A law of the iterated logarithm for sums of extreme values from a distribution with a regularly varying upper tail (Q1103952)

Let \(X_ 1\), \(X_ 2\),... be independent observations from a distribution with a regularly varying upper tail with index a greater than 2. For each \(n\geq 1\), let \(X_{1,n}\leq...\leq X_{n,n}\) denote the order statistics based on \(X_ 1\),..., \(X_ n\). Choose any sequence of integers \((k_ n)_{n\geq 1}\) such that \(1\leq k_ n\leq n\), \(k_ n\to \infty\), and \(k_ n/n\to 0.\) It has been recently shown by \textit{S. Csörgö} and \textit{D. M. Mason} [Ann. Probab. 14, 974-983 (1986; Zbl 0593.60034)] that the sum of the extreme values \(X_{n,n}+...+X_{n-k_ n,n}\), when properly centered and normalized, converges in distribution to a standard normal random variable. In this paper, we completely characterize such sequences \((k_ n)_{n\geq 1}\) for which the corresponding law of the iterated logarithm holds.