careful examination is necessary for its detection. Eor testing 

 amyl alcohol a special form of hydrometer, the " Tralles," can 

 be obtained. It should read 95-96 if the alcohol is of the proper 

 strength. A Reaumur thermometer is enclosed in the body of 

 this hydrometer, and it should be remembered that 12 R. cor- 

 responds to 15 C. Corrections for slight deviations from the 

 standard temperature are marked on each side of the scale of the 

 thermometer. 



SAMPLING. 



No part of the work requires more care than the taking of such 

 a sample of milk as shall truly represent the bulk. The mere 

 pouring of the milk into the weighing vessel does not ensure 

 thorough mixing. Prior to mixing, however, any particles of 

 fat which may have separated out in transit should be removed. 

 This done, the milk in each can should be agitated by means of 

 a plunger kept for the purpose on a rack alongside the platform. 

 The contents of the cans are then poured into the weighing 

 machine, and the sample immediately taken from the bulk. 

 When the amount of milk delivered by one supplier exceeds the 

 capacity of the weighing machine, a proportionate quantity 

 should be taken out of each can (after using the plunger) and 

 poured into a separate vessel, from which, after a further mix- 

 ing, the final sample is taken. 



The difficulty arising from clots of butter formed during 

 transit might to some extent be obviated if suppliers would 

 stir up any milk that has stood for a considerable time at the 

 farm before despatch to the creamery. 



The practice of paying for milk on the basis of a test made 

 at irregular intervals is not satisfactory. The proper method is 

 to take daily samples of each delivery and to let these accumu- 

 late for a fortnight or month, and to then test the composite 

 sample. This method has all the advantages of daily testing 

 and effects a considerable saving of labour. Moreover, each lot 

 of rich milk and of poor milk is represented in the composite 

 sample. For working composite sampling the following articles 

 should be provided : 



1. Two sets of bottles, of 8 oz. or 16 oz. capacity, according as 

 the testing is to be done fortnightly or monthly. 



2. A lock-up rack, in which the bottles are kept. 



3. A -oz. tinned steel dipper, to take the samples. 



The need for two sets of bottles arises from the fact that one 

 is required for collecting, while the other set is in use in the 

 testing room. The bottles should be of strong glass, without 

 shoulders, and should be furnished with loose glass covers. The 

 mouth should be 1^ inches, and the bottom 2 inches in 

 diameter. Ground glass stoppers, or stoppers with cork bands, 

 have proved unsatisfactory. The bottles must be securely 



