Tensile tests were performed on copper single crystals of various orientations at low temperatures. It was observed that the formation of twins was always preceded by the overshooting of the crystal axis. Also it was observed that just before twinning conjugate slip occurred. In both unirradiated and neutron irradiated crystals the twin plane was always parallel to the conjugate slip plane. Electron microscope transmission studies show prior to the twinning a dislocation distribution which is typical for single glide. This dislocation arrangement is not altered appreciably by the twinning. At least in the case of copper, all observations seem to confirm the twinning model proposed by Venables which is discussed in detail.