Determining the Hubble constant from the gravitational lens PG 1115+080

For the quadruple gravitational lens PG 1115+080, we combine recent measurements of the time delays with new lens models to determine the Hubble constant H-0. We explore the effects of systematic uncertainties in the lens models on the estimates of H-0, and we discuss how the uncertainties can be reduced by future observations. We find that the lens cannot be fit by an isolated lens galaxy, but that it can be well fit by including a perturbation from the nearby group of galaxies. To understand the full range of systematic uncertainties, it is crucial to use an ellipsoidal galaxy and to let the group position vary. In this case, the existing constraints cannot break degeneracies in the models with respect to the profiles of the galaxy and group and to the position of the group. Combining the known time delays with a range of lens models incorporating some of the plausible systematic effects yields H-0 = 51(-13)(+14) km s(-1) Mpc(-1). The constraints on the lens models, and hence on H-0, can be improved by reducing the standard errors in the lens galaxy position from 50 mas to similar to 10 mas, reducing the the uncertainties in the time delays to similar to 0.5 days, and constraining the lens mass distribution using Hubble Space Telescope photometry and the fundamental plane. In particular, the time delay ratio r(ABC) = Delta tau(AC)/Delta tau(BA) may provide the best constraint on the mass profile of the galaxy.