138 CONN, ATKINS, KLIGLER, NORTON AND HARMON 



forgotten, however, that there is one other equally important 

 object in counting bacteria, for it is only by this means that we 

 can determine the abundance of the organisms in any particular 

 habitat — a problem of value from the standpoint of pure science. 

 This latter aspect of the matter especially concerns this Society 

 rather than organizations interested in disease or public health. 

 The very fact that the subject is being so thoroughly investigated 

 from the standpoint of the sanitarian makes it all the more 

 important that it be studied by this Society also. Standardi- 

 zation of methods^ — which has been the chief aun of other organi- 

 zations taking up the matter — tends to prevent progress by 

 fixing the technic. To counteract this tendency, the committee 

 on technic plans to compare the various methods of counting 

 bacteria, laying chief stress upon their accuracy, rather than upon 

 their adaptabihty to routine use. 



There are three criteria by which methods of counting bacteria 

 can be judged: (1) agreement of duplicate determinations; (2) 

 size of the counts obtained ; and (3) actual accuracy of the counts. 

 The first of these, agreement of duplicate determinations, is the 

 object desired in control work, where incomplete counts are 

 entirely suitable, provided a uniform fraction of the total number 

 of bacteria is counted. Size of the counts is the most commonly 

 used criterion as to the relative merits of different methods 

 of counting, because it is generally recognized that ordinary 

 counts are but partial ones and the presumption is that the 

 higher count is the more nearly correct. Actual accuracy, how- 

 ever, although a far better criterion, is the hardest of all to apply, 

 because counts may be too high instead of too low, and there is 

 no absolute standard of comparison with which to check up 

 results. Even the best bacterial counts are but estimates because 

 the total numbers are too high to count with absolute accuracy, 

 and high magnification is necessarj^ to see the individuals; so 

 the only way the accuracy of any one method may be determined 

 is by comparing it with other methods and discounting the 

 probable sources of error in each method. 



There are three general types of methods by which bacteria 

 may be counted : the dilution method, the plate method, and the 



