92 THIRTIETH ANNUAL REPORT OF THE 



and then at once dip out approximately a gill (a gill cup on a 

 long handle works well — a small teacup will do) and pour into 

 the jar. Close the jar and keep it closed until the next milk- 

 ing. Proceed thus with each cow. At the next milking re- 

 peat the operation, adding a second gill of recently poured 

 milk from the first cow to the gill taken at the first milking, 

 and similarly with the other cows. Proceed thus for from 

 four to eight successive milking^ keeping the jar closed ex- 

 cept when putting in the milk. This makes what is known as 

 the composite sample, one which is much more trustworthy 

 than a sample taken from a single milking. If samples are 

 to' be transported, the last sub-sample of each composite 

 sample taken should be made to fill the jar absolutely full to 

 prevent churning on the way. 



Cows vary considerably in the quality of their milk at differ- 

 ent stages of lactation. If only infrequent samples are taken, 

 most nearly accurate results (that is, such as will most closely 

 indicate the average quality for the year) will be usually ob- 

 tained if samples are taken approximately as follows: 



Cows calving in the spring: One composite sample six 

 weeks and another six and a half to seven and a half months 

 after calving; or two composite samples, taken about two 

 weeks apart, six months after calving. 



Cows calving in the summer: One composite sample eight 

 weeks and another six to seven months after calving; or two 

 composite samples, taken about two weeks apart, from three 

 to five months after calving. 



Cows calving in the fall : One composite sample eight to 

 ten weeks and one five and a half to seven months after calv- 

 ing; or two composite samples, taken about two weeks apart, 

 from five to seven months after calving. 



Samples taken at other times may give satisfactory results. 

 Prolonged experience has shown, however, that greater like- 

 lihood of getting a correct average for the year is attained by 

 sampling at these times. 



II. To test the entire dairy as a whole — Prepare a fruit 

 jar as under I. If the churn will hold the entire milking, 

 pour it in and slowly revolve the churn for a couple of min- 

 utes, then draw out, taking a gill soon after starting the milk 

 out of the gate. Repeat for several milkings, as under I. 



If the milking is too big for the churn, pour the milk in 

 each large can three or more times back and forth and after 

 the last pouring of each can dip out at once a gill into a sec- 

 ond jar. Having gills from each can united in the jar, pour 

 these not less than three times. Take one gill and put into 

 jar as under I. The stirring method of sampling from large 

 cans should not be resorted to unless neither of those cited 



