Three letters to Walter Feit on group representations and quaternions. (Q2474469)

In 1974, in an unpublished note, answering a question of the author, W. Feit proved the following: If \(G\) is the quaternion group of order 8 and \(K=\mathbb{Q}[\sqrt{-35}]\) and \(\rho\colon G\to\text{GL}_2(K)\) is an irreducible representation of \(G\) over \(K\), then \(\rho\) cannot be written over \(O_K\), the ring of integers of \(K\). Generalizing this, the author proved that the above statement remains true if 35 is replaced by any \(N\equiv 3\pmod 8\) such that \(N\) is divisible by an odd power of a prime \(p\) with \(p\equiv\pm 1\pmod 8\). In 1997, he found a non-computational proof of this as a consequence of the ``genus theory'' of Gauss and Hilbert, and the paper contains three letters of the author to Feit explaining this proof.