AbstractThe quality of a series of objective analyses of synoptic charts has been examined and it has been found that statistically the discrepancies from manual analyses are of the same order as the variations which arise amongst conventional analyses by different individuals. A systematic fault in the computed analyses of surface maps (viz. an underestimation of the depth of intense depressions over the ocean) was found to be due to the use of the geostrophic approximation and to other limitations of the interpolation system. Modifications designed to overcome this shortcoming have raised the standard of the analysis of depressions to that of the analyses as a whole.