>o 



Sample n — Taken from the middle of D at a time when the 

 wagon had stood still for four minutes with the dipper resting on 

 the bottom of the can. Time, 8.15, and one mile traveled since 

 10 was taken. 



Sample 12 — Taken from the bottom of D. 



It will be noted that in three of the cans, A, C, and D, that the 

 samples taken are practically identical in quality, and that while 

 there is considerable variation in the samples taken from can B, 

 much the richest one of all was taken from the bottom of the can. 



That there might be no question as to the results, two more 

 trials were made. The samples were taken under the same gen- 

 eral conditions as in the previous trial, with the extra precau- 

 tion to take duplicate samples, marked a, and b, Of these as in 

 the other case, parallel determinations were made, and the figures 

 as before show the averages of the parallels in per cent, of fat : 



SECOND TRIAL. 



The results of the duplicates in samples 2, and 3, were unfortu- 

 nately lost during the course of analysis. Sample 1 was taken 

 from the top of the can after the milk had been thoroughly 

 mixed. Sample 5 was taken at the close of the route and the 

 few quarts remaining were thoroughly mixed before the samples 

 were taken. The other samples were taken at about equal in- 

 tervals. The variations in the results both as to samples and 

 duplicates fall entirely within the limits of error of chemical 

 analysis. Indeed, there was often as much variation in the par- 

 allel determinations of a and of b, as between a and b them- 

 selves, and the table shows that there is in general no more varia- 

 tion between samples, than between a and b. 



In the third trial the samples were taken, and the analyses 

 made exactly as in the second, the results are given below : 



