Measurements of the solar wind velocity have been compared at two widely separated locations using plasma data obtained in mid-1967 with Mariner 5, en route to Venus, and with the near-earth satellites Explorer 33, 34, and 35. A previously study of the propagation of interplanetary sector boundaries between Mariner and earth had implied the existence of a large-scale gradient in the velocity which we interpreted as a latitude gradient of approximately 13 km/s per degree of latitude. The present investigation extended this earlier study to the overall solar wind without regard to the presence of sector boundaries. Over 2000 hourly averages of the solar wind velocity which originated from essentially the same solar longitudes were compared at Mariner and earth. When Mariner 5 was still near earth, the velocity measurements were found to agree to within 7 km/s (≲2%). However, as the separation between Mariner and earth increased in radial distance, longitude, and latitude, the Explorer velocities became systematically larger than the corresponding Mariner velocities. By plotting both velocities as a function of the heliographic longitude at the instant that the solar wind passed the heliocentric distance of Mariner 5, it was possible to observe directly the systematic velocity differences. Several regression analyses of the Mariner and Explorer velocities were carried out to obtain a quantitative measure of the velocity differences. In addition to the 1-hour averages, 1-, 2-, and 4-day averages were analyzed to allow for the possible effect of persistence (i.e., a nonzero autocorrelation at lags of ?1 day). All of these analyses yielded statistically significant departures which implied higher velocities at earth than at Mariner. In 4 months these velocity differences had increased on the average to 80 km/s (?20%). The regression analysis results are in excellent agreement with the results obtained from the sector boundary propagatin study. This agreement shows conclusively that the gradient cannot be attributed to stream-stream interactions or other possible acceleration mechanisms operating in the vicinity of sector boundaries. It is evident that a large-scale velocity gradient existed during this interval and was a general property of the solar wind. The quantitative agreement between the two studies implies that the most plausible interpretation of the gradient is that the solar wind velocity is latitude dependent.