a pair of porous membranes kept from flying apart under flow pressure by an internal
cross-lacing of mucous fiber connections (see Flood’s original diagram), which create
the curving, organ-pipe shape of the filter. Flow is mostly parallel to the planes of the
filter, but, under sufficient pressure, water passes through all along their surfaces,
depositing particles on them. Pores are 0.24 × 1.43 μm in the lower surface, 0.18 ×
0.69 μm in the upper. Along the midline, the two curving mucous surfaces attach to
the sides of a central duct, communicating with it by valved pores. The central duct
attaches to the edges of the mouth.
(^) This sort of apparatus has also been devised by filtering engineers and is called a
tangential flow filter; particles are only held lightly against the mesh pores by the
outward component of flow, since the overall transfer is spread over such a wide area
and is locally very slow. When flow stops, the particles mostly drop back off the filter
into the central space, and slow flow from another direction will move them to an
accumulation point. The appendicularian does exactly that, the tail pumping stops and
the house deflates somewhat, shaking particles off the outer screen, cleaning it, and
freeing particles from the surfaces of the inner filter. A new, slower, flow moves water
sideways through the filter to the central duct and along to the mouth. Force for this
new flow comes from the cilia surrounding the spiracles (holes) in the sides of the
pharynx behind the mouth. Particles are filtered again by mucous sheeting across the
inside of the pharynx. Flood’s description of this complex mechanism is good reading.
(^) In another species, Oikopleura vanhoeffeni, the pharyngeal filter is much coarser
than the house filter, with typical pore openings of 3–5 μm (Deibel & Powell 1987).
Thus, the most numerous smaller particles of ∼1 μm may not be very efficiently
retained, and the mean size of prey actually reaching digestion may be >3 μm.
Apparently no actual determinations exist of retention efficiency as a function of
particle size. The bacteria-sized particles that are retained may be captured by direct
impact with filter fibers, possibly becoming embedded in the fluid boundary layers of
the fibers. The boundary layers should be thick at flow velocities measured by Acuña
et al. (1996), suggesting fiber Reynolds numbers ∼ 10 −5.
(^) Euthecosomes and pseudothecosomes (the two groups of particle-filtering
pteropods) are also mucous filter feeders, but they do not push water through a mesh;
rather they are dependent upon particle attraction to a mucous surface. According to
Gilmer and Harbison (1986), they expand a bubble of mucus (Plate 6.7) generated in
the pallial cavity (inside the shell in euthecosomes), hang from it for some minutes,
then ingest it with the particles it has attracted. It remains possible that ciliary–mucoid
filtering within the pallial cavity, described by Morton (1954), is used while the
animal actively swims. Cilia in the pallial cavity drive water and particles over the
pallial gland, which continually secretes mucus and gathers it with accumulated
particles into a string that is moved along ciliary tracts to the mouth for ingestion. In
either mode, hard, meshing teeth in a gizzard between esophagus and stomach break