Constraints on the gravitational constant at large distances

D. R. Long (1976) and others have speculated that the gravitational force between point masses in the Newtonian regime might not be exactly proportional to 1/r squared. Distance-dependent deviations from the 1/r squared law can be represented by a distance-dependent gravitational 'constant', G(r). Long has summarized the experimental evidence which constrains G(r) to be very nearly constant for 5 cm less than r less than 1 m. This paper presents observational evidence for constancy in the range 1000 km less than r less than 10 to the eighth km, and points out that the value of G(r) identical to Gc in this range has not been experimentally determined. Constraints on G(r) in the intermediate distance range 10 m less than r less than 1 km are so poor that one cannot rule out the possibility that Gc differs greatly from the laboratory value Go. Models of the earth and sun are used to argue that Gc differs from Go by not more than approximately 40%. Methods of improving the determination of Gc are suggested.