10.5c Graphs of Tangent and Cotangent

The last two graphs, and the two that behave least like the rest. Both have period pi rather than two pi, both have range all real numbers, and each branch climbs or falls monotonically between consecutive asymptotes instead of turning around. Covers the proof that the period really is pi using the tangent sum identity, the geometric reason behind it, the different fundamental cycles and quarter marks the two functions use, and graphing transformed versions.

Subject: Trigonometry · 67 slides · symbolic lesson

Open the interactive version of this deck

What this lesson covers

The lesson, slide by slide

1. Graphs of Tangent and Cotangent

Title

Trigonometry · Section 10.5.3

Two curves with asymptotes, no amplitude, and half the usual period

2. What this lesson gives you

Objectives

Four of the six circular functions are now drawn. The last two are the ones that behave least like the rest: they are quotients rather than reciprocals, their branches climb instead of turning around, and they finish a cycle in half a turn rather than a full one.

Stitz & Zeager, College Trigonometry, Ch. 10 Foundations of Trigonometry — §10.5 Graphs of the Trigonometric Functions §10.5, pp. 804-808 — subsection 10.5.3

You have drawn four of the six graphs. The cosine and the sine were bounded and smooth, and the secant and the cosecant added asymptotes but kept the shape of a curve that turns around. These last two keep the asymptotes and throw away the turning.

The reason is structural. Secant and cosecant are reciprocals of bounded functions, so their values are bounded away from zero but never small. Tangent and cotangent are quotients of two bounded functions, and a quotient of two numbers that are both close to zero can be anything at all. That single difference produces an unbounded range, branches that climb steadily, and the one surprise of the section: a period of half a turn instead of a whole one.

The proof of that shorter period is worth the price of the lesson on its own, because it is the first time the identities you learned in Section 10.4 do a job that nothing else could do. A picture can suggest a period; only an identity can establish one.

3. Before anything new: what happens as the cosine dies?

Warm-up

Discussion prompt

As an angle rises towards a quarter turn, what happens to its sine, to its cosine, and therefore to their quotient? Now answer the same three questions for an angle falling towards a quarter turn from above.

Write your three answers down before revealing anything. The point of the question is that only one of the three quantities is doing anything interesting.

The sine is calm. As the angle rises towards a quarter turn the sine rises towards one and stops there, so the numerator of the quotient is near one and going nowhere. Everything that happens, happens underneath. The cosine falls to zero, and crucially it falls to zero from above on one side and climbs to zero from below on the other.

Divide a number near one by a positive number that is nearly zero and you get an enormous positive number. Do the same with a negative number that is nearly zero and you get an enormous negative one. So the quotient does not merely blow up; it blows up in opposite directions on the two sides. That is the whole reason the tangent graph looks the way it does, and every asymptote in this lesson is this same paragraph with different numbers.

4. The tangent curve

Section

Part 1

5. A quotient, not a reciprocal

Concept

Secant and cosecant were reciprocals of bounded functions, which is why their ranges were bounded away from zero. These last two are quotients of two bounded functions, and a quotient of two small numbers can be anything at all.

\[ J(x) = \tan(x) = \frac{\sin(x)}{\cos(x)}, \qquad K(x) = \cot(x) = \frac{\cos(x)}{\sin(x)} \]

Figure (svg): Three curves over zero to pi on the same axes: the sine, gently rising then falling; the cosine, crossing zero at pi over two; and the tangent, shooting up to the left of pi over two and returning from below on the right, with a dashed vertical line at pi over two.

The tangent is the top curve divided by the middle one. Where the divisor crosses zero the quotient has nowhere to go but off the page, and it changes sign as it goes.

Every difference between these graphs and the four you have already drawn comes from that structure: an unbounded range, branches that climb rather than turn, and a period of half a turn rather than a full one.

Stitz & Zeager, College Trigonometry, Ch. 10 Foundations of Trigonometry — §10.5 Graphs of the Trigonometric Functions §10.5, p. 804 — the opening of 10.5.3

Look at the three curves and keep the arithmetic in mind. The top curve is the numerator, the middle one is the denominator, and the steep one is what you get by dividing the first by the second at each input.

On the left of the picture the denominator is close to one, so the quotient is close to the numerator and the tangent hugs the sine. As you move right, the denominator shrinks and the quotient grows faster and faster. Exactly at the dashed line the denominator is zero and there is no quotient at all.

Then look just past the dashed line. The numerator has barely changed, but the denominator has changed sign, so the quotient reappears from far below instead of far above. Nothing continuous happens across that line; the graph is simply not there, and the two pieces on either side of it are separate branches.

6. Why the tangent runs off at a quarter turn

Worked example

Four one-sided limits fix the shape of the graph at its first two asymptotes.

\[ J(x) = \tan(x) = \frac{\sin(x)}{\cos(x)} \]

Approach a quarter turn from below

Why: The sine climbs towards one; the cosine falls to zero but stays positive.

\[ x \to \tfrac{\pi}{2}^{-}: \quad \sin(x) \to 1^{-}, \quad \cos(x) \to 0^{+} \]

Divide

Why: A number near one over a tiny positive number is enormous and positive.

\[ \tan(x) \to +\infty \]

Approach the same point from above

Why: The sine is still near one, but the cosine has gone negative.

\[ x \to \tfrac{\pi}{2}^{+}: \quad \sin(x) \to 1^{-}, \quad \cos(x) \to 0^{-} \]

Divide again

Why: Same size, opposite sign.

\[ \tan(x) \to -\infty \]

Repeat at three quarters of a turn

Why: There the sine is near minus one and the cosine runs from negative to positive, so both signs flip and the two limits swap.

\[ x \to \tfrac{3\pi}{2}^{-}: \tan(x) \to +\infty, \qquad x \to \tfrac{3\pi}{2}^{+}: \tan(x) \to -\infty \]

Check with a number on each side

Why: A quarter turn is 1.570796, so take one reading a ten-thousandth below it and one a ten-thousandth above.

\[ \tan(1.5707) \approx 10381, \qquad \tan(1.5709) \approx -9646 \]

Stitz & Zeager, College Trigonometry, Ch. 10 Foundations of Trigonometry — §10.5 Graphs of the Trigonometric Functions §10.5, p. 804 — the limit analysis before the table

This is the argument the book makes before it draws anything, and it is worth being able to reproduce. Each of the four limits is the same two-step move: decide what the numerator approaches, decide what the denominator approaches and from which side, then divide.

The notation with the little minus and plus signs is about direction of approach, not about sign of the answer. Approaching from below means the input is slightly less than a quarter turn; approaching from above means it is slightly more. The sign that matters for the answer is the sign the cosine has while it is shrinking, and that is the thing that flips.

At three quarters of a turn both the numerator and the denominator have changed sign compared with the earlier case, so the two limits swap over relative to what you might guess. Do not memorise the four results; redo the two-step move each time, and you will never get the direction of a branch wrong.

The numerical check is a habit worth building. A calculator set to radians will happily give you the tangent of 1.5707 and of 1.5709, and seeing one answer above ten thousand and the next below minus nine thousand makes the asymptote real in a way a limit symbol does not.

7. Building the tangent's table over a full turn

Worked example

Take the eighth marks of a full turn and divide the sine by the cosine at each one.

\[ \tan(x) \text{ at } x = 0, \tfrac{\pi}{4}, \tfrac{\pi}{2}, \tfrac{3\pi}{4}, \pi, \tfrac{5\pi}{4}, \tfrac{3\pi}{2}, \tfrac{7\pi}{4}, 2\pi \]

Start at zero and a quarter of a quarter turn

Why: At zero the numerator is zero; at the next mark the two coordinates are equal.

\[ \tan(0) = \frac{0}{1} = 0, \qquad \tan\!\left(\tfrac{\pi}{4}\right) = \frac{\sqrt2/2}{\sqrt2/2} = 1 \]

Hit the first asymptote

Why: The cosine is zero there, so the quotient does not exist.

\[ \tan\!\left(\tfrac{\pi}{2}\right) = \frac{1}{0}: \text{ undefined} \]

Continue past it

Why: In the second quadrant the sine stays positive while the cosine turns negative.

\[ \tan\!\left(\tfrac{3\pi}{4}\right) = \frac{\sqrt2/2}{-\sqrt2/2} = -1, \qquad \tan(\pi) = \frac{0}{-1} = 0 \]

Finish the second half

Why: In the third quadrant both coordinates are negative, so the quotient is positive again.

\[ \tan\!\left(\tfrac{5\pi}{4}\right) = 1, \quad \tan\!\left(\tfrac{3\pi}{2}\right) \text{ undefined}, \quad \tan\!\left(\tfrac{7\pi}{4}\right) = -1, \quad \tan(2\pi) = 0 \]

x0π/4π/23π/4π5π/43π/27π/42π
tan x01undefined−101undefined−10

Figure (svg): The tangent curve over zero to two pi, with filled dots at the seven defined eighth-marks and dashed vertical asymptotes at pi over two and three pi over two.

Seven points and two asymptotes. Read the heights from left to right: they run 0, 1, up and over, minus 1, 0 — and then exactly the same list again.

Check: look for the repeat

Why: The list from 0 to pi reads 0, 1, undefined, minus 1, 0. The list from pi to two pi reads 0, 1, undefined, minus 1, 0. It is the same list, so the pattern has already closed after half a turn.

\[ \tan(x + \pi) = \tan(x) \text{ at every mark in the table} \]

Stitz & Zeager, College Trigonometry, Ch. 10 Foundations of Trigonometry — §10.5 Graphs of the Trigonometric Functions §10.5, p. 804 — the table for J(x) = tan(x)

Every entry here comes from the unit circle values you already know, divided in pairs. There is nothing new to memorise; if you can write down the cosine and the sine at the eighth marks, you can write down this table.

Two kinds of entry deserve attention. A zero happens when the numerator dies, which is where the sine is zero, so at the multiples of a half turn. An undefined entry happens when the denominator dies, which is where the cosine is zero, so at the odd multiples of a quarter turn. Zeros and asymptotes never coincide, because the cosine and the sine are never both zero.

Now do what the check step does: read the list of nine entries out loud and listen for a repeat. The five entries from zero to a half turn are exactly the five entries from a half turn to a full turn. The table has already given away that the pattern closes twice as fast as you expected, and the next part of the lesson turns that observation into a proof.

8. Every branch climbs through every real number

Concept

Between two consecutive asymptotes the tangent rises steadily, from far below the axis to far above it, crossing zero exactly once on the way. It never turns around.

\[ \text{on } \left(-\tfrac{\pi}{2}, \tfrac{\pi}{2}\right): \quad \tan(x) \text{ increases from } -\infty \text{ to } +\infty \]

That last line is worth pausing on. A continuous function that increases from below every bound to above every bound must pass through every value in between, and it can only do so once.

Stitz & Zeager, College Trigonometry, Ch. 10 Foundations of Trigonometry — §10.5 Graphs of the Trigonometric Functions §10.5, pp. 804-806 — the graph and Theorem 10.25

Take one branch and follow it from left to right. It starts arbitrarily low just after an asymptote, rises, crosses zero at the middle of the branch, keeps rising, and disappears arbitrarily high just before the next asymptote. It never pauses and it never comes back down.

That monotone climb is what makes the range every real number. A continuous function that gets below any number you name and above any number you name must, somewhere in between, equal every number you name. Because it is strictly increasing, it does so exactly once per branch.

Hold on to the phrase exactly once. It is the reason a tangent equation has a single family of solutions instead of two, and it is the reason the inverse tangent in the next lesson is so much better behaved than the inverse sine.

9. No amplitude, no maximum, no midline

Intuition

Amplitude is half the distance between the highest value and the lowest value. Tangent has neither, so there is no distance to halve and the word simply does not apply.

The same was true of secant and cosecant, but for a different reason. Those two had a gap in the middle of their range and no ceiling above it; these two have no gap and no ceiling. Both ways, the quantity amplitude measures does not exist.

Nor is there a midline. A sinusoid oscillates about a horizontal line and you can find it by averaging the extremes; with no extremes there is nothing to average. What survives is the line through the zeros, which a vertical shift moves.

Students lose marks here not because they cannot graph the function but because they answer a question that has no answer. If a problem asks for the amplitude of a tangent, the correct response is that there is none, and saying so is worth full credit.

It helps to remember what amplitude is for. On a sinusoid it tells you how far the curve strays from its centre line, and you compute it by taking half the gap between the top and the bottom. Both of those numbers have to exist. Here neither does.

There is one genuine number in the same place in the formula, and it deserves its own name: the vertical stretch. It scales the three plotted heights away from or towards the horizontal line that the vertical shift sets, and if it is negative it also turns the picture over. Call it a stretch, never an amplitude, and you will always be right.

10. One branch, copied forever

Picture it

Figure (svg): The tangent curve across five branches from minus five pi over two to five pi over two, with dashed asymptotes between them and the middle branch, from minus pi over two to pi over two, drawn thicker on a shaded background.

The shaded strip is the fundamental cycle, and the arrow measures it: exactly pi wide. Slide that one branch left or right by a half turn and it lands exactly on the next one.

The highlighted branch is the one the book takes as the fundamental cycle. Every other branch is that one translated by a whole number of half turns, which is exactly what it means to say the period is a half turn.

Stitz & Zeager, College Trigonometry, Ch. 10 Foundations of Trigonometry — §10.5 Graphs of the Trigonometric Functions §10.5, p. 804 — the graph of y = tan(x)

The shaded branch is the one the book singles out as the fundamental cycle, and the choice is a convention, not a fact: any branch would do. What matters is that once you know one branch, you know the whole function, because every other branch is that one slid sideways by a whole number of half turns.

Look at how the branches meet the dashed lines. Every asymptote has a branch running up its left side and a different branch coming up from far below on its right. The two pieces never join, and it would be a mistake to draw a curve crossing an asymptote.

Count the asymptotes in the picture and you will find six of them evenly spaced, one every half turn. That spacing is the period, visible directly. Whenever you graph a transformed tangent later in the lesson, the distance between two consecutive asymptotes is the number you are looking for.

11. Reading the domain of the tangent

Notation

Annotate

On: \( x \ne \frac{\pi}{2} + \pi k, \quad k \in \mathbb{Z} \)

  • The first place the cosine vanishes on the positive side. Everything excluded is measured from here.
  • Steps of a half turn, forwards and backwards. The cosine vanishes every half turn, not every full turn, which is already a hint about the period.
  • k runs over all of the whole numbers, negative ones included, so the exclusions continue forever in both directions.
  • Stitz and Zeager write the same set as a union of open intervals, from (2k+1)π/2 to (2k+3)π/2. That form lists the branches one at a time instead of listing the holes.

\[ \text{domain}(\tan) = \bigcup_{k=-\infty}^{\infty}\left(\frac{(2k+1)\pi}{2}, \frac{(2k+3)\pi}{2}\right) \]

Stitz & Zeager, College Trigonometry, Ch. 10 Foundations of Trigonometry — §10.5 Graphs of the Trigonometric Functions §10.5, p. 806 — Theorem 10.25, the domain of J

Step through the annotations and notice that the whole domain is described by saying which inputs are missing, rather than which are present. That is usually the shorter description when the holes are evenly spaced.

The detail that repays attention is the step size. The excluded inputs are a half turn apart, not a full turn apart, even though the cosine is a full-turn function. The cosine vanishes twice per turn, at a quarter turn and at three quarters of a turn, and those two are half a turn apart. So the first hint that the period is halved is already sitting in the domain.

The union form in the last annotation is the same set described the other way round: instead of listing the holes, it lists the branches. Both forms turn up in textbooks and in answers, and you should be able to move between them. Section 10.7 will need the union form when it solves inequalities.

12. Trap: expecting the tangent to turn around

Trap

The trap

Asked for the largest value this function takes, a student answers three:

\[ y = 3\tan(x) \]

\[ \max = 3? \]

Wrong. Nothing here has a largest value.

The fix

The three is reading the formula as though it were a sinusoid, where the number in front really is the largest height. Here the factor of three multiplies a quantity that already runs through every real number, and three times something unbounded is still unbounded.

\[ \text{range}(3\tan) = (-\infty, \infty) \]

A quick sanity check: the tangent of 1.57 is already 1255, so three times it is far past three.

This mistake comes from good instincts applied to the wrong family. In a sinusoid the number in front really is the maximum height above the centre, so reading it off is the right move there. Here the number in front multiplies something that has no maximum, and no amount of multiplying creates one.

The defence is to think about the range before you think about the formula. Ask yourself first what values the parent function takes, and only then ask what the transformation does to them. Multiplying every real number by three gives you every real number back.

The numerical check at the bottom is the fastest way to expose the error in an exam. If someone claims a maximum of three, find a single input where the function is bigger than three. The tangent of 1.57 alone is more than a thousand.

13. A tenth of a radian across the asymptote

Prediction

\[ \tan(1.5) \approx 14.1 \]

Predict first

Roughly what is the tangent of 1.6 radians?

  • About 14.9
  • About 15.5
  • About negative 34
  • About 0.03

Correct: About negative 34

Why: A quarter turn is about 1.5708 radians, so 1.5 sits just below the asymptote and 1.6 just above it. Crossing an asymptote throws the value from very large and positive to very large and negative, and the tangent of 1.6 is about negative 34.2. Nothing continuous happens across an asymptote, so no answer near 14 can be right.

Commit to one of the four options before revealing. Three of them are traps of the same kind: they assume that a small change in the input causes a small change in the output.

A quarter turn is 1.5708 to four places, so the two inputs in this question straddle it. One of them is a thousandth of a radian below the asymptote and the other is a hundredth above it. Values on opposite sides of an asymptote have nothing to do with each other, and in fact they have opposite signs.

There is a useful estimate hiding here. Near the asymptote the tangent is close to the reciprocal of the distance to the asymptote, with a sign taken from which side you are on. The input 1.6 is about 0.0292 past the asymptote, and one divided by 0.0292 is about 34. That matches the answer, and the next slide makes the estimate official.

14. How close to the asymptote before the tangent passes one hundred?

Estimation

\[ \tan\!\left(\tfrac{\pi}{2} - h\right) = \cot(h) = \frac{1}{\tan(h)} \approx \frac{1}{h} \quad \text{for small } h \]

Predict first

How close to a quarter turn must an angle be before its tangent exceeds one hundred?

  • Within about a tenth of a radian
  • Within about a hundredth of a radian
  • Within about a thousandth of a radian
  • Within about a millionth of a radian

Correct: Within about a hundredth of a radian

Why: For a small gap the tangent of the angle is close to the reciprocal of the gap, so a tangent of one hundred needs a gap of about one hundredth of a radian. The exact gap is 0.010000 to six places. Every value from one hundred upward is crammed into that last hundredth of a radian, which is why the curve looks vertical there.

\[ \arctan(100) = 1.560797, \qquad \frac{\pi}{2} - 1.560797 = 0.010000 \]

Guess before you compute. Most people guess far too small a gap, because the curve looks vertical and a vertical line seems to need no width at all.

The estimate at the top of the slide is worth knowing. The tangent of a quarter turn minus a small gap is the cotangent of that gap, and for a small gap the cotangent is close to the reciprocal of the gap. So a tangent of one hundred needs a gap of about one hundredth, a tangent of a thousand needs a gap of about a thousandth, and so on.

Turn that round and you have a striking fact. Every value the tangent takes above one hundred, which is to say infinitely many values, is crammed into the last hundredth of a radian before the asymptote. The curve is not actually vertical anywhere; it is just that a graph drawn on paper cannot show you a hundredth of a radian.

15. Zero, undefined, or plus or minus one?

Sorting

Sort into buckets

Sort each angle by what the tangent does there.

Tangent is 0
0; π
Tangent is undefined
π/2; 3π/2
Tangent is 1
π/4; 5π/4
Tangent is −1
3π/4
z
0 and π are where the sine vanishes, so the numerator is zero and the quotient is zero.
u
π/2 and 3π/2 are where the cosine vanishes, so the denominator is zero and there is no value at all.
p
At π/4 both coordinates are √2/2; at 5π/4 both are −√2/2. Either way the quotient is 1.
n
At 3π/4 the coordinates are −√2/2 and √2/2, of equal size and opposite sign, so the quotient is −1.

Sort these by looking at the fraction, not by trying to remember a picture. For each angle, write down the cosine and the sine first, then divide.

Two buckets are easy once you know which part of the fraction dies. The numerator is the sine, so the tangent is zero wherever the sine is, which is at the multiples of a half turn. The denominator is the cosine, so the tangent is undefined wherever the cosine is zero, at the odd multiples of a quarter turn.

The remaining angles all have coordinates of equal size, so the quotient is one or minus one depending only on whether the two signs agree. In the first and third quadrants they agree and the answer is one; in the second and fourth they disagree and the answer is minus one. That is another way of seeing that the pattern repeats every half turn rather than every full turn.

16. Why the period is pi

Section

Part 2

17. The graph suggests a half turn; that is not a proof

Concept

Reading the table showed the same five entries twice, and the picture shows branch after identical branch. Neither is an argument: a graph is drawn from finitely many points, and eyes are easy to fool.

A period is defined as the smallest positive number you can add to the input without changing the output. Proving a number is the period therefore takes two separate arguments.

  1. Show a half turn works. Then the period is at most a half turn.
  2. Show nothing smaller works. Then the period is exactly a half turn.

The first part is one line of the sum formula from Section 10.4. The second is a short argument that almost every student skips, and it is the half that actually pins the number down.

Stitz & Zeager, College Trigonometry, Ch. 10 Foundations of Trigonometry — §10.5 Graphs of the Trigonometric Functions §10.5, p. 805 — the paragraph after the graph

It is tempting to feel the case is already closed. The table repeated and the picture shows identical branches, so what more could anyone want? The answer is that both of those are finite pieces of evidence about an infinite claim, and you have already seen in this course how misleading a plotted curve can be.

But there is a second, subtler gap, and it is the one that catches people. Even if you knew for certain that adding a half turn never changes the value, you would not yet know the period. The period is the smallest positive shift that works, and something smaller might work too.

So the proof has two halves, and they are proved by completely different methods. The first half is a computation with the sum formula. The second half is a short logical argument that starts by assuming a period exists and then pins it down. Plan on writing both whenever a problem asks you to find a period.

18. Proving that a half turn is a period of the tangent

Worked example

One application of the tangent sum formula, Theorem 10.16.

\[ \text{Show that } \tan(x + \pi) = \tan(x). \]

Apply the sum formula with the second angle equal to a half turn

Why: The formula turns a tangent of a sum into an expression in the two separate tangents.

\[ \tan(x + \pi) = \frac{\tan(x) + \tan(\pi)}{1 - \tan(x)\tan(\pi)} \]

Evaluate the tangent of a half turn

Why: The sine of a half turn is zero and the cosine is minus one.

\[ \tan(\pi) = \frac{0}{-1} = 0 \]

Substitute

Why: The second term of the numerator disappears and the product in the denominator does too.

\[ \tan(x + \pi) = \frac{\tan(x) + 0}{1 - \tan(x)\cdot 0} \]

Simplify

Why: The denominator is one.

\[ \tan(x + \pi) = \tan(x) \]

Check at a value where both sides are easy

Why: A quarter of a quarter turn gives one, and five quarters of a quarter turn gives one as well.

\[ \tan\!\left(\tfrac{\pi}{4}\right) = 1 = \tan\!\left(\tfrac{5\pi}{4}\right) \]

Stitz & Zeager, College Trigonometry, Ch. 10 Foundations of Trigonometry — §10.5 Graphs of the Trigonometric Functions §10.5, p. 805 — the sum-formula computation; Exercise 41 asks for the same identity

This is four lines and it is the payoff for learning the sum formula for tangents in Section 10.4. Set the second angle to a half turn and everything collapses, because the tangent of a half turn is zero.

Watch where that zero does its work. In the numerator it removes the added term, and in the denominator it removes the product, leaving a denominator of exactly one. Both simplifications come from the same fact, which is why the whole identity falls out in one move.

The check is not decoration. An identity claimed for all inputs should be spot-checked at a convenient one, and a quarter of a quarter turn is the obvious choice, since you know both tangents by heart. If a proof of yours ever fails a spot check, the proof is wrong, no matter how convincing it looked.

One caution about what has been proved. This says a half turn is a period. It does not yet say it is the period, and the very next slide is about the difference.

19. Proving that nothing smaller than a half turn works

Worked example

The previous slide left open the possibility of a shorter period. This closes it.

\[ \text{Suppose } p > 0 \text{ and } \tan(x + p) = \tan(x) \text{ for every } x. \]

Specialise the assumption at zero

Why: A statement about every input is in particular a statement about the input zero.

\[ \tan(0 + p) = \tan(0) \]

Evaluate the right-hand side

Why: The tangent of zero is zero.

\[ \tan(p) = 0 \]

Solve that equation

Why: The tangent vanishes exactly where its numerator does, which is at the multiples of a half turn.

\[ p = \pi k, \quad k \in \mathbb{Z} \]

Take the smallest positive one

Why: The multiples are 0, plus or minus pi, plus or minus two pi, and p was required to be positive.

\[ p_{\min} = \pi \]

Check that a quarter turn really fails

Why: If a quarter turn were a period, these two values would have to agree, and they do not.

\[ \tan\!\left(\tfrac{\pi}{4} + \tfrac{\pi}{2}\right) = \tan\!\left(\tfrac{3\pi}{4}\right) = -1 \ne 1 = \tan\!\left(\tfrac{\pi}{4}\right) \]

Stitz & Zeager, College Trigonometry, Ch. 10 Foundations of Trigonometry — §10.5 Graphs of the Trigonometric Functions §10.5, p. 805 — the minimality argument

This is the half of the proof that most students skip, and it is the half that actually answers the question. The structure is worth learning as a template: assume there is some period, specialise the assumption at a convenient input, and see what that forces.

Zero is the convenient input here, because the tangent of zero is zero and the equation collapses to a statement about the tangent of the shift itself. Once you know the tangent of the shift is zero, you know the shift is a multiple of a half turn, because those are exactly the places the tangent vanishes.

The shift was assumed positive, so the smallest candidate left standing is a single half turn, and the first half of the proof already showed that one works. Together the two halves say the period is a half turn, and nothing weaker.

The final check tests the most plausible rival. If a quarter turn were a period, then two particular values would have to agree, and they come out as one and minus one. A single failing example destroys a claim about all inputs, which is why one line is enough.

20. The half-turn the tangent cannot feel

Picture it

Figure (svg): A unit circle with a straight line through the origin at thirty-five degrees. The line meets the circle at a point P in the first quadrant and at the opposite point Q in the third quadrant, and meets the vertical line one unit to the right at a marked height.

P and Q sit on the same straight line through the origin, so they give the same quotient of second coordinate to first. Half a turn moves the point but not the line.

Adding a half turn to an angle sends its terminal point to the diametrically opposite point of the circle, which negates both coordinates. The cosine notices and the sine notices, but their quotient cannot, because the two minus signs cancel.

\[ \tan(\theta + \pi) = \frac{-\sin(\theta)}{-\cos(\theta)} = \frac{\sin(\theta)}{\cos(\theta)} = \tan(\theta) \]

Stitz & Zeager, College Trigonometry, Ch. 10 Foundations of Trigonometry — §10.5 Graphs of the Trigonometric Functions §10.5, p. 805 — the geometric reading of the result

This picture is the reason behind the algebra, and it is the thing to remember when the identity has faded. Adding a half turn to an angle sends its point on the circle to the point directly opposite through the centre, which negates both coordinates.

Now think about who notices. The cosine reports the first coordinate, so it sees a sign change. The sine reports the second, so it sees a sign change too. But the tangent reports only the ratio of the two, and a ratio is completely blind to a sign change applied to both of its parts.

The dashed line makes the same point in a second way. The two arrows point in opposite directions along one straight line through the origin, and a line has a single slope. Since the tangent is that slope, it cannot tell the two directions apart, and the dot on the vertical line at the right shows the shared value.

21. Fill the middle of the half-turn argument

Fill the middle

Let the terminal point of the angle be the point with first coordinate a and second coordinate b.

Fill in the blanks

\tan(\theta) = \frac-b-a \qquad \tan(\theta + \pi) = \frac___}___} = \frac______

Why: Adding a half turn sends the point to the diametrically opposite one, so both coordinates change sign. Two minus signs in a quotient cancel, which is exactly why the tangent repeats after half a turn while the sine and the cosine, which see only one coordinate each, need a whole turn.

Fill both blanks before checking. This is the algebra of the previous picture, and writing it yourself is the difference between recognising the argument and owning it.

The point that was at the pair a and b is now at the pair with both signs flipped, so the new quotient has a minus sign top and bottom. Two minus signs in a quotient cancel, exactly as they would in any fraction of ordinary numbers.

Then notice what this argument does not prove. It shows a half turn is a period. It says nothing about whether a quarter turn might also work, which is why the minimality argument earlier in this part still has to be made separately.

22. The tangent is the slope of the terminal side

Concept

There is a second reading of the same picture, and it is the one worth carrying away. The terminal point has first coordinate the cosine and second coordinate the sine, so the ratio of the two is the slope of the line from the origin through that point.

\[ \text{slope} = \frac{\text{rise}}{\text{run}} = \frac{\sin(\theta)}{\cos(\theta)} = \tan(\theta) \]

Figure (svg): A unit circle with three rays at twenty, forty-five and sixty-five degrees, each extended by a dashed segment to the vertical line one unit to the right, where three dots sit at heights 0.36, 1 and 2.14.

The tangent of an angle is the height at which its terminal line crosses the vertical line one unit to the right. Stand the ray up and that height has no ceiling.

A line has one slope, and a line through the origin is determined by either of the two directions along it. So half a turn returns you to the same line and therefore to the same tangent, while a vertical line, reached at a quarter turn, has no slope at all. That is the asymptote.

Stitz & Zeager, College Trigonometry, Ch. 10 Foundations of Trigonometry — §10.5 Graphs of the Trigonometric Functions §10.5, p. 805 — the fundamental cycle of the tangent

This reading of the tangent will pay for itself many times over. The point on the circle has first coordinate the cosine and second coordinate the sine, so the line from the origin to it rises by the sine over a run of the cosine. That is the slope, and that is the tangent.

Look at the three dashed segments running out to the vertical line. Each terminal ray, continued outwards, crosses the line one unit to the right at a height equal to its own tangent, which is where the name of the function comes from. As the ray stands up, that crossing point climbs, slowly at first and then faster and faster.

The limiting case explains the asymptote in one sentence. At exactly a quarter turn the ray is vertical, a vertical line has no slope, and it never meets the vertical line one unit to the right at all. There is no value to plot, which is exactly what an asymptote is.

23. Why only the tangent family halves its period

Intuition

Cosine reports the first coordinate and sine reports the second. A half turn flips the sign of what each of them reports, so each of them needs a second half turn to get its sign back. That is why their period is a whole turn, and why secant and cosecant, being their reciprocals, inherit it.

Tangent and cotangent report a ratio of the two coordinates. A ratio is blind to a sign change applied to both parts of it, so half a turn is already enough. Nothing else in the chapter behaves this way.

Stated as a slogan: the four that read a coordinate repeat every full turn, and the two that read a ratio of coordinates repeat every half turn.

It is worth having a one-line reason for why these two functions are the exception, because otherwise the halved period feels arbitrary and gets forgotten.

The reason is about what each function reads off the circle. Cosine reads the first coordinate and sine reads the second, and a half turn flips the sign of both. A function that reports a signed coordinate therefore needs a second half turn to recover, which is why four of the six repeat only after a full turn.

Tangent and cotangent report a ratio, and a ratio cannot see a sign change applied to both of its parts. Half a turn is already enough for them. If you remember nothing else about the period, remember the slogan: read a coordinate and you need a whole turn, read a ratio of coordinates and half a turn will do.

24. Find the error: inheriting the period from the parents

Error analysis

A student reasons about the period of the tangent from what the numerator and the denominator do.

Annotate

On: \( \cos, \sin \text{ repeat every } 2\pi \;\Longrightarrow\; \tan \text{ repeats every } 2\pi \)

  • True: both the cosine and the sine are unchanged by adding a full turn.
  • Also true, as far as it goes. A full turn IS a period of the tangent, since adding it changes neither the numerator nor the denominator.
  • A period is the SMALLEST positive shift that works. Showing that 2π works says nothing about whether something smaller also works, and here something smaller does.
  • A combination of two functions can repeat sooner than either of them. Domains and parity can be inherited from the parents; the period cannot.

Stitz & Zeager, College Trigonometry, Ch. 10 Foundations of Trigonometry — §10.5 Graphs of the Trigonometric Functions §10.5, p. 805 — why the minimality half of the proof is needed

Read the claim and try to say what is wrong with it before revealing the annotations. It is harder than it looks, because nothing in it is actually false.

The premise is true and the conclusion is true. A full turn really is a period of the tangent. What fails is the word period in its technical sense, which means the smallest positive shift that works. The argument shows only that a full turn is one of the shifts that works, and never rules out a shorter one.

The general lesson is worth writing down. Some properties are inherited by a quotient from its parts, and some are not. A quotient inherits its domain restrictions and it inherits parity in a predictable way. It does not inherit its period, because a combination can repeat sooner than either part does. Whenever you think you know a period, check that nothing smaller works.

25. Which of these repeat every half turn?

Discrimination

Sort into buckets

Sort each function by whether adding pi to the input leaves it unchanged.

Repeats every π
y = tan x; y = cos(2x); y = cot x; y = sin(2x)
Needs more than π
y = sin x; y = sec x; y = tan(x/2)
y
tan x and cot x repeat every π because they read a ratio. cos(2x) and sin(2x) repeat every π because doubling the input halves the cycle: 2π divided by 2.
n
sin x and sec x both need a full turn, since a half turn flips the sign of the coordinate they read. tan(x/2) needs 2π, because halving the input doubles the cycle from π to 2π.

Sort these by asking one question of each: if I add a half turn to the input, do I get the same output? Do not sort by whether the function looks like a tangent.

Two of the items are in the yes bucket for a reason that has nothing to do with ratios. Doubling the input of a cosine or a sine halves its cycle, so those two finish in a half turn as well. A half turn is not the private property of the tangent family; it is just the only place it appears without help.

The last item is the one that catches people. Halving the input of a tangent doubles its cycle, from a half turn to a full one. Inside a function, a larger coefficient means a shorter period, and a smaller one means a longer period. That relationship is inverted from what the eye expects, so check it every time.

26. The cotangent curve

Section

Part 3

27. Swap the numerator and the denominator

Concept

Everything about the cotangent follows from the fact that it is the same two coordinates the other way up. Whatever made the numerator vanish now makes the denominator vanish, and the other way round.

\[ K(x) = \cot(x) = \frac{\cos(x)}{\sin(x)} \]

what vanisheseffect on the tangenteffect on the cotangent
sin x = 0, at the multiples of πa zeroan asymptote
cos x = 0, at the odd multiples of π/2an asymptotea zero

So the two graphs have their special points in exactly opposite places, and every feature you proved for one has a mirror statement for the other.

Stitz & Zeager, College Trigonometry, Ch. 10 Foundations of Trigonometry — §10.5 Graphs of the Trigonometric Functions §10.5, p. 805 — the introduction of K(x) = cot(x)

Everything about the cotangent follows from one sentence: it is the same two numbers the other way up. Get comfortable with that and you never have to learn the cotangent separately.

Work through the table on the slide slowly. Wherever the sine is zero, the tangent has a zero because its numerator dies, and the cotangent has an asymptote because its denominator dies. Wherever the cosine is zero, the two swap roles. Since the sine and the cosine are never zero at the same input, zeros and asymptotes never collide.

That is why the two graphs look like the same skeleton with the flesh redistributed. Every statement you proved about the tangent has a mirror statement about the cotangent, obtained by exchanging the words zero and asymptote and by exchanging the words rising and falling.

28. Building the cotangent's table over a full turn

Worked example

The same nine marks, with the quotient the other way up.

\[ \cot(x) = \frac{\cos(x)}{\sin(x)} \text{ at } x = 0, \tfrac{\pi}{4}, \tfrac{\pi}{2}, \ldots, 2\pi \]

Start at zero

Why: The sine is zero there, so this time the very first mark is an asymptote.

\[ \cot(0) = \frac{1}{0}: \text{ undefined} \]

Move a quarter of a quarter turn

Why: The two coordinates are equal, so the quotient is one, just as for the tangent.

\[ \cot\!\left(\tfrac{\pi}{4}\right) = \frac{\sqrt2/2}{\sqrt2/2} = 1 \]

Reach a quarter turn

Why: Now the cosine is zero, so the numerator dies and the value is zero.

\[ \cot\!\left(\tfrac{\pi}{2}\right) = \frac{0}{1} = 0 \]

Finish the first branch and start the second

Why: The second quadrant gives minus one, and the sine vanishes again at a half turn.

\[ \cot\!\left(\tfrac{3\pi}{4}\right) = -1, \qquad \cot(\pi) \text{ undefined} \]

x0π/4π/23π/4π5π/43π/27π/42π
cot xundefined10−1undefined10−1undefined

Figure (svg): The cotangent curve over zero to two pi, with filled dots at the six defined eighth-marks and dashed vertical asymptotes at zero, pi and two pi.

The same six heights as the tangent's table, in the opposite order: 1, 0, minus 1 on each branch. The asymptotes have moved to where the tangent had its zeros.

Check: read each branch in order

Why: Each branch runs 1, then 0, then minus 1, so the heights decrease from left to right, and the list from 0 to pi repeats exactly from pi to two pi.

\[ \cot(x + \pi) = \cot(x) \text{ at every mark in the table} \]

Stitz & Zeager, College Trigonometry, Ch. 10 Foundations of Trigonometry — §10.5 Graphs of the Trigonometric Functions §10.5, p. 805 — the table for K(x) = cot(x)

Notice how little work this is once the tangent's table exists. You are dividing the same pairs of numbers in the other order, so every entry is either the reciprocal of a tangent entry or a swap between zero and undefined.

The pattern within a branch is the thing to fix in memory. Each branch runs one, then zero, then minus one, in that order from left to right. The tangent's branches run minus one, zero, one. Same three heights, opposite order, which is the same as saying one falls where the other rises.

Do the repeat check again. The five entries from zero to a half turn match the five entries from a half turn to a full turn exactly, so the cotangent is also finished after a half turn. The next slide proves it properly, and it takes a different route from the tangent's proof because there is no convenient sum formula for the cotangent in the book.

29. Proving the period of the cotangent is a half turn

Worked example

The book leaves this to the reader. Do it straight from the definition rather than from a sum formula, which is quicker here.

\[ \text{Show that } \cot(x + \pi) = \cot(x), \text{ and that nothing smaller works.} \]

Write the cotangent as a quotient at the shifted input

Why: Definition, with the input x plus a half turn.

\[ \cot(x + \pi) = \frac{\cos(x + \pi)}{\sin(x + \pi)} \]

Apply the sum formulas to each part

Why: The cosine of a half turn is minus one and its sine is zero, so both expansions collapse to a single term.

\[ \cos(x + \pi) = -\cos(x), \qquad \sin(x + \pi) = -\sin(x) \]

Substitute and cancel the two minus signs

Why: A quotient of two negatives is the quotient of the positives.

\[ \cot(x + \pi) = \frac{-\cos(x)}{-\sin(x)} = \frac{\cos(x)}{\sin(x)} = \cot(x) \]

Now suppose p is any positive period and specialise at a quarter turn

Why: Zero is not in the domain of the cotangent, so use the first point where the cotangent is zero instead.

\[ \cot\!\left(\tfrac{\pi}{2} + p\right) = \cot\!\left(\tfrac{\pi}{2}\right) = 0 \]

Solve for p

Why: The cotangent vanishes exactly at the odd multiples of a quarter turn, so the shift must be a whole number of half turns.

\[ \tfrac{\pi}{2} + p = \tfrac{\pi}{2} + \pi k \;\Longrightarrow\; p = \pi k \;\Longrightarrow\; p_{\min} = \pi \]

Check numerically

Why: Pick an input at random and compare the two values.

\[ \cot(0.4) = 2.365222 = \cot(0.4 + \pi) \]

Stitz & Zeager, College Trigonometry, Ch. 10 Foundations of Trigonometry — §10.5 Graphs of the Trigonometric Functions §10.5, p. 805 — footnote 14, left to the reader

The book leaves this to you in a footnote, and offers two routes: copy the tangent's proof with a sum formula, or transform the tangent into the cotangent with an identity. The route taken here is a third and shorter one, straight from the definition.

The first half is three lines. Write the cotangent at the shifted input as a quotient, use the two facts that adding a half turn negates both the cosine and the sine, and cancel the two minus signs. That is the same cancellation that made the tangent work, which is no accident.

The second half needs one small adjustment, and it is the reason to look at this proof rather than simply assert that it is like the tangent's. You cannot specialise at zero, because zero is not in the cotangent's domain. Use a quarter turn instead, where the cotangent is zero, and the same argument goes through.

The numerical check picks an input at random on purpose. A check at a special angle can succeed by accident; a check at an ugly one like 0.4 radians is much harder to pass by luck.

30. Tangent up, cotangent down

Picture it

Figure (svg): Two panels. Left: one branch of the tangent between minus pi over two and pi over two, rising, with an upward arrow. Right: one branch of the cotangent between zero and pi, falling, with a downward arrow.

Same skeleton, opposite traffic. Each branch still passes through every real number exactly once, so both ranges are everything.

One branch of each, side by side and to the same scale. They occupy different intervals, they run in opposite directions, and they carry the same three marked heights in opposite orders.

Notice what is the same: both branches are continuous, both are strictly monotone, and both sweep out every real number exactly once. That is why both ranges are everything and neither has an amplitude.

Two branches drawn to the same scale, so you can compare them honestly. Cover the labels and ask yourself which is which; the direction of travel is the only clue you need.

The left branch lives between two asymptotes that are symmetric about the origin, and it passes through the origin itself. The right branch lives between an asymptote at zero and one at a half turn, and its zero is at the middle, a quarter turn along. Those are the two fundamental cycles, and the next part of the lesson builds the graphing method on top of them.

What is identical is more important than what differs. Both branches are continuous, both are strictly monotone, and both pass through every real height exactly once. That is why both ranges are everything, why neither has an amplitude, and why both will have well-behaved inverses.

31. Each graph is the other one, reflected

Intuition

Because the cofunction identity turns one into the other, the cotangent graph is the tangent graph read backwards. Reflect the tangent in the vertical axis and slide it a quarter turn, and you have the cotangent.

\[ \cot(x) = \tan\!\left(\frac{\pi}{2} - x\right) \]

Reading backwards is exactly what turns a rising branch into a falling one. That is the single structural difference between the two pictures, and it is the one students most often get wrong when sketching from memory.

Stitz & Zeager, College Trigonometry, Ch. 10 Foundations of Trigonometry — §10.5 Graphs of the Trigonometric Functions §10.4, p. 774 — Theorem 10.14, the cofunction identities

The cofunction identity from Section 10.4 says the cotangent of an input is the tangent of a quarter turn minus that input. Read the right-hand side carefully: the input appears with a minus sign in front of it, so the graph is being read backwards.

Reading a graph backwards is what turns rising into falling. That is the entire structural difference between these two pictures, and it is the detail people most often lose when sketching a cotangent from memory. If you find yourself drawing a rising cotangent, you have forgotten the reflection.

There is a practical consequence. If you can draw the tangent, you can always produce the cotangent without learning anything new: reflect left to right, then slide. Just do not try to produce it by sliding alone, which is the subject of a trap a few slides from now.

32. Tangent against cotangent

Comparison

Comparison matrix

featuretangentcotangent
zerosmultiples of πodd multiples of π/2
asymptotesodd multiples of π/2multiples of π
each branchincreasesdecreases
periodππ
rangeall real numbersall real numbers
parityoddodd
fundamental cycle(−π/2, π/2)(0, π)

Fill in every blank before checking, and try to fill each one by reasoning rather than by recall. Every row follows from one sentence: the cotangent is the same quotient upside down.

Two rows are the same in both columns, and that is worth noticing. The period is a half turn for both, because both read a ratio. The range is everything for both, because both branches are continuous and strictly monotone across an unbounded climb or fall.

The row on parity may surprise you. Both functions are odd, even though one is built from an even numerator and the other from an odd one. Work it out from the definitions: negating the input leaves the cosine alone and negates the sine, so either way exactly one of the two parts of the quotient changes sign.

33. Is the cotangent just one over the tangent?

Counterexample

\[ \cot(x) \overset{?}{=} \frac{1}{\tan(x)} \]

Discussion prompt

The reciprocal identity says these are equal. Find an input where the claim breaks down, say what each side does there, and state the correct version of the identity.

Write your answer before revealing. Finding where an identity breaks is a much better test of understanding than reciting it.

At a quarter turn the cotangent is a perfectly ordinary zero, because its numerator is zero and its denominator is one. The reciprocal of the tangent, on the other hand, asks you to divide by something that does not exist. One side of the claimed identity has a value and the other does not, so the two are not the same function.

The moral is about domains, and it applies to every reciprocal identity in the chapter. An identity is a statement about the inputs where both sides are defined, and the quotient definitions have larger domains than the reciprocal ones. When a question involves a quarter turn or a half turn, go back to the quotient definition rather than the reciprocal.

34. Trap: the cotangent is the tangent shifted

Trap

The trap

The two graphs look like translates of one another, so a shift gets written down:

\[ \cot(x) = \tan\!\left(x - \tfrac{\pi}{2}\right)? \]

Wrong, and by a sign.

\[ \tan\!\left(0.37 - \tfrac{\pi}{2}\right) = -2.578 \]

\[ \cot(0.37) = +2.578 \]

The fix

A translation slides a graph sideways; it cannot turn a rising branch into a falling one. Since the cotangent falls and the tangent rises, some reflection has to appear. Two correct versions:

\[ \cot(x) = -\tan\!\left(x + \tfrac{\pi}{2}\right) \]

\[ \cot(x) = \tan\!\left(\tfrac{\pi}{2} - x\right) \]

The first reflects vertically, the second reflects the input. Either way a minus sign is unavoidable, and if your remembered identity has no minus sign in it, it is the wrong one.

This is the most common wrong identity in the whole section, and it is wrong for a reason you can see rather than one you have to remember: a sideways slide cannot reverse the direction a branch travels.

So before you write any translation identity relating these two functions, ask whether a reflection is needed. Here it is, because one graph rises and the other falls, and the reflection has to show up as a minus sign somewhere. If your remembered identity has no minus sign in it, it is the wrong one.

The numerical check on the left is how to catch this in ten seconds. Pick any convenient input, evaluate both sides, and see that the answers are the negatives of one another. That immediately tells you not only that the identity is wrong but exactly how it is wrong, which is a minus sign.

35. Reading Theorem 10.25

Notation

Annotate

On: \( K(x) = \cot(x): \; \text{domain } x \ne \pi k, \; \text{range } (-\infty, \infty), \; \text{odd}, \; \text{period } \pi \)

  • Every multiple of a half turn is excluded, including zero itself, because the sine vanishes there. As a union of intervals that is the open interval from kπ to (k+1)π, one per branch.
  • Every real number, and each one is attained once on every branch. Nothing is missing and nothing is repeated within a branch.
  • Negating the input negates the output, because the cosine on top is even and the sine underneath is odd. So the graph has half-turn symmetry about the origin.
  • The same as the tangent's, and for the same reason: both are ratios of the two coordinates. Theorem 10.25 records the identical list of properties for both functions.

The theorem also says both functions are continuous and smooth on their domains. The asymptotes are not breaks in the curve; they are places where the curve was never defined.

Stitz & Zeager, College Trigonometry, Ch. 10 Foundations of Trigonometry — §10.5 Graphs of the Trigonometric Functions §10.5, p. 806 — Theorem 10.25

The theorem states the same four properties for both functions, and the annotations here walk through them for the cotangent because its domain is the easier of the two to write down.

Take the word smooth seriously. The theorem says both functions are continuous and smooth on their domains, and it is important that it says on their domains. An asymptote is not a break in a curve that is otherwise there; it is a place where the function was never defined at all, so continuity is not even being claimed there.

The parity claim is the one worth deriving yourself rather than memorising. For the cotangent, the numerator is even and the denominator is odd, so negating the input leaves the top alone and flips the bottom, and the quotient flips. For the tangent it is the other way round, and the quotient flips again. Both come out odd, so both graphs have half-turn symmetry about the origin.

36. Match each formula to its asymptotes and period

Matching

Match the pairs

  • a. y = tan x
  • b. y = cot x
  • c. y = tan(2x)
  • d. y = cot(x/2)
  • w. period π, asymptotes at the odd multiples of π/2
  • x. period π, asymptotes at the multiples of π
  • y. period π/2, asymptotes at the odd multiples of π/4
  • z. period 2π, asymptotes at the even multiples of π

Why: Find the asymptotes by setting the argument equal to whatever makes the parent undefined, then solve. For the tangent of two x, the argument is an odd multiple of a quarter turn when x is an odd multiple of an eighth of a turn, and consecutive asymptotes are then half of a half turn apart. For the cotangent of half of x, the argument is a multiple of a half turn when x is an even multiple of one, so consecutive asymptotes are a full turn apart.

Do these by solving, not by pattern matching. For each formula, set the argument equal to whatever makes the parent undefined and solve for the input.

The two transformed items are the interesting ones, and they move in opposite directions. Doubling the input compresses the graph, so the asymptotes come twice as often and the period halves. Halving the input stretches the graph, so the asymptotes come half as often and the period doubles.

If you are ever unsure which way a coefficient moves things, do not reason about it abstractly. Solve for two consecutive asymptotes and subtract. That computation is short, it never lies, and it gives you the period as a bonus.

37. The right quarter marks

Section

Part 4

38. Each function gets its own fundamental cycle

Concept

For a sinusoid you graphed one cycle by setting the argument equal to the five quarter marks of a full turn and solving. The method is the same here, but the marks are not.

Figure (svg): Two marked number lines. The upper one runs from minus pi over two to pi over two with hollow circles at the ends marked asymptote and three filled dots between them; the lower one runs from zero to pi with the same arrangement.

Both cycles are a half turn wide and both are open at the ends, but they start in different places, and the middle three heights run the opposite way round.

fundamental cycle — One complete copy of the graph, chosen once and for all so that every other copy is a translate of it. For the tangent the book takes the open interval from minus a quarter turn to a quarter turn; for the cotangent, the open interval from zero to a half turn.

Both cycles are a half turn wide and both are open at their ends, because the ends are asymptotes and the function has no value there. Only three of the five marks give points.

Stitz & Zeager, College Trigonometry, Ch. 10 Foundations of Trigonometry — §10.5 Graphs of the Trigonometric Functions §10.5, pp. 805-807 — the fundamental cycles and their quarter marks

The quarter-mark method itself is unchanged from the sinusoid lesson: name the argument, set it equal to each of five marks, solve for the input, substitute. What changes is the list of five marks, and it changes twice, once for each of these two functions.

Compare the two lines in the picture. Both cycles are a half turn wide, and both are open at the ends because the ends are asymptotes. The middle mark is a zero in both cases. The difference is where the cycle starts and, consequently, whether the heights either side of the zero read one then minus one or minus one then one.

The practical consequence is that only three of your five solved inputs give points. The other two give vertical lines. Get into the habit of writing asymptote beside those two rather than leaving them blank, because on an exam an unlabelled gap is worth nothing.

39. Why the sinusoid's quarter marks do not transfer

Intuition

The sinusoid marks were chosen to catch the interesting features of a cosine: a maximum, a crossing, a minimum, another crossing, and back to the maximum. Those features do not exist here.

The features that do exist are two asymptotes and one zero, with a point either side of the zero to fix the steepness. So the five marks were rechosen to catch an asymptote, a value of one, a zero, a value of minus one, and the next asymptote.

Use the old marks on a tangent and you will place the asymptotes at a quarter of the cycle in from each end, which is simply wrong. The marks are part of the function, not part of the method.

It is natural to reach for the marks you used last week, and the method looks so similar that the substitution is easy to make without noticing. The result is a graph with the right shape and every feature in the wrong place.

Think about what the marks are for. The sinusoid marks were chosen to catch a maximum, a crossing, a minimum, another crossing and the next maximum. Not one of those five features exists on a tangent. The marks here were rechosen to catch the features that do exist: an asymptote, a height of one, a zero, a height of minus one, and the next asymptote.

So treat the choice of marks as part of identifying the function, not as part of the method. The first thing to write on your page is which parent you are looking at; the five marks follow from that answer and from nothing else.

40. Example 10.5.5(1): one cycle of one minus the tangent of half of x

Worked example

Graph one cycle and find the period.

\[ f(x) = 1 - \tan\!\left(\frac{x}{2}\right) \]

Name the argument and list the tangent's quarter marks

Why: The argument is the whole expression inside the tangent.

\[ a = \frac{x}{2}; \quad a = -\tfrac{\pi}{2}, -\tfrac{\pi}{4}, 0, \tfrac{\pi}{4}, \tfrac{\pi}{2} \]

Set the argument equal to each mark and solve

Why: Multiply each mark by two.

\[ x = -\pi, \; -\tfrac{\pi}{2}, \; 0, \; \tfrac{\pi}{2}, \; \pi \]

Substitute the two end values

Why: The tangent is undefined at its own end marks, so these give the asymptotes.

\[ x = -\pi \text{ and } x = \pi: \quad f(x) \text{ undefined} \]

Substitute the three middle values

Why: Use the tangent values minus one, zero and one in turn.

\[ f\!\left(-\tfrac{\pi}{2}\right) = 1 - (-1) = 2, \quad f(0) = 1 - 0 = 1, \quad f\!\left(\tfrac{\pi}{2}\right) = 1 - 1 = 0 \]

Read the period from the two asymptotes

Why: The period is the distance between them.

\[ \pi - (-\pi) = 2\pi \]

Figure (svg): One cycle of one minus the tangent of half of x, from minus pi to pi. The branch falls from the top left asymptote to the bottom right one, passing through the marked points at height two, one and zero.

Halving the input doubles the cycle, and the minus sign turns the usual rising branch into a falling one. The centre of the branch has been lifted to height one.

Check the direction against the formula

Why: The coefficient in front of the tangent is minus one, so the branch must fall, and the three heights 2, 1, 0 do decrease from left to right.

\[ f(-\tfrac{\pi}{2}) = 2 > f(0) = 1 > f(\tfrac{\pi}{2}) = 0 \]

Stitz & Zeager, College Trigonometry, Ch. 10 Foundations of Trigonometry — §10.5 Graphs of the Trigonometric Functions §10.5, p. 807 — Example 10.5.5, part 1

Follow the order of operations here carefully, because it is the order you will use on every one of these problems. Name the argument, solve five small equations, then and only then substitute into the full function.

The halved input is what makes this cycle two full turns wide. That will feel backwards the first few times: a coefficient smaller than one inside the function makes the graph wider, because the input has to travel further before the argument has moved through a half turn.

The minus sign in front is the second thing to notice, and it is easy to lose. It turns the parent's rising branch into a falling one, so the three heights you compute must come out in decreasing order. They do, running two, one, zero, and the check at the end says so explicitly. If your heights come out increasing, you have dropped the sign.

41. Exercise 13: a tangent shifted to the right

Worked example

Graph one cycle and state the period.

\[ y = \tan\!\left(x - \frac{\pi}{3}\right) \]

Set the argument equal to the five tangent marks

Why: The argument is x minus a third of a half turn.

\[ x - \tfrac{\pi}{3} = -\tfrac{\pi}{2}, -\tfrac{\pi}{4}, 0, \tfrac{\pi}{4}, \tfrac{\pi}{2} \]

Solve each equation

Why: Add a third of a half turn to every mark; use a common denominator of twelve.

\[ x = -\tfrac{\pi}{6}, \; \tfrac{\pi}{12}, \; \tfrac{\pi}{3}, \; \tfrac{7\pi}{12}, \; \tfrac{5\pi}{6} \]

Name the asymptotes

Why: The two outer values, where the tangent's own end marks land.

\[ x = -\tfrac{\pi}{6} \quad \text{and} \quad x = \tfrac{5\pi}{6} \]

Substitute the three inner values

Why: The tangent of each inner mark is minus one, zero and one.

\[ \left(\tfrac{\pi}{12}, -1\right), \quad \left(\tfrac{\pi}{3}, 0\right), \quad \left(\tfrac{7\pi}{12}, 1\right) \]

Read the period

Why: Subtract the asymptotes.

\[ \frac{5\pi}{6} - \left(-\frac{\pi}{6}\right) = \frac{6\pi}{6} = \pi \]

Figure (svg): One cycle of the tangent of x minus pi over three, rising between dashed asymptotes at minus pi over six and five pi over six, through the points at heights minus one, zero and one.

Exactly the parent branch, moved a third of a half turn to the right. The width between the asymptotes is unchanged, so the period is still a half turn.

Check against the parent

Why: Nothing multiplies x inside, so the cycle cannot change width; it can only move. The period is the parent's pi and the whole picture has slid right by a third of a half turn, which is where the zero now sits.

\[ \text{zero at } x = \tfrac{\pi}{3}, \quad \text{period } \pi \]

Stitz & Zeager, College Trigonometry, Ch. 10 Foundations of Trigonometry — §10.5 Graphs of the Trigonometric Functions §10.5, p. 809 — Exercise 13

This is the simplest kind of transformation, and it is a good one to build the habit on because nothing multiplies the input. When nothing multiplies the input, the cycle cannot change width, so you already know the period before you start.

The arithmetic is all in twelfths. A third of a half turn is four twelfths, a quarter of a half turn is three twelfths, and adding those gives the seven twelfths in the list. Getting a common denominator before you add is what keeps this tidy.

Notice which way the graph moved. The quantity subtracted inside the function is positive, and the graph moves to the right. That is the direction most people get wrong, so anchor it with the zero: the parent's zero was at the origin, and the new zero is where the argument is zero, which is at a third of a half turn to the right.

42. Before you graph: which five marks?

Step zero

\[ y = \cot\!\left(x + \frac{\pi}{6}\right) \]

Discussion prompt

Before solving anything, write down the five quarter marks you will set the argument equal to, and say in one sentence how you know they are not the ones you used for a sinusoid.

Write your five marks before revealing, and then write one sentence justifying them. The justification is the part being assessed.

This is a cotangent, so the fundamental cycle starts at zero and ends at a half turn, and the marks are the five evenly spaced values across it. If you wrote the sinusoid's marks, you would have produced a cycle twice as wide with the asymptotes in the wrong places, and everything downstream would be wrong.

Thirty seconds spent on this decision saves several minutes of misdirected algebra. It is also the step that examiners can see: an answer that starts by naming the correct marks shows you know which parent you are working with, even if an arithmetic slip follows.

43. Exercise 14: a tangent with a long period

Worked example

Graph one cycle and state the period.

\[ y = 2\tan\!\left(\frac{x}{4}\right) - 3 \]

Set the argument equal to the five tangent marks

Why: The argument is a quarter of x.

\[ \frac{x}{4} = -\tfrac{\pi}{2}, -\tfrac{\pi}{4}, 0, \tfrac{\pi}{4}, \tfrac{\pi}{2} \]

Solve for x

Why: Multiply every mark by four.

\[ x = -2\pi, \; -\pi, \; 0, \; \pi, \; 2\pi \]

Substitute the three inner values

Why: Take the tangent, double it, then subtract three.

\[ y(-\pi) = 2(-1) - 3 = -5, \quad y(0) = -3, \quad y(\pi) = 2(1) - 3 = -1 \]

Read the period

Why: The distance between the outer two values.

\[ 2\pi - (-2\pi) = 4\pi \]

Figure (svg): One cycle of twice the tangent of a quarter of x, minus three, rising between asymptotes at minus two pi and two pi, through the points at heights minus five, minus three and minus one.

A quarter inside the tangent quadruples the cycle. The vertical stretch by two and the drop of three move the three marked heights but not the asymptotes.

Check the period against the frequency

Why: A quarter inside the tangent means the input must travel four times as far to move the argument through a half turn, and pi divided by one quarter is four pi.

\[ \text{period} = \frac{\pi}{1/4} = 4\pi \]

Stitz & Zeager, College Trigonometry, Ch. 10 Foundations of Trigonometry — §10.5 Graphs of the Trigonometric Functions §10.5, p. 809 — Exercise 14

Three separate transformations act here, and they act on different things, so keep them apart. The quarter inside the function changes the width of the cycle. The factor of two outside changes the three heights. The three subtracted at the end lowers everything.

Only the first of those three touches the asymptotes. That is worth saying to yourself as you work: outside changes cannot move an asymptote, because an asymptote marks an input where the parent has no value at all, and multiplying or shifting a missing value leaves it missing.

The check at the end does the period a second way, dividing the parent's half turn by the frequency of one quarter. Dividing by a quarter is multiplying by four, which is why the cycle is four times as wide. Doing the period twice, once from the asymptotes and once from the frequency, is cheap insurance.

44. Slide the frequency and watch the branches crowd

Tweak it

Parameter explorer

Slide the frequency. Do the branches crowd together or spread apart as it grows, and does the shape of each individual branch change?

\[ y = \tan({w}x) \]

  • w — from 0.5 to 4: frequency

Move the slider slowly and watch how many branches fit on the screen. At the smallest setting there is barely more than one; at the largest there are several, packed together.

What does not change is the shape of any individual branch. Each one still comes up from far below, crosses the middle, and runs off far above. The frequency controls only how much horizontal room each branch is given, and that room is a half turn divided by the frequency.

The picture has no axes or numbers, so do not try to read the period off it. What it is for is the qualitative relationship, and that relationship is the one people most often get backwards: bigger frequency, shorter period, more branches.

45. Example 10.5.5(2): one cycle of a transformed cotangent

Worked example

Graph one cycle and find the period. Note the frequency is not a whole number here.

\[ g(x) = 2\cot\!\left(\frac{\pi}{2}x + \pi\right) + 1 \]

List the cotangent's quarter marks and set the argument equal to each

Why: These are the cotangent's marks, starting at zero rather than at minus a quarter turn.

\[ \frac{\pi}{2}x + \pi = 0, \; \tfrac{\pi}{4}, \; \tfrac{\pi}{2}, \; \tfrac{3\pi}{4}, \; \pi \]

Solve the first and last equations

Why: Subtract a half turn, then multiply by two over pi.

\[ x = -2 \quad \text{and} \quad x = 0 \]

Solve the three middle equations

Why: The same two operations each time; every pi cancels.

\[ x = -\tfrac{3}{2}, \; -1, \; -\tfrac12 \]

Substitute the three middle values

Why: Take the cotangent value, double it, then add one.

\[ g\!\left(-\tfrac32\right) = 2(1) + 1 = 3, \quad g(-1) = 2(0) + 1 = 1, \quad g\!\left(-\tfrac12\right) = 2(-1) + 1 = -1 \]

Read the period

Why: The distance between the two asymptotes.

\[ 0 - (-2) = 2 \]

Figure (svg): One cycle of twice the cotangent of pi over two times x plus pi, plus one, falling between asymptotes at minus two and zero, through the points at heights three, one and minus one.

The input is measured in plain units now, not in multiples of pi: a frequency of pi over two makes the cycle exactly two units wide.

Check the period against the frequency

Why: A half turn divided by the frequency, and every pi cancels, which is why the answer is a plain number of units rather than a multiple of pi.

\[ \text{period} = \frac{\pi}{\pi/2} = 2 \]

Stitz & Zeager, College Trigonometry, Ch. 10 Foundations of Trigonometry — §10.5 Graphs of the Trigonometric Functions §10.5, pp. 807-808 — Example 10.5.5, part 2

The unfamiliar feature here is that the frequency itself contains a half turn, so all the half turns cancel and the answers come out as plain numbers. Do not let that unsettle you; the method has not changed at all.

Work through one of the five equations by hand to see the cancellation. Subtract a half turn from both sides, then multiply both sides by two over a half turn. Everything with a half turn in it disappears, and you are left with an ordinary number of units.

This is exactly how a tangent or cotangent turns up in an application. Time is measured in seconds, not in turns, so the frequency carries the half turn and the period comes out in seconds. The lighthouse slide later in the lesson is the same idea.

46. Exercise 23: a cotangent stretched and flipped

Worked example

Graph one cycle and state the period.

\[ y = -11\cot\!\left(\frac{x}{5}\right) \]

Set the argument equal to the five cotangent marks

Why: The argument is a fifth of x.

\[ \frac{x}{5} = 0, \; \tfrac{\pi}{4}, \; \tfrac{\pi}{2}, \; \tfrac{3\pi}{4}, \; \pi \]

Solve for x

Why: Multiply every mark by five.

\[ x = 0, \; \tfrac{5\pi}{4}, \; \tfrac{5\pi}{2}, \; \tfrac{15\pi}{4}, \; 5\pi \]

Substitute the three inner values

Why: Take the cotangent, then multiply by minus eleven.

\[ y\!\left(\tfrac{5\pi}{4}\right) = -11(1) = -11, \quad y\!\left(\tfrac{5\pi}{2}\right) = 0, \quad y\!\left(\tfrac{15\pi}{4}\right) = -11(-1) = 11 \]

Read the period

Why: The distance between the outer two values.

\[ 5\pi - 0 = 5\pi \]

Figure (svg): One cycle of minus eleven times the cotangent of a fifth of x, rising between asymptotes at zero and five pi, through the points at heights minus eleven, zero and eleven.

A cotangent branch that rises. Nothing about the shape changed; the whole picture was reflected in the horizontal axis by the factor of minus eleven.

Check the direction

Why: A cotangent branch normally falls, and the three heights here rise, which is exactly what the negative factor should do.

\[ -11 < 0 \;\Longrightarrow\; -11 < 0 < 11 \text{ from left to right} \]

Stitz & Zeager, College Trigonometry, Ch. 10 Foundations of Trigonometry — §10.5 Graphs of the Trigonometric Functions §10.5, p. 809 — Exercise 23

The large numbers here are deliberately distracting, and they are also harmless. Eleven and a fifth do completely different jobs, and separating them is the whole exercise.

The fifth is inside, so it stretches the cycle by a factor of five, from a half turn to five half turns. The eleven is outside, so it scales the three heights and, being negative, turns the picture over. Nothing about the eleven can move an asymptote.

Reversal is the thing to check at the end. A cotangent normally falls, and this one rises, which is exactly what the negative factor should do. Whenever the leading coefficient is negative, make the last line of your solution a statement about direction, because that is the part a marker looks for.

47. Finish it yourself: exercise 22

Faded example

The stepping-off point slide set up the marks for this one. Now carry them through: set the argument equal to the first, third and fifth marks and solve.

\[ y = \cot\!\left(x + \frac{\pi}{6}\right); \qquad x + \frac{\pi}{6} = 0, \; \frac{\pi}{2}, \; \pi \]

Fill in the blanks

\text-\frac{\pi}{6} x = \frac{\pi}{3}, \quad \text\frac{5\pi}{6} x = ___, \quad \text___ x = ___

Why: Subtract a sixth of a half turn from each mark. Zero gives minus one sixth; a half of a half turn is three sixths, and three sixths minus one sixth is two sixths, which is a third; a whole half turn is six sixths, and six sixths minus one sixth is five sixths. The two asymptotes are a half turn apart, so the period is a half turn, unchanged by a shift.

Stitz & Zeager, College Trigonometry, Ch. 10 Foundations of Trigonometry — §10.5 Graphs of the Trigonometric Functions §10.5, p. 809 — Exercise 22

The setup slide already chose the marks for this one, so all that remains is the arithmetic. Solve the first, third and fifth equations and fill in the three answers.

Work in sixths of a half turn throughout. Zero minus one sixth is minus one sixth. Three sixths minus one sixth is two sixths, which reduces to a third. Six sixths minus one sixth is five sixths. Keeping one denominator all the way through prevents most of the slips people make here.

Then look at your two asymptotes and check they are a half turn apart. They are, because nothing multiplies the input, so a shift alone cannot change the period. That check costs you five seconds and catches any arithmetic error you have just made.

48. Find the error: dividing a full turn by the frequency

Error analysis

A student computes the period of a stretched tangent with the formula that worked all through the sinusoid lesson.

Annotate

On: \( y = \tan(3x): \quad \text{period} = \frac{2\pi}{3} \)

  • 2π over the frequency is the sinusoid's period formula, and it is right for a sinusoid because the parent cosine takes a full turn to repeat.
  • The parent here takes only a half turn. The correct rule divides the PARENT'S period by the frequency, so it is π over 3.
  • Set 3x equal to plus and minus π/2: the asymptotes are at plus and minus π/6, which are π/3 apart. The graph agrees with π/3, not with 2π/3.
  • Never quote a period formula from memory on these two graphs. Solve for the two asymptotes and subtract; that works for every function in the chapter.

Read the line and see whether you can name the error before the annotations appear. The formula being used is a real formula that is correct in a different context, which is what makes this so easy to do by reflex.

The rule you actually want divides the parent's period by the frequency. For a cosine or a sine the parent takes a whole turn, so the numerator is a whole turn. For a tangent or a cotangent the parent takes only a half turn, so the numerator is a half turn. One rule, two different parents.

The third annotation shows the defence, and it is the same defence as everywhere else in this lesson. Solve for two consecutive asymptotes and subtract. That computation makes no assumption about which family you are in, so it cannot be fooled, and you should prefer it to any remembered formula.

49. Reading a transformed graph

Section

Part 5

50. Three parameters, and a direction

Concept

A sinusoid had four numbers to report. Here there are three, plus one yes-or-no question that the sinusoid never asked.

\[ y = A\tan(\omega x + \varphi) + B, \qquad \omega > 0 \]

quantityhow to get itwhy
periodπ divided by ωthe argument must travel π for the parent to repeat
phase shift−φ divided by ωthe value of x that makes the argument zero
vertical shiftBevery height moves up by B
branch directionrising if A > 0, falling if A < 0a negative factor reflects the graph

There is deliberately no amplitude row. The book declines to state a theorem for these two functions at all, on the grounds that applications of them are rare, and invites you to write one yourself.

Stitz & Zeager, College Trigonometry, Ch. 10 Foundations of Trigonometry — §10.5 Graphs of the Trigonometric Functions §10.5, p. 808 — the remark after Example 10.5.5

Compare this table with the one you built for sinusoids and notice both what is missing and what is new. Amplitude is gone, and in its place is a yes-or-no question about direction that a sinusoid never had to ask.

The period row is the one to internalise. A half turn divided by the frequency, not a whole turn divided by it, because the parent finishes twice as fast. The phase shift row is unchanged from the sinusoid: it is the input that makes the argument zero, and you find it by solving rather than by reading the formula.

The book itself declines to state a theorem for these two functions, saying that applications of them are rare compared with sinusoids, and invites the ambitious reader to write one. This table is that theorem, written out. If your course wants it stated formally, this is what to state.

51. The coefficient stretches, and can flip

Intuition

The number in front does two jobs, and only one of them is visible in the list of five marks. It multiplies every height measured from the vertical shift, which drags the three plotted points closer to that line when the number is small and away from it when the number is large.

What it cannot do is move an asymptote. The asymptotes are decided entirely by what is inside the function, because they are the inputs at which the parent has no value, and multiplying a non-existent value by anything leaves it non-existent.

If the number is negative it also reflects the graph vertically, which swaps rising for falling. That is the only way a tangent branch can be made to fall, and the only way a cotangent branch can be made to rise.

It helps to think of the leading coefficient as acting on distances measured from the vertical shift, rather than on heights measured from the horizontal axis. A coefficient of a third pulls every plotted point to a third of its distance from that line; a coefficient of five pushes it five times as far away.

That is also why the coefficient cannot affect the asymptotes. An asymptote is an input where the parent has no value, and a missing value multiplied by anything is still missing. So a stretch changes how steeply a branch passes through its three marked points, and nothing else about the picture.

When the coefficient is negative there is a second effect, and it is the one worth saying out loud in your written answer: the graph is reflected, so the branches change direction. It is the only mechanism in this lesson that can make a tangent fall or a cotangent rise.

52. Exercise 15: a negative inside the tangent

Worked example

Graph one cycle and state the period. The coefficient of x is negative, so rewrite before doing anything else.

\[ y = \frac13\tan(-2x - \pi) + 1 \]

Factor the minus sign out of the argument

Why: Both terms inside carry it.

\[ -2x - \pi = -(2x + \pi) \]

Use the fact that the tangent is odd

Why: Negating the input negates the output.

\[ \tan(-(2x + \pi)) = -\tan(2x + \pi) \]

Use the period to drop the half turn

Why: Adding a half turn to the argument changes nothing.

\[ \tan(2x + \pi) = \tan(2x) \]

Rewrite the whole function

Why: Now the coefficient of x is positive and there is no phase shift left.

\[ y = -\frac13\tan(2x) + 1 \]

Set the argument equal to the five tangent marks and solve

Why: Divide every mark by two.

\[ 2x = -\tfrac{\pi}{2}, -\tfrac{\pi}{4}, 0, \tfrac{\pi}{4}, \tfrac{\pi}{2} \;\Longrightarrow\; x = -\tfrac{\pi}{4}, -\tfrac{\pi}{8}, 0, \tfrac{\pi}{8}, \tfrac{\pi}{4} \]

Substitute the three inner values

Why: Take the tangent, multiply by minus a third, then add one.

\[ y\!\left(-\tfrac{\pi}{8}\right) = \tfrac43, \quad y(0) = 1, \quad y\!\left(\tfrac{\pi}{8}\right) = \tfrac23 \]

Figure (svg): One cycle of one minus a third of the tangent of two x, falling between asymptotes at minus pi over four and pi over four, through the points at heights four thirds, one and two thirds.

The three marked heights are squeezed towards the line y = 1 by the factor of one third, but the asymptotes are set by the two inside and do not move.

Check the period two ways

Why: The asymptotes are a quarter turn apart, and a half turn divided by the frequency of two gives the same number.

\[ \frac{\pi}{4} - \left(-\frac{\pi}{4}\right) = \frac{\pi}{2} = \frac{\pi}{2} \]

Stitz & Zeager, College Trigonometry, Ch. 10 Foundations of Trigonometry — §10.5 Graphs of the Trigonometric Functions §10.5, p. 809 — Exercise 15

The first four lines of this solution are all about avoiding trouble rather than making progress, and they are the most valuable part. A negative coefficient of the input makes the five solved values come out in reverse order, which is where mirrored graphs come from, so remove it before you start.

Two facts do the removing. The tangent is odd, so a minus sign inside can be pulled out to the front. The period is a half turn, so a half turn added inside can simply be dropped. After both moves the phase shift has vanished entirely and the function is as simple as it could be.

Once rewritten, the rest is routine. Note that the three heights, four thirds, one and two thirds, are close together: that is the factor of one third at work, squeezing them towards the line at height one. The asymptotes, set by the two inside, are unaffected by it.

53. Exercise 24: the same graph from a different formula

Worked example

Two exercises apart in the book, and the answer turns out to be a curve you have already drawn.

\[ y = \frac13\cot\!\left(2x + \frac{3\pi}{2}\right) + 1 \]

Reduce the phase using the period of the cotangent

Why: Subtracting a half turn from the argument changes nothing.

\[ \cot\!\left(2x + \tfrac{3\pi}{2}\right) = \cot\!\left(2x + \tfrac{\pi}{2}\right) \]

Set the argument equal to the five cotangent marks

Why: Use the reduced argument.

\[ 2x + \tfrac{\pi}{2} = 0, \; \tfrac{\pi}{4}, \; \tfrac{\pi}{2}, \; \tfrac{3\pi}{4}, \; \pi \]

Solve for x

Why: Subtract a quarter turn, then halve.

\[ x = -\tfrac{\pi}{4}, \; -\tfrac{\pi}{8}, \; 0, \; \tfrac{\pi}{8}, \; \tfrac{\pi}{4} \]

Substitute the three inner values

Why: Take the cotangent, multiply by a third, then add one.

\[ y\!\left(-\tfrac{\pi}{8}\right) = \tfrac43, \quad y(0) = 1, \quad y\!\left(\tfrac{\pi}{8}\right) = \tfrac23 \]

Read the period

Why: The distance between the outer two values.

\[ \frac{\pi}{4} - \left(-\frac{\pi}{4}\right) = \frac{\pi}{2} \]

Check against the previous exercise

Why: The cofunction identity says a cotangent of a quantity plus a quarter turn is minus the tangent of that quantity, which turns this formula into the previous one; the numbers agree at a test input as well.

\[ \tfrac13\cot\!\left(2x + \tfrac{\pi}{2}\right) + 1 = -\tfrac13\tan(2x) + 1; \quad \text{at } x = 0.2: \; 0.859069 \text{ both ways} \]

Stitz & Zeager, College Trigonometry, Ch. 10 Foundations of Trigonometry — §10.5 Graphs of the Trigonometric Functions §10.5, p. 809 — Exercise 24, compared with Exercise 15

This exercise sits nine numbers after the previous one in the book, and it is the same curve. That is a nice accident and also a useful lesson about how many different formulas can describe one graph.

The first move is the cheapest simplification available on these two functions: the argument contains three quarters of a full turn, and since the period is a half turn you may subtract one and work with a quarter turn instead. Always look for that reduction before you start solving equations.

The check at the end names the reason for the coincidence. A cotangent of a quantity plus a quarter turn is minus the tangent of that quantity, which converts this formula into the previous one exactly. The numerical agreement at a randomly chosen input confirms it, and it is a good habit whenever you claim two formulas describe one function.

54. Trap: calling the leading coefficient an amplitude

Trap

The trap

Asked for the amplitude of this cotangent, a student reads off the fraction in front:

\[ y = \tfrac13\cot(2x) + 1 \]

\[ A = \tfrac13? \]

Wrong. There is no amplitude to report at all.

The fix

Amplitude is half the distance between the highest value and the lowest value. This graph has no highest value and no lowest value, so that distance is not a number and the word does not apply. The correct answer to the question is that the function has no amplitude.

\[ \text{no max, no min} \]

\[ \Longrightarrow \; \text{no amplitude} \]

The one third is still doing something, and it deserves a name: it is a vertical stretch by a factor of one third, measured from the line the vertical shift puts at height one. Call it the vertical stretch and the answer is right.

This is the same trap as the earlier one about maxima, in the place where it actually costs marks: a question that asks for four parameters, one of which does not exist for this function.

The right answer is to say there is no amplitude, and to say why in one clause: the function has no largest and no smallest value, so there is no gap to halve. An answer that says none with a reason is complete. An answer that leaves the box blank looks like an omission.

Then give the coefficient the name it deserves. It is a vertical stretch, measured from the line the vertical shift sets, and together with its sign it tells you how steep the branches are and which way they run. Everything the number actually does is captured by that description.

55. One crossing per branch

Concept

Because every branch increases through every real number exactly once, a horizontal line cuts each branch in exactly one place. A sinusoid, which turns around inside each period, usually cuts twice.

Figure (svg): Two panels. Left: the tangent curve with a horizontal line at height two, crossing each branch exactly once. Right: the sine curve over two full periods with a horizontal line at height one half, crossing it twice in each period.

A curve that rises the whole way across a branch can meet a horizontal line only once there. A curve that turns around meets most lines twice.

\[ \tan(x) = 2 \;\Longrightarrow\; x = \arctan(2) + \pi k = 1.1071 + \pi k \]

That is why a tangent equation has one family of solutions and a sine equation usually has two. It is the first thing Section 10.7 will use, and the reason the inverse tangent in the next lesson needs no awkward choice of branch.

Stitz & Zeager, College Trigonometry, Ch. 10 Foundations of Trigonometry — §10.5 Graphs of the Trigonometric Functions §10.5, p. 806 — the range and monotonicity in Theorem 10.25

Compare the two panels and count. The horizontal line meets each tangent branch once and only once, because the branch rises the whole way across; the same line meets each period of the sine twice, because the sine turns around inside the period.

That counting difference is the practical reason the tangent is easier to solve equations with. One crossing per branch means one family of solutions, all spaced a half turn apart. Two crossings per period means two families, and you have to find both, which is where most errors in Section 10.7 come from.

It is also why the next lesson can invert the tangent so comfortably. A single branch already passes the horizontal line test with nothing thrown away, so restricting the tangent to one branch costs you no values at all, whereas restricting the sine costs you most of them.

56. The lighthouse beam

Real world

Figure (svg): A diagram: a straight shoreline drawn as a horizontal line across the top, a lighthouse marked below it, a dashed perpendicular of length d joining them, and three beams leaving the lighthouse at twenty, fifty-five and seventy degrees, striking the shore at points that are spaced ever further apart.

Equal turns of the beam do not sweep equal stretches of shore. Near the perpendicular the spot crawls; near parallel it sprints, because the tangent has an asymptote there.

Discussion prompt

A lighthouse stands one kilometre from a straight shoreline, and its beam turns at a steady rate. Write the distance of the lit spot from the nearest point of the shore in terms of the angle the beam makes with the perpendicular, and say what the asymptote means physically.

Write your formula before revealing. The geometry is a single right triangle: the perpendicular from the lighthouse to the shore is one leg, the stretch of shore is the other, and the beam is the hypotenuse.

Look at the three beams in the picture and at the two figures in the box. Turning the beam from twenty degrees to fifty-five moves the spot a certain distance along the shore. Turning it a much smaller amount further, from fifty-five to seventy, moves the spot almost twice as far. The beam turns at a constant rate; the spot does not travel at a constant speed.

The asymptote is the physical limit of that behaviour. As the beam approaches parallel with the shore the spot runs away without bound, and at exactly parallel the beam never meets the shore at all. That missing value is not a flaw in the model; it is the model telling you the truth about what happens at parallel.

57. How far round before the spot is ten kilometres out?

Estimation

\[ \tan(\theta) = 10 \]

Predict first

With the lighthouse one kilometre from the shore, at roughly what angle from the perpendicular is the lit spot ten kilometres away?

  • About 45 degrees
  • About 63 degrees
  • About 84 degrees
  • About 89.5 degrees

Correct: About 84 degrees

Why: The spot is ten kilometres out when the tangent of the angle is ten, which happens at 1.4711 radians, or 84.29 degrees. The beam has already turned most of the way to parallel, and the remaining six degrees carry the spot from ten kilometres out to beyond the horizon. Guessing 45 or 63 degrees badly underestimates how fast the tangent grows near its asymptote.

\[ \arctan(10) = 1.4711 \text{ rad} = 84.29^{\circ} \]

Commit to one of the four options. Most people choose one of the first two, because sixty-three degrees feels like a long way round and ten kilometres feels like a long way out.

The tangent grows far faster than intuition allows near its asymptote. At forty-five degrees the spot is only one kilometre out. At sixty-three it is about two. Ten kilometres needs eighty-four degrees, which is most of the way to parallel.

Then consider what is left. From eighty-four degrees to ninety is only six degrees of turning, and in those six degrees the spot travels from ten kilometres out to beyond any distance you care to name. That is what an asymptote feels like from inside a physical problem.

58. Putting it together

Section

Part 6

59. Pattern: graphing one cycle of a tangent or cotangent

Pattern

Figure (svg): A flow diagram: first make the coefficient of x positive, then branch on whether the function is a tangent or a cotangent to choose one of two sets of quarter marks, then solve the argument equations, substitute, and read the period off the two asymptotes.

Two decisions and three mechanical steps. The only branch point in the whole procedure is which function you are looking at.
  1. Make the coefficient of x positive first, using the odd identity and the period if you have to.
  2. Choose the quarter marks by function: minus a quarter turn to a quarter turn for tangent, zero to a half turn for cotangent.
  3. Set the argument equal to each of the five marks and solve for x. The outer two give asymptotes; the inner three give points.
  4. Substitute the three inner values into the whole function, remembering the stretch and the vertical shift.
  5. Read the period as the distance between the two asymptotes, and check it against the parent's period divided by the frequency.
  6. Never report an amplitude. Report the vertical stretch and the branch direction instead.

This is the order to work in every time, and the diagram shows that there is only one genuine decision in it: tangent or cotangent, which chooses your five marks.

Step one is the one people skip. If the coefficient of the input is negative, rewrite before doing anything else, using the fact that both functions are odd and that both have a period of a half turn. Skipping it does not always produce a wrong answer, but it produces a reversed list of values that is very easy to misread.

The last two lines are the ones markers look for. Reading the period off the two asymptotes and then confirming it against the parent's period divided by the frequency costs one line and catches almost every slip. And whatever the question asks, do not invent an amplitude; report the vertical stretch and the direction of the branches instead.

60. Order the graphing procedure

Ranking

Put in order

Put the steps of a complete graphing argument in the order you would carry them out.

  1. Rewrite so the coefficient of x is positive
  2. Choose the five quarter marks that belong to this parent function
  3. Set the argument equal to each mark and solve for x
  4. Substitute the three inner values into the whole function
  5. Read the period as the gap between the two asymptotes

Why: Rewriting comes first, because a negative coefficient of x makes the five solved values come out in reverse order and is the commonest source of a mirrored graph. Choosing the marks comes before any algebra, since it is a decision about which parent you are looking at. The period comes last because it is read off the asymptotes you have just found.

Drag the steps into order before checking, and pay particular attention to where you put the rewriting step.

Rewriting comes first because everything downstream depends on the coefficient of the input being positive. Choosing the marks comes second because it is a decision about which parent you are looking at, and it cannot be made from the algebra.

Reading the period comes last for a reason worth noticing: it is read off the asymptotes, which you only have once you have solved the five equations. Quoting a period at the start from a remembered formula is how the earlier error slide happened.

61. All six graphs at last

Comparison

Figure (svg): Six small graphs in two rows over one full turn: cosine and sine bounded between minus one and one; secant and cosecant in branches outside that band; tangent and cotangent running through every height between their asymptotes.

Four of the six repeat only after a full turn; the last two are done after half of one. Only the first two are bounded, and only the first two have an amplitude.

Comparison matrix

functionperiodrangeasymptotes at
cos x2π[−1, 1]none
sin x2π[−1, 1]none
sec x2πvalues of size at least 1odd multiples of π/2
csc x2πvalues of size at least 1multiples of π
tan xπall real numbersodd multiples of π/2
cot xπall real numbersmultiples of π

Take a moment over the six panels, because this is the end of a long chapter and the whole picture is now on one sheet. Two curves are bounded and smooth, two come in branches outside a forbidden band, and two run through every height between their asymptotes.

Fill in the blanks in the table from the pictures rather than from memory, and then check them against what you know. The last two rows are the only ones with a period of a half turn, and the last two rows are the only ones whose range has no gap and no bound.

Two columns are worth reading down. Only the first two functions have an amplitude, because only they have a largest and a smallest value. And every asymptote in the whole table sits where the relevant parent, cosine or sine, is zero, which is the single rule that generates all four of the lower graphs.

62. Check: the period of a stretched tangent

Check

Check your understanding

What is the period of y = 4 tan(3x) − 7?

  • A. 2π/3
  • B. π/3 (correct)
  • C. π
  • D. 3π

Answer: B

Why: The parent tangent has period π, and a frequency of 3 divides that by 3. Confirm it with the asymptotes: 3x equals plus or minus π/2 at x equals plus or minus π/6, and those are π/3 apart.

Why A tempts people
2π over the frequency is the sinusoid rule. The tangent's parent period is π, not 2π.
Why C tempts people
π is the parent's period; the 3 inside compresses the graph, so the period must be smaller.
Why D tempts people
Multiplying by the frequency instead of dividing by it. A larger frequency means a shorter period.

The question is testing one thing: whether you divide a half turn or a whole turn by the frequency. Three of the four options come from different ways of getting that wrong.

The parent tangent finishes in a half turn, so the period is a half turn divided by three. Confirm it independently by setting the argument equal to plus and minus a quarter turn, which puts the asymptotes a sixth of a half turn either side of the origin, a third of a half turn apart.

Neither the four in front nor the seven subtracted has any effect on the period. Outside changes cannot move an asymptote, so they cannot change the distance between two of them.

63. Check: where the asymptotes go

Check

Check your understanding

Where are the vertical asymptotes of y = cot(x + π/4)?

  • A. x = π/4 + πk
  • B. x = −π/4 + πk (correct)
  • C. x = π/4 + πk/2
  • D. x = −π/4 + 2πk

Answer: B

Why: A cotangent is undefined where its argument is a multiple of π. Set x + π/4 equal to πk and solve: x = −π/4 + πk. The shift is to the LEFT because the quantity added inside is positive.

Why A tempts people
Shifting in the wrong direction. Solving x + π/4 = πk subtracts π/4; it does not add it.
Why C tempts people
Steps of π/2 would need a frequency of 2 inside. There is no coefficient on x here, so the asymptotes stay a full π apart.
Why D tempts people
Steps of 2π would leave out half of the asymptotes. The cotangent has one every π.

Solve rather than guess. The cotangent is undefined where its argument is a whole number of half turns, so write that equation down and solve it for the input.

The direction of the shift is where the most tempting wrong answer comes from. A positive quantity added inside means the graph moves in the negative direction, so the asymptotes move left, not right. Solving the equation gets this right automatically, which is exactly why you should solve it.

The spacing is the second thing to check. There is no coefficient on the input here, so the asymptotes stay a full half turn apart. Any option with a different spacing can be eliminated before you even work out where the first one is.

64. Check: which branches fall

Check

Check your understanding

Which of these has branches that fall from left to right?

  • A. y = cot(x) + 5 (correct)
  • B. y = −cot(x)
  • C. y = tan(x) − 5
  • D. y = 3 tan(x)

Answer: A

Why: A cotangent branch falls by default, and adding 5 only lifts the whole picture, so the direction survives. Direction is decided by the parent together with the sign of the leading coefficient, and by nothing else.

Why B tempts people
The minus sign reflects the falling cotangent into a rising one.
Why C tempts people
A tangent branch rises, and subtracting 5 only lowers the picture; it does not turn it over.
Why D tempts people
A positive factor of 3 stretches the rising tangent but leaves it rising.

Direction is decided by two things and only two things: which parent you have, and the sign of the leading coefficient. Vertical shifts are irrelevant, and so is the size of the coefficient.

A cotangent branch falls by default, so a positive coefficient leaves it falling and a negative one makes it rise. A tangent branch rises by default, so the opposite applies. Two of the options here are sinusoid-style distractions, offering you a vertical shift as though it could turn a graph over.

If you hesitate, sketch the parent and then apply the sign. Five seconds of drawing beats any attempt to hold four transformed graphs in your head at once.

65. Explain why there is no amplitude

Explain it to yourself

Discussion prompt

In two or three sentences, say what amplitude measures, and explain why none of secant, cosecant, tangent and cotangent has one. Then say which of the sinusoid's other three parameters still make sense for a tangent.

Write your answer out before revealing it. The skill being practised is stating a limitation precisely, which is a different skill from computing.

A complete answer has three parts. Say what amplitude measures, which is half the gap between the largest and smallest values. Say why that fails here, which is that all four of these functions are unbounded in at least one direction. Then say what does still make sense, which is period, phase shift and vertical shift.

The last part is where the extra credit is. Naming the vertical stretch as the thing that replaces amplitude shows you understand that the number in the formula is still doing a job, even though the word attached to it in the sinusoid lesson no longer applies.

66. Exit ticket

Exit ticket

\[ y = 4\cot(3x) - 2 \]

Discussion prompt

Give the period, the two asymptotes bounding the cycle that starts at the origin, the three points between them, and say whether the branch rises or falls.

This one problem uses everything: the right quarter marks, the right period rule, the substitution, and the direction question that the sinusoid never asked.

The parent is a cotangent, so the marks run from zero to a half turn, and with a frequency of three the five solved inputs are a twelfth of a half turn apart. The two outer ones, at the origin and at a third of a half turn, are the asymptotes, and a third of a half turn is the period.

The direction is the part that is easiest to lose. A cotangent falls and the leading coefficient of four is positive, so the branch falls, and the three heights two, minus two and minus six confirm it. If you were asked for the amplitude, the answer would be that there is none.

67. Recap

Recap

tangentcotangent
quotientsine over cosinecosine over sine
asymptotesodd multiples of π/2multiples of π
zerosmultiples of πodd multiples of π/2
branchesrisingfalling
fundamental cycle(−π/2, π/2)(0, π)
periodππ
range and parityall reals, oddall reals, odd

\[ \text{period of } A\tan(\omega x + \varphi) + B \;=\; \frac{\pi}{\omega}, \qquad \omega > 0 \]

The whole section is now behind you: six functions, six graphs. Section 10.6 turns them round and asks for inverses, and the tangent goes first, because a single branch of it already passes the horizontal line test with nothing thrown away.

OpenStax Algebra and Trigonometry 2e, §8.2 Graphs of the Other Trigonometric Functions §8.2 — the same two graphs in OpenStax

One table holds the whole section now. Every row of it comes from the single sentence at the top: one function is the sine over the cosine and the other is the cosine over the sine, so wherever one has a zero the other has an asymptote.

Two facts are worth carrying out of this lesson above all the others. The period is a half turn, for both functions, because both read a ratio of coordinates rather than a coordinate, and a ratio is blind to the sign change that half a turn causes. And there is no amplitude, for either function, because neither has a largest or a smallest value to measure between.

Section 10.6 turns all six functions round and asks for their inverses. The tangent goes first and is the easiest of the six, and the reason is on this page: a single branch of it already climbs through every real number exactly once, so restricting it to one branch throws nothing away.

Sources

  1. Stitz & Zeager, College Trigonometry, Ch. 10 Foundations of Trigonometry — §10.5 Graphs of the Trigonometric Functions — Stitz & Zeager, College Trigonometry, Version 3 Corrected Edition, July 4 2013, pp. 804-809
  2. OpenStax Algebra and Trigonometry 2e, §8.2 Graphs of the Other Trigonometric Functions

Want this taught 1-on-1? Alexander tutors Trigonometry — $55/session, free consultation.

Book on Wyzant · Text (657) 465-8108