Which regression equation is best for predicting city fuel consumption? Why?

Refer to the accompanying table, which was obtained using the data from 21 cars listed in Data Set 20 “Car Measurements” in Appendix B. The response (y) variable is CITY (fuel consumption in mi / gal). The predictor (x) variables are WT (weight in pounds), DISP (engine displacement in liters), and HWY (highway fuel consumption in mi / gal).

Data Set 20: Car Measurements 

Data are from 21 cars (first five rows shown here). The same cars are used in Data Set 19. The data are measurements from cars that have automatic transmissions. WEIGHT is car weight (lb), LENGTH is car length (inches), BRAKING is braking distance (feet) from 60 mi/h, CYLINDERS is the number of cylinders, DISPLACEMENT is the engine displacement (liters), CITY is the fuel consumption (mi/gal) for city driving conditions, HIGHWAY is the fuel consumption (mi/gal) for highway driving conditions, and GHG is a measure of greenhouse gas emissions (in tons/year, expressed as CO2 equivalents). Data are from the National Highway Traffic Safety Administration.