The Schreckstoff-induced fright reaction (shift in fish depth distribution) was studied in roach from a European lowland lake in two twin vertical tubes of 0.85 m diameter and 11.5 m depth (Plon Plankton Towers) in an attempt to find possible reasons for the weakness of roach responses to the alarm substance in the field, as compared to those in laboratory conditions. Roach (fork length 15.4 +/- 1.6 cm), fed Daphnia and chironomids under diel 9/15 h light/dark periodicity, displayed a transitory but firm response to the alarm-substance treatment in one of the towers, while no change was seen in the reference one. The mean depth of fish within I h of dusk following Daphnia supply to the towers' epilimnia (0-2.5 m depth), as well as 3 h later in the dark, was significantly greater in the treatment than in the reference tower. The response was more pronounced in prey (Daphnia) density than in the depth distribution of roach. The indirect effects on Daphnia were extremely strong (Daphnia many-fold more abundant in the tower with alarm substance than in the reference one), but shortlasting. Frightened fish were more reluctant to feed than the reference fish in the daylight, but the initial evening difference in Daphnia abundance vanished overnight due to feeding in the dark.
