36o 



Journal of Agricultural Research 



Vol. IV, No. 4 



they are due to incidental differences in the bacterial quality of the air. 

 Accordingly three series of comparative tests were made, using the same 

 type of filter on each side of the comparison. The first series of 32 such 

 comparisons is recorded in Table VI. In this case two sand filters of the 

 type described on page 351 were used . Six liters of air artificially enriched 

 with dust were drawn through each filter at the same rate. The sand was 

 washed in 5 c. c. of sterile water, and i c. c. of the suspension was added 

 to each plate. 



Table VI. 



-Comparative counts obtained with two standard aeroscopes modified to 

 allow dry sterilization 



2 



3 



4 



5 



6 



7 



8 



9 



10 



II 



12 



13 



14 



15 



16 



17 



18 



19 



20 



21 



22 



23 



24 



25 



26 



27 



28 



29 



30 



31 



32 



Averacre 



Plate count. 



Count per liter of air. 



290 



239 



242 



As will be seen from Table VI, 32 tests gave an average of 287 colo- 

 nies per plate in one sample of air and 290 in the duplicate, or 239 

 colonies per liter in the sample and 242 in the duplicate. Such close 

 agreement of results, even in series as long as this, are rarely obtained 

 with the technique used. Marked variations occurred in individual 

 cases, however, showing the need of a long series of tests from which to 

 draw conclusions. 



