to more distant dealers, which in this 

 case are on the Boston Federal Order 

 Market (Figure 3). The plants were 

 located on a map by size groups and 

 by use of the serpentine technique 

 a sample was obtained for each size 

 group. The location of the secondary 

 market sample is shown in Figure 

 4 and the location of the sample of 

 plants shipping to the Boston Market- 

 ing Area is given in Figure 3. 



By this method the dealer sample 

 was representative by geographic lo- 

 cation and by size. As the sample was 

 taken, substitutes were drawn, to be 



used where needed, and thus to in- 

 sure that it was representative. 



All dealers with tank assembly in 

 1955 were interviewed. A sample of 

 dairy farms and truckers from these 

 tank operations was also studied by 

 means of interviews. For producers, 

 regardless of whether they used bulk 

 or can assembly, all sampling was of 

 dairy farms rather than of farmers. 

 A man with two dairy farms was 

 twice as likely to be interviewed in 

 the study as was a one-farm operator. 

 Hence the producer tables based on 

 the interviews refer to the number 



LOCATION OF RETAIL DEALER PLANTS 

 IN VERMONT, NEW HAMPSHIRE, a MAINE 



• I - 999 QTS. DAILY 

 X • 1,000-3,999 QTS. DAILY 

 ■ =4,000 QTS. a OVER DAILY 



Figure 2. 

 9 



