MATHEMATICAL INDUCTION

MATH 25A SECTION HANDOUT 2

BY GERARDO CON DIAZ

Imagine a bunch of dominoes on a table. They are set up in a straight line, and
you are about to push the first piece to set off the chain reaction that will bring all
the dominoes down. For this chain reaction to knock off every piece, you have to
make sure that every piece is close enough to the next one. Otherwise, the reaction
will stop. First, you have to be absolutely sure to have put all the domino pieces in
such a way that each of them, while falling, will knock the next one down. Next,
you just need to knock down the first piece. You can then be sure that, eventually,
every piece is definitely going to fall. Lets turn this into math.
The domino scenario we described above is a surprisingly accurate representa-
tion of a proof technique called mathematical induction. In this case, the proof
by induction consists in making sure that the first domino piece falls and that if
a domino piece falls, then it will knock the next one off. Then, by mathematical
induction, we will be sure that every single domino piece will fall. More formally,
this technique is summarized below.

Mathematical Induction: Let A(n) be an assertion involving an integer n. If

we can perform the following two steps:
(1) prove that A(1) is true,
(2) for a given arbitrary k, assume that A(k) is true, and prove that A(k + 1)
is also true,
then we can conclude that the assertion A(n) is true for every positive integer n.

In our domino analogy, A(n) is the statement that the nth piece of domino
falls. In addition, the first domino getting knocked off refers to the statement A(1)
is true and the condition that if a domino falls, then the next one falls as well is
analogous to if A(x) is true, then A(x + 1) is true as well. Finally, the idea of
all the dominoes falling is equivalent to the idea of A(n) being true for all positive
integers n.
Now we will show an example of a proof by induction. Please note how the different
steps are clearly marked.
Proposition 1. For every positive integer n, the following identity holds.
n(n + 1)
1 + 2 + ... + n = .
2
Proof. A proof by induction consists of three different steps. We will label these
steps along the proof for instructive purposes and will show an example without
the annotations in the next pages.

Base Step (Show that the first domino piece falls) Note that if n = 1, then
1
2 MATH 25A SECTION HANDOUT 2 BY GERARDO CON DIAZ

1 = 1(1+1)
2 equals the sum of the numbers from 1 to n = 1.
In this step, we make sure that whatever we want to prove works in the simplest
case of all. Sometimes the base step is one of the hardest ones in the proof (but
clearly not in this case!), but it is crucial to always check it.

Inductive Hypothesis (Assume that a domino piece falls) Assume that for some
n = k, we have
k(k + 1)
1 + 2 + ... + k = .
2
We will assume that the last equality is true. In general, once you assume the
inductive hypothesis, you can assume that it holds whenever you need it. We show
the explicit use of this hypothesis in the next step.

Inductive Step (Show that if a domino piece falls, then the next piece falls) We
assumed that for some k, the following equality holds.
k(k + 1)
1 + 2 + ... + k = .
2
We want to show that the domino piece following k falls. This piece is k + 1. To
show that it gets knocked off, we need to show that our equality holds if n = k + 1.
In other words, we need to show that
(k + 1)(k + 2)
1 + 2 + . . . + k + (k + 1) = .
2
Now look at our inductive hypothesis and at the equality we want to arrive to.
Examine them and ask yourself: What can we do to turn the inductive hypothesis
into what we want? In this case, if we look at the left sides of the equation, we see
that there is an extra term in the sum we want to prove. The term is k + 1, and
the inductive hypothesis doesnt have it!
We need this k + 1, so lets toss it into the inductive hypothesis. We do this in
the line below. Since
k(k + 1)
1 + 2 + ... + k = ,
2
we can add k + 1 to both sides of the equation and get
k(k + 1)
1 + 2 + . . . + k + (k + 1) = + (k + 1).
2
(k+1)(k+2)
Keeping in mind that we want the right side to read 2 , we note that the
last equation implies that
k(k + 1) + 2(k + 1)
1 + 2 + . . . + k + (k + 1) = .
2
But we can factor k + 1 on the right side to get
(k + 2)(k + 1)
1 + 2 + . . . + k + (k + 1) = ,
2
which is exactly what we wanted to arrive to.

We showed that n = 1 satisfies the property and that if some k satisfies it,
then k + 1 satisfies it as well. This shows, by the induction principle, that our
equation holds for every positive integer n, so we are done. 
 MATHEMATICAL INDUCTION 3

Now we will show another sum we can prove by induction. We will not annotate
it as often as we did before, but it is important to note that the basic elements of
an induction are still there, and have been clearly labeled.

Proposition 2. Let n be a positive integer. Show that the quantity

1 + 3 + 5 + 7 + . . . + (2n 1)
is a perfect square.
Proof. The sum equals 1 when n = 1, so our base case is satisfied. However, to
perform an induction, assuming that the sum equals any perfect square r2 is not
enough, since we would not be able to manipulate the expression
1 + 3 + 5 + 7 + . . . + (2n 1) = k 2
into
1 + 3 + 5 + 7 + . . . + (2n 1) + (2n + 1) = r2
by simply adding 2n + 1 to k 2 .
This means that we need to find something about which perfect square our sum
equals. The only way to do this is to plug in some values of n and find a pattern.
If n = 2, our sum is
1 + 3 = 4 = 22 .
If we keep trying, we will see that the values of the sum for n equal to 3, 4 and 5
are, respectively 32 , 42 and 52 . Therefore, it is safe to guess that
1 + 3 + 5 + 7 + . . . + (2n 1) = n2 .
Lets prove this by induction.
We already took care of the basis step. Lets assume that for some n, it is true that
1 + 3 + 5 + 7 + . . . + (2n 1) = n2 .
This last statement is our inductive hypothesis. We must now show that
1 + 3 + 5 + 7 + . . . + (2n 1) + (2n + 1) = (n + 1)2 .
To do this, we add (2n + 1) to both sides of our inductive hypothesis to get
1 + 3 + 5 + 7 + . . . + (2n 1) + (2n + 1) = n2 + (2n + 1).
But we can arrange the right side of the last equation to get
1 + 3 + 5 + 7 + . . . + (2n 1) + (2n + 1) = n2 + (2n + 1) = (n + 1)2 .
This is what we wanted to show, so our proof is complete. 
Consider this other exercise.
Proposition 3. For any integer n,
13 + 23 + . . . + n3 = (1 + 2 + 3 + . . . + n)2
.
We will not show this exercise, but you should try it at home. [Hint: Does the
right side look familiar?]
Induction is useful for more than just proving sums. It is a very versatile tool,
since we can construct millions of different statements depending on the natural
numbers. Consider the following proposition.
4 MATH 25A SECTION HANDOUT 2 BY GERARDO CON DIAZ

Proposition 4. Prove that 8n 1 is divisible by 7 for any positive integer n.

Proof. We can do this in two ways: One uses the properties of Z/nZ studied in
class and one uses mathematical induction.

Proof by induction When n = 1, the statement asks us to show that 7 is di-

visible by 7. This statement is clearly true. We proceed by induction. Assume that
for some n, 7 divides 8n 1. Then there exists an integer m such that 8n 1 = 7m.
We need to show that there exists some integer r such that 8n+1 1 = 7r.
We multiply the equation in the inductive hypothesis by 8 to turn the 8n in
it into 8n+1 in the equation we want to get. In other words, since 8n 1 = 7m,
we multiply by 8 to get 8n+1 8 = 56m. Now we need to rearrange the expres-
sion until we reach the expression we are looking for. In this case, we note that
56m = 8n+1 8 = 8n+1 7 1. This implies that 56m + 7 = 8n+1 1, which
implies that 7r = 8n+1 1 for r = 8m + 1. This is what we wanted to show, so our
proof is complete.

Proof with equivalence classes Showing that 7 divides 8n 1 is equivalent to

showing that the equivalence class of 8n 1 in Z/7Z is the equivalence class of 0.
But now we note that, because of the properties of equivalence classes in this field,
[8n 1] = [8n ] [1] = [8]n [1].
We conclude this proof by noting that [8] = [1], so that
[8n 1] = [1]n [1] = [1n ] [1] = [1] [1] = [0].
This concludes our proof. 
Although equivalence classes gave us a much more concise argument in this
case, the problem might be complicated enough to prevent us from finding a neat
argument using equivalence classes. This would force us to settle with a less elegant
yet equally valid proof by induction. This is the case in the following problem.
Proposition 5. Prove that for any integer n, the number 11n+1 +122n1 is divisible
by 133.
We will not prove this proposition since its technique is similar to the one used
above. However, it is a very instructive exercise. [Hint: Multiply by 12 and re-
arrange].
We have seen that induction gives us a way to prove statements about the set
of natural numbers or some of its subsets. But this is not an infallible method.
Consider the following proposition.
Proposition 6. Show that for any odd prime number p, either p 1 or p 3 is
divisible by 4.
Proof. We cannot proceed by induction (How do we know what the rth prime is?).
However, the proof is quite simple. Any number t must be contained in some
element of Z/4Z, so it must be of the form 4r, 4r + 1, 4r + 2 or 4r + 3 (Remember
that forming equivalence classes in Z/nZ is the same thing as grouping numbers
by their remainder when divided by n). But p may not be of the form 4r or 4r + 2,
since both of these numbers are even. Hence p is of the form 4r + 1 or 4r + 3. In
either case, either p 1 or p 3 is divisible by 4. 
 MATHEMATICAL INDUCTION 5

We hope that this short document provides a smooth introduction into what
an induction is. Pease do not hesitate to email any of your CAs if you have any
questions about this document or about any aspect of the class. Were here to help!
You will find some more challenging exercises at http://www.maths.ox.ac.uk/
prospective-students/undergraduate/practice-problems/pdf/allsheets2.pdf.