AGRICULTURAL CHEMISTRY AGROTECHNY. 209 



tion was principally to determine the efficiency of the various methods of 

 sampling in use and their relation to actual creamery practice. 



The kinds of samples taken wei'e individual, ordinary composite (obtaining a 

 sample consisting of a representative amount of cream from each delivery of a 

 single patron for 15 days), and proportionate (samples taken in proportion to 

 the amount of cream received at the creamery). For the whole experiment 40 

 per cent 18 gm. bottles, graduated to A per cent, were used. The samples were 

 weighed on a scale of the creamery torsion type. 



The duplicates obtained when testing the cream by the same or different 

 testers were only reasonably variable. " In samples reported 87.7 per cent 

 were exact duplicates or checked within 4 per cent, 9.8 per cent varied 1 per 

 cent from the individual, while 22 per cent showed a variation of more than 1 

 per cent." A greater latitude of variation between comiX)site and individual 

 samples was noted than between 2 sets of com|)osites. The variation, however, 

 was evenly divided. The comiwsites of one of the authors (Hepburn) showed 

 tests of 36.4 per cent above and 41.99 per cent below the individual samples. 



"Two series of composites taken in the same manner showed the following 

 tendency for variation: 03.73 per cent were exact duplicates, or varied but i 

 per cent; 17.G per cent varied 1 per cent; 6.33 iter cent varied 1.5 per cent; 

 7.52 varied 2 per cent ; 2.1 iier cent varied 2.5 per cent ; while 2.8 per cent varied 

 more than 2.5 per cent. During the winter months 41.87 iK'r cent of the com- 

 posites tested lower than the individual against 35.38 i)er cent testing higher. 

 During the summer months 33.91 per cent tested lower against 43.16 testing 

 higher." 



The results of long-time averages, that is, during a period of 6 months, show 

 that one of the testers (Lee) had only 15.7 per cent of the samples vary more 

 than i per cent, while with the other tester (Hepburn) the variation was 15.9 

 per cent. *' Under the proportionate system, for a period of 6 months, 24.75 

 per cent of the sami)les varied more than 4 per cent." When based on the 

 period of 1 year " only 5.95 per cent of Lee samples varied more than i iier cent, 

 while during the same length of time 7.78 per cent of Hepburn's varied more 

 than i per cent. Results from the yearly average butter fat show 14.52 per cent 

 of composites below the individual and 7.7 above by Lee, and by Hepburn 15.34 

 per cent below and 7.78 above. These results are closer than the results ob- 

 tained by duplicate testing." 



All of the above results were corroborated from the total weights of butter 

 fat, which showed that the variation between the individual and composite 

 samples was 0.27 (Lee samples) and 0.16 (Hepbui'n samples) percent. Accord- 

 ing to these findings, reasonable allowances should be made for the difCercnces 

 in duplicate samples obtained by the same or different testers. 



The aldehyde figure of butter, E. H. Miller (Ajialyst, 37 (1912), No. ^31, 

 pp. 50, 51). — Formaldehyde combines with milk proteins to form additive com- 

 pounds, pi-obably of the type R<^5;?^r' ■^^I'ich increase the acidity of the 



milk or cream in proportion to the total nitrogen present. This fact has been 

 employed successfully by Richmond (E. S. R., 18, p. 8; 22, p. 309) for estimat- 

 ing the proteins in milk, and the author sought to apply it to the protein 

 content of butter. The results obtained with butter 1 month old show that it 

 can be employed for this purpose. 



The method used in the investigation Avas as follows : "Approximately 10 

 gm. of butter is weighed into a tared beaker, which is placed in a water 

 bath at GO to 70° C, until the butter is completely melted. Twenty-five cc. 

 of water at about 65° is added, and followed by 1 cc. of a 0.5 per cent solution 

 of phenolphthalein. The contents are well agitated. Approximately twentieth- 



