A SHORT PROOF OF ERDÖS–STRAUS CONJECTURE FOR

EVERY n ≡ 13 mod 24
arXiv:2005.03273v3 [math.NT] 24 Aug 2021

MARIO GIONFRIDDO AND ELENA GUARDO

4
Abstract. The Erdös–Straus conjecture states that the equation n = x1 +
1 1
y
+ z has positive integer solutions x, y, z for every postive integers n ≥ 2. In
this short note we find explicity the solutions of the famous conjecture for the
case n ≡ 13 mod 24.

1. Introduction
In Number Theory there are many unsolvable problems that still attract
a lot of attention. Among all, there is a famous conjecture of Erdös–Straus
which states that, for all positive integers n ≥ 2, the rational number n4 can
be expressed as the sum of three positive unit fractions. Specifically, in 1948
P. Erdös and E. G. Straus formulated the following:
Conjecture 1.1. For every positive integer n ≥ 2 there exist positive inte-
gers x, y, z such that:
4 1 1 1
(1) = + + .
n x y z
This conjecture has attracted a lot of attention not only among reseachers
in Number Theory but also among many people involved among the different
areas in Mathematics, such as L. Bernstein [1], M. B. Crawford [3], M. Di
Giovanni, S. Gallipoli, M. Gionfriddo [4], C. Elsholtz and T. Tao [5], J.
Guanouchi [6, 7], S. K. Jeswal and S. Chakraverty [8], L. J. Mordell [9], D.
J. Negash [10], R. Obláth [11], L. A. Rosati [12], J. W. Sander [14], R. C.
Vaughan [16], K. Yamamoto [18], just to cite some of them. For example,
Swett in [15] has established validity of the conjecture for all n ≤ 1014 and it
also appears in The Penguin Dictionary of Curious and Interesting Numbers,
[17], or S. E. Salez [13] that proved that the conjecture holds for all n ≤ 1017 .
It is not clear if Conjecture 1.1 is true or not since there are many papers
in which some authors adfirm to have proved that the conjecture is true and
others in which the authors proved that it is false (see [19]). Since we have
found some mistakes in some of these published papers, it is our opinion
that this conjecture is still open.
L. J. Mordell, in [9], Chapter 30, Section 1, has only proven that the
conjecture is true for all n except possibly cases in which n is congruent
2010 Mathematics Subject Classification. 11D72, 11D45, 11P81 .
Key words and phrases. Conjecture Erdö–Straus, Diophantine equation .
Last updated: March 10, 2021.
1
2 M. GIONFRIDDO AND E. GUARDO

to 1, 121, 169, 289, 361, 529 mod 840, but he did not write explicity the so-
lution as we do in Theorem 3.1. In all the known literature, for the case
n 6≡ 1 mod 24 the authors prove only the existence of the solutions without
writing them explicity.
Since it is known that for n 6≡ 1 mod 24 solutions of (1) can be found
constructively and that for the remaining residue class n ≡ 1 mod 24 the
problem can be reduced to smaller sets with some exceptions (see [9] and
[14]), we have divided the problem into two parts, n ≡ 1 mod 24 and n ≡
13 mod 24. In Section 2, we recall all the known cases in the literature
and in Section 3 we propose our short proof of the conjecture for the case
n ≡ 13 mod 24. Therefore, to completely solve Conjecture 1.1, it remains
to find solutions for every n ≡ 1 mod 24.
Acknowledgement . The authors were supported by Università degli Studi
di Catania, “Piano della Ricerca PIACERI 2018/2020 Linea di intervento
2”. The second author is a member of GNSAGA of INdAM (Italy).

2. Known cases
2.1. n even. The case n even is easily solvable, as we can see in the following
theorem.
Theorem 2.1. Conjecture 1.1 is true for n even.
Proof. Let n ≡ 0, 2 mod 4.
(1) If n = 4k, then for: x = 2k, y = 4k, z = 4k, it is possible to verify
that the conjecture is true.
(2) If 4n = 4k + 2, then for: x = 2k + 1, y = 4k + 2, z = 4k + 2, it is
possible to verity that the conjecture is true.

2.2. n odd. The following results are already known for the case n odd.
Theorem 2.2. Conjecture 1.1 is true for n ≡ 3 mod 4.
Proof. If n = 4k + 3, then for: x = 2k + 2, y = 2k + 2, z = (k + 1)(4k + 3),
it is possible to verity that the conjecture is true (see [2]). 
Theorem 2.3. Conjecture 1.1 is true for n ≡ 3, 5 mod 6.
Proof. Let n ≡ 3, 5 mod 6 (see [2]).
If n = 6k + 3, then for: x = 6k + 3, y = 2k + 2, z = (2k + 1)(2k + 2),
it is possible to verify that the conjecture is true. If n = 6k + 5, then for:
x = 6k + 5, y = 2k + 2, z = (6k + 5)(2k + 2), it is possible to verify that the
conjecture is true. 
Collecting together the results proved in Theorems 2.2 and 2.3, we have
that:
Theorem 2.4. [2] Conjecture 1.1 is true for every n ≡ 3, 5, 7, 9, 11 mod 12.
 A SHORT PROOF OF ERDÖS–STRAUS CONJECTURE FOR EVERY n ≡ 13 mod 24 3

Proof. Indeed:
(1) if n ≡ 3 mod 12, then it is n ≡ 3 mod and also n ≡ 3 mod 6;
(2) if n ≡ 5 mod 12, then it is n ≡ 1 mod 4, but also n ≡ 5 mod 6;
(3) if n ≡ 7 mod 12, then it is n ≡ 1 mod 6, but also n ≡ 3 mod 4;
(4) if n ≡ 9 mod 12, then it is n ≡ 1 mod 4, but also n ≡ 3 mod 6;
(5) if n ≡ 11 mod 12, then it is n ≡ 3 mod 4 and also n ≡ 5 mod 6.

Therefore, it remains to examine the case n ≡ 1 mod 12. We have divided
the problem into two parts, n ≡ 1 mod 24 and n ≡ 13 mod 24, and get a
very short proof of the conjecture for the case n ≡ 13 mod 24.

3. Main Result: Solutions for n ≡ 13 mod 24.

Theorem 3.1. Conjecture 1.1 is true for every n ≡ 13 mod 24.
Proof. : Let n = 24k + 13, for any non negative integer k. If:
x = (2(k + 1)(24k + 13)),
y = (2(3k + 2)),
z = (2(k + 1)(24k + 13)(3k + 2)),
then it follows:
1 1 1
2(k+1)(24k+13)
+ 2(3k+2)
+ 2(k+1)(24k+13)(3k+2)
=
(3k+2)+(k+1)(24k+13)+1
= 2(k+1)(24k+13)(3k+2)
=
8(3k 2 +5k+2)
= 2(k+1)(24k+13)(3k+2)
=
8(3k+2)(k+1)
= 2(k+1)(24k+13)(3k+2)
=
4
= (24k+13)

which proves the statement. 

References
[1] L. Bernstein, Zur Losung der diophantischen Gieichung m/n = 1/x+1/y +1/z, insbesondere
im Fall m = 4, Journal fur die Reine und Angewandte Mathematik Vol. 211 (1962), 1–10.
[2] V. Castelli, M. Di Giovanni, S. Gallipoli, M. Gionfriddo, On a conjecture of Erdös-Straus,
Bulletin of Mathematics and Statistics Research, Vol.7, issue 4 (2019), 19–23.
[3] M. B. Crawford, On the Number of Representations of One as the Sum of Unit Fractions,
Masters of Science in Mathematics, Virginia Polytechnic Institute and State University,
(2019).
[4] M. Di Giovanni, S. Gallipoli, M. Gionfriddo, Historical origin and scientific development of
Graphs, Hypergraphs and Design Theory, Bulletin of Mathematics and Statistics Research,
Vol 7 (2019),
[5] C. Elsholtz, T. Tao, Counting the number of solutions to the Erdös-Straus equations on unit
fractions, J. of Australasian Math. Society, Vol 94, issue 1 (2013), pp 50–105
[6] J. Ghanouchi, An analytic approach of some conjectures related to diophantine
equations, Bulletin of Mathematical Sciences and Applications, Vol. 1, pp 29–40
doi:10.18052/www.scipress.com/BMSA.1.29
4 M. GIONFRIDDO AND E. GUARDO

[7] J. Ghanouchi, About the Erdös conjecture, International Journal of Sciinece and Research,
Vol. 4 (2), (2015), 341
[8] S. K. Jeswal, S. Chakraverty Connectionist based models for solving Diophan-
tine equation, Journal of Interdisciplinary Mathematics, 23:4, 825–841 (2020), DOI:
10.1080/09720502.2020.1712844
[9] L. J. Mordell, Diophantine Equations , Academic Press, London/New York, (1969)
[10] D. J. Negash, Solutions to Dophantine Equation of Erdös-Straus Conjecture,
https://arxiv.org/pdf/1812.05684.pdf
[11] R. Obláth, Sur l’equation diophantienne 4/n = 1/x1 + 1/x2 + 1/x3 , Mathesis, Vol 59 (1950),
pp 308–316
[12] L. A. Rosati, Sull’equazione diofantea 4/n = 1/x + 1/y + 1/z, Bollettino dell’Unione Matem-
atica Italiana Vol. 3 (1954), 59–63.
[13] S. E. Salez, The Erdös–Straus conjecture: New modular equations and checking up to n =
1017 , arXiv 1406.6307 (2014).
[14] J. W. Sander, 4/n = 1/x + 1/y + 1/z and Iwaniec’ Half Dimensional Sieve, J. of Number
Theory, Vol. 46, (1994), pp 123–136
[15] A. Swett, The Erdös-Straus Conjecture. Rev. 10/28/99.
http://math.uindy.edu/swett/esc.htm
[16] R. C. Vaughan, On a Problem of Erdös, Straus and Schinzel. Mathematika Vol 17, (1970)
pp. 193–198.
[17] D. Wells, The Penguin Dictionary of Curious and Interesting Numbers.. Middlesex, England:
Penguin Books, pp. 29, (1986).
[18] K. Yamamoto, On the Diophantine equation 4/n = 1/x + 1/y + 1/z, Memoirs of the Faculty
of Science, Kyushu University Vol. 19 (1965), 37–47.
[19] S. Zou, Erdös–Straus conjecture is wrong, Bulletin of Mathematics and Statistics Research
Vol. 7 (2019), 59–60

Dipartimento di Matematica e Informatica, Viale A. Doria, 6 - 95100 - Catania, Italy

Email address: gionfriddo@dmi.unict.it
URL: https://www.researchgate.net/profile/Mario Gionfriddo

Dipartimento di Matematica e Informatica, Viale A. Doria, 6 - 95100 - Catania, Italy

Email address: guardo@dmi.unict.it
URL: www.dmi.unict.it/guardo