With Yohkoh SXT and then SOHO/EIT and TRACE, we got our first good looks at motions of hot plasma in the low corona, both at the limb and on the disk. Thanks to this, we are learning a great deal about how CMEs begin. Best of all, we have the LASCO observations to "calibrate" these and many other relevant non-coronagraphic signatures.
Does one scenario explain all CMEs?