MICROORGANISMS AS A FACTOR IN SOIL FERTILITY 371 



by the microscope is probably, according to Conn, 5, 10 or over 20 times 

 greater than that indicated by the plate method. The discrepancy is 

 due to the failure of many cells to produce colonies rather than to the 

 occurrence of large clumps that o not break up in the process of 

 plating. 



QUANTITATIVE RELATIONS. Since the early work of Koch in 1881 

 many investigators have determined the number of bacteria in soil 

 samples, by means of the plate method. It is well known, however, 

 that on ordinary gelatin or agar plates kept under aerobic conditions 

 but a fraction of the soil organisms produce visible colonies. -The 

 anaerobic species do not appear, nor do aerobic Azotobacter, and nitro- 

 bacteria, while other common soil organisms form colonies sparingly 

 or not at all. By employing synthetic agar media instead of beef broth 

 gelatin or agar, Lipman and Brown have succeeded in securing the 

 growth of a much larger number of colonies from any given quantity 

 of soil, yet even these larger numbers were incomplete for reasons 

 mentioned above. 



H. Fischer recommends a simple medium of agar to which nothing 

 has been added but soil extract (prepared by extracting with a .1 

 per cent solution of Na 2 CO 3 ) and potassium phosphate. Following 

 the path of Lipman and Brown in reducing the content of organic 

 matter, Temple employed i g. of peptone per 1. as a culture medium 

 and obtained satisfactory results. Brown has further modified the 

 formula of Lipman and Brown by replacing the .05 g. of peptone with 

 .1 g. of albumin, and obtained results which were somewhat superior. 

 In a comparison of culture media, Conn considers the former media 

 undesirable for quantitative purposes because they contain substances 

 of indefinite chemical composition, and offers an agar medium con- 

 taining no organic matter except agar, dextrose and sodium asparag- 

 inate, and also a soil-extract gelatin which is valuable for qualitative 

 purposes. Other media that have been suggested, after a comparison 

 of all of the above-mentioned media, are the urea-ammonium nitrate 

 agar of R. C. Cook and the tap water gelatine and asparaginate agar of 

 Conn. It is evident, therefore, that the results secured in the counting 

 of soil bacteria have only a relative value. With the same media and 

 methods some information may be secured concerning the influence 

 of fertilization, tillage, liming, etc., on certain of the soil bacteria. But 

 even this information must be properly discounted, since equal numbers 



