MATH 324 Summer 2012
Elementary Number Theory
Notes on Pythagorean Triples
Pythagorean Triples
In the following we present a brief introduction to Pythagorean triples, and we leave as exercises several
questions which can be done as practice for the final examination.
Given a right triangle with legs of length a and b and hypotenuse of length c, the Pythagorean theorem tells
us that
a2 + b 2 = c 2 (∗)
and if a, b and c are positive integers, then the triplet (a, b, c) is called a Pythagorean triple.
The Pythagorean relationship was known even before Pythagoras. One of the Babylonian clay tablets in the
G. A. Plympton Collection at Columbia University indicates that the Babylonians knew of the Pythagorean
theroem more than 3500 years ago.
Note that (3, 4, 5) is a Pythagorean triple, moreover, for any positive integer n, we have
(3n)2 + (4n)2 = 9n2 + 16n2 = 25n2 = (5n)2 ,
so that (3n, 4n, 5n) is also a Pythagorean triple. Thus, there are infinitely many Pythagorean triples.
If (a, b, c) is a Pythagorean triple, then we say that it is a primitive Pythagorean triple if and only if
a, b, and c have no common factors, that is,
gcd(a, b) = gcd(a, c) = gcd(b, c) = 1.
There are many different ways to construct Pythagorean triples.
• We can use the Fibonacci numbers
F0 F1 F2 F3 F4 F5 F6 F7 F8 ··· Fn
0 1 1 2 3 5 8 13 21 ··· Fn−2 + Fn−1
where each term in the sequence is the sum of the previous two terms, to construct Pythagorean triples
as follows. Let n be a positive integer and define
a = Fn Fn+3
b = 2Fn+1 Fn+2
2 2
c = Fn+1 + Fn+2 ,
where Fn is the nth Fibonacci number, then it is easy to show that a2 + b2 = c2 , so that (a, b, c) is a
Pythagorean triple. As an exercise, determine if it is a primitive Pythagorean triple.
Also, show that we cannot generate all Pythagorean triples in this way.
• Another method, discovered by the Pythagoreans, is as follows. If n is any odd positive integer, and
we define
a=n
n2 − 1
b=
2
n2 + 1
c=
2
then (a, b, c) is a Pythagorean triple. Verify this as an exercise, and show also that we cannot generate
all Pythagorean triples in this way.
• Yet another method, discovered by the Greek philosopher Plato, is as follows. If n is any positive
integer, and we define
a = 2n
b = n2 − 1
c = n2 + 1
then (a, b, c) is a Pythagorean triple. Again, verify that this is so, and that we cannot generate all
Pythagorean triples this way.
• In 1934, M. Willey and E. C. Kennedy gave the following method for generating infinitely many
Pythagorean triples.
a b c
21 220 221
201 20200 20201
2001 2002000 2002001
20001 200020000 200020001
200001 20000200000 20000200001
2000001 2000002000000 2000002000001
.. .. ..
. . .
You are invited to find a formula for these a, b and c, and show that it does indeed generate infinitely
many Pythagorean triplets, but again, not all of them.
• In his book The Elements, Euclid gave the following method for generating Pythagorean triplets. For
any positive integers m and n which are relatively prime, that is, gcd(m, n) = 1, and of different parity,
that is, one is even and the other is odd, if we define for m > n,
a = 2mn
b = m 2 − n2
c = m 2 + n2
then (a, b, c) is a primitive Pythagorean triple. Moreover, every primitive Pythagorean triple arises in
this way. In order to prove this, we need the following lemmas.
Lemma 1. If (a, b, c) is a primitive Pythagorean triple, then a and b are of different parity, that is,
one is even and the other is odd.
Proof. If a and b are both even, then gcd(a, b) ≥ 2, which is a contradiction, since (a, b, c) is a primitive
Pythagorean triple. If a and b are both odd, then
a2 ≡ 1 (mod 4) and b2 ≡ 1 (mod 4),
so that
c2 = a2 + b2 ≡ 2 (mod 4),
which is impossible, since the square of any integer is congruent to 0 or 1 modulo 4. Therefore, a and
b have opposite parity.
Lemma 2. If u and v are relatively prime positive integers such that u · v is a perfect square, then
both u and v are perfect squares.
Proof. Let p be a prime that divides u and let α be the exact power of p in u. (In symbols, we denote
this by pα ||u.) Since u and v are relatively prime, then p does not divide v, and therefore pα ||u · v.
However, u · v is a perfect square, so that α must be even. Since this holds for all primes p dividing u,
then u is a perfect square. Similarly, v is a perfect square.
Theorem. If (a, b, c) is a primitive Pythagorean triple, with b even, then there exist positive integers
m and n which are relatively prime and of opposite parity with m > n such that
a = m 2 − n2
b = 2mn
c = m 2 + n2
Proof. If (a, b, c) is a primitive Pythagorean triple with b even, then a is odd, and since c 2 = a2 + b2 ,
then c is also odd. Therefore, c − a and c + a are both even, so that
2
c+a c−a b
· = .
2 2 2
Any common divisor of the two factors on the left divides both their sum c and their difference a, and
since gcd(a, c) = 1, then the two factors on the left have no common divisors except 1, that is, they
are relatively prime. From the lemma, they must both be squares, so there exist positive integers m
and n such that
c+a
= m2
2
c−a
= n2
2
b
= mn.
2
We note that m and n are relatively prime since c + a and c − a are relatively prime, and that m > n.
Also, since c2 = m2 + n2 and c is odd, then m and n have different parity.