This paper reviews literature on the 10-plus years of Yohkoh soft X-ray observations of coronal holes (CHs), which span almost half a Hale cycle. They thus extend and generalize the results of Skylab, providing better sampling and duration. These modern X-ray data confirm the tendency towards rigid rotation of equatorward extensions of polar holes, but show no clear sign of the mechanism of magnetic reconnection that makes this possible. Coronal-hole boundary evolution does not seem to depend upon transient magnetic activity such as X-ray bright points or large-scale arcade events associated with transient coronal holes. The morphology of the coronal holes as seen in X-rays differs from that seen in He I lambda10830: X-rays generally do not do so well with the polar holes or other holes near the limb, because of foreground confusion, but they show narrow features better.