See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/266528994

Two Topics in Number Theory A Formula for Primes and an Algebraic Identity

Article · January 2011

CITATION READS
1 2,217

1 author:

Rafael Jakimczuk
Universidad Nacional de Luján
231 PUBLICATIONS   324 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Integers View project

Diophantine equations View project

All content following this page was uploaded by Rafael Jakimczuk on 08 December 2017.

The user has requested enhancement of the downloaded file.

International Mathematical Forum, Vol. 6, 2011, no. 33, 1651 - 1654

Two Topics in Number Theory

A Formula for Primes and
an Algebraic Identity
Rafael Jakimczuk

División Matemática
Universidad Nacional de Luján
Buenos Aires, Argentina
Jakimczu@mail.unlu.edu.ar

In memory of my sister Fedra Marina Jakimczuk (1970-2010)

Abstract

In this article we present formulas for the n-th prime pn in terms of

the former primes p1 , p2 ,. . ., pn−1 .
We also obtain an algebraic identity for the sum of eight squares.

Mathematics Subject Classification: 11A41, 11A99

Keywords: Prime numbers. Formulas. Sums of eight squares

1 The Formulas for the n-th Prime Number

The following theorem is sometimes called the principle of cross-classification
(see [4], page 84). We now enunciate the principle.

Theorem 1.1 Let S be a set of N distinct elements, and let S1 , . . . , Sr

be arbitrary subsets of S containing N1 , . . . , Nr elements, respectively. For
1 ≤ i < j < . . . < l ≤ r, let Sij...l be the intersection of Si , Sj , . . . , Sl and let
Nij...l be the number of elements of Sij...l . Then the number of elements of S
not in any of S1 , . . . , Sr is
  
K=N− Ni + Nij − Nijk + . . . + (−1)r N12...r (1)
1≤i≤r 1≤i<j≤r 1≤i<j<k≤r
1652 R. Jakimczuk

Proof. See, for example, [4] (page 84) or [2] (page 233).

In 1971, J. M. Gandhi gave a formula for the n-th prime pn in terms of

p1 , p2 , . . . , pn−1 . Let Pn−1 = p1 p2 . . . pn−1 . The Gandhi’s formula is,
⎡ ⎛ ⎞⎤
1 1  µ(d)
pn = ⎣1 − log ⎝− + ⎠⎦
log 2 2 d/Pn−1 2d − 1
Where µ(d) is the Mobius function and [x] denotes the largest integer that
does not exceed x . The Mobius function is defined as follows: µ(1) = 1, if n
is the product of r distinct primes, then µ(n) = (−1)r , if the square of a prime
divides n, then µ(n) = 0.
A more simple proof of this formula was given by Vanden Eynden in 1972
and in 1974 Golomb gave another proof (see [7], Chapter 3).
There exist (former to Gandhi) others formulas. However, these formulas
are useless, because to calculate pn one need to know not only p1 , p2 , . . . , pn−1 ,
but also pn . . .(see [7], Chapter 3, and [2], Chapter XXII).
We shall use the principle of cross-classification in the proof of some simple
formulas for pn in terms of p1 , p2 , . . . , pn−1 .
As in the Gandhi’s formula, let Pn−1 = p1 p2 . . . pn−1 . We have the following
theorem.
Theorem 1.2 The following formulas hold
⎡ ⎤
∞

⎣
1
pn = 1 +  ⎦ (2)
i
i=1 d/Pn−1 µ(d) d
⎡ ⎤
∞

⎣
1
pn = pn−1 + 1 +  ⎦ (3)
pn−1 +i
i=1 d/Pn−1 µ(d) d
n
⎡ ⎤
2

⎣
1
pn = 1 +  ⎦ (4)
i
i=1 d/Pn−1 µ(d) d
n
⎡ ⎤
2

⎣
1
pn = pn−1 + 1 +   ⎦ (5)
pn−1 +i
i=1 d/Pn−1 µ(d) d

Proof. Let us consider the primes p1 , p2 , . . . , pn−1 (n ≥ 2) and a positive integer

N ≥ 1.
The number A(N) of numbers which do not exceed N and are not divisible
by any of the first n − 1 primes p1 , p2 , . . . , pn−1 will be (see (1))
     
 N  N  N
A(N) = N − + −
1≤i≤n−1 pi 1≤i<j≤n−1 pi pj 1≤i<j<k≤n−1
pi pj pk
   
N  N
n−1
+ . . . + (−1) = µ(d)
p1 p2 . . . pn−1 d/Pn−1
d
Formulas for primes and an algebraic identity 1653

Note that A(N) is nondecreasing and if 1 ≤ N ≤ (pn − 1) then A(N) = 1. On

the other hand if N = pn then A(N) = 2.
Consequently  
∞
1
pn = 1 +
i=1 A(i)

That is, formula (2).

Formula (2) can be written in the following way
∞
  pn−1 ∞
  
 1  1  1
pn = 1 + =1+ +
i=1 A(i) i=1 A(i) i=1 A(pn−1 + i)
∞
 
 1
= 1 + pn−1 +
i=1 A(pn−1 + i)

That is, formula (3). Formulas (4) and (5) are an immediate consequence
of formulas (2) and (3) since pn ≤ 2n (n ≥ 1) (see [2], chapter XXII). The
theorem is proved.

2 The Algebraic Identity

The following algebraic identity was known to Leonardo Fibonacci or Pisano
and was used implicity by Diophantos of Alexandria (see [6], page 196).

(a2 + b2 )(c2 + d2 ) = (ac + bd)2 + (ad − bc)2

By means of this identity one can represent the product of any two numbers
that are sums of two squares as the sum of two others squares.
The following algebraic identity is due to Euler (see [3], pages 213-214).

(a2 + b2 + c2 + d2 )(e2 + f 2 + g 2 + h2 ) =
(ae + bf + cg + dh)2 + (af − be − ch + dg)2
+ (ag + bh − ce − df )2 + (ah − bg + cf − de)2

By means of this identity one can represent the product of any two numbers
that are sums of four squares as the sum of four others squares. This identity
can be used in the proof of the famous result of Lagrange ”every positive
integer is a sum of four nonnegative integral squares” (see [1], page 144).
It was proved by Nathanson (see [5], page 34) that there does not exist an
algebraic identity of the form

(x21 + x22 + x23 )(y12 + y22 + y32 ) = z12 + z22 + z32

where z1 , z2 , z3 are polynomials in x1 , x2 , x3 , y1 , y2 , y3 with integral coefficients.

 1654 R. Jakimczuk

We present the following algebraic identity for the sum of eight squares.

(x21 + x22 + x23 + x24 + x25 + x26 + x27 + x28 )

(y12 + y22 + y32 + y42 + y52 + y62 + y72 + y82) =
(x1 y1 + x2 y2 + x3 y3 + x4 y4 + x5 y5 + x6 y6 + x7 y7 + x8 y8 )2
+ (x1 y2 − x2 y1 + x3 y4 − x4 y3 − x5 y6 + x6 y5 + x7 y8 − x8 y7 )2
+ (x1 y3 − x2 y4 − x3 y1 + x4 y2 + x5 y7 + x6 y8 − x7 y5 − x8 y6 )2
+ (x1 y4 + x2 y3 − x3 y2 − x4 y1 − x5 y8 + x6 y7 − x7 y6 + x8 y5 )2
+ (x1 y5 + x2 y6 − x3 y7 + x4 y8 − x5 y1 − x6 y2 + x7 y3 − x8 y4 )2
+ (x1 y6 − x2 y5 − x3 y8 − x4 y7 + x5 y2 − x6 y1 + x7 y4 + x8 y3 )2
+ (x1 y7 − x2 y8 + x3 y5 + x4 y6 − x5 y3 − x6 y4 − x7 y1 + x8 y2 )2
+ (x1 y8 + x2 y7 + x3 y6 − x4 y5 + x5 y4 − x6 y3 − x7 y2 − x8 y1 )2

Use for example the software Maple 9.5 of Maplesoft (a division of Waterloo
Maple inc.) in the verification of this identity.

ACKNOWLEDGEMENTS. The author is very grateful to Universidad

Nacional de Luján.

References
[1] G. E. Andrews, Number Theory, Dover Publications, Inc. New York.
1994.
[2] G. H. Hardy and E. M. Wright, An Introduction to The Theory of Num-
bers, Fourth edition. Oxford. 1960.
[3] A. M. Legendre, Théorie des Nombres, Tome I. Troisiéme Édition. Paris.
1830.
[4] W. J. LeVeque, Topics in Number Theory, Volume 1. First edition.
Addison-Wesley. 1958.
[5] M. B. Nathanson, Additive Number Theory. The Classical Bases,
Springer. 1996.
[6] Oystein Ore, Number Theory and Its History, Dover Publications, Inc.
New York. 1988.
[7] P. Ribenboim, The New Book of Prime Number Records, Springer. 1996.

Received: December, 2010

View publication stats