We have estimated global elevations on Titan using a combination of nadir-looking altimetry and SAR monopulse measurements acquired by the Cassini spacecraft radar on multiple Titan encounters. These data correspond to a set of one-dimensional tracks spread over much of the moon's surface, but with fewer observations in the southern hemisphere than in the north. We fit the measured points with spheres, biaxial ellipsoids, and a set of spherical harmonic functions to produce global elevation estimates.The best-fit sphere has radius 2574.95 km, but the sphere center is displaced 270 m southward so that the north polar radius is about 500 m less than the south polar radius, referenced to Titan's barycenter. A biaxial ellipsoid fit yields a polar radius of 2574.20 km and an equatorial radius of 2575.08, so that the moon appears to be slightly oblate. In this case the ellipse seems to be displaced northward by about 140 m, so both poles are still low compared to the equator but in this solution the north may be a bit further from the barycenter than the south pole. Rotation of the solution from an ideal oblate ellipsoid is negligible (<4 deg). The spherical expansion approach also gives a north pole radius several hundred meters less than the equatorial radius, and suggest that the south pole radius is comparable to the northern value, but it has larger uncertainty. For example, a 6th order fit yields north, equatorial, and south pole radii of 2574.35, 2574.85, and 2574.27 km, respectively, although thiese values are dependent on the selection of a parameter for a smoothing constraint. The paucity of data from the far south lead to greater uncertainty in the south polar radius and at this time it is not clear whether the south pole is closer to Titan's barycenter than is the north pole. However, the solutions for north polar radius and mean equatorial radius are robust over our analysis cases and are more trustworthy. Our early error analysis implies uncertainties in all of these values of perhaps +/-200 m. Both of our data types are limited in accuracy by knowledge of the spacecraft attitude, which affects the incidence angle of the observations. These data are consistent with the observation that liquid lakes are seen predominantly near Titan's north pole, which all of our measurements show to be depressed as compared with the equatorial bulge. The presence of the lakes may be linked with a subsurface methane table level dependent on Titan's geoid, so that lakes might be expected to appear on the surface in the lowest lying regions. Lakes seem to be less extensive at the south pole, so it will be crucial to determine the south polar radius accurately if we are to be able to distinguish among competing hypotheses for their formation.