MICHIGAN ACADEMY OF SCIENCE. 125 



from 2 to 5 times as large with tlie microscope as with the petri plate. The 

 reason for this is not clear, though our experience with sample No. 10 sug- 

 gests a partial explanation. Representative colonies on the plates were fished 

 grown in milk, and stained. It was found that the bacteria took the stain very 

 poorly. Various stains were tried, but with little better success except with 

 Giemsa's stain, which differentiated the bacteria, and the casein of the milk 

 very well, when the culture was grown in milk. But when this stain was 

 applied to the butter smears it did not stain the bacteria more successfully 

 than other stains. Another possible cause for overlooking bacteria in counting 

 the smears is the small size of the usual flora of butter, which consists prin- 

 cipally of Bacterium lactis acidi. Neither of these explanations seems to fully 

 explain the discrepant results and further search is being made for the reason. 

 In two of the experiments the non-aqueous portion was examined for bacteria, 

 but none were found. It is realized that the foregoing discussion is based on 

 very limited data. The main reason for presenting it at the present time is 

 in the hope that some suggestion may be received here that will help to solve 

 the difficulties encountered. 



Experiment Station, 

 Michigan Agr. College. 

 East Lansing, Mich. 



REFERENCES. 

 ^Tanner, F. W. Bacteriology and Mycology of Foods. 



John Wiley and Sons, 1st ed., page 428. 

 'Lohnis, F. Laboratory Methods in Agricultural Bacteriology. 



Chas. Griffin & Co., Ltd., page 81. 

 ^Bowhill, Thos. Manual of Bacteriological Technique and Special Bacteriologv 



Wm. Wood & Co., 2nd ed., page 26. 



