Threefield identities and simultaneous representations of primes by binary quadratic forms (Q740919): Difference between revisions

It has been proven by \textit{I. Kaplansky} in [Proc. Am. Math. Soc. 131, No. 7, 2299--2300 (2003; Zbl 1055.11026)] that a prime \(p\), where \(p \equiv 1 \pmod{16}\), is representable by both or none of the quadratic forms \(x^2 + 32y^2\) and \(x^2 + 64y^2\), and that a prime \(p\), where \(p \equiv 9 \pmod{16}\), is representable by exactly one of the above stated quadratic forms. Using class field theory, \textit{D. Brink} [ibid. 129, No. 2, 464--468 (2009; Zbl 1209.11042)] was able to prove five more theorems similar to that of Kaplansky and claimed that these are the only results of their kind giving a heuristic argument as support. Recall that for a negative integer \(\Delta\) with \(\Delta \equiv 0 \pmod 4\) (resp. \(\Delta \equiv 1 \pmod 4\)) the principal binary quadratic form \(F (x, y)\) of discriminant \(\Delta\) is defined to be \(x^2 - \frac{\Delta}{4} y^2\) (resp. \(x^2 +xy+ \frac{1-\Delta}{4} y^2\)). It turns out that there are at least two pairs of discriminants for Kaplansky-like theorems on principal binary quadratic forms that are not on Brink's list. Namely, the author shows that a prime \(p \equiv 1 \pmod{48}\) is representable by both or none of \(x^2 + 64y^2\) and \(x^2 + 288y^2\), whereas a prime \(p \equiv 25 \pmod{48}\), is representable by exactly one of the quadratic forms. Moreover, a prime \(p \equiv 1,65,81 \pmod{112}\) is representable by both or none of \(x^2 + 56y^2\) and \(x^2 + 448y^2\), whereas a prime \(p \equiv 9, 25, 57 \pmod{112}\) is representable by exactly one of these forms.