The present experiment was designed in order to investigate the effects of noise produced intermittently during the input stage of serially displayed visual materials (6-digit lists). These materials were to be memorized and subsequently reproduced on a specially designed keyboard. The four experimental conditions were the following: (1) quiet (Q);

(2) noise uninterrupted during the presentation of each list (CN);

(3) noise emitted simultaneously with each list element (500 ms) (ND);

(4) noise emitted in the interval (140 ms) between each element (NA).

Record was taken of the errors and omissions, of key response initiating times and of hand movement times. As an additional variable, subjects were required to estimate the annoyance induced by each of the four conditions. All noise conditions yielded more errors and omissions than the Q-condition, with the highest rates occurring in the two conditions where noise was superimposed on the digits. Close analysis of the errors showed that the primacy end of the lists was the most affected by noise, whereas keying speed for correct reproductions remained relatively unaffected.

On the other hand, when the speed variable was split into a decision and a motor component, it turned out that the more a condition was judged as annoying, the more the time for response decision was delayed, while the reverse occurred for keying movement times. The latter effect could possibly be ascribed to arousal induced by the three noise conditions. It is argued that noise is liable to affect differently the various components of a task performance, either directly by interference with the intake of pertinent information, or indirectly through induced annoyance or through noise modulated arousal.

They particularly would like to thank Dr M. Coltheart for his great help in the preparation of the definitive version of the paper.