Cumulative effects of a two‐second burst of thermal noise on the threshold shift for 35‐millisecond tone pips were examined by tracing the threshold shift over the course of 100 successive noise bursts. When bursts were presented at a rate of 6 per minute, a systematic increase in the threshold shift for the 35‐millisecond tone pip was observed over the 100 burst series. Changing the burst‐pip interval from 90 to 445 to 1000 milliseconds had little effect on the degree of this accumulation, but changing the SPL of the burst had a marked effect. At a burst level of 110 db (SPL re 0.0002 microbar) the mean threshold shift increased by about 6 db between the 15th and the 100th burst. Decreasing the burst level to 100 db reduced the mean accumulation to about 4 db, and at a burst level of 90 db to about 1 db. Changing the burst presentation rate from 6 per minute to 3 per minute reduced, but did not eliminate, the cumulative effect. The degree of this cumulative phenomenon showed wide individual differences among subjects.It is concluded that, unless proper allowances are made, the repetitive stimulation required by burst‐pip audiometry may be a source of serious error. On the other hand, the reaction of an ear to repeated stimuli may prove to be a useful index of the ear's susceptibility to irreversible acoustic trauma.