Comment—Revised estimates of the bias in the 'minimum number alive' estimator