COMPILATION METHODSSteps to improve quality during the compilation of existing gravity data were to delete duplicate data, to correct or delete erroneous data, and to evaluate the accuracy of accepted data. Duplicate data could not be easily identify when previous compilers modified values of observed gravity, modified locations, or copied earlier data into their sets and then substituted different station names. The pattern of two sets of sequence numbers at associated data points ofter could be recognized except in cases where a later compiler modified locations or sorted data points by their latitude and then replaced original station numbers with numerical sequence numbers. Unless a less accurate data point is close! ' surrounded by data points of a less accurate data set or the data set with generally higher accuracy apparently cipied the less accurate data point, the less accurate data point was discarded.Data points with the same station names as other data points or without a station name, which occurred in subsets of data from the National Geophysical Data Center (table 2), were assigned unique station names. Data points located near spot elevations or bench marks, near cultural features such as roads or marked section corners, or arranged along regularly-spaced lines or grids generally were accepted with minor modifications. The last method for identifying data points to be rejected, corrected, or accepted was to look for conspicuous disagreements (forming one-station "bnllseye"