The meaning of a limit, one-sided limits, techniques for resolving the 0/0 indeterminate form, sin x / x, the Squeeze Theorem and vertical asymptotes. The first exam question almost always comes from here.
The limit is the basic concept of calculus. The derivative is a limit, the integral is a limit, and continuity is defined with a limit. On exams the limit questions are usually on the first page: quick points. But each one hides a small trap. This week’s algorithm is simple:
Substitute first. If you get a finite number, you are done.
If you get 00, simplify with algebra: factor, multiply by the conjugate, or build the xsinx pattern.
If you get 0c (c=0), the answer is ±∞ or “does not exist”. Check the sign from each side.
If there is an absolute value, a piecewise function or a root, compute the right and left limits separately.
If there is an oscillating expression (like sinx1), use the Squeeze Theorem.
The meaning of a limit
limx→af(x)=L
means “when x gets close enough to a, f(x) gets as close to L as you want”. The critical point: the value of f(a) does not affect the limit.f(a) may be undefined, or it may be some other number. The limit only looks arounda.
For example, f(x)=x−1x2−1 is undefined at x=1. But for x=0.999 it gives 1.999, and for x=1.001 it gives 2.001. The limit is 2.
One-sided limits
x→a−limf(x): as x approaches afrom the left (x<a).
x→a+limf(x): from the right (x>a).
limx→af(x)=L⟺limx→a−f(x)=limx→a+f(x)=L
Example 1. The function x∣x∣ is 1 for x>0 and −1 for x<0. The right limit is 1 and the left limit is −1. limx→0 does not exist.
Example 1b (piecewise function).limx→1f(x) for f(x)={x2+1,3x−1,x<1x≥1.
From the left the piece x<1 is used: 12+1=2. From the right the piece x≥1: 3−1=2. Both sides are equal, so the limit is 2. At a junction point always work like this: for each direction, use the formula of that direction.
Three cases where the limit does not exist
The left and right limits differ (a jump): x∣x∣, x→0.
Unbounded growth: x1, x→0. +∞ on one side, −∞ on the other.
Oscillation: sinx1, x→0. At the points x=nπ1 the value is 0, and at the points x=(4n+1)π2 it is 1. As it approaches 0, the function swings between −1 and 1 infinitely many times and never settles on a single number.
Guessing a limit with a numerical table is misleading in this third case. You see different numbers depending on which points you choose. A table gives an idea, not a proof.
Limit laws
If limf=L and limg=M: the limit of a sum is L+M, of a product LM, of a quotient L/M (if M=0), lim[f]n=Ln, limnf=nL. Consequence: for polynomials and rational functions (as long as the denominator is not zero), substitution is enough.
Example 2.x→2lim(x2+3x)=4+6=10.
All the laws work under the condition “if the limits exist”. If one of the two limits does not exist, you cannot use the sum law; this also does not mean that the limit of the sum does not exist. For example limx→0sinx1 does not exist and limx→0(−sinx1) does not exist, but their sum is 0 everywhere and its limit is 0.
The typical flow on an exam is:
substitute ├─ got a number -> answer ├─ c/0 (c ≠ 0) -> ±inf or does not exist: left/right sign analysis └─ 0/0 -> algebra: ├─ polynomial/rational -> factor ├─ difference of roots -> multiply by the conjugate ├─ sin, tan, cos -> sin(u)/u pattern └─ oscillation × 0 -> squeeze
Resolving 0/0
00 is not an answer, it is a sign that says “work more” (an indeterminate form). There are three main techniques.
1. Factoring
Example 3.x→−1limx+1x2+4x+3=x→−1limx+1(x+1)(x+3)=x→−1lim(x+3)=2.
If the denominator is zero as x→a and the numerator is also zero, both have a factor (x−a). Find it and cancel it.
2. Multiplying by the conjugate (rationalizing)
For differences that contain roots, multiply the expression A−B by A+B to get (A)2−B2.
Derived from it: x→0limxtanx=1 and x→0limx1−cosx=0.
Example 5.x→0limxsin4x=x→0lim4⋅4xsin4x=4.
Rule: the expression inside the sine must also appear exactly in the denominator. If it does not, create it by multiplying and dividing. Short result: bxsinax→ba.
Example 5b.x→0limsin5xsin3x. Put the numerator and denominator into the pattern separately:
For expressions with tan, write tanx=cosxsinx; since cosx→1, bxtanax→ba also holds.
PITFALL · the pattern does not work unless x→0.x→πlimxsinx=π0=0, not 1. The pattern holds only when the inside of the sine goes to zero.
The Squeeze Theorem
If g(x)≤f(x)≤h(x) near a and limg=limh=L, then limf=L.
The most common use: a bounded expression times an expression that goes to zero.
Example 6.x→0limx2cosx1.
cosx1 has no limit (it oscillates constantly). But −1≤cosx1≤1. Multiply by x2≥0:
−x2≤x2cosx1≤x2
Both ends go to 0, so the middle one also goes to 0.
PITFALL · using a limit law in the wrong place. You cannot write "limx2⋅limcosx1=0⋅(?)"; the product law requires both limits to exist. Use the squeeze.
Infinite limits and vertical asymptotes
If f(x) grows without bound as x→a, we write limf=+∞. This does not mean the limit “exists”; it describes the behavior. The line x=a is a vertical asymptote.
In the case 0c (c=0), look at the sign of the denominator to determine the sign:
Example 7.x→2−limx−2x+1. The numerator →3>0. For x<2, x−2<0. The result is 0−3=−∞. From the right it is +∞.
Example 8.x→3lim(x−3)21=+∞. The square is positive on both sides, so the two-sided limit is +∞.
Algorithm for finding vertical asymptotes (rational function):
1. Find the x values that make the denominator zero.2. Factor the numerator and denominator, and cancel common factors.3. A point that cancels -> a hole (removable), NOT an asymptote.4. A point that still makes the denominator zero after cancelling -> vertical asymptote.
The precise definition of a limit (epsilon-delta)
x→alimf(x)=L means: for every ε>0 there is a δ>0 such that 0<∣x−a∣<δ implies ∣f(x)−L∣<ε.
ε is “how much error I accept in the output”, and δ is “how close I must keep the input”. For linear functions the computation is short:
Example 9.x→1lim(3x+2)=5. ∣(3x+2)−5∣=3∣x−1∣. For this to be smaller than ε, it is enough that ∣x−1∣<ε/3. Choose δ=ε/3.
In Adams this definition comes with numerical questions of the type “find δ for ε=0.01”. Rearrange the difference so that it contains a factor ∣x−a∣, and divide by the coefficient.
Common mistakes
Wrong thinking
Correct
Got 00, the answer is 0 (or 1)
00 is indeterminate; simplify
Got 05, the answer is 0
Infinite limit; check the sign from each side
f(a) is undefined, so the limit does not exist either
The limit does not look at f(a), it looks around it
xsin3x→1
→3; the 3x inside must also be in the denominator
limx→πxsinx=1
The pattern is only for x→0; here it is 0
A point that cancels is also a vertical asymptote
A point that cancels is a hole
A single computation for an expression with ∣x−2∣
Compute separately for x<2 and x>2
limx2sinx1=0⋅limsinx1
The second limit does not exist; use the squeeze
Calculator in degree mode
xsinx→1 holds only in radians
Test yourself
Exam-style questions. Solve them on paper first, then reveal the answer.
Question 1easy
Evaluate x→3limx2−5x+6x2−9.
Show answer
6
Substitution gives 00. Factor: (x−3)(x−2)(x−3)(x+3)=x−2x+3 (x=3). Now substitute: 16=6.
Question 2easy
Evaluate x→4limx−4x−2.
Show answer
41
The 00 indeterminate form. Notice that x−4=(x−2)(x+2), or multiply by the conjugate: (x−2)(x+2)x−2=x+21. At x=4: 41.
Question 3easy
Evaluate x→0lim3xsin5x.
Show answer
35
Build the expression inside the sine into the denominator: 3xsin5x=35⋅5xsin5x. As x→0, 5x→0, and 5xsin5x→1. The result is 35.
Question 4medium
Evaluate h→0limh(2+h)3−8.
Show answer
12
Expand: (2+h)3=8+12h+6h2+h3. The numerator is 12h+6h2+h3; divide by h: 12+6h+h2. As h→0: 12. (This is the derivative of x3 at x=2; it will look familiar in week 5.)
Question 5medium
Evaluate x→0limx21−cosx.
Show answer
21
Multiply by the conjugate: x2(1+cosx)(1−cosx)(1+cosx)=x2(1+cosx)sin2x=(xsinx)2⋅1+cosx1. Limit: 12⋅21=21.
Question 6medium
Find x→2−limx2−4∣x−2∣ and x→2+limx2−4∣x−2∣. Does x→2limx2−4∣x−2∣ exist?
Show answer
−41 and 41; the limit does not exist
x2−4=(x−2)(x+2). From the left (x<2): ∣x−2∣=−(x−2), and the ratio x+2−1→−41. From the right (x>2): ∣x−2∣=x−2, and the ratio x+21→41. The one-sided limits differ, so the two-sided limit does not exist.
Question 7easy
Find x→1+limx−1x+2 and x→1−limx−1x+2.
Show answer
+∞ and −∞
As x→1 the numerator goes to 3 (positive). The denominator goes to zero. From the right, x−1 is a small positive number: 0+3=+∞. From the left, x−1 is a small negative number: 0−3=−∞. x=1 is a vertical asymptote.
Question 8medium
Find all vertical asymptotes of f(x)=x2+x−2x2−1.
Show answer
x=−2 only (x=1 is a removable hole)
f(x)=(x−1)(x+2)(x−1)(x+1). The denominator is zero at x=1 and x=−2. At x=1 the factor cancels: limx→1x+2x+1=32, which is finite, so this is a hole. At x=−2 the numerator is 3=0 and the denominator is zero: the limit is infinite, so it is a vertical asymptote. Answer: only x=−2.
Question 9hard
Use the Squeeze Theorem to find x→0+limxesin(π/x).
Show answer
0
−1≤sin(π/x)≤1 and et is increasing, so e−1≤esin(π/x)≤e. Multiply by x≥0: ex≤xesin(π/x)≤ex. As x→0+ both bounds go to 0. By the Squeeze Theorem the limit is 0.
Question 10medium
For x→3lim(4x−5)=7, find the largest δ>0 such that ∣x−3∣<δ guarantees ∣(4x−5)−7∣<0.01.
Show answer
δ=0.0025
∣(4x−5)−7∣=∣4x−12∣=4∣x−3∣. For 4∣x−3∣<0.01 it is enough that ∣x−3∣<0.0025. In general δ=ε/4.
Question 11hard
Evaluate x→0lim1−cosxxsinx.
Show answer
2
Multiply by the conjugate: 1−cos2xxsinx(1+cosx)=sin2xxsinx(1+cosx)=sinxx(1+cosx). As x→0: 1⋅2=2.
Question 12hard
Find the constant a for which x→2limx−2x2+ax−6 exists, and find the value of the limit.
Show answer
a=1, limit =5
The denominator goes to 0. For the limit to be finite, the numerator must also go to 0 (otherwise it is 0constant, which is infinite). 4+2a−6=0⇒a=1. Then x−2x2+x−6=x−2(x−2)(x+3)=x+3→5.
Verification: All limits were checked with node (verify/MAT1051/03.mjs): the function was evaluated on both sides of the point at distances 1e-4 to 1e-7; for infinite limits the size and sign of the value, and for the epsilon-delta question by scanning the interval.
Review card
First SUBSTITUTE. If you get a number, that is the answer.
0/0 -> factor / multiply by the conjugate / the sin x / x pattern
c/0 (c ≠ 0) -> infinite limit, check the sign from each side
lim sin x / x = 1 lim tan x / x = 1 lim (1 - cos x)/x^2 = 1/2 (x -> 0, radians)
sin(ax)/(bx) -> a/b
The limit exists <=> left limit = right limit
If you see |x - a|, compute the right and left limits separately
Squeeze: g <= f <= h, lim g = lim h = L implies lim f = L (x^2 sin(1/x) -> 0)
Vertical asymptote x = a: denominator 0, numerator ≠ 0 (if it cancels, it is a hole)
ε-δ: |f(x) - L| < ε when |x - a| < δ; for a linear function δ = ε/|slope|