Asymptotic distribution of classes of positive binary quadratic forms with conditions of divisibility of coefficients (Q950799)

The author's aim is to develop Yu. V. Linnik's investigation of asymptotic computation of the number of classes of positive binary forms with the condition of divisibility of coefficients in connection with applications of the discrete ergodic method extended by Linnik to arithmetic forms. Let \(h_1(-m, q)\) denote the number of classes of integral properly primitive positive binary quadratic forms of determinant \(m\), the initial coefficients of which are divisible by a given number \(q\). The first result on the asymptotic behavior of \(h_1(-m, q)\) as \(m\to\infty\) for the case where \(q\) is a square of a prime number and \(m > 3\) is odd has been obtained by Linnik in 1955. With the help of the ergodic method, this result was extended by \textit{A. V. Malyshev} and the author [Algebra and Number Theory, Nal'chik, 4, 53--67 (1979)] to the case of composite \(q\) without the assumption that \(m > 3\) is odd. In the paper under review the following theorem is proved: Theorem 2: Let \(h^*_{1,3}(-m, q)\) be the number of classes of properly primitive positive binary quadratic forms of determinant m such that the product of their extreme coefficients is divisible by a given odd number \(q\geq 3\). Moreover, let \(\left(\frac{-m}{p}\right)=\)1 for all prime \(p\mid q\). Then, as \(m\to\infty\), \[ h^*_{1,3}(-m, q)\sim \frac{2^{\nu(q)}\tau(q)}{\sigma_0(q)} h(-m), \tag{1} \] where \(\nu(q)\) is the number of different prime divisors of \(q\), \(\tau(q)\) is the number of divisors of \(q\), \(\sigma_0(q) =q\prod_{p\mid q}\left(1 + \frac1p\right)\), and \(h(-m)\) is the number of classes of properly primitive positive binary quadratic forms of determinant \(m\). The constant values present in the asymptotic formula (1) depend only on \(q\). This is a special case of a more general Theorem 3 which is the only one proved in the paper (whereas (similar) proofs for the related Theorems 1, 2, 4, and 5 have been omitted). The proof (of Theorem 3) makes use of the author's previous paper in [ J. Sov. Math. 19, 1122--1155 (1982); translation from Zap. Nauchn. Semin. Leningr. Otd. Mat. Inst. Steklova 93, 87--141 (1980; Zbl 0457.10028)].