Congruences via Partitions with Exactly Two Part Sizes
Sittinon Jirattikansakul, Teeradej Kittipassorn, Kraiwich Kongsiri, Nitipon Moonwichit, Kirati Sriamorn · 2026
Read Congruences via Partitions with Exactly Two Part Sizes by Sittinon Jirattikansakul, Teeradej Kittipassorn, Kraiwich Kongsiri, Nitipon Moonwichit, Kirati Sriamorn (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2604-25394.txt.
Introduction
Jirattikansakul and four colleagues prove a congruence for sums of divisor counts over squares. For N of five set forms, the sum of the divisor count of N minus a square, over squares below N, is a multiple of 4. Their proof leans on a result of Keith on partitions of N with exactly two part sizes. Following Dewitt and Keith, they join counting arguments with modular arithmetic. The work links partition counts to divisor sums through congruences.
Catalog record
| CREATOR | Sittinon Jirattikansakul, Teeradej Kittipassorn, Kraiwich Kongsiri, Nitipon Moonwichit, Kirati Sriamorn |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 24942 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2604.25394
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Congruences via Partitions with Exactly Two Part Sizes Sittinon Jirattikansakul, Teeradej Kittipassorn, Kraiwich Kongsiri, Nitipon Moonwichit, Kirati Sriamorn Abstract . We prove the congruence ∑ 1 ≤ k < N σ 0 ( N − k 2 ) ≡ 0 ( mod 4 ) \sum_{1\leq k<\sqrt{N}}\sigma_{0}(N-k^{2})\equiv 0\pmod{4} , where σ 0 ( m ) \sigma_{0}(m) denotes the number of positive divisors of m m , for N = A n + B N=An+B with ( A , B ) ∈ { ( 16 , 14 ) , (A,B)\in\{(16,14), ( 36 , 30 ) , (36,30), ( 72 , 42 ) , (72,42), ( 196 , 70 ) , (196,70), ( 252,114 ) } (252,114)\} . Our proof relies on a result of Keith which states that ν 2 ( N ) ≡ 0 ( mod 4 ) \nu_{2}(N)\equiv 0\pmod{4} , where ν 2 ( N ) \nu_{2}(N) is the number of partitions of N N with exactly two part sizes. Inspired by Dewitt and Keith, our approach combines combinatorial arguments with modular arithmetic techniques. 1. Introduction Integer partitions play a central role in number theory and combinatorics. Since the foundational work of Euler [ 4 ] , partitions have been studied from a variety of perspectives, including generating functions by Andrews [ 2 ] , asymptotic analysis by Hardy and Ramanujan [ 5 ] , and geometric interpretations via Ferrers and Young diagrams by MacMahon [ 8 ] , among many others. Define \nu_{k}(N) to be the number of partitions of N in which exactly k sizes of part appear . For example, \nu_{2}(4)=2 since all partitions of 4 with exactly two part sizes are (3,1) and (2,1,1) , and \nu_{3}(4)=0 since no partition of 4 has exactly three part sizes. This counting function has been studied by MacMahon [ 9 ] , Andrews [ 1 ] , and Tani and Bouroubi [ 10 ] where the first two independently derived a compact formula for the case k=2 as \displaystyle\nu_{2}(N)=\frac{1}{2}\left(\sum_{k=1}^{N-1}\sigma_{0}(k)\sigma_{0}(N-k)-\sigma_{1}(N)+\sigma_{0}(N)\right) where \sigma_{0}(N) is the number of positive divisors of N and \sigma_{1}(N) is the sum of the positive divisors of N . Keith [ 7 ] then later proved that \nu_{2}(An+B)\equiv 0\pmod{4} when (A,B) is one of (16,14), (36,30), (72,42), (196,70) and (252,114). This inspires us to revisit several results of Keith [ 3 ] using a finer combinatorial approach based on multiset partition. In particular, we study configurations of Young diagrams of partitions of N with exactly two part sizes in the arithmetic sequences considered by Keith [ 7 ] , and obtain the following congruence. Theorem 1 . Let (A,B)\in\{(16,14),(36,30),(72,42),(196,70),(252,114)\} and N=An+B for some nonnegative integer n . Then \sum\limits_{1\leq k<\sqrt{N}}\sigma_{0}(N-k^{2})\equiv 0\pmod{4}. The sum \sum_{1\leq k<\sqrt{N}}\sigma_{0}(N-k^{2}) in Theorem 1 and its asymptotic behavior were investigated by Hooley [ 6 ] in his study of representations of N as the sum of a square and a product where the number of such representations is equal to \sum_{-\sqrt{N}<k<\sqrt{N}}\sigma_{0}(N-k^{2}) . The following corollaries are consequences of Theorem 1 . Corollary 2 . Let N be a positive integer such that N\equiv 14\pmod{16} or N\equiv 70\pmod{196} . Then the number of positive odd integers k such that k^{2}<N and \sigma_{0}(N-k^{2})\equiv 2\pmod{4} is always even. Corollary 3 . Let N be a positive integer such that N\equiv B\pmod{A} where (A,B) is one of (36,30),(72,42), and (252,114) . Then \sum\limits_{\begin{subarray}{c}1\leq k<\sqrt{N}\\ 3\nmid k\end{subarray}}\sigma_{0}(N-k^{2})\equiv 0\pmod{4}. To prove Theorem 1 , we first prove an auxiliary congruence relation in Theorem 4 by assuming a weaker assumption on the natural number N . Theorem 4 . Let N be a natural number satisfying • N cannot be written as a sum of two squares, and • N=2m for some odd integer m . Then \nu_{2}(N)+\sum\limits_{1\leq k<\sqrt{N}}\sigma_{0}(N-k^{2})+\frac{1}{2}\sigma_{1}(N)\equiv 0\pmod{4}. Note that according to the formula for \nu_{2}(N) given in the introduction, the congruence in Theorem 4 can be written as \displaystyle\frac{1}{2}\left(\sigma_{0}(N)+\sum_{k=1}^{N-1}\sigma_{0}(k)\sigma_{0}(N-k)\right)+\sum_{1\leq k<\sqrt{N}}\sigma_{0}(N-k^{2})\equiv 0\pmod{4}. To the best of our knowledge, this congruence does not appear explicitly in the existing literature. We organize the rest of the paper as follows. In Section 2 , we discuss background and terminology. Section 3 is devoted to the proof of Theorem 1 , which serves as our main result. In Section 4 , we apply Theorem 1 and use some elementary number theory to prove Corollaries 2 and 3 . Finally, in Section 5 , we conclude the paper with some remarks and a conjecture inspired by our findings. 2. Background Fix a natural number n . A partition of n with exactly one part size is a tuple \vec{v}=(n_{1},n_{1},\ldots,n_{1}) of k copies of n_{1} for some k\geq 1 satisfying kn_{1}=n . A partition of n with exactly two part sizes is a tuple \vec{v}=(n_{1},n_{1},\ldots,n_{1},n_{2},n_{2},\ldots,n_{2}) where n_{1}>n_{2}\geq 1 such that if n_{1} appears k_{1} times and n_{2} appears k_{2} times, then k_{1}n_{1}+k_{2}n_{2}=n . In this paper, when we mention a rectangle , we mean a usual rectangular grid where each cell is a square. An m\times n rectangle is a rectangle consisting of m rows and n columns. If X is an m\times n rectangle, denote by X^{T} the n\times m rectangle. A Young diagram for a partition is the unique grid of cells where the cells are arranged in left-justified rows with the row lengths in non-increasing order. For each Young diagram consisting of n cells in k rows, we associate the diagram with the tuple (n_{1},n_{2},\ldots,n_{k}) where n_{i} is the number of cells lying in the i -th row. By this interpretation, each Young diagram in rectangular shape consisting of n cells uniquely determines the partition of n with exactly one part size. Likewise, each Young diagram in L-shape uniquely determines the partition of n with exactly two part sizes. For example, Figure 1(a) is interpreted as the partition (3,3,3,3) of 12 , and Figure 1(b) is interpreted as the partition (5,5,5,2,2) of 19 . (a) (b) Figure 1 . The Young diagrams for the partitions: (a) (3,3,3,3) of 12 , and (b) (5,5,5,2,2) of 19 Any Young diagram corresponding to a partition with exactly two part sizes can be viewed as obtained by gluing two rectangles vertically, with the rectangle having the wider base placed on top. Conversely, every such L-shaped diagram can be decomposed in this way into a unique pair of rectangles. This observation motivates our approach and leads to a proof based on counting pairs of rectangles. 3. Proof of Theorem 1 Let N be a natural number that cannot be written as a sum of two squares. A canonical pair for N is an unordered pair of rectangles \{X,Y\} such that the numbers of rows of X and Y are fewer than or equal to the numbers of columns of X and Y , respectively, and the sum of numbers of cells of X and Y is N . Let \mathcal{A} be the multiset consisting of \{X,Y\},\{X,Y^{T}\},\{X^{T},Y\},\{X^{T},Y^{T}\} for all canonical pairs \{X,Y\} for N . We define subsets \mathcal{B},\mathcal{C},\mathcal{D} , and \mathcal{E} of \mathcal{A} as follows: \displaystyle\mathcal{B} \displaystyle=\{\{X,Y\}\in\mathcal{A}\ |\ \text{the numbers of columns of $X$ and $Y$ are different}\}, \displaystyle\mathcal{C} \displaystyle=\{\{X,Y\}\in\mathcal{A}\ |\ \text{$X$ or $Y$ is a square}\}, \displaystyle\mathcal{D} \displaystyle=\{\{X,Y\}\in\mathcal{A}\ |\ \text{the numbers of columns of $X$ and $Y$ are the same}\},