Sixteen years of solar wind observations via the technique of interplanetary scintillation (IPS) are presented. By an ecliptic comparison with in situ spacecraft observations, these data are shown to be valuable estimates of the large-scale slowly evolving structures in the solar wind speed, but to underestimate the speed in small-scale or rapidly evolving structures. These IPS observations allow the large structures to be studied over solar latitudes from 60¿ north to 60¿ south over more than a solar cycle. The observations are presented as half-yearly averages in Carrington longitude and latitude. These are compared with plots of coronal density estimated from coronameter observations and radial magnetic field estimated from solar magnetograph observations and the potential field method. In low and declining solar activity, there is the expected relationship between fast wind and low-density open-field regions, and between slow wind, dense corona and proximity to the neutral sheet. The dipolar field component wanders up to 30¿ from the rotation axis and is followed by dipolar distributions of density and velocity. Near solar maximum, average wind speeds are uniformly slow over the entire range of latitudes covered, the coronal density becomes more spherically distributed (though it retains a somewhat lower density over the solar poles), and the neutral sheet ranges over all latitudes aned sometimes forms disconnected surfaces. The relationship between wind speed, density and angular separation from the neutral sheet are then broken. The results confirm the controlling influence of the coronal magnetic field in determining the three-dimensional structure of the solar wind. ¿American Geophysical Union 1990