UNITED STATES DEPARTMENT OF AGRICULTURE 



BULLETIN No. 563 



Contribution from Bureau of Animal Industry. 

 A. D. MELVIN, Chief. 



Washington, D. C. 



PROFESSIONAL PAPER 



June 19, 1917 



THE DETERMINATION OF BACTERIA IN ICE CREAM. 



By S. Henry Ayers and W. T. Johnson, Jr., of the Dairy Division. 

 CONTENTS. 



Difficulty of making accurate bacteriological 



analyses. ' 1 



Method of sampling and plating the ice cream . 2 

 Variation in the bacterial content of com- 

 mercial ice cream 3 



Variation when held in an ice-cream cabinet. 8 



Variation when held in storage 9 



Variation in samples taken directly from 



freezer 10 



Comparison of incubation at 37" C. for two 

 days and at 30° C. for five days 12 



Number of colonies most desirable on Petri 

 plates ., r 13 



Variation between duplicate counts from 

 same sample 14 



Interpreting differences in bacterial counts. . . 15 



Summary and conclusions 16 



DIFFICULTY OF MAKING ACCURATE BACTERIOLOGICAL ANALYSES. 



Statements have been made that the distribution of bacteria in 

 ice cream is markedly uneven, that there is great variability in the 

 bacterial counts of different portions of the same container, and that 

 this variability is so great that any small sample selected for analysis 

 will not represent the whole mass of the ice cream. 



It must be remembered that the accuracy of a bacterio- 

 logical analysis can never be so great as that of a chemical 

 analysis. In making bacterial counts we are dealing with living 

 organisms which are distributed in the material under examination. 

 The method of analysis follows the assumption that the bacteria, as 

 individual cells, are distributed evenly throughout the sample and 

 that the portion removed for analysis contains a number in exact 

 proportion to the total number in the sample. Having removed a 

 definite part, it must then be placed in a medium suitable for plating 

 in which the individual bacterial cells can multiply and form visible 

 colonies. The inaccuracy of such a method must be evident at once. 



We know that some bacteria are in clumps or chains, and many 

 organisms may then develop into one colony which must be counted 

 as a single colony. The removal of a quantity of material which will 

 contain the same number of bacteria in suspension as another like 

 quantity is known to be impossible. Since we are dealing with 



95793°— Bull. 563—17 1 



