118 FOODS AND FOOD ADULTF.RANTS. 



koumiss, found its percentage of lactose to be 7.20. On the other hand, 

 ordinary mares that were kept at work gave a milk containing only 5.95 

 per cent, sugar. The quantity of milk-sugar in mare's milk is great, 

 but there is a deficiency of fat and other solids. It appears to contain 

 fully 89 per cent, water, while cow's milk does not have more than 87 

 per cent. 



The process of manufacture is not uniform. In the East the mare's 

 milk is placed in leathern vessels ; to it is added a portion of a previ- 

 ous brewing, and also a little yeast. In thirty to forty-eight hours the 

 process is complete. During this time the vessels are frequently shaken. 



In the samples analyzed by me the milk was treated with a lactic 

 ferment and yeast. After twenty-four to forty-eight hours' fermenta- 

 tion the koumiss was bottled. The bottles were kept in a cool place, 

 not above 50 P., and in a horizontal position. When shipped to me 

 they were packed in ice. After they were received in the laboratory 

 they were kept on ice until analyzed. 



METHOD OF ANALYSIS. 



Carbonic dioxide. The estimation of the carbonic dioxide was a prob- 

 lem of considerable difficulty. It was evidently impracticable to at- 

 tempt opening the bottle and determining the gas in a portion of the 

 contents. Fortunately I had access to a large balance which would turn 

 with a milligram. On this I weighed the whole bottle, into the cork 

 of which I had inserted a stop-cock such as is sometimes used with a 

 champagne bottle. With the bottle of koumiss were also weighed two 

 drying flasks containing concentrated sulphuric acid with their connec- 

 tions. 



Having obtained the weight of Iriie whole, the gas was allowed to 

 escape slowly from the stop-cock and to bubble through the sulphuric 

 acid in the washing flasks. 



These flasks, previously to being weighed, were filled with the gas 

 from an ordinary carbonic dioxide generator. After the gas had almost 

 ceased to flow the bottle of koumiss was frequently shaken. It was also 

 placed in a pail of water having a temperature of 30 C. After half 

 an hour ihe gas ceased to come over. 



The whole apparatus was again weighed. The loss of weight gave 

 the quantity of free carbonic dioxide in the sample. After the analysis 

 was completed the volume of the bottle was measured. It is fair to 

 assume that at 30 C. the koumiss still contained an equal volume of dis- 

 solved CO 2 . In determining the total CO 2 this volume, or its equivalent 

 weight, was added to that obtained by direct determination. 



By this method the OO 2 dissolved under pressure in the bottles is esti- 

 mated separately from that which the koumiss contains in solution under 

 the weight of one atmosphere. Since it is of no importance to separate \ 

 the gas into these two portions, I have given it altogether in the tables, 

 iu volume, by weight; and in percentage by weight. 



