The free-field speech-reception threshold (SRT) for sentences was investigated in quiet and under nine conditions involving noise or competing speech for a group of 20 elderly subjects (ten male, age 75-85; ten female, age 76-88) and a reference group of ten young normal-hearing subjects. The noise source had the same long-term average spectrum as the competing speech. The interfering signals were presented at a constant level of 55 dBA. All elderly subjects had moderate, nearly symmetrical pure-tone hearing losses with an average loss at 500, 1000, and 2000 Hz of between 9 and 40 dB re: ISO-389. The main results are (1) the SRT values in noise and competing speech are about equal, whereas the normal-hearing subjects showed a lower SRT (7 dB lower for the condition that both sound sources are in front) in competing speech than in noise; apparently, the elderly subjects do not benefit from the relatively silent periods in competing speech; (2) the gain obtained by moving the interfering noise source from the front to the lateral position is only 2.5 dB, in contrast to a gain of 9.6 dB for the young subjects; apparently, the elderly are unable to make full use of the spatial divergence between primary speaker and noise source.