QUANTITATIVE ASPECTS OF FILTER FEEDING IN INVERTEBRATES

SUMMARY A description is given of the structure and function of the feeding organs in various aquatic invertebrate filter feeders (suspension feeders), especially in such forms as have also been used in experiments on feeding rate, on efficiency of feeding organs in retaining particles of different sizes, etc. Sponges ingest indiscriminately particles with and without food value. Particles that are too big to enter through the pores of the surface may be phagocytized by the cells of the epithelium. In primitive sponges with large flagellated chambers intake and digestion of food particles is mainly performed by the choanocytes, whereas in highly developed sponges a large part of the particles are phagocytized by the walls of the incurrent canals before they reach the flagellated chambers. The tube‐living polychaete Serpulimorpha are suspension feeders. They feed by means of the ciliated branchial crown which surrounds the mouth. The burrowing, tube‐living Chaetopterus variopedatus feeds by filtering water through a mucus bag. A similar feeding method has also been described in Nereis diversicolor. Within the Echiuroidea Urechis caupo likewise feeds by means of a mucus net. The net is attached to the walls of the burrow in which the worm is living. In suspension‐feeding lamellibranchs the gills both propel and filter the water. Most investigators assume that the filtration is performed mainly by the laterofrontal cilia of the gill filaments, whereas MacGinitie states that during normal feeding, water is filtered through sheets of mucus which cover the surfaces of the gills. A variety of sorting devices, especially on the gills, are developed in different lamellibranchs. Sorting is performed according to size, shape and density of particles. Qualitative sorting has, however, also been demonstrated in the oyster. Filter‐feeding habits have been adopted independently by several gastropod families, both sessile and free‐living. As in the lamellibranchs, it is generally the cleansing mechanisms of the unmodified gill and of the mantle cavity that have been developed into food‐collecting mechanisms. In Crepidula and other highly specialized suspension‐feeding Prosobranchia the filtering of the water is performed by mucus sheets, which are continuously carried over the gill surface. In vermetids from still water the importance of the gill as a food‐collecting organ is reduced, but long mucus threads, produced by the pedal gland and floating freely in the water, are used to catch food particles which adhere to the mucus. In suspension‐feeding copepods a filter chamber is enclosed between the ventral body wall and the maxillae which project ventroanteriorly. The maxillae carry long plumose setae extending antero‐medially towards the mouth and forming the lateral walls of the filter chamber. The feeding currents are produced chiefly by rapid vibratory movements of the antennae. Most if not all of the filter‐feeding copepods can feed in other ways too, e.g. by scraping or by catching larger food. Tunicates possessing a branchial sac feed by means of mucus sheets, which are produced by the endostyle and continuously carried across the inside of the branchial wall towards the dorsal lamina. Here the sheets are rolled up into a food string which is transported down into the oesophagus. The ciliary through‐current is thus filtered by the mucus sheets. In Doliolida and Salpida, where the branchial sac is reduced or absent, the feeding mucus is formed into a net which is supported by the peripharyngeal grooves. The net is continuously carried backwards, especially by the cilia of the oesophagus which twist the mucus net into a string. In Salpida the through‐current is produced by contractions of circular muscles in the body wall. Most appendicularians do not directly inspire the surrounding water, but concentrate the suspended food particles in external trap‐like filters. Amphioxus feeds similarly to tunicates, with well‐developed branchial sac. Sponges effectively strain 0–5‐1μ1 particles from the feeding current. The mucus nets of Chaetopterus and Urechis retain large protein molecules completely. Adsorption to the mucus is of negligible importance for the retention of proteins. In Mytilus and Ostrea the critical particle size was found to range round about 1–3μ. In different species and stages of filter‐feeding copepods the minimum distances between the barbs of the setae of the filters vary from 1–5 to 8‐2μ. Ascidians retain completely particles of 1 μ, but only inefficiently proteins such as haemoglobin and haemocyanin. Measurements of filtration rates in sponges vary from 45 to 170 ml./hr./mg. nitrogen. Pressures in the flagellated chambers range from o‐8 mm. water in primitive species with large chambers, to a maximum of 4 mm. in species with small spherical chambers. Veligers of O. edulis have been found to filter on an average 800 ml./hr./mg. nitrogen; adult O. virginica 19–48 ml./hr./mg. nitrogen; young Mytilus edulis 120–160 ml.; and medium‐sized to large M. californianus 5–8 ml./hr./mg. nitrogen. The optimum temperature for water propulsion in M. californianus varies with the mean temperature of the surroundings. The mussels maintain under natural conditions about the same pumping rate in all localities irrespective of the differences which may exist in mean temperatures. A quantity of o‐1 g./l. of silt, calcium carbonate, or kaolin depressed the pumping rate of Ostrea virginica by about 50%, and 1 g./l. by more than 80%. Sperm of O. virginica carries a substance, diantlin, which stimulates the flow rate. A correlation has been observed in O. virginica between rate of water propulsion and the concentration of a carbohydrate factor in the water. This factor is stated to be probably a mixture of rhamnoside and dehydro‐ascorbic acid. It is found in amounts ranging from 2 to 100 mg./l. Calanus fintnarchicus filters some 20 ml./hr./mg. dry weight, and Centropages hamatus some 85 ml./hr./mg. dry weight. Calanus is about 20 times as big as Centropages.Ciona intestinalis , weighing about 2–6 g., filtered no ml./hr./mg. nitrogen, and Molgula sp. (0–5 g.) 150 ml./hr./mg. nitrogen. Urechis feeds on an average about 13 min./hr. Lamellibranchs, copepods, and probably also sponges and ascidians, feed practically continuously when food is present in the water. Most sponges, lamellibranchs and ascidians filter 13–16 1. of water for each millilitre of oxygen taken up, and copepods 4–9 1. Suspension feeders, which filter about 15 l./ml. oxygen uptake, can cover their food requirements for maintenance and optimal growth when 0–15‐0‐20 mg. of utilizable organic matter is available per litre of water. The amount of organic matter in solution in unpolluted sea water is fairly constant, the measurements varying from 2‐2 to 4–6 mg./l. in different localities and depths. About one‐third to one‐half of the dissolved matter is ‘protein’. This fraction of the organic matter in the sea is probably not accessible as food to the filter feeders. Phytoplankton organic matter in sea water varies from zero to more than 3 mg./l. Representative seasonal variations of coastal waters are 60–1160/μg./l. in the English Channel, 30–480μ/1. on Georges Bank, 6–920μg./l. off Miami, Florida. The amounts of organic matter contained in the phytoplankton in coastal waters, where all the suspension feeders so far investigated live, are therefore sufficient to cover the food requirements of suspension feeders during parts of the year. The amounts of phytoplankton may be even 10–20 times the amounts required for optimal growth, if 151. are filtered for each millilitre of oxygen intake. When production is low, however, the amounts of phytoplankton may be too small to provide enough food even for maintenance. During such periods detritus may be of importance as a food source. Suspension feeders living at great depths in the oceans, where the average content of particulate detritus is about 25 jug./l., must be able to filter some 100 1. of water or more for each millilitre of oxygen consumed.