86 THIRTIETH ANNUAL REPORT OF THE 



accurate sample and it may not. Fresh milk, not creamed, 

 well aerated and stirred, carted over rough roads and drawn 

 from cows not giving large fat globules, may be accurately 

 sampled thus with a minimum amount of stirring. On the 

 contrary, milk which has creamed, which is a day or more 

 old, from Jersey or Guernsey cows, but slightly shaken in 

 transportation if, in considerable quantity, cannot be mixed 

 with sufficient thoroughness to insure accurate sampling by 

 superficial stirring. The Vermont station several years ago 

 did much work in investigating methods of milk sampling, as a 

 result of which we are prepared to say with a fair degree of 

 assurance, that when five hundred pounds of milk somewhat 

 creamed is delivered at the factory, there is no surety of the 

 accuracy of the sample taken therefrom by the dipper meth- 

 od, unless it be stirred for from two to four minutes, round 

 and round and up and down. Hence it is wise to consider the 

 advisibility of choosing some method which is more likely 

 than this one to insure accurate sampling. 

 Jr While there is no method of sampling which is not open to 

 defeat through improper handling, there are methods wherein 

 there is a greater proportion of automatic action than in the 

 one just considered. The coring method is one of these. Sever- 

 al devices designed to core milk are used. The Scovell sampler, 

 which was used in the world's fair tests in 1893, is a fair type 

 of this class of implement. It consists of a small brass tube 

 with a perforated sliding cap at the bottom. It is lowered into 

 the milk slowly so that it will flow into the tube until it 

 strikes the bottom, when the perforated cap slides over and 

 closes the tube, thus procuring a core of milk. This method 

 of sampling, provided the cream is not separated in clots and 

 the milk is neither loppered or frozen, will take a correct sam- 

 ple if carefully used. It is more likely to take a correct 

 sample than the dipper method, or, rather, is less likely to 

 take an incorrect one. 



There is, however, a method which suits me better than 

 either of these, known as the automatic method. The ap- 

 paratus for this consists of a weigh can covered by a cone- 

 shaped wire cloth or wire mesh, and some means of with- 

 drawing a small stream from the outflowing milk. This 

 small stream may be abstracted by means of a small faucet, 

 set at the bottom of the can near the outlet gate, or by means 

 of a hole punched in the conductor head or spout. The pet- 

 cock or faucet modification of this device on the whole approves 

 itself to me rather than the other. • 



The milk being weighed, both the gate and the pet cock 

 are opened and remain open until all the milk has run out. 

 A small proportion, varying according to the size of the ori- 



