124 TWENTY-FIRST REPORT. 



One gram of butter was weighed out and transferred to a clean separa- 

 tory funnel. One cubic centimeter of hot water was added to the butter and 

 the whole mass agitated until the butter was melted and thoroughly mixed 

 with the water. Then 50 c.c. to 100 c.c. of ether, gasoline or other fat solvent 

 was added to dissolve the fat, when it was allowed to stand until the two 

 liquids had separated. The aqueous portion was drawn off into a graduated 

 cylinder and an equal amount of sterile skim milk which had been allowed to 

 setliment for a week or more after sterilization was added to the aqueous 

 extract of the butter. After mixing 1/50 of a cubic centimeter of the mix- 

 ture was spread out with a needle on a glass slide so that it covered an area of 

 1 square centimeter. Sometimes a larger amount of butter was used but in 

 those cases a larger amount of warm water was used so that in every case 

 1/50 c.c. of the final mixture represented approximately 1/100 gram of the 

 butter. The smears were then treated in the same manner as the milk smears 

 are treated in the Breed method of counting bacteria in milk ; namely, the 

 smears were dried quickly on a warm plate, immersed in xylol for a minute 

 to dissolve any trace of fat, fixed for one minute in alcohol, stained with a 

 one per cent aqueous methylene blue solution, destained in alcohol and dried. 

 Two duplicate smears were always made for each sample. The counting was 

 done imder a compound microscope in the same manner as in the Breed 

 method. In this work an E. Leitz 1/12 oil immersion objective was used with 

 a No. 1 ocular and the draw tube adjusted to 152 mm. There was also used 

 a special ocular micrometer marked into quadrants with cross-hairs and a 

 circle 8 mm. in diameter. This combination give a factor of 650,000 for 

 every bacterium seen in one field. From 10 to 40 fields were counted and 

 averaged per sample. The following results were secured (see table) : 



Butter 



Sample Plate Microscopic Count. 



No. Kind of Cream. Counts. Groups. Individuals. 



1. Pasteurized, ripened by starter 620.000 5,035,000 7,962,500 



2. Pasteurized, ripened naturally 168,000 9,250,000 14,925,000 



3. Pasteurized, ripened naturally 101,000 666,250 980,000 



4. Pasteurized, ripened by starter 5,000,000 9,600,000 14,880,000 



5. Pasteurized, ripened by starter 1,240,000 25,350,000 40,950,000 



6. Raw, ripened naturally 15,100,000 6,435,000 11,115,000 



7. Kaw, ripened naturally 16,600,000 10,010,000 17,680,000 



8. Unknown 3,165,000 3,022,750 5.590,000 



9. Raw, ripened naturally. 9,100,000 5,622,500 8,582,500 



10. Unknown, dairy butter 3,597,000 2,762,500 4,290,000 



325,000 325,000 



torulae torulae 



In every case when the cream was first pasteurized the results by the 

 microscopic group count were higher than by the plating method, which is 

 not surprising since the dead cells would stain as well as the living ones. 

 But where raw cream was used the results by the plating method are higher 

 than the microscopic group count. This is opposed to' all experience with 

 the microscopic technique when applied to milk, where one usually gets results 



