On the 2-part of the class numbers of cyclotomic fields of prime power conductors (Q763990): Difference between revisions

Let \(p\) be an odd prime, \(n \in \mathbb N\) and \(K_n = \mathbb Q(\zeta_{p^{n+1}})\) be the \(p^{n+1}\)-st cyclotomic field. Denote by \(h_n\) and \(h_n^-\) the class number and the relative class number of \(K_n\), respectively. Now fix a prime \(l \not= p\). Then a classical result of \textit{L. C. Washington} [Invent. Math. 49, 87--97 (1978; Zbl 0403.12007)] states that the quotient \(h_n/h_{n-1}\) is not divisible by \(l\) for \(n > m_p\) for \(m_p\) sufficiently large. In the case \(l=2\) the authors of the paper under review determine an \textit{explicit} bound \(m_p\) such that the quotient \(h_n^- / h_{n-1}^-\) is odd for all \(n > m_p\). When \(l>2\) the authors obtain a similar result on the \(l\)-parts of the relative class numbers of \(K_n^+(\zeta_l)\) under additional assumptions. Note that \(h_n^+\) is odd if \(h_n^-\) is by a result of \textit{K. Iwasawa} [Nagoya Math. J. 27, 239--247 (1966; Zbl 0139.28104)]. As mentioned in Remark 1(I) of this paper, one can similarly show that \(h_n^+ / h_{n-1}^+\) is odd if \(h_n^- / h_{n-1}^-\) is. Using these results and computational methods, the authors obtain the following theorem: For any odd prime \(p \leq 509\) the ratio \(h_n / h_0\) is odd for all \(n \geq 1\).