DIRTY MILK 63 



accurately with those of another, but they are fairly uniform when 

 compared with each other. 



(6) The direct examination of milk by the Prescott-Breed 

 method. The details of this method are found in the " Cent, f . 

 Bact.," 11, XXX, p. 337, 1911, but since it is somewhat new, a 

 brief description of it will be here given. A capillary tube is pre- 

 pared, arranged to receive a rubber bulb at one end and marked 

 carefully to deliver one hundredth of a cubic centimetre. After 

 a most thorough mixing of the milk, one hundredth of a cubic 

 centimetre is removed with the sterilized pipette and spread 

 uniformly over a square centimetre on an ordinary microscopic 

 slide. It is allowed to dry and fixed with methyl alcohol, after 

 which the fat is dissolved from it by the use of xylol. The smear 

 is then stained either with methylene blue or preferably with one 

 of the blood stains, the Jenner stain or Wright stain being useful 

 for this purpose. If the staining is so deep as to make the speci- 

 men too opaque for proper study, it is slightly decolorized with 

 alcohol, which removes the stain from the general sediment more 

 readily than it does from the bacteria or the tissue cells. The 

 stained smear is studied under a twelfth inch immersion. The 

 draw tube of the immersion is adjusted so that the field of the 

 microscope covers exactly fifteen millimetres and under these 

 circumstances the number of bacteria present in the one hun- 

 dredth of a cubic centimetre is exactly five thousand times the 

 number found in a microscopic field. The counting of a large 

 number of fields (one hundred fields) and averaging the results, 

 multiplied by this number, will therefore give approximately the 

 number of cells or bacteria contained in a one hundredth 

 cubic centimetre of milk. This method has the advantage of 

 eliminating all of the errors that are associated with the irregu- 

 larity in the action of the centrifugal machine and many of the 

 other individual errors connected with the Stewart-Slack 

 method. On the other hand, it has the disadvantage of examin- 

 ing only a small quantity of milk, one hundredth of a cubic 

 centimetre being hardly a fair sample. This disadvantage, how- 

 ever, may be met by making a very thorough mixture of the 

 milk before sampling and by making the examination in dupli- 

 cate. A second disadvantage is the large multiplier that has to 

 be used to reduce the results to one cubic centimetre. This of 



