Ramanujan J (2008) 17: 145–153 DOI 10.1007/s11139-007-9023-y
Extensions of representations of integral quadratic forms Wai Kiu Chan · Byeong Moon Kim · Myung-Hwan Kim · Byeong-Kweon Oh
Received: 5 April 2006 / Accepted: 14 March 2007 / Published online: 18 September 2007 © Springer Science+Business Media, LLC 2007
Abstract Let N and M be quadratic Z-lattices, and K be a sublattice of N. A representation σ : K → M is said to be extensible to N if there exists a representation ρ : N → M such that ρ|K = σ . We prove in this paper a local–global principle for extensibility of representation, which is a generalization of the main theorems on representations by positive definite Z-lattices by Hsia, Kitaoka and Kneser (J. Reine Angew. Math. 301:132–141, 1978) and Jöchner and Kitaoka (J. Number Theory 48:88–101, 1994). Applications to almost n-universal lattices and systems of quadratic equations with linear conditions are discussed. Keywords Extension of representations · Integral quadratic forms Mathematics Subject Classification (2000) 11E12 · 11E20
Research of the first author was partially supported by the National Science Foundation. The third author was partially supported by KRF Research Fund (2003-070-C00001). W.K. Chan Department of Mathematics and Computer Science, Wesleyan University, Middletown, CT 06459, USA e-mail:
[email protected] B.M. Kim Department of Mathematics, Kangnung National University, Kangwondo 210-702, Korea e-mail:
[email protected] M.-H. Kim Department of Mathematical Science, Seoul National University, Seoul 151-747, Korea e-mail:
[email protected] B.-K. Oh () Department of Applied Mathematics, Sejong University, Seoul 143-747, Korea e-mail:
[email protected]
146
K.W. Chan et al.
1 Introduction Let f (x) = m i,j =1 aij xi xj be a quadratic form in m variables x1 , . . . , xm , where (aij ) is an m × m symmetric matrix over Q. We say that f represents a quadratic form g(y) = ni,j =1 bij yi yj in n variables y1 , . . . , yn over Z if there exist b1 , . . . , bn ∈ Zm such that f (y1 b1 + · · · + yn bn ) = g(y). The representation problem asks for a complete determination of the set of forms that are represented by f . This is equivalent to deciding when the matrix equation (which can be viewed as a system of Diophantine equations) t
T (aij )T = (bij )
has a solution T ∈ Mm×n (Z). For the recent development of this problem, we refer the readers to the recent surveys [3] and [12] and the references cited therein. Using the geometric language of quadratic spaces and lattices (see [9, 11]), the representation problem of quadratic forms can be rephrased as follows. The equivalence class of the quadratic form f (x) corresponds to the isometry class of a Z-lattice M with a basis {v1 , . . . , vm } and a symmetric bilinear form B such that B(vi , vj ) = aij . A Z-lattice N of rank n with bilinear form B is said to be represented by M, written N → M, if there exists a Z-linear map σ : N → M such that B(σ (x), σ (y)) = B (x, y) for all x, y ∈ N . This map σ is called a representation of N by M. If g(y) is the quadratic form corresponding to N , then N → M if and only if g(y) is represented by f (x). Therefore, the representation problem of quadratic forms is equivalent to deciding when a Z-lattice is represented by another Z-lattice. Throughout this paper, we shall conduct our discussion in the above geometric language. Any unexplained terminologies and notations can be found in [9] and [11]. The term lattice always refers to a Z-lattice on a non-degenerate quadratic space with a bilinear form B and its associated quadratic map Q. Note that under this condition every representation of a lattice by another lattice must be injective. Let p be a prime. For a lattice M, its p-adic completion Mp is the Zp -lattice Zp ⊗ M, where Zp is the ring of p-adic integers. The genus of M, denoted gen(M), is the set of all lattices L on the space QM such that Lp ∼ = Mp for all primes p. A necessary condition for N → M is that Np → Mp for all p and R ⊗ N → R ⊗ M, and we say that N is represented by gen(M). If this is the case, then Hasse-Minkowski’s local– global principle asserts that there will be a representation of the space QN by QM. Therefore, when discussing the representation problem, there is no harm to assume that N is a lattice in the space QM. Furthermore, Witt’s theorem implies that any representation of N by M is the restriction of an isometry of QM. Let K be a sublattice of N and σ : K → M be a representation. We say that a representation ρ : N → M is an extension of σ if ρ|K = σ . This definition has its analog for Zp -lattices. In the next section, we shall prove a result concerning extension of representation which can be viewed as a generalization of the main representation theorems for positive definite lattices proved by Hsia, Kitaoka and Kneser [4] and by Jöchner and Kitaoka [7]. In Sects. 3 and 4 we shall apply our results to study almost 2-universal lattices and quadratic equations with linear conditions.
Extensions of representations of integral quadratic forms
147
We close this section with some additional notations and terminologies which will be used throughout this paper. If M is a positive definite lattice of rank m, for any i = 1, . . . , m, the i-th successive minimum of M is denoted by μi (M). By reduction theory ([2], Chap. 12) there exists a constant α, depending only on m, such that d(M) > αμ1 (M) · · · μm (M). Let R be Z or Zp . For a basis {v1 , . . . , vn } of a R-lattice M, we write M∼ = (B(vi , vj )). The matrix on the right hand side is called a matrix presentation of M, which we often identify with M itself. If M admits an orthogonal basis {v1 , . . . , vn }, we call M diagonal and simply write M∼ = Q(v1 ), . . . , Q(vn ). The determinant of the matrix (B(xi , xj )) is called the discriminant of M, denoted d(M). Unless stated otherwise, every R-lattice M is assumed to be integral in the sense that B(M, M) ⊆ R. A representation σ : N → M of R-lattices is said to be primitive if σ (N) is a direct summand of M as R-modules. If φ and ψ are representations from N to M and a ∈ R, we write φ ≡ ψ mod aM if φ(x) − ψ(x) ∈ aM for all x ∈ N .
2 Extensibility of representations Throughout this section, K, N and M are lattices of rank k, n and m, respectively, in the quadratic space V = QM. Let U be the space QK and W be the orthogonal complement of U in V . Let π be the orthogonal projection from V onto W . Theorem 2.1 Let σ : K→M be a representation. Suppose that m ≥ k + 2(n − k) + 3. Let s be a positive integer and T be a finite set of primes which contains all primes that divide 2d(M)d(K). If the orthogonal complement of σ (K) in M is positive definite, then there exists a constant C1 = C1 (K, M, n, T , s) with the following property: Suppose that for each prime p, there exists a representation ρ(p) : Np →Mp which extends σ . If μ1 (π(N )) > C1 , then there exists a representation ρ : N →M which extends σ and ρ ≡ ρ(p) mod p s Mp for all p ∈ T . Proof Suppose that the theorem holds for the special case when K ⊆ N ∩ M and σ is the inclusion map. Now, K ⊆ N ∩ σ −1 (M) and for each prime p, the restriction of the representation σ −1 ρ(p) : Np →σ −1 (M)p to Kp is the inclusion map. Let C1 be the constant for K, σ −1 (M), n, T and s. If μ1 (π(N )) > C1 , then there exists a representation γ : N →σ −1 (M), which is an extension of the inclusion K → σ −1 (M), such that γ ≡ σ −1 ρ(p) mod p s σ −1 (Mp ) for all p ∈ T . Let ρ be the isometry σ γ . Then ρ|K = σ and hence ρ extends σ . Moreover, ρ ≡ ρ(p) mod p s Mp for all p ∈ T . So, from now on we assume that σ : K→M is the inclusion map. We may enlarge T to contain all the primes dividing d(π(M)). Then select a positive integer a such that p a π(Mp ) ⊆ p s Mp for all p ∈ T . Pick a prime q outside T and a positive
148
K.W. Chan et al.
integer e. Let C1 be the constant C1 (m, n, π(M), T , a, q, e) obtained from the main theorem in [7], p. 96. The prime q and the integer e will not play any role in the subsequent discussion; they are chosen so that the constant C1 can be obtained. Note that C1 does not depend on N . For each p ∈ T , write ρ(p) = 1Up ⊥ τ(p) for some τ(p) ∈ O(Wp ). Now suppose that μ1 (π(N )) > C. Since rank(π(M)) = m − k ≥ 2(n − k) + 3 = 2 rank(π(N )) + 3, there exists an isometry γ on W such that γ : π(N)→π(M), and that γ ≡ τ(p) mod p a π(Mp ) for all p ∈ T . Let ρ be the isometry 1U ⊥ γ ∈ O(V ). Let x be any vector in N . If p ∈ T , then Qp Kp ∩ Np = Kp and π(Mp ) ⊆ Mp . Therefore, ρ(x) = (x − π(x)) + γ (π(x)) ∈ Mp . If p ∈ T , then ρ(x) − ρ(p) (x) = γ (π(x)) − τ(p) (π(x)) ∈ p a π(Mp ) ⊆ p s Mp . As a result, ρ(N ) ⊆ M and ρ ≡ ρ(p) mod p s Mp for all p ∈ T . It is clear that ρ extends σ . Remark 2.2 Theorem IV of [1] gives a version of the above theorem with additional primitivity conditions imposed on both the ρ(p) and ρ. When M itself is positive definite, the following version of Theorem 2.1 will be found more useful in later discussion. Theorem 2.3 Let σ : K→M be a representation. Suppose that M is positive definite and that m ≥ k + 2(n − k) + 3. Let s be a positive integer and T be a finite set of primes which contains all primes that divide 2d(M)d(K). There exists a constant C2 = C2 (K, M, n, T , s) with the following property: Suppose that for each prime p, there exists a representation ρ(p) : Np → Mp which extends σ . If μk+1 (N ) > C2 , then there exists a representation ρ : N→M which extends σ and ρ ≡ ρ(p) mod p s Mp for all p ∈ T . Proof This is a direct consequence of the next lemma.
Lemma 2.4 Suppose that M is positive definite and K ⊆ N ∩ M. Then there exists a constant C = C (K) such that μ1 (π(N )) > C μk+1 (N ). Proof The lemma is clear when k = 0. Suppose that k ≥ 1, and fix a basis {e1 , . . . , ek } of K. Let x ∈ N such that π(x) is a minimal vector of π(N ). Let be the lattice spanned by {e1 , . . . , ek , x}. Then d() = d(K)Q(π(x)) = d(K)μ1 (π(N )). However, since {e1 , . . . , ek , x} is a set of k + 1 linearly independent vectors in N , there exists a constant α (depending only on k) such that d() ≥ αμ1 (N ) · · · μk (N )μk+1 (N ). Therefore, μ1 (π(N )) ≥
αμ1 (N ) · · · μk (N ) μk+1 (N ). d(K)
Extensions of representations of integral quadratic forms
149
)···μk (N ) Take C to be the quantity αμ1 (Nd(K) . It is independent of N because N contains K whose rank is k and so μ1 (N ), . . . , μk (N ) are bounded above by a constant independent of N .
Remark 2.5 Let F be a totally real number field and O be its ring of integers. We define the minimum of a totally positive O-lattice L by μ1 (L) = min{tr(Q(v)) : 0 = v ∈ L}, where tr is the trace from F to Q. Since the main theorem in [7] holds in this general setting (see [7], Remark (ii), p. 100), Theorem 2.1 holds for O-lattices as well.
3 New almost 2-universal lattices of rank 6 In this section, all lattices are assumed to be positive definite. A lattice M is called n-universal if M represents all lattices of rank n, and is called almost n-universal if it represents all but finitely many lattices of rank n. It is not hard to see that an almost nuniversal lattice must have rank at least n + 3. It was shown in [10] that there are only finitely many isometry classes of almost n-universal lattices of rank n + 3 for n ≥ 2. When n = 2, there are exactly 11 isometry classes of 2-universal lattices of rank 5; see [8]. Therefore there are infinitely many isometry classes of almost 2-universal lattices of rank 6 or higher. A natural question is whether every almost 2-universal lattice of rank 6 contains some almost 2-universal lattices of rank 5. This leads to the following definition. Definition 3.1 An almost n-universal lattice M is called new if it does not contain an almost n-universal sublattice of smaller rank. As an application of Theorem 2.3, we shall prove in this section that there are infinitely many new almost 2-universal lattices of rank 6. Lemma 3.2 Let a, b be a pair of relatively prime positive odd integers such that 5 ab, and let M(a, b) be the rank 6 lattice 1, 1, 2, 5, a, b. For any positive integer α, there exists z ∈ M(a, b) such that (1) Q(z) = α; (2) if N is a binary lattice which contains z primitively, then for each p there is a representation of Np by M(a, b)p which extends Zp [z]. Proof Suppose that M(a, b) has the matrix presentation 1, 1, 2, 5, a, b with respect to some orthogonal basis {x1 , . . . , x6 }. Since a and b are fixed throughout this proof, we shall denote M(a, b) simply by M. It is easy to verify that the lemma is true when α = 1; so we assume in the following that α > 1. We claim that there exists ∈ Z with the following properties: (a) α − 2 is represented by Z[x2 , x3 , x4 ] ∼ = 1, 2, 5; whenever p | α. (b) ∈ Z× p
150
K.W. Chan et al.
A direct computation shows that the above claim holds for all α ≤ 25. For the convenience of the readers, we list the choices of for all α ≤ 25 in the following table:
α
1 2 3 4
2, 3, 4, 5, 6, 7, 8, 9, 10, 12, 18, 22, 24 11, 13, 15 14, 16, 20, 22 17, 19, 21, 23, 25
For α ≥ 26, here are the choices for : 2 or 4
if 5 α and 2 α
5
if 5 α but 2 | α
2
if 5 | α but 2 α
1
if 5 | α and 2 | α
Since 1, 2, 5 has class number 1 (see, for example, [5]), it suffices to check that α − 2 is represented by the genus of 1, 2, 5, which can be done in a routine manner. So, there exists a vector z ∈ Z[x1 , x2 , x3 , x4 ] such that Q(z) = α
and = B(z, x1 ).
Let N be a binary lattice which contains z primitively. If α ∈ Z× p , then z extends to ⊥ an orthogonal basis {z, u} for Np . Moreover, Zp [z] in Mp represents all elements in Zp . Therefore, Mp contains a vector e such that B(z, e) = 0 and Q(e) = Q(u), and hence there is a representation of Np by Mp which sends z to itself and u to e. From now on, α is assumed to be divisible by p. Let J = Z[z, x1 ] and K be the orthogonal complement of J in M. Note that Jp is unimodular because d(Jp ) =
α − 2 ∈ Z× p . Suppose that N has a matrix presentation β γ . It suffices to show that for each p, there exists w ∈ Jp satisfying B(z, z) B(z, w) α β = and γ − δ→Kp . B(z, w) B(w, w) β δ α β
Since z is a primitive vector in the unimodular lattice Jp , there exists w ∈ Jp such that B(z, w) = β; see [11], 82:17. It remains to show that γ − Q(w) is represented by Kp . It suffices to show that Kp is universal, that is, Kp represents all elements in Zp . Note that Kp is split by a unimodular sublattice of rank at least 3. Therefore, Kp is universal when p > 2. Suppose now that p = 2. Then K2 ∼ = a, b, c1 , 2c2 for , where H is the hyperbolic plane and ∼ . So, K some c1 , c2 ∈ Z× H ⊥ d ⊥ 2d = 2 2 d, d ∈ Z× 2 . This shows that K2 is also universal.
Extensions of representations of integral quadratic forms
151
Theorem 3.3 There are infinitely many isometry classes of new almost 2-universal lattices of rank 6. Proof Let a, b be a pair of integers satisfying the hypothesis of Lemma 3.2. Let N be α β a binary lattice with matrix presentation β γ in reduced form, that is, 2|β| ≤ α ≤ γ . If α is sufficiently large, then N→M(a, b) by [6]. Furthermore, for any fixed α, if γ is sufficiently large, then N→M(a, b) by Theorem 2.1 and Lemma 3.2. Consequently, M(a, b) is almost 2-universal. Since there are only finitely many isometry classes of almost 2-universal lattices of rank 5 by [10], there exists an a such that 1, 1, 2, 5, a does not contain any almost 2-universal lattice. Therefore, for all sufficiently large b, M(a, b) does not contain any almost 2-universal lattice of rank 5. 4 Quadratic equations with linear conditions Let Mm,n (Z) be the set of m × n integral matrices and Sn+ (Z) be the set of positive definite symmetric matrices of rank n. The following is a matrix version of Theorem 2.3. + (Z), H ∈ S + (Z), A ∈ M Theorem 4.1 Let M ∈ Sm k,m (Z) and B ∈ Mk,n−k (Z) n−k with k < n such that m ≥ k + 2(n − k) + 3. Then there exists a constant C = C(M, A, B) > 0 satisfying the following property: If the system of Diophantine equations
X t MX = H
and AX = B
(4.1)
has a solution X(p) ∈ Mm,n−k (Zp ) for every prime p and μ1 (H ) > C, then it has a solution X ∈ Mm,n−k (Z). Proof Let T := t (A · adj(M)) and K := t T MT . Then, any T˜ ∈ Mm,n−k (Z) which satisfies the matrix equation K det(M)B t ˜ ˜ (T , T )M(T , T ) = t (det(M)B) H also satisfies (4.1). Therefore, this theorem follows from Theorem 2.3.
As an example to illustrate the above theorem, let us consider the simplest case in which n = 2 and k = 1. Corollary 4.2 Let a1 , . . . , am be pairwise relatively prime positive odd integers and 1 , . . . , m be any integers such that gcd(1 , . . . , m ) = 1. Suppose that m ≥ 6 and b is a fixed integer. There exists a constant C, depending only on b, the ai s and the j s, such that the system of Diophantine equations 2 = h, a1 x12 + a2 x22 + · · · + am xm 1 x1 + 2 x2 + · · · + m xm = b
152
K.W. Chan et al.
has an integral solution provided (i) h > C and (ii) (h − b)
m
i=1 i
is even.
Proof Let M be the lattice a1 , . . . , am and {x1 , . . . , xm } be an orthogonal basis of M such that Q(xi ) = ai for all i. Let A = a1 · · · am , and for any i = 1, . . . , m, let Ai := A/ai and M(i) be the orthogonal complement of Z[xi ] in M. Let x := 1 A1 x1 + · · · + m Am xm and Q(x) = α. Then, by virtue of Theorem 2.3, it suffices to show that for each prime p, there exists a vector y ∈ Mp such that α bA . (4.2) Zp [x, y] ∼ = bA h We first treat the case when p is an odd prime. Without loss of generality, we may assume that p 1 . There exists a primitive vector z ∈ M(1)p such that x − 1 A1 x1 = p t z for a nonnegative integer t. Suppose that p A1 . Then M(1)p is unimodular of rank ≥ 5, and hence it contains a pair of primitive vectors e, f such that Q(e) = 2 Q(z), Q(f) = h − −2 1 b a1 and B(e, f) = 0. By [9], Corollary 5.4.1, there exists ρ ∈ O(M(1)p ) such that ρ(e) = z. We may then take y to be the vector −1 1 bx1 + ρ(f). If p | A1 , we may assume that p | am and p Am . In this case, M(m)p is unimodular of rank ≥ 5. Let z be a primitive vector in M(m)p such that p k z = x − m Am xm . By a similar reasoning as above, there exists a vector y ∈ M(m)p such that B(z , y) = p −k bA and Q(y) = h. One can check easily that this vector y satisfies (4.2). Now, we consider the case when p = 2. Note that M2 is unimodular. Suppose that of Z2 [x] in M(1)2 α ∈ Z× 2 . One can readily check that the orthogonal complement m is even if and only if m i=1 i is odd. Since (h − b) i=1 i is even, there exists a vector w ∈ Z2 [x]⊥ such that Q(w) = h − b2 A2 α −1 . We can then take y to be vector w + bAα −1 x. Finally, assume that α ∈ 2Z2 . Without loss of generality, we may assume that 1 × ⊥ is odd. Let v be the vector x − 1 A1 x1 ∈ M(1)2 . Since Q(v) ∈ Z 2m, Z2 [v] in M(1)2 is a unimodular Z2 -lattice. Moreover, it is even if and only if i=2 i is odd. As is explained in the last paragraph, there exists a vector w in the orthogonal comple2 ment of Z2 [v] in M(1)2 such that Q(w) = h − −2 1 b a1 . Then we may take y to be −1 1 bx1 + w. References 1. Böcherer, S., Raghavan, S.: On Fourier coefficients of Siegel modular forms. J. Reine Angew. Math. 384, 80–101 (1988) 2. Cassels, J.W.S.: Rational Quadratic Forms. Academic Press, London (1978) 3. Hsia, J.S.: Arithmetic of indefinite quadratic forms. In: Contemporary Mathematics, vol. 249, pp. 1–15. Am. Math. Soc., Providence (1999) 4. Hsia, J., Kitaoka, Y., Kneser, M.: Representations of positive definite quadratic forms. J. Reine Angew. Math. 301, 132–141 (1978)
Extensions of representations of integral quadratic forms
153
5. Jagy, W.C., Kaplansky, I., Schiemann, A.: There are 913 regular ternary forms. Mathematika 44, 332–341 (1997) 6. Jöchner, M.: On the representation theory of positive definite quadratic forms. In: Contemporary Mathematics, vol. 249, pp. 73–86. Am. Math. Soc., Providence (1999) 7. Jöchner, M., Kitaoka, Y.: Representation of positive definite quadratic forms with congruence and primitive conditions. J. Number Theory 48, 88–101 (1994) 8. Kim, B.M., Kim, M.-H., Oh, B.-K.: 2-universal positive definite integral quinary quadratic forms. In: Contemporary Mathematics, vol. 249, pp. 51–62. Am. Math. Soc., Providence (1999) 9. Kitaoka, Y.: Arithmetic of Quadratic Forms. Cambridge Tracts in Mathematics, vol. 106. Cambridge University Press, Cambridge (1993) 10. Oh, B.-K.: The representation of quadratic forms by almost universal forms of higher rank. Math. Z. 244, 399–413 (2003) 11. O’Meara, O.T.: Introduction to Quadratic Forms. Grundlehren der mathematischen Wissenschaften, vol. 117. Springer, Berlin (1963) 12. Schulze-Pillot, R.: Representation by integral quadratic forms—a survey. In: Contemporary Mathematics, vol. 344, pp. 303–321. Am. Math. Soc., Providence (2004)