Generalization of the method of maximum likelihood [comment and reply]

For original paper see ibid., vol. 28, p. 65-72 (1993). The approximations connected with the ML in its 'classic' form (hrMLH) are very well pointed out, introducing the terms S1 and S2 which provide a generalized, unbiased form of the maximum likelihood method, called grMLH. The original author replies to the discussion.>