I have been comparing published NCDC US temperatures to the raw HCN temperatures which they are supposed to be derived from. The graph below plots the difference between NCDC published US temperatures and HCN temperatures for all stations which have been continuously active since at least 1920. Note that the 1930s have suffered about 1.4F tampering relative to 2012.
Paul Homewood has approached this from a different angle. He compared the original weather bureau documents of published temperatures to the current NCDC temperatures, and found the same thing.
I reported a few days ago how the original CONUS temperatures, as recorded State by State in 1934, were 1.23F higher than NCDC now show. The original figure of 55.37F has since been adjusted down to 54.14F. As a result, 2012, at 55.34F, is officially the hottest year on record for the CONUS. However, reader Daryl spotted that I had overstated the size of Maine. The result is that I had underestimated the original 1934 temperature by 0.16F, thus making NOAA’s adjustment even greater, at 1.38F.
This tampering is almost 3X larger than the USHCN documented tampering, and is clearly fraudulent.