Creepers: Real quadratic orders with large class number

Shanks's sequence of quadratic fields $\Q(\sqrt{S_{n}})$ where $S_{n}=(2^n+1)^2 + 2^{n+2}$ instances a class of quadratic fields for which the class number is large and, therefore, the continued fraction period is relatively short. Indeed, that period length increases linearly with $n$, that is: in arithmetic progression. The fields have regulator $O(n^2)$. In the late nineties, these matters intrigued Irving Kaplansky, and led him to compute period length of the square root of sequences $a^2x^{2n}+bx^{n}+c$ for integers $a$, $b$, $c$, and $x$. In brief, Kap found unsurprisingly that, generically, triples $(a,b,c)$ are `leapers': they yield sequences with period length increasing at exponential rate. But there are triples yielding sequences with constant period length, Kap's `sleepers'. Finally, there are triples, as exemplified by the Shanks's sequence, for which the period lengths increase in arithmetic progression. Felicitously, Kaplansky called these `creepers'. It seems that the sleepers and creepers are precisely those for which one is able to detail the explicit continued fraction expansion for all $n$. Inter alia, this thesis noticeably extends the known classes of creepers and finds that not all are `kreepers' (of the shape identified by Kaplansky) and therefore not of the shape of examples studied by earlier authors looking for families of quadratic number fields with explicitly computable unit and of relatively large regulator. The work of this thesis includes the discovery of old and new families of hyperelliptic curves of increasing genus $g$ and torsion divisor of order $O(g^2)$. It follows that the apparent trichotomy leaper/sleeper/creeper coincides with the folk belief that the just-mentioned torsion is maximum possible.