In an attempt to test the effectiveness of sounds in deterring harbor porpoises from nets and reducing porpoise bycatch in gill net fisheries, two harbor porpoises, kept in a large floating pen at Neeltje Jans, The Netherlands, were subjected to 3 different underwater sounds. The effect of each sound was judged by comparing the animals' behavior during a 15-min test period with that during a 15-min baseline period immediately before the test and a 15-min recovery period immediately after the test. The effects of the alarms were quantified as the distance between the porpoises' surfacings and the alarm and the animals' respiration rates. Each alarm was tested in two positions in the pen. The behavior observed was related to the sound-pressure-level distribution in the pen. All three alarms: the standard Dukane alarm (a commercially available alarm with a regular pulse interval of 4.3 sec used to deter dolphins from fishing nets), the random Dukane alarm (the same alarm with random pulse interval of between 2 and 30 sec), and the “bird alarm” (a sound from a generator) resulted in increases in both the distance of the animals' surfacings from the alarms and their respiration rates. The standard Dukane alarm and the bird alarm were more effective than the random Dukane alarm in inducing the animals to swim away from the sound source.