While tying a new project into published control points yesterday I ran into a odd situation. My intent was to tie the project to control points close to my site which are owned by the local county. I found two of the three points destroyed. Looking for a quick way to verify the remaining control point, I located two TDOT gps control points located along the state highway in front of my project. When I returned to my office, I pulled the data sheets for the two TDOT points and used their published values (state plane) to translate and rotate my data file. The horizontal and vertical error between the TDOT points was about .01'. I then created a point at the published coordinates of the local county control point and inversed between it and the shot position. The horizontal error was 127 feet. The coordinates of the local control point are only given in state plane coordinates NAD83(90). The TDOT data sheets give lats and longs in NAD83(95) and state plane coordinates in both meters and survey feet. I ran the TDOT lats and longs through Corpscon and converted to state plane in meters and survey feet and found that the resulting coordinates do not match the published info. I suspect that TDOT's lats and longs are correct, but that they've used a custom transformation to reference the gps control points which were set in 2002 to the original state plane coordinates on the ROW plans which date to the 1960's. I guess that's why they have a big disclaimer on their data sheets that basically says "use at your own risk".