Trigonometry Enrichment Program

Karl Hahn

c 2010, 2012 March 2012 version adds the Cotangent rule and a new identity proof

How do you solve trig identities and trig equations? Each one can seem
like a completely new puzzle, on which the methods you used on the last
one no longer apply. But there are some basics. Your tools are the following
identities which you should memorize.
Symmetry Identities

cos (−x) = cos (x) sin (−x) = − sin (x) tan (−x) = − tan (x)

sec (−x) = sec(x) csc (−x) = − csc (x) cot (−x) = − cot (x)
The Pythagorean Identity

sin2 (x)+cos2 (x) = 1 hence cos2 (x) = 1−sin2 (x) and sin2 (x) = 1−cos2 (x)

Cosine of sums and differences Identities

cos (x + y) = cos (x) cos (y) − sin (x) sin (y)

cos (x − y) = cos (x) cos (y) + sin (x) sin (y)

Sine of sums and differences Identities

sin (x + y) = sin (x) cos (y) + cos (x) sin (y)

sin (x − y) = sin (x) cos (y) − cos (x) sin (y)

Definitions of tangent, cotangent, secant, and cosecant

sin (x) cos (x) 1

tan (x) = cot (x) = =
cos (x) sin (x) tan (x)
1 1
sec (x) = csc (x) =
cos (x) sin (x)

1
From the above you can easily derive these:
1 sin2 (x) + cos2 (x) sin2 (x) cos2 (x)
sec2 (x) = = = + = tan2 (x) + 1
cos2 (x) cos2 (x) cos2 (x) cos2 (x)
Derivation of a similar identity involving csc2 and cot2 is left as an exercise
to the reader.
Double angle formulas:
sin (2x) = sin (x + x) = sin (x) cos (x) + sin (x) cos (x) = 2 sin (x) cos (x)
cos (2x) = cos (x + x) = cos (x) cos (x) − sin (x) sin (x) = cos2 (x) − sin2 (x)
If you replace sin2 (x) with 1 − cos2 (x) in the latter, you get
cos (2x) = 2 cos2 (x) − 1
Likewise if you replace cos2 (x) with 1 − sin2 (x) in the same equation, you
get
cos (2x) = 1 − 2 sin2 (x)
For tangent of the double angle:
sin(x)
sin (2x) 2 sin (x) cos (x) 2 cos(x) 2 tan (x)
tan (2x) = = = =
cos (2x) cos2 (x) − sin2 (x) cos2 (x)
− sin2 (x) 1 − tan2 (x)
cos2 (x) cos2 (x)

You can use a similar approach to show that for any x and y:
tan (x) + tan (y)
tan (x + y) =
1 − tan (x) tan (y)

So much for the ones you need to memorize. From these and the application
of algebra you can prove all the other identities and solve all the trig equations
that you might be assigned to do as homework problems.
Triple angle formulas:
sin (3x) = sin (x + 2x) = sin (x) cos (2x)

+ sin (2x) cos (x)
2 2 2
= sin (x) cos (x) − sin
(x) + 2 sin (x) cos (x)

= sin (x) 1 − 2 sin2 (x) + 2 sin (x) 1 − sin2 (x)
= sin (x) − 2 sin3 (x) + 2 sin (x) − 2 sin3 (x)
= 3 sin (x) − 4 sin3 (x)

cos (3x) = cos (x + 2x) = cos (x) cos (2x)

− sin (x) sin (2x)
= cos (x) cos2 (x) − sin2 (x) − sin (x) (2 sin (x) cos (x))
= cos (x) (2 cos2 (x) − 1) − 2 sin2 (x) cos (x)
= 2 cos3 (x) − cos (x) − 2 (1 − cos2 (x)) cos (x)
= 4 cos3 (x) − 3 cos (x)

2
Problem: Find the coordinates of the vertices of a pentagon.
Assume the circle has ra-
dius of 1. If θ is the an-
gle from one vertex to the
next then the angle from
the x-axis going counter-
clockwise to (x1 , y1) is 2θ.
The angle from the x-axis
going counterclockwise to
(x2 , y2 ) is 3θ. Observe that
x1 = x2 . Since x1 = cos (2x)
and x2 = cos (3θ), it follows
that cos (2θ) = cos (3θ). We
apply the double and triple
angle formulas to form an
equation that we can solve
for cos (θ).
cos (2θ) = cos (3θ)
2 cos2 (θ) − 1 = 4 cos3 (θ) − 3 cos (θ)
0 = 4 cos3 (θ) − 2 cos2 (θ) − 3 cos (θ) + 1
If you replace cos (θ) with x, you have the cubic polynomial,
0 = 4x3 − 2x2 − 3x + 1. Observe that θ = 0 must be a solution because
twice zero is equal to three times zero. Since cos (0) = 1, it must be that
x = 1 is a solution. So we can divide it out using polynomial long division.
4x2 + 2x − 1
4x3 − 2x2 − 3x + 1


x−1
− 4x3 + 4x2
2x2 − 3x
− 2x2 + 2x
−x+1
x−1
0
It’s easy enough to apply the quadratic formula to the resuling quotient:
√ √
−2 ± 4 + 16 ± 5−1
x= =
8 4
So why two solutions? We expected that solving this would give us x in the
(x, y) shown in the figure. That would be the positive solution. The other

3
solution is x1 , which we already determined is the same as x2 . Our original
setup was to solve for the cosine of the angle whose double and triple angles
have the same cosine. Notice if you take twice the angle from the x-axis to
(x1 , y1 ) (which takes you nearly around the circle to the lower-right vertex),
and three times that same angle (which takes you fully around to (x, y)),
both of those destinations have the same x coordinate and hence the same
cosine. So the negative solution is legitimate also.
To find the y coordinate of (x, y), we use the Pythagorean identity,
sin2 (θ) = 1 − cos2 (θ). Hence
√ !2 √
5−1 5− 5
y 2 = 1 − x2 = 1 − =
4 8

By taking the the square root of the above result you arrive at y.
From the diagram you can see that y2 = −y1 . So we could have solved for
θ for which sin (2θ) = − sin (3θ). By applying the double and triple angle
formulas for sine:

2 sin (θ) cos (θ) = 4 sin3 (θ) − 3 sin (θ)

Clearly again θ = 0 is a solution because sin (0) = 0. But we eliminate that

solution by dividing the common factor of sin (θ) out of the above.

2 cos (θ) = 4 sin2 (θ) − 3

Now replace sin2 (θ) with 1 − cos2 (θ)

2 cos (θ) = 4−4 cos2 (θ)−3 = 1−4 cos2 (θ) hence 4 cos2 (θ)+2 cos (θ)−1 = 0

If you again replace cos (θ) with x, you get the same quadratic we got doing
it the first way, and therefore the same solutions.

4
Quad-angle formulas:

cos (4x) = cos (2x + 2x)

= 2 cos2 (2x) − 1
2
= 2 (2 cos2 (x) − 1) − 1
= 8 cos4 (x) − 8 cos2 (x) + 2 − 1
= 8 cos4 (x) − 8 cos2 (x) + 1

sin (4x) = sin (2x + 2x)

= 2 sin (2x) cos (2x)
= 4 sin (x) cos (x) (2 cos2 (x) − 1)
= 8 sin (x) cos3 (x) − 4 sin (x) cos (x)
or
4 sin (x) cos (x) 1 − 2 sin2 (x)


=
= 4 sin (x) cos (x) − 8 sin3 (x) cos (x)

We can use quad-angle formulas to solve the pentagon problem also. If θ is

the angle from the x-axis to (x, y), then 4θ takes you from the x-axis to the
lower right vertex, which has the same x value. So by solving for θ where
cos (θ) = cos (4θ), we should arrive at the same solutions as before. Applying
the quad-angle formula for cosine:

cos (θ) = 8 cos4 (θ) − 8 cos2 (θ) + 1

0 = 8 cos4 (θ) − 8 cos2 (θ) − cos (θ) + 1

This time the polynomial to solve is 8x4 − 8x2 − x + 1 = 0. As before, θ = 0

is a trivial solution that corresponds to x = cos (0) = 1. Dividing out that
root,
8x3 + 8x2 −1
4 2


x−1 8x − 8x − x + 1
− 8x4 + 8x3
8x3 − 8x2
− 8x3 + 8x2
−x+1
x−1
0
We see that the quotient is a cubic. We expect that we should get the same
two solutions solving the pentagon problem this way as we did solving it
using the double and triple angle formulas. Yet a cubic must have either one
or three real roots – never two. So is there a solution to cos (θ) = cos (4θ)

5
that we haven’t accounted for? You have to think about it for a few minutes.
The unaccounted for solution, it turns out, has nothing to do with pentagons.
It arises from the fact that going around the circle 13 times ( 2π
3
radians or
◦ 4 8π ◦
120 ) is the same as going around 3 times
( 3 1radians or 480 ). It means
2π 2π
that θ = 3 is a solution. Since cos 3 = − 2 , it must be the case that
2x + 1 evenly divides 8x3 + 8x2 − 1.
4x2 + 2x − 1
8x3 + 8x2


2x + 1 −1
3 2
− 8x − 4x
4x2
− 4x2 − 2x
− 2x − 1
2x + 1
0
This, as you can see, results in the same quadratic as we got solving the
pentagon problem twice before.
Application to compass and straight-edge construction: Algebra
and geometry are more intertwined than perhaps you had imagined. The
quadratic solution we arrived at for the vertices of a pentagon implies that
a regular pentagon can be constructed using a compass and straight-edge.
The reason has to do with constructable lengths. If you choose some unit
length, then any integer multiple of that length is constructable simply by
walking the compass down the straight-edge that number of times. Likewise
the sum or difference of any two constructable lengths is also constructable.
In addition the square root of any constructable length is constructable. This
is because if x is a constructable length then so are x + 1 and |x − 1|. By
constructing two perpendicular lines then measuring |x − 1| from their in-
teresection along one of them, it is possible to construct a right triangle
whose one side is |x − 1| and√whose hypotenuse√ is x + 1. By Pythagoras,
the remaining side must be 2 x. You find x by bisecting that remaining
side. Indeed if you have any constructable length, that length divided by any
power of 2 is also constructable by repeated bisection (note that it is also
possible to n-sect a line segment for any positive integer n, so any rational
fraction of a constructable length is also constructable).
The consequence is that the solutions we found to the quadratic,
4x2 + 2x − 1,
√ are constructable using compass and straight-edge. This is
because 1 ± 5 is constructable, as is that value divided by 4 by virtue of
4 being an integer divisor. It follows that a regular pentagon is also con-

6
structable.
But suppose we were to try the same approach to finding the vertices of a
regular heptagon (a figure with 7 equal sides and angles). We would use the
triple angle formula together with the quad-angle formula to find θ where
cos (3θ) = cos (4θ). Applying the formulas:

4 cos3 (θ) − 3 cos (θ) = 8 cos4 (x) − 8 cos2 (θ) + 1

This results in having to solve the polynomial, 8x4 − 4x3 − 8x2 + 3x + 1 = 0.

Even when you divide out the trivial root, x − 1, you still have the cubic,
8x3 + 4x2 −4x−1 = 0. This cubic has three real roots that can be found only
by taking cube roots of values derived from the cubic’s coefficients. Compass
and straight-edge constructions can make constructable lengths of square
roots of other constructable lengths, but no combination of summing and
taking square roots can ever find the cube root of any value other than zero
and one. It follows that no matter how clever you are, you will never be able
to construct a perfect regular heptagon using only compass and straight-edge.
Half-angle formulas: Because we know that cos (2θ) = 2 cos2 (θ) − 1, we
can derive a half-angle formula:
 
θ θ
cos (θ) = cos +
2 2
 
θ 2
= 2 cos −1
2
 
1 θ
2
(cos (θ) + 1) = cos
2 2
r  
1 θ
(cos (θ) + 1) = cos
2 2

Likewise using cos (2θ) = 1 − 2 sin2 (θ) we can derive,

r  
1 θ
(1 − cos (θ)) = sin
2 2

The impossibility of trisecting angles. We now extend our discussion of

compass and straight-edge constructions. You can see from the cosine
half-
θ
angle formula that if cos (θ) is a constructable length, then cos 2 is also

7
a constructable length. Why? Because it can be calculated using a sum of
constructable lengths, division by an integer, and taking a square root. This
implies that it is possible to bisect any angle using compass and straight-
edge alone. But suppose we used the triple angle formula in the same way
to derive a formula for cosine of 31 of a known angle, θ?
 
θ
cos (θ) = cos 3
3
   
θ
3 θ
= 4 cos − 3 cos
3 3

Suppose a constructable
length, `, is known to be equal to cos (θ) and we

want to find x = cos θ3 . Then we would have to solve the polynomial,

4x3 − 3x − ` = 0

Except for special values of ` such as ` = 0, solving this cubic for x will
involve extracting cube roots. Again no amount of summing, dividing by
integers, and taking square roots can ever result in the cube root of any
value except zero and one. This means that, except in special cases, x is an
unconstructable length. But if angle 3θ were constructable, its cosine would
also be constructable. It follows that it is impossible to trisect an arbitrary
angle using only compass and straight-edge.
Challenge problem: Apply the double angle and triple angle formulas for
sine and cosine to cos (5θ) = cos (3θ + 2θ) to arrive at a quint-angle formula
for cosine that is strictly in terms of powers of cos (θ). When you have done
the algebra, see [Link]
to compare your result with the correct answer. Use this result together
with the quad-angle formula for cosine to arrive at a polynomial that must
be solved to determine the vertices of a regular nonagon (the 9-sided figure
with equal sides and angles). Do this by expanding cos (5θ) = cos (4θ).
Product formulas for Sine and Cosine and their applications. From
the double angle formula for cosine, we can easily derive another formula for
the square of the cosine in terms of the double angle.

cos (2θ) = 2 cos2 (θ) − 1

1
cos2 (θ) = (1 + cos (2θ))
2
If you replace 1 with cos (0), you have

8
 1
cos2 (θ) = (cos (0) + cos (2θ))
2
Look at this result carefully. We are taking the product, cos (θ) × cos (θ),
and ending up with a sum involving cosine of θ − θ and cosine of θ + θ. That
is the product of the cosines of two angles (which in this case are identical)
yields a sum involving cosines of the sum of the two angles and the difference
of two angles. Of course so far we have seen this to be the case only when
those two angles are equal. But this pattern is generally true. Here’s the
proof based upon the cosine-of-a-sum formula.

cos (θ − φ) = cos (θ) cos (φ) + sin (θ) sin (φ)

cos (θ + φ) = cos (θ) cos (φ) − sin (θ) sin (φ)
cos (θ − φ) + cos (θ + φ) = 2 cos (θ) cos (φ)

If you take the difference of the above equations you get the formula for
the product of sines. To get the cross-product of sine with cosine, use the
sine-of-a-sum formula:

sin (θ + φ) = sin (θ) cos (φ) + cos (θ) sin (φ)

sin (θ − φ) = sin (θ) cos (φ) − cos (θ) sin (φ)
sin (θ + φ) + sin (θ − φ) = 2 sin (θ) cos (φ)

In every case you find that the product formula is half the sum (or difference)
of sines or cosines of the sum and difference of θ and φ.
In engineering it is common for a signal to be a sinusoidal function of time.
The typical formula is y (t) = A sin (ωt + φ), where A is called the amplitude,
ω the radian frequency, and φ the phase. An experiment you can do if you
can find two identical tuning forks is this. Wrap rubber bands tightly around
each of the tines of one of the tuning forks. This will lower its frequency
slightly. Strike both tuning forks simultaneously, touch their handles to the
same surface, and listen. You will hear them beating. That is you will hear
the tone grow louder and softer at a fixed rate. If you can’t find tuning forks,
you can also do this experiment with musical instruments – two trumpets for
example. Tune one of them so it is slightly out of tune with the other. Then
have two trumpet players each blow a middle C. You will hear them beating
against each other. On a standard B-flat trumpet, middle C is around 465
Hz. Suppose one trumpet is tuned for middle C at 463 Hz. and the other
tuned to 467 Hz. The graph below shows what happens when both trumpets
are blown at once.

9
You can see that near the center of the graph, the signals from the two
trumpets are aligned. They re-enforce each other. On each side near the
outside of the graph, the two trumpets fight each other. When you add the
two signals you get a graph that looks like this:

To get the radian frequency of a signal, you multiply the [Link] by 2π.
You can imagine the signals from the two trumpets being A sin (2π (465 + 2) t)
and A sin (2π (465 − 2) t). From the formulas we developed, we expect this
sum to be equivalent to 2A sin ((2π × 465) t) cos ((2π × 2) t). Can you see
how the latter function is expressed in the graph? Near the middle (where
t = 0), the cosine function is close to unity. So at that time you hear the
trumpets loudly. Near the outside of the graph, the cosine function is close
to zero. At that time the two trumpets cancel each other. The frequency of
the beats is the frequency with which cos (2π × 2t) crosses zero – that is 4
beats per second.
AM radio broadcasts a signal whose frequency is in the range of 520 kHz to
1610 kHz. When you tune an AM radio to a station, you are selecting the
frequency in that range on which that station transmits. The program – say
the announcer’s voice – is encoded on the radio transmission by making the
signal stronger and weaker in time with the vibrations of the announcer’s
voice. Suppose an AM radio station were to transmit a signal that would
cause your radio receiver to produce a sinusoidal sound from the speaker.
The radio signal would look something like this:

We can describe this signal mathematically using the expression,

A sin (ω1 t) (1 + h cos (ω2 t)), where ω1 is the station’s broadcast radian fre-
quency and ω2 is the radian frequency of the sound the speaker is produc-
ing. Note that ω1 is much greater than ω2 . h, which must be between
0 and 1, is called the modulation index (in the graph, h = 0.4). When

10
you multiply this function out you get a component, A sin (ω1 t), which ra-
dio engineers call the carrier, and another component, Ah sin (ω1 t) cos (ω2 t).
Using the product formula we see that the latter is equivalent to
1
2
Ah sin ((ω1 + ω2 ) t) + 21 Ah sin ((ω1 − ω2 ) t). Radio engineers call these two
terms the sidebands because one component is ω2 higher than the carrier,
the other ω2 below the carrier – that is they are on either side of the car-
rier. The sidebands are not just theoretical. Their frequencies are broadcast
over the air along with the carrier. In the United States, regulations restrict
each AM broadcast station to 20 kHz. of the radio spectrum. This means
that the highest frequency of sound that an AM radio station can send to
the speaker of a radio receiver is 10 kHz. Attempting to broadcast higher
frequency sounds would result in sidebands that fall outside of the station’s
designated channel.
If you paid a lot of money for a home audio system, you would expect that
it faithfully reproduce the sounds played through it. This means that the
output signal (what comes out of the speakers) should be strictly a constant
multiple of the input signal (e.g., what’s on your CD or the program coming
from a radio station). That is, we want input, G (t), and output, H (t), to
have this relationship:

H (t) = kG (t)
where k is the amplification factor. Another way of putting this is that we
want the output to be a linear polynomial of the input. The degree to which
the output deviates from being a linear polynomial of the input is called
the distortion. So how might an audio engineer measure the distortion of
a sound system? We can employ the product formulas developed in the
previous paragraphs to see how. Supposing we were to allow the input to be
the sum of two sinusoidal signals, one at 2000 Hz., the other at 2200 Hz.

G (t) = A (sin (2π × 2000t) + sin (2π × 2200t))

If the system were perfect, we expect that the output should contain these
two frequencies and nothing else.

H (t) = kG (t) = kA (sin (2π × 2000t) +

sin (2π × 2200t))
But if the system were not perfect, then the transfer function – that is the
function that gives you the output in terms of the input – would be a higher
degree polynomial. Suppose we add a second degree term to the transfer
function:

11
 H (t) = kG (t) + ξ (G (t))2
Even if ξ is very small compared to k, there is a strategy by which we can
measure it. Expanding our sum of 2000 Hz. and 2200 Hz. using the transfer
function we get

H (t) = kA (sin (2π × 2000t) + sin (2π × 2200t)) +

ξA2 (sin (2π × 2000t) + sin (2π × 2200t))2
When you multiply out the square term you find that you have sine squared of
both the 2000 [Link] the 2200 [Link] present (which produce components
at 4000 Hz. and 4400 Hz. respectively). But you also find that you have a
term, 2ξA2 sin (2π × 2000t) sin (2π × 2200t). Our formula tells us that this is
the same as ξA2 (cos (2π × 200t) − cos (2π × 4200t)). Using electronic filters
it is possible to eliminate all of the higher frequencies and single out the 200
Hz. signal only. If the system were perfect (that is, if ξ = 0), there should be
no 200 Hz. component in the output whatsoever. The strength of whatever
200 Hz. signal is present in the output will be in proportion to magnitude of
ξ. And ξ is a measurement of the distortion produced by the system. This
type of test of an amplifier is called an intermodulation test.
Challenge: Suppose you were a technician evaluating an audio system.
You have already determined that ξ (G (t))2 = 0 to within the sensitivity of
your measurement. You suspect though that there is a higher order term,
ζ (G (t))3 , present in the system’s output that is measurably nonzero. You
need to set up an intermodulation test to measure this. What pair of fre-
quencies might you use as input in order to measure ζ by singling out and
measuring the amplitude of 200 Hz in the output?

Many Worked Trig Identities

What follows is a sampling of the many trig identity problems that people
have emailed me over time along with my worked solutions. Each solution is
on the page following the problem. Try to work each of them on your own
before looking at its solution. In each case where there are tangents, cotan-
gents, secants, and/or cosecants, I begin with the method of reconstructing
the problem in terms of sines and cosines only. This is not always the short-
est path to a result, but it always works, and so it is what I recommend to
use always as a first step to students who need a consistent methodology for
solving trig identity problems. Students who are able on their own to spot

12
the shorter method, where it can be found, are not likely to be in need of
the review provided here.
Building a Bridge Backward. The objective when you prove a trig iden-
tity is to build a bridge from some pair of expressions whose equality is obvi-
ous, step by step, to the equation whose equality you are wanting to prove.
But at the outset, you do not know what that pair of equal expressions is.
That is, you don’t know where the bridge begins. You find it by building the
bridge backward. That is you start with the equation whose equality you are
trying to prove, and step by step, work toward a pair of expressions whose
equality is obvious. But always keep in mind that you have to be able to
travel the bridge you build in the reverse direction from the way you build it.
That means that each step you take must be reversible. Here is an extreme
example of why that is. If you allow irreversible operations, you can prove
that 1 = 2 simply by multiplying both sides of that equation by zero. That
step is not reversible because you can never divide by zero. But multiplying
both sides of an equation by, say, 2, is allowable because the reverse, dividing
by 2, is always possible. Multiplying both sides by, say, sin (θ), is also allow-
able, but only with the understanding that it does not apply to values of θ
for which sin (θ) = 0, that is for values of θ that are integer multiples of π.
That understanding includes the requirement that the identity be proved by
other means for those θ’s. As you observe the steps that prove the identities
that follow, see if you can identify the discrete points for which the proof
might not apply – that is those values of θ for which some step of the proof
is irreversible. Challenge: What can you determine about the reversibility of
squaring both sides of an equation as part of a proof? That is, under what
conditions should that step be allowed?

1) Prove that
1
= sec2 (θ) − tan (θ) sec (θ)
sin (θ) + 1

13
 1
Replace sec (θ) wherever it occurs with . Replace tan (θ) wherever it
cos (θ)
sin (θ)
occurs with .
cos (θ)
1 1 sin (θ)
= −
sin (θ) + 1 cos (θ) cos2 (θ)
2

Multiply top and bottom of the left side by 1 − sin (θ). Use the difference of
squares on the bottom.

1 − sin (θ) 1 sin (θ)

2 = −
1 − sin (θ) cos (θ) cos2 (θ)
2

Use the Pythagorean identity on the bottom of the left side.

1 − sin (θ) 1 sin (θ)

2
= −
cos (θ) cos (θ) cos2 (θ)
2

Now simply observe the common denominator among the two terms to the
right of the equal to complete the proof.

2) Prove that

sin (θ) 1
=
sin (θ) − cos (θ) 1 − cot (θ)

14
 cos (θ)
Replace cot (θ) with
sin (θ)
sin (θ) 1
=
sin (θ) − cos (θ) 1 − cos(θ)
sin(θ)

Multiply top and bottom of the right-hand side by sin (θ)

sin (θ) sin (θ)

=
sin (θ) − cos (θ) sin (θ) − cos (θ)
Done.

3) Prove that

csc (θ) + cot (θ)

= cot (θ) csc (θ)
tan (θ) + sin (θ)

15
 sin (θ) cos (θ)
Replace tan (θ) with . Replace cot (θ) with . Replace csc (θ)
cos (θ) sin (θ)
1
with .
sin (θ)
1
sin(θ)
+ cos(θ)
sin(θ) cos (θ) 1
sin(θ)
=
+ sin (θ) sin (θ) sin (θ)
cos(θ)

Multiply top and bottom of the left side by sin (θ) cos (θ).

cos (θ) + cos2 (θ) cos (θ) 1

2 2 =
sin (θ) + sin (θ) cos (θ) sin (θ) sin (θ)
Find the common factors in the left side. On the top you can factor out
cos (θ). On the bottom you can factor out sin2 (θ).

cos (θ) (1 + cos (θ)) cos (θ) 1

2 =
sin (θ) (1 + cos (θ)) sin (θ) sin (θ)
Cancel the common factor on top and bottom of 1 + cos (θ). Multiply the
two terms of right hand side, and you’re done.
(( (
(
(1(+(cos
cos (θ) ( (θ)) cos (θ)
( ( (=
(
2
(1(+(cos (θ))
sin (θ) ( sin2 (θ)

Observe in the previous two problems the technique of multiplying the top
and bottom of a nasty fractions-inside-a-fraction in order to simplify it. Of
the denominators of the fractions inside the fraction, find the least common
multiple of them (often by multiplying all of those denominators all together).
That product will be what you multiply top and bottom of the big fraction
by.

4) Prove that

(sin (θ) − cos (θ))2 = 1 − 2 sin (θ) cos (θ)

16
Square out the left hand side:

sin2 (θ) − 2 sin (θ) cos (θ) + cos2 (θ) = 1 − 2 sin (θ) cos (θ)
Rearrange the terms on the left and then apply the Pythagorean identity,
and you’re done.

sin2 (θ) + cos2 (θ) − 2 sin (θ) cos (θ) = 1 − 2 sin (θ) cos (θ)
Observe that if you apply the double-angle formula for sine to the right, you
have the further identity:

(sin (θ) − cos (θ))2 = 1 − sin (2θ)

5) Prove that

1 + cos (2θ)
= cos (θ)
2 cos (θ)

17
Apply the double-angle formula for cosine. You may have observed it has
three different forms. When you try each one, you find that the one that is
useful is cos (2θ) = 2 cos2 (θ) − 1.

1 + 2 cos2 (θ) − 1
= cos (θ)
2 cos (θ)
Take the cancellation of the plus 1 with the minus 1, then cancel the common
factor of 2 cos (θ) from top and bottom, and you’re done.

1
+ 2 cos2 (θ) 
−1
= cos (θ)
2 cos (θ)

2
cos
2 (θ)
2
 
(θ)
cos  = cos (θ)

6) Prove that

1 − tan2 (θ)
= 1 − sec2 (θ)
1 − cot2 (θ)

18
 sin (θ) cos (θ)
Replace tan (θ) with . Replace cot (θ) with . Replace sec (θ)
cos (θ) sin (θ)
1
with .
cos (θ)
sin2 (θ)
1− cos2 (θ) 1
cos2 (θ)
=1−
1− cos2 (θ)
sin2 (θ)

Multiply both sides by cos2 (θ). This will get rid of the fraction on the right.
It will also get rid of the fraction in the numerator of the left hand side.

cos2 (θ) − sin2 (θ)

cos2 (θ)
= cos2 (θ) − 1
1− sin2 (θ)

Multiplying the denominator of the left side by sin2 (θ) is the same as dividing
that entire left side by the same. So we do that to the left as well as the
right.

cos2 (θ) − sin2 (θ) cos2 (θ) 1

2 2
= 2 − 2
sin (θ) − cos (θ) sin (θ) sin (θ)
The left hand side is clearly equal to −1. The right has two terms over a com-
mon denominator. So combine them. Note that cos2 (θ)−1 = − (1 − cos2 (θ)).
Apply the Pythagorean identity to finish it.

7) Prove that

tan (θ) sin (2θ) = 1 − cos (2θ)

19
 sin (θ)
Replace tan (θ) with .
cos (θ)
sin (θ)
sin (2θ) = 1 − cos (2θ)
cos (θ)
Apply the double angle formula for sine.

sin (θ)
(2 sin (θ) cos (θ)) = 1 − cos (2θ)
cos (θ)
Combine terms to simplify the left hand side.

2 sin2 (θ) = 1 − cos (2θ)

Apply the double angle formula for cosine to the right. Again there are three
forms of this identity. The one that is the most useful in this problem is
cos (2θ) = 2 cos2 (θ) − 1.

2 sin2 (θ) = 1 − 2 cos2 (θ) + 1 = 2 − 2 cos2 (θ)

Factor the 2 out of the right hand side. Apply the Pythagorean identity to
the result, and you’re done.

8) Prove that

1 − tan2 (θ)
= 1 − sec2 (θ)
1 − cot2 (θ)

20
 sin (θ) cos (θ)
Replace tan (θ) with . Replace cot (θ) with . Replace sec (θ)
cos (θ) sin (θ)
1
with .
cos (θ)
sin2 (θ)
1− cos2 (θ) 1
cos2 (θ)
=1−
1− cos2 (θ)
sin2 (θ)

Multiply both sides by cos2 (θ) to remove the cosines from both the denon-
minators in which they occur.

cos2 (θ) − sin2 (θ)

cos2 (θ)
= cos2 (θ) − 1
1− sin2 (θ)

Multiply top and bottom of the left side by sin2 (θ) to remove the sine from
the denominator in which it occurs.

sin2 (θ) cos2 (θ) − sin2 (θ)

2 2
= cos2 (θ) − 1
sin (θ) − cos (θ)
Observe that cos (θ)−sin2 (θ) = − sin2 (θ) − cos2 (θ) . Applying this to the
2

left gives a common factor of sin2 (θ) − cos2 (θ) on top and bottom. Cancel
them.
(
((((
− sin2 (θ) (sin ((−(cos2 (θ)
2


(((θ)
2 ((((2(
( = − sin2 (θ) = cos2 (θ) − 1
sin((
( (θ) − cos (θ)
Observe that cos2 (θ) − 1 = − (1 − cos2 (θ)). Apply the Pythagorean identity
and you’re done.
Alternatively, if you remember your identities among tangent, cotangent, and
secant – in particular that tan (θ) cot (θ) = 1 and that 1 + tan2 (θ) = sec2 (θ)
– you can multiply top and bottom of the left by tan2 (θ).

tan2 (θ) (1 − tan2 (θ))

= 1 − sec2 (θ)
tan2 (θ) − 1
You get a cancellation on the left of the nasty stuff, leaving
− tan2 (θ) = 1 − sec2 (θ). Add tan2 (θ) + sec2 (θ) to both sides, apply the
identity, and you’re done.

9) Prove that  
θ
csc (θ) − tan = cot (θ)
2

21
 θ


sin
 
θ 2
cos (θ)
Replace tan with θ
. Replace cot (θ) with . Replace csc (θ)
2 cos 2
sin (θ)
1
with .
sin (θ)
θ


1 1 sin 2
cos (θ)
− θ
=
sin (θ) 2 cos 2
sin (θ)
When there is a mix of half-angles and full-angles, the best next step is to
θ
make the substitution, 2u = θ or equivalently u = .
2
1 sin (u) cos (2u)
− =
sin (2u) cos (u) sin (2u)
Now apply the double angle formulas for sine and cosine. Again the best
version of the cosine formula is cos2 (2u) = 2 cos (u) − 1.

1 sin (u) 2 cos2 (u) − 1

− =
2 sin (u) cos (u) cos (u) 2 sin (u) cos (u)
1
Add to both sides.
2 sin (u) cos (u)
1 sin (u) 2
cos
2 (u)
− = 
sin (u) cos (u) cos (u) 2
sin (u)  (u)
cos
Observe the cancellations you get subsequent to that last step. Now add
sin (u)
to both sides.
cos (u)
1 cos (u) sin (u)
= +
sin (u) cos (u) sin (u) cos (u)
Put the two terms on the right over the common denominator of sin (u) cos (u).

1 cos2 (u) + sin2 (u)

=
sin (u) cos (u) sin (u) cos (u)
Finally apply the Pythagorean identity to the numerator of the right side,
and you’re done.

10) Prove that

4 sin (θ) cos (θ) cos (2θ) = sin (4θ)

22
Apply the double angle formula for cosine to the left. This time the best
version is cos (2θ) = 1 − 2 sin2 (θ).

4 sin (θ) cos (θ) 1 − 2 sin2 (θ) = sin (4θ)

Apply the quad-angle formula for sine to the right.

4 sin (θ) cos (θ) 1 − 2 sin2 (θ) = 4 sin (θ) cos (θ) − 8 sin3 (θ) cos (θ)

Factor sin (θ) cos (θ) out of the two terms of the right-hand side.

(( ( (( (
(( (θ) 1 − 2 sin2 (θ) = 4 − 8 sin2 (θ) ( ((



sin((θ)
4( cos sin((θ) cos (θ)

Observe the cancellation of the common factor from both sides. All that’s
left is to factor a 4 from the right-hand side, and you’re done.

11) Prove that

2 csc (2θ) = tan (θ) + cot (θ)

23
 1 sin (θ)
Replace csc (2θ) with . Replace tan (θ) with . Replace cot (θ)
sin (2θ) cos (θ)
cos (θ)
with .
sin (θ)
2 sin (θ) cos (θ)
= +
sin (2θ) cos (θ) sin (θ)
Apply the double angle formula for sine to the left side.

2
sin (θ) cos (θ)

= +
2
sin (θ) cos (θ) cos (θ) sin (θ)
Put the two terms of the right side over the common denominator, sin (θ) cos (θ).

1 sin2 (θ) + cos2 (θ)

=
sin (θ) cos (θ) sin (θ) cos (θ)
Now apply the Pythagorean identity, and you’re done.

12) Simplify

2 csc2 (θ) − csc4 (θ) + cot4 (θ)

24
 1 cos (θ)
Replace csc (θ) with . Replace cot (θ) with .
sin (θ) sin (θ)
2 1 cos4 (θ)
− +
sin2 (θ) sin4 (θ) sin4 (θ)
Observe that the common denominator here is sin4 (θ).

2 sin2 (θ) − 1 + cos4 (θ)

sin4 (θ)
Replace sin2 (θ) with 1 − cos2 (θ).

1 − 2 cos2 (θ) + cos4 (θ)

sin4 (θ)
Observe that the numerator is a perfect square. Factor it.
2
(1 − cos2 (θ))
sin4 (θ)
1
Replace 1 − cos2 (θ) with sin2 (θ), then take the cancellation. Replace
sin (θ)
with csc (θ) to get your final answer.

13) Find the solutions to

sin (2θ) + sin (4θ) = 0

25
Clearly θ = 0 is a solution, as is any value of θ for which sin (2θ) = 0. This
happens when

2θ = kπ where k is any integer

Or in other words, whenever θ is an integer multiple or half-integer multiple
of π. But there are more solutions. Use the double angle formula on sin (4θ):

sin (2θ) + 2 sin (2θ) cos (2θ) = 0

Extract the common factor.

sin (2θ) (1 + 2 cos (2θ)) = 0

Since we have already accounted for the solutions arising when sin (2θ) = 0,
divide that out and you are left with

1 + 2 cos (2θ) = 0
This is true for all values of θ for which cos (2θ) = − 12 . So which are those?
You need to look at the graph of the cosine function and recall your facts
about cosine of multiples of π3 . When
2π 4π 8π 10π
2θ = , , , , ...
3 3 3 3
you find that it solves the last equation. See the pattern? Whenever the
integer multiplier in the numerator is even but not a multiple of 6. If you
divide the whole thing by 2 to solve for θ, you have
kπ
θ= for integer k that is not a multiple of 3
3
Combine this solution set with the first solution set we found and you have
the whole solutions set.
14) Prove that

sec (θ) sec (θ)

− = 2 csc2 (θ)
1 + sec (θ) 1 − sec (θ)

26
 1 1
Replace sec (θ) with . Replace csc (θ) with .
cos (θ) sin (θ)
1 1
cos(θ) cos(θ) 2
1 − 1 = 2
1 + cos(θ) 1 − cos(θ) sin (θ)
On the left, multiply top and bottom of both fractions by cos (θ).
1 1 2
− = 2
cos (θ) + 1 cos (θ) − 1 sin (θ)
Put both of the fractions on the left over a common denominator. That
common denominator is (cos (θ) + 1) (cos (θ) − 1) = cos2 (θ) − 1.

cos (θ) − 1 − (cos (θ) + 1) 2

= 2
cos2 (θ) − 1 sin (θ)
Take the sum on the left. Also apply the Pythagorean identity to the de-
nominator on the left, and you’re done.

15) A bird flies horizontally out of a tree 5.6 meters directly above the
hunter’s gun and flies at a constant velocity of 28 meters/second. If the
velocity of the bullet is 35 meters/second, and the gun is fired at the moment
the bird leaves the tree:
a) What is the angle (with respect to the vertical) at which the gun must be
pointed to hit the bird? and
b) What is the flight time of the bullet?

27
Let the angle be θ. Let the horizontal distance the bullet travels be x and
the vertical distance it travels (which is given as 5.6 meters) be y. Then,
according to Pythagoras, the total distance, s, that the bullet travels is given
by

s2 = x2 + y 2
And from what we know about right triangles,

x = s sin (θ) and y = s cos (θ)

If t is the time of flight, vA the velocity of the bird,

and vB the velocity of the bullet, then

x = vA t and s = vB t

Both vA and vB are given in the problem.

Substituting, x = vB t sin (θ) and consequently
vA = vB sin (θ). From this we know the sine of the
angle. Since
sin (θ)
tan (θ) = q
1 − sin2 (θ)
we have, by doing a little algebra,
vA
tan (θ) = q
vB2 − vA2

Putting the numbers in you get θ = arctan 34 or about 53.13◦ . To determine

the flight time, we take advantage of y being given. Again substituting from
prior equations, y = s cos (θ) = vB t cos (θ). Since y, vB , and θ are all now
known,
y 5.6 meters 4
t= = = seconds = 0.2667 seconds
vB cos (θ) 21 meters/sec 15

16) Prove that

1 − csc (θ) sin (3θ) = 2 cos (2θ)

28
 1
Replace csc (θ) with :
sin (θ)
sin (3θ)
1− = 2 cos (2θ)
sin (θ)
Apply the triple-angle formula for sine from page 2 of this booklet to the
numerator:

3 sin (θ) − 4 sin3 (θ)

1− = 2 cos (2θ)
sin (θ)
Cancel the common factor of sin (θ) from top and bottom of the fraction:

1 − 3 − 4 sin2 (θ) = 2 cos (2θ)

Simplify the left side by combining the constant terms and factoring out the
common factor of 2:

2 −1 + 2 sin2 (θ) = 2 cos (2θ)

Apply the appropriate double-angle formula for cosine from page 2 to the
right to complete the proof.

29
Challenge problem: Prove that

cos (3θ) − cos (5θ)

= tan (θ)
sin (3θ) + sin (5θ)
Clearly you must apply the triple angle and quint-angle formulas for both
sine and cosine. In a previous challenge problem you were asked to derive the
quint-angle formula for cosine. You are encouraged to do the same for sine.
But you can also look both quint-angle formulas up at
[Link] There are sev-
eral forms of each quint-angle formula. I have applied the most useful for
this problem below. But first a rearrangement of the triple angle for sine.
Earlier I listed it as sin (3θ) = 3 sin (θ) − 4 sin3 (θ). But

3 sin (θ) − 4 sin3 (θ) = sin (θ) 3 − 4 sin2 (θ)

= sin (θ) (3 − 4 (1 − cos2 (θ)))

= sin (θ) (4 cos2 (θ) − 1)
Substituting all the multiple angle formulas into the original equation:

(4 cos3 (θ) − 3 cos (θ)) − (16 cos5 (θ) − 20 cos3 (θ) + 5 cos (θ))
= tan (θ)
sin (θ) (4 cos2 (θ) − 1
+ 16 cos4 (θ) − 12 cos2 (θ) + 1
)
Now gather terms.

−16 cos5 (θ) + 24 cos3 (θ) − 8 cos (θ)

= tan (θ)
sin (θ) (16 cos4 (θ) − 8 cos2 (θ))
Ok. Now let’s do some factoring. There are plenty of common factors to pull
out. Factor −8 cos (θ) from the top and 8 cos2 (θ) from the bottom.

−8
cos (θ) (2 cos4 (θ) − 3 cos2 (θ) + 1)

= tan (θ)
8
sin (θ) cos2 (θ) (2 cos2 (θ) − 1)
Observe that cos4 (θ)−3 cos2 (θ)+1 factors into (2 cos2 (θ) − 1) (cos2 (θ) − 1).
That results in a common factor of 2 cos2 (θ) − 1 on the top and bottom. So
cancel that out and you get

− cos (θ) (cos2 (θ) − 1)

= tan (θ)
sin (θ) cos2 (θ)
But − (cos2 (θ) − 1) = 1 − cos2 (θ) = sin2 (θ). So you have

cos  sin
2 (θ)


(θ)

 cos
2 (θ)

= tan (θ)
sin(θ)

Surely you can do the rest.

30
 The Cotangent Rule:

The diagram shows a triangle with

a line segment from its apex to
its base splitting it into two.
The contangent rule states that
(m + n) cot (c) = m cot (a)−n cot (b)
The trick, of course, is to prove this.
At first it’s hard to see the connec-
tion. But by transforming the prob-
lem it becomes solvable. We have
formulas that apply trig functions to
right triangles, so we transform this
into a right triangle problem. We do
this by constructing a line perpen-
dicular to the base of the triangle
that passes through its apex. This
gives us new lengths and angles,
which we designate with “primed”
symbols. Each of the primed sym-
bols is related by a simple equation
to its unprimed counterpart. We solve the problem in the primed context,
then use those simple equation to related it back to the original unprimed
problem.
The new diagram shows the trans-
formed problem. The new symbol,
p, is the distance along the base
from where the original line seg-
ment intersects the base to where
the perpendicular, h, intersects it.
m0 and n0 are the lengths along the
base to the left and right of the
perpendicular. c0 is the angle be-
tween the splitting line segment in
the first diagram and the perpen-
dicular. Between the original prob-
lem and the transformed problem
we have the following relationships
along the base: m = m0 − p, and
n = n0 + p. The angles constituting

31
the apex have these relationships:

a = a0 − c0 and b = b0 + c0

Look carefully at the diagram and observe the various right triangles that
the new perpendicular line segment forms. Recall the rule of right triangles
that the tangent of either of its non-right angles is equal to the length of its
opposite side divided by that of the adjacent side. Hence

0 m0
tan (a ) =
h

0 n0
tan (b ) =
h
p
tan (c0 ) =
h
Recall also from page 2 of this booklet the formula for the tangent of a sum.
Applying the relationships we’ve already established,
m −p 0
tan (a0 ) − tan (c0 ) h
tan (a) = =
1 + mh2p
0
1 + tan (a0 ) tan (c0 )
n +p0
tan (b0 ) + tan (c0 ) h
tan (b) = =
1 − mh2p
0 0 0
1 − tan (b ) tan (c )
On the right side of each of the above equations, multiply top and bottom
by h2 and replace m0 − p and n0 + p with m and n respectively:
hm
tan (a) =
h2 + m0 p
hn
tan (b) =
h2 − n0 p
Now recall that the cotangent of an angle is the reciprocal of its tangent. So
flipping both equations and multiplying through by m and n respectively:

h2 + m0 p
m cot (a) =
h
h2 − n0 p
n cot (b) =
h

32
Taking the difference between these two equations:

p (m0 + n0 )
m cot (a) − n cot (b) =
h
p
But as we’ve already seen, h = tan (c ), and because c and c0 are com-
0

plementary angles (that is they add up to a right angle) we know that,

tan (c0 ) = cot (c). Observe also that since m + n and m0 + n0 both add up to
the base of the original triangle, it follows that m0 + n0 = m + n. Making all
of those substitutions completes the proof.

33
 Trig functions and complex numbers

If you have never studied complex numbers, you probably believe that −1 has
no square root. And among the real numbers it indeed has no square root.
But in the 16th century Italian mathematicians, Niccolo Fontana Tartaglia
and Gerolamo Cardano, asked the question, what happens if we imagine the
existence of a square root of −1. They called this imaginary value, i, such
that i2 = (−i)2 = −1. Of course you have to be able to add and multiply
this imaginary value with any of the real numbers. This gives rise to a whole
self-consistent number system known as the complex numbers. In general a
complex number is in the form, a + ib, where a and b are both real numbers.
Clearly the real numbers is a subset of the complex numbers where b is held
to zero. In the example, a is called the real part of a + ib, and ib is called
its imaginary part. To add two complex numbers you simply add the real
parts to get the real part of the sum and add the imaginary parts to get the
imaginary part of the sum. To multiply two complex numbers we use the
distributive law.

(a + ib) (c + id) = ac+iad+ibc+i2bd = ac+iad+ibc−bd = ac−bd+i (ad + bc)

You can diagram the complex num-

bers in the same way you diagram
the x−y plane, where the horizontal
axis is the real axis and the vertical
axis is the imaginary axis. The fig-
ure shows a + ib (where, in this case,
both a and b are positive) plotted
on the complex plane. The distance
from 0 (the origin) to a + ib is shown
as r. The angle that a + ib makes
with the positive real axis is shown as θ. Note that if a + ib is not in quad-
rant I or quadrant IV, you have to add the quadrant adjustment of π or −π
for quadrants II or III respectively to the arctangent value.
From the parameters shown in the diagram, we see that we can represent
a + ib as r cos (θ) + ir sin (θ). Likewise if another complex number, c + id,
is s √
distance from the origin and makes angle φ with the real axis (that is
s = c2 + d2 and φ = arctan dc with the proper quadrant adjustment), then
we can represent this complex number as s cos (φ) + is sin (φ). Observe what

34
happens when we multiply the two complex numbers together.

(a + ib) (c + id) = ac − bd + i (ad + bc)

= rs cos (θ) cos (φ) − rs sin (θ) sin (φ)
+ i (rs cos (θ) sin (φ) + rs sin (θ) cos (φ))
= rs cos (θ + φ) + irs sin (θ + φ)

We arrive at the last line of the above by applying the sum formulas for
sine and cosine. What this means is that when you multiply two complex
numbers, the distance the product is from the origin will be the product of
the distances from the origin of the two factors, and the angle the product
makes with the real axis will be the sum of the angles to the real axis of the
two factors. From these relationships it is easy to see that

(r cos (θ) + ir sin (θ))n = r n cos (nθ) + ir n sin (nθ)

which is known as De Moivre’s formula.
Notice that if you define f (x) = cos (x)+i sin (x), then f (x + y) = f (x) f (y).
Recall that this is a property of exponential functions.
Observe that i is distance of 1
from the origin and makes angle,
π
2
radians, with the real axis. If
you multiply a + ib by i you get
−b + ia. You can see in the fig-
ure that this has the effect ro-
tating a + ib π2 radians (that is
90◦ ) counterclockwise as would
be expected from the multipli-
cation formula for complex num-
bers we have just seen.
Did you know that sine and cosine are actually exponential func-
tions? What are the critical characteristics of the function, y = ekx , for
some arbitrary constant, k? They are these: a) This function passes through
the point, (0, 1) regardless of the value of k. b) The slope of the tangent line
at (0, 1) is equal to k.

35
The figure shows this to be
so for the examples of k = 1
and k = 2. c) At any x, the
slope of the tangent line is
equal to kekx . If you have al-
ready studied some calculus
you will know why the latter
fact is always true. So what
do these three facts imply
about k = i? In the figure
when k = 1, if you want to
approximate e∆x , where ∆x
is very small, you can do so
by adding ∆x to e0 . Hence
e∆x ≈ 1 + ∆x. Likewise
if you want to approximate
e2∆x , you could do so with
e2∆x ≈ 1 + 2∆x. And in gen-
eral, if you want to approx-
imate ek∆x , you could do so
with ek∆x ≈ 1 + k∆x. Note
that these approximations get better and better for smaller and smaller ∆x.
By this pattern we see that if k = i, then we approximate ei∆x with 1 + i∆x.
That is from (0, 1) if you increase x by just a little, eix advances by that
same little bit but in the i direction. Note that the distance to the origin of
1 + i∆x is within ∆x2 of 1.
By the third fact, if we
2x
know the value of e for some
value of x, then we approximate
e2(x+∆x) ≈ e2x + 2∆xe2x , and in gen-
eral, ek(x+∆x) ≈ ekx + k∆xekx .
Now suppose at some value for x we
know where eix is. We would ap-
proximate ei(x+∆x) ≈ eix + i∆xeix .
Both approximations – for ei∆x and
for ei(x+∆x) – are shown in the di-
agram. For the latter, recall the
observation several paragraphs back

36
that multiplying by i rotates a point on the complex plane π2 counterclock-
wise. You can see that the short line segment of length, ∆x, that starts at
eix is at right angles to the line segment connecting eix to the origin.
The point of all this is that as x increases, the value of eix always moves at
right angles to the line segment connecting eix to the origin. That means
that eix follows a circular path, centered at the origin, around the complex
plane. We know that the distance to the origin of eix at x = 0 is 1, so we
know that the radius of the circle is also 1. In addition, at x the function,
eix , has advanced an arc length around the circle from (0, 1) of exactly x.
This leads to the following identity:

eix = cos (x) + i sin (x)

This equation is known as Euler’s formula. In electrical and mechanical
engineering, as well as in physics, it arises all the time. By applying symmetry
identities to Euler’s formula you also have

e−ix = cos (x) − i sin (x)

By taking the sum and difference these two and dividing each by 2

eix + e−ix eix − e−ix

cos (x) = sin (x) =
2 2i
This is how sine and cosine are, in reality, exponential functions. So what is
an immediate way we can put these formulas to work? Suppose you wanted
to know what cosn (θ) was in terms of multiples of θ.
 iθ n
n e + e−iθ
cos (θ) =
2
We can apply the binomial formula to the right-hand side:

n  
n −n
X n
cos (θ) = 2 eikθ e−i(n−k)θ
k=0
k
n  
X n
= 2−n ei(2k−n)θ
k=0
k

Recall that
 
n n!
=
k k! (n − k)!

37
So applying this to n = 6 we have

1
cos6 (θ) = e−6iθ + 6e−4iθ + 15e−2iθ + 20e0 + 15e2iθ + 6e4iθ + e6iθ


64
Rearranging terms
 6iθ
e + e−6iθ e4iθ + e−4iθ e2iθ + e−2iθ

6 1
cos (θ) = +6 + 15 + 10
32 2 2 2
1
= (cos (6θ) + 6 cos (4θ) + 15 cos (2θ) + 10)
32
See if you can use the same method to arrive at
1
cos4 (θ) =
(cos (4θ) + 4 cos (2θ) + 3)
8
Standing the expansion for cos6 on its head along with formulas similarly
derived for cos4 (θ) and cos2 (θ), we can readily arrive at a formula for the
cosine of the hex-angle. We can get the cos (6θ) term by multiplying our
expansion of cos6 by 32. We can eliminate the cos (4θ) term from that by
subtracting 48 times the expansion for cos4 .

32 cos6 (θ) = cos (6θ) + 6 cos (4θ) + 15 cos (2θ) + 10

−48 cos4 (θ) = −6 cos (4θ) − 24 cos (2θ) − 18
= cos (6θ) − 9 cos (2θ) − 8
Now we can eliminate the cos (2θ) term from the above result by adding 18
times the expansion for cos2 .

32 cos6 (θ) − 48 cos4 (θ) = cos (6θ) − 9 cos (2θ) − 8

+18 cos2 (θ) = 9 cos (2θ) + 9
cos (6θ) + 1
Now subtract 1 from this result to get

32 cos6 (θ) − 48 cos4 (θ) + 18 cos2 (θ) − 1 = cos (6θ)

Deriving a formula for sinn (θ) is a higher degree of difficulty than what we
just did.

n n  
eiθ − e−iθ

−n
X n ikθ
n
sin (θ) = = (2i) e (−1)n−k e−i(n−k)θ
2i k=0
k

38
or, combining terms on the right
n  
−n
X n i(2k−n)
n
sin (θ) = (2i) e (−1)n−k
k=0
k
You can see that we have the nasty business of raising i to the −n power. But
this, it turns out, is not hard at all. Powers of i follow a repeating pattern of
length, 4. i0 = 1; i1 = i; i2 = −1; i3 = −i; then, at n = 4 it repeats from the
beginning. When n is negative, its the same pattern going backward. The
real nastiness here is that when you resolve this back to trig functions, as we
did with the cosine power, you get a sum of cosines when n is even and a
sum of sines when n is odd. As an exercise, work it out for n = 3 and n = 4
to see why this happens.
Challenge problem: Use the exponential forms of sine and cosine to prove
that

 
1 i−x
y = arctan (x) = ln note that ln is natural log
2i i+x

Step 1: Observe from the above that x = tan (y). Derive a formula for
x = tan (y) in terms of exponentials

eiy − e−iy
sin (y) 2i
x = tan (y) = =
cos (y) eiy + e−iy
2

1 eiy − e−iy
= now multiply top and bottom by eiy
i eiy + e−iy

1 e2iy − 1 1
= observing that = −i
i e2iy + 1 i

e2iy − 1
x = tan (y) = −i
e2iy + 1
Step 2: We now need to solve for y in terms of x. To avoid algebra errors,
let u = eiy . We have

39
 u2 − 1
x = −i
u2 + 1
Step 3: Multiply through by u2 + 1 to get rid of the fraction.

x(u2 + 1) = u2 x + x = −iu2 + i
Step 4: Get all the terms with a factor of u2 to one side of the equation and
all the terms without to the other.

iu2 + u2 x = i − x
Step 5: Factor i + x out of the left side.

u2 (i + x) = i − x
Step 6: Divide through by i + x.
i−x
u2 =
i+x
Step 6: Back-substitute u = eiy .
i−x
e2iy =
i+x
Step 7: Take the natural log of both sides.
 
i−x
2iy = ln
i+x
Step 8: Divide by 2i. Since y = arctan (x), this gets you to the final result of

 
1 i−x
y = arctan (x) = ln which is what we set out to prove.
2i i+x

The tractor-pull problem: Suppose you have n tractors arranged at equal

increments around a circle (to form the vertices of a regular n-gon). Each
tractor is chained to a post at the center of the circle. The tractors all
pull directly outward from the circle, each with equal force. Intuitively we
expect that such equal forces arranged in this way will all cancel resulting
in a net force on the post of zero. Indeed when n is even, the symmetry
of the arrangement makes it clear that they do all cancel. But it’s not so
clear when n is odd. Using complex numbers, though, we can easily prove
that no matter what integer, n > 1, you choose, the forces always cancel

40
exactly. Here’s the proof. Place the first tractor at 1 + i0. Place the next
at z = cos (θ) + i sin (θ), where θ = 2π
n
. Place the remaining n − 2 tractors
at cos (kθ) + i sin (kθ), where k ranges from 2 to n − 1. By De Moivre’s
formula, the n tractors are at points, z k , where z = cos (θ) + i sin (θ) and k
ranges from zero to n − 1. We also know that if you go a full n steps of 2π n
around the circle, you end up where you started. If you start at 1 + i0, you
will end up there again after encountering n tractors. Hence we know that
z n = 1 + i0. All we need to do now is prove that the sum of z k , for k going
from 0 to n − 1 is always zero.
n−1
X
z 0 + z 1 + z 2 + · · · + z n−1 = z k = 0 to be proved
k=0

What happens if we multiply the polynomial above by z − 1?

z 1 + z 2 + · · · + z n−1 + z n
z 0 + z 1 + z 2 + · · · + z n−1
−z 0 + zn
That is, the sum of all those powers of z times z − 1 is equal to z n − 1. But
we already established that z n = 1. Hence z n − 1 = 0. It follows that one of
the two factors, either z − 1 or the sum of all the powers of z, must be zero.
But z = cos (θ) + i sin (θ), where θ = 2π
n
. If 0 < θ < 2π, it is impossible for
z to be equal to 1. Therefore z − 1 cannot be zero, and it follows that the
sum of the powers of z from zero to n − 1 must be zero.

41