Learn About Sponges: Feeding

Research study 1

photograph of typical bacterial-type food of sponge courtesy John Smit, University of British ColumbiaSponges feed mostly on bacteria and small organic particles that they filter from seawater being pumped through their internal chambers. About 80% of all water-borne bacteria are removed on a single passage through a sponge. The bacteria are captured on apical collars of special “sieve” cells known as choanocytes.

NOTE collar cells closely resemble choanoflagellate protists, and the two groups, choanoflagellates and sponges, are considered to be closely related taxonomically

The collars are comprised of tightly spaced, rodlike microvilli that surround a long flagellum. The flagellum beats, drawing water through the collar, and the particles adhere to the microvilli. The particles somehow move to the cell surface, pass inside, and are enclosed within digestive vacuoles. The choanocyte may now resorb its collar and crawl away amoeboid fashion, or the food vacuole may be passed to a wandering amoebocyte. The same or similar cells may carry undigested residues to exhalent canals and release them for transport outside of the sponge. Reiswig 1975 Can J Zool 53: 582.
first of 4 drawings to illustrate the capture of a food bacterium by a choanocyte of a sponge second of 4 drawings to illustrate the capture of a food bacterium by a choanocyte of a sponge
This series of 4 drawings shows one
food-transfer option. After a bacterium
is phagocytosed the choanocyte resorbs
its collar and crawls off into the mesohyl
to distribute the digested food products,
possibly to one or more amoebocytes
third of 4 drawings to illustrate the capture of a food bacterium by a choanocyte of a sponge last of 4 drawings to illustrate the capture of a food bacterium by a choanocyte of a sponge

Research Study2

Studies on the syconid calcareous sponge Sycon coactum at the Bamfield Marine Sciences Centre, British Columbia show that small particles (e.g., 0.1µm latex microspheres and bacteria) are filtered from seawater by the collar microvilli. In the sycon form of sponge, numerous choanocyte chambers extend at right angles from the central spongocoel. In S. coactum each chamber is about 450x100µm in size and contains 10,000 choanocytes.
hotograph of syconid sponge Sycon coactum Sycon and scanning e-micr photos courtesy Sally Leys, U Alberta

Syconid sponge Sycon coactum 1X

X-sectional photograph of a sponge Sycon coactum Sycon and scanning e-micr photos courtesy Sally Leys, U Alberta
X-section of body wall ofS. coactum as delineated by the white box in Left photo
The choanocytes are closely packed and their microvilli may touch and sometimes fuse, appearing to form a continuous filtering surface. Collar length varies greatly within a single sponge, from 0.5-4.0µm. The scanning e-micrographs show many 0.1µm latex microspheres adhering to the microvilli. Leys & Eerkes-Medrano 2006 Biol Bull 211: 157. Photos of Sycon and scanning e-micr photos courtesy Sally Leys, U Alberta
photograph of a choanocyte of the sponge Sycon coactum with latex beads on microvilli Sycon and scanning e-micr photos courtesy Sally Leys, U Alberta

Choanocyte with latex beads on microvilli

photograph of close view of the microvilli of a sponge Sycon coactum showing latex beads filtered out by the sponge Sycon and scanning e-micr photos courtesy Sally Leys, U Alberta

Closer view of latex beads on microvilli as delineated by the white box in photo on L

Research Study3

In the calcareous sponge Sycon coactum larger particles (0.5-1.0µm latex microspheres and large bacteria) are not captured by the sieving function of the collar microvilli in the way that smaller particles are (0.1µm latex microspheres and small bacteria). Rather, studies at the Bamfield Marine Sciences Centre, British Columbia show that clumps of large particles are engulfed by pseudopodia several micrometers from the cell surfaces, while single large particles are phagocytosed by lamellipodia at the surface of the choanocytes. photograph and drawing of flagella and pseudopodia of the calcareous sponge Sycon coactum Sycon and scanning e-micr photos courtesy Sally Leys, U AlbertaThe pseudopodia can reach out more than twice the height of the collar microvilli to snag particles. After the particles (latex microspheres, diatoms) are phagocytosed, the choanocytes resorb their collars and flagella, adopt amoeboid shapes, and crawl away, presumably to dump their loads of indigestible matter, or themselves along with them, into the spongocoel. A unique part of the study is that the experimenters "feed" some sponges are in situ by enclosing them in zip-loc plastic bags and then injecting the spheres in seawater solution into the bags. The study adds a new dimension to our knowledge of feeding in sponges, at least in syconid sponges – employment of feeding modes that are independent of the sieving action of the collar microvilli. Leys & Eerkes-Medrano 2006 Biol Bull 211: 157. Photos courtesy Sally Leys, U Alberta


NOTE  lit. “plate-like foot” G. The authors use these names to differentiate between a flat, wrapping type of cytoplasmic extension (lamellipodium) and a taller, engulfing type (pseudopodium)

photograph of the pseudopodia of a sponge Sycon coactum Sycon and scanning e-micr photos courtesy Sally Leys, U Alberta Pseudopodia reaching out from the surfaces of choanocytes to engulf 0.5mm microspheres

photograph of a lamellipodium of a calcareous sponge Sycon coactumLamellipodia wrapping around larger beads (1um) at the choanocyte surfaces

Research Study4

photograph of a boot spong Rhabdocalyptus dawsoni A novel approach to the study of feeding in sponges employs the use of sandwich cultures. These are, with some variation, glass microscope slides with coverslips raised on spacers to create a 0.25mm space within which sponge fragments can live. Bits of hexactinellid sponge Rhabdocalyptus dawsoni collected from Saanich Inlet, British Columbia can be maintained for several weeks in the laboratory. The authors observe that food particles (flagellates and bacteria) and latex beads (1µm diameter) are taken up almost exclusively in flagellated areas where water currents generated by the flagellae of collar bodies trap and hold particles against the sponge tissue. The collar bodies do not take up particles directly; rather, they are phagocytised across the membranes enclosing the syncytium. Although the cultures fail to grow, cytoplasmic streams that flow abundantly throughout the preparation appear to be distributing the phagocytised matter. The authors point out that the features of flagellated chambers with beating cilia, cytoplasmic streaming, and feeding displayed in their cultures indicate that the technique can serve to create valid models of intact sponges. The main difference noted by the authors between feeding in these hexactinellids versus demosponges is the lack of phagocytosis of particles by collar bodies in the former (with the possible exception of small colloidal-sized particles). Wyeth et al. 1996 Acta Zool 77: 227; Wyeth 1999 Invert Biol 118: 236.

NOTE  the preparation is sealed with vaseline save for the ends. When placed edge on in a water flow, these openings allow water to flow through the sandwich

NOTE  choanocytes are absent in hexactinellid sponges. Rather, “collar bodies” function in pumping. Collar bodies have a collar and flagellum, but a nucleus is absent and many collar bodies are interconnected by “stolons”. There is a suggestion from earlier studies that the collar bodies of R. dawsoni may be involved in capture of small colloidal particles (0.05µm). Mackie & Singla 1983 Phil Trans Roy Soc Lond Ser B 301: 365

Boot sponge Rhabdocalyptus dawsoni 0.3X

Research Study5

Scientists in Barkley Sound, British Columbia have employed an underwater remotely operated vehicle to sample inhalent and exhalent water simultaneously from two species of deep-inhabiting (120-160m) hexactinellid sponges Aphrocallistes vastus and Rhabdocalyptus dawsoni to determine aspects of feeding and metabolism.  Both species are primarily bacteriovores, removing on average 80% of bacteria from their inhalent streams. Values on the graph below the dotted line indicate feeding.  Both species respond linearly to elevated bacterial concentrations over the full range encountered in situYahel et al. 2007 Limnol Oceanogr 52 (1): 428.

NOTE  an earlier study on feeding in these species in Saanich Inlet, British Columbia employing SCUBA divers for in situ sampling of surrounding and exhalent water leads to the conclusion that hexactinellids feed principally on colloidal organic matter, with little reliance photograph showing in situ feeding measurements being taken of a deep-water spongeon bacteria.  The author notes a high variance in the data, however, perhaps owing to contamination of samples by stirred-up sediments and to the feeding activities of many plants and invertebrates inhabiting the thick sediment on the surface of Rhabdocalyptus dawsoni, and advises that the results be treated with caution.  Reiswig 1985 p. 504 In, New perspectives in sponge biology (Rützler, ed.) Smithsonian Inst Press, Washington.

In situ feeding measurements being taken of the sponge A. vastus at a depth of 160m. One probe is sampling exhalent water from within the main
osculum, while another is squirting a dye solution.
The depth of the photo is 1m; hence, the 2 scale bars

graph showing bacterial retention by sponges Aphrocallistes vastus and Rhabdocalyptus dawsoni

Research Study 6

About the same extent of protists are also removed, but their absolute mass is consderably less than the bacteria. Silica uptake is below detection levels (0.3µmol . l-1),  supporting previous notions of low growth rates in hexactinellid sponges.  Interestingly, while detritus accounts for a large portion (about 60%) of particulate organic carbon at the site, its contribution to feeding of these glass sponges is minimal.  The study is unique in that it provides the first data on feeding and metabolism of glass sponges in their natural habitat. Yahel et al. 2007 Limnol Oceanogr 52: 428.

NOTE  the siliceous skeleton comprises about 80% of the dry mass of A. vastus, with living tissue forming only a thin veneer on its surface


The 2 species occupy different positions on the rock wall, with Rhabdocalyptus
hanging parallel to it and Aphrocallistes vastus projecting up to 1m from it. 
The positions carry different risks of sedimentation and dislodgment, but appear to
allow partitioning of food resources. Intake of suspended inorganic matter is considerable,
but seems to pass through the sponges and be eliminated in the exhalent flows.

Percentage change in bacteria and protist amounts during
a single passage of water through the 2 hexactinellids


Research Study 7

photograph of 3 collar bodies of an hexactinellid sponge courtesy Yahel et al. 2006 Aquat Microb Ecol 45: 181.
View of 3 collar bodies with flagella and microvilli collars (these are directed up and to the L). Elements of the trabecular syncytium are the chunky bits on the lower R and the parts visible behind the microvilli towards the upper L. The trabecular syncytial tissue actually lines (and forms) the water canals. An inhalent canal opens on the upper R and water moves in the direction of the blue arrows through the channel and up and through the microvilli collars. A 1um latex bead (green) is being moved along in the water flow

Another study by the same research group at the Bamfield Marine Sciences Centre, British Columbia shows that the glass sponges Aphrocallistes vastus and Rhabdocalyptus dawsoni are highly selective in what they retain from seawater being pumped throuth their bodies. Thus, while 99% of the most abundant bacterial cells are removed and retained, all silt (clay) and other “debris” particles are expelled via the exhalent streams.  Moreover, while some bacteria and algal cells (<0.4–5µm) are retained with high efficiency (>86%), others are not. For example, one type of bacterium, efficiently removed (92%) in winter when overall plankton concentration is low, is less favoured (28%) in summer when overall plankton concentration is high. This indicates selective feeding. Detailed morphological study of the filtering units and simultaneous “feeding” of dye and indigestible microspheres suggests that no “bypass” routes exist, and that all water and particles entering the sponge pass through the collar bodies. The type of selective filtration exhibited by these hexactinellid sponges appears to be post-capture, requiring that all particles be individually processed, recognised, sorted, and likely transported through the syncytial tissue of the sponge.This would seem to be quite a “chore” for the collar bodies to undertake, were it not for their enormous density. Based on their numbers, average density of bacteria and silt particles in the water, and known pumping rates, the authors estimate that each collar unit would be expected to encounter 1 bacterium and 10 indigestible particles per day. This low encounter rate could presumably give the collar bodies and syncytial tissue sufficient time for processing of individual particles.  Yahel et al. 2006 Aquat Microb Ecol 45: 181.


NOTE  if bypass routes were to exist, then the dye and microspheres should exit the sponge together in the exhalent streams, but this does not happen

Research Study 8

As if the presence of extensive deep-sea sponge reefs were not surprise enough, now scientists from the Monterey Bay Aquarium and University of Victoria have described 4 new species of deep-sea sponges in the northeast Pacific that are thought to be carnivorous.  Common features of most of such species (F. Cladorhizidae) are the absence of water-canals and choanocyte chambers.  Most individuals examined of 3 of the 4 new species are found to have numerous crustaceans in various stages of decomposition attached to them (see photographs below). At this early stage of investigation it is not clear whether these crustaceans are truly prey and, if they are, how they are captured and how digestion and absorption occur.  Lundsten et al. 2014 Zootaxa 3786 (2): 101. Photographs courtesy the authors, Monterey Bay Aquarium Research Institute, pilots of the deep-sea ROVs Doc Ricketts, crews of the surface support vessels Western Flyer and Rachel Carson, and the Magnolia Press.

NOTE  the first carnivorous sponge species was discovered in the Mediterranean in 1995.  At least 11 species are now known from the Pacific west coast, all in deep water, and several being associated with deep-rift chemosynthetic environments

NOTE  another possibility might the sponge defending itself, especially with sponges associated with active hydrothermal vents and perhaps deriving their nutrition from consumption of methane-oxidising bacteria as has been proposed by other researchers

photograph of a deep-sea sponge Asbestopluma monticola
Deep-sea glass sponge Asbestopluma monticola at 1280m depth fastened to the substratum by a holdfast disc (Davidson Seamount off the coast of central California).  Monticola means “mountain dweller” in L.
photograph of what appears to be a decomposing amphipod on a deep-sea sponge Asbestopluma monticola
Asbestopluma monticola has an arborescent growth form with many fine-diameter filaments extending from its body. The red arrow shows a small crustacean that may be in the process of digestion and absorption
photograph of what appears to be a decomposing amphipod on a deep-sea sponge Cladorhiza evae
Another small crustacean caught up in the deep-sea sponge Cladorhiza evae appears to be an amphipod. It seems to wrapped in some sort of covering which strongly supports the notion that it is being digested