This site is a work in progress. New lessons are added regularly. Contact us
Chapter Review · a rapid pre-test review (speedrun)

Exponential and Logarithmic Functions: Chapter Review

A rapid review before the test: the chapter's vocabulary and notation, every formula with the conditions to use it, the standard problem types step by step, and the traps that cost points.

Vocabulary and notation

Growth factor bb and growth rate rr
The rate is the fraction added per step; the factor is what you multiply by, with b=1+rb = 1 + r. A factor of 0.850.85 means a 15%15\% decrease.
Effective annual rate
The once-a-year rate matching a schedule, (1+rn)n1\left(1 + \tfrac{r}{n}\right)^{n} - 1 or er1e^{r} - 1, which is how two offers are compared.
Common logarithm logx\log x
Base 1010, written with no base at all; a calculator's log\log key.
ee
The irrational number 2.7182818282.718281828\ldots, so every decimal you write for it is a rounding.
Argument of a logarithm
The quantity inside it. Positivity restricts the ARGUMENT, never the variable.
Extraneous solution
A candidate satisfying an equation you derived but not the one you were given: a step widened the domain.

Formulas and theorems

  • Exponential function, and the constant-ratio test

    f(x)=abx,f(x+s)f(x)=bsf(x) = a \cdot b^{x}, \qquad \frac{f(x+s)}{f(x)} = b^{s}

    Use when a0a \neq 0, b>0b > 0, b1b \neq 1. Test a table by DIVIDING neighbours: with spacing ss the constant quotient is bsb^{s}, the base only when s=1s = 1. Then b0=1b^{0} = 1, bn=1bnb^{-n} = \tfrac{1}{b^{n}}, bp/q=bpqb^{p/q} = \sqrt[q]{b^{\,p}} and bx>0b^{x} > 0 are forced.

    e.g. Outputs 3,12,483, 12, 48 at x=0,2,4x = 0, 2, 4: b2=4b^{2} = 4, so b=2b = 2.

  • Graph of y=abxh+ky = a \cdot b^{\,x-h} + k

    y=abxh+k,bxh=bhbxy = a \cdot b^{\,x-h} + k, \qquad b^{\,x-h} = b^{-h} \cdot b^{x}

    Use when Parent y=bxy = b^{x}: domain all reals, range y>0y > 0, points (0,1)(0,1) and (1,b)(1,b), asymptote y=0y = 0, increasing exactly when b>1b > 1. Shifted: asymptote y=ky = k, range y>ky > k for a>0a > 0, y<ky < k for a<0a < 0. Intercepts: substitute x=0x = 0, or solve y=0y = 0.

  • Equating exponents

    bu=bv    u=vb^{u} = b^{v} \iff u = v

    Use when b>0b > 0, b1b \neq 1, with no restriction on uu and vv. It fails at b=1b = 1 (12=151^{2} = 1^{5}) and needs ONE shared base: 3x=5x13^{x} = 5^{x-1} does not give x=x1x = x - 1.

  • Definition of a logarithm

    y=logbx    by=xy = \log_b x \iff b^{y} = x
    The logarithm is the exponential reflected across the line y = xTwo curves on one pair of axes. The exponential rises from just above the horizontal axis on the left to steeply upward on the right, never touching that axis. The logarithm climbs away from the vertical axis at the bottom and flattens to the right, never touching that axis. A dashed diagonal line y equals x runs between them. A dot at (0, 1) on the exponential and a dot at (1, 0) on the logarithm are joined by a short dashed segment that meets the diagonal at a right angle and is cut in half by it.(0, 1)(1, 0)y = bxy = logb xy = x
    Text description

    The exponential and the logarithm are mirror images across the line y = x, so their points, intercepts, and asymptotes all swap coordinates.

    Use when b>0b > 0, b1b \neq 1, x>0x > 0. Inverse of bxb^{x}: range all reals, crosses (1,0)(1,0), passes (b,1)(b,1), vertical asymptote x=0x = 0, increasing exactly when b>1b > 1. Free: logb1=0\log_b 1 = 0, logbb=1\log_b b = 1. An unknown BASE gives xk=Nx^{k} = N: keep only the positive root, never 11.

    e.g. log5125=2\log_5 \tfrac{1}{25} = -2, because 52=1255^{-2} = \tfrac{1}{25}.

  • Cancellation laws

    logb ⁣(bt)=t,blogbx=x\log_b\!\left(b^{t}\right) = t, \qquad b^{\log_b x} = x

    Use when The first for EVERY real tt; the second only for x>0x > 0, since otherwise logbx\log_b x does not exist.

    e.g. 9log35=(3log35)2=259^{\log_3 5} = \left(3^{\log_3 5}\right)^{2} = 25.

  • Product, quotient, and power rules

    logb(MN)=logbM+logbNlogb ⁣(MN)=logbMlogbNlogb ⁣(Mp)=plogbM\begin{gathered} \log_b(MN) = \log_b M + \log_b N \\ \log_b\!\left(\tfrac{M}{N}\right) = \log_b M - \log_b N \\ \log_b\!\left(M^{p}\right) = p\log_b M \end{gathered}

    Use when M>0M > 0 and N>0N > 0 SEPARATELY, not merely MN>0MN > 0; pp any real. Also logb1N=logbN\log_b \tfrac{1}{N} = -\log_b N, and logb ⁣(x2)=2logbx\log_b\!\left(x^{2}\right) = 2\log_b |x| is good for every x0x \neq 0. The power rule never fires on (logbM)p\left(\log_b M\right)^{p}.

    e.g. log26+log223=log24=2\log_2 6 + \log_2 \tfrac{2}{3} = \log_2 4 = 2.

  • Change of base

    logbM=logaMlogab=logMlogb=lnMlnb\log_b M = \frac{\log_a M}{\log_a b} = \frac{\log M}{\log b} = \frac{\ln M}{\ln b}

    Use when M>0M > 0, both bases positive and unequal to 11, which keeps logab0\log_a b \neq 0. Argument on top, old base underneath. Special case logbM=1logMb\log_b M = \dfrac{1}{\log_M b}, needing M1M \neq 1.

    e.g. log832=log232log28=53\log_8 32 = \dfrac{\log_2 32}{\log_2 8} = \dfrac{5}{3}.

  • Bracketing a logarithm

    bm<x<bm+1    m<logbx<m+1\begin{gathered} b^{m} < x < b^{m+1} \\ \implies m < \log_b x < m+1 \end{gathered}

    Use when Needs b>1b > 1; for 0<b<10 < b < 1 the logarithm decreases and the inequalities reverse. In base 1010, a positive integer NN has dd digits exactly when d1logN<dd - 1 \le \log N < d.

    e.g. 32<50<6432 < 50 < 64 gives 5<log250<65 < \log_2 50 < 6; log ⁣(240)12.04\log\!\left(2^{40}\right) \approx 12.04 means 1313 digits.

  • Constant-percent model

    A(t)=P(1+r)t=PbtA(t) = P(1 + r)^{t} = P\,b^{\,t}

    Use when P>0P > 0, b>0b > 0. Grows when b>1b > 1, holds at b=1b = 1, decays toward zero without reaching it when 0<b<10 < b < 1. Chained percent changes multiply their factors rather than adding.

    e.g. Falling 12%12\% a year gives A=P(0.88)tA = P(0.88)^{t}, not P(0.12)tP(0.12)^{t}.

  • Compounding nn times a year

    A=P(1+rn)ntA = P\left(1 + \frac{r}{n}\right)^{nt}

    Use when rr as a decimal (6%6\% is 0.060.06), tt in the same unit as rr. Divide the rate AND multiply the time; n=1n = 1 recovers A=P(1+r)tA = P(1+r)^{t}.

    e.g. 20002000 dollars at 6%6\% monthly for 55 years: 2000(1.005)602697.702000(1.005)^{60} \approx 2697.70.

  • Doubling time and half-life

    doubling: t=log2logbhalf-life: t=log12logbA(t)=P2t/dor  A(t)=P(12)t/h\begin{gathered} \text{doubling:}\ t = \frac{\log 2}{\log b} \\ \text{half-life:}\ t = \frac{\log \tfrac{1}{2}}{\log b} \\ A(t) = P \cdot 2^{\,t/d} \\ \text{or}\ \ A(t) = P\left(\tfrac{1}{2}\right)^{t/h} \end{gathered}

    Use when b>0b > 0, b1b \neq 1; doubling needs b>1b > 1, half-life 0<b<10 < b < 1. PP cancels first, so neither time depends on the starting amount. Both logarithms in the decay quotient are negative, so the time is positive. Elapsed time goes in the NUMERATOR, over dd or hh, and the two model forms are ALTERNATIVES, never equal.

    e.g. At 4%4\% a year, t=log2log1.0417.7t = \dfrac{\log 2}{\log 1.04} \approx 17.7 years.

  • Continuous growth model

    (1+1n)ne,A=Pert\left(1 + \frac{1}{n}\right)^{n} \longrightarrow e, \qquad A = Pe^{rt}

    Use when rr as a decimal, tt in the same unit. Those values rise with nn yet never reach 33, so PertPe^{rt} is a CEILING no schedule beats. Decay uses a negative exponent coefficient.

    e.g. 20002000 dollars at 5%5\% for 88 years: 2000e0.42983.652000e^{0.4} \approx 2983.65, against 2976.262976.26 quarterly.

  • Natural logarithm

    lnx=logexelnx=xln ⁣(ex)=x\begin{gathered} \ln x = \log_e x \\ e^{\ln x} = x \\ \ln\!\left(e^{x}\right) = x \end{gathered}

    Use when The first two need x>0x > 0; the third holds for every real xx. Base ee, so every logarithm rule carries over, with ln1=0\ln 1 = 0, lne=1\ln e = 1.

    e.g. e0.04t=2e^{0.04t} = 2 gives 0.04t=ln20.04t = \ln 2, so t17.33t \approx 17.33.

  • Exponentials outrun every polynomial

    2nn2for every integer n42^{n} \ge n^{2} \quad \text{for every integer } n \ge 4

    Use when For x0x \ge 0, y=2xy = 2^{x} and y=x2y = x^{2} meet exactly twice, at (2,4)(2,4) and (4,16)(4,16), and the parabola leads only between them. Higher degree merely delays the crossing.

    e.g. 210=10242^{10} = 1024 while 102=10010^{2} = 100.

Problem types, step by step

Find the exponential through two points

  1. Divide the two values; aa cancels, leaving a pure power of bb.
  2. Take the POSITIVE root for bb, then substitute back for aa.
  3. Check BOTH points: a wrong bb can still satisfy one.

e.g. f(2)=18f(2) = 18, f(5)=486f(5) = 486: b3=27b^{3} = 27, so b=3b = 3, a=2a = 2.

Evaluate a logarithm by hand

  1. Ask the defining question: this base to what power gives the argument?
  2. Write base and argument as powers of one common number, then equate exponents.
  3. An argument below 11 needs a negative exponent.
  4. Exponentiate your answer back to confirm it.

e.g. log84\log_8 4: 23y=222^{3y} = 2^{2}, so y=23y = \tfrac{2}{3}, and 82/3=48^{2/3} = 4.

Find the domain of a logarithmic expression

  1. Set every argument strictly greater than zero, never the variable itself.
  2. For a quadratic argument, factor and test the product's sign per interval.
  3. Intersect the conditions when several logarithms appear together.

e.g. log5 ⁣(x29)\log_5\!\left(x^{2} - 9\right) needs (x3)(x+3)>0(x-3)(x+3) > 0, so x<3x < -3 or x>3x > 3.

Expand or condense a logarithmic expression

  1. To expand: split the quotient, then the products, then bring exponents down; roots are fractional exponents.
  2. To condense: send coefficients up as exponents FIRST, since the other rules need a coefficient of 11.
  3. Gather added logarithms into one numerator, subtracted ones into one denominator.
  4. Evaluate whatever is exact, assuming every variable inside is positive.

e.g. log5 ⁣(25x3y)=2+3log5x12log5y\log_5\!\left(\dfrac{25x^{3}}{\sqrt{y}}\right) = 2 + 3\log_5 x - \tfrac{1}{2}\log_5 y.

Solve an exponential equation

  1. Isolate the power, then write both sides over one base and equate exponents if you can.
  2. Otherwise take log\log or ln\ln of both sides, both positive, and use the power rule to bring the unknown down.
  3. Divide by the constant logarithm, and give the exact form and the rounded decimal separately.

e.g. 47x1=3004 \cdot 7^{\,x-1} = 300: 7x1=757^{\,x-1} = 75, so x=1+log75log73.219x = 1 + \dfrac{\log 75}{\log 7} \approx 3.219.

Solve a logarithmic equation

  1. Write the domain first: every argument in the ORIGINAL equation must be positive.
  2. Condense each side to a single logarithm.
  3. For logbu=k\log_b u = k write u=bku = b^{k}; for logbM=logbN\log_b M = \log_b N set M=NM = N; then solve.
  4. Test every candidate in the ORIGINAL and discard any making an argument zero or negative.

e.g. log2x+log2(x2)=3\log_2 x + \log_2(x-2) = 3: x(x2)=8x(x-2) = 8 gives x=4x = 4 or 2-2; only 44 survives.

Solve an exponential equation that hides a quadratic

  1. Spot that one power is the square of the other, as 4x=(2x)24^{x} = \left(2^{x}\right)^{2}.
  2. Substitute uu for the smaller power and solve the ordinary quadratic.
  3. Undo the substitution; a root u0u \le 0 yields no xx, since an exponential is never zero or negative.
  4. Check each surviving xx in the original equation.

e.g. 4x32x4=04^{x} - 3 \cdot 2^{x} - 4 = 0: u=4u = 4 or 1-1, so 2x=42^{x} = 4 and x=2x = 2.

Build a growth or decay model and evaluate it

  1. Turn the wording into a factor: rises 30%30\% gives 1.301.30, falls 30%30\% gives 0.700.70.
  2. From two readings, divide them so the starting amount cancels, then take the root matching the elapsed time.
  3. Pick the matching form: per step, nn periods a year, continuous, or written from a doubling time or half-life.

e.g. 500500 grows to 800800 in 66 years: b=1.61/61.0815b = 1.6^{1/6} \approx 1.0815, about 8.15%8.15\% a year.

Find how long a quantity takes to reach a target

  1. Set the model equal to the target and divide by the starting amount, which cancels.
  2. Take a logarithm of both sides, choosing ln\ln when the base is ee.
  3. Bring the exponent down with the power rule, divide by the constant logarithm or by rr, then check by substituting back.

e.g. 60(12)t/12=1060\left(\tfrac{1}{2}\right)^{t/12} = 10: t12=log6log22.585\dfrac{t}{12} = \dfrac{\log 6}{\log 2} \approx 2.585, so t31.0t \approx 31.0 years.

Exam traps

  • Trap Splitting the logarithm of a sum: writing log(A+B)=logA+logB\log(A + B) = \log A + \log B.

    Fix The product rule wants a PRODUCT inside; a sum inside does not break up at all. Test it: log2(4+4)=3\log_2(4+4) = 3, while log24+log24=4\log_2 4 + \log_2 4 = 4.

  • Trap Reading logMlogN\dfrac{\log M}{\log N} as logMlogN\log M - \log N.

    Fix A quotient of logarithms is a change of base, equal to logNM\log_N M; the difference is logMN\log \tfrac{M}{N}. Compare log40log52.292\dfrac{\log 40}{\log 5} \approx 2.292 with log80.903\log 8 \approx 0.903.

  • Trap Skipping the check after combining logarithms, or discarding a candidate for being negative.

    Fix Combining widens the domain, so test each candidate in the original: every ARGUMENT must be positive. In log2 ⁣(x29)=4\log_2\!\left(x^{2} - 9\right) = 4 the root x=5x = -5 is genuine.

  • Trap Applying logb ⁣(x2)=2logbx\log_b\!\left(x^{2}\right) = 2\log_b x without knowing the sign of xx.

    Fix It is false for every x<0x < 0, so it silently drops the negative half of a solution set. Use 2logbx2\log_b |x| when xx could be negative.

  • Trap Putting the rate where the factor belongs.

    Fix "Grows 6%6\% a year" gives b=1.06b = 1.06. The model A=P(0.06)tA = P(0.06)^{t} describes something losing 94%94\% of itself yearly.

  • Trap Adding percent changes instead of multiplying the factors.

    Fix A 30%30\% rise then a 30%30\% fall is 1.30×0.70=0.911.30 \times 0.70 = 0.91, a 9%9\% net loss rather than a wash, and (1+r)(1r)=1r2<1(1+r)(1-r) = 1 - r^{2} < 1 for every r0r \neq 0.

  • Trap Dividing an inequality by logb\log b without flipping it, for a decaying model.

    Fix For 0<b<10 < b < 1, logb\log b is negative, so the inequality REVERSES: (0.75)t<0.05(0.75)^{t} < 0.05 gives t>log0.05log0.7510.41t > \dfrac{\log 0.05}{\log 0.75} \approx 10.41, so the first whole year is 1111.

  • Trap Reading 23x-2 \cdot 3^{x} as (6)x(-6)^{x}, or insisting every exponential has range y>0y > 0.

    Fix An exponent binds only to the base beneath it, so evaluate 3x3^{x} then multiply. A negative aa makes every value negative: y=2xy = -2^{x} stays below the axis.

  • Trap Mixing up ln\ln and log\log, or feeding a percent straight into erte^{rt}.

    Fix log100=2\log 100 = 2 but ln1004.605\ln 100 \approx 4.605. And 6%6\% enters as r=0.06r = 0.06: e6te^{6t} inflates one year's growth from about 1.061.06 to over 400400.

Chapter test Questions from across the chapter