The N=16n+6N=16n+6 divisor-sum congruence conjecture

From papers

Let N=16n+6N=16n+6 for some nonnegative integer nn. Define \a0σ0(m)\a0\sigma_0(m) to be the number of positive divisors of mm. The N=16n+6N=16n+6 congruence conjecture.

1k<Nσ0(Nk2)0(mod4).\sum_{1\leq k<\sqrt{N}}\sigma_0(N-k^2)\equiv0\pmod4.

The authors report computational evidence for this congruence for small values of NN and conjecture that it extends the theorem, whose currently established conditions on NN are more restrictive. The general case remains open.

Progress summary

Open

No publicly verified progress on this divisor-sum congruence was found.

No public discussion or published progress concerning this conjecture was found in the retrieved sources.

Current status (as of August 2026): The conjecture remains open, with no recorded proof, counterexample, or verified advance.

Sources & referencesView supporting material

Primary source

Sittinon Jirattikansakul, Teeradej Kittipassorn, Kraiwich Kongsiri, Nitipon Moonwichit and Kirati Sriamorn, “Congruences via Partitions with Exactly Two Part Sizes”, arXiv:2604.25394 (2026).

Solutions 1

Proof

In fact the conjecture holds on the larger progression N6(mod8)N\equiv6\pmod8, simultaneously covering both N=16n+6N=16n+6 and N=16n+14N=16n+14.

For such NN, define

D(N)=#{(e,a)Z>02:e even, a odd, e2+2a2=N}.D(N) = \#\{(e,a)\in\mathbb Z_{>0}^2: e\text{ even},\ a\text{ odd},\ e^2+2a^2=N\}.

We prove separately that the odd- and even-index contributions both equal 2D(N)(mod4)2D(N)\pmod4.

Let χ4\chi_4 be the nontrivial character modulo 44. If m5(mod8)m\equiv5\pmod8, then mm is not a square and its divisors pair without fixed points as {d,m/d}\{d,m/d\}. Because m1(mod4)m\equiv1\pmod4, both divisors in each pair have the same χ4\chi_4-value. Thus each pair contributes 22 to the divisor count and either 22 or 2-2 to the character sum. Therefore

d(m)dmχ4(d)=r2(m)4(mod4),d(m)\equiv\sum_{d\mid m}\chi_4(d) =\frac{r_2(m)}4\pmod4,

where

r2(m)=#{(u,v)Z2:u2+v2=m}r_2(m)=\#\{(u,v)\in\mathbb Z^2:u^2+v^2=m\}

and the equality is Jacobi's two-square formula. For m5(mod8)m\equiv5\pmod8, exactly one coordinate is odd and the other is nonzero even. Hence r2(m)/4r_2(m)/4 counts ordered positive representations.

Write

Sodd=1j<N\j oddd(Nj2).S_{\mathrm{odd}} = \sum_{\substack{1\le j<\sqrt N\j\ \mathrm{odd}}} d(N-j^2).

For odd jj, Nj25(mod8)N-j^2\equiv5\pmod8. The preceding identity gives

Sodd2C(N)(mod4),S_{\mathrm{odd}}\equiv2C(N)\pmod4,

where

C(N)=#{(j,a,e)Z>03:j,a odd, e even, j2+a2+e2=N}.C(N) = \#\{(j,a,e)\in\mathbb Z_{>0}^3: j,a\text{ odd},\ e\text{ even},\ j^2+a^2+e^2=N\}.

The involution (j,a,e)(a,j,e)(j,a,e)\mapsto(a,j,e) pairs all off-diagonal triples. Its fixed points satisfy j=aj=a and are counted by D(N)D(N). Consequently

Sodd2D(N)(mod4).S_{\mathrm{odd}}\equiv2D(N)\pmod4.

Now let j>0j>0 be even. Since Nj22N-j^2\equiv2 or 6(mod8)6\pmod8, write Nj2=2wN-j^2=2w with ww odd. Then

d(Nj2)=2d(w){2(mod4),w is a square,0(mod4),otherwise.d(N-j^2)=2d(w) \equiv \begin{cases} 2\pmod4,&w\text{ is a square},\\ 0\pmod4,&\text{otherwise}. \end{cases}

The square cases are exactly j2+2a2=Nj^2+2a^2=N with jj even and aa positive odd. Therefore

Seven2D(N)(mod4).S_{\mathrm{even}}\equiv2D(N)\pmod4.

Adding the two contributions yields

1j<Nd(Nj2)4D(N)0(mod4)(N6(mod8)).\boxed{ \sum_{1\le j<\sqrt N}d(N-j^2) \equiv4D(N)\equiv0\pmod4 \qquad(N\equiv6\pmod8). }

Taking N=16n+6N=16n+6 proves Conjecture 7 for every n0n\ge0; the same theorem also recovers the previously established progression N=16n+14N=16n+14.

Source: “Congruences via Partitions with Exactly Two Part Sizes,” arXiv:2604.25394, Conjecture 7.

0 endorsements
Shivam Patel ·