July 15. 1915 Methods of Bacterial Analyses of Air 351 



Thus far the results were contradictory, since the tests where the aero- 

 scopes were placed in tandem had shown a greater efficiency for the Rett- 

 ger filters than for the sand filters, while in the duplicate tests of the 

 two types of aeroscopes the sand filters gave better results. Moreover, 

 there were several unsatisfactory features of the tests. Possible leakage 

 had not been perfectly eliminated, the number of bacteria present in the 

 air was so low that slight errors due to contaminations were greatly mag- 

 nified, and the sterilization of the sand filters with steam produced unsat- 

 isfactory results. The latter caused the sand to cake more or less, as 

 already pointed out, and soon injured the rubber stoppers and connec- 

 tions to such an extent that leakage was always a possibility. 



At this point a modification of the standard method was tested. A 

 cotton layer sufficiently firm to support the column of sand was inserted 

 just above the lower rubber stopper. The lower rubber stopper was then 

 removed until the aeroscope proper had been sterilized with dry heat. 

 After sterilization the rubber stopper was carefully inserted. Two series 

 of tests of this modified form of aeroscope were made, one of 43 and the 

 other of 38 trials, but the results were so contradictory and inconclusive 

 that they are not given in detail. At this time the modification described 

 and illustrated (fig. 2) was devised. 



COMPARISON OF THE EFFICIENCY OF THE MODIFIED STANDARD AND THE STANDARD 



AEROSCOPE 



In comparing the efficiency of the aeroscopes of the standard type with 

 those of the modified type, the aeroscopes were set up in tandem and the 

 tandem couples of each type of filter were used for duplicate analyses. 

 The openings of the aeroscopes were always placed side by side, and 

 similar aspirators were used so that the rate of filtration could be kept 

 the same. The samples of air were all taken in the stable when the cows 

 were in their stalls and during such operations as feeding or milking. 

 This was done because previous experience had shown that the number 

 of bacteria in the air of the stable when it was empty or when everything 

 was quiet was so low that very few colonies appeared on the plates. In 

 every case the sand was shaken out into 10 c. c. of sterile physiological 

 salt solution. One c. c. of the resulting bacterial suspension was added 

 to each plate. Plating was done in duplicate in the first series given in 

 Table II and in triplicate in the second series. 



