434 ANNUAL REPORTS OF DEPARTMENT OF AGRICULTURE. 



use of milk ferments and kindred preparations is a subject now under 

 consideration. 



Much valuable information has been gained from the inspection of 

 various factories in different sections of the country where foods are 

 handled and prepared for human consumption. Conditions in milk 

 depots, ice-cream factories, bakeries, candy kitchens, hotels, restau- 

 rants, boats, dining cars, etc., have been observed. 



The number and nature of the samples examined is shown by the 

 following tabulation of interstate samples and the statement as to 

 research work : 



Butter 12 



Ketchup and ketchup material 65 



Corn meals 18 



Cream, raw and pasteurized 158 



Cream puffs 18 



Eggs: 



Desiccated 50 



Frozen 50 



Figs 10 



Flour. 34 



Ice 21 



Ice cream 42 



Milk : 



Raw and pasteurized 756 



Fermented-milk preparations 11 



Miscellaneous 57 



Oysters 69 



Soft drinks 40 



Water : 



Bottled 89 



All others 22 



Total 1, 522 



The research samples included 92 ciders, 209 samples of desiccated 

 and frozen eggs, 145 samples of oysters, 287 samples of water from 

 various sources, 115 samples of figs, and other miscellaneous materials 

 aggregating 1,068, a total of 2,590 samples examined during the year. 



SPECIAL INVESTIGATIONS OF THE DIVISION OF FOODS. 



Determination of colors in foods. — Under the direction of Dr. 

 B. C. Hesse, collaborating expert, New York City, a comprehensive 

 study is in progress by A. M. Doyle, of the Washington Food Inspec- 

 tion Laboratory, looking to the classification of food colors and the 

 construction of analytical trees for their identification. This investi- 

 gation is extensive and intricate in the extreme and will require some 

 time for its completion. The construction of the trees for greens, 

 violets, and browns alone, which has been practically completed, in- 

 volved the making of about 11,000 fundamental observations on 284 

 samples, in addition to which many hundred tests are made to assure 

 the accurac}'^ and reliability of the trees after their construction. 

 These trees may be relied upon for concentrations ranging from 1 to 

 1,000 to 1 to 5,000 ; either side of this range their value has not been 

 established. As far as the work has gone it does not appear that the 

 impuritisa accompanying different makes of the same Green Table 

 number interfere at all with the use of the trees; the only difficulty 

 observed in this respect was in distinguishing a clean penta methyl 



