Correction to a paper by A. G. Pakes

Starting from a probability a on the half-line with moments of any order A. G. Pakes has defined probabilities a, by length biasing of order r and g(r) by the stationary-excess operation of order r, r = 1, 2,.... Examples are given to show that a can be determined in the Stieltjes sense while sigma(1), and g(1) are indeterminate in the Stieltjes sense. This shows that a statement in a recent paper by Pakes does not hold.