In early 1998 the Ulysses spacecraft made history by completing an orbit around the Sun inclined to the solar equator by 80¿. It observed the properties of the high-latitude heliosphere at a time of low solar activity when the configuration of the solar magnetic field was relatively simple. The solar wind data acquired by Ulysses have placed firm observational constraints on theories of the acceleration of the fast wind from the polar regions. The speed of the polar solar wind was in the range 750--800 km s-1, with a slow increase toward the poles. The polar proton flux of ~2¿108 cm-2 s-1 was only two-thirds the low-latitude flux but still sufficiently high to require energy sources in addition to the conduction of heat from the hot solar corona. In comparison with the slow, low-latitude solar wind, the heavy ions in the fast, high-latitude wind had less elemental fractionation relative to the solar surface and a lower state of ionization. Fine structures in the polar solar wind, named microstreams, exhibited a correlation between speed, proton temperature, and alpha particle abundance. The absence of a latitude dependence of the widths of the microstreams suggests that they are caused by temporal, rather than purely spatial variations at the Sun. Magnetic field observations revealed a latitude-independent radial field strength of ~3 nT (normalized to 1 AU) and a large flux of outward propagating Alfv¿n waves. The most probable direction of the interplanetary magnetic field was close to the Archimedes (Parker) spiral predicted on the basis of radial fields close to the Sun combined with the effects of solar rotation. The fluctuations about the average direction were so large, however, that there was little latitudinal variation of the average value of |BR|/B, where BR is the radial component of the field with magnitude B. Latitudinal transport of field lines and nonradial field directions near the Sun appear to be other necessary additions to the simple Parker model. Some unexpected differences were found between the low- and the high-latitude properties of the plasma from transient events called coronal mass ejections (CMEs). At high latitudes, the speeds of the CME plasma clouds were close to the speed of the ambient fast wind, shocks were observed both ahead of and behind the CMEs, and enhancements of the helium abundance and of high charge states of heavy ions were not observed, as they often are at low latitudes. The review closes with a summary of arguments in support of the view that the fast solar wind observed by Ulysses at intermediate latitudes has its origin at higher latitudes in the polar coronal holes. ¿ 1999 American Geophysical Union