132 



INVESTIGATION OF DIFFERENT METHODS OF TESTING 

 BABCOCK MILK BOTTLES. 



BY 



B. J. Smit, B a. 

 Division of Chemistry, Dcpuitiiient of AyricuU arc , Pretoria. 



Read July 10, 1922. 



The usual quick method of testing Babcock Milk Bottles 

 used in the Cedara Laboratory is a volumetric one, but before 

 finally rejecting any particular bottle as being outside the limits 

 of error allowed, it has been usual to carry out a confirmatory 

 test by a gravimetric method, using either water or mercury. 

 Having found in several instances serious discrepancies between 

 the two sets of results, I decided to investigate the matter further 

 in order to see what differences one would expect in results 

 obtained by the different methods of calibration, when only taking 

 the ordinary amount of care and observing the usual precautions in 

 each case. 



The following details are given of the three methods of 

 calibration employed. 



I. Volumetric Method. 



For this method, 1 c.c. standard graduated pipettes are used, 

 which have been carefully calibrated by running their contents 

 of distilled water into a bottle and weighing, after allowing to 

 drain for three minutes, and then carefully blowing out the last 

 drop. The calibration of each pipette was done for the whole 

 1 cc. and the en"or in delivery was found to be fairly constant 

 under the conditions of working, being 0-008 c.c. This gives a 

 correction of 0"08 per cent, over a range of 10 per cent, on the 

 Babcock scale for milk bottles. 



When testing, the milk bottle is first filled with water exactly 

 to the zero mark, and the neck carefully dried with a I'oU of 

 blotting or filter paper. Then the contents of the graduated 

 pipette up to O'B c.c. are carefully run into the neck of the bottle 

 allowing the pipette about three minutes to drain intO' the bottle. 

 At the end of that period the last drop is blown out, and the 

 reading on the neck taken, adding a correction of 0"03 per cent. 

 This is repeated, and the reading then taken for the whole range 

 of the scale (that is, 8 per cent.), allowing now 0'06 per cent, as a 

 correction. 



It is realised that the result of the calibration of the lower 

 half of the graduated scale is vitiated somewhat by the fact that 

 a small proportion of the water run in clings as a film on the 

 top half of the neck. 



