(2) Cutler, D. W., and Crump, L. M. (1920), Ann. App. Biol., vol. vii., 11-24. The discovery of this method at once puts into the hands of the investigator a much more efficient instrument for studying the activities of the soil micro-population, especially since at a slightly later date Thornton’s method for counting bacteria was devised. Early in 1920 Cutler and Crump[6] decided to make a preliminary survey of the protozoon and bacterial populations of one of the Rothamsted field soils (Broadbalk dunged plot). The investigation was continued for 28 days, daily soil samples being taken. The results so obtained showed that an extended investigation of the micro-population of field soil would yield interesting and important results, especially as it was evident that certain views held by soil biologists required modification. In July of the same year, therefore, it was decided to start an extended investigation of the soil protozoa and bacteria. The method adopted was to make counts of the numbers of bacteria and of six[G] species of protozoa in soil samples taken daily direct from the field (Barnfield dunged plot) and by statistical methods to correlate these counts one with another and with the data for external conditions. Observations at shorter periods than 24 hours could not be made, but it was found possible to continue the research for 365 days.[7] [G] Actual counts were made of six species, though, as stated on p. 10, observations were made on seventeen. [Illustration: FIG. 10.--Daily numbers of active amœbæ (Dimastigamœba and Species α) and bacteria in 1 gram of field soil, from August 29 to October 8, 1920. (From Phil. Trans. Roy. Soc., vol. ccxi.) X-axis: August September October Y-axis (left): Amoebae Active numbers per gramme of soil Y-axis (right): Bacteria in millions per gramme of soil Legend: Dimastigamoeba Species α Bacteria] The number of all the organisms showed large fluctuations of two kinds, daily and seasonal. The size of the changes that took place within so short a period as 24 hours was, perhaps, the most surprising fact that the experiment revealed. Thus three consecutive samples gave 58·0, 14·25 and 26·25 millions of bacteria per gram respectively; and the changes exhibited by any of the species of protozoa were at times even larger. This fact is of extreme importance, since in the past it has always been assumed that the number of bacteria remained fairly constant from day to day, and investigators have not hesitated to separate the taking of soil samples by long periods. It is now obvious that such a procedure is of little use for comparative purposes (Fig. 10).