Open Journal of Mathematical Sciences

Congruence properties of indices of triangular numbers multiple of other triangular numbers

Vladimir Pletser
European Space Agency (ret.); pletservladimir@gmail.com

Abstract

For any non-square integer multiplier \(k\), there is an infinity of triangular numbers multiple of other triangular numbers. We analyze the congruence properties of indices \(\xi\) of triangular numbers multiple of triangular numbers. Remainders in congruence relations \(\xi\) modulo \(k\) come always in pairs whose sum always equal \((k-1)\), always include 0 and \((k-1)\), and only 0 and \((k-1)\) if \(k\) is prime, or an odd power of a prime, or an even square plus one or an odd square minus one or minus two. If the multiplier \(k\) is twice the triangular number of \(n\), the set of remainders includes also \(n\) and \((n^{2}-1)\) and if \(k\) has integer factors, the set of remainders include multiples of a factor following certain rules. Algebraic expressions are found for remainders in function of \(k\) and its factors, with several exceptions. This approach eliminates those \(\xi\) values not providing solutions.

Keywords:

Triangular numbers; Multiple of triangular numbers; Recurrent relations; Congruence properties.

1. Introduction

Triangular numbers \(T_{t}=\frac{t\left(t+1\right)}{2}\) are one of the figurate numbers enjoying many properties; see, e.g., [1,2] for relations and formulas. Triangular numbers \(T_{\xi}\) that are multiples of other triangular number \(T_{t}\)

\begin{equation} T_{\xi}=kT_{t}\label{eq:1} \end{equation}
(1)
are investigated. Only solutions for \(k>1\) are considered as the cases \(k=0\) and \(k=1\) yield respectively \(\xi=0\) and \(\xi=t,\forall t\). Accounts of previous attempts to characterize these triangular numbers multiple of other triangular numbers can be found in [3,4,5,6,7,8,9]. Recently, Pletser [9] showed that, for non-square integer values of \(k\), there are infinitely many solutions that can be represented simply by recurrent relations of the four variables \(t,\xi,Tt\) and \(T_{\xi}\), involving a rank \(r\) and parameters \(\kappa\) and \(\gamma\), which are respectively the sum and the product of the \(\left(r-1\right)^{\text{th}}\) and the \(r^{\text{th}}\) values of \(t\). The rank \(r\) is being defined as the number of successive values of \(t\) solutions of (1) such that their successive ratios are slowly decreasing without jumps.

In this paper, we present a method based on the congruent properties of \(\xi\left(\text{mod}\,k\right)\), searching for expressions of the remainders in function of \(k\) or its factors. This approach accelerates the numerical search of the values of \(t_{n}\) and \(\xi_{n}\) that solve (1), as it eliminates values of \(\xi\) that are known not to provide solutions to (1). The gain is typically in the order of \(k/\upsilon\) where \(\upsilon\) is the number of remainders, which is usually such that \(\upsilon\ll k\).

2. Rank and Recurrent Equations

Sequences of solutions of (1) are known for \(k=2,3,5,6,7,8,10\) and are listed in the Online Encyclopedia of Integer Sequences (OEIS) [10], with references given in Table 1.

Among all solutions, \(t=0\) is always a first solution of (1) for all non-square integer value of \(k\), yielding \(\xi=0\).

Table 1. OEIS [10] references of sequences of integer solutions of (1) for \(k=2,3,5,6,7,8,10\).
\(k\) 2 3 5 6 7 8 10
\(t\) A053141 A061278 A077259 A077288 A077398 A336623 A341893
\(\xi\) A001652 A001571 A077262 A077291 A077401 A336625 A341895
\(T_{t}\) A075528 A076139 A077260 A077289 A077399 A336624 A068085
\(T_{\xi}\) A029549 A076140 A077261 A077290 A077400 A336626

Let's consider the two cases of \(k=2\) and \(k=7\) yielding the successive solution pairs as shown in Table 2. We indicate also the ratios \(t_{n}/t_{n-1}\) for both cases and \(t_{n}/t_{n-2}\) for \(k=7\). It is seen that for \(k=2\), the ratio \(t_{n}/t_{n-1}\) varies between close values, from 7 down to 5.829, while for \(k=7\), the ratio \(t_{n}/t_{n-1}\) alternates between values 2.5 ... 2.216 and 7.8 ... 7.23, while the ratio \(t_{n}/t_{n-2}\) decreases regularly from 19.5 to 16.023 (corresponding approximately to the product of the alternating values of the ratio \(t_{n}/t_{n-1}\)). We call rank \(r\) the integer value such that \(t_{n}/t_{n-r}\) is approximately constant or, better, decreases regularly without jumps (a more precise definition is given further). So, here, the case \(k=2\) has rank \(r=1\) and the case \(k=7\) has rank \(r=2\).

Table 2. Solutions of (1) for \(k=2,7\).
n \(k=2\) \(k=7\)
\(t_{n}\) \(\xi_{n}\) \(\frac{t_{n}}{t_{n-1}}\) \(t_{n}\) \(\xi_{n}\) \(\frac{t_{n}}{t_{n-1}}\) \(\frac{t_{n}}{t_{n-2}}\)
0 0 0 0 0
1 2 3 -- 2 6 -- --
2 14 20 7 5 14 2.5 --
3 84 119 6 39 104 7.8 19.5
4 492 696 5.857 87 231 2.231 17.4
5 2870 4059 5.833 629 1665 7.230 16.128
6 16730 23660 5.829 1394 3689 2.216 16.023

In [9], we showed that the rank \(r\) is the index of \(t_{r}\) and \(\xi_{r}\) solutions of (1) such that

\begin{equation} \kappa=t_{r}+t_{r-1}=\xi_{r}-\xi_{r-1}-1,\label{eq:3.2} \end{equation}
(2)
and that the ratio \(t_{2r}/t_{r}\), corrected by the ratio \(t_{r-1}/t_{r}\), is equal to a constant \(2\kappa+3\)
\begin{equation} \frac{t_{2r}-t_{r-1}}{t_{r}}=2\kappa+3.\label{eq:3-0} \end{equation}
(3)
For example, for \(k=7\) and \(r=2\), (2) and (3) yield respectively, \(\kappa=7\) and \(2\kappa+3=17\).

Four recurrent equations for \(t_{n},\xi_{n},T_{t_{n}}\) and \(T_{\xi_{n}}\) are given in [9] for each non-square integer value of \(k\)

\begin{align} t_{n} & =2\left(\kappa+1\right)t_{n-r}-t_{n-2r}+\kappa,\label{eq:3.3}\\ \end{align}
(4)
\begin{align} \xi_{n} & =2\left(\kappa+1\right)\xi_{n-r}-\xi_{n-2r}+\kappa,\label{eq:3.3-1}\\ \end{align}
(5)
\begin{align} T_{t_{n}} & =\left(4\left(\kappa+1\right)^{2}-2\right)T_{t_{n-r}}-T_{t_{n-2r}}+\left(T_{\kappa}-\gamma\right),\label{eq:3.3-2}\\ \end{align}
(6)
\begin{align} T_{\xi_{n}} & =\left(4\left(\kappa+1\right)^{2}-2\right)T_{\xi_{n-r}}-T_{\xi_{n-2r}}+k\left(T_{\kappa}-\gamma\right),\label{eq:3.3-3} \end{align}
(7)
where coefficients are functions of two constants \(\kappa\) and \(\gamma\), respectively the sum \(\kappa\) and the product \(\gamma=t_{r-1}t_{r}\) of the first two sequential values of \(t_{r}\) and \(t_{r-1}\). Note that the first three relations (4) to (6) are independent of the value of \(k\).

3. Congruence of \(\xi\) modulo \(k\)

We use the following notations: for \(A,B,C\in\mathbb{Z},B< C,C>1\), \(A\equiv B\left(\text{mod}\,C\right)\) means that \(\exists D\in\mathbb{Z}\) such that \(A=DC+B\), where \(B\) and \(C\) are called respectively the remainder and the modulus. To search numerically for the values of \(t_{n}\) and \(\xi_{n}\) that solve (1), one can use the congruent properties of \(\xi\left(\text{mod}\,k\right)\) given in the following propositions. In other words, we search in the following propositions for expressions of the remainders in function of \(k\) or its factors.

Proposition 1. For \(\forall s,k\in\mathbb{Z}^{+}\), \(k\) non-square, \(\exists\ \ \xi,\mu,\upsilon,i,j\in\mathbb{Z}^{+}\), such that if \(\xi_{i}\) are solutions of (1), then for \(\xi_{i}\equiv\mu_{j}\left(\text{mod}\,k\right)\) with \(1\leq j\leq\upsilon\), the number \(\upsilon\) of remainders is always even, \(\upsilon\equiv0\left(\text{mod}\,2\right)\), the remainders come in pairs whose sum is always equal to \(\left(k-1\right)\), and the sum of all remainders is always equal to the product of \(\left(k-1\right)\) and the number of remainder pairs, \(\sum\limits_{j=1}^{\upsilon}\mu_{j}=\left(k-1\right)\upsilon/2\).

Proof. Let \(s,i,j,k,\xi,\mu,\upsilon,\alpha,\beta\in\mathbb{Z}^{+}\), \(k\) non-square, and \(\xi_{i}\) solutions of (1). Rewriting (1) as \(T_{t_{i}}=T_{\xi_{i}}/k\), for \(T_{t_{i}}\) to be integer, \(k\) must divide exactly \(T_{\xi_{i}}=\xi_{i}\left(\xi_{i}+1\right)/2\), i.e., among all possibilities, \(k\) divides either \(\xi_{i}\) or \(\left(\xi_{i}+1\right)\), yielding two possible solutions \(\xi_{i}\equiv0\left(\text{mod}\,k\right)\) or \(\xi_{i}\equiv-1\left(\text{mod}\,k\right)\), i.e., \(\upsilon=2\) and the set of \(\mu_{j}\) includes \(\left\{ 0,\left(k-1\right)\right\} \). This means that \(\xi_{i}\) are always congruent to either \(0\) or \(\left(k-1\right)\) modulo \(k\) for all non-square values of \(k\).

Furthermore, if some \(\xi_{i}\) are congruent to \(\alpha\) modulo \(k\), then other \(\xi_{i}\) are also congruent to \(\beta\) modulo \(k\) with \(\beta=\left(k-\alpha-1\right)\). As \(\xi_{i}\equiv\alpha\left(\text{mod}\,k\right)\), then \(\xi_{i}\left(\xi_{i}+1\right)/2\equiv\left(\alpha\left(\alpha+1\right)/2\right)\left(\text{mod}\,k\right)\) and replacing \(\alpha\) by \(\alpha=\left(k-\beta-1\right)\) yields \(\left(\alpha\left(\alpha+1\right)/2\right)=\left(\left(k-\beta-1\right)\left(k-\beta\right)/2\right)\), giving \(\xi_{i}\left(\xi_{i}+1\right)/2\equiv\left(\left(k-\beta-1\right)\left(k-\beta\right)/2\right)\left(\text{mod}\,k\right)\equiv\left(\beta\left(\beta+1\right)/2\right)\left(\text{mod}\,k\right)\). In this case, \(\upsilon=4\) and the set of \(\mu_{j}\) includes, but not necessarily limits to, \(\left\{ 0,\alpha,\left(k-\alpha-1\right),\left(k-1\right)\right\} \).

Note that in some cases, \(\upsilon>4\), as for \(k=66,70,78,105,...\), \(\nu=8\). However, in some other cases, \(\upsilon=2\) only and the set of \(\mu_{j}\) contains only \(\left\{ 0,\left(k-1\right)\right\} \), as shown in the next proposition. In this proposition, several rules (R) are given constraining the congruence characteristics of \(\xi_{i}\).

Proposition 2. For \(\forall\ \ s,k,\alpha,n\in\mathbb{Z}^{+}\), \(k\) non-square, \(\alpha>1\), \(\exists\ \ \xi,\mu,\upsilon,i\in\mathbb{Z}^{+}\), such that if \(\xi_{i}\) are solutions of (1), then \(\xi_{i}\) are always only congruent to \(0\) and \(\left(k-1\right)\) modulo \(k\), and \(\upsilon=2\) if either

  • (R1) \(k\) is prime, or
  • (R2) \(k=\alpha^{n}\) with \(\alpha\) prime and \(n\) odd, or
  • (R3) \(k=s^{2}+1\) with \(s\) even, or
  • (R4) \(k=s^{\prime2}-1\) or (R5) \(k=s^{\prime2}-2\) with \(s^{\prime}\) odd.

Proof. Let \(s,s^{\prime},k,\alpha>1,n,i,\xi,\mu,\upsilon\in\mathbb{Z}^{+}\), \(k\) non-square, and \(\xi_{i}\) are solutions of (1).

(R1)+(R2): If \(k\) is prime or if \(k=\alpha^{n}\) (with \(\alpha\) prime and \(n\) odd as \(k\) is non-square), then, in both cases, \(k\) can only divide either \(\xi_{i}\) or \(\left(\xi_{i}+1\right)\), yielding the two congruences \(\xi_{i}\equiv0\left(\text{mod}\,k\right)\) and \(\xi_{i}\equiv-1\left(\text{mod}\,k\right)\).

(R3): If \(k=s^{2}+1\) with \(s\) even, the rank \(r\) is always \(r=2\) [11], and the only two sets of solutions are

\begin{align} \left(t_{1},\xi_{1}\right) & =\left(s\left(s-1\right),\left(s^{2}+1\right)\left(s-1\right)\right)\label{eq:2.8}\\ \end{align}
(8)
\begin{align} \left(t_{2},\xi_{2}\right) & =\left(s\left(s+1\right),\left(s^{2}+1\right)\left(s+1\right)-1\right)\label{eq:2.9} \end{align}
(9)
as can be easily shown. For \(t_{1}\), forming \begin{align*} kT_{t_{1}} & =\frac{1}{2}\left(s^{2}+1\right)\left(s\left(s-1\right)\right)\left(s\left(s-1\right)+1\right)\\ & =\frac{1}{2}\left[\left(s^{2}+1\right)\left(s-1\right)\right]\left[\left(s^{2}+1\right)\left(s-1\right)+1\right]=T_{\xi_{1}} \end{align*} which is the triangular number of \(\xi_{1}\). One obtains similarly \(\xi_{2}\) from \(t_{2}\). These two relations (8) and (9) show respectively that \(\xi_{1}\) is congruent to \(0\) modulo \(k\) and \(\xi_{2}\) is congruent to \(\left(k-1\right)\) modulo \(k\).

(R4): For \(k=s^{\prime2}-1\) with \(s^{\prime}\) odd, the rank \(r=2\) [11], and the only two sets of solutions are

\begin{align} \left(t_{1},\xi_{1}\right) & =\left(\left(s^{\prime}-1\right)s^{\prime}-1,\left(s^{\prime2}-1\right)\left(s^{\prime}-1\right)-1\right)\label{eq:2.13}\\ \end{align}
(10)
\begin{align} \left(t_{2},\xi_{2}\right) & =\left(\left(s^{\prime}-1\right)\left(s^{\prime}+2\right)+1,\left(s^{\prime2}-1\right)\left(s^{\prime}+1\right)\right)\label{eq:2.14} \end{align}
(11)
as can be easily demonstrated as above. These two relations (10) and (11) show that \(\xi_{1}\) and \(\xi_{2}\) are congruent respectively to \(\left(k-1\right)\) and \(0\) modulo \(k\).

(R5): For \(k=s^{\prime2}-2\) with \(s^{\prime}\) odd, the rank \(r=2\) [11], and the only two sets of solutions are

\begin{align} \left(t_{1},\xi_{1}\right) & =\left(\frac{1}{2}\left(s^{\prime}-2\right)\left(s^{\prime}+1\right),\frac{1}{2}\left(s^{\prime2}-2\right)\left(s^{\prime}-1\right)-1\right)\label{eq:2.15}\\ \end{align}
(12)
\begin{align} \left(t_{2},\xi_{2}\right) & =\left(\frac{s^{\prime}}{2}\left(s^{\prime}+1\right)-1,\frac{1}{2}\left(s^{\prime2}-2\right)\left(s^{\prime}+1\right)\right)\label{eq.2.16} \end{align}
(13)
as can easily be shown as above. These two relations (12) and (13) show that \(\xi_{1}\) and \(\xi_{2}\) are congruent respectively to \(\left(k-1\right)\) and \(0\) modulo \(k\).

There are other cases of interest as shown in the next two Propositions:

Proposition 3. For \(\forall\ \ n\in\mathbb{Z}^{+}\), \(\exists\ \ k,\xi,\mu< k,i,j\in\mathbb{Z}^{+}\), \(k\) non-square, such that if \(\xi_{i}\) are solutions of (1) with \(\xi_{i}\equiv\mu_{j}\left(\text{mod}\,k\right)\), and (R6) if \(k\) is twice a triangular number \(k=n\left(n+1\right)=2T_{n}\), then the set of \(\mu_{j}\) includes \(\left\{ 0,n,\left(n^{2}-1\right),\left(k-1\right)\right\} \), with \(1\leq j\leq\upsilon\).

Proof. Let \(n,k,\xi,\mu< k,i,j\in\mathbb{Z}^{+}\), \(k\) non-square, and \(\xi_{i}\) solutions of (1). Let \(\xi_{i}\equiv\mu_{j}\left(\text{mod}\,k\right)\) with \(1\leq j\leq\upsilon\). As the ratio \(\xi_{i}\left(\xi_{i}+1\right)/k\) must be integer, \(\xi_{i}\left(\xi_{i}+1\right)\equiv0\left(\text{mod}\,k\right)\) or \(\mu_{j}\left(\mu_{j}+1\right)\equiv0\left(\text{mod}\,n\left(n+1\right)\right)\) which is obviously satisfied if \(\mu_{j}=n\) or \(\mu_{j}=\left(n^{2}-1\right)\).

Finally, this last proposition gives a general expression of the congruence \(\xi_{i}\left(\text{mod}\,k\right)\) for most cases to find the remainders \(\mu_{j}\) other than \(0\) and \(\left(k-1\right)\).

Proposition 4. For \(\forall n>1\in\mathbb{Z}^{+}\), \(\exists k,f,\xi,\nu< n< k,\mu< k,m< n,i,j\in\mathbb{Z}^{+}\), \(k\) non-square, let \(\xi_{i}\) be solutions of (1) with \(\xi_{i}\equiv\mu_{j}\left(\text{mod}\,k\right)\), let \(f\) be a factor of \(k\) such that \(f=k/n\) with \(f\equiv\nu\left(\text{mod}\,n\right)\) and \(k\equiv\nu n\left(\text{mod}\,n^{2}\right)\), then the set of \(\mu_{j}\) includes either \(\left\{ 0,mf,\left(\left(n-m\right)f-1\right),\left(k-1\right)\right\} \) or \(\left\{ 0,\left(mf-1\right),\left(n-m\right)f,\left(k-1\right)\right\} \), where \(m\) is an integer multiplier of \(f\) in the congruence relation and such that \(m< n/2\) or \(m< \left(n+1\right)/2\) for \(n\) being even or odd respectively, and \(1\leq j\leq\upsilon\).

Proof. Let \(n>1,k,f,\xi,\mu< k,m< n,i,j< n< k\in\mathbb{Z}^{+}\), \(k\) non-square, and \(\xi_{i}\) a solution of (1). Let \(\xi_{i}\equiv\mu_{j}\left(\text{mod}\,k\right)\) with \(1\leq j\leq\upsilon\). As the ratio \(\xi_{i}\left(\xi_{i}+1\right)/k\) must be integer, \(\xi_{i}\left(\xi_{i}+1\right)\equiv0\left(\text{mod}\,k\right)\) or \(\mu_{j}\left(\mu_{k}+1\right)\equiv0\left(\text{mod}\,fn\right)\). For a proper choice of the factor \(f\) of \(k\), let \(\mu_{j}\) be a multiple of \(f\), \(\mu_{j}=mf\), then \(m\left(mf+1\right)\equiv0\left(\text{mod}\,n\right)\). As \(f\equiv\nu\left(\text{mod}\,n\right)\), one has

\begin{equation} m\left(m\nu+1\right)\equiv0\left(\text{mod}\,n\right).\label{eq:120} \end{equation}
(14)
Let now \(\left(\mu_{j}+1\right)\) be a multiple of \(f\), \(\mu_{j}+1=mf\), then \(m\left(mf-1\right)\equiv0\left(\text{mod}\,n\right)\) or
\begin{equation} m\left(m\nu-1\right)\equiv0\left(\text{mod}\,n\right).\label{eq:121} \end{equation}
(15)

An appropriate combination of integer parameters \(m\) and \(\nu\) guarantees that (14) and (15) are satisfied. Proposition 1 yields the other remainder value as \(mf+\left(n-m\right)f-1=k-1\) and \(\left(mf-1\right)+\left(n-m\right)f=k-1\).

The appropriate combinations of integer parameters \(m\) and \(\nu\) are given in Table 3 for \(2\leq n\leq12\). The sign \(-\) in subscript corresponds to the remainder \(\left(mf-1\right)\); the sign \(/\) indicates an absence of combination.

Table 3. Combination of parameters \(m\) and \(\nu\) for \(2\leq n\leq12\).
\(m\) \(\nu\)
\(\searrow\) 1 2 3 4 5 6 7 8 9 10 11
\(n\) 2 1\_
3 1\_ 1
4 1\_ / 1
5 1\_ 2 2\_ 1
6 1\_ / / / 1
7 1\_ 3 2 2\_ 3\_ 1
8 1\_ / 3\_ / 3 / 1
9 1\_ 4 / 2 2\_ / 4\_ 1
10 1\_ / 3 / 5\_ / 3\_ / 1
11 1\_ 5 4\_ 3\_ 2 2\_ 3 4 5\_ 1
12 1\_ / / / 3 / 4\_ / / / 1

One deduces from Table 3 the following simple rules:

  • 1) \(\forall n\in\mathbb{Z}^{+}\), only those values of \(\nu\) that are co-prime with \(n\) must be kept, all other combinations (indicated by \(/\) in Table 3) must be discarded as they correspond to combinations with smaller values of \(n\) and \(\nu\); for \(n\) even, this means that all even values of \(\nu\) must be discarded. For example, \(\nu=2\) and \(n=4\) are not co-prime and their combination corresponds to \(\nu=1\) and \(n=2\).
  • 2) For \(\nu=1\) and \(\nu=n-1\), all values of \(m\) are \(m=1\) with respectively the remainders \(\left(mf-1\right)\) and \(mf\).
  • 3) For \(\forall n,i\in\mathbb{Z}^{+}\), \(n\) odd, \(2\leq i\leq\left(n-1\right)/2\), and for \(\nu=\left(n-\left(2i-3\right)\right)/2\) and \(\nu=\left(n+\left(2i-3\right)\right)/2\), all the values of \(m\) are \(m=i\).
  • 4) For \(\forall n\in\mathbb{Z}^{+}\), \(n\) odd, and for \(\nu=2\) and \(\nu=n-2\), the remainders are respectively \(mf\) and \(\left(mf-1\right)\).
  • 5) For \(\forall n,i\in\mathbb{Z}^{+}\), \(n\) even, \(2\leq i\leq n/2\), and for \(\nu=\left(n-\left(2i-3\right)\right)/2\) and \(\nu=\left(n+\left(2i-3\right)\right)/2\), all the values of \(m\) are \(m=i\).

Expressions of \(\mu_{i}\) are given in Table 4 for \(2\leq n\leq12\) (with codes E\(n\nu\)). For example, for \(k\equiv12\nu\left(\text{mod}\,12^{2}\right)\) and \(\nu=5\) (code E125), i.e. \(k=60,204,348,...\), \(\xi_{i}\equiv\mu_{j}\left(\text{mod}\,k\right)\) with the set of remainders \(\mu_{j}\) including \(\left\{ 0,mf,\left(\left(n-m\right)f-1\right),\left(k-1\right)\right\} \) with \(m=3\) (see Table 3) and \(f=k/12=5,17,29...\)respectively.

Table 4. Expressions of \(\mu_{j}\) for \(2\leq n\leq12\).
\(n\) \(\nu\) \(m\) \(k\equiv\) \(f\) \(\mu_{j}\) Code
2 1 1 \(2\left(\text{mod}\,4\right)\) \(k/2\) \(0,(k/2)-1,k/2,k-1\) E21
3 1 1 \(3\left(\text{mod}\,9\right)\) \(k/3\) \(0,\left(k/3\right)-1,2k/3,k-1\) E31
2 1 \(6\left(\text{mod}\,9\right)\) \(0,k/3,\left(2k/3\right)-1,k-1\) E32
4 1 1 \(4\left(\text{mod}\,16\right)\) \(k/4\) \(0,\left(k/4\right)-1,3k/4,k-1\) E41
3 1 \(12\left(\text{mod}\,16\right)\) \(0,k/4,\left(3k/4\right)-1,k-1\) E43
5 1 1 \(5\left(\text{mod}\,25\right)\) \(k/5\) \(0,\left(k/5\right)-1,4k/5,k-1\) E51
2 2 \(10\left(\text{mod}\,25\right)\) \(0,2k/5,\left(3k/5\right)-1,k-1\) E52
3 2 \(15\left(\text{mod}\,25\right)\) \(0,\left(2k/5\right)-1,3k/5,k-1\) E53
4 1 \(20\left(\text{mod}\,25\right)\) \(0,k/5,\left(4k/5\right)-1,k-1\) E54
6 1 1 \(6\left(\text{mod}\,36\right)\) \(k/6\) \(0,\left(k/6\right)-1,5k/6,k-1\) E61
5 1 \(30\left(\text{mod}\,36\right)\) \(0,k/6,\left(5k/6\right)-1,k-1\) E65
7 1 1 \(7\left(\text{mod}\,49\right)\) \(k/7\) \(0,\left(k/7\right)-1,6k/7,k-1\) E71
2 2 \(14\left(\text{mod}\,49\right)\) \(0,3k/7,\left(4k/7\right)-1,k-1\) E72
3 3 \(21\left(\text{mod}\,49\right)\) \(0,2k/7,\left(5k/7\right)-1,k-1\) E73
4 3 \(28\left(\text{mod}\,49\right)\) \(0,\left(2k/7\right)-1,5k/7,k-1\) E74
5 2 \(35\left(\text{mod}\,49\right)\) \(0,\left(3k/7\right)-1,4k/7,k-1\) E75
6 1 \(42\left(\text{mod}\,49\right)\) \(0,k/7,\left(6k/7\right)-1,k-1\) E76
8 1 1 \(8\left(\text{mod}\,64\right)\) \(k/8\) \(0,\left(k/8\right)-1,7k/8,k-1\) E81
3 3 \(24\left(\text{mod}\,64\right)\) \(0,\left(3k/8\right)-1,5k/8,k-1\) E83
5 3 \(40\left(\text{mod}\,64\right)\) \(0,3k/8,\left(5k/8\right)-1,k-1\) E85
7 1 \(56\left(\text{mod}\,64\right)\) \(0,k/8,\left(7k/8\right)-1,k-1\) E87
9 1 1 \(9\left(\text{mod}\,81\right)\) \(k/9\) \(0,(k/9)-1,8k/9,k-1\) E91
2 4 \(18\left(\text{mod}\,81\right)\) \(0,4k/9,(5k/9)-1,k-1\) E92
4 2 \(36\left(\text{mod}\,81\right)\) \(0,2k/9,(7k/9)-1,k-1\) E94
5 2 \(45\left(\text{mod}\,81\right)\) \(0,(2k/9)-1,7k/9,k-1\) E95
7 4 \(63\left(\text{mod}\,81\right)\) \(0,(4k/9)-1,5k/9,k-1\) E97
8 1 \(72\left(\text{mod}\,81\right)\) \(0,k/9,(8k/9)-1,k-1\) E98
10 1 1 \(10\left(\text{mod}\,100\right)\) \(k/10\) \(0,(k/10)-1,9k/10,k-1\) E101
3 3 \(30\left(\text{mod}\,100\right)\) \(0,3k/10,(7k/10)-1,k-1\) E103
7 3 \(70\left(\text{mod}\,100\right)\) \(0,(3k/10)-1,7k/10,k-1\) E107
9 1 \(90\left(\text{mod}\,100\right)\) \(0,k/10,(9k/10)-1,k-1\) E109
11 1 1 \(11\left(\text{mod}\,121\right)\) \(k/11\) \(0,(k/11)-1,10k/11,k-1\) E111
2 5 \(22\left(\text{mod}\,121\right)\) \(0,5k/11,(6k/11)-1,k-1\) E112
3 4 \(33\left(\text{mod}\,121\right)\) \(0,(4k/11)-1,7k/11,k-1\) E113
4 3 \(44\left(\text{mod}\,121\right)\) \(0,(3k/11)-1,8k/11,k-1\) E114
5 2 \(55\left(\text{mod}\,121\right)\) \(0,2k/11,(9k/11)-1,k-1\) E115
6 2 \(66\left(\text{mod}\,121\right)\) \(0,(2k/11)-1,9k/11,k-1\) E116
7 3 \(77\left(\text{mod}\,121\right)\) \(0,3k/11,(8k/11)-1,k-1\) E117
8 4 \(88\left(\text{mod}\,121\right)\) \(0,4k/11,(7k/11)-1,k-1\) E118
9 5 \(99\left(\text{mod}\,121\right)\) \(0,(5k/11)-1,6k/11,k-1\) E119
10 1 \(110\left(\text{mod}\,121\right)\) \(0,k/11,(10k/11)-1,k-1\) E1110
12 1 1 \(12\left(\text{mod}\,144\right)\) \(k/12\) \(0,(k/12)-1,11k/12,k-1\) E121
5 3 \(60\left(\text{mod}\,144\right)\) \(0,3k/12,(9k/12)-1,k-1\) E125
7 4 \(84\left(\text{mod}\,144\right)\) \(0,(4k/12)-1,8k/12,k-1\) E127
11 1 \(132\left(\text{mod}\,144\right)\) \(0,k/12,(11k/12)-1,k-1\) E1211
Table 5. Values of \(\mu_{j}\) for \(2\leq k\leq120\).
\(k\) \(\mu_{j}\) References \(k\) \(\mu_{j}\) References
2 0,1 R1,R6,E21 63 0,27,35,62 E72,E97
3 0,2 R1,E31 65 0,64 R3
5 0,4 R1,R3,E51 66 0,11,21,32,33,44,54,65 E21+E31+E65+E116
6 0,2,3,5 R6,E21,E32,E61 67 0,66 R1
7 0,6 R1,R5,E71 68 0,16,51,67 E41
8 0,7 R2,R4,E81 69 0,23,45,68 E32
10 0,4,5,9 E21,E52,E101 70 0,14,20,34,35,49,55,69 E21+E54+E73+E107
11 0,10 R1,E111 71 0,70 R1
12 0,3,8,11 R6,E31,E43,E121 72 0,8,63,71 R6,E81,E98
13 0,12 R1 73 0,72 R1
14 0,6,7,13 E21,E72 74 0,73 ?
15 0,5,9,14 E32,E53 75 0,24,50,74 E31
17 0,16 R1,R3 76 0,19,56,75 E43
18 0,8,9,17 E21,E92 77 0,21,55,76 E74,E117
19 0,18 R1 78 0,12,26,38,39,51,65,77 E21+E32+E61
20 0,4,15,19 R6,E41,E54 79 0,78 R1,R5
21 0,6,14,20 E31,E73 80 0,79 R4
22 0,10,11,21 E21,E112 82 0,40,41,81 E21
23 0,22 R1,R5 83 0,82 R1
24 0,23 R4 84 0,27,56,83 E31,E127
26 0,12,13,25 E21 85 0,34,50,84 E52
27 0,26 R2 86 0,42,43,85 E21
28 0,7,20,27 E43,E74 87 0,29,57,86 E32
29 0,28 R1 88 0,32,55,87 E83,E118
30 0,5,24,29 R6,E51,E65 89 0,88 R1
31 0,30 R1 90 0,9,80,89 R6,E91,E109
32 0,31 R2 91 0,13,77,90 E75
33 0,11,21,32 E32,E113 92 0,23,68,91 E43
34 0,16,17,33 E21 93 0,30,62,92 E31
35 0,14,20,34 E52,E75 94 0,46,47,93 E21
37 0,36 R1,R3 95 0,19,75,94 E54
38 0,18,19,37 E21 96 0,32,63,95 E32
39 0,12,26,38 E31 97 0,96 R1
40 0,15,24,39 E53,E85 98 0,48,49,97 E21
41 0,40 R1 99 0,44,54,98 E92,E119
42 0,6,35,41 R6,E61,E76 101 0,100 R1,R3
43 0,42 R1 102 0,50,51,102 E21
44 0,11,32,43 E43,E114 103 0,102 R1
45 0,9,35,44 E54,E95 104 0,103 ?
46 0,22,23,245 E21 105 0,14,20,35,69,84,90,104 E32+E51+E71
47 0,46 R1,R5 106 0,52,53,105 E21
48 0,47 R4 107 0,106 R1
50 0,24,25,49 E21 108 0,27,80,107 E43
51 0,17,33,50 E32 109 0,108 R1
52 0,12,39,51 E41 110 0,10,99,109 R6,E101,E1110
53 0,52 R1 111 0,36,74,110 E31
54 0,26,27,53 E21 112 0,48,63,111 E72
55 0,10,44,54 E51,E115 113 0,112 R1
56 0,7,48,55 R6,E71,E87 114 0,56,57,113 E21
57 0,18,38,56 E31 115 0,45,69,114 E53
58 0,28,29,57 E21 116 0,28,87,115 E41
59 0,58 R1 117 0,26,90,116 E94
60 0,15,44,59 E43,E125 118 0,58,59,117 E21
61 0,60 R1 119 0,118 R1,R5
62 0,30,31,61 E21 120 0,15,104,119 E87

Values of the remainders \(\mu_{j}\) are given in Table 5 for \(2\leq k\leq120\), with rule (R) and expression (E) codes as references. R and E codes separated by commas imply that all references apply simultaneously to the case; E codes separated by + mean that all expressions apply to the case; some expression references are sometimes missing. One observes that in two cases (for \(k=74\) and 104), expressions could not be found (indicated by question marks).

Table 5 gives correctly the values of the remainder pairs in most of the cases. There are although some exceptions and some values missing.

Among the exceptions to the values given in Table 5, for \(n=2\), remainders values for \(k=30,42,74,90,110,\ldots\) are different from the theoretical ones in Table 4. Furthermore, for \(k=66,70,78,105,...\), additional remainders exist. Expressions are missing for \(k=74\) (E21) and 104 (E85). Finally, one observes also that for 16 cases, some Rules or Expressions supersede some other Expressions (indicated by Ra > Exy or Exy > Ezt), as reported in Table 6. For example, Rule 6 supersedes Expression 21 (R6 > E21) for \(k=30,42,90,110\), i.e., \(k=2T_{5},2T_{6},2T_{9},2T_{10},...\) and more generally for all \(k=2T_{i}\) for \(i\equiv1,2\left(\text{mod}4\right)\).

Table 6. Rules and Expressions superseding other Rules and Expressions.
\(k\)
24 R4 > E32; R4 > E83
30 R6 > E21; R6 > E31; R6 > E103; E51 > E103; E65 > E103
42 R6 > E21; R6 > E32
48 R4 > E31
56 R6 > E43
60 E43 > E32; E43 > E52
65 R3 > E53
72 R6 > E43
80 R4 > E51
84 E31 > E41; E31 > E75
90 R6 > E21; R6 > E53
102 E21 > E31; E21 > E65
110 R6 > E21; R6 > E52
114 E21 > E32; E21 > E61
119 R1 > E73; R5 > E73
120 E87 > R4; E87 > E31; E87 > E54

Note that 11 of these 16 values of \(k\) are multiples of 6, the others are 2 mod 6 and 5 mod 6 for, respectively three and two cases. One notices as well, that generally, Ra and Exy supersede Ezt with \(x< z\) and \(t< y\), except for \(k=60\) and \(120\).

4. Conclusions

We have shown that, for indices \(\xi\) of triangular numbers multiples of other triangular numbers, the remainders in the congruence relations of \(\xi\) modulo \(k\) always come in pairs whose sum always equal \(\left(k-1\right)\), always include 0 and \(\left(k-1\right)\), and only 0 and \(\left(k-1\right)\) if \(k\) is prime, or an odd power of a prime, or an even square plus one or an odd square minus one or minus two. If the multiplier \(k\) is twice a triangular number of \(n\), the set of remainders includes also \(n\) and \(\left(n^{2}-1\right)\) and if \(k\) has integer factors, the set of remainders include multiple of a factor following certain rules. Finally, algebraic expressions are found for remainders in function of \(k\) and its factors. Several exceptions are noticed as well as reported above and it appears that there are superseding rules between the various rules and expressions.

This approach allows eliminating in numerical searches those \(\left(k-\upsilon\right)\) values of \(\xi_{i}\) that are known not to provide solutions of (1), where \(\upsilon\) is the even number of remainders. The gain is typically in the order of \(k/\upsilon\), with \(\upsilon\ll k\) for large values of \(k\).

Conflicts of Interest

The author declares no conflict of interest.

Data Availability

Data required for this research is included within this paper.

Funding Information

The author received no financial support for the research authorship and / or publication of this article.

References

  1. Andrews, G. E. (1971). Number Theory. Dover Publications, New York. [Google Scholor]
  2. Weisstein, E. W. (2021). Triangular Number, in MathWorld-A Wolfram Web Resource. Last accessed 14 February 2021. [Google Scholor]
  3. Cunningham, A. (1901). Mathematical Questions and Solutions in Continuation of the Mathematical Columns of "the Educational Times", Volume 75, F. Hodgson, 87-88.
  4. de Joncourt, E. (1762). The Nature and Notable Use of the Most Simple Trigonal Numbers. The Hague: Husson. [Google Scholor]
  5. Roegel, D. (2013). A reconstruction of Joncourt's table of triangular numbers (1762). LOCOMAT project. Last accessed 14 February 2021. https://locomat.loria.fr/joncourt1762/joncourt1762doc.pdf. [Google Scholor]
  6. Dickson, L. E. (2005). History of the Theory of Numbers, Vol. II: Diophantine Analysis. Dover Publications, New York, p. 587. [Google Scholor]
  7. Chahal, J. S., & D'Souza, H. (1993). Some remarks on Triangular Numbers, in Number Theory with an Emphasis on the Markoff Spectrum, A. D. Pollington and W. Mran, eds., Marcel Dekker Inc., New York.
  8. Breiteig, T. (2015). Quotients of triangular numbers. The Mathematical Gazette, 99(545), 243-255. [Google Scholor]
  9. Pletser, V. (2021). Recurrent relations for multiple of triangular numbers being triangular numbers. ArXiv 2101.00998. Last accessed 14 February 2021. http://arxiv.org/abs/2101.00998. [Google Scholor]
  10. Sloane, N. J. A., (2018). The On-Line Encyclopedia of Integer Sequences. Last accessed 14 February 2021. Published electronically at https://oeis.org. [Google Scholor]
  11. Pletser V. (2021). Searching for multiple of triangular numbers being triangular numbers. Research Gate, DOI: 10.13140/RG.2.2.35428.91527. https://www.researchgate.net/publication/349788977 . [Google Scholor]