The SMM coronagraph/polarimeter obtained images of the solar corona in 1980 and from 1984 to 1989. Approximately 1300 coronal mass ejections have been identified in this data set; accurate measurements of angular widths and apparent central latitudes have been made for 1209 of them. The distribution of observed angular widths is broad and slightly skewed toward large values; the average width is 47¿ (in position angle measured around the limb of the Sun), the median width is 44¿. There is no evidence in this data set for any significant or systematic change in angular widths during the epoch of SMM observations. The distribution of apparent central latitudes for all 1209 measurements is roughly symmetric about the heliographic equator, with a root-mean-square average latitude of 35¿. The latitude distributions for different calendar years show significant changes in the spread about the equator; mass ejections occurred over a wide range of latitudes at times of high solar activity but were largely confined to near-equatorial latitudes at times of low activity. For example, the root-mean-square average latitude was 41¿ in 1980, 38¿ in 1989 (both years near maxima in sunspot number) but only 13¿ in 1986 (the year of minimum sunspot number).

The changes in the distribution of mass ejection latitudes do not correspond to those for solar features or activity related to small-scale magnetic structures such as sunspots, active regions, or H&agr; flares; they do resemble those of features related to large-scale magnetic structures, such as prominences and bright coronal regions. In 1984, when the ''quiet'' or background corona suggested the presence of a magnetic dipole structure tilted at ~30¿ with respect to the solar rotation axis, mass ejection latitudes were clumped about the tilted ''heliomagnetic equator'' rather than the heliographic equator. Approximately half of the mass ejections that occurred during 1984 were preceded by several days of brightening and spreading of the bright, background corona at the mass ejection site, and produced a conspicuous disruption of the preexisting structure. These observations strengthen the arguments for a close connection between mass ejections and large-scale, closed magnetic structures in the corona. ¿ American Geophysical Union 1993