ABSTRACT. Particle retention efficiency in 6 species of suspension-feeding Northeast American bivalves was determined by simultaneous measurement of clearance of different sized particles.Geukensia dernissa, Spisula solidissirna, Brachidontes exustus, and Mercenaria rnercenana, which all possess large laterofrontal cirri, completely retained particles above 4 pm. Below this the retention efficiency gradually decreased to between 35 and 70 % for 2 kim particles. Crassostrea virginica, which has small laterofrontal cirri, and Argopecten irradians, which has none, entirely retained particles above 5 to 6 pm. Below 5 pm the retention efficiency gradually decreased to 50 '0 for 2 Iim particles in C. virginica while it sharply dropped to 15 % for 2 pm particles in A. irradians. At 27 to 29OC filtration rates (F, I h-') as a function of tissue dry weight ( W , g ) could be described by the allometric equation: F = a~" .In C. virginica, G. dernissa and M mercenaria the equations were: 6 . ? 9~O . '~, 6.15 w o H 3 and 1.24 W'.'', respectively. B. exustus and S. sohdissin~a showed intermedial-y filtration rates. In G. demlssa gill area increased with sizc at the same rate as filtration rate.

Filtration rate (measured as clearance of algal cells) was measured at different temperatures in the sponge Halichondna panicea. An increase in water temperature from 6 to 12 "C caused the mean flltration rate to increase 4.3 * 2.3 tlmes. Thls value was higher than previously found for other marine ciliary suspension-feeding anlmals. Filtration rate at 12 "C was also measured in Haliclona urceolus by means of a n indirect clearance method In addition to a direct technique for measuring pumping rate. It was found that the 2 sponge specles had near-identical flltration rates, with maximum rates of approximately 60 m1 m~n ' (g dry weight).' at 12 T. The normal pump pressure, or operating point 0,, of a standard sponge (based on our own measurements and calculat~ons from l~terature data for a 0.1 g dry weight Haliclona sp.) was estimated as the sum of maln contributions to head losses along the flow path from entry (ostia) to exit (osculum) The head losses were as follows: ostia 0.0373 mm H,O, inhalant canal 0 1205 to 0.013 mm H,O, prosopyles 0.1153 to 0 02321 mm H,O, collar-fllter 0.122 mm H,O; exhalant canals = inhalant canals; and osculum 0.1576 mm H20. The (maximal) 0, was found to be 0 673 mm H,O and the power output P, from the sponge pump was 0.677 pW. The pump work, defined as P,R.' where R is the respiratory output, was 0.85 % The low energy cost of filtration and the temperature effect are discussed and compared w~t h recent data for other ciliary suspension feeders. It is argued that passive current-induced filtration may be of insignificant importance for sponges.

Properties of the blvalve pump were assessed by investigating, in the mussel R4yhlus eduhs, effects on pump characteristics of reduced valve gape, serotonin stlrnulation of cihary activity, as well as spawning. Reduced valve gaping caused a proportional dechne both in flow rate at zero back pressure, V', and in pump pressure at zero flow, AH?, ( = AH",,, the coefficient C I Z = L,.Hq2N0 remaining constant. Enhanced beating rate of lateral cilia, induced by serotonin stimulation, did not affect flow rate or pump pressure. Excessive stimulation that eliminated the laterofrontal clrri from the flow pathway had eri-atlc, non-significant effects. Pump pressure and flow rate vaned with the extension of mantle edges and expiratory siphon, as correlated with the valve gape. Retraction of mantle edges and siphon at reduced valve gape results in a reduction in width of the interfilament canals and thus in the distance between opposing bands of lateral cilia. This distance appears to be the main factor in determining pump pressure and flow rate in the mussel and other bivalves The distance tends to be maximal when mantle edges and siphon are fully extended, as they normally are in undisturbed mussels. It is concluded that the rate of water pumping constitutes an emergent property of the spatial geometry of the interfllament canals and the mantle cavlty, rather than a physiologically regulated process. Based partly on the experimental results and kartly on theoretical as_sumptions the pump could

A stepwise-reduction-cold vapor atomic absorption spectrometry (CVAA) method has been further developed. The method is relatively simple and quick and allows measurements of both inorganic and organic mercury in the same small biological sample. It provides an effective technique for studying biomagnification of mercury in a marine grazing food-chain (algae/mussels/fish). In laboratory experiments with mussels exposed to both inorganic mercury (965 ng I-') and organic mercury (methylmercury) (35 ng I-'), and fed algal cells at a natural low concentration from a chemostat for 80 d, uptake of both mercury species was linear. Organic mercury was taken u p 15 times more readily than inorganic mercury. Flounders force-fed with food contaminated with equal concentrations of organic and inorganic mercury for 46 d readily accumulated organic mercury in blood cells, liver, kidney and muscle-tissue while inorganic mercury was only accumulated in measurable amounts in Liver and kidney. During an elimination experiment (48 d) with both force-fed and starved flounders (loaded with mercury), organic mercury concentration did not decrease in the muscle-tissue, but significantly decreased in the liver, kidney and blood cells, while inorganic mercury tended to increase in the liver. Thus organic mercury may be biotransformed to inorganic mercury in the liver. Relatively low concentrations of organic mercury in the gall bladder indicate that enterohepatic recirculation may not play a similar role in fish as reported for mammals. Higher mercury concentrations in the proximal part of the intestine compared to the distal part show, however, that some mercury may be secreted via the bile and reabsorbed, but no differences between fed and starved fish were found Though only about 1 % of the total mercury in chronically polluted waters may be in the organic form, an enhanced uptake rate of organic mercury in mussels leads to differential partitioning of the 2 mercury forms. Since fish such as flounder may primarily feed on juvenile bivalves about 75 % of the total mercury ingested by fish may be in the inorganic form. The efficient accumulation of organic mercury and the lack of elimination results in an increasing organic mercury concentration with both age and trophic level (i.e. biomagnification) in the marine grazing food-chain.

The energetics of the ciliary crown-filament pump were studied for the suspensionfeeding polychaete Sabella penicillus. Maxlmum filtration rate expressed as the clearance capacity (F, 1 h-' ind.-') as a function of body slze (W, g dry wt) was: F = 1 3 . 6 2 W~.~~. The filtration rate was high and constant at algal (Rhodomonas sp.) concentrations below about 4 X 103 cells ml-', but at higher concentrations the gut capacity was probably exceeded thus leading to a reduced filtration rate. Oxygen consumption (R, m1 O2 h-' ind.-l) as a function of size was: R = 0.13W'.~~. The water-processing capacity of a 'standard' 65 mg dry wt S. penicillus was estimated as 354 1 of water filtered per m1 of oxygen consumed. This suggests that the polychaete is adapted to live in waters with extremely low algal concentrations. Filtration rate as a function of temperature was measured in 2 size groups of worms. The relationship fitted straight lines and it was found that the viscosity effect may explain the whole correlation between filtration rate and temperatures between 5 and 20 "C. The operating point, 0,, of the crown-filament pump was determined by equating pump characteristic and system characteristic, AHp = AH,. The system characteristic was calculated as the sum of the 2 major contributions, namely the pressure drop across the pinnule-lattice of the crown-filaments, AH,,,, and the kinetic loss, AHkex, in the water leaving each crown-filament, which was regarded as one of a series of parallel 'pump units' The calculated operating point and components for e.g. 15 "C were: 0, = AHipc + AHkex = 0.0222865 + 0.000065 = 0.0224 mm H20. The mechanical work done by the pump (pumping power) was 0.451 yW, compared to a total metabolic energy expenditure of R = 112 pW.