The basic limit laws and what they require, direct substitution for polynomials and rational functions, the five algebraic moves for resolving an indeterminate quotient, one-sided laws for piecewise functions, and the Squeeze Theorem with the fundamental trigonometric limits it delivers.
Subject: Calculus I · 70 slides · symbolic lesson
Open the interactive version of this deck
Title
Calculus I · Section 2.3
From estimating limits to computing them
Objectives
Section 2.2 found limits by reading graphs and tables, which is slow and sometimes wrong. This lesson replaces guessing with rules: a short list of laws that let you compute a limit line by line, and a few algebraic moves for when those laws seem to fail.
OpenStax Calculus Volume 1, §2.3 The Limit Laws §2.3, pp. 140-152 — learning objectives 2.3.1 to 2.3.6
In the last section you found limits by staring at graphs and filling in tables of values. That works, but it is slow, and you saw that a table can mislead you when a function oscillates. This lesson gives you rules instead, so that most limits take a line or two of algebra and no guessing at all.
The rules come in layers. First, a handful of limit laws that let you split a complicated limit into simple pieces. Second, the observation that for polynomials and most rational functions those laws boil down to plugging the number in. Third, a toolkit for the cases where plugging in produces nonsense such as zero over zero.
The last part meets a function that no amount of algebra can simplify, and handles it by trapping it between two simpler functions. That idea, the Squeeze Theorem, gives you the most important trigonometric limit in the course.
Warm-up
\[ \lim_{x\to 1}\frac{x^2-1}{x-1} \]
Discussion prompt
In Section 2.2 you estimated this limit from a table. The function is undefined at x = 1. Why does that not stop the limit from existing? Answer in one sentence using the definition of a limit.
Write your sentence before you reveal the answer. If you wrote something like the function gets close to 2, you have the right picture, but the key word to find is near: a limit is about inputs near the point, not at it.
That idea is the hinge of this whole lesson. Over and over you will replace a function by a simpler one that differs from it at exactly one point, the point you are approaching. That replacement is legal only because the limit never looks at that point.
Keep the table in mind too. The values at 0.99 and 1.01 sit on either side of 2 and hug it. In a minute you will get the same answer in one line, by cancelling a factor, and the table becomes a check rather than the method.
Section
Part 1
Concept
Figure (svg): Two panels. Left: the line y equals x, with dots at inputs 1, 1.5, 1.8, 1.95 and 3, 2.5, 2.2, 2.05 closing in on the point (2, 2). Right: the horizontal line y equals 5, with dots at the same inputs, all at height 5.
\[ \lim_{x\to a} x = a \qquad \lim_{x\to a} c = c \]
These two results (Theorem 2.4) are the bricks. Every limit law builds bigger limits out of these two. Example 2.13 is exactly this: as x approaches 2, x approaches 2, and the constant 5 stays 5.
\[ \lim_{x\to 2} x = 2, \qquad \lim_{x\to 2} 5 = 5 \]
OpenStax Calculus Volume 1, §2.3 The Limit Laws §2.3, pp. 140-141 — Theorem 2.4 and Example 2.13
Everything in this lesson is built from these two facts, so it is worth seeing why they are true rather than memorising them. On the left, the function simply hands back its input. If the inputs close in on 2, the outputs do too, because they are the same numbers.
On the right, the function ignores its input completely and always returns 5. However you approach 2, the output is 5, so the limit is 5. Notice that the dots are at the same inputs in both panels; only the rule for turning an input into an output differs.
These look almost too simple to state. The point is that they are the bricks: the limit laws on the next slide let you build any polynomial out of x and constants using sums and products, and so these two facts, plus the laws, will give you the limit of every polynomial.
Concept
Suppose f and g both have limits at a, say L and M. Then combining the functions combines the limits in the same way (Theorem 2.5).
| law | statement | condition |
|---|---|---|
| sum and difference | lim (f ± g) = L ± M | both limits exist |
| constant multiple | lim c·f = cL | the limit of f exists |
| product | lim f·g = L·M | both limits exist |
| quotient | lim f/g = L/M | both exist and M ≠ 0 |
| power | lim fⁿ = Lⁿ | n a positive integer |
| root | lim ⁿ√f = ⁿ√L | if n is even, L ≥ 0 and f ≥ 0 |
Every law has the same hidden first condition: the limits on the right-hand side must exist. The laws say nothing when they do not.
OpenStax Calculus Volume 1, §2.3 The Limit Laws §2.3, p. 141 — Theorem 2.5
Read the table as a single idea stated six times: if you know the limits of the pieces, you know the limit of anything built from them with ordinary arithmetic. Sum, difference, constant multiple, product, quotient, power, root.
Now look at the condition column, because that is where the mistakes live. Every law assumes that the limits of f and g exist. The quotient law adds that the bottom's limit is not zero, since you cannot divide by zero. The root law adds that for an even root, such as a square root, the inside must not be negative.
When a condition fails, the law does not say the limit fails to exist. It says nothing at all. You then need another method, and most of this lesson is about those other methods.
Intuition
If f(x) is within a hair of L and g(x) is within a hair of M, then their sum is within two hairs of L plus M. Closeness survives adding, multiplying and dividing, as long as you never divide by something close to zero.
\[ f(x) \approx 1.5, \quad g(x) \approx 1 \;\Longrightarrow\; f(x)+g(x) \approx 2.5, \quad f(x)g(x) \approx 1.5 \]
The book omits the proofs; Section 2.5 builds the precise language (epsilon and delta) in which they are written.
The book leaves the proofs of the laws out, and so will we, but you should have a reason to believe them. Think of being close as being within a small error. If f is within a small error of 1.5 and g is within a small error of 1, then their sum is within the two errors added together of 2.5. As the errors shrink, so does their total.
Products work the same way, with a little more bookkeeping. Quotients also work, provided the bottom stays away from zero. If the bottom is close to zero, a tiny change in it can make the quotient enormous, and the closeness argument collapses. That is the intuition behind the quotient law's extra condition.
In Section 2.5 you will meet the precise definition of a limit, and these arguments become real proofs.
Notation
Annotate
On: \( \lim_{x\to a}\frac{f(x)}{g(x)} = \frac{\lim_{x\to a} f(x)}{\lim_{x\to a} g(x)} = \frac{L}{M} \quad (M \ne 0) \)
OpenStax Calculus Volume 1, §2.3 The Limit Laws §2.3, p. 141 — Theorem 2.5, quotient law
Step through the notes one at a time. The law turns one limit of a fraction into a fraction of two limits. That is only a useful move if each of those two limits is easy, and it is only a legal move if each exists and the bottom one is not zero.
The last note is the one to take seriously. Almost every interesting limit in this lesson has a denominator that heads to zero, which is exactly when this law refuses to help. When you see that, do not write the law down and hope; switch to one of the techniques in Parts 3 to 5.
A good habit is to check the bottom first. Put the point into the denominator before anything else. If you get a nonzero number, the quotient law applies and the problem is nearly finished.
Picture it
Figure (svg): Three curves on one set of axes near x equals 2: f rising through height 1.5, g a shallow parabola touching height 1, and their sum f plus g passing through height 2.5, with the three heights marked by dots on the dashed line x equals 2.
Near 2 the blue curve sits near 1.5 and the orange curve near 1, so the green curve, their sum, sits near 2.5. Adding functions adds their heights point by point, and so it adds their limits.
Look at the dashed vertical line at x equal to 2. The blue curve crosses it at height 1.5, the orange curve at height 1, and the green curve at height 2.5. That is not a coincidence: the green curve is defined as the blue plus the orange, so at every x its height is the sum of their heights.
Now move your eye a little to either side of the dashed line. The blue and orange heights change a little, and the green height changes by the sum of those small changes. So as x approaches 2, the green height approaches 1.5 plus 1.
That is the whole sum law, drawn. The other laws have the same flavour: the operation you do to the functions is the operation you do to their limits.
Worked example
Use the limit laws to evaluate the limit, naming each law.
\[ \lim_{x\to -3}(4x+2) \]
Apply the sum law
Why: Both pieces have limits, so the limit of the sum splits.
\[ = \lim_{x\to -3} 4x + \lim_{x\to -3} 2 \]
Apply the constant multiple law
Why: The 4 comes outside the first limit.
\[ = 4\cdot \lim_{x\to -3} x + \lim_{x\to -3} 2 \]
Apply the basic limit results
Why: The limit of x is the point; the limit of a constant is the constant.
\[ = 4(-3) + 2 \]
Simplify
Why: Negative twelve plus two.
\[ = -10 \]
Check with a nearby input
Why: At x equal to minus 2.999 the function is minus 9.996, already very close to minus 10.
\[ 4(-2.999) + 2 = -9.996 \approx -10 \]
OpenStax Calculus Volume 1, §2.3 The Limit Laws §2.3, p. 141 — Example 2.14
Do this once, slowly, so that you know exactly which law justifies each line. First the sum law splits 4x plus 2 into two limits. Then the constant multiple law pulls the 4 out. Then the two basic results from Theorem 2.4 turn the limit of x into minus 3 and the limit of 2 into 2.
Before each step, ask whether the new limits exist. Here they obviously do, but the habit is what protects you later when they do not.
Notice where you ended up: minus 10 is exactly what you get by putting minus 3 into 4x plus 2. The laws have proved that substitution is legitimate for this function. After Theorem 2.6 you will be allowed to jump straight there, but only because this chain of laws stands behind the jump. The check at minus 2.999 shows the values really are closing in on minus 10.
Worked example
\[ \lim_{x\to 2}\frac{2x^2-3x+1}{x^3+4} \]
Apply the quotient law
Why: Allowed because the bottom's limit, 8 plus 4, is 12, not 0.
\[ = \frac{\lim_{x\to 2}(2x^2-3x+1)}{\lim_{x\to 2}(x^3+4)} \]
Apply the sum and constant multiple laws
Why: Every limit from here on is as x approaches 2.
\[ = \frac{2\lim x^2 - 3\lim x + \lim 1}{\lim x^3 + \lim 4} \]
Apply the power law
Why: The limit of a power is the power of the limit.
\[ = \frac{2\left(\lim x\right)^2 - 3\lim x + \lim 1}{\left(\lim x\right)^3 + \lim 4} \]
Apply the basic limit results
Why: Each limit of x becomes 2; each constant stays.
\[ = \frac{2(2)^2 - 3(2) + 1}{(2)^3 + 4} \]
Simplify
Why: Eight minus six plus one, over eight plus four.
\[ = \frac{3}{12} = \frac14 \]
Check numerically
Why: Inputs either side of 2 give values either side of one quarter.
\[ x = 1.999: \; 0.24983 \qquad x = 2.001: \; 0.25017 \]
OpenStax Calculus Volume 1, §2.3 The Limit Laws §2.3, p. 142 — Example 2.15
The very first thing to do is check the denominator, because the quotient law is the first law you want to use. At 2, x cubed plus 4 is 12, not zero, so the law applies and the fraction splits into a limit on top and a limit below.
From there it is the same routine as the last example, just longer. The sum and constant multiple laws break each polynomial into single terms, the power law turns the limit of x squared into the square of the limit of x, and the basic results finish it off. The shorthand of writing the limit without the little x approaches 2 underneath is only to keep the lines on the screen; every limit is still as x approaches 2.
The answer is a quarter, and again it equals what you get by substituting 2 directly. The numerical check brackets a quarter from both sides, which is the honest test that a limit is right.
Worked example
\[ \lim_{x\to 6}(2x-1)\sqrt{x+4} \]
Apply the product law
Why: Both factors will turn out to have limits.
\[ = \lim_{x\to 6}(2x-1)\cdot\lim_{x\to 6}\sqrt{x+4} \]
Apply the root law
Why: A square root is an even root, so check that the inside tends to something nonnegative: it tends to 10.
\[ = \lim_{x\to 6}(2x-1)\cdot\sqrt{\lim_{x\to 6}(x+4)} \]
Apply the difference, sum and constant multiple laws
Why: Then the basic results put 6 in for x.
\[ = (2\cdot 6 - 1)\sqrt{6+4} \]
Simplify
Why: Eleven times the root of 10.
\[ = 11\sqrt{10} \approx 34.785 \]
Check with a nearby input
Why: At x equal to 6.001 the product is 34.793, close to 34.785 and closing in.
\[ (11.002)\sqrt{10.001} \approx 34.793 \]
OpenStax Calculus Volume 1, §2.3 The Limit Laws §2.3, p. 142 — Checkpoint 2.11
This checkpoint brings in the two laws not used yet: the product law and the root law. The product law splits the expression into the limit of 2x minus 1 times the limit of the square root.
The root law carries a condition that is easy to skip. A square root is an even root, so the law needs the inside to tend to something that is not negative. Here the inside, x plus 4, tends to 10, so you are safe, and the limit can move inside the root.
After that the basic results give 11 times the root of 10, about 34.785. The check at 6.001 gives 34.793, a little above, as you would expect because the function is increasing there. In your own write-ups, name each law as you use it; that is what the book asks for in Exercises 83 to 86, and it is the fastest way to catch a step that is not allowed.
Trap
A tempting line:
\[ \lim_{x\to 0} x\cdot\frac1x \]
\[ = \lim_{x\to 0}x \cdot \lim_{x\to 0}\frac1x \]
\[ = 0 \cdot (\text{huge}) = 0 \]
Wrong. The second line is not allowed.
The product law needs both limits to exist, and one over x has no limit at 0. Simplify first, for x not equal to 0:
\[ x\cdot\frac1x = 1 \]
\[ \lim_{x\to 0} 1 = 1 \]
This mistake is tempting because each individual step looks like a law. The product law does say the limit of a product is the product of the limits. But it says so only when both of those limits exist, and one over x has no limit at zero: it runs off to plus infinity on one side and minus infinity on the other.
Writing zero times huge equals zero is the giveaway. Zero times something enormous can be anything, which is why the law will not touch it. The correct move is to simplify first: for every x other than zero, x times one over x is exactly 1, and the limit of the constant 1 is 1.
So the rule of thumb is: before splitting a limit with a law, make sure every piece you create has a limit of its own. If one does not, simplify the expression instead of splitting it.
Section
Part 2
Concept
Take any polynomial and push the limit through it with the laws. Nothing can go wrong: there is no division and no root.
\[ \lim_{x\to a}\left(c_n x^n + \cdots + c_1 x + c_0\right) \]
\[ = c_n\left(\lim_{x\to a} x\right)^n + \cdots + c_1\lim_{x\to a} x + \lim_{x\to a} c_0 \]
\[ = c_n a^n + \cdots + c_1 a + c_0 = p(a) \]
The laws have turned the limit into the value. For a quotient of polynomials the quotient law adds one condition: the bottom must not be 0 at a.
OpenStax Calculus Volume 1, §2.3 The Limit Laws §2.3, p. 142 — Theorem 2.6 and its proof
This slide turns the pattern you noticed in Examples 2.14 and 2.15 into a theorem. Take a general polynomial, with coefficients c and powers of x. The sum law splits it into terms, the constant multiple law pulls out each coefficient, and the power law turns the limit of x to the n into the limit of x, raised to the n.
Every limit of x is a, so every term becomes the coefficient times a power of a, and the whole expression is the polynomial evaluated at a. Nothing could have gone wrong along the way, because a polynomial involves no division and no roots, so none of the laws' conditions can fail.
For a rational function, a polynomial over a polynomial, you add the quotient law at the end, and that brings back one condition: the bottom polynomial must not be zero at a. That single condition is the dividing line between the easy limits and the interesting ones.
Notation
Annotate
On: \( \lim_{x\to a} p(x) = p(a), \qquad \lim_{x\to a}\frac{p(x)}{q(x)} = \frac{p(a)}{q(a)} \;\text{ when } q(a) \ne 0 \)
Go through the notes in order. The first statement says the limit of a polynomial is its value, at every real number. The second says the same for a rational function, with fine print.
The fine print is where you should slow down. If the denominator is zero at a, then substituting gives either zero over zero or a nonzero number over zero. Those are completely different situations. The first usually hides a finite limit that some algebra will reveal. The second always means the function is blowing up near a.
In practice this theorem means your first move on any limit is to substitute, because if it produces a number, that number is the answer and you are done.
Worked example
\[ \text{(a) } \lim_{x\to 3}\frac{2x^2-3x+1}{5x+4} \qquad \text{(b) } \lim_{x\to -2}(3x^3-2x+7) \]
Check the denominator in (a)
Why: Theorem 2.6 needs the bottom to be nonzero at 3.
\[ 5(3)+4 = 19 \ne 0 \]
Substitute into the numerator
Why: Eighteen minus nine plus one.
\[ 2(3)^2 - 3(3) + 1 = 10 \]
Write the limit in (a)
Why: The value is the limit.
\[ \lim_{x\to 3}\frac{2x^2-3x+1}{5x+4} = \frac{10}{19} \]
Substitute in (b)
Why: A polynomial: no condition to check.
\[ 3(-2)^3 - 2(-2) + 7 = -24 + 4 + 7 = -13 \]
Check (a) with a nearby input
Why: Ten nineteenths is about 0.52632; at x equal to 3.001 the function is 0.52665.
\[ \frac{10}{19} \approx 0.52632, \qquad f(3.001) \approx 0.52665 \]
OpenStax Calculus Volume 1, §2.3 The Limit Laws §2.3, p. 143 — Example 2.16 and Checkpoint 2.12
These two show what a finished problem looks like when Theorem 2.6 applies. For part (a), check the denominator first: 5 times 3 plus 4 is 19, which is not zero, so substitution is allowed. The numerator at 3 is 18 minus 9 plus 1, which is 10, so the limit is ten nineteenths.
Part (b) is a polynomial, so there is no condition to check at all. Be careful with the signs: minus 2 cubed is minus 8, times 3 is minus 24, and minus 2 times minus 2 is plus 4. Together with the 7 that makes minus 13.
The written answer can be a single line in each case. What earns full marks is the check on the denominator in part (a), stated before you substitute, because that is what makes the one line legitimate.
Sorting
Sort into buckets
Substitute the point into each expression. Sort by what comes out.
Substitute each point into each expression before you sort. You are not finding the limits yet, only sorting by what the first move produces, because that tells you which tool comes next.
Two expressions give ordinary numbers, a quarter and minus 13, and for those substitution is the whole answer. Two give zero over zero: in each, both top and bottom vanish at the point, which means there is more algebra to do and the limit could turn out to be any number. The last two give a nonzero number over zero, which is a different situation entirely: the quotient's size grows without bound.
Learn to see these three outcomes as three different road signs. A number means stop. Zero over zero means simplify. Nonzero over zero means think about infinity and signs.
Section
Part 3
Picture it
Figure (svg): Two panels. Left: the line g of x equals x plus 1, complete. Right: the same line for f of x equals x squared minus 1 over x minus 1, with an open circle at (1, 2) where f is undefined.
For every x except 1 the two functions agree, because the factor x minus 1 cancels. The limit never looks at x equal to 1, so it cannot tell them apart.
\[ \lim_{x\to 1}\frac{x^2-1}{x-1} = \lim_{x\to 1}\frac{(x-1)(x+1)}{x-1} = \lim_{x\to 1}(x+1) = 2 \]
OpenStax Calculus Volume 1, §2.3 The Limit Laws §2.3, pp. 143-144 — Figure 2.24
Compare the two panels. They are the same line, except that the right-hand one has a single missing point at (1, 2), because x squared minus 1 over x minus 1 is undefined at x equal to 1. Everywhere else the factor x minus 1 cancels and the two functions are identical.
Now ask what the limit at 1 sees. It looks at inputs near 1, never at 1 itself, and near 1 the two graphs are the same. So the two functions must have the same limit, and for the line x plus 1 the limit is easy: substitute 1 and get 2.
The chain of equals signs at the bottom is the model for the rest of this part. Factor, cancel the troublesome factor, and substitute into what remains. The hole is still there in the original function, but it has no effect on the limit.
Concept
Agreement principle — If f(x) equals g(x) for every x other than a in some open interval around a, then f and g have the same limit at a.
\[ f(x) = g(x) \text{ for all } x \ne a \;\Longrightarrow\; \lim_{x\to a} f(x) = \lim_{x\to a} g(x) \]
So the plan for a stuck quotient is always the same: find a simpler function that agrees with it everywhere except at a, then substitute into that.
OpenStax Calculus Volume 1, §2.3 The Limit Laws §2.3, p. 143 — the observation before Figure 2.24
This principle is the justification you should be able to give whenever you cancel inside a limit. If two functions agree at every point near a except possibly a itself, then they have the same limit at a. That follows straight from the definition: a limit only ever consults the values at points other than a.
It turns every zero over zero problem into a search. You are looking for a simpler function that agrees with yours except at a, and that you can evaluate at a by substitution. Factoring, conjugates and combining fractions are three ways of finding one.
Be clear about what cancelling does not do. It does not make the original function defined at a. It produces a different function, one that happens to have the same limit.
Picture it
Figure (svg): Three panels, each a quotient that gives zero over zero at x equals 2. Left: a line with a hole at height 4. Middle: the curve one over x plus 2 with a hole at height one quarter. Right: the curve one over x minus 2 splitting at a vertical asymptote at x equals 2.
Zero over zero is called an indeterminate form: it names the situation, not the answer. The limit can be a number, a different number, or not exist at all.
All three of these quotients give zero over zero at x equal to 2, and yet their limits are completely different. The first cancels to x plus 2, a line with a hole at height 4. The second cancels to one over x plus 2, a curve with a hole at height one quarter. The third cancels to one over x minus 2, which has a vertical asymptote at 2 and no limit at all.
This is what indeterminate means. The form zero over zero does not determine the answer; it only tells you that the top and bottom are racing each other to zero, and the result depends on which one gets there faster.
So never write zero over zero equals zero, or equals one, or does not exist. Write that the form is indeterminate, and then do the algebra that reveals the actual limit.
Intuition
If a polynomial is 0 at a, then x minus a is one of its factors. That is the factor theorem from algebra. So when both the top and the bottom vanish at a, both contain the factor x minus a, and it can be cancelled.
\[ p(a) = 0 \iff p(x) = (x-a)\,r(x) \]
\[ x^2 - 3x \text{ at } 3 = 0 \;\Longrightarrow\; x^2-3x = (x-3)\,x \]
Here is why factoring almost always works for rational functions. The factor theorem from algebra says a polynomial is zero at a exactly when x minus a is one of its factors. So if substituting a makes the top zero, the top contains x minus a. If it also makes the bottom zero, the bottom contains x minus a too.
That is a guarantee, not luck. When a rational function gives zero over zero at a, you know in advance which factor to look for, and you know it will cancel. In Example 2.17, the numerator x squared minus 3x is zero at 3, so it must contain x minus 3, and indeed it is x times x minus 3.
If after cancelling you still get zero over zero, the factor appeared more than once. Cancel it again.
Concept
OpenStax Calculus Volume 1, §2.3 The Limit Laws §2.3, p. 144 — Problem-Solving Strategy
This is the book's Problem-Solving Strategy, and it is worth memorising as three steps. First substitute, to confirm you really have zero over zero and are not looking at an easy limit or a nonzero-over-zero one.
Second, find a function equal to yours except at a. The shape of the expression tells you how. Polynomials over polynomials: factor both and cancel. A difference involving a square root: multiply top and bottom by the conjugate. A fraction inside a fraction: combine into a single fraction.
Third, apply the limit laws to the new function, which usually means substituting. If the new function still gives zero over zero, go around the loop again.
Worked example
\[ \lim_{x\to 3}\frac{x^2-3x}{2x^2-5x-3} \]
Substitute to identify the form
Why: Nine minus nine on top; eighteen minus fifteen minus three below.
\[ \frac{9-9}{18-15-3} = \frac{0}{0} \]
Factor the numerator
Why: Take out the common x.
\[ x^2 - 3x = x(x-3) \]
Factor the denominator
Why: It must contain x minus 3, since it is 0 at 3.
\[ 2x^2 - 5x - 3 = (x-3)(2x+1) \]
Cancel the common factor
Why: Allowed because x is not 3 anywhere the limit looks.
\[ \frac{x(x-3)}{(x-3)(2x+1)} = \frac{x}{2x+1} \]
Substitute into the simplified form
Why: Three over seven.
\[ \lim_{x\to 3}\frac{x}{2x+1} = \frac{3}{7} \]
Check the factoring and the value
Why: Expanding restores the denominator, and x equal to 3.001 gives 0.428592, next to 3/7 = 0.428571.
\[ (x-3)(2x+1) = 2x^2-5x-3, \quad f(3.001) \approx 0.428592 \]
OpenStax Calculus Volume 1, §2.3 The Limit Laws §2.3, p. 144 — Example 2.17
Start by substituting 3 to see the form. The top is 9 minus 9 and the bottom is 18 minus 15 minus 3, so you have zero over zero, and by the factor theorem both parts must contain x minus 3.
The numerator factors by pulling out x. For the denominator you already know one factor is x minus 3, so the other must multiply with it to give 2x squared and minus 3, which forces 2x plus 1. Cancelling x minus 3 leaves x over 2x plus 1, and substituting 3 gives three sevenths.
The check does two jobs. Multiplying x minus 3 by 2x plus 1 gives back the original denominator, so the factoring was right. And the original expression at 3.001 is 0.428592, right beside three sevenths, 0.428571. If those two numbers had disagreed, you would know to look for an algebra slip.
Fill the middle
\[ \lim_{x\to -3}\frac{x^2+4x+3}{x^2-9} \]
Fill in the blanks
\frac31/3})} = \frac______ \;\to\; ___
Why: The denominator is a difference of squares, x plus 3 times x minus 3. Cancelling x plus 3 leaves x plus 1 over x minus 3, and substituting minus 3 gives minus 2 over minus 6, which is one third.
OpenStax Calculus Volume 1, §2.3 The Limit Laws §2.3, p. 144 — Checkpoint 2.13
Fill in the blanks before checking. Both top and bottom vanish at minus 3, so both contain x plus 3. The top factors as x plus 3 times x plus 1. The bottom is a difference of squares, x plus 3 times x minus 3.
After cancelling x plus 3 you have x plus 1 over x minus 3. Substituting minus 3 gives minus 2 over minus 6, which is one third. A common slip is to lose a sign here and write minus one third; count the minus signs carefully.
If you want to check, try x equal to minus 2.999 in the original expression. You get about 0.33322, which is one third to three decimal places.
Error analysis
Annotate
On: \( \frac{x^2-3x}{2x^2-5x-3} = \frac{-3x}{x^2-5x-3} \quad \text{(cancel an }x^2\text{ top and bottom)} \)
Read the line and see whether you can spot the problem before revealing the notes. The fraction on the right looks simpler, and that is exactly the danger.
Cancelling is division. You may divide the top and the bottom by the same thing only when that thing multiplies the whole top and the whole bottom, which is what a factor is. An x squared that is merely added into the top and into the bottom is a term, not a factor, and striking it out changes the value of the fraction. Testing at x equal to 1 proves it: the two fractions give one third and three sevenths.
The protection is a routine: factor completely, then cancel only whole factors. And whenever you are unsure, put a convenient number into both versions. If they disagree, the step was illegal.
Worked example
A difference quotient: the shape every derivative in Chapter 3 will take.
\[ \lim_{h\to 0}\frac{(1+h)^2-1}{h} \]
Substitute to identify the form
Why: One minus one, over zero.
\[ \frac{1-1}{0} = \frac00 \]
Expand the square
Why: The only way to expose a factor of h.
\[ (1+h)^2 - 1 = 1 + 2h + h^2 - 1 \]
Simplify the numerator
Why: The ones cancel.
\[ = 2h + h^2 \]
Factor out h
Why: Now h multiplies the whole numerator.
\[ \frac{2h+h^2}{h} = \frac{h(2+h)}{h} \]
Cancel and substitute
Why: Legal since h is not 0.
\[ \lim_{h\to 0}(2+h) = 2 \]
Figure (svg): The parabola y equals x squared with the point (1, 1) marked, two secant lines from it to the points at x equals 1.8 and x equals 1.4, and a dashed tangent line of slope 2 through (1, 1).
Check with a small h
Why: At h equal to 0.001 the quotient is 2.001. It is the slope of a secant of y equals x squared, heading for slope 2.
\[ \frac{(1.001)^2-1}{0.001} = 2.001 \]
OpenStax Calculus Volume 1, §2.3 The Limit Laws §2.3, p. 153 — Exercise 96
This exercise is worth your attention because its shape, something like f of 1 plus h minus f of 1, all over h, is the shape of every derivative in Chapter 3. Substitution gives zero over zero, but nothing factors until you expand the square.
Expanding gives 1 plus 2h plus h squared, and subtracting 1 removes the constant. That cancellation is the crucial step: without it there would be no common factor of h. Once the numerator is 2h plus h squared, pull out h, cancel it with the bottom, and substitute h equal to 0 into 2 plus h.
The picture shows what the quotient measures. It is the slope of the line through (1, 1) and a second point on the parabola y equals x squared, h units to the right. As h shrinks, those secant lines swing toward the dashed tangent line of slope 2. The limit you just computed is that tangent slope.
Section
Part 4
Concept
A square root hides the factor that should cancel. Multiplying a difference by its sum squares the root away, because the middle terms cancel:
\[ (\sqrt{A} - B)(\sqrt{A} + B) = A - B^2 \]
\[ (\sqrt{x+2} - 1)(\sqrt{x+2} + 1) = x + 2 - 1 = x + 1 \]
The factor x plus 1 now appears in plain sight, ready to cancel with the denominator.
OpenStax Calculus Volume 1, §2.3 The Limit Laws §2.3, p. 144 — Problem-Solving Strategy, step 2b
When a square root is involved, factoring does not work directly, because the factor you need is hidden inside the root. The conjugate is the trick that brings it out. The conjugate of the root of A minus B is the root of A plus B: the same two pieces with the sign between them flipped.
Multiplying the two together is a difference of squares. The cross terms cancel and the root is squared away, leaving A minus B squared with no root in it. For the root of x plus 2 minus 1, that gives x plus 2 minus 1, which is x plus 1.
And x plus 1 is exactly the factor in the denominator of Example 2.18. The conjugate has turned a factor you could not see into one you can cancel.
Worked example
\[ \lim_{x\to -1}\frac{\sqrt{x+2}-1}{x+1} \]
Substitute to identify the form
Why: The root of 1, minus 1, over 0.
\[ \frac{\sqrt{1}-1}{-1+1} = \frac00 \]
Multiply top and bottom by the conjugate
Why: Multiplying by a fraction equal to 1 changes nothing.
\[ \frac{\sqrt{x+2}-1}{x+1}\cdot\frac{\sqrt{x+2}+1}{\sqrt{x+2}+1} \]
Multiply out the numerator only
Why: Leave the denominator factored; you hope x plus 1 will cancel.
\[ = \frac{x+1}{(x+1)\left(\sqrt{x+2}+1\right)} \]
Cancel
Why: The factor x plus 1 is not zero for x not equal to minus 1.
\[ = \frac{1}{\sqrt{x+2}+1} \]
Apply the limit laws
Why: Substitute minus 1.
\[ \lim_{x\to -1}\frac{1}{\sqrt{x+2}+1} = \frac{1}{1+1} = \frac12 \]
Figure (svg): The graph of the square root of x plus 2, minus 1, all over x plus 1, starting at (−2, 1) and falling slowly to the right, with an open circle at (−1, one half).
Check with a nearby input
Why: At x equal to minus 0.999 the original expression is 0.49988.
\[ \frac{\sqrt{1.001}-1}{0.001} \approx 0.49988 \]
OpenStax Calculus Volume 1, §2.3 The Limit Laws §2.3, p. 145 — Example 2.18
Substituting minus 1 gives the root of 1, minus 1, over 0, which is zero over zero. The numerator is a difference involving a square root, so reach for the conjugate: multiply top and bottom by the root of x plus 2, plus 1. That is multiplying by a fraction equal to 1, so the value does not change.
Here is the key tactical point: multiply out the numerator, but leave the denominator in factored form. The numerator becomes x plus 1, and you want to see that same factor sitting in the denominator so you can cancel it. Multiplying out the denominator would bury it.
After cancelling you have one over the root of x plus 2, plus 1, which is perfectly well behaved at minus 1 and gives one half. The graph shows the open circle at height one half, and the check at minus 0.999 gives 0.49988.
Prediction
\[ \lim_{x\to 5}\frac{\sqrt{x-1}-2}{x-5} \]
Predict first
Multiply by the conjugate and cancel. What is the limit?
Correct: 1/4
Why: Multiplying by the square root of x minus 1, plus 2, turns the top into x minus 1 minus 4, which is x minus 5. That cancels with the bottom, leaving one over the root of x minus 1 plus 2. At x equal to 5 this is one over 2 plus 2, a quarter. Zero is what you get if you only look at the top.
\[ \frac{(x-1)-4}{(x-5)(\sqrt{x-1}+2)} = \frac{1}{\sqrt{x-1}+2} \to \frac{1}{2+2} = \frac14 \]
OpenStax Calculus Volume 1, §2.3 The Limit Laws §2.3, p. 145 — Checkpoint 2.14
Commit to an answer before revealing. This is the same pattern as Example 2.18, so the conjugate is the root of x minus 1, plus 2.
Multiplying out the top gives x minus 1 minus 4, which is x minus 5, and that cancels with the bottom. What is left is one over the root of x minus 1 plus 2, and at 5 the root is 2, so the limit is one over 4.
If you chose zero, you probably noticed the top goes to zero and stopped there. The bottom goes to zero too, and it is the race between them that decides the answer. If you chose one half, check that you added the 2 inside the final denominator rather than using only the root.
Worked example
\[ \lim_{x\to 1}\frac{\frac{1}{x+1}-\frac12}{x-1} \]
Substitute to identify the form
Why: One half minus one half, over zero.
\[ \frac{\frac12 - \frac12}{0} = \frac00 \]
Multiply by 2(x+1) over itself
Why: That clears both small denominators at once.
\[ \frac{\frac{1}{x+1}-\frac12}{x-1}\cdot\frac{2(x+1)}{2(x+1)} \]
Multiply through the numerator
Why: Keep the denominator factored.
\[ = \frac{2-(x+1)}{2(x-1)(x+1)} \]
Simplify the numerator
Why: Two minus x minus one.
\[ = \frac{-x+1}{2(x-1)(x+1)} \]
Factor out minus 1
Why: This exposes x minus 1.
\[ = \frac{-(x-1)}{2(x-1)(x+1)} \]
Cancel and substitute
Why: Then put in x equal to 1.
\[ \lim_{x\to 1}\frac{-1}{2(x+1)} = \frac{-1}{2(2)} = -\frac14 \]
Check with a nearby input
Why: At x equal to 1.001 the original is minus 0.24988.
\[ \frac{\frac{1}{2.001}-\frac12}{0.001} \approx -0.24988 \]
OpenStax Calculus Volume 1, §2.3 The Limit Laws §2.3, pp. 145-146 — Example 2.19
Substituting 1 gives one half minus one half on top and zero on the bottom, so zero over zero. The top is itself a difference of fractions, which makes this a complex fraction, and the move is to clear the small denominators.
Multiplying top and bottom by 2 times x plus 1 does that in one step: on top, the first fraction becomes 2 and the second becomes x plus 1. Keep the bottom factored so that you can see what might cancel. The numerator simplifies to minus x plus 1, and factoring out minus 1 reveals minus the quantity x minus 1, which cancels with the x minus 1 below.
What remains is minus 1 over 2 times x plus 1, and at x equal to 1 that is minus one quarter. Watch the sign when you factor out the minus 1; losing it is the most common error here, and the check at 1.001, which comes out negative, would catch it.
Step zero
\[ \lim_{x\to -3}\frac{\frac{1}{x+2}+1}{x+3} \]
Discussion prompt
Before any algebra: what form does substitution give, and which of the strategy's moves does this expression call for? Then carry it out.
OpenStax Calculus Volume 1, §2.3 The Limit Laws §2.3, p. 146 — Checkpoint 2.15
Write your plan before you reveal. Substituting minus 3 into the top gives one over minus 1, plus 1, which is 0, and the bottom is 0 too. So the form is zero over zero, and the expression has a fraction sitting inside the numerator. That points to combining.
Rewrite 1 as x plus 2 over x plus 2 and add: the top becomes x plus 3 over x plus 2. Now the whole expression is x plus 3 over x plus 2, divided by x plus 3, and the factor x plus 3 cancels.
You are left with one over x plus 2, and at minus 3 that is one over minus 1, which is minus 1. Noticing the shape first, before doing any algebra, is what keeps these problems short.
Worked example
\[ \lim_{x\to 0}\left(\frac1x + \frac{5}{x(x-5)}\right) \]
Notice that neither piece has a limit
Why: Both blow up at 0, so the sum law does not apply.
\[ \frac{1}{0.001} = 1000, \qquad \frac{5}{0.001(0.001-5)} \approx -1000.2 \]
Combine over the common denominator
Why: Multiply the first fraction by x minus 5 over itself.
\[ \frac1x + \frac{5}{x(x-5)} = \frac{x-5+5}{x(x-5)} \]
Simplify the numerator
Why: The fives cancel.
\[ = \frac{x}{x(x-5)} \]
Cancel x
Why: Allowed since x is not 0.
\[ = \frac{1}{x-5} \]
Substitute
Why: Now the laws apply.
\[ \lim_{x\to 0}\frac{1}{x-5} = -\frac15 \]
Figure (svg): Three curves on the window from −2 to 3: one over x and five over x times x minus 5, both shooting off to infinity at x equals 0 in opposite directions, and their sum, one over x minus 5, passing smoothly through an open circle at (0, −0.2).
Check with the two pieces added
Why: At x equal to 0.001 the pieces are 1000 and about minus 1000.2, and their sum is minus 0.20004.
\[ 1000 - 1000.20004 = -0.20004 \approx -0.2 \]
OpenStax Calculus Volume 1, §2.3 The Limit Laws §2.3, p. 146 — Example 2.20
This example does not fit the patterns so far. It is a sum, so your first instinct is the sum law. But both pieces blow up at zero, as the first step's numbers show: one over x is 1000 at x equal to 0.001, and the second piece is about minus 1000.2. The sum law needs both limits to exist, so it cannot be used.
Look at the graph. The blue and orange curves run off in opposite directions at zero, and their blow-ups cancel almost perfectly. Their sum, the green curve, passes calmly through height minus one fifth. So the limit exists even though the pieces have none.
The algebra finds it: put both fractions over x times x minus 5, simplify the top to x, cancel x, and substitute into one over x minus 5. The check adds the two huge numbers and gets minus 0.20004, which is the whole story in one line.
Fill the middle
\[ \lim_{x\to 3}\left(\frac{1}{x-3} - \frac{4}{x^2-2x-3}\right) \]
Fill in the blanks
\frac11/4 = \frac______ = \frac______}} \;\to\; ___
Why: Factor the second denominator as x minus 3 times x plus 1, then write both fractions over it. The numerator becomes x plus 1 minus 4, which is x minus 3, and it cancels. What remains is one over x plus 1, and at x equal to 3 that is one quarter.
OpenStax Calculus Volume 1, §2.3 The Limit Laws §2.3, p. 146 — Checkpoint 2.16
Fill in the blanks before checking. The key observation is that x squared minus 2x minus 3 factors as x minus 3 times x plus 1, so it already contains the first denominator. That makes it the common denominator.
Rewrite the first fraction with that denominator by multiplying top and bottom by x plus 1. The combined numerator is x plus 1 minus 4, which is x minus 3, and it cancels with the x minus 3 below.
You are left with one over x plus 1, which at x equal to 3 is one quarter. As in Example 2.20, the two separate pieces each blow up at 3, so you had no choice but to combine them first.
Matching
Match the pairs
Why: The shape picks the move. A quotient of polynomials factors; a difference involving a square root wants its conjugate; a fraction inside a fraction wants one common denominator; a power of a binomial must be expanded before h can be seen as a factor. All four give 0/0, and their limits are 4, 1/4, −1/4 and 6.
Match each expression to its move before checking. The skill here is recognition: looking at the shape of an expression and knowing the technique before you start writing.
A quotient of polynomials calls for factoring. A difference involving a square root calls for the conjugate. A fraction sitting inside the numerator calls for combining over a common denominator. A binomial raised to a power, with h in the bottom, calls for expanding so that h can come out as a factor.
All four give zero over zero, and all four have finite limits: 4, one quarter, minus one quarter, and 6. Try each one after you match it; each takes three or four lines.
Section
Part 5
Concept
Every law has a one-sided version. For a limit from the left, the function only needs to be defined on an interval just to the left of a; for a limit from the right, just to the right.
\[ \lim_{x\to a^-} h(x): \; h \text{ defined on } (b, a) \qquad \lim_{x\to a^+} h(x): \; h \text{ defined on } (a, c) \]
And the two-sided limit exists exactly when both one-sided limits exist and are equal.
OpenStax Calculus Volume 1, §2.3 The Limit Laws §2.3, p. 146 — one-sided limit laws
One-sided limits obey the same laws, with one adjustment. For a limit from the left you only care about inputs slightly less than a, so the function only has to be defined on some interval just to the left of a. For a limit from the right, just to the right.
This matters in two situations you will meet constantly. The first is a function that simply does not exist on one side, like a square root at the edge of its domain. The second is a piecewise function, whose rule changes at a: each side then has its own formula, and you use the laws on that formula.
And remember the link to two-sided limits from Section 2.2. The two-sided limit exists exactly when both one-sided limits exist and agree, and then it is their common value.
Worked example
\[ \text{(a) } \lim_{x\to 3^-}\sqrt{x-3} \qquad \text{(b) } \lim_{x\to 3^+}\sqrt{x-3} \]
Find the domain
Why: The inside must be nonnegative.
\[ x - 3 \ge 0 \iff x \in [3, \infty) \]
Decide (a)
Why: There are no inputs just left of 3, so there is nothing for the limit to look at.
\[ \lim_{x\to 3^-}\sqrt{x-3} \text{ does not exist} \]
Decide (b) with the root law
Why: The inside tends to 0, which is allowed for an even root.
\[ \lim_{x\to 3^+}\sqrt{x-3} = \sqrt{\lim_{x\to 3^+}(x-3)} = \sqrt{0} = 0 \]
Figure (svg): The curve y equals the square root of x minus 3, starting at the point (3, 0) and rising to the right; the region left of x equals 3 is shaded red and marked as having no graph.
Check (b) with an input
Why: At x equal to 3.0001 the root is 0.01; the heights shrink to 0.
\[ \sqrt{3.0001-3} = \sqrt{0.0001} = 0.01 \]
OpenStax Calculus Volume 1, §2.3 The Limit Laws §2.3, pp. 146-147 — Example 2.21 and Figure 2.25
Start with the domain, because it decides everything. The square root needs x minus 3 to be at least zero, so the function lives only on the interval from 3 to infinity. The red region in the graph is where there is nothing at all.
For part (a), a limit from the left would need inputs just below 3, and there are none. The laws cannot be applied, and in fact the left-hand limit does not exist. For part (b), inputs just above 3 are fine, and the root law applies because the inside tends to zero, which is not negative. So the right-hand limit is the root of zero, which is zero.
The check shows how it approaches: at 3.0001 the function is 0.01. The graph starts at the point (3, 0) and rises steeply, so the heights shrink to zero as you slide down toward 3 from the right.
Worked example
\[ f(x) = \begin{cases} 4x-3 & x < 2 \\ (x-3)^2 & x \ge 2 \end{cases} \]
Left of 2, use the first rule
Why: For every x in the interval up to 2, f is 4x minus 3.
\[ \lim_{x\to 2^-} f(x) = \lim_{x\to 2^-}(4x-3) = 5 \]
Right of 2, use the second rule
Why: For every x beyond 2, f is x minus 3 squared.
\[ \lim_{x\to 2^+} f(x) = \lim_{x\to 2^+}(x-3)^2 = 1 \]
Compare the two sides
Why: Five and one are different.
\[ 5 \ne 1 \;\Longrightarrow\; \lim_{x\to 2} f(x) \text{ does not exist} \]
Figure (svg): A piecewise graph: the line 4x minus 3 for x less than 2, ending in an open circle at (2, 5); and the parabola x minus 3 squared for x at least 2, starting at a filled dot at (2, 1) and dipping to 0 at x equals 3.
Check with inputs on each side
Why: Just left of 2 the function is near 5; just right, near 1.
\[ f(1.999) = 4.996, \qquad f(2.001) = 0.998 \]
OpenStax Calculus Volume 1, §2.3 The Limit Laws §2.3, pp. 147-148 — Example 2.22 and Figure 2.26
The function has one rule for x less than 2 and another for x at least 2, so the two sides of 2 must be handled separately. For the left-hand limit, every input is less than 2, so f is 4x minus 3 there, and the limit is 8 minus 3, which is 5.
For the right-hand limit, every input is greater than 2, so f is x minus 3 squared, and the limit is minus 1 squared, which is 1. Look at the graph: the line ends at an open circle at height 5, and the parabola starts at a filled dot at height 1.
Since 5 and 1 are different, the two-sided limit does not exist. The inputs 1.999 and 2.001 are almost the same number, yet the function gives about 5 at one and about 1 at the other, which is exactly what a jump looks like.
Prediction
\[ f(x) = \begin{cases} -x-2 & x < -1 \\ 2 & x = -1 \\ x^3 & x > -1 \end{cases} \qquad \lim_{x\to -1^-} f(x) = ? \]
Predict first
What is the limit of f as x approaches −1 from the left?
Correct: −1
Why: From the left, x is less than minus 1, so only the first rule matters: minus x minus 2 at x equal to minus 1 is 1 minus 2, which is minus 1. The value 2 belongs to the point itself, which a limit ignores. The cubic also heads for minus 1 from the right, so even the two-sided limit is minus 1.
Figure (svg): A piecewise graph: the line minus x minus 2 for x less than minus 1, ending in an open circle at (−1, −1); an isolated filled dot at (−1, 2); and the cubic x cubed for x greater than minus 1, starting from an open circle at (−1, −1).
OpenStax Calculus Volume 1, §2.3 The Limit Laws §2.3, p. 148 — Checkpoint 2.17
Pick an answer before revealing. The only rule that matters for a left-hand limit at minus 1 is the one that holds for x less than minus 1, which is minus x minus 2. Substituting minus 1 gives 1 minus 2, which is minus 1.
The value 2 is the tempting wrong answer, because it is right there in the definition at x equal to minus 1. But it is the value at the point, and a limit ignores the value at the point. In the graph it is the lone filled dot, sitting well above where the curves are heading.
In fact the cubic on the right also heads for minus 1, since minus 1 cubed is minus 1. So both one-sided limits are minus 1 and the two-sided limit exists, even though the function's value there is 2. A limit and a value can disagree, and here they do.
Trap
For Example 2.22, someone reasons: f(2) uses the second rule, so
\[ \lim_{x\to 2^-} f(x) = (2-3)^2 = 1 \]
Wrong. The left-hand limit never uses the second rule.
Approaching from the left, x is always less than 2, and there f is 4x minus 3. The rule at the point itself is irrelevant to a limit.
\[ \lim_{x\to 2^-}(4x-3) = 5 \]
This error comes from letting the value at the point leak into the limit. The rule at exactly x equal to 2 is the second one, so it feels natural to use it. But the left-hand limit only ever looks at inputs slightly less than 2, and for those the first rule, 4x minus 3, is the one in force.
A reliable habit is to write the inequality before you write the rule: for x less than 2, f of x is 4x minus 3. Once that sentence is on the page, you cannot pick the wrong formula. Then do the same for the right side, and only at the end compare the two answers.
Concept
When the top tends to a nonzero number K and the bottom tends to 0, the quotient's size grows without bound. The only question is the sign on each side.
\[ \frac{f(x)}{g(x)}, \quad f(x) \to K \ne 0, \quad g(x) \to 0 \;\Longrightarrow\; \left|\frac{f}{g}\right| \to \infty \]
Figure (svg): A sign chart on a number line with marks at 0, 2 and 3. Rows give the signs of x minus 3, x, x minus 2, and the quotient on each interval; the interval from 0 to 2 is highlighted, where the quotient is positive.
A sign chart settles the sign: mark where each factor is zero, and read the signs on the interval next to the point.
OpenStax Calculus Volume 1, §2.3 The Limit Laws §2.3, p. 148 — the form K over 0
When substitution gives a nonzero number over zero, the situation is different from zero over zero. No cancellation can rescue it, because the top is not approaching zero. Dividing a number near K by a number near zero gives something enormous, so the size of the quotient grows without bound.
What is left to decide is the sign, and it can be different on the two sides. That is what the sign chart is for. Mark the points where each factor is zero, then read the sign of each factor on the interval you care about. For Example 2.23, between 0 and 2, the numerator x minus 3 is negative, x is positive, and x minus 2 is negative, so the quotient is negative over negative, which is positive.
So as x approaches 2 from the left, the quotient heads to plus infinity. From the right, between 2 and 3, the signs give a negative quotient, heading to minus infinity.
Worked example
\[ \lim_{x\to 2^-}\frac{x-3}{x^2-2x} \]
Substitute to identify the form
Why: Top minus 1, bottom 0.
\[ \frac{2-3}{4-4} = \frac{-1}{0} \]
Factor the denominator
Why: Find which factor is zero at 2.
\[ = \lim_{x\to 2^-}\frac{x-3}{x(x-2)} \]
Separate the factor that vanishes
Why: Only x minus 2 is zero at 2.
\[ = \lim_{x\to 2^-}\frac{x-3}{x}\cdot\frac{1}{x-2} \]
Take the limit of the tame part
Why: Minus 1 over 2.
\[ \lim_{x\to 2^-}\frac{x-3}{x} = -\frac12 \]
Take the limit of the wild part
Why: Just left of 2, x minus 2 is a tiny negative number.
\[ \lim_{x\to 2^-}\frac{1}{x-2} = -\infty \]
Multiply the signs
Why: A negative number times something going to minus infinity.
\[ \lim_{x\to 2^-}\frac{x-3}{x^2-2x} = +\infty \]
Check with inputs approaching 2 from the left
Why: The values are positive and growing tenfold with each step.
\[ f(1.99) \approx 50.75, \qquad f(1.999) \approx 500.75 \]
OpenStax Calculus Volume 1, §2.3 The Limit Laws §2.3, p. 148 — Example 2.23
Substituting 2 gives minus 1 over 0: a nonzero number over zero, so the answer is infinite and the work is all about the sign. Factor the denominator as x times x minus 2 to see which piece is causing the trouble. Only x minus 2 is zero at 2.
Split the expression into a tame part, x minus 3 over x, which tends to minus one half, and a wild part, one over x minus 2. Just to the left of 2, x minus 2 is a tiny negative number, so one over it is a huge negative number, heading to minus infinity.
A negative number times something going to minus infinity goes to plus infinity. The numbers confirm it: 50.75 at 1.99, and 500.75 at 1.999, positive and growing about tenfold each time you move ten times closer. Writing plus infinity is a precise description of how the limit fails to exist, not a number.
Prediction
\[ \lim_{x\to 1}\frac{x+2}{(x-1)^2} \]
Predict first
Substitution gives 3/0. What is the two-sided limit?
Correct: +∞
Why: The top tends to 3, which is positive. The bottom is a square, so it is positive on both sides of 1 and shrinks to 0. A positive number over a tiny positive number is huge and positive from both sides, so the limit is plus infinity. The sides would disagree only if the bottom changed sign at 1, which a square cannot.
\[ f(0.99) = 29{,}900, \qquad f(1.01) = 30{,}100 \]
OpenStax Calculus Volume 1, §2.3 The Limit Laws §2.3, p. 148 — Checkpoint 2.18
Commit before revealing. Substitution gives 3 over 0, a nonzero number over zero, so the quotient blows up; the only question is the sign on each side.
The numerator tends to 3, which is positive. The denominator is a square, and a square of a nonzero number is always positive, whichever side of 1 you are on. So the quotient is positive and huge on both sides, and the two-sided limit is plus infinity.
The numbers show it: about 29,900 just left of 1 and 30,100 just right. Compare Example 2.23, where the denominator changed sign at the point and the two sides went to opposite infinities. A squared factor is what prevents that.
Section
Part 6
Concept
Figure (svg): Three curves pinched together at x equals 2: an upper curve h, a lower curve f, and between them an oscillating curve g; all three pass through the point (2, 1).
Squeeze Theorem (Theorem 2.7) — If f(x) is at most g(x), which is at most h(x), for all x other than a near a, and f and h both have limit L at a, then g also has limit L at a.
The algebraic moves cannot touch something like x times cosine x at 0. The squeeze does not simplify g at all; it only fences it in.
OpenStax Calculus Volume 1, §2.3 The Limit Laws §2.3, pp. 148-149 — Theorem 2.7 and Figure 2.27
The algebraic methods so far all work by rewriting the function. Some functions cannot be rewritten into anything simpler, such as x times cosine of x, or x squared times sine of one over x. The Squeeze Theorem handles them without simplifying at all.
Look at the picture. The green curve g wriggles, but it is always between the blue curve h above and the orange curve f below. Those two outer curves close together at x equal to 2 and both reach height 1 there. Whatever g does, it has no room left: it is pinned to height 1 as well.
The theorem makes that precise. You need the inequality to hold near a on both sides, and you need the two outer functions to have the same limit. Then the middle function has that limit too.
Notation
Annotate
On: \( f(x) \le g(x) \le h(x), \quad \lim_{x\to a} f = L = \lim_{x\to a} h \;\Longrightarrow\; \lim_{x\to a} g = L \)
Step through the three notes. The first is the fence: g lies between f and h. It does not have to hold everywhere, only near a, and it does not have to hold at a itself. But it must hold on both sides of a, or you have only squeezed a one-sided limit.
The second note is the one people forget. The fences must tend to the same number. If the lower fence tends to 0 and the upper to 1, the middle function can still wander anywhere between them. You will see exactly that failure on the counterexample slide.
The third note is the payoff. You learn the limit of g without evaluating g near a at all. All the work goes into finding good fences, and the usual source of fences is a simple bound such as cosine or sine lying between minus 1 and 1.
Worked example
\[ \lim_{x\to 0} x\cos x \]
Bound the wild factor
Why: Cosine always lies between minus 1 and 1.
\[ -1 \le \cos x \le 1 \]
Multiply through by the absolute value of x
Why: A nonnegative multiplier keeps the inequalities pointing the same way.
\[ -|x| \le |x|\cos x \le |x| \]
Replace the middle by x cos x
Why: It equals plus or minus the middle term, and both lie inside the same fence.
\[ -|x| \le x\cos x \le |x| \]
Find the limits of the fences
Why: Both absolute values tend to 0.
\[ \lim_{x\to 0}(-|x|) = 0 = \lim_{x\to 0}|x| \]
Apply the Squeeze Theorem
Why: The middle is forced to the same limit.
\[ \lim_{x\to 0} x\cos x = 0 \]
Figure (svg): The lines y equals the absolute value of x and y equals minus the absolute value of x forming a sideways X through the origin, with the curve y equals x cos x weaving between them and passing through the origin.
Check with a small input
Why: At x equal to 0.01 the product is 0.0099995, inside the fence of width 0.01.
\[ 0.01\cos(0.01) \approx 0.0099995 \]
OpenStax Calculus Volume 1, §2.3 The Limit Laws §2.3, pp. 149-150 — Example 2.24 and Figure 2.28
The factor cosine of x is the awkward one, but it is bounded: it always lies between minus 1 and 1. Multiplying that inequality by the absolute value of x, which is never negative, keeps the inequalities pointing the same way and fences the product between minus the absolute value of x and the absolute value of x.
The middle step deserves a moment. The product x times cosine x equals either the absolute value of x times cosine x, or minus that, depending on the sign of x. Both of those lie inside the same symmetric fence, so x times cosine x does too.
Now both fences tend to 0, so the Squeeze Theorem forces x times cosine x to 0. In the graph, the two lines form an X through the origin, and the green curve is trapped inside it. The check at 0.01 gives 0.0099995, comfortably inside a fence of half-width 0.01.
Tweak it
Parameter explorer
The curve is the absolute value of x to the power k, times sin(1/x). Slide k from 0 upward. For which k does the curve get squeezed to 0 at x = 0? At k = 2 this is Checkpoint 2.19.
\[ g(x) = |x|^{{k}}\sin\left(\frac1x\right) \]
\[ -x^2 \le x^2\sin\frac1x \le x^2 \;\Longrightarrow\; \lim_{x\to 0} x^2\sin\frac1x = 0 \]
OpenStax Calculus Volume 1, §2.3 The Limit Laws §2.3, p. 150 — Checkpoint 2.19
Start at k equal to 2, which is Checkpoint 2.19. The curve is squashed flat near the origin: however fast sine of one over x oscillates, the factor x squared crushes it. The fences are minus x squared and x squared, both tending to 0, so the limit is 0.
Now slide k down toward 0. The fence opens wider near the origin, but as long as k is positive it still closes to a point at x equal to 0, and the limit is still 0. You can see the oscillations getting taller, yet they are always pinched at the centre.
At k equal to 0 the fence stops closing: the curve is plain sine of one over x, swinging between minus 1 and 1 forever as x approaches 0. There is no limit. The slider shows the exact moment the squeeze stops working.
Counterexample
\[ -1 \le \sin\frac1x \le 1 \]
Discussion prompt
Sine of one over x is trapped between −1 and 1 for every x near 0. Why does the Squeeze Theorem not give it a limit at 0? Name the hypothesis that fails and say what the function does instead.
Write your answer before revealing. The function sine of one over x is certainly trapped: it is always between minus 1 and 1. It is tempting to think that a function trapped between two fences must have a limit.
But the Squeeze Theorem has two hypotheses, and only one of them holds here. The fences are the constants minus 1 and 1, and their limits are minus 1 and 1. They do not meet, so the theorem says nothing. And in fact the function has no limit: as x approaches 0, one over x runs through every multiple of pi, so the sine swings from top to bottom infinitely often.
This is the counterexample to remember for the phrase bounded implies a limit. Boundedness only keeps a function from blowing up. To force a limit you need the fence itself to shrink to a single value.
Concept
Figure (svg): A unit circle with an angle theta of about 0.9 radians at the centre. The point (cos theta, sin theta) is marked; a vertical blue segment of length sin theta drops from it to the x-axis, and the yellow arc of length theta runs along the circle from (1, 0) up to the point.
On the unit circle, sine of theta is the height of the point and theta is the length of the arc. A straight drop is shorter than the arc, so for small positive theta:
\[ 0 < \sin\theta < \theta \;\Longrightarrow\; \lim_{\theta\to 0^+}\sin\theta = 0 \]
Sine is odd, so the left side follows by symmetry. Then the identity cosine equals the root of 1 minus sine squared finishes cosine.
\[ \lim_{\theta\to 0}\sin\theta = 0, \qquad \lim_{\theta\to 0}\cos\theta = \lim_{\theta\to 0}\sqrt{1-\sin^2\theta} = 1 \]
OpenStax Calculus Volume 1, §2.3 The Limit Laws §2.3, pp. 150-151 — Figure 2.29, equations 2.16 and 2.17
Before you can squeeze sine theta over theta, you need the limits of sine and cosine themselves, and the unit circle gives them. The angle theta, measured in radians, is the length of the yellow arc. The height of the point, sine of theta, is the blue segment.
The blue segment drops straight down to the axis, while the arc curves around to the same point, so the segment is shorter. That gives sine of theta between 0 and theta for small positive angles, and squeezing with 0 and theta shows that sine of theta tends to 0 from the right. From the left, sine of minus theta is minus sine of theta, so the same argument works.
For cosine, use the identity that cosine equals the square root of 1 minus sine squared, valid for angles between minus a right angle and a right angle. Sine tends to 0, so the root tends to 1 by the root law.
Worked example
\[ \lim_{\theta\to 0}\frac{\sin\theta}{\theta} \]
Figure (svg): A unit circle with an angle theta of 0.8 radians. The blue segment sin theta drops from the circle to the x-axis; the yellow arc theta runs along the circle; the red segment tan theta rises from (1, 0) up the tangent line to where the extended radius meets it.
Compare three lengths
Why: For a small positive angle: the vertical drop, the arc, the tangent segment, in that order.
\[ \sin\theta < \theta < \tan\theta \]
Divide by sine theta
Why: It is positive, so the inequalities keep their direction.
\[ 1 < \frac{\theta}{\sin\theta} < \frac{1}{\cos\theta} \]
Take reciprocals
Why: Flipping positive quantities reverses the inequalities.
\[ 1 > \frac{\sin\theta}{\theta} > \cos\theta \]
Squeeze from the right
Why: Both fences tend to 1.
\[ \lim_{\theta\to 0^+}\cos\theta = 1 = \lim_{\theta\to 0^+} 1 \]
Handle the left by symmetry
Why: Replacing theta by minus theta leaves the ratio unchanged.
\[ \frac{\sin(-\theta)}{-\theta} = \frac{-\sin\theta}{-\theta} = \frac{\sin\theta}{\theta} \]
Conclude
Why: Both one-sided limits are 1.
\[ \lim_{\theta\to 0}\frac{\sin\theta}{\theta} = 1 \]
Figure (svg): Three curves for theta from −1.5 to 1.5: the flat line at height 1, the curve cos theta below it, and between them the curve sin theta over theta, which has an open circle at (0, 1) where all three meet.
Check with numbers
Why: At theta equal to 0.1 the ratio is 0.998334, between cos 0.1 = 0.995004 and 1.
\[ 0.995004 < 0.998334 < 1 \]
OpenStax Calculus Volume 1, §2.3 The Limit Laws §2.3, p. 151 — Figure 2.30 and equation 2.18
This is the most important limit in the lesson, so follow each line. Substitution gives zero over zero, but no algebra will simplify sine of theta over theta. Instead, look at the figure: for a small positive angle, the blue drop, the yellow arc and the red tangent segment come in increasing order. The red segment is tangent of theta, because it is the height where the extended radius meets the line x equals 1.
Dividing all three by sine of theta, which is positive, gives 1, then theta over sine theta, then one over cosine theta. Taking reciprocals of positive quantities reverses the inequalities, and now sine theta over theta sits between cosine theta and 1. Both fences tend to 1, so the ratio does too from the right. The left side follows because the ratio is unchanged when theta is replaced by minus theta.
The second figure shows the squeeze directly: the green curve is caught between cosine and the line at 1, all meeting at the open circle. And the numbers agree: at a tenth of a radian the ratio is 0.998334, between 0.995004 and 1. The book's text here refers to Figure 2.30 adding a triangle to Figure 2.30; it means Figure 2.29, the previous unit circle.
Estimation
Predict first
Using radians, sin(0.1)/0.1 is closest to which value?
Correct: 0.998
Why: The ratio is 0.998334. Since the ratio sits between cosine of 0.1 (0.995) and 1, it cannot be 0.9 or 1.1, and a tenth of a radian is already small enough to get within two thousandths of 1.
| θ (radians) | 0.5 | 0.1 | 0.01 |
|---|---|---|---|
| sin θ / θ | 0.958851 | 0.998334 | 0.999983 |
| cos θ | 0.877583 | 0.995004 | 0.999950 |
Make your estimate before revealing. You do not need a calculator: the squeeze says the ratio lies between cosine of 0.1 and 1, and cosine of a tenth is about 0.995. That already rules out 0.9 and 1.1, and leaves 0.99 and 0.998 as candidates, with the true value just inside the upper end.
The table shows how fast the ratio approaches 1. At half a radian it is still about 0.96, at a tenth it is 0.998, and at a hundredth it is within about two hundred-thousandths of 1. The gap from 1 shrinks roughly like the square of theta, which is why halving the angle quarters the error.
This is also your first taste of an approximation that engineers use every day: for small angles in radians, sine of theta is very nearly theta.
Trap
Working on a calculator in degree mode:
\[ \lim_{x\to 0}\frac{\sin(x^\circ)}{x} = 1 \]
Wrong. The limit 1 was proved for radians only.
The proof used arc length equals angle, which is true only in radians. An angle of x degrees is pi x over 180 radians:
\[ \frac{\sin(x^\circ)}{x} = \frac{\sin\left(\frac{\pi x}{180}\right)}{x} \]
\[ \to \frac{\pi}{180} \approx 0.01745 \]
The limit of sine theta over theta equals 1 rests on one fact: that the angle theta equals the arc length on the unit circle. That is the definition of a radian. Measured in degrees, the angle is a different number, and the proof falls apart.
To see what really happens, convert: x degrees is pi times x over 180 radians. So sine of x degrees over x is sine of pi x over 180, divided by x, and by the radian limit that tends to pi over 180, about 0.01745, not 1.
This is why calculus always uses radians. With degrees, every formula for the derivatives of sine and cosine would carry an awkward factor of pi over 180. If a trig limit on your calculator comes out near 0.0175 instead of 1, check the angle mode.
Worked example
\[ \lim_{\theta\to 0}\frac{1-\cos\theta}{\theta} \]
Multiply by the conjugate
Why: One plus cosine over itself, aiming at a Pythagorean identity.
\[ = \lim_{\theta\to 0}\frac{1-\cos\theta}{\theta}\cdot\frac{1+\cos\theta}{1+\cos\theta} \]
Multiply out the numerator
Why: A difference of squares.
\[ = \lim_{\theta\to 0}\frac{1-\cos^2\theta}{\theta(1+\cos\theta)} \]
Use the identity
Why: One minus cosine squared is sine squared.
\[ = \lim_{\theta\to 0}\frac{\sin^2\theta}{\theta(1+\cos\theta)} \]
Split into two known limits
Why: One of them is the limit just proved.
\[ = \lim_{\theta\to 0}\frac{\sin\theta}{\theta}\cdot\frac{\sin\theta}{1+\cos\theta} \]
Apply the product law
Why: Both limits exist now.
\[ = 1\cdot\frac{0}{2} = 0 \]
Check with a small angle
Why: At theta equal to 0.01 the ratio is about 0.005, half of theta, and shrinking with it.
\[ \frac{1-\cos(0.01)}{0.01} \approx 0.0049999 \]
OpenStax Calculus Volume 1, §2.3 The Limit Laws §2.3, pp. 151-152 — Example 2.25 and equation 2.19
Substitution gives zero over zero, and none of the algebraic tricks apply directly, because there is no factor to cancel. The move is borrowed from the conjugate method: multiply by one plus cosine over itself. The numerator becomes one minus cosine squared, which the Pythagorean identity turns into sine squared.
Now split sine squared over theta times one plus cosine into two factors: sine theta over theta, which you just proved tends to 1, and sine theta over one plus cosine theta, which you can evaluate by substitution as 0 over 2. Both limits exist, so the product law applies, and the answer is 1 times 0, which is 0.
The check at a hundredth of a radian gives about 0.005, which is half of theta. That is not an accident: the ratio behaves like theta over 2 for small theta, which is a preview of something you will make precise with Taylor polynomials in Calculus II.
Intuition
Chapter 3 finds the derivative of sine as a limit of a difference quotient. After the sine addition formula, that quotient splits into exactly the two limits you now have:
\[ \frac{\sin(x+h)-\sin x}{h} = \sin x\cdot\frac{\cos h - 1}{h} + \cos x\cdot\frac{\sin h}{h} \]
\[ \to \sin x\cdot 0 + \cos x\cdot 1 = \cos x \]
So the fact that the derivative of sine is cosine rests on the unit-circle squeeze you just did.
It may seem odd to spend so long on two specific limits, but these are the ones that the whole of trigonometric calculus stands on. In Chapter 3 you will compute the derivative of sine directly from its definition, as the limit of sine of x plus h minus sine of x, all over h.
The addition formula for sine breaks that quotient into two pieces. One contains cosine of h minus 1 over h, which is minus the limit from Example 2.25, and tends to 0. The other contains sine of h over h, which tends to 1. What is left is cosine of x.
So every time you use the fact that the derivative of sine is cosine, you are using the unit-circle squeeze from this lesson. And because that squeeze needed radians, the derivative formula needs radians too.
Worked example
Inscribe a regular polygon of n sides in a circle of radius r, as Archimedes did, and let the triangles get thin.
Figure (svg): Two circles of radius r. The left one has a regular 8-sided polygon inscribed, with one triangle shaded: two sides of length r meeting at the centre at angle theta, height h to the middle of the base b. The right one has a 16-sided polygon, hugging the circle more closely.
Express h and b
Why: Half the vertex angle sits in each half of the triangle.
\[ h = r\cos\tfrac{\theta}{2}, \qquad b = 2r\sin\tfrac{\theta}{2} \]
Area of one triangle
Why: Use the double-angle identity to tidy it.
\[ \tfrac12 bh = r^2\sin\tfrac{\theta}{2}\cos\tfrac{\theta}{2} = \tfrac12 r^2\sin\theta \]
Relate n and theta
Why: The n angles fill a full turn.
\[ n\theta = 2\pi \;\Longrightarrow\; n = \frac{2\pi}{\theta} \]
Area of the polygon
Why: n triangles.
\[ A = \frac{2\pi}{\theta}\cdot\tfrac12 r^2\sin\theta = \pi r^2\cdot\frac{\sin\theta}{\theta} \]
Let the angle shrink
Why: The limit from equation 2.18.
\[ \lim_{\theta\to 0}\pi r^2\frac{\sin\theta}{\theta} = \pi r^2 \]
Check with Archimedes' polygon
Why: With r equal to 1 and 96 sides, the area is 3.13935, against pi, 3.14159.
\[ \tfrac{96}{2}\sin\tfrac{2\pi}{96} \approx 3.13935 \]
OpenStax Calculus Volume 1, §2.3 The Limit Laws §2.3, p. 152 — Student Project, Figure 2.31
This is the book's Student Project, done in full. Archimedes estimated pi by fitting regular polygons inside a circle. Each polygon of n sides is n identical isosceles triangles meeting at the centre, with two sides of length r and a vertex angle theta. The shaded triangle in the figure is one of them.
Cutting the triangle in half down its height gives two right triangles with angle half of theta, so the height is r cosine of half theta and the base is 2r sine of half theta. The triangle's area is half base times height, and the double-angle identity tidies it to half r squared sine theta. The n angles fill a full turn of 2 pi, so n is 2 pi over theta, and the polygon's area is pi r squared times sine theta over theta.
As the number of sides grows, theta shrinks to 0, and the limit you proved gives pi r squared. The check uses Archimedes' own choice of 96 sides: the area is 3.13935 for a unit circle, against pi at 3.14159. The figure shows the same trend with 8 and 16 sides.
Real world
Figure (svg): The line y equals theta and the curve y equals sin theta for theta from 0 to 1.2 radians; they are indistinguishable near 0 and separate slowly, with a dashed marker at 0.349 radians, which is 20 degrees.
Discussion prompt
Physicists replace sine of theta by theta for a swinging pendulum, which makes its equation solvable. Use the limit to explain why that is reasonable for small swings, and estimate the percentage error at a swing of 10 degrees.
Write your explanation and estimate before revealing. The limit says sine theta over theta tends to 1, so for small angles the two numbers are almost equal. In the pendulum equation that lets you replace sine of theta by theta, which turns an equation nobody can solve in simple terms into the equation of simple harmonic motion.
Convert the angle first: 10 degrees is about 0.1745 radians, and its sine is about 0.1736. The relative difference is about half a percent, small enough to ignore in most experiments. The graph shows why: up to about 20 degrees, the line and the sine curve are nearly on top of each other.
By 20 degrees the error has grown to about 2 percent, and it keeps growing. That is why physics books say the pendulum formula holds for small oscillations, and now you know exactly what small means.
Section
Part 7
Pattern
Figure (svg): A flow diagram. Start: substitute x equals a. Four arrows lead out: to 'a number: that is the limit'; to 'K over 0: infinite, use a sign chart'; to '0 over 0: factor, conjugate, combine, expand'; and to 'bounded times something going to 0: squeeze'.
This is the order to work in for every limit in this chapter. Substitution comes first, always, because it costs one line and its result tells you what to do next.
A number means you are done. A nonzero number over zero means the answer is infinite or does not exist, and a sign chart on each side settles which. Zero over zero means the limit is hiding, and the shape of the expression chooses the move: factor, conjugate, combine or expand. If the function is piecewise or you are at the edge of a domain, work each side separately.
And if no algebra helps but the function is bounded, fence it in with the Squeeze Theorem, or rearrange it into sine theta over theta. With this list you can do nearly every limit in the exercise set without a table.
Comparison
Comparison matrix
| what you see | what it means | the move |
|---|---|---|
| substitution gives 10/19 | Theorem 2.6 applies | stop |
| 0/0, polynomials | a shared factor x − a | factor |
| 0/0, √ in a difference | the factor is hidden | conjugate |
| 0/0, fraction on top | a complex fraction | combine |
| −1/0 | infinite | sign chart |
| x cos x at 0 | bounded times small | squeeze |
Fill in each blank before checking. Each row pairs a symptom with its cure, and the aim is to make the pairing automatic.
When substitution gives a number, stop. Zero over zero with polynomials means factor. Zero over zero with a square root in a difference means use the conjugate. A fraction inside a fraction means combine. A nonzero number over zero means the limit is infinite on at least one side. And a bounded factor multiplied by something going to zero is a squeeze.
Ranking
Put in order
Order the steps of Example 2.18.
Why: Substitution comes first because it may finish the problem. The conjugate is applied to top and bottom, only the top is multiplied out so that the factor stays visible, and only after cancelling may you substitute.
Drag the steps into order before checking. The one place people go wrong is the third step: multiplying out the denominator as well as the numerator. That hides the factor x plus 1 inside a longer expression and makes the cancellation hard to see.
The other trap is substituting before you cancel. Until the factor x plus 1 is gone, putting in minus 1 still gives zero over zero. Cancel first, then substitute into the simplified expression.
Check
Check your understanding
Suppose f(x) → 4, g(x) → 9 and h(x) → 6 as x → 6 (Exercise 108). What is the limit of (g(x) − 1)/f(x) as x → 6?
Answer: A
Why: The difference law gives 9 minus 1, which is 8, for the top; the bottom tends to 4, which is not 0, so the quotient law applies and the limit is 8 over 4, which is 2.
This is the style of Exercises 107 to 114: you are given the limits of three functions but not the functions themselves. The laws are exactly what you need, because they only ever use the limits of the pieces.
The top, g minus 1, tends to 9 minus 1, which is 8 by the difference law. The bottom tends to 4, which is not zero, so the quotient law applies, and the answer is 2. The most common slip is forgetting to subtract the 1 before dividing.
Check
Check your understanding
Checkpoint 2.20: what is the limit of (1 − cos θ)/sin θ as θ → 0?
Answer: A
Why: Multiply by one plus cosine over itself: the top becomes sine squared, so the quotient simplifies to sine theta over one plus cosine theta, which tends to 0 over 2, which is 0.
OpenStax Calculus Volume 1, §2.3 The Limit Laws §2.3, p. 152 — Checkpoint 2.20
This is Checkpoint 2.20. Substitution gives zero over zero, and the same conjugate trick as Example 2.25 works: multiply by one plus cosine over itself. The top becomes sine squared.
Cancelling one sine leaves sine theta over one plus cosine theta. That can be evaluated by substitution: 0 over 2, which is 0. No squeeze is needed at the last step, because the expression is now continuous at 0.
Check
Check your understanding
What is the limit of |x − 3|/(x − 3) as x → 3?
Answer: D
Why: For x greater than 3 the absolute value is x minus 3 and the quotient is 1; for x less than 3 it is minus the quantity x minus 3 and the quotient is minus 1. The one-sided limits are 1 and minus 1, which differ, so there is no two-sided limit.
Split the absolute value at the point where its inside changes sign. For x greater than 3, the absolute value of x minus 3 is just x minus 3, so the quotient is 1. For x less than 3, it is minus the quantity x minus 3, so the quotient is minus 1.
The one-sided limits are therefore 1 and minus 1. They disagree, so the two-sided limit does not exist. The graph of this function is two horizontal rays with a jump of 2 at x equal to 3.
Explain it to yourself
Discussion prompt
Cancelling x − 3 in Example 2.17 changes the function: the new one is defined at 3 and the old one is not. Explain in two sentences why the limit is nevertheless unchanged, naming the part of the definition of a limit that allows it.
Write your two sentences before revealing. This is the idea that makes most of the lesson work, so being able to state it clearly is worth more than any single calculation.
The answer rests on one phrase in the definition of a limit: x close to a, but not equal to a. Cancelling changes the function only at a, and the limit never looks at a. So the original function and the simplified one have identical values everywhere the limit looks, and therefore identical limits. The simplified one can be evaluated by substitution, and that value is the limit of both.
Exit ticket
\[ \lim_{x\to 4}\frac{\sqrt{x}-2}{x-4} \]
Discussion prompt
Identify the form, name the move, and evaluate the limit.
Try this one on your own before revealing, naming each step as you go. It uses the whole routine: substitute, recognise the form, choose the move, simplify, substitute again.
Substituting 4 gives the root of 4 minus 2, over 0, which is zero over zero. The numerator is a difference involving a square root, so multiply top and bottom by the root of x plus 2. The top becomes x minus 4, which cancels with the bottom, leaving one over the root of x plus 2. At 4 that is one over 4.
The check at 4.01 gives 0.24984, right beside a quarter. If you could do this in under two minutes, you have the central skill of the lesson.
Recap
| tool | use it when | watch out for |
|---|---|---|
| limit laws | the pieces have limits | every piece's limit must exist; quotient needs M ≠ 0 |
| direct substitution | polynomial, or rational with q(a) ≠ 0 | it is a theorem, not a habit |
| factor / conjugate / combine | substitution gives 0/0 | cancel factors, never terms |
| one-sided laws | piecewise functions, domain edges | use the rule that holds on that side |
| K/0 analysis | nonzero over zero | the sign on each side decides +∞ or −∞ |
| Squeeze Theorem | bounded oscillation, trig limits | the fences must share the limit |
\[ \lim_{\theta\to 0}\frac{\sin\theta}{\theta} = 1, \qquad \lim_{\theta\to 0}\frac{1-\cos\theta}{\theta} = 0 \]
Next, Section 2.4 names the property that made substitution work, continuity, and extends it far beyond polynomials.
Stewart, Calculus: Early Transcendentals 8e, §2.3 Calculating Limits Using the Limit Laws §2.3, pp. 95-103 — the same material in Stewart
The limit laws let you take a limit apart, as long as every piece has a limit and no denominator heads to zero. For polynomials, and for rational functions away from the zeros of their denominators, the laws reduce every limit to substitution.
When substitution gives zero over zero, the limit is hiding, and factoring, conjugates or combining fractions will find a function that agrees except at the point. When it gives a nonzero number over zero, the limit is infinite on each side with a sign you read from a sign chart. Piecewise functions are done one side at a time.
When nothing algebraic works, squeeze. That gave you the two trigonometric limits on this slide, which the derivative of sine will need in Chapter 3. Next, Section 2.4 gives a name to the property that made substitution work, continuity.
Want this taught 1-on-1? Alexander tutors Calculus I — $55/session, free consultation.