MATH 203 Calculus I • Concordia University, Montreal
Corrected exercises: derivatives of inverse functions and logarithms (MATH 203)
This is the corrected exercise set for section 3.8 of Thomas' Calculus, derivatives of inverse functions and logarithms, in MATH 203, Differential and Integral Calculus I, at Concordia University.It completes the differentiation rules of the course: after it, ln, loga, ax and powers with a variable exponent can all be differentiated.
What the set covers, the traps it targets
It completes the differentiation rules of the course: after it, ln, loga, ax and powers with a variable exponent can all be differentiated.
A scientific calculator is allowed, as on the exams, so exact answers come first and decimals only where the question announces a rounding. Every solution names the rule it applies, the inner function of each chain rule, and the domain it works on.
The thread running through the set: every derivative of this chapter comes from UNDOING. The slope of f−1 at b is the reciprocal of the slope of f at the SWAPPED point a=f−1(b); and ln undoes products, quotients and powers, so the laws of logarithms are applied BEFORE differentiating, the true ones only, and every base other than e leaves its constant lna. The marks of this chapter are lost in that algebra, not in the rules: a cube root that must become the power 31, a denominator that comes out with a minus sign, ax rewritten as exlna, 81−3/4 taken root first.
The traps named in the solutions: f′ evaluated at b instead of a, ln(x2+9) split like a product, the domain lost when ln((x−1)(x−4)) is expanded, ∣x∣1 for the derivative of ln∣x∣, the power rule applied to 3x, 41 in place of ln41, logxe differentiated as if its base were constant, logarithmic differentiation of a sum, stopping at yy′, the missing term of xx2, the answer 1 to a form 1∞, and a 15% yearly loss read as a relative rate of −0.15.
•ax=exlna, so dxdau=aulna⋅u′; the slope of ax at 0 is lna, equal to 1 only for a=e.
•Laws, for a,b>0: ln(ab)=lna+lnb, lnba=lna−lnb, ln(ar)=rlna, lneu=u. Nothing splits ln(a+b).
•Logarithmic differentiation: ln∣y∣, expand, differentiate (yy′ on the left), multiply by y. Required when x is in the base AND the exponent of one power.
•limh→0hln(1+h)=1, hence e=limx→0(1+x)1/x.
Part A: the basics (/50)
Exercise 1: The derivative of ln: the domain first, the inner function named
9 answers to check
Section 3.8 of Thomas gives dxdlnx=x1 for x>0 and, with the chain rule, dxdlnu=uu′ wherever u>0. With an absolute value, dxdln∣u∣=uu′ wherever u=0: the formula keeps no absolute value.
Every question on a logarithm starts with its DOMAIN, read on the function and not on its derivative. The figure shows y=ln(4−x2) between its two dashed boundaries.
a) Give the domain of f(x)=ln(4−x2), differentiate f, compute f′(1) and find where the tangent is horizontal.
b) Differentiate g(x)=ln∣2x−5∣ and compute g′(2). Why would ln(2x−5) give no answer at x=2?
c) Differentiate h(x)=xlnx−x and compute h′(e2).
d) Give the domain of k(x)=lnx, differentiate it, and compute k′(e) to four decimals.
e) Thomas obtains dxdlnx from the identity elnx=x. Redo this derivation, then write the tangent line to y=lnx at x=e and show that it passes through the origin.
Show the solution
Answers
a)Domain (−2,2); f′(x)=−4−x22x, f′(1)=−32; horizontal tangent at x=0
b)g′(x)=2x−52, g′(2)=−2; ln(2x−5) is undefined for x≤25
c)h′(x)=lnx, h′(e2)=2
d)Domain [1,∞); k′(x)=2xlnx1 for x>1, k′(e)=2e1≈0.1839
e)y=lnx⟺ey=x, so eyy′=1 and y′=x1; tangent y=ex, intercept 0
a)
The logarithm needs 4−x2>0, that is x2<4, that is −2<x<2. The algebra slip that costs the domain mark is to write x<±2 or x<2: a square smaller than 4 traps x between −2 and 2, which is what the dashed lines show. Chain rule with inner function u=4−x2, u′=−2x: f′(x)=uu′=−4−x22x. At x=1: f′(1)=−32. The denominator is positive on the whole domain, so f′(x)=0 exactly when x=0: horizontal tangent at (0,ln4), the top of the arch on the figure. Writing 4−x21, the derivative of ln without the derivative of the inside, is the most frequent loss.
b)
g is defined wherever 2x−5=0, that is x=25. By dxdln∣u∣=uu′ with u=2x−5, u′=2: g′(x)=2x−52, with NO absolute value in the answer. At x=2: g′(2)=−12=−2. The student who writes ∣2x−5∣2 gets +2, the wrong sign: on the left branch ∣2x−5∣=5−2x decreases as x increases, and so does its logarithm. Without the absolute value, ln(2x−5) exists only for x>25, and g′(2) would be a number about a point where there is no function.
c)
Product rule on xlnx: (x)′lnx+x(lnx)′=lnx+x⋅x1=lnx+1. Subtract the derivative of x: h′(x)=lnx+1−1=lnx, for x>0. At x=e2: h′(e2)=lne2=2, since ln and exp undo each other. The simplification x⋅x1=1 looks trivial; forgetting it leaves an answer the marker has to decode.
d)
The square root needs lnx≥0, that is x≥1, and ln itself needs x>0: domain [1,∞). Chain rule, outer function the square root, inner function u=lnx: k′(x)=2lnx1⋅x1=2xlnx1, valid for x>1 only: at x=1 the denominator is 0, and the graph leaves (1,0) with a vertical tangent. At x=e: lne=1, so k′(e)=2e1≈0.1839. Rounding comes LAST: the exact value 2e1 is the answer, the decimal is what the calculator reports.
e)
Let y=lnx for x>0. By definition of ln as the inverse of ex, ey=x. Differentiate both sides with respect to x, y being a function of x (implicit differentiation, section 3.7): ey⋅y′=1. Since ey>0, divide: y′=ey1=x1, the last step replacing ey by x. At x=e: the point is (e,lne)=(e,1) and the slope is e1. Tangent: y=1+e1(x−e)=1+ex−1=ex. Its intercept is 0: it passes through the origin. The algebra step e1⋅e=1 is the whole proof; a student who leaves y−1=e1(x−e) unsimplified has the right line and cannot see the origin in it.
Points are earned before the solution; after it, you can keep practising.
For you:
Tick the exercises you have done or want to review: a free account, no password, keeps your ticks and your right answers from one visit to the next, tells you which chapter to tackle next and lets you ask for the exercise you are missing. Create your space, an email is enough.
Exercise 2: The laws of logarithms BEFORE differentiating, and only the true ones
9 answers to check
For a>0 and b>0: ln(ab)=lna+lnb, lnba=lna−lnb, ln(ar)=rlna, and lneu=u. A root is a fractional power, 3x=x1/3; a factor in the denominator comes out with a MINUS sign. Expanding first turns one quotient rule inside a chain rule into a sum of short terms.
Nothing splits ln(a+b). And the laws need POSITIVE arguments: with factors of unknown sign they are written with absolute values.
a) For x>0, differentiate f(x)=ln(5x+1)43x(x2+3)2 and compute f′(1).
b) Differentiate g(x)=ln1+e2xe2x, simplify to a single fraction, and compute g′(0).
c) For x>0, differentiate h(x)=log2x+18x3 and compute h′(1) to four decimals.
d) Say whether each rewriting is legitimate:(i)ln(x2+9)=2lnx+ln9(ii)lnx2+9=21ln(x2+9)(iii)ln(9x2)=ln9+2ln∣x∣ for x=0.Then compute the true derivative of ln(x2+9) at x=3 and of ln(9x2) at x=−3.
e) Let F(x)=ln((x−1)(x−4)). Give its domain, explain why F(x)=ln(x−1)+ln(x−4) is wrong at x=0, write the correct expansion, and compute F′(0).
d)(i) not legitimate, (ii) and (iii) legitimate; derivatives 31 at x=3 and −32 at x=−3
e)Domain x<1 or x>4; F(x)=ln∣x−1∣+ln∣x−4∣, F′(0)=−45
a)
For x>0 every factor is positive, so the laws apply as they are. The cube root is the power 31 and the denominator comes out with a minus sign: f(x)=31lnx+2ln(x2+3)−4ln(5x+1). This line is the algebra where the marks are: a student who writes 3lnx for the cube root, or +4ln(5x+1), has lost the question before differentiating anything. Term by term, chain rule on the last two with inner functions x2+3 and 5x+1: f′(x)=3x1+x2+32⋅2x−5x+14⋅5=3x1+x2+34x−5x+120. At x=1: 31+1−620=31+1−310=−2.
b)
ln1+e2xe2x=lne2x−ln(1+e2x)=2x−ln(1+e2x), since lneu=u; both arguments are positive for every x. Chain rule on the second term, inner function 1+e2x with derivative 2e2x: g′(x)=2−1+e2x2e2x. Common denominator: g′(x)=1+e2x2(1+e2x)−2e2x=1+e2x2. At x=0: g′(0)=22=1. The fake law ln(1+e2x)=ln1+lne2x=2x would make g identically 0, a function whose graph is the x axis: a quick sanity check kills it, since g(0)=ln21=0.
c)
For x>0: h(x)=log28+3log2x−21log2(x+1)=3+3log2x−21log2(x+1), because log28=3 is a CONSTANT. Change of base, log2u=ln2lnu, so every term carries ln21: h′(x)=ln21(x3−2(x+1)1). At x=1: ln21(3−41)=4ln211≈3.9674. Two frequent losses: differentiating log28 as if it were a function, and dropping the ln21, which gives 2.75.
d)
(i) NOT legitimate: x2+9 is a SUM, and no law splits the logarithm of a sum; 2lnx+ln9=ln(9x2), a different function. The true derivative, chain rule with inner function x2+9: x2+92x, at x=3: 186=31, while the fake version gives x2=32. (ii) Legitimate: the power law with exponent 21, and x2+9>0 for every x. (iii) Legitimate: 9x2=9∣x∣2 with 9>0 and ∣x∣>0, so ln(9x2)=ln9+2ln∣x∣ for every x=0; writing 2lnx would delete the whole left half of the domain. Its derivative is x2, at x=−3: −32.
e)
The logarithm needs (x−1)(x−4)>0: both factors of the same sign, so x<1 or x>4. At x=0 the product is (−1)(−4)=4>0 and F(0)=ln4 exists, but ln(0−1) and ln(0−4) do not: the law ln(ab)=lna+lnb needs a>0 and b>0, which holds only for x>4. On the whole domain: F(x)=ln∣x−1∣+ln∣x−4∣. Then F′(x)=x−11+x−41 and F′(0)=−1−41=−45. Check without expanding: F(x)=ln(x2−5x+4), F′(x)=x2−5x+42x−5, and at 0: −45. The two routes agree, as they must.
Points are earned before the solution; after it, you can keep practising.
For you:
Exercise 3: The derivative of an inverse function at a point, without a formula for the inverse
8 answers to check
Thomas 3.8: if f is differentiable on an interval and f′ is never zero there, then f−1 is differentiable and, at b=f(a), (f−1)′(b)=f′(a)1=f′(f−1(b))1. The graphs of f and f−1 are mirror images in the line y=x: the point (a,b) of f becomes the point (b,a) of f−1, and the slope is inverted.
Let f(x)=ex+2x. The equation ex+2x=y cannot be solved for x with algebra, so no formula for f−1 exists on the page. The figure shows f, f−1 and the line y=x.
a) Explain why f is one-to-one, with no derivative test, and find f−1(1).
b) Compute (f−1)′(1), then the tangent line to y=f−1(x) at the point (1,0).
c) A student answers (f−1)′(1)=f′(1)1=e+21. What did he do, and which point is his number actually about?
d) Let q(x)=x2−6x restricted to x≥3. Compute (q−1)′(16) with the theorem, then find q−1 explicitly by completing the square and check. What happens at b=−9?
e) For p(x)=x4 on x>0, p−1(x)=x1/4. Compute (p−1)′(81) with the theorem, then with the power rule.
Show the solution
Answers
a)ex and 2x are strictly increasing, so is their sum: one-to-one; f(0)=1, so f−1(1)=0
b)(f−1)′(1)=f′(0)1=31; tangent y=31x−31
c)He evaluated f′ at b=1 instead of a=0; e+21 is the slope of f−1 at b=e+2
d)q(8)=16, (q−1)′(16)=q′(8)1=101; q−1(x)=3+x+9; at b=−9, q′(3)=0: vertical tangent
e)(p−1)′(81)=p′(3)1=1081, and 41⋅81−3/4=1081
a)
If x1<x2, then ex1<ex2 and 2x1<2x2; adding, f(x1)<f(x2). So f is strictly increasing, and a horizontal line meets its graph at most once: f is one-to-one (horizontal line test, Thomas 1.6). No derivative is needed, and none is allowed for this step at this point of the course. Since f(0)=e0+0=1, the point (0,1) is on f and f−1(1)=0. Finding a by inspection is the normal way: try x=0, x=1, x=−1 before anything else.
b)
f′(x)=ex+2, never zero, so the theorem applies. With b=1 and a=f−1(1)=0: f′(0)=1+2=3, and (f−1)′(1)=f′(0)1=31. The tangent to f−1 at (1,0): y=0+31(x−1)=31x−31. On the solution figure, the tangent to f at (0,1) has slope 3 and its mirror image, the tangent to f−1 at (1,0), has slope 31: reflecting a line in y=x swaps rise and run.
c)
He evaluated f′ at b=1, the input of f−1, instead of at a=f−1(1)=0, the input of f. The theorem pairs the two graphs point by point: the slope of f−1 at (b,a) is the reciprocal of the slope of f at the SWAPPED point (a,b). His number f′(1)1=e+21≈0.2119 is a true slope of f−1, but at b=f(1)=e+2, a different point. Writing the pair (a,b) on the copy before any derivative is the gesture that prevents this loss, which costs the whole question.
d)
q(x)=(x−3)2−9 is a parabola with vertex at x=3, so on x≥3 it is one-to-one (restricting the domain, Thomas 1.6). Solve q(a)=16: a2−6a−16=0, (a−8)(a+2)=0, and a=−2 is rejected since a≥3: a=8. q′(x)=2x−6, q′(8)=10, so (q−1)′(16)=101. Explicitly: y=x2−6x gives y+9=(x−3)2, so x=3+y+9, with the PLUS sign because x≥3; hence q−1(x)=3+x+9. Its derivative is 2x+91, and at 16: 2⋅51=101. The theorem and the explicit route agree. At b=−9=q(3), q′(3)=0: the hypothesis fails, and indeed 2x+91 blows up at x=−9, where q−1 has a VERTICAL tangent, the mirror image of the horizontal tangent of q at its vertex.
e)
p−1(81)=3, since 34=81 and 3>0. p′(x)=4x3, p′(3)=108, so (p−1)′(81)=1081. Power rule: dxdx1/4=41x−3/4, and 81−3/4=(811/4)−3=3−3=271, so the value is 41⋅271=1081. The rewriting of 81−3/4, root first and power second, is where a MATH 203 copy loses the mark: taking 813 first gives a number no one roots by hand, and the negative exponent must end up in the denominator.
Points are earned before the solution; after it, you can keep practising.
For you:
Exercise 4: Logarithms in other bases: the constant ln a, even when the base moves
9 answers to check
For a>0, a=1: logax=lnalnx, so dxdlogau=ulnau′. The change of base is the whole method: it turns every logarithm into ln divided by a constant, or, when the base itself contains x, into a QUOTIENT of two natural logarithms. The base never disappears: it survives as lna, positive for a>1, negative for 0<a<1; and dxdax=axlna (section 3.6) shows the same constant as a slope.
The figure shows y=2x, y=ex and y=4x, all through (0,1), and the dashed line y=1+x.
a) Differentiate k(x)=log4(x2+12) and compute k′(2) to four decimals.
b) For x>0, differentiate g(x)=x2log3x, factor the result, and compute g′(3) to four decimals.
c) For x>1, rewrite h(x)=logxe with natural logarithms, differentiate it, and compute h′(e) to four decimals.
d) The dashed line y=1+x is tangent to one of the three curves at (0,1). Which one? Give the slopes of the other two at x=0, and the point where the tangent to y=4x at x=0 cuts the x axis.
e) Simplify, then differentiate: m(x)=log2x⋅logx8 for x>0, x=1, and n(x)=log3(9x).
d)y=ex; slopes ln2≈0.6931 and ln4≈1.3863; the tangent y=1+xln4 cuts the axis at x=−ln41≈−0.7213
e)m(x)=3, m′(x)=0; n(x)=2x, n′(x)=2
a)
Change of base: k(x)=ln4ln(x2+12), and ln41 is a constant factor. Chain rule with inner function x2+12: k′(x)=ln41⋅x2+122x. At x=2: 16ln44=4ln41, and since ln4=ln22=2ln2, this is 8ln21≈0.1803. The domain is all of R, since x2+12>0. Writing 4(x2+12)2x, the base in place of its logarithm, gives 0.25 and costs the mark.
b)
Product rule, with (log3x)′=xln31: g′(x)=2xlog3x+x2⋅xln31=2xlog3x+ln3x, the simplification xx2=x being the step to write. Factor x: g′(x)=x(2log3x+ln31). At x=3: log33=1, so g′(3)=3(2+ln31)=6+ln33≈8.7307. The two frequent losses: 3x1 for the derivative of log3x, which gives 7, and log33 left unevaluated, as if it were a mystery number.
c)
Here the BASE is the variable, so the rule xlna1 does not apply: it needs a constant base. Change of base: h(x)=logxe=lnxlne=lnx1, defined for x>0, x=1, and here x>1. Rewrite as a power, h=(lnx)−1, and use the chain rule with inner function lnx: h′(x)=−(lnx)−2⋅x1=−x(lnx)21. At x=e: −e⋅11=−e1≈−0.3679. The copy that answers elnx1, treating x as the constant and e as the variable, has read the rule backwards: before any rule, ask which of base and argument contains x.
d)
The slope of ax at x=0 is a0lna=lna. It equals 1 exactly when a=e: the line y=1+x, of slope 1 through (0,1), is the tangent to y=ex. For 2x the slope at 0 is ln2≈0.6931<1, flatter than the dashed line; for 4x it is ln4≈1.3863>1, steeper, as the figure shows. The tangent to 4x at 0 is y=1+xln4; it cuts the x axis where 1+xln4=0, at x=−ln41=−2ln21≈−0.7213, to the right of −1, where the dashed line cuts it: a steeper tangent reaches the axis sooner. This is how Thomas characterizes e: the one base whose exponential crosses the y axis with slope exactly 1, which is why dxdex=ex with no constant.
e)
By change of base, m(x)=ln2lnx⋅lnxln8=ln2ln8=ln23ln2=3 for x>0, x=1: m is constant and m′(x)=0 on its domain. And n(x)=log3(9x)=xlog39=2x by the power law, so n′(x)=2. The product rule on m and the chain rule on n lead to the same answers after half a page of fractions; rewriting first is not elegance, it removes the places where marks are lost.
Points are earned before the solution; after it, you can keep practising.
For you:
Exercise 5: Logarithmic differentiation of heavy products and quotients
10 answers to check
Logarithmic differentiation, Thomas 3.8: take ln∣y∣ of both sides, expand with the laws of logarithms, differentiate both sides (the left side gives yy′ by the chain rule), then multiply by y. The absolute value makes the method valid wherever y=0.
The expansion is the algebra step where the marks are: each exponent in front, roots as fractional powers, the denominator with minus signs, and lneu=u.
a) Let y=(x2+1)3(x+1)42x+7 for x>−1. Find yy′, then y(1), y′(1) and the tangent line at x=1.
b) Let y=(x+3)2(x−1)(x−5)3. Explain why lny cannot be used at x=3, and compute y′(3) by logarithmic differentiation.
c) Let y=3x2+7ex(2x+1)5 for x>−21. Compute y(1)y′(1) exactly, then y′(1) to two decimals.
d) A student applies the method to y=(x+1)4+2x+7 and writes lny=4ln(x+1)+21ln(2x+7). Is the method needed here? Find the true y′(1).
e) For x>2, let y=(x−2)2x3. By logarithmic differentiation, write y′ as a single fraction and find the point where the tangent is horizontal.
d)Not needed (a SUM: no law applies); y′=4(x+1)3+2x+71, y′(1)=397
e)y′=(x−2)3x2(x−6); horizontal tangent at (6,227)
a)
For x>−1 every factor is positive, so lny=4ln(x+1)+21ln(2x+7)−3ln(x2+1). Differentiate both sides, chain rule on each term: yy′=x+14+21⋅2x+72−x2+13⋅2x=x+14+2x+71−x2+16x. At x=1: y(1)=816⋅3=6 and yy′=2+91−3=−98, so y′(1)=6⋅(−98)=−316. Tangent: y=6−316(x−1). Stopping at −98 and calling it the slope forgets the last step: the method ends by MULTIPLYING BY y.
b)
At x=3: y(3)=62⋅2(−2)3=−728=−91<0, so lny does not exist there, and ln(x−5)=ln(−2) neither. With absolute values, valid wherever y=0: ln∣y∣=3ln∣x−5∣−2ln∣x+3∣−ln∣x−1∣, so yy′=x−53−x+32−x−11. At 3: −23−62−21=−69+2+3=−37. Then y′(3)=−91⋅(−37)=277. Two negative signs meet in the last product: a sign lost in y(3) gives −277.
c)
All factors are positive for x>−21. Expand, with the cube root as the power 31 and lnex=x: lny=5ln(2x+1)−31ln(x2+7)−x. Differentiate: yy′=2x+110−3(x2+7)2x−1. At x=1: 310−242−1=1240−1−12=1227=49. And y(1)=38e35=2e243, so y′(1)=2e243⋅49=8e2187≈100.57. The quotient rule here needs a product rule and two chain rules inside it; the logarithm reduces it to three one-line terms.
d)
Not needed, and not even legal as written. y is a SUM, and no law splits the logarithm of a sum: ln((x+1)4+2x+7) is not 4ln(x+1)+21ln(2x+7), which is the logarithm of the PRODUCT. Differentiate term by term: y′=4(x+1)3+2x+71, so y′(1)=32+31=397. The student's line leads to y′=y(2+91)=19⋅919=9361 at x=1, a wrong answer reached by a correct-looking method. Logarithmic differentiation is for products, quotients and powers only.
e)
For x>2, y>0 and lny=3lnx−2ln(x−2). Then yy′=x3−x−22. Common denominator, the algebra step that makes the zero visible: x(x−2)3(x−2)−2x=x(x−2)x−6. Multiply by y: y′=(x−2)2x3⋅x(x−2)x−6=(x−2)3x2(x−6). On x>2 the factors x2 and (x−2)3 are positive, so y′=0 exactly at x=6, where y=16216=227. Horizontal tangent at (6,227). Whether this is a minimum is a question for chapter 4.
Points are earned before the solution; after it, you can keep practising.
For you:
Part B: problems and reasoning (/50)
Exercise 6: A variable in the base AND in the exponent of the same power
9 answers to check
The power rule needs a constant exponent, the rule for ax a constant base. When x sits in both places of ONE power uv, neither applies: take lny=vlnu, which turns the power into a product, differentiate with the product rule, and multiply by y. The same computation can be read as uv=evlnu.
The figure shows y=x1/x on (0,10]: it seems to have a high point somewhere near x=3.
a) For x>0, differentiate y=x1/x. Find exactly the point where the tangent is horizontal, and the tangent line at x=1.
b) Differentiate y=(x2+1)x and compute y′(1) to four decimals.
c) Differentiate y=xx2 for x>0 and compute y′(2) exactly. A student writes y′=x2⋅xx2−1: what does he get at x=2, and which term is missing?
d) For x>1, differentiate y=(lnx)x and compute y′(e).
e) Does y=3xx3 need logarithmic differentiation? Compute y′(1) to four decimals.
Show the solution
Answers
a)y′=x1/xx21−lnx; horizontal tangent at (e,e1/e)≈(2.7183,1.4447); tangent at x=1: y=x
c)y′=xx2(2xlnx+x), y′(2)=64ln2+32≈76.3614; the student gets 32, missing 64ln2
d)y′=(lnx)x(ln(lnx)+lnx1), y′(e)=1
e)No: x is in the exponent of one factor and the base of another, so the product rule suffices; y′(1)=9+3ln3≈12.2958
a)
For x>0, y>0 and lny=x1lnx=xlnx. Quotient rule on the right: yy′=x2x1⋅x−lnx⋅1=x21−lnx, the simplification x1⋅x=1 being the step to write. So y′=x1/xx21−lnx. Since x1/x>0 and x2>0, y′=0 exactly when lnx=1, that is x=e: the horizontal tangent is at (e,e1/e)≈(2.7183,1.4447). The figure suggested 3; the computation says e. At x=1: y=1 and y′=1⋅11−0=1, so the tangent is y=1+(x−1)=x.
b)
(x2+1)x>0 for every x, so lny=xln(x2+1). Product rule, with the chain rule on ln(x2+1), inner function x2+1: yy′=ln(x2+1)+x⋅x2+12x. So y′=(x2+1)x(ln(x2+1)+x2+12x2). At x=1: y(1)=2 and yy′=ln2+1, so y′(1)=2ln2+2≈3.3863. Reading it as exln(x2+1) and applying the chain rule gives the same line: the two methods are one.
c)
lny=x2lnx. Product rule: yy′=2xlnx+x2⋅x1, and the algebra step is x2⋅x1=x, a cancellation to write out on the copy rather than carry as xx2. So y′=xx2(2xlnx+x). At x=2: y=24=16, the exponent x2=4 computed FIRST, and yy′=4ln2+2, so y′(2)=64ln2+32≈76.3614. The student treats x2 as a constant exponent: 4⋅23=32. He gets exactly the term 16⋅2=32 that comes from the BASE, and misses 16⋅4ln2=64ln2, the term that comes from the EXPONENT. Reading xx2 as (xx)2=x2x is the other algebra slip: at x=2 it gives 16 by coincidence, and at x=3 it gives 729 instead of 39=19683.
d)
The base lnx must be positive, hence x>1. lny=xln(lnx). Product rule, with the chain rule on ln(lnx), inner function lnx: yy′=ln(lnx)+x⋅lnx1⋅x1=ln(lnx)+lnx1. At x=e: lne=1, so ln(lne)=ln1=0 and lne1=1; also y(e)=1e=1. Hence y′(e)=1⋅(0+1)=1. The two logarithms stacked are the trap: ln(lnx) is not (lnx)2 and not lnx⋅lnx.
e)
No. The rule of thumb is about ONE power: here x is the exponent of the factor 3x and the base of the factor x3, and each factor has a rule of its own. Product rule: y′=3xln3⋅x3+3x⋅3x2=3xx2(xln3+3), after factoring 3xx2. At x=1: 3(ln3+3)=9+3ln3≈12.2958. Logarithmic differentiation would also work, but it answers a question nobody asked: the decision is read on the SHAPE, variable in the base and the exponent of the same power, or not.
Points are earned before the solution; after it, you can keep practising.
For you:
Exercise 7: The number e as a limit: a calculator table, then a proof
12 answers to check
Thomas 3.8 proves that e=limx→0(1+x)1/x, using only the derivative of ln at 1. The figure shows y=(1+x)1/x: the function is not defined at x=0, and the hole is the limit.
A scientific calculator computes the values; it cannot compute the limit. A table suggests, a proof decides.
a) With a calculator, compute (1+x)1/x to four decimals for x=0.1, x=0.01, x=0.001 and x=−0.01. What do the values suggest?
b) Write dxdlnxx=1 as a limit, deduce limh→0hln(1+h), and prove that limx→0(1+x)1/x=e.
c) Find limx→0(1+2x)3/x and limx→0(1−3x)2/x, each in the form ek.
d) Recognize each limit as a derivative at a point and evaluate it: limh→0h5h−1 and limh→0hlog3(9+h)−2.
e) Study (1+x)1/x2 as x→0+ and as x→0−. Does the limit as x→0 exist?
Show the solution
Answers
a)2.5937, 2.7048, 2.7169, 2.7320: the values approach e≈2.7183, from below on the right and from above on the left
b)limh→0hln(1+h)=1, so ln[(1+x)1/x]→1 and (1+x)1/x→e
c)e3/2 and e−6
d)ln5≈1.6094 and 9ln31≈0.1011
e)+∞ from the right, 0 from the left: the limit does not exist
a)
1.110≈2.5937, 1.01100≈2.7048, 1.0011000≈2.7169 and 0.99−100≈2.7320, with the yx key. The values from the right increase toward 2.718…, the value from the left is slightly above it: both sides seem to close in on e≈2.7183. A table is evidence, not proof: it cannot rule out that the values turn away for x=10−9, and a calculator pushed too far even returns 1, because it rounds 1+10−15 to 1. That is why b) is needed.
b)
By the definition of the derivative, dxdlnxx=1=limh→0hln(1+h)−ln1=limh→0hln(1+h), and this derivative is known: 11=1. So hln(1+h)→1. The power law read from right to left turns this into the logarithm of our function: h1ln(1+h)=ln[(1+h)1/h]. Hence ln[(1+x)1/x]→1 as x→0. Since (1+x)1/x=eln[(1+x)1/x] and the exponential is continuous, (1+x)1/x→e1=e. Stopping at ln[…]→1 and answering 1 is the classic lost mark: that is the limit of the logarithm.
c)
ln[(1+2x)3/x]=x3ln(1+2x)=23⋅uln(1+u) with u=2x→0; it tends to 23, so the limit is e3/2. The algebra is to multiply and divide by the right number: x3=23⋅u1. Similarly x2ln(1−3x)=−6⋅uln(1+u) with u=−3x→0: it tends to −6, and the limit is e−6. The sign of −3 goes to the exponent; losing it gives e6, a huge number for a base below 1 raised to large powers, which a calculator check at x=0.001 exposes at once.
d)
Since 50=1, h5h−1=h50+h−50 is the difference quotient of 5x at 0; its limit is dxd5xx=0=50ln5=ln5≈1.6094. Since log39=2, hlog3(9+h)−2=hlog3(9+h)−log39 is the difference quotient of log3 at 9, and its limit is 9ln31≈0.1011. The skill is to spot the three pieces: a function, a point, and the value at that point subtracted, here hidden as 1 and as 2.
e)
ln[(1+x)1/x2]=x2ln(1+x)=x1⋅xln(1+x). The second factor tends to 1 by b). As x→0+, x1→+∞, so the logarithm tends to +∞ and the function to +∞. As x→0−, x1→−∞, the logarithm tends to −∞ and the function to 0. The one-sided limits differ, so the limit does not exist. Calculator check: 1.0110000 is about 1.6×1043, 0.9910000 about 2×10−44. The form 1∞ is the same as in b), and the outcome is completely different: the form decides nothing, the logarithm decides.
Points are earned before the solution; after it, you can keep practising.
For you:
Exercise 8: Five statements to correct
10 answers to check
Each statement below was written on a MATH 203 assignment, and each is false. Say what is wrong, give the correct result, and check it on a value.
a) For f(x)=2x+lnx, (f−1)′(2)=f′(2)1=52.
b) dxd3x=x3x−1.
c) dxdln(ex2)=ex21.
d) ln(x−1)−ln(x+1)=ln(x+1)ln(x−1), so its derivative follows from the quotient rule.
e) dxdln∣x∣=∣x∣1.
Show the solution
Answers
a)False: f(1)=2, so (f−1)′(2)=f′(1)1=31
b)False: dxd3x=3xln3, equal to 3ln3≈3.2958 at x=1
c)False: ln(ex2)=x2, derivative 2x
d)False: the difference is lnx+1x−1; derivative x−11−x+11=x2−12, 41 at x=3
e)False: dxdln∣x∣=x1, equal to −21 at x=−2
a)
FALSE. The theorem evaluates f′ at a=f−1(b), not at b. Here f(1)=2+ln1=2, so a=1. f′(x)=2+x1, f′(1)=3, and (f−1)′(2)=31. The number f′(2)1=5/21=52 is the slope of f−1 at b=f(2)=4+ln2, another point. Correct: find a first, write the pair (a,b)=(1,2), then invert f′(a).
b)
FALSE. The power rule needs a constant exponent; in 3x the exponent is the variable and the base is the constant. Rewrite 3x=exln3: its derivative is exln3ln3=3xln3. At x=1: 3ln3≈3.2958, while the statement gives 1⋅30=1. Another check: at x=0 the statement gives slope 0, while 3x crosses the y axis rising with slope ln3>0, as 2x and 4x do on the figure of exercise 4.
c)
FALSE twice. First, ln and exp undo each other: ln(ex2)=x2, whose derivative is 2x. Second, even without simplifying, the chain rule gives uu′ with u=ex2, u′=2xex2, so ex22xex2=2x; the statement forgot the derivative of the inside. At x=1: 2, not e1. Simplifying first leaves nothing to forget.
d)
FALSE. The law is lna−lnb=lnba, the logarithm of a quotient, not a quotient of logarithms. And no rewriting is needed: differentiate the difference term by term, x−11−x+11=(x−1)(x+1)(x+1)−(x−1)=x2−12, for x>1. At x=3: 82=41. Numerically, at x=3 the difference is ln2−ln4=−ln2≈−0.69 while the quotient is ln4ln2=21: not even the same sign.
e)
FALSE. For x<0, ln∣x∣=ln(−x), and the chain rule with inner function −x gives −x−1=x1. So dxdln∣x∣=x1 for every x=0, negative on the left branch. At x=−2: −21, not 21. The graph agrees: as x increases from −5 to −1, ∣x∣ decreases, so ln∣x∣ decreases, and a decreasing function cannot have a positive slope there.
Points are earned before the solution; after it, you can keep practising.
For you:
Exercise 9: Rates on logarithmic scales: decibels and pH
6 answers to check
The sound level of a sound of intensity I, in W/m², is β=10log10I0I decibels, with I0=10−12 W/m². The pH of a solution is pH=−log10[H+], the concentration in mol/L. Both are logarithms in base 10, so every rate on these scales carries the factor ln101.
In open air, a loudspeaker produces at a distance of r metres, r≥1, the intensity I(r)=r210−2 W/m². The figure shows its sound level against the distance. Separately, 50 mL of an acid solution with [H+]=0.01 mol/L is diluted with V mL of pure water; the number of moles of H+ does not change, so [H+]=50+V0.5 mol/L.
a) Using the laws of logarithms, show that β(r)=100−20log10r, and compute β(10).
b) Compute β′(r), then β′(10) to four decimals, with its unit and the meaning of its sign.
c) At what distance does the sound level drop at a rate of exactly 1 dB per metre? Give it to two decimals.
d) Show that pH(V)=log10(100+2V), check pH(0)=2, and compute the rate of change of the pH at V=0 and at V=450 mL. By what factor has the rate dropped?
e) For which volume of added water does the pH rise at the rate of 0.001 pH unit per mL? Give it to two decimals.
Show the solution
Answers
a)β=10log10r21010=100−20log10r; β(10)=80 dB
b)β′(r)=−rln1020; β′(10)=−ln102≈−0.8686 dB per metre: the level falls as the listener moves away
c)r=ln1020≈8.69 m
d)pH′(V)=(50+V)ln101; ≈0.0087 per mL at V=0, ≈0.00087 at V=450: ten times smaller
e)V=ln101000−50≈384.29 mL
a)
I0I=r2⋅10−1210−2=r21010: the powers of ten combine first, 10−2/10−12=10−2+12, the exponent rule where copies most often go wrong. Then, for r≥1, log10r21010=log101010−log10r2=10−2log10r, and β(r)=10(10−2log10r)=100−20log10r. At r=10: β=100−20=80 dB. Every tenfold increase of the distance costs 20 dB, which is the shape of the curve on the figure: steep near the speaker, almost flat far away.
b)
log10r=ln10lnr, so β′(r)=−20⋅rln101=−rln1020 dB per metre. At r=10: β′(10)=−ln102≈−0.8686 dB per metre. The sign says the level DECREASES as the distance increases, and the unit is dB per metre, the unit of β over the unit of r. Dropping the ln101 gives −2 dB per metre, more than twice the truth. Differentiating the original 10log10r2⋅10−1210−2 directly works too, through a quotient and a chain rule; the expansion of a) made it one line.
c)
The level drops at 1 dB per metre when β′(r)=−1: rln1020=1, so r=ln1020≈8.69 m. Closer to the speaker the drop is faster, farther it is slower, since ∣β′(r)∣ is inversely proportional to r: at 5 m it is ln104≈1.74 dB per metre, at 20 m only ln101≈0.43.
d)
pH=−log1050+V0.5=log100.550+V=log10(100+2V), using −log10ba=log10ab and 0.51=2. At V=0: log10100=2. Chain rule with inner function 100+2V: pH′(V)=(100+2V)ln102=(50+V)ln101 pH unit per mL. At V=0: 50ln101≈0.0087; at V=450: 500ln101≈0.00087, where the pH is log101000=3. The rate has dropped by a factor of 10, the ratio of the total volumes 50500: each new millilitre dilutes a larger volume and changes the concentration by a smaller fraction.
e)
(50+V)ln101=0.001 gives 50+V=ln101000≈434.29, so V≈384.29 mL. Round only at the end: rounding ln10 to 2.3 first gives 384.78, off by half a millilitre. At that moment the pH is log10(100+768.59)≈2.94.
Points are earned before the solution; after it, you can keep practising.
For you:
Exercise 10: A car, a line of credit, and the relative rate hidden in ln a
8 answers to check
A quantity Q(t)=Q0at changes at the rate Q′(t)=Q0atlna=(lna)Q(t): its RELATIVE rate QQ′ is the constant lna, not a−1. For Q(t)=Q0ert, continuous compounding, the relative rate is r itself.
A car bought for 24000 dollars loses 15% of its value each year: V(t)=24000(0.85)t dollars, t in years. A line of credit of 6000 dollars grows at 7.2% compounded continuously: B(t)=6000e0.072t dollars. Give rates in dollars per year, to the cent, and times to two decimals.
The figure below shows the value V of the car.
a) Compute V′(3) and the relative rate VV′ to four decimals. Why is it not −0.15?
b) When is the car losing value at the rate of 1500 dollars per year?
c) When is interest accruing on the line of credit at the rate of 600 dollars per year?
d) Another lender quotes 7.2% compounded ANNUALLY: C(t)=6000(1.072)t. Give its relative rate to four decimals, and the annual rate, in percent to two decimals, that is equivalent to 7.2% compounded continuously.
e) Find the doubling time of B and of C, and explain which number decides it.
Show the solution
Answers
a)V′(3)=24000(0.85)3ln0.85≈−2395.37 dollars per year; VV′=ln0.85≈−0.1625 per year
b)t=ln0.85ln(1500/(−24000ln0.85))≈5.88 years
c)B=0.072600=325000, so t=0.072ln(25/18)≈4.56 years
d)ln1.072≈0.0695 per year; e0.072−1≈7.47%
e)B: 0.072ln2≈9.63 years; C: ln1.072ln2≈9.97 years; the relative rate decides
a)
By the rule for au: V′(t)=24000(0.85)tln0.85. At t=3: (0.85)3=0.614125, V(3)=14739 dollars, and V′(3)=14739ln0.85≈−2395.37 dollars per year. The sign says the car is LOSING value, since ln0.85<0 because 0.85<1. The relative rate is VV′=ln0.85≈−0.1625 per year, at every t. It is not −0.15: the 15% is the loss over one whole year, measured against the value at the start of the year, while ln0.85 is the instantaneous rate; since the value shrinks during the year, a larger instantaneous rate is needed to lose the same 15%. The solution figure shows the tangent at t=3.
b)
V′(t)=−1500: 24000(0.85)tln0.85=−1500, so (0.85)t=−24000ln0.851500≈0.3846. Take ln of both sides and use the power law: tln0.85=ln0.3846, so t≈−0.1625−0.9556≈5.88 years. Both logarithms are negative and their quotient is positive: a negative time here would mean a sign lost when the equation was divided by ln0.85.
c)
B′(t)=0.072B(t), so interest accrues at 600 dollars per year when B=0.072600=325000≈8333.33 dollars. Then 6000e0.072t=325000 gives e0.072t=1800025000=1825, and t=0.072ln(25/18)≈4.56 years. The rate in dollars per year grows with the balance; the relative rate stays 0.072.
d)
C′(t)=6000(1.072)tln1.072, so CC′=ln1.072≈0.0695 per year, a little LESS than 0.072: annual compounding at 7.2% is slower than continuous compounding at 7.2%. Conversely, in one year B is multiplied by e0.072≈1.07466, so the equivalent annual rate is e0.072−1≈7.47%. The quoted 7.2% is the same number in both contracts; the relative rates, 0.072 and ln1.072, are not.
e)
e0.072t=2 gives t=0.072ln2≈9.63 years; (1.072)t=2 gives t=ln1.072ln2≈9.97 years. In both cases the doubling time is relative rateln2, and that is the number that decides: 0.072 for continuous compounding, ln1.072≈0.0695 for annual compounding. A doubling time computed as 0.072ln2 for the annual contract confuses the quoted rate with the relative rate, the same confusion as −0.15 in a).
Points are earned before the solution; after it, you can keep practising.
Something missing from this set?A type of exercise your teacher gives, a problem you are stuck on, a mistake you spotted: tell me, I read every request.
Sending goes through your free space: that is how I can answer you and tell you when the exercise is added. An email is enough, no password. Create your space.
Rather not create an account? Write to me directly.