Abstract
We employ Ramanujan's formula to prove three conjectures of R. S. Melham on representation of an integer as sums of polygonal numbers.
1. Introduction
Jacobi’s classical two-square theorem is as follows.
Theorem 1 (see [1]). Let denote the number of representations of as a sum of two squares, counting order and sign, and let denote the number of positive divisors of congruent to modulo . Then
The above theorem can also be recasted in terms of Lambert series as Similar representation theorems involving squares and triangular numbers were found by Dirichlet [2], Lorenz [3], Legendre [4], and Berndt [5]. For example, the following two theorems are due to Lorenz and Ramanujan, respectively.
Theorem 2 (see [3]). Let denote the number of representations of as a sum of times a square and times a square. Then
Theorem 3 (see [5]). Let denote the number of representations of as a sum of times a triangular number and times a triangular number. Then
Hirschhorn [6, 7] obtained forty-five similar identities (including those obtained by Legendre and Ramanujan) involving squares, triangular numbers, pentagonal numbers, and octagonal numbers employing dissection of the -series representations of the identities obtained by Jacobi, Dirichlet, and Lorenz. In [8], Baruah and the author obtained twenty-five more such identities involving squares, triangular numbers, pentagonal numbers, heptagonal numbers, octagonal numbers, decagonal numbers, hendecagonal numbers, dodecagonal numbers, and octadecagonal numbers. More works on this topic have been done in [9–11]. In [11], Melham presented 21 conjectured analogues of Jacobi’s two-square theorem which are verified using computer algorithms. In [12], Toh offered a uniform approach to prove these conjectures using known formulae for . In this paper, we show that some of these conjectures can also be proved by using Ramanujan’s famous formula. We prove three conjectures enlisted in the following theorem which have appeared as , , and , respectively, in [11].
Theorem 4. Consider
The next section of this paper is devoted to notations, definitions, and preliminary results.
2. Notations and Preliminary Results
Ramanujan’s general theta-function is defined by [5, page 34, (18.1)] Jacobi’s famous triple product identity takes the simple form [5, page 35, Entry 19] where, here and throughout the paper, for ,
We also use the following notation for the sake of brevity of expressions: One important special case of is where the product representations in (12) arise from (9).
From [5, page 46, Entry 30] we find that Also if , then [5, page 45, Entry 29]
We recall Ramanujan’s formula in the form [5, page 34, (17.6)] where
Finally, we note that
3. Proof of the Conjectures (5)–(7)
Proof of Conjecture (5). In view of (17), we rewrite (5) as
Replacing by and then setting , , and in (15), we have
Expanding the -products on the left side of (19) and noting that , we obtain
Also
Thus, from (20) and (21), we have
Replacing by and then setting , , and in (15), and proceeding as in case of (22), we find that
Multiplying (23) by and adding to (22), we obtain
Setting and in (13), we find that
Thus, (24) reduces to
Now
We have
Similarly
Hence from (27), (28), and (29) we have
Using (30) in (26), we find that
This completes the proof of (18).
Proof of Conjecture (6). In view of (17), we rewrite (6) as
Replacing by and then setting , , and in (15), and proceeding as in case of (22), we find that
Replacing by and then setting , , and in (15), and proceeding in a way similar to obtaining (22), we have
Multiplying (34) by and then adding to (33), we find that
Setting , , , and in (14), we find that
Using (36) in (35), we have
Now
Using (38) in (37), we have
This completes the proof of (32).
Proof of Conjecture (7). In view of (17), we rewrite (7) as
Replacing by and then setting , , and in (15), and proceeding as in case of (22), we find that
Also, replacing by and then setting , , and in (15), and proceeding similar to (22), we obtain
Multiplying (42) by and then adding to (41), we find that
Setting , , , and in (14), we find that
Hence, (43) reduces to
Now
Using (46) in (45), we have
This completes the proof of (40).
4. Concluding Remarks
All conjectures in [11] can easily be reformulated as theta function identity using Ramanujan’s formula. However, these identities might be too complicated to actually have a proof.
Conflict of Interests
The author declares that there is no conflict of interests regarding the publication of this paper.