A convexity property of zeros of Bessel functions (Q809236)

For \(\nu \geq 0\) and \(\kappa >0\) let \(j_{\nu \kappa}\) be the solution of the differential equation \[ \frac{dj}{d\nu}=2j\int^{\infty}_{0}K_ 0(2j \sinh t)e^{-2\nu t} dt, \] where \(K_ 0(x)\) is the modified function of order zero, with the boundary condition \(\lim_{\nu \to - \kappa +0}j(\nu)=0\). For \(\kappa =k=1,2,..\). we obtain the positive zeros \(j_{\nu k}\) of \(J_{\nu}(x)\). The authors present this new interesting property. Let f(\(\nu\)) be defined by \[ f(\nu)=\log \frac{2}{\nu +1}- \frac{1}{4},\quad 0\leq \nu \leq 1, \] \[ f(\nu)=\frac{1}{4}(\nu \log \nu - \nu),\quad \nu \geq 1. \] Then the function \(j_{\nu \kappa}+f(\nu)\) is convex with respect to \(\nu \geq 0\) provided \(\kappa \geq \kappa_ 0=0.7070... \). The paper contains a misprint in the definition of f(\(\nu\)).