Four undergraduate classes were randomly selected from a southeastern university; two classes served as experimental groups and two classes served as controls. A male actor played the part of a disgruntled student who was upset with the professor over a failing grade. He was in front of all classes for approximately 30 seconds. Control groups witnessed the incident with the actor wearing all black clothing, while experimental groups saw the actor in colored clothing. Students were immediately debriefed and questioned after the incident. Findings revealed that the presence of color cues on the suspect in the eyewitness situation reduced the accuracy of suspect descriptions. Significant differences occurred in accuracy between experimental and control groups, with controls being more accurate. Overall performance (experimental and control groups combined) and within-group performance differed significantly on the three types of questionnaire variables (color, detail, and descriptive). Accuracy was also affected by student grade point average, seating position, and gender. Study limitations are noted, and research on eyewitness identification of suspects is reviewed. 44 references, 3 notes, and 5 tables