It has been hypothesized that letters in orthographic strings (those that follow the rules of English) are more accurately identified than letters in nonorthographic strings because confusions about the positions of letters are more likely in nonorthographic strings. This hypothesis was tested by giving 16 college students a set of targets, one of which appeared in each three-letter string, and having subjects report the position of the target as well as its identity after each exposure. One-half of the subjects reported the position first. Thirty-two trigrams, one-half of which were orthographic, were each shown once per block in five different test blocks given to each of the subjects. The target was the first or last letter in each trigram. Performance was found to be much better on orthographic strings than on nonorthographic strings, and most of the difference occurred when the position of the target letter was reported correctly. When the position was misreported, identity errors still were slightly more frequent in nonorthographic strings. The results suggest that position errors are not a strong determinant of the word-superiority effect. (NB)