The purpose of this study was twofold: 1) to replicate an experiment in the research of Rhodes (1987) which found that reaction times (RT) necessary to localize a sound increased linearly up to 90 degree eccentricities from the focus of attention; and, 2) to test the hypothesis that if shifting auditory attention is a resource dependent, serial process, then a measure of capacity--the task-evoked pupillary response (TEPR)--should reflect the relationship between capacity and auditory attention shifts. More specifically, progressively larger TEPRs should be produced as the angle of auditory attention shift increases. Twenty undergraduates (10 of each gender) received course credit for participation. Pupil and RT data were obtained while subjects made attention shifts in 45 degree increments from 0--180 degrees. Results from the TEPR data did not support that hypothesis; however, a very unusual but highly significant interaction between pupil diameter and direction of speaker numbering (clockwise vs. counterclockwise) was found. Nonetheless, Rhodes' (1987) findings in terms of RT were replicated and seem to represent a quite robust phenomenon.