In the former paper, it was laid down, as the basis of any improvements I might be able to suggest as to life contingency calculation, that, in place of premises of a specific nature leading to special conclusions, we should rather, from the very commencement, recognise the variable nature of the results we might have to encounter, and accordingly construct a system of limits rather than one of particular values.