Wiki
Wiki

Research notes on every problem, and a library of the papers behind them. Built from the open erdos repository.

Updated


Statement

Setting: (1) is the equation a/n=1/x+1/y+1/za/n=1/x+1/y+1/z in positive integers x,y,zx,y,z, repetition allowed; q,r,s,tq,r,s,t denote positive integers, and the paper assumes a>3a>3 throughout (p. 193). See the Theorem for the setting in full.

Lemma 1 (p. 193). "Suppose that rn+s≡0 (mod arst−1)rn+s\equiv0\ (\mathrm{mod}\ arst-1). Then (1) is soluble."

The proof (p. 193) names the solution: if rn+s+q=arstqrn+s+q=arstq, then x=stqx=stq, y=nrtqy=nrtq, z=nrstz=nrst solve (1). The congruence supplies such a qq, namely q=(rn+s)/(arst−1)q=(rn+s)/(arst-1), which is a positive integer.

Source. R. C. Vaughan, On a problem of Erdős, Straus and Schinzel, Mathematika 17 (1970), 193--198, doi:10.1112/S0025579300002886; Lemma 1 and its proof on p. 193. The edition is identified on the source card.

Read depth. Claims checked: the statement was read clause by clause on the page image, and the one-line proof was followed: with rn+s+q=arstqrn+s+q=arstq, 1/(stq)+1/(nrtq)+1/(nrst)=(rn+s+q)/(nrstq)=a/n1/(stq)+1/(nrtq)+1/(nrst)=(rn+s+q)/(nrstq)=a/n. Nothing here is independently reviewed.

Proof pointer

P. 193, one line: the identity above. For a=4a=4 the paper points to Chapter 30, § 1 of Mordell's Diophantine equations (1969) for solutions of the same kind.

Dependencies

None.

Bears on

  • Problem 242: at a=4a=4 the lemma is a sufficient condition for 4/n4/n to be a sum of three unit fractions, possibly with repeated denominators: nn is representable whenever rn+s≡0(mod4rst−1)rn+s\equiv0\pmod{4rst-1} for some positive integers r,s,tr,s,t. It gives a representation for each nn in the residue classes it covers and decides no case outside them. The paper uses it, through Lemma 2, to sieve the exceptions in the Theorem.