roads, this would mean that 78 radial roads fan out from the Connecticut 

 towns studied. The North Carolina study reported a b value of 1.16. 

 With the same assumptions as mentioned ahove this means 10 radial 

 roads fan out from processing plants in that state. Finally, the New 

 Hampshire empirical study found a b value of 1.351. Again with the same 

 assumptions this means nearly five radial roads from Concord, which on 

 inspection seems about right. 



The a values in the three empirical equations move the intercepts 

 of the regression lines away from zero. A positive a value wovild mean 

 that fewer radial roads are available to reach farms near the center than 

 away from it. A negative a value means that more radial roads are avail- 

 able to reach farms near the center. The latter seems more realistic and 

 is true for the New Hampshire regression. 



It was decided to use the New Hampshire empirical regression for 



all distances greater than 10 radial miles. For distances less than 10 



radial miles the inclusion of any a value other than zero does not seem 



correct, especially when that a value is based on observations extending 



out 50 miles. In addition, eight radial roads were counted within a ten 



mile radius of Concord rather than the five extending out 50 miles. For 



these reasons a theoretical regression is used but is based on eight radial 



77r 

 roads. This would mean a b value of r-: — |- r or 1.19635, and an a value 



16 



of zero. At 10 radial miles the New Hampshire empirical regression 



yields 11.976 road miles while the theoretical one, assuming eight radial 



roads, yields 11.964. 



44 



