The Chung-Smirnov law of the iterated logarithm and the Finkelstein functio
nal law of the iterated logarithm for empirical processes are used to estab
lish new results on the central limit theorem, the law of the iterated loga
rithm, and the strong law of largo numbers for L-statistics with certain bo
unded and smooth weight functions. These results are used to obtain necessa
ry and sufficient conditions for almost sure convergence and for convergenc
e in distribution of some well-known L-statistics and U-statistics, includi
ng Gini's mean difference statistic. A law of the logarithm for weighted su
ms of order statistics is also presented. (C) 2001 Academic Press.