The three-dimensional structure of the solar wind is often studied by searching for solar latitude variations in observed physical properties that have been averaged over solar longitude. The effects of longitude averaging are examined here by using general qualitative arguments and quantitative calculations for a simple class of assumed spatial variations in the solar wind speed. We show that unless the solar wind spatial structure is simply organized about the solar equator, its presence is extremely difficult to infer from longitude averages unless the observations extend to very high solar latitudes. Thus the absence of semiannual variations in ecliptic observations or the absence of large latitude gradients in radio scintillation or comet tail results do not constitute evidence agaist the presence of strong spatial variations (or large spatial gradients) in the solar wind.