One of the principal aims of the Lick proper motion program, as conceived and initiated by Wright (1950), was to derive the correction to precession. Ideally, proper motions of stars from a fundamental catalogue should be measured with respect to galaxies. Unfortunately, these stars are too bright for a direct reference to faint galaxies, even with an objective grating and two systems of exposures (2 h and 1 min) on every plate, as employed at Lick. For this reason a cooperative project with the U.S. Naval Observatory was initiated in 1953 (Scott, 1954; Vasilevskis, 1954), with an intention to establish a direct relationship between meridian circle and Lick observations. When Heckmann (1954) proposed the formation of the AGK3, it was agreed to discontinue the cooperation mentioned, so as to make the resources of the U.S. Naval Observatory available for observing the AGK3 reference stars, and then to use the AGK3 as an intermediary for relating the Lick proper motions to a fundamental system. An obvious advantage of this change was offered by the abundance of AGK stars for measurement on Lick plates; a disadvantage is the absence of the AGK3 data south of declination –2.°5.