422 SECTIONAL TRANSACTIONS.— M. 



comparisons between individual farms are made within each of these groups. One 

 aim is to widen the farmer's outlook and to show him what others are doing. Under 

 this system the farm is treated as a unit, no attempt being made to allocate costs, 

 income or profits between various departments of the farm. Until lately the analyses 

 and reports have been made upon an empirical basis, but now attempts are being 

 made to use the statistical method upon the data collected during the past five years. 

 One object is to see if there is any kind of correlation between the factors, capital, 

 production, and cost. If so, then it may be possible to classify farms in a different 

 way from which they have been grouped in the past. This statistical work is still 

 in its experimental stages and, although it promises to give valuable results, it is too 

 early to be definite about anything at present. 



Considerable work has been done in ascertaining the cost of growing sugar beet 

 in Hereford and Worcester. 



Mr. F. Hirst. — Problems in connection with the Establishment of the 

 Canning Industry. 



Friday, September 5. 



Discussion (Sections K, M) on Mineral Elements in Plant Nutrition. 

 (See Section K.) 



Dr. J. F. Tocher. — The Adulteration of Milk with Water. 



The detection of the adulteration of milk by the addition of water is diflBcult 

 because apart from breed and selection, (1) milk is composed very largely of water, 



(2) the proportions of the solid constituents vary with every individual cow, and 



(3) the number of cows whose milk is bulked varies in commercial milk. It is very 

 desirable for the protection of the public that a criterion should be obtained which 

 would enable the public analyst to state definitely, within certain limits, the 

 probability that a particular sample of milk has been adulterated by the addition of 

 water. In the opinion of the writer no such criterion has yet been found. Certain 

 workers depend on the supposed constancy of the freezing point of milk in genuine 

 milk. Others on (1) the supposed constancy of the refractive index, (2) the real or 

 supposed shortage in ash, (3) the supposed minimum of sohds-not-fat, and (4) the 

 supposed minimum and low variability of total solids. In all these cases the computors 

 have not taken into account the natural variations in the proportions of the con- 

 stituents of genuine milk and the fact that the composition of milk must bear a relation- 

 to the number of cows whose milk is bulked. 



The supposed constancy of the freezing point has been specially stressed by 

 prominent workers in America and in the Dominions, but most American and 

 Dominions data are from selected cows and not from a general population of both 

 selected and unselected cows whose milk constitutes the supply for this country. 

 It is natural, therefore, that the bulked milk from selected cows should show greater 

 constancy in the values of the depression of the freezing point and in other characters 

 than the values found from commercial milk in this country. 



In a paper shortly to be published, the writer shows that the proportions of the 

 various constituents of milk in samples low in solids-not-fat are quite different from 

 the proportions of these same constituents in samples high in solids-not-fat. He 

 further shows the nature of the relationships which exist between the proportions of 

 the various constituents and finds non-linearity to prevail. 



The writer's data consist of results of analyses of over 500 samples of milk from 

 individual cows for all the constituents of milk. 



As a first approximation to a criterion the following equation was obtained : — 



Jo= - 18-41 + 5-522;, - ■28x,'^ + -bSiz^ - 4.64) - -SSix^ - 2-42) 



+ •40{a;3 - -74) - l-12(a;4 - -84) 



where J„ = predicted solids-not-fat ; x^ = solids-not-fat ; x-^ = lactose ; x-i = casein ; 

 a:, = albumin and globulin ; and x^ = ash. 



When the number of cows whose milk has been bulked is known, the difference 

 (*o — io) can be compared with S (a; —a;)- Relative differences greater than those 

 found from genuine samples show that the sample has been watered. A better 



