ABSTRACT. The visibility of herring Clupea harengus and mysids Praunus flexuosus was studied in the shallow water of the Kiel Fjord (Baltic Sea). The apparent brightness contrast of the predator (hernng) and its prey (mysid) were measured from various visual angles with a n underwater video system. The camera looked at the prey from the predator's point of view and vice versa. During a head-on encounter, the contrast of the prey against the ocean background was very poor. With increasing visual angles (relative to a horizontal plane), the visibility of the prey unproved. The predator was most visible from below, but was excellently camouflaged when observed from above or from a horizontal line of vision. When the herring was located 30 to 90' below the mysid, the prey was fairly visible to the predator but the predator was nearly invisible to the prey. By incorporating these ocean optics and taking into account the limits of the maneuvering potential of cruis~ng herring, a n optimal attack angle can be postulated: a t oblique predator-prey configurations between 30 and 60°, herring are still capable of modulating their normal cruising to a n upward directed position while the preconditions of a successful attack -to see and not to be seen -are fulfilled. In Kiel Aquarium, adult herring attacked mysids from a n average angle of 48" from below. In sltu measurements from inside a feeding school of juvenile herring showed a mean attack angle of 43"