Abstract
Recent research has demonstrated that when participants are required to draw from memory a standard clock with numbers represented by Roman numerals, they typically misrepresent the four as “IV” rather than as the correct “IIII.” Participants do not make this mistake when they simply copy the clock while it remains in full view. A reanalysis of the data presented by Richards, French, and Harris (1996) is reported in which we investigated the representation of the orientation of the Roman numerals on the clockface and found that participants in the memory conditions (surprise and forewarned) were more likely than participants in the copy (control) condition to draw the numerals vertically rather than in their correct centripetal orientation. In addition, we found this error was even more likely to occur for participants when they had no warning that a memory task would take place.