Fibonacci and Pythagoras

The first few terms of the Fibonacci sequence are as follows: F 1 = 1 , F 2 = 1 , F 3 = 2 , F 4 = 3 , F 5 = 5 , F 6 = 8 , F 7 = 13 , F 8 = 21 , F 9 = 34 , \begin{aligned} F_1 = 1,&\, F_2 = 1,\, F_3 = 2,\, F_4 = 3,\,\\\\ {\color{#D61F06}F_5 = 5},&\, F_6 = 8,\, {\color{#D61F06}F_ 7 = 13},\, F_8 = 21, {\color{#D61F06}F_9 = 34},\, … \end{aligned} It just so happens that 5 5 , 13 13 , and 34 34 are also the hypotenuses of right triangles with all integer sides: ( 3 , 4 , 5 ) , ( 5 , 12 , 13 ) , ( 16 , 30 , 34 ) . (3, 4, {\color{#D61F06}5}),\ (5, 12, {\color{#D61F06}13}),\ (16, 30, {\color{#D61F06}34}). Is every Fibonacci number F 2 n 1 F_{2n - 1} for all n 3 n \geq 3 the hypotenuse of a right triangle with integer sides?

No Yes

This section requires Javascript.
You are seeing this because something didn't load right. We suggest you, (a) try refreshing the page, (b) enabling javascript if it is disabled on your browser and, finally, (c) loading the non-javascript version of this page . We're sorry about the hassle.

5 solutions

299792458 萬
Sep 16, 2018

It's easy to discover that F 2 n 1 = F n 2 + F n 1 2 F_{2n-1} = {F_n}^2 + {F_{n-1}}^2 , thus we have F 2 n 1 2 = ( F n 2 ) 2 + ( F n 1 2 ) 2 + 2 F n 2 F n 1 2 = ( F n 2 F n 1 2 ) 2 + ( 2 F n F n 1 ) 2 {F_{2n-1}}^2 = {({F_n}^2)}^2 + {({F_{n-1}}^2)}^2 + 2{{F_n}^2}{{F_{n-1}}^2} \\ = {({F_n}^2 - {F_{n-1}}^2)}^2 + {(2{F_n}{F_{n-1}})}^2

Since all terms here appear as squares of positive integers, the statement is proved.

(We can prove that F 2 n 1 = F n 2 + F n 1 2 F_{2n-1} = {F_n}^2 + {F_{n-1}}^2 in this way:

F 2 n 1 = F 2 n 3 + F 2 n 2 = F 1 F 2 n 3 + F 2 F 2 n 2 = F 1 F 2 n 3 + F 2 ( F 2 n 3 + F 2 n 4 ) = ( F 1 + F 2 ) F 2 n 3 + F 2 F 2 n 4 = F 2 F 2 n 4 + F 3 F 2 n 3 = F 3 F 2 n 5 + F 4 F 2 n 4 = = F n 1 2 + F n 2 . F_{2n-1} = F_{2n-3} + F_{2n-2} \\ = {F_1}{F_{2n-3}} + {F_2}{F_{2n-2}} \\ = {F_1}{F_{2n-3}} + {F_2}{({F_{2n-3}} + {F_{2n-4}})}\\ = {(F_1 + F_2)}{F_{2n-3}} + {F_2}{F_{2n-4}} \\ = {F_2}{F_{2n-4}} + {F_3}{F_{2n-3}}\\ = {F_3}{F_{2n-5}} + {F_4}{F_{2n-4}}\\ = …\\ = {F_{n-1}}^2 + {F_n}^2 .

Or simply take the closed-form formula (Binet's formula) and verify it.)

---------Update---------

Another way to prove that F 2 n 1 = F n 2 + F n 1 2 F_{2n-1} = {F_n}^2 + {F_{n-1}}^2 using induction. Less elegant than the method mentioned above, but maybe easier to think of and more natural.

Suppose that F 2 n 1 = F n 2 + F n 1 2 F_{2n-1} = {F_n}^2 + {F_{n-1}}^2 holds for all n k n ≤ k . Thus we have F 2 k 1 = F k 2 + F k 1 2 F_{2k-1} = {F_k}^2 + {F_{k-1}}^2 and F 2 k 3 = F k 1 2 + F k 2 2 F_{2k-3} = {F_{k-1}}^2 + {F_{k-2}}^2 . It is to show that the statement also holds for n = k + 1 n = k+1 , that is, F 2 k + 1 = F k + 1 2 + F k 2 F_{2k+1} = {F_{k+1}}^2 + {F_k}^2 .

The first step is to find the relation between F 2 k + 1 F_{2k+1} , F 2 k 1 F_{2k-1} , and F 2 k 3 F_{2k-3} .

F 2 k + 1 = F 2 k + F 2 k 1 = 2 F 2 k 1 + F 2 k 2 = 3 F 2 k 2 + 2 F 2 k 3 F_{2k+1} = F_{2k} + F_{2k-1}\\ = 2 F_{2k-1} + F_{2k-2}\\ = 3 F_{2k-2} + 2 F_{2k-3}

From the two equations above we can eliminate F 2 k 2 F_{2k-2} , and get

F 2 k + 1 = 3 F 2 k 1 F 2 k 3 = 3 ( F k 2 + F k 1 2 ) ( F k 1 2 + F k 2 2 ) = 3 F k 2 + 2 F k 1 2 F k 2 2 = 3 F k 2 + 2 F k 1 2 ( F k F k 1 ) 2 = 2 F k 2 + F k 1 2 + 2 F k F k 1 = ( F k + F k 1 ) 2 + F k 2 = F k + 1 2 + F k 2 . F_{2k+1} = 3 F_{2k-1} - F_{2k-3}\\ = 3({F_k}^2 + {F_{k-1}}^2) - ({F_{k-1}}^2 + {F_{k-2}}^2)\\ = 3{F_k}^2 + 2{F_{k-1}}^2 - {F_{k-2}}^2\\ = 3{F_k}^2 + 2{F_{k-1}}^2 - (F_k - F_{k-1})^2\\ = 2{F_k}^2 + {F_{k-1}}^2 + 2{F_k}{F_{k-1}}\\ = (F_k + F_{k-1})^2 + {F_k}^2\\ = {F_{k+1}}^2 + {F_k}^2 .

And so we have our statement proved.

This is more clever, using { a = s 2 t 2 b = 2 s t c = s 2 + t 2 a 2 + b 2 = c 2 \begin{cases}a=s^2-t^2\\b=2st\\c=s^2+t^2\\a^2+b^2=c^2\end{cases} To reduce the problem into partition F 2 n 1 F_{2n-1} into sum of squares.

Kelvin Hong - 2 years, 8 months ago

Log in to reply

Yep, that's my approach.

Arjen Vreugdenhil - 2 years, 8 months ago

I found it quite remarkable to discover that alternate numbers in the sequence were always the sum of two squares and even more remarkable that the sum of squares of consecutive numbers in the sequence produced all these numbers.

Proving it made it all rather unremarkable.

Malcolm Rich - 2 years, 8 months ago

Log in to reply

When learning Math I occasionally get astonished by some bizzare statements, but soon later their proofs just wash away such feelings and leave me an impression that every step just goes so naturally. Maybe it is part of the mathematical beauty, to finally change something unbelievable to a matter of course?

299792458 萬 - 2 years, 8 months ago

Log in to reply

Totally agree. Prior to today, if someone were to suggest there was a sequence of numbers Fn, which contained every combination of the sum-product of every pair of consecutive numbers - ie F(x)F(y) + F(x+1)F (y+1) - I would have expected that sequence to have quite few gaps.

It's just another interesting property of the Fibonacci series.

Malcolm Rich - 2 years, 8 months ago

Lovely indeed. Curiously, but I suppose expectedly, the ratio of the sides of the triangles as n increases approaches 1,2, sqrt(5)

Paul Stancioff - 2 years, 8 months ago

Log in to reply

Exactly! This follows directly from the fact that the ratio of two neighbouring terms in the Fibonacci Series has a limit of the golden ratio: lim (F n / F (n-1)) = φ = (1 + √5)/2, n->∞.

299792458 萬 - 2 years, 8 months ago
Mark Hennings
Sep 15, 2018

Applying Binet's Formula, we have F n = 1 5 [ ϕ n ψ n ] n 0 F_n \; = \; \tfrac{1}{\sqrt{5}}\big[\phi^n - \psi^n\big] \hspace{2cm} n \ge 0 where ϕ = 1 2 ( 1 + 5 ) ψ = 1 2 ( 1 5 ) \phi \; = \; \tfrac12\big(1 + \sqrt{5}\big) \hspace{2cm} \psi \; = \; \tfrac12\big(1 - \sqrt{5}\big) so that ϕ + ψ = 1 \phi + \psi = 1 , ϕ ψ = 1 \phi\psi = -1 , and ϕ 2 ϕ 1 = ψ 2 ψ 1 = 0 \phi^2 - \phi - 1 = \psi^2 - \psi-1 = 0 . Then 2 F n F n 1 = 2 5 ( ϕ 2 n + 1 + ψ 2 n + 1 ϕ n ψ n + 1 ϕ n + 1 ψ n ) = 2 5 [ ϕ 2 n + 1 + ψ 2 n + 1 + ( 1 ) n + 1 ] F n 1 F n + 2 = 1 5 ( ϕ 2 n + 1 + ψ 2 n + 1 ϕ n 1 ψ n + 2 ϕ n + 2 ψ n 1 ] = 1 5 [ ϕ 2 n + 1 + ψ 2 n + 1 + ( 1 ) n ( ϕ 3 + ψ 3 ) ] = 1 5 [ ϕ 2 n + 1 + ψ 2 n + 1 + 4 ( 1 ) n ] \begin{aligned} 2F_nF_{n-1} & = \; \tfrac25\big(\phi^{2n+1} + \psi^{2n+1} - \phi^n\psi^{n+1} - \phi^{n+1}\psi^n\big) \; = \; \tfrac25\big[\phi^{2n+1} + \psi^{2n+1} + (-1)^{n+1}\big] \\ F_{n-1}F_{n+2} & = \; \tfrac15\big(\phi^{2n+1} + \psi^{2n+1} - \phi^{n-1}\psi^{n+2} - \phi^{n+2}\psi^{n-1}\big] \; = \; \tfrac15\big[\phi^{2n+1} + \psi^{2n+1} + (-1)^n(\phi^3 + \psi^3)\big] \; = \; \tfrac15\big[\phi^{2n+1} + \psi^{2n+1} + 4(-1)^n\big] \end{aligned} so that ( 2 F n F n + 1 ) 2 + ( F n 1 F n + 2 ) 2 = 1 25 [ 4 ( ϕ 2 n + 1 + ψ 2 n + 1 ) 2 + 8 ( 1 ) n + 1 ( ϕ 2 n + 1 + ψ 2 n + 1 ) + 4 + ( ϕ 2 n + 1 + ψ 2 n + 1 ) 2 + 8 ( 1 ) n ( ϕ 2 n + 1 + ψ 2 n + 1 ) + 16 ] = 1 5 [ ( ϕ 2 n + 1 + ψ 2 n + 1 ) 2 + 4 ] = 1 5 [ ϕ 2 n + 1 ψ 2 n + 1 ] 2 \begin{aligned} \big(2F_nF_{n+1}\big)^2 + \big(F_{n-1}F_{n+2}\big)^2 & = \; \tfrac{1}{25}\left[ \begin{array}{l} 4(\phi^{2n+1}+\psi^{2n+1})^2 + 8(-1)^{n+1}(\phi^{2n+1} + \psi^{2n+1}) + 4 \\ + (\phi^{2n+1} + \psi^{2n+1})^2 + 8(-1)^n(\phi^{2n+1} + \psi^{2n+1}) + 16\end{array}\right] \\ & = \; \tfrac15\big[(\phi^{2n+1} + \psi^{2n+1}\big)^2 + 4\big] \; = \; \tfrac15\big[\phi^{2n+1} - \psi^{2n+1}\big]^2 \end{aligned} and hence ( 2 F n F n + 1 ) 2 + ( F n 1 F n + 2 ) 2 = F 2 n + 1 2 \big(2F_nF_{n+1}\big)^2 + \big(F_{n-1}F_{n+2}\big)^2 \; = \; F_{2n+1}^2 for all n 1 n \ge 1 , and the two terms 2 F n F n + 1 2F_nF_{n+1} and F n 1 F n + 2 F_{n-1}F_{n+2} are nonzero for n 2 n \ge 2 , and hence F 2 n + 1 F_{2n+1} is the hypotenuse of a Pythagorean triad for all n 2 n \ge 2 , making the answer Y e s \boxed{\mathrm{Yes}} .

Rather more simply: F(N)^2 + F(N+1)^2 = [{phi^n - (-1/phi)^n}^2 + {phi^(n+1) - (-1/phi)^(n+1)}^2]/5, which boils down to [phi^(2n+1).{phi + 1/phi} + [phi^-(2n+1}.{phi + 1/phi}}/5, which equals F(2N+1) since phi + 1/phi = sqrt(5)

A Former Brilliant Member - 2 years, 8 months ago
Arjen Vreugdenhil
Sep 17, 2018

Any sum of two squares p 2 + q 2 p^2 + q^2 is the hypotenuse in a Pythagorean triple: ( a , b , c ) = ( p 2 q 2 , 2 p q , p 2 + q 2 ) (a,b,c) = (p^2-q^2, 2pq, p^2+q^2) for some integers p , q p,q .

Since F 2 n 1 = F n 2 + F n 1 2 F_{2n-1} = F_n^2 + F_{n-1}^2 , we simple set p = F n p = F_n and q = F n + 1 q = F_{n+1} to obtain the desired conclusion.

I think the second statement requires some sort of proof.

Malcolm Rich - 2 years, 8 months ago
Srikanth Tupurani
Sep 20, 2018

We can derive it directly using the formula. Most elegant way is to use linear algebra.

Bob Ewell
Sep 20, 2018

I love the proofs. I solved the problem by observation. Unfortunately I didn’t notice that F(2n-1) = Fn^2 + F(n-1)^2. I saw a recursion that produced all the Pythagorean triples. Given Pythagorean triples (a(n-2), b(n-2), F(n-2)) and (a(n), b(n) and F(n)), the next triple is given by a(n+2) = a(n-2) + F(n), b(n+2) = b(n)+F(n+1) - a(n-2), and, of course, F(n+2).

0 pending reports

×

Problem Loading...

Note Loading...

Set Loading...