The vertical distribution of the phototrophic ciliate Mesodinium rubrum (= Myrionecta rubra) was examined under relatively stable conditions in Inre Verkviken, a brackish inlet on the h a n d Islands, Finland, in early summer 1991. Distribution was characterized by a population maximum situated persistently within the thermocline (8 to 11 m depth) just above the nitracline, and at an irradiance (-1 to 10 pm01 photons m-' S-') below the estimated irradiance5 required either for photosynthetic

more »

... uration (Ik = 275 pm01 photons m-' S-') or even compensation (I, = 15 pm01 photons m-2 S-'). This population maximum often co-occurred with a second deeper maximum at around 14 to 16 m, with transient evidence of a third poorly defined peak within the upper 5 m. Maximum cell numbers reached over 200 cells rnl-' at 15 m, but were more typically observed at around 50 to 100 cells ml-l within the 2 lower population maxima. Cells were usually present at all depths down to the anoxic boundary (15 to 16 m), but were occasionally almost absent from the surface few metres. Total depth integrated population (over 0 to 20 m) was of the order 550 X 106 to 900 X 106 cells m-', only a factor of about 2 lower than typical red-water integrated abundances. Despite the distinct vertical separation between irradiance and dissolved nutrients, no consistent diel pattern of vertical migration was discerned. Apparent movements of maxima were observed but were not associated with diel changes in the depth of either Ik or I,. Given the exceptional swimming speed and photosynthetic efficiency of M. rubrum, and these complicated changes in distribution, a diel migration pattern may not be necessary, and we speculate that migration might occur on an 'individual needs' basis However, this migration potential was a constant logistical complication hampering temporal and spatial sampling design, and artefactual changes in distribution cannot be ruled out. Improved future understanding of the migration of this species will only be realized with a much more dynamic approach to sampling in order to minimize the possibility that migration might be quicker than data collection.