x\sin^2 x = 1 - \cos^2 x, clear a complex fraction, rewrite a negative exponent.

  • a) Where is f(x)=x2+2x−15x2−9f(x) = \frac{x^2 + 2x - 15}{x^2 - 9} undefined? Give its continuous extension at each such point, or show that none exists.
  • b) Extend g(x)=3+x−3−xxg(x) = \frac{\sqrt{3 + x} - \sqrt{3 - x}}{x} continuously at x=0x = 0.
  • c) Find every point where h(x)=sin⁡2x1−cos⁡xh(x) = \frac{\sin^2 x}{1 - \cos x} is undefined, and the continuous extension of hh.
  • d) Find the constant aa such that m(x)=1x+a−14x−1m(x) = \frac{\frac{1}{x + a} - \frac{1}{4}}{x - 1} can be extended continuously at x=1x = 1, then give the value of the extension.
  • e) Extend n(x)=x−1−xx−1n(x) = \frac{x^{-1} - x}{x - 1} continuously wherever it is possible.

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

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

Answers

  • a) Undefined at ±3\pm 3; extension f(3)=43f(3) = \frac{4}{3}; none at −3-3 (infinite).
  • b) g(0)=13≈0.5774g(0) = \frac{1}{\sqrt 3} \approx 0.5774
  • c) Undefined at x=2kπx = 2k\pi; the extension is 1+cos⁡x1 + \cos x, with the value 22 at each 2kπ2k\pi.
  • d) a=3a = 3, extension value −116-\frac{1}{16}
  • e) n(1)=−2n(1) = -2; no extension at 00 (infinite).

a) x2−9=0x^2 - 9 = 0 at x=±3x = \pm 3. Factor: x2+2x−15=(x+5)(x−3)x^2 + 2x - 15 = (x + 5)(x - 3) and x2−9=(x−3)(x+3)x^2 - 9 = (x - 3)(x + 3), so f(x)=x+5x+3f(x) = \frac{x + 5}{x + 3} for x≠±3x \ne \pm 3. At 33: the limit is 86=43\frac{8}{6} = \frac{4}{3}, and F(3)=43F(3) = \frac{4}{3} extends ff continuously. At −3-3 the numerator is 9−6−15=−12≠09 - 6 - 15 = -12 \ne 0 while the denominator tends to 00: an infinite discontinuity, no extension. Note that 86\frac{8}{6} must be reduced; a calculator check, f(3.001)≈1.3333f(3.001) \approx 1.3333, confirms the value.

b) The domain is [−3,3][-3, 3] minus {0}\{0\}, and at 00 the form is 00\frac{0}{0}. Multiply by the conjugate of the DIFFERENCE of roots, 3+x+3−x\sqrt{3 + x} + \sqrt{3 - x}: the numerator becomes (3+x)−(3−x)=2x(3 + x) - (3 - x) = 2x. For x≠0x \ne 0, g(x)=23+x+3−xg(x) = \frac{2}{\sqrt{3 + x} + \sqrt{3 - x}}, continuous at 00, so lim⁡x→0g(x)=223=13≈0.5774\lim_{x\to 0} g(x) = \frac{2}{2\sqrt 3} = \frac{1}{\sqrt 3} \approx 0.5774, and g(0)=13g(0) = \frac{1}{\sqrt 3} is the extension. The sign in (3+x)−(3−x)(3 + x) - (3 - x) is where the marks go: 3+x−3−x=03 + x - 3 - x = 0 loses the parenthesis and the whole question.

c) hh is undefined where cos⁡x=1\cos x = 1, that is at x=2kπx = 2k\pi, kk an integer. Use sin⁡2x=1−cos⁡2x=(1−cos⁡x)(1+cos⁡x)\sin^2 x = 1 - \cos^2 x = (1 - \cos x)(1 + \cos x): for cos⁡x≠1\cos x \ne 1, h(x)=1+cos⁡xh(x) = 1 + \cos x. The cosine is continuous, so at every 2kπ2k\pi the limit is 1+1=21 + 1 = 2, and the continuous extension is simply 1+cos⁡x1 + \cos x on all of R\mathbb{R}. Every one of the infinitely many holes is removable. Without the identity the expression cannot be simplified: the trigonometry of chapter 1 is the algebra here, and splitting sin⁡2x1−cos⁡x\frac{\sin^2 x}{1 - \cos x} into two fractions is not an identity at all.

d) As x→1x \to 1 the denominator tends to 00. If the limit LL were finite, the numerator, equal to (x−1)m(x)(x - 1)m(x), would tend to 0⋅L=00 \cdot L = 0; since it is continuous at 11, its value at 11 must be 00: 11+a=14\frac{1}{1 + a} = \frac{1}{4}, so a=3a = 3. Then clear the complex fraction by multiplying top and bottom by 4(x+3)4(x + 3): m(x)=4−(x+3)4(x+3)(x−1)=1−x4(x+3)(x−1)=−14(x+3)m(x) = \frac{4 - (x + 3)}{4(x + 3)(x - 1)} = \frac{1 - x}{4(x + 3)(x - 1)} = -\frac{1}{4(x + 3)}, which tends to −116-\frac{1}{16}. The extension takes the value −116-\frac{1}{16} at 11. The parenthesis around x+3x + 3 matters: 4−x+34 - x + 3 gives 7−x7 - x, which does not vanish at 11, and the method falls apart.

e) Rewrite the negative exponent first: x−1=1xx^{-1} = \frac{1}{x}, so the numerator is 1x−x=1−x2x\frac{1}{x} - x = \frac{1 - x^2}{x}, and n(x)=1−x2x(x−1)=(1−x)(1+x)x(x−1)=−x+1xn(x) = \frac{1 - x^2}{x(x - 1)} = \frac{(1 - x)(1 + x)}{x(x - 1)} = -\frac{x + 1}{x} for x≠0,1x \ne 0, 1. At 11: the limit is −2-2, and n(1)=−2n(1) = -2 is the continuous extension. At 00: −x+1x-\frac{x + 1}{x} is unbounded, an infinite discontinuity, no extension. Reading x−1x^{-1} as −x-x is the error the tutorials of the course target: x−1x^{-1} is a reciprocal, never a negative.

Exercise 7: What a table of values proves about a continuous function

A function ff is continuous on [0,5][0, 5], and all that is known about it is the table of the figure. The IVT turns each pair of neighbouring values into a guarantee, and nothing more: between two values of the table, ff may wander as it likes, as long as it does not break.

A guarantee needs continuity on the closed interval between the two points, and a value y0y_0 strictly between the two values of ff (or equal to one of them, which is then reached at the endpoint).

xf(x)031-122324-451
  • a) What is the least number of zeros of ff in [0,5][0, 5] that the table guarantees? Locate each between two integers.
  • b) What is the least number of solutions of f(x)=2f(x) = 2 that the table guarantees?
  • c) Same question for f(x)=−2f(x) = -2.
  • d) Must ff have a zero in (2,3)(2, 3)? Can it? Can ff have exactly four zeros in [0,5][0, 5]?
  • e) Suppose now that ff is continuous on [0,5][0, 5] EXCEPT at x=4x = 4, with the same table. How many zeros are still guaranteed?

Type your answers, the page tells you right or wrong 0/7

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

Answers

  • a) At least 44: in (0,1)(0, 1), (1,2)(1, 2), (3,4)(3, 4) and (4,5)(4, 5).
  • b) At least 33: x=2x = 2, x=3x = 3, and one in (0,1)(0, 1).
  • c) At least 22: in (3,4)(3, 4) and in (4,5)(4, 5).
  • d) Not necessarily; it can; yes, exactly four is possible.
  • e) Only 22: in (0,1)(0, 1) and (1,2)(1, 2).

a) The signs of the table are +,−,+,+,−,++, -, +, +, -, +. There are four changes of sign between neighbours: on [0,1][0, 1] (33 to −1-1), on [1,2][1, 2] (−1-1 to 22), on [3,4][3, 4] (22 to −4-4) and on [4,5][4, 5] (−4-4 to 11). On each of these closed intervals ff is continuous, so the IVT gives a zero inside each OPEN interval. The four open intervals are disjoint, so the four zeros are distinct: at least four zeros. On [2,3][2, 3] the values 22 and 22 have the same sign, and nothing is guaranteed there.

b) Two solutions are read directly: f(2)=2f(2) = 2 and f(3)=2f(3) = 2. On [0,1][0, 1], 22 lies strictly between f(0)=3f(0) = 3 and f(1)=−1f(1) = -1, so the IVT gives a third solution in (0,1)(0, 1). On [1,2][1, 2] the value 22 is only reached at the endpoint x=2x = 2, already counted; on [3,4][3, 4] it is only reached at x=3x = 3; on [4,5][4, 5], 22 is not between −4-4 and 11. At least three solutions. Counting the endpoint values twice, once as a table entry and once as an IVT root, is the classic overcount.

c) The value −2-2 lies between f(3)=2f(3) = 2 and f(4)=−4f(4) = -4, and between f(4)=−4f(4) = -4 and f(5)=1f(5) = 1: one solution in (3,4)(3, 4) and one in (4,5)(4, 5). On [0,1][0, 1] the values are 33 and −1-1, and −2-2 is not between them; the same on [1,2][1, 2]. At least two solutions. The question for a value y0≠0y_0 \ne 0 is the same question for the function f−y0f - y_0: shift the table by 22 and count sign changes again.

d) Must: no. With f(2)=f(3)=2f(2) = f(3) = 2, a continuous ff can stay above 11 on the whole of [2,3][2, 3], for instance the constant 22 there. Can: yes, ff may dip below the axis between 22 and 33 and come back, which adds two zeros. Exactly four: yes, join the points of the table by straight segments; each segment that changes sign crosses the axis once, the segment on [2,3][2, 3] stays at height 22, and there are exactly four zeros. The IVT gives lower bounds, never an exact count.

e) The guarantees on [0,1][0, 1] and [1,2][1, 2] survive, since ff is still continuous on those closed intervals: two zeros. On [3,4][3, 4] and [4,5][4, 5] the IVT needs continuity at the endpoint 44 (from the left for [3,4][3, 4], from the right for [4,5][4, 5]), and a discontinuity at 44 means that at least one of these fails, without saying which. So nothing is guaranteed on either interval: ff could jump from negative to positive values without ever vanishing. At least two zeros. One bad point costs the guarantee on BOTH intervals that touch it.

Exercise 8: Five statements to correct

Each statement below was written by a student on a MATH 203 assignment, and each one is false. Say what is wrong, give a counterexample, and write a correct statement.

  • a) If ff is continuous at cc and gg is not, then fgfg is not continuous at cc.
  • b) A function that is continuous at every point of its domain has a graph that can be drawn without lifting the pencil.
  • c) If ff is discontinuous at cc, then f(c)f(c) is undefined.
  • d) If ff and gg are continuous everywhere, then fg\frac{f}{g} is continuous everywhere.
  • e) If f(0)<0<f(2)f(0) < 0 < f(2) and ff is discontinuous at x=1x = 1, then ff has no zero in (0,2)(0, 2).

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

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

Answers

  • a) False: xx times the sign function (11 for x≥0x \ge 0, −1-1 otherwise) is ∣x∣|x|, continuous at 00. True for f+gf + g.
  • b) False: 1x\frac{1}{x} is continuous on its domain, in two pieces. True on an INTERVAL.
  • c) False: x2x^2 for x≠1x \ne 1 with f(1)=5f(1) = 5 is defined at 11 and discontinuous there.
  • d) False: 1x2−1\frac{1}{x^2 - 1} fails at ±1\pm 1. True wherever g≠0g \ne 0.
  • e) False: x−12x - \frac12 for x<1x < 1, xx for x≥1x \ge 1, vanishes at 12\frac12. The IVT just gives nothing.

a) FALSE. Take c=0c = 0, f(x)=xf(x) = x (continuous) and g(x)=1g(x) = 1 for x≥0x \ge 0, g(x)=−1g(x) = -1 for x<0x < 0 (a jump at 00). Then (fg)(x)=x(fg)(x) = x for x≥0x \ge 0 and −x-x for x<0x < 0, that is ∣x∣|x|, continuous at 00. The factor f(0)=0f(0) = 0 crushes the jump. Correct statements: f+gf + g is not continuous at cc (otherwise g=(f+g)−fg = (f + g) - f would be continuous at cc), and fgfg is not continuous at cc when f(c)≠0f(c) \ne 0 (otherwise g=fgfg = \frac{fg}{f} would be continuous at cc).

b) FALSE. Thomas calls 1x\frac{1}{x} a continuous function: it is continuous at every point of its domain, which excludes 00. Its graph is two separate branches, and no pencil draws both without lifting. The pencil picture describes continuity on an INTERVAL. Correct statement: a function continuous on an interval has a graph in one unbroken piece over that interval. A continuous function can still have a domain with holes, and the graph breaks exactly at those holes.

c) FALSE. Let f(x)=x2f(x) = x^2 for x≠1x \ne 1 and f(1)=5f(1) = 5. Then f(1)f(1) is defined, lim⁡x→1f(x)=1\lim_{x\to 1} f(x) = 1 exists, and 1≠51 \ne 5: condition 3 fails and ff is discontinuous at 11 (removable). The isolated value of Exercise 1 at −2-2 is another example, with a jump. Correct statement: ff is discontinuous at cc when at least ONE of the three conditions fails; an undefined value is only the first way to fail.

d) FALSE. f(x)=1f(x) = 1 and g(x)=x2−1g(x) = x^2 - 1 are continuous everywhere, but fg=1x2−1=1(x−1)(x+1)\frac{f}{g} = \frac{1}{x^2 - 1} = \frac{1}{(x - 1)(x + 1)} is undefined at x=1x = 1 and x=−1x = -1, with infinite discontinuities there. Correct statement: fg\frac{f}{g} is continuous at every point cc where g(c)≠0g(c) \ne 0. The quotient rule has a hypothesis, and the hypothesis is solved as an equation, g(x)=0g(x) = 0, by factoring.

e) FALSE. Let f(x)=x−12f(x) = x - \frac{1}{2} for x<1x < 1 and f(x)=xf(x) = x for x≥1x \ge 1. Then f(0)=−12<0<2=f(2)f(0) = -\frac{1}{2} < 0 < 2 = f(2), ff jumps at 11 (left limit 12\frac{1}{2}, value 11), and yet f(12)=0f\left(\frac{1}{2}\right) = 0. When a hypothesis of the IVT fails, the theorem says NOTHING: it neither guarantees a zero nor forbids one. Correct statement: if ff is continuous on the whole closed interval [0,2][0, 2] and f(0)<0<f(2)f(0) < 0 < f(2), then ff has a zero in (0,2)(0, 2); without continuity, a zero may or may not exist.

Exercise 9: A room thermostat: a temperature that cannot jump

A room is heated one winter morning. Time tt is measured in hours after 5:00, for 0≤t≤100 \le t \le 10. The heater runs at full power until 11:00 (t=6t = 6), when the room reaches the setpoint of 2020 degrees Celsius, then switches to a holding power: the figure shows its power H(t)H(t) in kilowatts.

The temperature of the room, in degrees Celsius, is modelled by T(t)=15+atT(t) = 15 + at for 0≤t<20 \le t < 2, T(t)=21−bt−1T(t) = 21 - \frac{b}{t - 1} for 2≤t≤62 \le t \le 6, and T(t)=20T(t) = 20 for 6<t≤106 < t \le 10, where aa and bb are constants.

123456789100.511.522.5H(t)t (hours after 5:00)power (kW)
  • a) Classify the discontinuity of HH at t=6t = 6, and say from which side HH is continuous there. Why, on the other hand, must a model of the TEMPERATURE be continuous?
  • b) Find bb, then aa, so that TT is continuous on [0,10][0, 10].
  • c) With these values, prove that the room is at exactly 1818 degrees at some moment between 6:00 and 8:00, naming each hypothesis. Then find that moment exactly, as a clock time.
  • d) A colleague keeps a=12a = \frac{1}{2} but takes b=6b = 6. Where is his model discontinuous, and how large is each jump? Does the IVT still guarantee 1717 degrees between 6:00 and 8:00? Does his model reach 1717 degrees anyway?
  • e) The thermostat displays D(t)=⌊T(t)⌋D(t) = \lfloor T(t) \rfloor, the temperature rounded DOWN to a whole degree. With the model of b), at which moments is DD discontinuous? Can the display ever show 18.518.5?

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

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

Answers

  • a) A jump (22 to 0.50.5 kW), continuous from the right. A temperature varies without jumps: heat takes time to flow.
  • b) b=5b = 5, a=12a = \frac{1}{2}
  • c) T(1)=15.5<18<18.5=T(3)T(1) = 15.5 < 18 < 18.5 = T(3); exactly at t=83t = \frac{8}{3}, that is 7:40.
  • d) Jumps at t=2t = 2 (down by 11 degree) and t=6t = 6 (by 0.20.2 degree). No guarantee; it reaches 1717 anyway at 7:30.
  • e) At 7:00, 7:15, 7:40, 8:30 and 11:00. Never 18.518.5: DD takes whole values and is not continuous.

a) At t=6t = 6 the power drops from 22 kW (left limit) to 0.50.5 kW, and H(6)=0.5H(6) = 0.5 (filled dot): the one-sided limits are finite and different, a jump discontinuity, and since lim⁡t→6+H(t)=0.5=H(6)\lim_{t\to 6^+} H(t) = 0.5 = H(6), HH is continuous from the right. A power can be switched instantly. A temperature cannot: changing the temperature of the air and the walls takes energy, and a finite power delivers energy gradually, so TT passes through every intermediate temperature. A discontinuous temperature model is a model with an error in it, and this is what makes the IVT usable on it.

b) Each piece is continuous on its open interval (on (2,6)(2, 6), t−1≠0t - 1 \ne 0). Seam at 66: T(6)=21−b5T(6) = 21 - \frac{b}{5} must equal lim⁡t→6+T(t)=20\lim_{t\to 6^+} T(t) = 20, so b5=1\frac{b}{5} = 1 and b=5b = 5. Seam at 22: T(2)=21−51=16T(2) = 21 - \frac{5}{1} = 16 (the equality sign is in the middle piece), and lim⁡t→2−T(t)=15+2a\lim_{t\to 2^-} T(t) = 15 + 2a. Condition: 15+2a=1615 + 2a = 16, so a=12a = \frac{1}{2}. The order matters: bb must be found first, because the value T(2)T(2) depends on it. With these values the heater warms the room by half a degree per hour, then more slowly as it approaches 2020 degrees.

c) Work on [1,3][1, 3], from 6:00 to 8:00. TT is continuous on [1,3][1, 3]: each piece is, and b) made the seam at 22 continuous; without b) the IVT could not be used across t=2t = 2. T(1)=15+12=15.5T(1) = 15 + \frac{1}{2} = 15.5 and T(3)=21−52=18.5T(3) = 21 - \frac{5}{2} = 18.5, and 1818 lies between them, so by the IVT T(c)=18T(c) = 18 for some cc in (1,3)(1, 3). Exact moment: on [0,2)[0, 2), 15+t2=1815 + \frac{t}{2} = 18 gives t=6t = 6, outside the piece; on [2,6][2, 6], 21−5t−1=1821 - \frac{5}{t - 1} = 18 gives 5t−1=3\frac{5}{t - 1} = 3, t−1=53t - 1 = \frac{5}{3}, t=83t = \frac{8}{3} hours, that is 22 h 4040 min after 5:00: 7:40. The solution figure shows the crossing. A solution of the equation must be checked against the interval of its piece: t=6t = 6 is a solution of the wrong formula.

d) With b=6b = 6: lim⁡t→2−T(t)=16\lim_{t\to 2^-} T(t) = 16 but T(2)=21−6=15T(2) = 21 - 6 = 15, a jump down by 11 degree at 7:00; and T(6)=21−65=19.8T(6) = 21 - \frac{6}{5} = 19.8 while lim⁡t→6+T(t)=20\lim_{t\to 6^+} T(t) = 20, a jump of 0.20.2 degree at 11:00. Such a model makes the room lose a full degree in an instant. On [1,3][1, 3], T(1)=15.5T(1) = 15.5 and T(3)=21−3=18T(3) = 21 - 3 = 18, but TT is not continuous on [1,3][1, 3], so the IVT guarantees nothing. It happens anyway: 21−6t−1=1721 - \frac{6}{t - 1} = 17 gives t−1=32t - 1 = \frac{3}{2}, t=2.5t = 2.5, at 7:30, inside the middle piece. As in Exercise 8 e), a failed hypothesis gives no conclusion, in either direction.

e) The floor function jumps at each integer, so DD jumps exactly when TT reaches a whole degree coming from below. TT rises from 1515 to 2020 on [0,6][0, 6], then stays at 2020. Solve T(t)=nT(t) = n: n=16n = 16 at t=2t = 2 (end of the first piece); 21−5t−1=1721 - \frac{5}{t - 1} = 17 gives t=94t = \frac{9}{4}; =18= 18 gives t=83t = \frac{8}{3}; =19= 19 gives t=72t = \frac{7}{2}; =20= 20 gives t=6t = 6. So DD is discontinuous at 7:00, 7:15, 7:40, 8:30 and 11:00, five jumps of one degree, each continuous from the right (at 11:00, D=19D = 19 just before and D(6)=20D(6) = 20). DD never shows 18.518.5: it takes whole values only, and the IVT does not apply to it because it is not continuous. The temperature passes through 18.518.5 (at 8:00); the display does not.

123456789101415161718192021T(t)T = 18t = 8/3t (hours after 5:00)°C

Exercise 10: A final exam question: transcendental equations from physics

Planck's radiation law puts the peak of the light emitted by a hot body at the wavelength λmax=hcxkT\lambda_{max} = \frac{hc}{xkT}, where TT is the temperature in kelvins, hck=0.014388\frac{hc}{k} = 0.014388 m K, and x>0x > 0 solves the transcendental equation 5(1−e−x)=x5(1 - e^{-x}) = x. No algebra solves it exactly: its root is located by the IVT and a bisection table. The figure shows both sides of the equation.

In the second half, the bright fringes of a single-slit diffraction pattern, beyond the central one, lie very close to the positive solutions of tan⁡β=β\tan\beta = \beta (β\beta in radians). Calculator in radian mode throughout, four decimals.

123456123456y = 5(1 - e^(-x))y = xx
  • a) Check that x=0x = 0 solves the equation and say why it is excluded. Then prove that the equation has a solution in (4,5)(4, 5) with g(x)=5(1−e−x)−xg(x) = 5(1 - e^{-x}) - x.
  • b) Three steps of bisection on [4,5][4, 5], in a table: give an interval of length 18\frac{1}{8} that contains the root.
  • c) Deduce an interval for λmax\lambda_{max} of the Sun, T=5800T = 5800 K, in nanometres (11 nm =10−9= 10^{-9} m), to the nearest 0.10.1 nm.
  • d) A student writes: h(β)=tan⁡β−βh(\beta) = \tan\beta - \beta has h(1)≈0.5574>0h(1) \approx 0.5574 > 0 and h(2)≈−4.1850<0h(2) \approx -4.1850 < 0, so there is a bright fringe at some β\beta in (1,2)(1, 2). Is the argument valid? Is there a solution of tan⁡β=β\tan\beta = \beta in (1,2)(1, 2)?
  • e) Prove that tan⁡β=β\tan\beta = \beta has a solution in (4.4,4.5)(4.4, 4.5), justifying the continuity carefully.

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

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

Answers

  • a) 5(1−1)=05(1 - 1) = 0; x=0x = 0 gives no finite wavelength. g(4)≈0.9084>0g(4) \approx 0.9084 > 0, g(5)≈−0.0337<0g(5) \approx -0.0337 < 0: a root in (4,5)(4, 5).
  • b) g(4.5)g(4.5), g(4.75)g(4.75), g(4.875)g(4.875) all positive: [4.875,5][4.875, 5].
  • c) 496.1496.1 nm <λmax<508.9< \lambda_{max} < 508.9 nm
  • d) Not valid: tan⁡\tan is undefined at π2\frac{\pi}{2}, inside [1,2][1, 2]. No solution in (1,2)(1, 2).
  • e) h(4.4)≈−1.3037<0<0.1373≈h(4.5)h(4.4) \approx -1.3037 < 0 < 0.1373 \approx h(4.5), and tan⁡\tan is continuous on (π2,3π2)⊃[4.4,4.5]\left(\frac{\pi}{2}, \frac{3\pi}{2}\right) \supset [4.4, 4.5].

a) At x=0x = 0: 5(1−e0)=5(1−1)=05(1 - e^0) = 5(1 - 1) = 0, so x=0x = 0 is a solution; it would give λmax=hc0⋅kT\lambda_{max} = \frac{hc}{0 \cdot kT}, which is undefined, and the physics needs x>0x > 0. The function g(x)=5(1−e−x)−xg(x) = 5(1 - e^{-x}) - x is built from an exponential and polynomials, so it is continuous on R\mathbb{R}, and in particular on [4,5][4, 5]. In radians or not (no angle here), g(4)=5−5e−4−4=1−5e−4≈0.9084>0g(4) = 5 - 5e^{-4} - 4 = 1 - 5e^{-4} \approx 0.9084 > 0 and g(5)=5−5e−5−5=−5e−5≈−0.0337<0g(5) = 5 - 5e^{-5} - 5 = -5e^{-5} \approx -0.0337 < 0; the sign of g(5)g(5) is even exact, since e−5>0e^{-5} > 0. By the IVT there is cc in (4,5)(4, 5) with g(c)=0g(c) = 0. Distributing the 55 correctly is the gesture: 5(1−e−x)=5−5e−x5(1 - e^{-x}) = 5 - 5e^{-x}, not 5−e−x5 - e^{-x}.

b) Table. Midpoint 4.54.5: g(4.5)=0.5−5e−4.5≈0.4445>0g(4.5) = 0.5 - 5e^{-4.5} \approx 0.4445 > 0, keep [4.5,5][4.5, 5]. Midpoint 4.754.75: g(4.75)=0.25−5e−4.75≈0.2067>0g(4.75) = 0.25 - 5e^{-4.75} \approx 0.2067 > 0, keep [4.75,5][4.75, 5]. Midpoint 4.8754.875: g(4.875)=0.125−5e−4.875≈0.0868>0g(4.875) = 0.125 - 5e^{-4.875} \approx 0.0868 > 0, keep [4.875,5][4.875, 5], of length 18\frac{1}{8}, which contains the root by the IVT. Three positive values in a row are not a failure of the method: the root sits close to 55 (it is 4.96514.9651 to four decimals), and each step still halves the interval.

c) λmax=0.0143885800 x\lambda_{max} = \frac{0.014388}{5800\,x} m decreases when xx increases, so the bracket for xx is REVERSED for λ\lambda: x=5x = 5 gives the smallest wavelength, 0.0143885×5800≈4.961×10−7\frac{0.014388}{5 \times 5800} \approx 4.961 \times 10^{-7} m =496.1= 496.1 nm, and x=4.875x = 4.875 the largest, 0.0143884.875×5800≈5.089×10−7\frac{0.014388}{4.875 \times 5800} \approx 5.089 \times 10^{-7} m =508.9= 508.9 nm. So 496.1<λmax<508.9496.1 < \lambda_{max} < 508.9 nm: green light, which is why the peak of sunlight is quoted near 500500 nm. Copying the order of the xx bracket gives an interval upside down; a quantity of the form 1x\frac{1}{x} reverses inequalities between positive numbers.

d) Not valid. The IVT needs hh continuous on [1,2][1, 2], and tan⁡β=sin⁡βcos⁡β\tan\beta = \frac{\sin\beta}{\cos\beta} is undefined at β=π2≈1.5708\beta = \frac{\pi}{2} \approx 1.5708, which lies in [1,2][1, 2]. The sign changes by a jump through infinity. And there is in fact no solution in (1,2)(1, 2): on (1,π2)\left(1, \frac{\pi}{2}\right), the inequality β<tan⁡β\beta < \tan\beta for 0<β<π20 < \beta < \frac{\pi}{2} (the one Thomas 2.4 uses to find lim⁡sin⁡θθ\lim\frac{\sin\theta}{\theta}) gives h(β)>0h(\beta) > 0; on (π2,2]\left(\frac{\pi}{2}, 2\right], tan⁡β<0<β\tan\beta < 0 < \beta gives h(β)<0h(\beta) < 0. The solution figure shows it: the line y=βy = \beta passes between the two branches.

e) The tangent is continuous on each interval where cos⁡β≠0\cos\beta \ne 0, in particular on (π2,3π2)≈(1.5708,4.7124)\left(\frac{\pi}{2}, \frac{3\pi}{2}\right) \approx (1.5708, 4.7124), which contains [4.4,4.5][4.4, 4.5]. So h(β)=tan⁡β−βh(\beta) = \tan\beta - \beta is continuous on [4.4,4.5][4.4, 4.5], as a difference of continuous functions. In radians, h(4.4)≈3.0963−4.4=−1.3037<0h(4.4) \approx 3.0963 - 4.4 = -1.3037 < 0 and h(4.5)≈4.6373−4.5=0.1373>0h(4.5) \approx 4.6373 - 4.5 = 0.1373 > 0. By the IVT there is β\beta in (4.4,4.5)(4.4, 4.5) with tan⁡β=β\tan\beta = \beta: the first bright fringe beyond the central one (the root is 4.49344.4934). Writing the interval of continuity of tan⁡\tan, with the decimal value of 3π2\frac{3\pi}{2}, is the sentence that d) was missing.

0.511.522.533.544.55-6-4-22468y = tan xy = xπ/23π/2x

© Ahmed Squalli Houssaini. Exercise set published at www.letuteurscientifique.ca/en/exercices/math203-continuity. Free for personal and classroom use; republishing it elsewhere requires written permission (legal notice).

See also

Struggling with MATH 203?

I tutor first-year calculus at Concordia and McGill, in English or in French, in Montreal or online, with the algebra that the course takes for granted. Get in touch for a first session.

Site by Studio Squalli