“…An essentially more modem treatment of the problem is due to Rainal (1987) and Mimaki (1978Mimaki ( , 1981. In these works, the estimation of the conditional density was formulated in a different way, and by making the assumption that a given individual interval between threshold crossings is statistically independent of nearly all the possible sums of preceding intervals, the problem was reduced to the solution of a Volterra integral equation involving the numerical results fi-om the approximation just described.…”