This is a question about the geoid and geodesy. It would seem that geologist G K Gilbert and glaciologist T C Chamberlin asked Robert Woodward to look into Gilbert's observation of shorelines at varying heights on Lake Bonneville. Their interest was in whether it was feasible for the shorelines at the two points of interest to be contemporaneous, which would imply a sloping lake surface. They wondered whether the variation in height might be due to the gravitational attraction of an overlying glacier.
In response, Woodward formulated a general procedure for calculating gravity and geoid "anomalies" which was a landmark paper of its time and is consistent with modern methods. This theory is one of several achievements for which Woodward is noted. Looking at his biography, I see no suggestion that he confirmed the geologists' conjecture, and can confirm that the gravitational attraction due to a glacier would not suffice to account for such a substantial variation in elevation. The simplest explanation of the geologists' observation is that the shorelines correspond to different times in the development of the lake. There are a number of ways (perhaps, for example, using isotope geochemistry) that we might set about confirming that today. See the biography at http://www.nasonline.org/publications/biographical-memoirs/memoir-pdfs/woodward-robert-s.pdf
These days we do know that the height of the geoid (more or less, the natural level of the sea surface) varies quite substantially – probably in the region of −106 to +85 m - but over distances of a few thousand km rather than the few km of Lake Bonneville. This is due to large-scale variations in the density of the Earth. There are plenty of maps on the WWW such as at http://earth-info.nga.mil/GandG/wgs84/gravitymod/egm2008/egm08_gis.html. The slopes represented on these maps (which a still, tideless, water surface would adopt) are of order 1:10 million. Shorter-wavelength variations are also present over features such as the deep trenches in the western Pacific.
I cannot confirm the figure, but I would be surprised if the maximum known slope of the geoid were as steep as 1:100,000. That's still a thousand times too small to confirm Gilbert's conjecture.’