THE TRIANGULAR THEOREM OF EIGHT BEN KANE Abstract. We investigate here sums of triangular numbers f (x) := P bi Txi where i Tn is the n-th triangular number. We show that, fixing bi ≥ 0, f (x) represents every (nonnegative) integer if and only if it represents 1, 2, 4, 5, and 8, with the standard P application to sums of odd squares (2xi + 1)2 . Moreover, we show that no finite i subset will suffice if cross terms are included. 1. Introduction In 1638 Fermat wrote that every number is a sum of at most three triangular numbers, four square numbers, and in general n polygonal numbers of order n. Here the triangular , where we include x = 0 for simplicity. The claim for four numbers are Tx := x(x+1) 2 squares was shown by Lagrange in 1772, while Gauss famously wrote “Eureka, 4 + 4 + 4 = n” in his mathematical diary on July 10, 1796. Theorem (Gauss, 1796). Every positive integer is the sum of three triangular numbers. The first proof of the full assertion of Fermat was given by Cauchy in 1813. In 1917, Ramanujan extended the question about four squares to consider which choices of b = (b1 , b2 , b3 , b4 ) have the property that b1 x21 + b2 x22 + b3 x23 + b4 x24 represents every positive integer. We shall refer to such forms as universal diagonal forms. He gives a list of 55 possible choices of b which he then claims are the complete list of universal quarternary diagonal forms (54 forms actually turned out to be universal). In light of Ramanujan’s generalization of Lagrange’s Theorem, we will consider the following generalization of Gauss’s “Eureka” Theorem. We will investigate determining every choice of b such that the sum f (x) := fb (x) := k X bi Txi i=1 represents every integer. We will see the following simple condition to determine whether fb represents every integer. Date: February 5, 2008. 2000 Mathematics Subject Classification. 11E25, 11E20, 11E45. Key words and phrases. Triangular Numbers, Quadratic Forms, Sums of Odd Squares. 1 2 BEN KANE Theorem 1.1. Fix a sequence of integers b1 , . . . , bk . Then k P 1 The triangular form f (x) = bi Txi represents every integer if and only if it i=1 represents the integers 1, 2, 4, 5, and 8. 2 The corresponding diagonal quadratic form Q(x) = k P bi x2i with xi all odd rep- i=1 resents every integer of the form 8n + k X bi i=1 if and only if it represents 8+ k P i=1 bi , 16+ k P bi , 32+ i=1 k P bi , 40+ i=1 k P i=1 bi , and 64+ k P bi . i=1 Recently, Conway and Schneeberger proved a similar condition for positive definite quadratic forms whose corresponding matrix has integer entries, but without publishing their results. Theorem (Conway-Schneeberger). A positive definite quadratic form Q(x) = xt Ax, where A is a positive symmetric matrix with integer coefficients, represents every positive integer if and only if it represents the integers 1, 2, 3, 5, 6, 7, 10, 14, and 15. Bhargava gave a simpler proof of the Conway-Schneeberger 15 theorem in [1], in addition to showing more generally that for any set S it is always sufficient to check a finite subset S0 , and showed the set S0 for S all odd integers and S all primes. More recently, Bhargava and Hanke have shown the 290 theorem, stating the necessary set when the corresponding matrix is half integral, the largest of which is 290 [2]. For a more complete history of related questions about sums of figurate numbers and some new results, please see [4]. For further information and background about quadratic forms, a good source is [5]. There is a natural mapping from representations by sums of triangular numbers to quadratic forms, namely the mapping which takes xi to 2xi + 1 after multiplying by 8. In light of the Conway-Schneeberger 15 theorem and the Bhargava-Hanke 290 theorem, one next considers whether a similar result holds for the preimage of the above map. This preimage is unique up to a constant, so we will first make a cannonical mapping which will allow us to define a certain metric and then later add this metric to our function to obtain a form for which we can address an analogous question to the ConwaySchneeberger 15 theorem. To this end, we define the canonical preimage of a positive definite quadratic form Q whose corresponding matrix A has even non-diagonal coefficients as f˜(x) := f˜Q (x) := k X i=1 bi Txi + X i<j,cij ≥0 cij (2xi xj + xi + xj ) + X i<j,cij <0 cij (2xi xj + xi + xj + 1), THE TRIANGULAR THEOREM OF EIGHT where here Tx = x(x+1) 2 3 extends to x negative and Q will be defined below. Remark 1.2. Our choice of adding the constant cij every time cij < 0 may not seem canonical at first, but notice that if Q0 is the equivalent quadratic form obtained by replacing x1 with −x1 , then we find that this choice leads to f˜Q = f˜Q0 . Note that f˜(x) represents the integer n if and only if the quadratic form k X Q(x) := bi x2i + i=1 represents the integer 8n + k P i=1 bi + P X 4cij xi xj i<j 4|cij | with xi all odd. i<j However, note that if we have Q̃ and Q̃0 are two equivalent quadratic forms with the mapping between them preserving xi odd (we will refer to such forms as equivalently odd, then it is not necessarily the case that f˜Q = f˜Q0 , and that f˜ may represent (finitely many) negative integers. Thus we define the natural metric m̃ ˜ := | min f˜(x)| f x and define (1.1) f (x) := f[Q]o (x) := f˜(x) + m̃f˜, where [Q]o denotes the equivalently odd equivalence class of Q. Clearly, f is independent of the choice of Q0 in the equivalently odd equivalence class of Q, and f only represents nonnegative integers. This brings one to consider whether a condition similar to the 15 theorem or the 290 theorem holds for determining whether fQ represents every positive integer. However, no such subset exists. Theorem 1.3. For a set S, there is no proper subset S0 such that it holds for every Q with even non-diagonal coefficients that fQ represents S0 if and only if fQ represents S. In particular, there is no proper subset of N. Our proof is by construction of forms f which represent every integer other than any arbitrary integer n. Remark 1.4. The corresponding statement about the quadratic form Q is that there is no finite subset to determine if Q represents every integer from its minimum value which is congruent modulo 8 to the minimum value. Due to the congruence conditions modulo 8 implied by having all xi odd, these are the only possible values that Q may represent. One may consider the question of whether the form Q represents every integer congruent to 1 mod 8, or 2 mod 8, etc., with xi odd. Using a simple mapping to forms with x1 odd and xi even otherwise, one can continue with escalator lattices as in Bhargava’s work, escalating after the first dimension only to choices of vectors which would occur 4 BEN KANE with xi even. There should only be finitely many escalators in the resulting tree, and hence there would be a finite subset in this case. We will not work out the details of this observation here, however. However, under a certain restriction, we will find again that a finite subset will suffice. Consider f˜ equivalent to f˜0 , written f˜ ∼ f˜0 , if the corresponding f = f 0 . We will denote the equivalence class by [f ]. Then one can define the following metric, depending only on the equivalence class [f ], or, correspondingly, the equivalently odd equivalence class [Q]o , m[Q]o := m[f ] := min m̃f˜. f˜∈[f ] Then we will be able to find a finite subset of the positive integers will suffice when m[f ] is bounded. Theorem 1.5. Fix an integer m and a subset S of the positive integers. Then there is a finite subset S0,m ⊂ S depending only on m and S such that if m[f ] ≤ m then f represents S if and only if f represents S0,m . Moreover, for S = N, we find that N0,m m2 . It may be of interest to investigate the growth of N0,m in terms of m. Acknowledgements 2. Theorem of Eight Here we prove Theorem 1.1. We will proceed by showing by a standard argument that the theorem is equivalent to a statement about (diagonal) quadratic forms, and then prove the corresponding result for quadratic forms. We will only need some elementary results about quadratic forms and a theorem of Siegel to show the desired result. Proof. Consider the generating function F (q) := Fb (q) := X q f (x) = x ∞ X t(n)q n , n=0 where t(n) is the number of solutions to f (x) = n. Then we see that k P q i=1 bi 8 F (q ) = X k P q i=1 bi (2xi +1)2 , x so that t(n) = ro (8n − k P bi ), where ro (n) is the number of representations of n by the i=1 corresponding (diagonal) quadratic form with xi odd. We proceed as with escalator lattices in [1]. Without loss of generality, b1 ≤ b2 ≤ · · · ≤ bk . Fixing b = [b1 , . . . , bk−1 ], we will escalate to [b1 , . . . , bk ] by choosing all possible choices of bk ≥ bk−1 for which it is possible to represent the next largest integer not already represented. We will then THE TRIANGULAR THEOREM OF EIGHT 5 develop an escalator tree by forming an edge between b and [b1 , . . . , bk ], with ∅ as the root. If b represents every integer, then b will be a leaf of our tree. Since t(1) > 0, it follows that b1 = 1. We need t(2) > 0, so b2 = 1 or b2 = 2. If b2 = 1, then we need t(5) > 0, so 1 ≤ b3 ≤ 5. For b3 = 3, we need t(8) > 0, so 3 ≤ b4 ≤ 8. Likewise, if b2 = 2, then 2 ≤ b3 ≤ 4. Therefore, if t(n) > 0 for every n, then we must have one of the above choices of bi as a sublattice. By showing that each of these choices of bi satisfies t(n) > 0 for every n, we will see that this condition is both necessary and sufficient. For ease of notation, we will denote the triangular form corresponding to b with [b1 , b2 , . . . , bk ] and the corresponding quadratic form by (b1 , . . . , bk ). For the forms [1, 1, 1], [1, 1, 4], [1, 1, 5], [1, 2, 2], and [1, 2, 4], ro (n) = r(n), where r(n) is the number of representations of n without the restriction of xi odd. For each of these choices of b, (b1 , b2 , b3 ) is a Genus 1 quadratic form. Therefore, n is represented globally if and only if n is represented locally, and for these choices of b, a quick check of the local density shows that each integer 8n + b1 + b2 + b3 is locally represented for every n. Our proofs for [1, 1, 2], [1, 2, 3], and [1, 1, 3] will be essentially the same. For [1, 1, 2], we note that if x2 + y 2 + 2z 2 = 8n + 4 has a solution with x, y, and z not all odd, then taking each side modulo 8 leads us to the conclusion that x, y, and z must all be even. Therefore, the solutions without x, y, and z odd correspond to solutions of 4x2 + 4y 2 + 8z 2 = 8n + 4, or, x2 + y 2 + 2z 2 = 2n + 1. Using Siegel’s theorem to compare the local density at 2, we see that the average of the number of representations over the genus is twice as large for 8n + 4 as 2n + 1. However, (1, 1, 2) is again a Genus 1 quadratic form, so r(8n + 4) = 2r(2n + 1), and hence t(n) = ro (8n + 4) = r(8n + 4) − r(2n + 1) = r(2n + 1). Since (1, 1, 2) is Genus 1, we need again only check local conditions. Similar arguments show that t[1,2,3] (n) = ro,(1,2,3) (8n + 6) = r(1,2,3) (8n + 6) − r(4,2,12) (8n + 6) = r(1,2,3) (8n + 6) − r(1,2,6) (4n + 3) = r(1,2,6) (4n + 3). Again (1, 2, 6) is Genus 1, so checking local conditions shows that every integer n is represented. For [1, 1, 3] we see analogously that t[1,1,3] (n) = r0,(1,1,3) = r(1,1,3) − r(1,1,12) = r(1,1,12) , and again (1, 1, 12) is Genus 1. However, in this case, local conditions indicate that integers congruent to 8 modulo 9 are note represented by [1, 1, 3]. Considering [1, 1, 3, k] for k ∈ 3..8, we see that for n ≡ 8 mod 9, taking the fourth variable to be 1 will give 6 BEN KANE us n − k 6≡ 8 mod 9, and hence n − k is represented by [1, 1, 3], so that n is represented by [1, 1, 3, k]. Having seen that each of our choices of b is indeed a leaf to the tree, we conclude that the representing the integers 1, 2, 4, 5, and 8 suffices. 3. Cross Terms We will now consider the “triangular forms” f defined by equation (1.1) corresponding to quadratic forms whose corresponding matrix has all non-diagonal entries even. We shall begin by showing that no finite subset as in section 2 will suffice for forms of this type. Proof of Theorem 1.3. Let a set S ⊆ N be given. We will proceed by explicit construction of a triangular form f which represents exactly every integer other than an arbitrary fixed m ∈ S, so that no checking no proper subset of S will suffice. First note that if the smallest positive integer represented by a form f is m, then since the sum of three triangular numbers represents every integer, we can add m + 1 times the sum of three triangular numbers plus m + 2 times the sum of three triangular numbers so that there is a form which represents exactly every integer other than m. Consider the quadratic form QN (x, y) := N x2 + N y 2 + 4xy, and denote to the corresponding triangular form as fN . We first show that it is sufficient to show that the generating function for fN is 2 + 2q + O(q N −6 ). (3.1) Assuming equation (3.1), then the generating function for f := m M fN i=1 is m m m 2 (1 + q + ··· + q ) + O(q N −6 ), 1 m since m[f ] is additive across independent sums. If we choose N large enough, then the first integer not represented by f is m + 1. Therefore, since m was arbitrary, there is a form which represents every integer other than m + 1 (this also sufficies for m = 0, by the argument above). We now show that the generating function satisfies (3.1). Since m fN = N Tx + N Ty + (2xy + x + y) + 1, taking x = 0, y = −1 or x = −1, y = 0 represents 0, and x = 0, y = 0 or x = −1, y = −1 represents 1. Now, without loss of generality assume that |x| ≥ |y| and one of THE TRIANGULAR THEOREM OF EIGHT 7 them is not 0 or −1. Then |2xy + x + y| ≤ 2|x|2 + 2|x| = 4T|x| , so that fN ≥ N Tx − 4T|x| + N Ty . When x ≤ −2 it is easy to check that N Tx − 4T|x| ≥ (N − 6)T|x|−1 ≥ N − 6. and when x > 0 N Tx − 4T|x| ≥ (N − 4)Tx ≥ N − 4, since |x| ≥ |y| and one of them is not 0 or −1, so that the terms Tx ≥ 1 or T|x|−1 are positive. Since Ty ≥ 0, or assertion is verified. Remark 3.1. It is important here to note how the above counterexamples differ from the proof when we only have diagonal terms, since this observation will lead us to the following proof when m[f ] is bounded. If, at each step of the escalator lattice we escalate by what will here be referred to as a block, a form whose corresponding quadratic form has an irreducible matrix, which is independent of the other variables, then, after the depth is at least 4, we can use known bounds for modular forms to show that the depth from a given branch off of this depth 4 escalator will be finite. In Section 2, the breadth each time we escalate will be finite, so that the overall tree is finite, and hence there is a largest truant, so there is a finite subset S0 which will suffice. In the above proof, however, there are infinitely many inequivalent forms f which represent the integer 1, so that the breadth is infinite. Proof of Theorem 1.5. Fix a positive integer m. We will start with a small overview of the proof. As in the above remark, we will escalate with blocks. We will first show that when m[f ] ≤ m, the number of blocks that are not dimension 1 in any branch of the escalator tree is bounded, and that there are only finitely many choices for the cross terms c. We will then proceed by defining N (M1 , M2 , . . . , Mk , c) to be the smallest integer not represented by the form corresponding to f˜(x) := k X i=1 Mi Txi + X cij (2xi xj + xi + xj ) + i<j,cij ≥0 X cij (2xi xj + xi + xj + 1). i<j,cij <0 Our claim is then equivalent to showing that in the escalator tree sup N (M1 , M2 , . . . , Mk , c) M1 ,...Mk ,c is finite. To do so, we will effectively show that with c fixed, the supremum with Mi sufficiently large is finite and independent of the choice of Mi , and then fix M1 ≤ m1 , and again show that the resulting supremum is independent of M2 , . . . , Mk , and so forth. 8 BEN KANE Since there are only finitely many such choices of c, the result comes from taking the maximum of each of these supremums. We begin with a lemma that will show that there are only finitely many choices of the cross terms. Lemma 3.2. If m[f ] ≤ m, then there are only finitely many choices of the cross terms cij , up to equivalent forms. Proof. First note that m[f1 ⊕f2 ] = m[f1 ] +m[f2 ] , so that we can only have at most m blocks. It therefore sufficies to show that each block has bounded dimension and bounded entries. Fix the cross terms c of a block f˜ with dimension k such that m̃f˜ = m[f ] , namely a minimal element. We will recursively show a particular choice of xi such that f˜(x) ≤ − max{max |cij |, k − 1}, i,j so that the max of the cij is bounded by m, and the dimension is bounded by m + 1. First set x1 = 0. Since f˜ is a block, then we know at step j that there is some i < j such that cij 6= 0. Choose i < j such that |cij | is maximal. If xi = 0, then we set xj = −1 if cij > 0 and xj = 0 otherwise. If xi = −1 then we set xj = 0 if cij > 0 and xj = −1 otherwise. Since all of our choices of xi are 0 or −1, we note that the integer represented is independent of Mi , since T−1 = T0 = 0. Now we note that if xi = xj = 0 or xi = xj = −1 then 2xi xj + xi + xj = 0 and 2xi xj + xi + xj = −1 otherwise. Therefore, if xi = xj , then from our definition of f˜, the cross term corresponding to cij adds 0 if cij ≥ 0 and adds −|cij | otherwise. If xi = 0 and xj = −1, then the cross term adds −|cij | if cij ≥ 0 and adds 0 otherwise. Therefore by our construction above, we know that for |cij | maximal, we have added −|cij | to our sum, and we never add a positive integer, so the sum is at most −|cij |. Moreover, since the block is connected, we have added at most −1 at each inductive step, so that the sum is at most −(k − 1). For simplicity, in our escalator tree, we will “push” up all of the blocks tothe top of the tree which are not dimension 1. To do so, we will first build the tree with all possible choices of blocks which are not dimension 1, and then escalate with only dimension 1 blocks from each of the nodes of the tree, including the root (the empty set). Thus, every possible form will show up in our representation. This tree is depth at most m in the number of blocks, but is of infinite breadth. Henceforth, we can consider c to be fixed, and take the maximum over all choices of c. We will now see that the subtree from each fixed node is of finite depth. Consider the corresponding quadratic form Q. First note that the generating function for Q when all xi are odd is the generating function for Q minus the generating functino with some xi even, and the others arbitrary, which is simply another quadratic form without any restrictions, taking xi → 2xi . Thus, we have the generating function of a difference of finitely many quadratic forms, and hence we have the Fourier expansion of a modular form. Now we simply note that any quadratic form can be decomposed into an Eisenstein THE TRIANGULAR THEOREM OF EIGHT 9 series and a cusp form (cf. [6]). Using the bounds of Tartakowsky [8] and Deligne [3], as long as the Eisenstein series is non-zero, the growth of the coefficients of the Eisenstein series can be shown to grow more quickly than the coefficients of the cusp form whenever the dimension is greater than or equal to 4, other than finitely many congruences classes for which the coefficients of both the Eisenstein series and the cusp form are zero. Therefore, as long as the Eisenstein series is non-zero, there are only finitely many congruence classes and finitely many “sporadic” integers which are not represented by the quadratic form. Thus, after dimension 4, there are only finitely many congruence classes and finitely many sporadic integers not represented by the form f . If at any step of the escalation, any of the integers in these congruence classes is represented, then we have less congruence classes, and only finitely many more sporadic integers which are not represented, so that the resulting depth is bounded. For the dimension 1 blocks, it is clear that the breadth of each escalation is finite, so there are only finitely many escalators coming from this node. Therefore, it suffices to show that the Eisenstein series is non-zero. Again using Siegel’s theorem [7], the Eisenstein series is simply a difference of the local densities. At every prime other than p = 2, the local densities are equal, so we only need to show that the difference of the local densities at p = 2 is positive. However, the difference of the number of local representations at a fixed 2 power must be positive, since the integer is locally represented with xi odd, except for finitely many congruence classes. Therefore, we can define Ñ (M1 , . . . , Mk , c) to be the maximum of N (M1 , . . . , Mk , Mk+1 , . . . Ml , c), where Mk+1 to Ml are the dimension 1 blocks coming from this node. We will show that Ñ (M1 , . . . , Mk , c) is independent of the choice of Mi whenever Mi is sufficiently large by showing that the resulting subtrees are identical. We need the following lemma to obtain this goal. We will need some notation before we proceed. For a set T , define X q T := qt. t∈T For fixed sets S, T ⊆ N, we will say that a form f (x) := k P bi Txi represents S/T if the i=1 generating function g(x) for f (x) satisfies q T g(x) having for each s ∈ S, the coefficient of q s is positive. Lemma 3.3. Let a (diagonal) triangular form f be given. Fix S, T1 , T2 ⊆ N and M ∈ N such that minn∈T2 ≥ M . Define T := T1 ∪T2 . Then there exists a bound MT1 and a finite subset S0 ⊆ S, depending only on T1 and S such that if M > MT1 ,S , then f represents S/T if and only if f represents S0 /T1 . 10 BEN KANE Proof. We will escalate as in [1] with a slight deviation. At each escalation node, there is a least element s ∈ S such that S/T1 is not represented by the form f corresponding to this node. As in [1], we shall refer to s as the truant of f . To represent {s}/T1 , we must have some t1 ∈ T1 such that s − t1 is represented by f + bTx . Therefore, for each t1 < s we escalate with finitely many choices of b, and there are only finitely many choices of t1 . Thus, the breadth at each escalation is finite, and our argument above using modular forms shows that the depth is also finite, so there are only finitely many choices of s ∈ S which are truants in the escalation tree. Take S0 to be the set of truants in the escalation tree and define MT1 ,S := max s ∈ S0 s + 1. The argument above shows that representing S/T1 is equivalent to representing S0 /T1 . When following the above process with T instead of T1 whenever M > MT1 ,S , we will have the same truants at each step, so that representing S/T is equivalent to representing S/T1 , and hence representing S/T is equivalent to representing S0 /T1 . Remark 3.4. It is of interest to note that if we replace “(diagonal) triangular form” with “quadratic form” (without the odd condition), then the proof follows verbatim, since the breadth is also finite, so that this can be considered a generalization of Bhargava’s result that there is always a finite subset S0 of S such that the quadratic form represents S if and only if it represents S, since this is obtained by taking T1 = T = {0}. Now consider Xj := {x : xi arbitrary for i ≤ j, xi ∈ {0, −1} otherwise} and define T1,j := {f (x) : x ∈ Xj } and T2,j := {f (x) : x ∈ / Xj } and We will use Lemma 3.3 with T1 = T1,j and T2 = T2,j . To use the lemma effectively, we will show the following lemma. (i) Lemma 3.5. There exist bounds MXj depending only on M1 , . . . , Mj , c such that if (i) Mi ≥ mXj for every i > j, then the smallest element of T2,j is greater than MT1 ,N , where MT1 ,N is as defined in lemma 3.3. Proof. We will proceed by induction. For j = 0, we will take X (i) MX0 = MT1 ,N + 6 |cij |. j Noting that cij (2xi xj + xi + xj ) ≥ −|cij |(2T|xi | + 2T|xj | ) Again, our bounds for T|xi | in terms of T|xi | give the desired result, noting that xi ∈ / {0, −1}. We now continue by induction on j. For the corresponding quadratic form, we note P − that plugging in x1 = c1j j 2M1 gives the minimal value over the reals, and a quadratic THE TRIANGULAR THEOREM OF EIGHT 11 form Q0 with rational coefficients and denominator dividing by 2M1 . We therefore can consider Q̃ := 4M1 · Q0 , which is a quadratic form of the desired type. Thus, we can use the inductive step for Q̃. But this gives a bound which minimizes Q̃, and hence Q0 , but our choice of x1 must give a value greater than or equal to this, so the result follows. Now, by our choice of Xj , T1,j is independent of Mi for i > j, since Txi = 0. Thus, (i) fix c and take Mi ≥ MX0 . Then we have the corresponding subtrees identical, so that (1) sup Ñ (M1 , . . . , Mk , c) is this unique element. We may now fix M1 ≤ MX0 , since there are only finitely many such choices. With this M1 fixed, we define T1,1 as above, and again find bounds for the other Mi . Continuing recursively gives the desired result, since we know that k ≤ m, so there are only finitely many supremums that we take. We finally would like to show that NN,m m2 . To do so, we consider again the Lm construction of our counterexamples. Consider f (x) := i=1 fN ⊕ Tx . For N sufficiently large, the smallest integer not represented by f is clearly Tm+1 − 1 m2 . References [1] M. Bhargava, On the Conway-Schneeberger fifteen theorem, Quadratic Forms and their applications (Dublin, 1999), 27-37, Contemp. Math 272, Amer. Math. Soc., Providence, RI, 2000. [2] M. Bhargava, J. Hanke, Universal Quadratic Forms and the 290-Theorem, preprint. [3] P. Deligne, La Conjecture de Weil I, Inst. Hautes Études Sci. Publ. Math., 43, 273-307, 1974. [4] B. Duke, Some Old Problems and New Results about Quadratic Forms, Notices Amer. Math. Soc. 44, no. 2, 190–196, 1997. [5] B. Jones, The Regularity of a Genus of Postiive Ternary Quadratic Forms, Trans. Amer. Math. Soc., 111-124, 1931. [6] K. Ono, Web of Modularity: Arithmetic of the Coefficients of Modular Forms and Q-series, Amer. Math. Society, 2003. [7] C. Siegel, ber die analytische Theorie der quadratischen Formen. Ann. Math. 36, 527-606, 1935. [8] W. Tartakowsky, Die Gesamtheit der Zahlen, die durch eine quadratische form F (x1 , x2 , . . . xs ), (s ≥ 4) darstellbar sind., Izv. Akad. Nauk SSSR, 7, 111-122, 165-196, 1929. Department of Mathematics, Radboud Universiteit, Toernooiveld 1, 6525 Nijmegen, Netherlands E-mail address: [email protected]

© Copyright 2021 Paperzz