The long and active San Andreas fault was revealed by the San Francisco earthquake of 1906. Strike-slip movement on a major crustal fracture was first established by that event. The elastic rebound theory was developed in an analysis of this earthquake. It was proposed in 1926 that cumulative horizontal movement on the San Andreas amounted to several miles, but such a great displacement was generally agreed to be unreasonable. In 1953, new evidence of cross-fault stratigraphic correlations of Pleistocene to Cretaceous rocks was presented which seemed to require tens to hundreds of miles of strike-slip displacement. Controversy and additional studies ensued, resulting in general acceptance of such movements by 1968. Since the 1965 proposal that the San Andreas is a transform fault, within a plate-tectonics mechanism, reservations about great horizontal movements of the Earth's crust have been essentially eliminated. The single most important factor in delaying acceptance of miles of strike-slip on the San Andreas has been the long-continued confusion between fault separation and fault slip. Lawson, Noble, Taliaferro, Hill and Dibblee, Wilson, and a few others played the more leading roles in interpretations of the fault. Post-earthquake studies by Gilbert again confirmed his reputation as a great geologist. The San Francisco earthquake was the chief contributor to knowledge about the San Andreas, but now there are more questions than ever regarding the nature, geologic history, and significance of this important crustal structure. The present consensus about the role of the fault in local and global tectonics surely will be modified by revolutionary new conceptual models.