THE MILK SUPPLY OF MASSACHUSETTS 127 



Aside from the inaccuracy of tlie sales data reported by the dealers, two 

 serious statistical errors are involved in this method of computation, each 

 error involving an over-estimate of the average per capita consumption. 

 Dealers customarily sell milk in the towns or cities surrounding the city 

 in which their principal business is located and frequently report their 

 total sales as pertaining entirely to the central city. This inflates the 

 figures on milk sold within the city limits. Moreover, the milk sold in any 

 large city is consumed by a population much larger than those counted 

 b\ the census as residents. The number of those who reside outside of 

 the city, but eat one or more meals there, regularly or occasionally, is 

 very large, especially in the case of the chief cities. 



In this study every effort was made to avoid these inaccuracies. The 

 area was carefully chosen and the data on receipts and consumption were 

 carefully scrutinized. It was assumed, however, tiiat the number of visit- 

 ors entering the area was not greater than the number leaving it during the 

 year. For this reason abnormally large milk handlings of the week of 

 the Eastern States Exposition were excluded. 



With the assistance of the milk inspectors in tliese markets the amount 

 of cream for the fluid cream trade and the jimount of milk handled by 

 each dealer was ascertained. Tlie sources of this milk and cream were 

 found to be three: that produced in the local area; that which is secured 

 from elsewhere in the state; and that which is secured outside the state. 

 Special statements based directly upon their records were secured from all 

 the large dealers, and care was taken to avoid duplication of data when 

 milk was sold at wholesale. 



Data were obtained for conditions as they existed in the late fall of 

 lf^2G during which period practically no dealer had any surplus and it was 

 certain that only small and measurable quantities of local milk failed to 

 go into fluid milk trade. 



The milk and cream available for consumption included the local produc- 

 tion as well as the supplies shipped in from the outside. Local production 

 was estimated for October by multiplying the total number of cows in 

 the area, as shown by the tax assessors' data for 1925, by the average 

 milk production per cow in Hampden County, as estimated by the New 

 England Crop Reporting Service for October 1925. From this five per 

 cent was subtracted to cover wastage and use in feeding on the farms. 

 The result gave the daily supply available in October. To obtain a more 

 accurate estimate of the daily average for the year, data were secured 

 from the leading dealers showing the sales in each month of the year, and 

 this October supply was multiplied by the ratio between the average for 

 all months and the October sales. The resulting average was then divided 

 by the population of the area as shown by the state census for 1925. 



Data for fluid cream consumption were carefully secured in addition 

 to the data for milk. Such estimates as have been available hitherto for 

 cream consumption in Massachusetts appear to be even less exact than those 

 for milk. Part of the supply counted has been used for manufacturing 

 ii>to ice-cream and a considerable portion of this product has been shipped 

 tc other consuming centers. 



In this study the cream included was all used for the fluid cream trade, 

 including the supplies used by restaurants and soda fountains in the form 

 of fresh cream. The small amount sold as "light cream" was converted to 

 the equivalent of 40 per cent or "heavy" cream, which' is the ordinary 



