Key points are not available for this paper at this time.
SUMMARY: The use of a conditional probability model for the identification of bacteria is illustrated by a pilot study on a group consisting mainly of bacteria of the family Enterobacteriacae. It is shown that such models may be untrustworthy unless more stringent criteria are used about the basic data and about the results than is normally the case in studies of this kind, including the present one.
DYBOWSKI et al. (1968) studied this question.