On Some High-Indices Theorems II (Q4948567)

[For Part I see Trans. Am. Math. Soc. 69, 468-478 (1950; Zbl 0040.03601).] Let \(f(s)= \sum^\infty_1 a_ne^{-\lambda_ns}\) with \(0< \lambda_1 <\lambda_2< \dots\) and \(\lambda_n \to\infty\) and the Hadamard gap condition \(\lambda_{n+1}/ \lambda_n \geq q>1\), be a lacunary series for \(n\geq 1\). The existence and finiteness of \(\lim_{s\to 0^+} f(s)\) imply the convergence of \(\sum a_n\). Zygmund and Hardy and Littlewood had shown that if \(f(s)\) is of bounded variation in \((0,\infty)\) then \(\sum a_n\) converges absolutely, i.e., they proved the extreme cases corresponding to \(m=1\) and \(m=\infty\), respectively, of the inequalities \[ \sum|a_n |^m \lambda_n^{m-1} \leq A_{q,m} \int^\infty_0 \bigl|f(s)\bigr |^mds \] which hold for \(m\geq 1\). In the present paper the author proves two theorems generalizing the above result for \(0<m<1\) and weakening the requirement that \(f(s)\) is of bounded variation in \((0,\infty)\) assumed by Zygmund. The second theorem is a generalization of Zygmund's result for generalized variation, the \(\lambda\)-variation.