Question is below: need urgent help

Doctors studying how the human body assimilates medication inject some patients with penicillin, and then monitor the concentration of the drug (in units/cc) in the patients' blood for several hours. The data are shown in the table. Time Elapsed Concentration (Hours) (Units/cc) AWNH 42 28 19 13 vau 9 6 A First they tried to fit a LINEAR model. The regression analysis is shown. Regression of Concentration on Time Elapsed Dependent Variable: Concentration R-squared = 90.05% Variable Coeff S.E. of Coeff Constant 41.2857 2.1602 Time -6 1.8766 Not being satisfied with those results the researchers try a new model, using the reexpression log(Concentration). Examine the regression analysis below. Dependent Variable: log (Concentration) R-squared = 99.98% Variable Coefficient S.E. of Coefficient Constant 1.7891 0.045 Time -0.1688 0.010 Is this model better than the original linear model? Why or why not? Using the new nonlinear model, estimate the concentration of penicillin after 4 hours...remember the transformation used is x, logy