8o MAINE AGRICULTURAL EXPERIMENT STATION. 



the reason that the percentage of fat is subject to considerable 

 variation from day to day. 



To get a reliable result, at least 6 consecutive milkings should 

 be tested. This is most easily done by making a composite 

 sample as follows. After the milk is drawn from the udder, turn 

 it into a deep and narrow can or pail, then lower the sampling 

 tube, described on page 65 and take out one tube full. Have 

 at hand a half pint fruit jar labeled with the cow's name or num- 

 ber, and run the contents of the tube into it, closing the jar 

 tightly to prevent evaporation. Proceed in the same manner 

 each time for 6 consecutive milkings. Care must be used to 

 lower the sampling tube with the lower end open to the bottom 

 of the pail, so as to secure a column of milk that will represent 

 the whole. By this method a proportional part of each milking 

 is taken and a very accurate sample obtained.* A small piece 

 of bichromate of potash, about the size of a pea bean, should be 

 put in the jar and dissolved in the milk, to prevent its souring 

 in warm weather. After the last tube full has been put in the 

 jar and the whole thoroughly mixed, the composite sample is 

 complete and can be tested. The percentage of fat found will 

 be a fair representation of the quality of milk the cow is giving. 



If one wishes to learn the true value of a cow as a butter pro- 

 ducer, it is necessary to know the quantity of butter fat given. 

 This can be estimated approximately by testing the cow every 

 two or three months and keeping a record of the number of 

 pounds of milk she gives. To get an accurate result, the milk 

 should be tested each month, for as the period of lactation 

 advances, the quantity of milk decreases, while the percentage 

 of fat increases. With a fat test of the composite sample and a 

 record of the number of pounds of milk given for each month, 

 the total fat yield is very simply calculated by multiplying the 

 per cent of fat found by the number of pounds of milk for that 

 month. The yield for the whole year will be the sum of these 

 monthly products. If several cows are being tested, a sample 

 jar, properly labeled, should be supplied for each one. 



A very convenient form of keeping records is shown on page 

 82, and explains itself. Fools-cap paper can be ruled oft to 



*In the absence of a sampling tube tie sample can be obtained by pouring the 

 freshly dra^mi milk from ore vessel to another Three or four times and quickly 

 dipping up a small cup or ladle full, and pouring this into the jar. 



