An assessment is made of (1) the reliability of the changes in the absolute gravity datum and standard incorporated in the IGSN 71 values of Morelli et al (1974), and (2) the effect of these changes and those incorporated in the new international gravity formula (Geodetic Reference System 1967) defining theoretical gravity values at sea level on gravity anomaly values. Pendulum-interval comparisons with modern absolute gravity intervals between Washington, D.C.; Teddington, England; Paris, France; and Potsdam, East Germany indicate that the correction of 14 mgal to the Potsdam datum is correct to within .03 + or - .016 mgal. Although there appear to be occasional discrepancies in the IGSN 71 values of the order of 0.1 mgal, in general their reliability appears to approximate + or -0.05 mgal. A discrepancy of approximately -0.03 mgal per 1000 mgal of absolute gravity is indicated, however, in the gravity standard defined by the IGSN 71 values over the range of modern absolute gravity determinations between Fairbanks, Alaska and Bogota, Colombia. An evaluation of the earlier attempt to standardize gravity on a global scale (Woollard and Rose, 1963) indicates a mean datum difference of 14.7 mgal (standard deviation 0.2 mgal) relative to IGSN 71 values at 776 sites having a worldwide distribution. No discernible difference in gravity standard from that of the IGSN 71 values is indicated except in South America, where the Woollard and Rose values are found to incorporate a difference in standard of 0.2 mgal per 1000 mgal. It is shown that anomaly conversions from the old to the new gravity system can be effected directly using the old anomaly values. Two correction terms are required, specifically, (1) a constant representing the difference in base value used relative to the IGSN 71 value for the base, and (2) a latitude-dependent term representing the difference in theoretical sea level gravity using the old International Gravity Formula (IGF) and the new GRS 67 formula. Test conversions of anomaly values on the old gravity system for regional areas involving data from several sources indicate that, in general, anomaly conversion to the new gravity system can be obtained with a reliability of about 0.3 to 0.5 mgal. The limiting factor is the unknown variation in gravimeter calibration standards represented.