Wiki
Wiki

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

Updated


Statement

Notation (p. 11): nk(p)n_k(p) is the least kk-th power nonresidue of the prime pp. The paper gives no further convention; in particular it does not say what nk(p)n_k(p) is for a prime pp with gcd⁡(k,p−1)=1\gcd(k,p-1)=1, which has no kk-th power nonresidue.

Conjecture (4) (p. 11, as printed; introduced as likely, not proved):

∑p<xnk(p)=(1+o(1)) ck xlog⁡x.\sum_{p<x}n_k(p)=\bigl(1+o(1)\bigr)\,c_k\,\frac{x}{\log x}.

The paper states no range of kk and nothing about ckc_k beyond its dependence on kk. The English summary (p. 17) words it "It is very likely true that ∑p<xnk(p)=(1+o(1))ckxlog⁡x\sum_{p<x}n_k(p)=(1+o(1))\frac{c_kx}{\log x}." For k=2k=2 the paper proves it, with c2=∑j≥1pj/2jc_2=\sum_{j\ge1}p_j/2^j: equation (3).

The obstruction (p. 11). For k>2k>2 the author has no good upper bound for the number of primes p<xp<x with nk(p)>A(x)n_k(p)>A(x), where A(x)→∞A(x)\to\infty with xx. The paper says that (4) would follow if that number were shown to be less than

c1 xlog⁡x A(x)2+c2,c_1\,\frac{x}{\log x\,A(x)^{2+c_2}},

and that this had not been done. Here c1c_1 and c2c_2 are unspecified constants; this c2c_2 is not the ckc_k of (4) at k=2k=2.

Source. P. Erdős, Számelméleti megjegyzések, I. (Remarks on number theory, I.; in Hungarian, with Russian and English summaries on p. 17), Mat. Lapok 12 (1961), 10--17; MR 26 #2410, Zbl 0154.294. The definition, conjecture (4) and the obstruction on printed p. 11, the English summary on p. 17; read on the page images of the edition identified on the source card.

Read depth. Claims checked: the definition, the display (4), the sufficient condition and the English summary were read clause by clause on the page images. The sufficiency claim is stated in the paper without proof and was not checked here.

Proof pointer

None: (4) is a conjecture in this paper, proved here only for k=2k=2, as equation (3).

Bears on

  • Problem 980: (4) is the problem's question, posed here for the least kk-th power nonresidue with no convention for the primes that have none; the problem page cites it at p. 11 and records how it reads the sum. This paper proves only the case k=2k=2.