Abstract An experiment was conducted to extend some findings reported by Hockey (1970 a, b) who employed a primary tracking and secondary watchkeeping task. He found that when signals were more probable in the centre and subjects were asked to perform a central tracking task, they missed peripheral signals while in noise. When signals were equally probable at all loci this effect did not occur. Hockey concluded that the effect reflects a funnelling of attention rather than of vision. In the present experiment similar tracking and watchkeeping tasks were employed, but for separate groups either the watchkeeping or tracking was primary and the other task secondary. In the primary tracking group there was a subgroup for which signal probability was higher in the centre and one in which there was no bias; for the primary watchkeeping group there was a subgroup for which signal probability was higher in the periphery and a subgroup for which there was no bias. In each subgroup there were 12 subjects, who performed the tasks on different days under three Noise Conditions (75 dBA background noise, 136 dB (SPL) peak periodic impulse noise, and 105 dBA recorded continuous noise). Results quite different from those reported by Hockey were obtained: there was no effect of noise on quality of monitoring, but tracking performance was impaired. Possible explanations are advanced.