MATH 140 practice final exam with full solutions (McGill)
This is a practice final examination for MATH 140, Calculus 1, the differential calculus course of the first year at McGill University. It covers the whole course and is weighted like a real cumulative final: about a quarter on limits and continuity, a quarter on the derivative and its rules, and half on the applications of the derivative, from related rates and linear approximation to the Mean Value Theorem, curve sketching, optimization and antiderivatives. Twelve questions, one hundred points, three parts.
Sit it as an exam: three hours on a timer, no calculator, no notes. Every answer is exact, −2π, e−1, 274, 1+90π, 753, and every mark goes to the method: name the algebra that opens a 00, write the form before each round of L'Hospital's rule, name the inner function of every chain, substitute a value only AFTER differentiating, check every hypothesis of a theorem on its interval, and justify that a maximum is absolute. Not one question repeats an exercise of the twenty chapter sets of this site: the gestures are those of every MATH 140 final, the functions, curves and situations are new, so the paper measures what you can do and not what you remember. Each question carries its Answers box for a first quick marking, and the full reasoning underneath.
The traps named in the solutions: writing x2=x at −∞, using θsinθ→1 with an angle that does not tend to 0, forgetting that a differentiable seam must first be continuous, applying L'Hospital's rule to a form that is no longer indeterminate, reading 1∞ as 1, losing the factor of an inner function, evaluating (f−1)′ at b instead of at f−1(b), giving a falling level a positive rate, linearizing with a degree instead of a radian, quoting a theorem without its hypotheses, forgetting the endpoint of a closed interval, and dropping the constant of an antiderivative.
Self-checking setType 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.
Exercise 1: Four limits, and the algebra that opens each one
Evaluate each limit exactly, or show that it is infinite. L'Hospital's rule is NOT allowed in this question: two points each, and the first line must name the move (conjugate, substitution, division by the dominant power, squeeze).
a) x→−1limx3+1x+5−2
b) x→1limx2−1sin(πx)
c) x→∞lim3−x16x2+5+2x and x→−∞lim3−x16x2+5+2x. Give the horizontal asymptotes of the graph.
d) x→∞limx+32x−cos(x2)
Show the solution
Answers
a)121
b)−2π
c)−6 at +∞ and 2 at −∞: asymptotes y=−6 and y=2
d)2: xcos(x2)→0 by the squeeze theorem
a) Direct substitution gives −1+14−2=00: the numerator and the denominator share the factor x+1, and the root hides it on top. Two moves. Multiply top and bottom by the conjugate x+5+2: the numerator becomes (x+5)−4=x+1. Factor the sum of cubes below: x3+1=(x+1)(x2−x+1). For x=−1 the quotient is (x+1)(x2−x+1)(x+5+2)x+1=(x2−x+1)(x+5+2)1, continuous at −1, and direct substitution gives 3⋅41=121.
b) Substitution gives 0sinπ=00. The limit θsinθ→1 needs an angle that tends to 0, and πx tends to π: writing πxsin(πx)→1 here is the classic error. Move: let t=x−1, so t→0. Then sin(πx)=sin(π+πt)=−sin(πt) and x2−1=(x−1)(x+1)=t(t+2). The quotient is −πtsin(πt)⋅t+2π→−1⋅2π=−2π.
c) Move: take x2 out of the root, remembering that x2=∣x∣: 16x2+5=∣x∣16+x25, then divide top and bottom by x. As x→∞, ∣x∣=x: the quotient is 3/x−116+5/x2+2→−14+2=−6. As x→−∞, ∣x∣=−x, so the numerator is −x16+x25+2x=x(2−16+x25), and the quotient is 3/x−12−16+5/x2→−12−4=2.
The graph has TWO horizontal asymptotes, y=−6 on the right and y=2 on the left. A student who writes x16x2+5=16+x25 for every x finds −6 at both ends: that step is false for x<0, where the left side is negative and the right side positive.
d) Move: divide top and bottom by x, the dominant power: 1+x32−xcos(x2). The quotient law needs the limit of xcos(x2), and cos(x2) itself has none: it oscillates between −1 and 1 forever. But it is bounded, and that is enough: for x>0, −x1≤xcos(x2)≤x1, and both bounds tend to 0, so by the squeeze theorem xcos(x2)→0. Now every piece has a limit and the quotient law applies: the limit is 1+02−0=2. Writing limcos(x2) as if it existed, or dropping the cosine without the squeeze, costs the mark.
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: The derivative as a limit, and a seam that must be smooth
Parts a) and b) to d) are independent. In b) to d), a and b are constants and f(x)=ax2+b for x<4, f(x)=2x+1 for x≥4. The one-sided derivatives at 4 must be computed from the DEFINITION of the derivative, as limits of difference quotients.
a) Recognize each limit as a derivative g′(a), naming g and a, then evaluate it: h→0limh327+h−3 and x→2limx−22x−4.
b) Find the condition on a and b for f to be continuous on R.
c) Find a and b such that f is differentiable at 4.
d) Take a=0 and b=3. Is f continuous at 4? Differentiable? Name the defect of the graph. What happens to the left difference quotient when 16a+b=3?
Show the solution
Answers
a)g(x)=3x, a=27: 271; g(x)=2x, a=2: 4ln2
b)16a+b=3
c)Right derivative 31, left derivative 8a: a=241, b=37
d)Continuous, not differentiable: a corner, slopes 0 and 31; if 16a+b=3 the left quotient is infinite
a) The first limit has the shape hg(a+h)−g(a) with g(x)=3x and a=27, since the subtracted constant 3 is exactly g(27). So it equals g′(27), and g′(x)=31x−2/3 gives 31⋅91=271. The second has the shape x−ag(x)−g(a) with g(x)=2x, a=2, and g(2)=4 is indeed the subtracted constant. So it equals g′(2)=22ln2=4ln2, since (2x)′=2xln2. Checking the subtracted constant is the whole test: limx→2x−22x−3 is not a derivative at all.
b) On (−∞,4), f is a polynomial; on (4,∞) it is 2x+1 with 2x+1>0: continuous on both open pieces. Only the seam remains. f(4)=9=3, the right-hand limit is 3, and the left-hand limit is limx→4−(ax2+b)=16a+b. So f is continuous on R if and only if 16a+b=3.
c) A differentiable function is continuous, so 16a+b=3 is required first. Right quotient, h>0: hf(4+h)−f(4)=h9+2h−3=h(9+2h+3)(9+2h)−9=9+2h+32→62=31 (conjugate). Left quotient, h<0: f(4+h)=a(16+8h+h2)+b=3+8ah+ah2 by the condition of b), so hf(4+h)−f(4)=h8ah+ah2=8a+ah→8a.
The two one-sided derivatives are equal if and only if 8a=31: a=241, and then b=3−2416=37. Check with the rules: the parabola 24x2+37 has slope 12x, which is 31 at 4, and (2x+1)′=2x+11 is 31 at 4: the two pieces share the point (4,3) AND the tangent there.
d) With a=0, b=3: 16a+b=3, so f is continuous at 4. The left quotient is 8a=0 (the left piece is the constant 3), the right one is 31: two finite, different one-sided derivatives, so f is not differentiable at 4 and the graph has a CORNER at (4,3). If 16a+b=3, the numerator of the left quotient tends to 16a+b−3=0 while h→0−: the quotient is infinite, and f has no derivative, as it must, since it is not even continuous. Continuity is necessary for differentiability, never sufficient.
Exercise 3: L'Hospital's rule: the form first, then the power forms
Evaluate each limit. Two points each: before EVERY application of L'Hospital's rule, write the form and check that it is indeterminate; for a product or a power, write the rewritten quotient and its form first.
a) x→0limx2ex−1−sinx
b) x→∞limx(2π−arctanx)
c) x→1limx1/(1−x)
d) x→0+lim(ex−1)x
Show the solution
Answers
a)00 twice, then 2e0+sin0=21
b)0⋅∞ rewritten as 00: limit 1
c)1±∞; lny=1−xlnx→−1: limit e−1
d)00; lny=xln(ex−1)→0: limit 1
a) Form: 01−1−0=00, indeterminate. Round 1, top and bottom differentiated separately: 2xex−cosx, whose form is 01−1=00 again. Round 2: 2ex+sinx, whose form is 21+0: NOT indeterminate, so stop and substitute: the limit is 21. A third round would differentiate the constant 2 into 0 and produce nonsense: the rule is applied only to an indeterminate form.
b) Form: ∞⋅0, since arctanx→2π. Rewrite as a quotient, putting the factor x in the denominator as x1: 1/x2π−arctanx, of the form 00. Differentiate: −x21−1+x21=1+x2x2→1. The limit is 1: the gap 2π−arctanx shrinks exactly like x1. The other rewriting, 1/(π/2−arctanx)x, leads to a heavier quotient.
c) As x→1 the base tends to 1 and the exponent 1−x1 to +∞ from the left and −∞ from the right: the form 1±∞ is indeterminate, and the answer is NOT 1. Near 1, x>0, so let y=x1/(1−x) and lny=1−xlnx, of the form 00. Differentiate: −11/x→−1, from both sides. Since the exponential is continuous, y=elny→e−1.
d) As x→0+, ex−1→0+ and the exponent x→0: the form 00. Let y=(ex−1)x, so lny=xln(ex−1), of the form 0⋅(−∞). Rewrite: 1/xln(ex−1), of the form ∞−∞. Differentiate, with the chain rule on the top: −1/x2ex/(ex−1)=−xex⋅ex−1x. Now ex−1x→1 (it is the reciprocal of x−0ex−e0→(ex)′(0)=1), and −xex→0: so lny→0 and y→e0=1. Simplifying after the round avoids a second, heavier round.
Part B: derivatives and their rules (/26)
Exercise 4: Three derivatives, every rule named
Three points each. Name every rule as you use it (product, quotient, chain with the inner function written down), then simplify: the factored or reduced form is part of the answer.
a) Differentiate f(x)=x4e−x/2, factor f′(x), and find the points of the graph where the tangent is horizontal.
b) Differentiate g(x)=cosx+xsinxsinx−xcosx and show that g′(x)=(cosx+xsinx)2x2. Find the tangent line at x=2π.
c) Differentiate h(x)=x+x, list its layers, and compute h′(1). Where is h defined, and where is it differentiable?
Show the solution
Answers
a)f′(x)=21x3e−x/2(8−x); horizontal tangents at (0,0) and (8,4096e−4)
b)g′(x)=(cosx+xsinx)2x2; tangent y=π2+x−2π
c)h′(x)=4xx+x2x+1, h′(1)=832; defined on [0,∞), differentiable on (0,∞)
a) Product rule with u=x4 and v=e−x/2. For v′, the chain rule with inner function w=−2x: v′=ew⋅w′=−21e−x/2. So f′(x)=4x3e−x/2−21x4e−x/2=21x3e−x/2(8−x). Horizontal tangents where f′(x)=0: the exponential is never 0, so x3=0 or 8−x=0, that is x=0 or x=8. The points are (0,0) and (8,84e−4)=(8,4096e−4). The common slip is v′=e−x/2, forgetting the factor −21 of the inner function: it moves the second point to x=−4.
b) Quotient rule with N=sinx−xcosx and D=cosx+xsinx. Each needs the product rule on its second term: N′=cosx−(cosx−xsinx)=xsinx and D′=−sinx+(sinx+xcosx)=xcosx. Then g′=D2N′D−ND′=D2xsinx(cosx+xsinx)−(sinx−xcosx)xcosx=D2x[sinxcosx+xsin2x−sinxcosx+xcos2x]=D2x2(sin2x+cos2x)=(cosx+xsinx)2x2, wherever D=0.
At x=2π: N=1−0=1 and D=0+2π=2π, so g(2π)=π2 and g′(2π)=π2/4π2/4=1. Tangent: y=π2+(x−2π). Expanding D2 in the denominator, or stopping before sin2x+cos2x=1, leaves an answer that cannot be evaluated at a glance: the simplification is what the question tests.
c) Layers, from the outside: the square root u, with u=x+x, and inside u the root x. Chain rule on the outer root, then the sum rule inside: h′(x)=2x+x1⋅(1+2x1)=4xx+x2x+1. At x=1: 4⋅1⋅23=423=832.
h is defined when x≥0 (for x), and then x+x≥0 holds automatically: domain [0,∞). The formula of h′ needs x>0; at 0 the difference quotient th(t)−h(0)=tt+t≥tt1/4→∞, a vertical tangent. So h is differentiable on (0,∞) only.
Exercise 5: A curve with no formula for y, and an inverse with no formula at all
Parts a) to c) concern the curve C of equation xey+y=2, shown in the figure with its tangent at P(2,0) (dashed) and a point V of part b). Part d) is independent.
a) Check that P(2,0) is on C, find dxdy by implicit differentiation, and write the tangent line at P.
b) Show that C has no horizontal tangent, and find the exact point V where its tangent is vertical.
c) Find dx2d2y at P. Is the curve above or below its tangent near P?
d) Let f(x)=x+arctanx. Explain why f is one-to-one, then find (f−1)′(0), the tangent line to y=f−1(x) at the origin, and (f−1)′(1+4π).
Show the solution
Answers
a)dxdy=−xey+1ey; slope −31 at P; tangent y=−31(x−2)
b)The numerator −ey never vanishes; V=(−e−3,3)
c)y′′=274>0 at P: the curve is above its tangent
d)(f−1)′(0)=21, tangent y=2x; (f−1)′(1+4π)=32
a) At (2,0): 2e0+0=2, so P is on C. Differentiate both sides with respect to x, y being a function of x. The term xey is a product, and ey is a composition with inner function y: dxd(xey)=ey+xeyy′. So ey+xeyy′+y′=0, hence y′(xey+1)=−ey and dxdy=−xey+1ey wherever xey+1=0. At P: −2+11=−31. Tangent: y=−31(x−2). The formula contains y: it gives the slope at a POINT, which is all that is needed.
b) A horizontal tangent needs the numerator −ey to vanish, which never happens: no horizontal tangent anywhere. A vertical tangent needs xey+1=0 at a point of C, with the numerator nonzero. On C, xey=2−y, so the condition reads 2−y+1=0: y=3, then x=(2−3)e−3=−e−3. There the numerator is −e3=0: V=(−e−3,3) is the only point with a vertical tangent. It lies a hair to the left of the y-axis, e−3 being about 0.05, which is why the figure barely shows the curve turning back. Check: C is the graph of x=(2−y)e−y, and dydx=−(3−y)e−y vanishes exactly at y=3.
c) Differentiate ey+xeyy′+y′=0 once more, remembering that y and y′ are both functions of x: eyy′+[eyy′+xey(y′)2+xeyy′′]+y′′=0, that is 2eyy′+xey(y′)2+(xey+1)y′′=0. At P, with y′=−31: −32+92+3y′′=0, so 3y′′=94 and y′′=274. It is positive: near P the curve is concave up and lies ABOVE its tangent, as in the figure.
d) x and arctanx are both increasing on R, so their sum is increasing, and an increasing function never takes the same value twice: f is one-to-one. No formula for f−1 exists in terms of elementary functions, and none is needed. f(0)=0, so f−1(0)=0; f′(x)=1+1+x21, the derivative of arctan being 1+x21, and f′(0)=2=0. Hence (f−1)′(0)=f′(f−1(0))1=21, and the tangent to y=f−1(x) at the origin is y=2x, the reflection in y=x of the tangent y=2x of f.
For b=1+4π, first find the point where f takes that value: f(1)=1+4π, so f−1(b)=1. Then (f−1)′(b)=f′(1)1=1+211=32. The classic error is f′(b)1: the derivative of f must be taken at f−1(b), not at b.
Exercise 6: Logarithmic differentiation: a moving exponent, then a heavy quotient
Parts a) and b) are independent. Four points each.
a) Let y=xx for x>0. Explain why neither the power rule nor the exponential rule applies, find y′, compute y′(4), and find the exact point where the tangent is horizontal.
b) Let y=(1+x)3e2x1+x2 for x>−1. Find yy′ by logarithmic differentiation, write the tangent line at x=0, and show that y′(1)=y(1).
Show the solution
Answers
a)y′=xx2xlnx+2; y′(4)=8(ln2+1); horizontal tangent at (e−2,e−2/e)
b)yy′=2+1+x2x−1+x3; tangent y=1−x; at x=1, yy′=1
a) The power rule (xn)′=nxn−1 needs a CONSTANT exponent, and the exponential rule (bx)′=bxlnb a CONSTANT base; here both move. Take logarithms, y>0: lny=xlnx. Differentiate both sides, the left by the chain rule, the right by the product rule: yy′=2x1lnx+x⋅x1=2xlnx+2. So y′=xx2xlnx+2.
At x=4: y(4)=42=16 and y′(4)=16⋅42ln2+2=8(ln2+1). Horizontal tangent: xx>0 and 2x>0, so y′=0 exactly when lnx=−2, x=e−2. Then x=e−1 and y=(e−2)e−1=e−2/e: the point is (e−2,e−2/e).
b) On (−1,∞) every factor is positive (e2x, the root, and 1+x), so y>0 and lny can be taken without absolute values. Expand BEFORE differentiating: lny=2x+21ln(1+x2)−3ln(1+x). Differentiate: yy′=2+21⋅1+x22x−1+x3=2+1+x2x−1+x3. The product and quotient rules would need three factors and a cube in the denominator; here each factor contributes one relative rate.
At x=0: y(0)=11⋅1=1 and yy′=2+0−3=−1, so y′(0)=−1 and the tangent is y=1−x. At x=1: yy′=2+21−23=1, so y′(1)=y(1)=8e22. Forgetting to multiply by y at the end, and answering 1 for y′(1), is the usual loss of marks.
Part C: applications of the derivative (/50)
Exercise 7: Filling a trough with a triangular cross-section
A water trough is 4 m long. Its two ends are isosceles triangles, vertex down, 80 cm across the top and 40 cm deep, as in the figure. Water is pumped in at the constant rate of 0.08 m3/min. Let h be the depth of the water and w the width of its surface, in metres, functions of the time t in minutes.
a) Use similar triangles to express w in terms of h, then the volume V of water in terms of h alone.
b) How fast is the water level rising when the water is 20 cm deep?
c) How fast is the width w of the water surface increasing at that instant? How fast does the level rise just as the trough becomes full, and why is it slower?
d) Later, a crack opens in the trough and water leaks out while the pump keeps running. At the instant the depth is 30 cm, the level is observed to FALL at 1 cm/min. At what rate is water leaking out?
Show the solution
Answers
a)w=2h, V=4h2 m3 for 0≤h≤0.4
b)dtdh=0.05 m/min, that is 5 cm/min
c)dtdw=0.1 m/min; at h=0.4, dtdh=0.025 m/min
d)0.104 m3/min, about 104 litres per minute
a) The water fills a triangle similar to the end of the trough: same apex at the bottom, same angles. Corresponding sides are proportional: hw=0.40.8, so w=2h at every instant. The cross-section of the water has area 21wh=21(2h)h=h2, and the volume is that area times the length: V=4h2 m3, for 0≤h≤0.4. Eliminating w BEFORE differentiating is what leaves one unknown rate.
b) Differentiate V=4h2 with respect to t, by the chain rule: dtdV=8hdtdh. With dtdV=0.08 and, only now, h=0.2: 0.08=1.6dtdh, so dtdh=0.05 m/min, 5 cm/min. Check without related rates: starting from empty, V=0.08t, so h=0.02t; the depth is 0.2 at t=2, and dtdh=0.02t0.01=0.20.01=0.05 there.
c) From w=2h: dtdw=2dtdh=0.1 m/min, 10 cm/min at that instant. Just as the trough becomes full, h=0.4: 0.08=8(0.4)dtdh=3.2dtdh, so dtdh=0.025 m/min, half the rate of b). The same 0.08 m3 per minute now spreads over a surface twice as wide (80 cm against 40 cm), so it raises the level half as much: dtdV is constant, dtdh is not.
d) Let L be the leak rate, in m3/min. The volume now changes at dtdV=0.08−L. At the instant h=0.3 with dtdh=−0.01 m/min (the level falls, so the rate is NEGATIVE, and 1 cm is 0.01 m): dtdV=8(0.3)(−0.01)=−0.024. Hence 0.08−L=−0.024 and L=0.104 m3/min, about 104 litres per minute. Using +0.01 gives L=0.056, a leak smaller than the pump, which could never make the level fall.
Exercise 8: Linear approximation: a tangent in degrees, and an artery that narrows
Parts a) and b), c) are independent. In b) and c), Poiseuille's law says that the flow of blood through an artery of radius r is F=kr4, where k is a positive constant.
a) Use a linear approximation of tanx to estimate tan46∘ exactly, then as a short decimal. Is the estimate too large or too small?
b) Plaque reduces the radius of an artery by 5%. Use differentials to estimate the percentage change of the flow. Compute the exact change, and explain from the concavity of F why the linear estimate exaggerates the drop.
c) By what percentage must the radius shrink to halve the flow? Give the linear estimate, then the exact answer, bracketed without a calculator using 1.184<2<1.194.
Show the solution
Answers
a)tan46∘≈1+90π≈1.035, too small
b)FdF=4rdr=−20%; exact 0.954=0.81450625, a drop of 18.55%
c)Linear: 12.5%; exact: 1−2−1/4, between 15.2% and 16.0%
a) Calculus works in radians: 46∘=4π+180π. Linearize f(x)=tanx at a=4π, where everything is exact: f(a)=1, f′(x)=sec2x, f′(4π)=2. So L(x)=1+2(x−4π) and tan46∘≈1+2⋅180π=1+90π≈1.035. Taking dx=1 because the angle grew by one degree gives 3, an absurd value: the increment must be in radians. Side: f′′(x)=2sec2xtanx>0 on (0,2π), so tan is concave up there and lies ABOVE its tangent line: the estimate is too small.
b) dF=4kr3dr, so the relative change is FdF=kr44kr3dr=4rdr: a power 4 multiplies relative changes by 4. With rdr=−0.05: FdF=−0.20, a drop of about 20%. Exact: F(r)F(0.95r)=0.954=0.90252=0.81450625, a drop of 18.549375%. The linear estimate exaggerates the drop because F′′=12kr2>0: F is concave up in r, its tangent lies BELOW the curve, so the tangent predicts a smaller flow than the true one.
c) Linear: 4rdr=−0.5 gives rdr=−0.125, a 12.5% reduction. Exact: (1−s)4=21, so s=1−2−1/4. From 1.184<2<1.194 (indeed 1.184=1.39242=1.93877… and 1.194=1.41612=2.00533…): 1.18<21/4<1.19, so 1−1.181<s<1−1.191, that is 1.180.18<s<1.190.19, between about 15.2% and 16.0%. The linear estimate is off by a quarter of itself: a differential is a tool for SMALL changes, and halving a flow is not one.
Exercise 9: Three theorems, their hypotheses checked one by one
Parts a), b) and c) are independent. For each theorem you use, name it and check EVERY hypothesis, on the right interval, before quoting the conclusion. The figure, for part b), shows y=x+2x, the chord AB over [0,2] and a tangent parallel to it at C.
a) Prove that the equation ex+x3=4 has exactly one real solution, and that it lies in (1,2).
b) Check the hypotheses of the Mean Value Theorem for f(x)=x+2x on [0,2], and find the number c exactly. Prove without a calculator that it lies in (0,2).
c) Apply the Mean Value Theorem to x on [49,50] to prove that 7+161<50<7+141. Deduce that 80113<2<7099.
Show the solution
Answers
a)IVT on [1,2]: f(1)=e−3<0<f(2)=e2+4; f′(x)=ex+3x2>0 and Rolle forbid a second root
b)Chord slope 41; c=22−2
c)50−7=2c1 with 7<c<8; then 50=52
a) Let f(x)=ex+x3−4. Existence, by the Intermediate Value Theorem: f is continuous on [1,2] (an exponential plus a polynomial), f(1)=e−3<0 because e<3, and f(2)=e2+4>0. So f(c)=0 for some c in (1,2). At most one root, by Rolle's theorem and contradiction: suppose f(r1)=f(r2)=0 with r1<r2. Then f is continuous on [r1,r2], differentiable on (r1,r2) and f(r1)=f(r2), so f′(c)=0 for some c between them. But f′(x)=ex+3x2>0 for EVERY x, since ex>0 and 3x2≥0. Contradiction: the equation has exactly one real solution, and it lies in (1,2).
b) f is a rational function whose denominator x+2≥2>0 on [0,2]: it is continuous on [0,2] and differentiable on (0,2), the two hypotheses. The chord has slope 2−0f(2)−f(0)=21/2−0=41. By the quotient rule, f′(x)=(x+2)2(x+2)−x=(x+2)22. Solve (c+2)22=41: (c+2)2=8, c+2=±22. The root c=−2−22 is outside (0,2) and rejected; c=22−2 remains. Since 1<2<2: 0<22−2<2. The figure's point C is at x≈0.83.
c) g(x)=x is continuous on [49,50] and differentiable on (49,50), where x>0. The Mean Value Theorem gives c in (49,50) with 50−49=g′(c)(50−49), that is 50−7=2c1. The number c stays unknown; only its interval is used: 49<c<50<64, so 7<c<8 and 161<2c1<141. Hence 7+161<50<7+141, both inequalities strict.
Now 50=52: dividing by 5, 80113<2<7099, that is 1.4125<2<1.41429. Check by squaring: (80113)2=640012769<2 since 12769<12800, and (7099)2=49009801>2 since 9801>9800. The upper bound is famous: 7099 misses 2 by less than 10−4.
Exercise 10: A complete study: x minus twice its arctangent
Let f(x)=x−2arctanx. The figure shows its graph and two dashed lines; every feature of it must be justified by a computation, as on the final. No calculator: π stays π.
a) Give the domain and the symmetry of f. Find limx→∞[f(x)−(x−π)] and limx→−∞[f(x)−(x+π)], name the two asymptotes, and say on which side of each the graph lies.
b) Find f′, the intervals of increase and decrease, and the local extreme values, exactly.
c) Find f′′, the intervals of concavity, the inflection point and the tangent line there.
d) Prove that f has exactly three zeros, 0 and ±r, with 2<r<π.
e) Give the range of f. For which real numbers k does the equation x−2arctanx=k have three solutions? Two? One?
Show the solution
Answers
a)R, odd; both limits 0: asymptotes y=x−π (graph above) and y=x+π (graph below)
b)f′(x)=x2+1x2−1; increasing on (−∞,−1] and [1,∞); local max 2π−1 at −1, local min 1−2π at 1
c)f′′(x)=(1+x2)24x; concave down on (−∞,0), up on (0,∞); inflection (0,0), tangent y=−x
d)f<0 on (0,1], f increasing on [1,∞), f(2)<0<f(π); then oddness
e)Range R; three solutions for ∣k∣<2π−1, two for ∣k∣=2π−1, one otherwise
a) arctan is defined on R: the domain is R, and f is continuous everywhere, so there is no vertical asymptote. f(−x)=−x−2arctan(−x)=−x+2arctanx=−f(x), since arctan is odd: f is odd, its graph symmetric about the origin. Asymptotes: f(x)−(x−π)=π−2arctanx→π−2⋅2π=0 as x→∞, and f(x)−(x+π)=−π−2arctanx→−π+π=0 as x→−∞. So y=x−π is a slant asymptote on the right and y=x+π on the left: TWO different slant asymptotes, because arctan has two different limits.
Position: arctanx<2π for every x, so f(x)−(x−π)=π−2arctanx>0: the graph is always ABOVE y=x−π. And arctanx>−2π, so f(x)−(x+π)=−(π+2arctanx)<0: always BELOW y=x+π. The whole graph lives in the band between the two dashed lines.
b) f′(x)=1−1+x22=x2+1x2−1, defined everywhere; the denominator is positive, so f′ has the sign of x2−1. Critical numbers: ±1. f′>0 on (−∞,−1) and (1,∞), f′<0 on (−1,1): f increases on (−∞,−1], decreases on [−1,1], increases on [1,∞). By the First Derivative Test, local maximum f(−1)=−1−2(−4π)=2π−1 and local minimum f(1)=1−2π, the two values opposite as oddness requires.
c) f′′(x)=dxd(1−2(1+x2)−1)=2(1+x2)−2⋅2x=(1+x2)24x, with the sign of x. Concave down on (−∞,0), concave up on (0,∞): the concavity CHANGES at 0, where f is continuous, so (0,0) is an inflection point. There f′(0)=−1: the tangent is y=−x, and the graph crosses it at the origin.
d) f(0)=0. On (0,1], f is decreasing from f(0)=0, so f(x)<0: no zero there. On [1,∞), f is increasing, so it has at most one zero. Existence: f is continuous on [2,π]. f(2)=2−2arctan2<0 amounts to arctan2>1, that is tan1<2: true, since 0<1<3π (as π>3) and tan increases on (0,2π), so tan1<tan3π=3<2. And f(π)=π−2arctanπ>π−2⋅2π=0. By the Intermediate Value Theorem, f(r)=0 for exactly one r in (2,π). By oddness f(−r)=0, and there is no other zero on (−∞,0). Exactly three zeros.
e) f is continuous, f(x)→∞ as x→∞ (it stays above x−π) and f(x)→−∞ as x→−∞: by the IVT its range is R. The number of solutions of f(x)=k is the number of times the horizontal line y=k meets the graph. With m=2π−1, the local maximum value: for −m<k<m the line meets each of the three monotonic pieces once, three solutions; for k=±m it touches at an extremum and crosses once elsewhere, two solutions; for ∣k∣>m, one solution. Sign table: f′ is +,0,−,0,+ at −1 and 1; f′′ is −,0,+ at 0.
Exercise 11: Optimization: the rain gutter folded at the best angle
A rain gutter is made from a long sheet of metal 30 cm wide, by folding up 10 cm on each side. The base stays 10 cm wide, and each side makes the same angle θ with the horizontal, 0≤θ≤2π, as in the cross-section of the figure. The gutter carries the most water when the area A of its cross-section is largest.
a) Show that A(θ)=100sinθ(1+cosθ), in cm2.
b) Find the critical numbers of A in (0,2π), factoring A′(θ).
c) Find the angle that gives the largest cross-section, justifying that the maximum is ABSOLUTE, and give that area exactly. Compare with a rectangular gutter, θ=2π.
d) A rectangular gutter could also be folded along other lines: fold up x cm on each side, at a right angle. Find the best x and the largest area, and show without a calculator that the best gutter of c) still carries more.
Show the solution
Answers
a)Height 10sinθ, top width 10+20cosθ: A=100sinθ(1+cosθ)
b)A′(θ)=100(2cosθ−1)(cosθ+1); only θ=3π
c)A(0)=0, A(3π)=753, A(2π)=100: maximum 753 cm2 at θ=3π
d)A=x(30−2x), best x=7.5, area 112.5 cm2<753
a) The cross-section is a trapezoid. Each side, of length 10, rises by 10sinθ and moves outward by 10cosθ: the height is h=10sinθ, the base 10 and the top 10+2⋅10cosθ. Its area is the mean of the parallel sides times the height: A=210+(10+20cosθ)⋅10sinθ=(10+10cosθ)⋅10sinθ=100sinθ(1+cosθ). The domain is the closed interval [0,2π] given by the statement, and A is continuous on it.
b) Product rule: A′(θ)=100[cosθ(1+cosθ)+sinθ(−sinθ)]=100(cosθ+cos2θ−sin2θ). Replace sin2θ by 1−cos2θ to get a quadratic in cosθ: A′(θ)=100(2cos2θ+cosθ−1)=100(2cosθ−1)(cosθ+1). On (0,2π), cosθ+1>0, so A′=0 only when cosθ=21: θ=3π. The other factor would give θ=π, outside the domain.
c) Closed interval method, A being continuous on [0,2π]: A(0)=0 (a flat sheet), A(3π)=100⋅23⋅23=753, and A(2π)=100⋅1⋅1=100, the rectangular gutter. Compare exactly: 753>100 amounts to 3>34, that is 3>916, true. The absolute maximum is 753 cm2, about 130 cm2, at θ=3π: the sides lean outward at 60∘, the top is 20 cm wide and the depth 53 cm. The sign of A′, that of 2cosθ−1, confirms it: positive before 3π, negative after. Stopping at A′=0, without the endpoint 2π whose value is not 0, is where the mark is lost.
d) Folding x cm on each side at right angles leaves a base of 30−2x: A(x)=x(30−2x)=30x−2x2 on [0,15]. A′(x)=30−4x=0 at x=7.5; the closed interval method gives A(0)=A(15)=0 and A(7.5)=7.5⋅15=112.5 cm2, the absolute maximum. And 753>112.5 amounts to 3>1.5, that is 3>2.25: the best trapezoid beats the best rectangle.
Exercise 12: Antiderivatives: rewrite first, one condition per constant
Two points each. Check every answer by differentiating it.
a) Find f if f′(x)=x+2x−4 for x>0 and f(9)=1.
b) Find f if f′′(x)=x21 for x>0, f(1)=0 and f(e)=0.
c) The area of a growing leaf increases at the rate A′(t)=3t cm2 per day, t in days, and A(0)=2 cm2. Find A(t), and the day on which the area reaches 18 cm2.
Show the solution
Answers
a)f(x)=32x3/2−2x+1
b)f(x)=e−1x−1−lnx
c)A(t)=2t3/2+2; 18 cm2 at t=4 days
a) No quotient rule runs backwards: rewrite first. x−4=(x−2)(x+2) for x>0, so f′(x)=x−2=x1/2−2. The general antiderivative on (0,∞) is f(x)=32x3/2−2x+C. Condition: f(9)=32⋅27−18+C=C=1. So f(x)=32x3/2−2x+1. Check: f′(x)=x1/2−2, and (x−2)(x+2)=x−4.
b) First antiderivative: f′(x)=−x1+C. Second: f(x)=−lnx+Cx+D, on the interval (0,∞), where ln∣x∣=lnx. Two constants, two conditions: f(1)=C+D=0 and f(e)=−1+Ce+D=0. Subtracting, C(e−1)=1, so C=e−11 and D=−e−11. Hence f(x)=e−1x−1−lnx. Check: f′′=x21, f(1)=0−0, f(e)=1−1. The two conditions are VALUES at two points, not a value and a slope: they still fix the two constants, because the resulting system has a unique solution.
c) A(t)=3⋅32t3/2+C=2t3/2+C, and A(0)=C=2: A(t)=2t3/2+2. Then 2t3/2+2=18 gives t3/2=8, t=82/3=4: the area reaches 18 cm2 after 4 days. Forgetting the constant, A=2t3/2, gives t=92/3, a leaf with no area at t=0.