The relative errors of bacteriological plate counting methods

It is generally conceded that the degree of precision ordinarily attained with dilution plate counting methods of enumerating viable bacteria is not so great as could be desired. The extent of variation encountered even under the best con-ditions is considerably larger than the maximum permitted, for example, of routine chemical analyses. Coefficients of variation consistently as high as 30 per cent of the mean have been reported (Mudge and Lawler, 1928; Ziegler and Halvorson, 1935), although considerably lower values have also been encountered (Jennison, 1937). Some of the principal sources of variation which may be identified at the present time are: (1) The presence of clumps, consisting of various numbers of organisms, in the original suspension (Jennison, 1937). (2) The error involved in preparing a dilute suspension (Jennison and Wads-worth, 1940). (3) The error involved in measuring aliquots of this suspension into plates. (4) The distribution of organisms in these aliquots (Fisher, Thornton, an