COLLECTION OF SAMPLES 203 



which this vessel has had very appreciable effects upon the bac- 

 terial count and the number of coliform bacteria. For the col- 

 lection of combined chemical and bacteriological samples the 

 author has used for several years rectangular, narrow-necked, 

 six-ounce glass-stoppered bottles, 16 of which can be placed in a 

 tray, 10 by 6| inches. This tray is surrounded with ice and 

 water, and the whole contained in a water-tight galvanised-iron 

 box 14J by 10J by 7 inches. In cold climates the cooling 

 mixture can be dispensed with in winter and when there is 

 any possibility of the milk freezing, wide-mouthed bottles 

 should be used to prevent freezing of the sample and so blocking 

 the neck of the bottle during the transfer of the sample. All 

 milk retailed in bottle should be delivered to the laboratory 

 in the original container unopened as the only other method of 

 satisfactorily sampling such milk is to transfer the sample to a 

 sterile bottle and then back to the original container, this being 

 repeated several times.- The sterile bottle necessary for the 

 success of this method cannot usually be obtained so that this 

 system should not be encouraged. 



All samples should be labelled in such a way that there can 

 be no possibility of doubt as to the identity of each sample 

 and a complete record of the sampling data made immediately 

 after the sample is taken. This should include name of vendor, 

 date, time and- place, temperature, character of container 

 and name of collector. The temperature of milk in bulk is 

 observed immediately after the sample has been taken whilst 

 that of bottled milk should be obtained from a second bottle. 

 A quickly reacting Fahrenheit thermometer is suitable for this 

 purpose. 



If the object of the examination of samples is to obtain 

 figures representative of the total milk supply and from which 

 averages can be calculated which are strictly comparative from 

 month to month or from year to year, the collection of samples 

 must be carried out as scientifically as possible and not in the 

 usual haphazard fashion. The output of each vendor should 

 be estimated and the number of samples varied in proportion 



