The theorem of Vijayaraghavan and Hardy (Q5946018)

The paper improves upon the headline theorem, which is a Tauberian type one that infers---possibly instrumental for proper Tauberian theorems---boundedness of a real sequence from boundedness of its transform by regular nonnegative matrices \(C=(c_{xk})\) \((x\geq 0\), \(k=0,\;1, \dots)\) with row sums 1. G. H. Hardy's version is Theorem 238 in [Divergent series (Oxford: At the Clarendon Press) (1949; Zbl 0032.05801)]; the authors put it, incorporating Theorem 239, into a refined form which we will refer to (Satz VH). Thus, the Tauberian condition (TB) on \(s=(s_k)\) says \(s_q-s_p>- (\Phi_q- \Phi_p)-1\) \((q\geq p)\) with a function \(\Phi\) such that \(0< \Phi_t\uparrow \infty\) \((0\leq t\to\infty)\) and \(\Phi_{t+1}-\Phi_t\) is bounded above. In order to result in the Tauberian conclusion \((\text{B}C\Phi)\): ``if \(Cs\) is bounded and (TB) holds then \(s\) is bounded'', Satz VH required the sections \(\sum_{k\leq M}c_{xk}\), \(\sum_{k\geq N}c_{xk} (\Phi_k-\Phi_N)\) (see for reduction of (3.1)) to allow for some kind of simultaneous approach to zero. It is being sharpened in that the sections and their parameters are subject to related bound conditions (Satz 4.1), as is appropriate for a ``bounded-Tauberian'' theorem. Variants are given by way of theorems and notes; \((\text{B} C\Phi)\) is applied to the matrix \(C\) of some Valiron methods or the standard Borel one and to \(\Phi_t= \sqrt t\) in each case.