![]() ![]() ![]() |
Table 1 gives the values of
for which the predicted moments become negative. The finite differences formed from the moments with
do not in this case yield stronger information. However, M.R. Pennington has pointed out to us that if one considers all moments but eliminates those with
which are not determined by the theory, then the constraints obtained are stronger than those resulting from the positivity of the first
moments and their finite differences alone [9]. The additional conditions hardly affect the lower boundary of the allowed region in table 1 but the upper boundary is reduced to
for all
[10].