title used in learnabout sections of A SNAIL'S ODYSSEY
  Foods & feeding
  black dot


  Examples of studies on suspension-feeding are presented here, while ones dealing with DEPOSIT-FEEDING and ABSORPTION OF DOM are presented elsewhere. This section has a smaller subsection on THEFT OF FOOD BY SNAILS.
Research study 1

Suspension-feeding sabellids eat particulate organics, including phytoplankton, detritus, small zooplankters, and other material caught in tentacles or filaments of the branchial crown. The filaments comprising the branchial crown of the sabellid tubeworm Eudistylia sp. are arranged in 2 spirals of equal photograph of sabellid worm Eudistylia polymorphasize.  Pinnules are symmetrically arranged along each filament.  Water currents generated by ciliary beating flow through the pinnules.  As particles are driven up through the pinnules they enter a zone of decreased pressure (Bernouli’s principle) and drop out of suspension into the ciliated groove of the pinnule.  From there they are moved to the food groove on the filament and thence to the mouth where they are sorted as to edibility.  If inedible the particles are shunted to ciliary tracks leading to rejection or used as construction material for building the tube. 

Double set of filaments arising from the head of a sabellid worm
sp. The pinnules, visible here as fine, whiteish-coloured hairs,
are arranged in a double row on each filament . The food groove
of each filament is located on the opposite side of the filament as seen
from this angle. The mouth is located between the filament clusters just
below the lip of the tube, but is not visible in the photo 2X

photo composite of sabellid tubeworm showing parts
Expanded view of sabellid tubeworm Eudistylia showing parts involved in tube construction
photograph showing close view of filaments and pinnules of a sabellid tubeworm
Close view of filaments showing pinnules used to capture particulate food from the plankton
photo composite showing water flow through the pinnules of the feeding crown of a sabellid tubeworm
Diagram of water flow through the pinnules of a sabellid tubeworm. The flow is generated by beating cilia
photo composite showing simulated food particles impinging on the pinnules of a sabellid tubeworm during feeding
The arrow shows simulated food particles caught up in the food grooves of the pinnules
photo composite showing food particles being moved down the pinnules to the food groove of a filament in the branchial crown of a sabellid tubeworm
Food particles being moved to the central food groove of the filament
photo composite of food particles being moved along the food groove a filament towards the mouth in a sabellid tubeworm
Food particles being moved along the central food groove of a filament towards the mouth
Research study 2

drawing of parchment worm Chaetopterus in its tubeA study at the William G. Kerckhoff Marine Laboratory, Newport Beach, California provides some of the earliest details of how  suspension-feeding occurs in the parchment worm Chaetopterus variopedatus.  The worm lives in a leathery U-shaped tube through which it pumps seawater by rhythmic beating of 3 large fans (see drawing).  The worm begins by secreting at a rate of about 1mm per second a cylindrical-shaped mucous bag from the inner edges of a pair of aliform notopodia that is spread out against the sides of the tube such that its free ends meet as a circle.  The bag moves posteriorad through action of beating cilia located in a dorsal groove until its free end comes to rest in a cup-shaped organ known as the dorsal cupule.  Once joined top and bottom in this way, the water being pumped through the tube now has to pass through the bag where it is sieved of particulate matter.  As the bag is formed at the front end the action of cilia in the cupule rolls it into a ball at the same rate at the back end.  The notopodia now cease secreting mucus and the food bolus is moved from the cupule to the mouth in the dorsal groove by cilia that reverse their direction of beat.  After the bolus is consumed, the process is repeated.  All small particulate matter is retained and ingested.  Inorganic particles that are too large to be ingested are sensed by the notopodia, which lift up and allow the particles to be carried past the worm and out the tube.  The digestible component of the bolus consists of detritus material (organic matter with protists, fungi, bacteria, and so on) that is moved along the sea bottom by water currents.  MacGinitie 1939 Biol Bull 77 (1): 115.

NOTE  in an earlier publication another researcher notes that the fans in Chaetopterus beat in order, starting from the most anterior one, and this acts as a pacemaker. If the 3 segments are removed from the rest of the worm they continue to beat rhytmically, but if isolated from one another they beat independently. Berrill 1927 Nature 119 (2998): 564

NOTE  the bolus size is proportional to the size of the worm.  A worm of 15cm length produces a bolus of about 3mm diameter.  Under average temperatures of southern California the time taken from initial secretion of the bag to its ingestion as a bolus is 16-18min

  Research study 2.1

drawings of worm Spiochaetopterus costatum showing body parts involved in suspension-feedingA detailed study on feeding in the cosmopolitan species Spiochaetopterus costatum at Stazione Zoologica Napoli is facilitated by providing food in the form of red-dyed boiled pasta ground to a fine consistency. The worms are removed from their straight, natural parchment tubes and transferred to glass capillary tubes of 1-2mm diameter, where their behaviour can be observed for several months. There are 3 regions to a chaetopterid’s body. The 9 anterior segments bear lobe-like notopodia (see drawing). On the first of these segments are paired long palps whose total length may equal or exceed half the body length (8cm total BL; see drawing). The middle body region contains about 30 segments, most or all of which bear broad, rounded lobes used for anchoring the body in the tube. The mucous bag for feeding is secreted by a mid-dorsal ciliated region on an anterior one of these segments. As it forms, the bag is carried posteriorad in the water flow through the tube and is caught up in a large ciliated cupule located mid-dorsally on this same segment. The cilia in the cupule beat in a posteriorad direction, rolling the gathered end of the mucous bag into a ball. When a certain size is reached the ball with its load of plankton and fine detrital particles is detached from the cupule and carried anteriorad to the mouth in the dorsla ciliated food groove where it is ingested. The large palps bear a ciliated groove along the mid-line of their lengths that may be used as a secondary, but minor, means of food-collection. Food particles are trapped within these grooves and carried proximally directly to the mouth. On the outer side of each palp is a smaller groove that runs along the palp to its tip, within which the cilia beat distally. This groove collects any overly large or otherwise inedible particles coming from the palps or carried in by water flow and ejects them. The posterior body region consists of a variable number of photograph of chaetopterid tubeworm Spiochaetopterus costatumundifferentiated segments ending with a terminal anus. The palps function also to eject feces. After being expelled from the anus the fecal pellets are caught up by cilia in the dorsal ciliated groove that runs the entire length of the body and moved to the ejection grooves on the palps. If the worm happens to be within its tube during either of these ejection events, it crawls upwards until the palps entend out of the tube opening, allowing the unwanted materials to fall outside of the tube. The efficacy of this housekeeping is evidenced by the quick and almost total removal of unpalatable carmine particles squirted by the experimenter into the tube opening. Water currents for gas exchange and ridding the body of excretory wastes are generated by ciliary beating on special areas of the segments of the middle body region. Barnes 1965 Biol Bull 129 (2): 217. Photograph courtesy Linda Schroeder, PNSC, Olympia WA and Bruce Kerwin.

NOTE this species is also present along the west coast of North America from British Columbia to California. In contrast toChaetopterus featured in Research Study 2 above, which inhabits a U-shaped tube, this species lives in a straight tube

Apart from the superb quality of this photo showing a tubeworm Spiochaetopterus
bearing eggs of a snail Lacuna variegata being eaten by another snail Margaritus
, you should be struck by the slenderness of the worm in its tube. This slenderness
enables the worm to pull in and turn around inside its tube, should it need to do so


Research study 3

photograph of sabellid worms Eudistylia sp.diagram of flow direction and speed in the branchial crown of a sabellid tubeworm Eudistylia vancouveri during feedingStudies on feeding in clusters of tubeworms Eudistylia vancouveri in San Juan Island, Washington show that feeding currents generated by ciliary beating on the tentacles (0.2-0.8 mm . sec-1) are augmented more than 3 orders of magnitude (340 mm . sec-1) by enhanced velocity1 of current flow over the cluster (see drawings on Right).

Because the current flow increases in speed as it flows over a cluster of tubeworms, individuals within a cluster experience higher current velocities and thus have access to more food than solitary individuals2.  Sampling of the current for suspended particles upstream and downstream of a worm cluster shows that about 45-65% of the particles are removed.  About 70% of the particles are 3-6µm in diameter3  Thus, in creating mounds or hummocks, Eudistylia is not only minimising turbulent drag, but also maximising feeding effectiveness. Merz 1984 Biol Bull 167: 200.

graph showing particle removal from water passing through a cluster of tubeworms Eudistylia vancouveriNOTE1  rates and patterns of current are monitored in the laboratory using a thermistor flow meter and fluoroscein dye, and in the field using a commercial electromagnetic water-current meter

NOTE2 downstream of a cluster the current vortices create chaotic, recirculating flow patterns that enhance particle capture for individuals located there 

NOTE3  additionally, bacteria of 0.5µm diameter are known to be filtered by some sabellids.  Within this size range are bacterio-, myco-, and small-phytoplankton, all of which appear to be readily digested by the worm

Research study 4

comparison of collar types on the tubes of sabellarid worm Phragmatopoma californicaAn observation by a researcher at University of California, Berkeley that tube morphology of sabellariids Phragmatopoma californica varies with degree of wave exposure leads to an interesting study on water flow, tentacle deflection, and feeding in the species.  In Monterey Bay, California, the worms fashion their tubes with raised extensions (collars in the form of either flares or hoods) that extend about 7mm above the upper tube rim.  In wave-protected habitats these extensions (hoods) only half-encircle the circumference of the apertures and they open towards the predominant direction of water flow over the aggregation.  In wave-exposed habitats, however, the collars (flares)extend completely around the circumference of the tube (see drawing on Left).  Most or all individuals in an aggregation have the same type of collar.  Field and laboratory (Long Marine Laboratory and Hopkins Marine Station) experiments with individuals of each type subjected to currents impinging from different directions provide information on the functional significance of the collars.  First, both types of collar tend to decrease the rate of exchange between the aggregation and the mainstream water flow, although hoods appear to do this to a greater extent than flares.  Second, although both types of collars reduce deflection of the feeding tentacles downstream by water currents, the effect is greater in hooded individuals by the tentacles being more protected (see graph below.  Note in the graph that there is little deflection of the tentacles at any orientation of the hoods at low-flow velocity, but at higher velocities, as expected, orientation does become significant.  Although not measured by the researcher in this study, these observations suggest that feeding may be more efficient in hooded than in flared individuals.  If a collar is damaged experimentally in the field, it is repaired, with the new hood orientated in the same direction as the original.  Interestingly, replacement of a damage collar is an order of magnitude slower for flares than for hoods (240h vs. 24h, respectively), a difference not completely related to the greater mass of repair needed to be done.  Thomas 1994 Mar Biol 119: 525.

NOTE  the author terms these collars flares if they completely encircle the aperture, or hoods if they only half-encircle the aperture.  By the nature of their packing, an aggregation of flared worms has a honeycomb appearance (see drawing on Right)

NOTE  this reduction in flow velocity has several implications, not only for feeding but also for degree of fertilisation success within the aggregation

Research study 5

graph showing temporal variability in food for tubeworms Serpula columbiana in Indian Arm, British ColumbiaField measurements of particulate carbon in habitats of Serpula columbiana in Indian Arm, British Columbia show that food levels can change by up to 600%1 over a matter of minutes (see graph on Right).  The authors note that if a worm were to pull back into its tube in response to disturbance and remain there for some time, it could experience a greatly reduced level of food availability when it re-emerges as compared with when it withdrew. In its effect on reduced growth and reproductive output, decreased food intake could have critical effects on the worm's fitness. Will a hungry worm adjust its hiding time relative to food levels it was experiencing prior to withdrawal? What if the initial withdrawal were in response to a predator? if the worm emerges too early (in other words, before the predator that caused the worm to withdraw has left the scene) it risks being eaten.

graph showing effect of predator simulation on withdrawal times of tubeworms Serpula columbiana Experiments on Serpula from Indian Arm, British Columbia show that hiding time varies inversely with food availability (see graph lower Right). The graph is read as follows: one group of 35 worms fed on a high food ration2 for 5d and then stimulated to withdraw with a vibration device during this 5-d period, hides on average for 61sec.  When tested a few days later while being fed on a low food ration, the worms hide on average for 117sec.  A second group of 35 worms maintained initially on a low food ration for 5d hides on average for 132sec.  When tested a few days later after being shifted to a high food ration, the same worms hide on average for 70sec.  Results3 of similar experiments suggest to the authors that the worms in their tubes somehow remember what the food conditions are outside and judge their hiding times accordingly.  Dill & Fraser 1997 Behav Ecol 8: 186.

NOTE1  the data presented here show only about a 60% change; however, much larger changes are indicated in other data sets presented by the authors. Note that sampling interval is 5minphotograph of a cluster of serpulid tubeworms Serpula columbiana courtesy Ron Long, SFU, British Columbia







NOTE2  the food provided to the worms in this experiment is powdered Spirulina algae purchased commercially and made up into a food-stock of 1g powder in 200ml filtered seawater.  A “low” food ration is 1ml of this suspension in 7 liters of seawater in the worms’ aquarium tank; a high ration is 10ml per tank.  In other experiments, not reported here, the food is powdered yeast made up in the same concentration as the algal food

NOTE3  the authors admit that their data suggesting a capability for memory are statistically weak; however, the notion is a good one and is deserving of further research

Cluster of serpulid tubeworms Serpula columbiana.
Note that 2 individuals are withdrawn, while a
third individual is partly withdrawn 0.7X

  black dot

What other explanation(s) might there be for the data? Think about the options given below, then CLICK HERE.

The worms can sense through their tube what the feeding/predator conditions are like outside and adjust their hiding times accordingly. 

The worms are more tired on a low-ration diet and simply stay in their tubes longer after withdrawal. 

The withdrawn worm samples the outside seawater for the presence of food and predators via the operculum and adjusts its hiding time accordingly. 

Greater concentration of food in the water reduces oxygen levels; thus, a worm that withdraws at this time will have incurred an oxygen debt that must be repaid and so it emerges quicker.

Research study 6

photograph of pectinariid worm Pectinaria californiensis courtesy Lovell & Libby Langstroth and CALPHOTOS.photograph and drawing of larval tubeworm Pectinaria californiensis within its mucous houseAs adults, “ice-cream cone” worms Pectinaria californiensis inhabit sandy tubes within sediments.  The worm orients with the large, or head end, downwards and feeds on organic matter within the sediment.  The narrow, posterior end projects upwards, out of the sediment, and is a respiratory tube for ventilation of the gills.  The larval stage is free-living and, as discovered by a researcher at Friday Harbor Laboratories, Washington, suspension-feeds by constructing a relatively large mucous “house” (see drawing and photograph on Right).  The house is hollow, about 800µm maximum diameter, and acts as a filtering basket.  Seawater is pumped by the larva past its body, and into and through the house.  Particles greater than about 6µm are retained on the inner walls of the house.  A recirculating current generated within the house by cilia on the oral hood surrounding the larva’s mouth drives the captured particles to the mouth where they are eaten.  A house is produced about 3d post-fertilisation.  When a house is abandoned, perhaps when it become laden with inedible particles, a new one is created within about 10min.  Older houses are more visible to the eye than newer ones.  Pernet 2004 Science 306: 1757.  Photograph of adult worm courtesy Lovell & Libby Langstroth, California and CALPHOTOS.


Theft of food by snails

Research study 1

photograph of snail Trichotropis cancellata courtesy Linda Schroedergraph showing growth of snail Trichotropis cancellata with and without its tubeworm host Serpula columbianaSeveral species of west-coast tubeworms, including the sabellids Schizobranchia insignis and Potamilla ocellata, the serpulid Serpula columbiana, and the sabellariid Sabellaria cementarium, are subject to food theft by snails Trichotropis cancellata. The snail itself is basically a suspension-feeder, filtering out particulate food on its ctenidium and transferring it from there to its mouth using an extensible part of its lower lip termed by the authors a pseudoproboscis. When behaving as a food thief, the snail creeps close to the worm, inserts its proboscis into the worm’s mouth, and channels the food streams into its own mouth (see photos below).  The strategy definitely enhances fitness, as snails kept for 8mo in the laboratory with access to feeding tubeworms grow faster and survive more than ones without access (see graph on Right).  Note in the graph that most of the juvenile snails die when placed together with just empty tubes of their worm hosts. Other experiments show that small snails are unable to meet basic metabolic needs solely by suspension-feeding and thus tend to be obligate parasites.  In contrast, larger snails tend to be facultative parasites.  The authors are actually able to count particles, in this case, cysts of brine shrimps Artemia, captured by a tubeworm (e.g., Serpula columbiana) and in the process of being eaten, and compare these with the number of cysts being diverted to the snail’s mouth.  In some experiments 100% of cysts captured by a worm are diverted to the thieving snail.  Snails in the field are commonly seen positioned near the tube openings of sabellariid and sabellid worms.  A bout of food theiving may last for hours or even weeks, and throughout this the worm seems unaware of, or unresponsive to, the snail’s presence. Pernet & Kohn 1998 Biol Bull 195: 349. Photograph of Trichotropis courtesy Linda Schroeder, Pacific Northwest Shell Club, Seattle, Washington.

NOTE  known as kleptoparasitism, a behaviour more commonly reported for vertebrates (birds and mammals) and insects, than for marine invertebrates

NOTE  one wonders about the need for this particular bit of jargon.  The simpler word “proboscis” (lit. “in front of” “feed” G.) would seem to be perfectly suitable. Probosces do tend to be tubular and sucking, while the device on Trichotropis seems more for lapping, and that may be the explanation

photograph of parasitic snail Trichotropus cancellata crawling on its host serpulid Serpula colulmbiana photograph of parasitic snail Trichotropus cancellata crawling on its host serpulid Serpula colulmbiana photograph of parasitic snail Trichotropus cancellata crawling on its host serpulid Serpula colulmbiana
Parasitic snail Trichotropis cancellata crawls on the tube of its host serpulid Serpula columbiana On another host, the sabellid Potamilla ocellata, the snail inserts its proboscis into the crown of tentacles This snail has its proboscis inserted into the mouth of P. ocellata. The snail's tentacles can be seen on either side
Research study 2

histogram comparing growth rates of snails Trichotropis cancellata on different worm hostsThese observations raise the questions as to which host species provide Trichotropis cancellata with best growth, and are these species  preferred by the snail when it is given a choice.  Research on these topics at Friday Harbor Laboratories, Washington involve maintaining worm/snail arrays in the field and laboratory for periods of from 3-13wk, then measuring how much the snails have grown (increase in shell length).  Overall, growth of Trichotropis is significantly enhanced on all hosts over that of control snails (tethered to an empty worm-tube), but only in spring/summer when phytoplankton levels are high are there no differences among hosts in their growth-promoting abilities (see histogram).  In autumn, when phytoplankton is less common, snails grow more on sabellid hosts than on the serpulid one.  The snails spend about the same amount of time thieving from each host species, so the difference must owe to lower quantity or quality of food taken from the serpulid host.  Despite this, Trichotropis does not appear to preferentially select sabellid hosts over serpulid hosts in controlled choice experiments.  The author concludes that choice of hosts by Trichotropis is governed by factors other than growth rate.  Iyengar 2004 Oecologia 138 (4): 628.

NOTE  species used include the serpulid Serpula columbiana and the sabellids Pseudopotamilla ocellata, Schizobranchia insignis, and Eudistylia vancouveri.  A brachiopod is included for interest sake, but the results are not presented here

NOTE  in fact, contrary to expectation, the author does mention finding some preference for serpulids over sabellids in field surveys, but does not mention whether this owes to more serpulids being present in the particular habitat being surveyed

Research study 3

histogram showing effect of snail parasites Trichotropis cancellata on growth of tubeworm Serpula columbianaIn a related paper, the same author focuses on a single host tubeworm-species Serpula columbiana and assesses the benefits and costs  to both participants in the symbiosis.  Thus, in several field experiments in which snails Trichotropis cancellata are tethered with threads to rocks alone (as a suspension-feeding treatment) or with host worms (parasite treatment) for a 29d period at 18m depth, the following results are obtained: 1) snails of all sizes grow more quickly with a host worm than without (up to 16-fold more), 2) growth-enhancing benefits are significantly greater for smaller-sized snails than for larger, suggesting that smaller individuals are more food-limited than larger ones, and 3) adding snails of medium size to a single host significantly reduces growth of all, and the effect is additive.  With respect to this last experiment the author additionally comments that although more than one snail can feed simultaneously, there never appears to be any interference competiton between them,  and no one member of a group  emerges as competitively dominant.  As for the worms, tube-growth is significantly reduced in the presence of a food-thieving snail, and the effect is additive as more snails (of medium size) are tethered to the same host (see histogram).  This last is the first definitive evidence that the relationship is a photograph of snail Trichotropis cancellata close view Dave Cowlesparasitic one.  Other negative effects to the host could include extra energy expended in manufacturing a thicker tube to support the added mass of the parasite, compounded by increased drag associated with a snail’s presence.  The study is especially valuable in that it is done completely in the field. Iyengar 2002 Mar Ecol Prog Ser 244: 153.Photograph courtesy Dave Cowles, Walla Walla University, Washington.

NOTE  Serpula’s tube is calcareous and can be marked by soaking it in a solution of fluorescent calcein dye for 48h.  This produces a permanent yellow ring around the outer edge of the tube, easily visible under ambient light

Worm's-eye view of a Trichotropis
cancellata snail looking back 3X

Research study 4

graph showing correlation between density of parasitic snails Trichotropic cancellata and their tubeworm hosts in the San Juan Islands, Washingtongraph showing seasonal variation in parasitic activity of snails Trichotropis cancellata in San Juan Islands, WashingtonTo what extent does parasitism of snails Trichotropis cancellata vary seasonally.  This is investigated by a researcher at Friday Harbor Laboratories, Washington for 6 sites in the San Juan Islands area over a 2yr period.  Results show that parasitism is the dominant feeding mode in summer, with about 65% of members of the 6 snail populations engaging in the activity with about 25% of the entire worm population (see graph on Left).  Lowest incidences of parasitism occur in winter, which is the species mating season, and the season of low plankton abundance.  Females return to parasitic feeding in Mar-Apr after their egg-masses hatch, while males have returned somewhat earlier.  In this region, 5 species of tube-dwelling worms from 3 families are Trichotropis’ most common hosts, and their densities correlate well with extent of parasitism (see graph on Right).  The author provides other information on the parasitic relationship, such as host/parasite loads, host sizes, and so on.  Iyengar 2005 Can J Zool 83: 1097.

NOTE  the studies are done in the field using SCUBA

NOTE  the species are Eudistylia vancouveri, Pseudopotamilla ocellata, and Schizobranchia insignis (F. Sabellidae), Serpula Columbiana (F. Serpulidae), and Sabellaria cementarium (F. Sabellariidae)