Locomotory waves of Koruga and Deltotrichonympha: flagella wag the cell.

We investigated the nature of the locomotory waves of Koruga and Deltotrichonympha, flagellates living symbiotically in the hindgut of the Australian termite Mastotermes darwiniensis. The locomotory waves consist of two components: metachronal waves of flagellar beating and undulations of the cell surface, which propagate synchronously with the same wavelength, frequency, and velocity. We asked, do body waves cause flagellar waves, or vice versa? Using video microscopy and selective inhibitors and drugs, we found that (1) the amplitude of flagellar waves remains constant independent of variations in the amplitude of body waves, (2) flagellar waves can occur in the complete absence of body waves, (3) flagellar waves can induce body waves on swollen regions, (4) inhibition of flagellar beating by dynein inhibitors causes disappearance of body waves, and (5) cytochalasin D induces changes in cell shape but does not inhibit locomotory waves. Therefore, flagellar waves are not produced passively by an active contractile system in the cell cortex; instead, metachronally beating flagella exert waves of pressure that induce passive undulations of a pliant cell surface. These results support Machemer's [1974] theoretical analysis of the data of Cleveland and Cleveland [1966: Arch. Protistenk. 109:39-63], who believed the opposite.

Dynamic control of reversible cell adhesion and actin cytoskeleton in the mouth of Beroë.

Cell-cell adhesion in the various types of intercellular junctions of differentiated tissues is relatively stable and permanent. In migrating cells of embryos, or in wound closure, inflammatory responses and tumors of adult tissues, however, bonds between cells are made and broken and made again, i.e., cell-cell adhesions are transient and reversible. These nonjunctional contacts lack the organized structure of intercellular junctions, but may initiate their tissue-specific formation during development. Investigation of dynamic, nonjunctional cell-cell adhesions has been hampered by the asynchronous and heterogeneous distribution of these transient contacts among groups of moving cells. We recently discovered a novel system of reversible cell adhesion in a differentiated tissue that overcomes this difficulty. Here I review our current knowledge of this system, particularly its unique experimental advantages for investigating the mechanisms and control of dynamic cell adhesion.

A giant nerve net with multi-effector synapses underlying epithelial adhesive strips in the mouth of Beroë (Ctenophora).

We present ultrastructural evidence for the first known example of a giant nerve net in the phylum Ctenophora. The giant fibre system in Beroë underlies paired strips of adherent epithelial cells that run inside the lips. Interlocking actin-lined cell junctions between opposing adhesive strips keep Beroë's large mouth closed while the ctenophore searches for prey. The giant neurons, up to 6-8 microns in diameter, form a continuous lattice-like plexus rich in vesicles, microtubules, and 'presynaptic triads'. A novel feature is that individual giant axons make synaptic contacts with more than one type of effector, i.e. longitudinal muscle fibres and epithelial adhesive cells. Contact of prey with sensory receptors on the lips of Beroë induces rapid disappearance of the actin-lined adhesive cell junctions, and muscular opening of the mouth to ingest prey. Electron microscopy of food-opened mouths shows local thickening of longitudinal muscles and widening of the basal ends of epithelial cells in the adhesive strip, correlated with retraction of the adhesive epithelium into the mesoglea. Addition of 1% Triton X-100 to formaldehyde fixative in the absence of prey also elicits regional thickening of longitudinal muscles at the location of the adhesive strips (visualized by rhodamine-phalloidin staining). The giant neuron system may serve as a final common pathway to rapidly signal disassembly of actin-based junctions between adhesive cells as well as contractions of longitudinal muscles underlying the adhesive strips, thereby enabling Beroë to open its mouth rapidly to engulf prey.

Visualization of calcium transients controlling orientation of ciliary beat.

To image changes in intraciliary Ca controlling ciliary motility, we microinjected Ca Green dextran, a visible wavelength fluorescent Ca indicator, into eggs or two cell stages of the ctenophore Mnemiopsis leidyi. The embryos developed normally into free-swimming, approximately 0.5 mm cydippid larvae with cells and ciliary comb plates (approximately 100 microns long) loaded with the dye. Comb plates of larvae, like those of adult ctenophores, undergo spontaneous or electrically stimulated reversal of beat direction, triggered by Ca influx through voltage-sensitive Ca channels. Comb plates of larvae loaded with Ca Green dextran emit spontaneous or electrically stimulated fluorescent flashes along the entire length of their cilia, correlated with ciliary reversal. Fluorescence intensity peaks rapidly (34-50 ms), then slowly falls to resting level in approximately 1 s. Electrically stimulated Ca Green emissions often increase in steps to a maximum value near the end of the stimulus pulse train, and slowly decline in 1-2 s. In both spontaneous and electrically stimulated flashes, measurements at multiple sites along a single comb plate show that Ca Green fluorescence rises within 17 ms (1 video field) and to a similar relative extent above resting level from base to tip of the cilia. The decline of fluorescence intensity also begins simultaneously and proceeds at similar rates along the ciliary length. Ca-free sea water reversibly abolishes spontaneous and electrically stimulated Ca Green ciliary emissions as well as reversed beating. Calculations of Ca diffusion from the ciliary base show that Ca must enter the comb plate along the entire length of the ciliary membranes. The voltage-dependent Ca channels mediating changes in beat direction are therefore distributed over the length of the comb plate cilia. The observed rapid and virtually instantaneous Ca signal throughout the intraciliary space may be necessary for reprogramming the pattern of dynein activity responsible for reorientation of the ciliary beat cycle.

Dynamic control of cell-cell adhesion and membrane-associated actin during food-induced mouth opening in Beroë.

We used rhodamine-phalloidin and ultrastructural methods to follow dynamic changes in adhesive cell junctions and associated actin filaments during reversible epithelial adhesion in the mouth of the ctenophore Beroë. A cruising Beroë keeps its mouth closed by interdigitated actin-coated appositions between paired strips of cells lining the lips. The mouth opens rapidly (in 0.2-0.3 s) by muscular action to engulf prey (other ctenophores), then re-seals after ingestion. We found that the interlocking surface architecture of the adhesive cells, including the actin-coated junctions, rapidly disappears after food-induced opening of the mouth. In contrast, forcible separation of the lips in the absence of food rips the junctions, still intact, from the surfaces of the cells. The prey-stimulated loss of adhesive cell junctions and associated actin cytoskeleton is one of the most rapid changes in actin-based junctions yet observed. This system provides unique experimental advantages for investigating the dynamic control of reversible cell adhesions and membrane-associated actin filaments.

Mechanical properties of ciliary axonemes and membranes as shown by paddle cilia.

Cilia with a distal membrane expansion enclosing a coiled end of the axoneme (paddle cilia or discocilia) have been commonly reported in marine invertebrates. We recently showed that paddle cilia in molluscan veligers are artifacts of non-physiological conditions. Here we investigated the possible mechanisms of formation of paddle cilia under hypotonic conditions; particularly, whether a helical conformational change of doublet microtubules induced by Ca or proton flux is responsible. Typical paddle cilia are induced by hypotonic Ca-free solutions at normal or low pH, showing that axonemal coiling does not require Ca influx or proton efflux. In addition, Triton-demembranated straight axonemes do not coil in high Ca solutions. Most decisively, complete removal of paddle ciliary membranes with detergents, but not mere permeabilization, causes immediate uncoiling and straightening of the axonemes to approximately their original length before hypotonic treatment. These findings and other data show that axonemal coiling in paddles is due to membrane tensile stress acting on an elastic axoneme. Light and electron microscopy of paddles show that axonemes coil uniformly toward the direction of the effective stroke (doublets nos 5-6), even when beating is inhibited by sodium azide or glutaraldehyde before hypotonic treatment. This indicates that axonemes possess an intrinsic asymmetry of stiffness within the beat plane, independent of active microtubule sliding. Paddle cilia thus reveal important mechanical properties of ciliary axonemes and membranes that should be useful for understanding ciliary function.

Patterns of electrical activity in comb plates of feeding Pleurobrachia (Ctenophora).

The electromotor behaviour of ciliary comb plates was studied during prey-stimulated and electrically stimulated feeding by intact Pleurobrachia pileus (Müller). Comb plate electrical activity was recorded by extracellular electrodes attached directly to the cilia; comb plate motility was recorded by high-speed video microscopy. Comb plate electrical activity fell into two distinct classes, identified by waveform and amplitude: (i) excitatory postsynaptic potentials (EPSPS) in the comb plate (polster) cells and (ii) regenerative potentials in the cilia, as described previously (Moss & Tamm 1987). Slow phasic bursts of regenerative potentials (reversal volleys) were observed in comb plates of rows undergoing reversed beating during capture of prey or by rhythmic electrical stimulation of the tentacles. All plates of a given comb row exhibited virtually identical electrical activity. Timing and development of electrical activity in comb plates of the subtentacular (ST) rows were nearly identical even though separated by several centimetres; onset of the reversal volleys of plates of subsagittal (ss) rows were delayed on average by about 0.5 s relative to the ST rows, although individual EPSPS displayed very similar timing. Microsurgery, combined with extracellular recording from comb plates and the tentacle and associated basal structures, revealed the presence of an integrative center in the tentacular bulb. This communicates with the comb plates by means of a diffuse pathway, presumably the nerve net, which itself is maximally sensitive to rhythmic input. The pathway underlying the reversal volley may innervate only the stimulated hemisphere. In addition to the rhythmic pathway, a through-conducting pathway runs from distal regions of the tentacle to the comb plate cells. Yet another excitatory pathway, possibly distinct from the tentacular through-conducting pathway, may mediate certain cases of global postsynaptic activity. The pathway that controls mouth movements during feeding is entirely independent of any comb plate pathway.

On the Nature of Paddle Cilia and Discocilia.

Cilia with paddle-shaped or disc-shaped tips enclosing a curved end of the axoneme (paddle cilia or discocilia) have been described in a variety of marine invertebrates. Although numerous studies, in which fixed specimens were used, claimed that paddle cilia and discocilia are genuine structures of unknown function, several studies, in which fresh living material was used, reported that modified cilia are artifacts. We have re-investigated a recent SEM report that paddle cilia are genuine organelles in veliger larvae of marine bivalves (Campos and Mann, 1988). Using high-speed video and electronic flash DIC microscopy, we find no paddle cilia in living larvae of Spisula solidissima and Lyrodus pedicellatus. Hypotonic seawater, however, induces formation of paddle cilia and vesiculations of the ciliary membrane in these veligers, as does the hypotonic SEM fixative used by Campos and Mann (1988). Fixatives that are isosmotic with seawater, on the other hand, do not induce paddle cilia. We conclude that paddle cilia are artifacts, and we propose a unifying mechanism to explain their production in various animals under different conditions.

Reversible Epithelial Adhesion Closes the Mouth of Beroe, a Carnivorous Marine Jelly.

We investigated how the ctenophore Beroe, a carnivore of the marine zooplankton, keeps its mouth shut to maintain a streamlined body shape during forward swimming in search of prey. In big-mouthed, thin-walled beroids we found that mouth closure requires neither muscular nor nervous activity. Instead, mouth adhesion is due to paired strips of adhesive epithelial cells on opposing stomodaeal walls. The two joined epithelial layers make numerous close appositions interrupted by vacuolar intercellular spaces. At regions of apposition, the plasma membranes are highly folded and interdigitated with each other, and are separated by a uniform distance of about 15 nm. A dense cytoplasmic coat underlies the membranes at such appositions. Synapses of neurites are found on the basal ends of the adhesive cells. We found two orthogonally different orientations of the stomodaeal adhesive strips in B. sp. vs. B. forskali, correlated with different distributions of feeding macrocilia inside the stomodaeum. Mouth opening in response to food requires muscular contractions of the lips. However, the stomodaeal adhesive strips are not pulled apart all at once, but are peeled apart starting from a site of vigorous muscular tension. The mouth re-seals after feeding, or after being experimentally pulled open, showing that tissue adhesion is functionally reversible. Epithelial adhesion in Beroe appears to be a useful method for closing the mouth and streamlining the body of a gelatinous predator that spends most of its time swimming mouth-forward in search of prey. Opening of the mouth appears to be an efficient process as well, because peeling apart of the adhesive strips requires a smaller applied force than does separating them all at once. Tissue adhesion in Beroe shares many structural and functional properties with transient adhesions formed between moving cells in embryos and in culture, and may be a useful experimental system for studying the mechanisms and regulation of dynamic cell adhesions. "Loose lips Sink ships." --U. S. Navy slogan, WW II.

Extracellular ciliary axonemes associated with the surface of smooth muscle cells of ctenophores.

We describe the first example of bare ciliary axonemes existing outside eukaryotic cells. The axonemes run in longitudinal invaginations of the surface membrane of giant smooth muscle cells in ctenophores. No motility of the surface-associated axonemes has been detected in living muscles. The axonemes are truly extracellular and in direct contact with the extracellular matrix (mesoglea), as shown by the ultrastructural tracer horseradish peroxidase. The axonemes appear partially degraded and disorganized, and individual doublet microtubules are difficult to distinguish. Nevertheless, immunofluorescence microscopy shows that the axonemes retain antigenic sites reacting with mouse monoclonal anti-beta-tubulin. The origin of the extracellular axonemes is unknown: no attached basal bodies (extracellular or intracellular) have been found. The muscle-associated axonemes may play a unique role in smooth muscle function and/or development, and may be related to the evolution of muscle cells in soft-bodied invertebrates that exploit cilia for a wide variety of functions.

Calcium sensitivity extends the length of ATP-reactivated ciliary axonemes.

We use the Ca-dependent activation response of macrocilia of the ctenophore Beroë to map the distribution of Ca sensitivity along axonemes of detergent-extracted ATP-reactivated models. Local iontophoretic application of Ca (or Sr or Ba) to any site along the length of demembranated macrocilia in ATP-Mg solution elicits oscillatory bending. Bending responses are localized to the site of application of these cations and do not propagate. Ca sensitivity for initiating bends is, therefore, distributed along the entire length of the axonemes. Since Ca triggers ATP-dependent microtubule sliding disintegration of macrociliary axonemes, a Ca-sensitive mechanism for activating microtubule sliding extends the length of the axonemes. In contrast, local application of Ca to living dissociated macrociliary cells elicits beating only when applied to the base of the macrocilium, indicating that the effective site of Ca entry is localized to the membrane at the ciliary base. Therefore, the spatial distributions of membrane Ca permeability and axonemal Ca sensors do not coincide.

Control of reactivation and microtubule sliding by calcium, strontium, and barium in detergent-extracted macrocilia of Beroë.

Macrocilia of the ctenophore Beroë are activated to beat continuously in the normal direction by membrane-mediated Ca2+ influx (Tamm: Journal of Comparative Physiology [A] 163:23-31, 1988a). Using saponin or Brij-58 permeabilized models of macrocilia, we show that ATP-reactivation of beating requires microM levels of free Ca2+, Ba2+, or Sr2+. Isolated macrocilia beat initially in reactivation solution (RS) containing Ca2+, Ba2+, or Sr2+ and then undergo microtubule sliding disintegration without added proteases. Addition of protease inhibitors to RS + 10(-5) M Ca2+ prevents sliding disruption. Pretreatment in wash solution (containing 1 mM EGTA) without protease inhibitors, followed by RS + 10(-5) M Ca2+ with protease inhibitors results in extensive sliding disintegration. However, treatment in wash solution followed by RS + protease inhibitors does not induce sliding. Therefore, Ca2+ is not required for proteolysis by endogenous proteases, but is necessary for sliding disintegration. Local iontophoretic application of Ca2+, Ba2+, or Sr2+ to permeabilized macrocilia in RS lacking these cations triggers motility and/or sliding disintegration. Extrusion of microtubules occurs from the tip or the base, depending on whether or not the macrocilium remains attached to its large actin bundle. Thin sheets of microtubules telescope out initially, due to synchronized sliding of subsets of doublet microtubules from parallel rows of axonemes. Macrocilia are one of the first examples of ATP-induced microtubule sliding which retains Ca2+ sensitivity. In addition, the finding that Ba2+ and Sr2+ also trigger active sliding provides an additional method for investigating the control of dynein-powered microtubule movements.

Calcium activation of macrocilia in the ctenophore Beroë.

1. Macrocilia on the lips of the ctenophore Beroë are usually quiescent, but can be activated to beat rapidly and continuously by various stimuli. 2. During feeding, macrocilia beat actively and serve to spread the lips of Beroë over its prey. 3. Vigorous, repetitive mechanical stimulation of the lips evokes widespread activation of macrocilia via a pathway that is probably neural. 4. Extracellular electrical stimulation (DC or bipolar pulse-trains) elicits immediate activation of macrocilia on lip pieces, but not on dissociated cells. 5. Macrocilia on lip pieces are activated to beat by high KCl artificial sea water (ASW), but not by high KCl Ca-free ASW. Continuous beating for long periods is also elicited by high Ca ASW or Mg-free ASW, but not by Ca-Mg-free ASW. Addition of La, Cd, Co or Mn (10 mM) to high KCl ASW reversibly blocks activation. Verapamil, D-600, nifedipine, or BAY K 8644 (10 microM) has no effect on KC1-induced activation, but the anticalmodulin drug W-7 (10 microM) reversibly inhibits beating. 6. Mild heat treatment dissociates macrociliary cells from lip tissue. Such isolated macrociliary cells usually beat continuously in normal sea water, and swim in circular paths. Ca-free ASW, or addition of Co or Mn to ASW, inhibits beating of dissociated cells. High KCl ASW activates beating of quiescent, isolated macrociliary cells. 7. Ca-Mg-free ASW inhibits beating of dissociated macrociliary cells, and return to Mg-free ASW activates motility, allowing one to activate macrocilia on isolated cells simply by addition of Ca.(ABSTRACT TRUNCATED AT 250 WORDS)

Iontophoretic localization of Ca-sensitive sites controlling activation of ciliary beating in macrocilia of Beroë: the ciliary rete.

Macrocilia are thick compound ciliary organelles found on the lips of the ctenophore Beroë. Each macrocilium contains several hundred axonemes enclosed by a single common membrane around the shaft of the organelle. Macrocilia are activated to beat rapidly and continuously in the normal direction by stimulus-triggered Ca influx through voltage-dependent Ca channels (Tamm, 1988). Heat-dissociated macrociliary cells are spontaneously active without depolarizing stimuli, providing Ca is present (Tamm, 1988). Here we investigate the spatial distribution of macrociliary Ca channels by iontophoretic application of extracellular Ca to different sites along quiescent, "potentially activated" macrocilia of dissociated cells in Ca-free medium. We find that Ca sensitivity for eliciting motility is highest or resides exclusively on the basal portion of the macrociliary surface. This is the first demonstration of local differences in Ca sensitivity along living cilia or flagella. The Ca-sensitive region coincides morphologically with a reticulum of unfused ciliary membranes at the base of the macrocilium. This ciliary rete is in direct communication with the surrounding sea water. It is likely that the ciliary rete provides the necessary Ca influx to trigger beating by virtue of its greater Ca conductance (i.e., density of Ca channels) and/or greater total membrane area.

Development of macrociliary cells in Beroë. I. Actin bundles and centriole migration.

Differentiation of macrociliary cells on regenerating lips of the ctenophore Beroë was studied by transmission electron microscopy. In this study of early development, we found that basal bodies for macrocilia arise by an acentriolar pathway near the nucleus and Golgi apparatus, in close association with plaques of dense fibrogranular bodies. Procentrioles are often aligned side-by-side in double layers with the cartwheel ends facing outward toward the surrounding plaques of dense granules. Newly formed basal bodies then disband from groups and develop a long striated rootlet at one end. At the same time, an array of microfilaments arises in the basal cytoplasm. The microfilaments are arranged in parallel strands oriented toward the cell surface. The basal body-rootlet units are transported to the apical surface in close association with the assembling actin filament bundle. Microfilaments run parallel to and alongside the striated rootlets, to which they often appear attached. Basal body-rootlet units migrate at the heads of trails of microfilaments, as if they are pushed upwards by elongation of their attached actin filaments. Near the apical surface the actin bundle curves and runs below the cell membrane. Newly arrived basal body-rootlets tilt upwards out of the microfilament bundle to contact the cell membrane and initiate ciliogenesis. The basal bodies tilt parallel to the flat sides of the rootlets, and away from the direction in which the basal feet point. The actin bundle continues to enlarge during ciliogenesis. These results suggest that basal body migration may be driven by the directed assembly of attached actin filaments.

Development of macrociliary cells in Beroë. II. Formation of macrocilia.

Two patterns of macrociliary growth occur in Beroë. Early differentiation described previously (Tamm & Tamm, 1988) leads to the first pattern of ciliogenesis. A tuft of 10-20 single cilia initially grows out from basal bodies that have migrated to the cell surface and are axially aligned. Ciliary membranes then begin to fuse along their length, except at the base, resulting in thicker groups of cilia on each cell. Progressive fusion of ciliary membranes, together with addition and elongation of new axonemes, finally results in mature macrocilia, 5 microns thick and 40 microns long, enclosed by a single membrane distally. The second pattern of ciliogenesis begins with the simultaneous appearance of several hundred ciliary buds on the apical surface. The short cilia possess individual membranes with bulbous tips, and are not axially aligned. Subsequent elongation is accompanied by progressive fusion of neighbouring ciliary membranes, except at the base, leading to flat-topped 'stumps' surrounded by a single membrane distally. Further elongation then proceeds asymmetrically within each stump. Axonemes on the aboral side of the macrocilium stop elongating, while those towards the oral side increase progressively in height, resulting in a slanted profile. Basal feet and central-pair microtubules are now uniformly aligned. Unequal elongation of axonemes on the oral and aboral sides of the macrocilium continues until the macrocilium resembles a lobster's claw, with a long slender shaft projecting from a broad base. Finally, the polarity of unequal growth reverses: the shorter axonemes on the aboral side elongate and almost catch up with the longer ones on the opposite side, resulting in a mature macrocilium of uniform diameter. The unusual membrane architecture of the macrocilium is thus a consequence of selective fusion of the distal regions of originally separate ciliary membranes. The polarized, asymmetrical growth of axonemes on the two sides of the macrocilium illustrates a remarkable control of microtubule elongation at the subcellular level.

A calcium regenerative potential controlling ciliary reversal is propagated along the length of ctenophore comb plates.

We have used the giant ciliary comb plates of ctenophores to record electrical activity directly from cilia. A compound action potential was recorded extracellularly over most of the length of the comb plate cilia in response to electrical stimulation of the ectodermal nerve net. The ciliary action potential was correlated with intracellularly recorded action potentials, selectively blocked by Ca2+-channel antagonists, and correlated with ciliary reorientation and reversed beating. Dual-electrode recording from different sites on the same comb plate showed that, unlike protistan cilia, the approximately 1-mm-long cilia of comb plates are not isopotential. Rather, action potentials are generated 150-200 microns from the base and propagate to the tip of the cilia. These results indicate that voltage-dependent channels that mediate increases in intraciliary Ca2+ concentration are distributed over most of the length of the cilia. Consequently, the Ca2+-sensitive machinery controlling ciliary motor responses is also likely to be located along the length of the axoneme.

Massive actin bundle couples macrocilia to muscles in the ctenophore Beroë.

Macrocilia are thick compound ciliary organelles arising individually from elongated epithelial cells on the lips of beroid ctenophores. A giant wedge-shaped bundle of microfilaments extends 25-30 microns from the base of each macrocilium to the lower end of the cell, terminating at a junction with an underlying smooth muscle cell. The broad end of the microfilament bundle is anchored to the macrocilium by striated rootlet fibers that extend from the basal bodies into the bundle and are linked to the microfilaments by periodic bridges. Fluorescence microscopy of rhodamine-phalloidin stained intact tissue, dissociated macrociliary cells, and Triton/glycerol-isolated bundles shows that the microfilaments contain actin. The microfilaments run generally parallel to the long axis of the bundle but are not highly ordered. Filaments decorated with myosin S1 show a uniform polarity with arrowheads pointing away from the tapered membrane-associated end of the bundle. No variations in bundle length (nor changes in rootlet periodicity) were observed in tissue fixed under conditions of calcium activation. Isolated bundles did not contract in Mg-ATP, even though detached macrocilia underwent reactivated beating and sliding disintegration. Macrocilia are used to bite through food organisms or transport prey into the stomach. The actin filament bundles probably play a supporting role as a structural linker between macrocilia and subepithelial muscle fibers and may serve as intracellular tendons to mechanically coordinate the motor activities of macrocilia and muscles during prey ingestion.

Electrophysiological control of ciliary motor responses in the ctenophore Pleurobrachia.

Prey capture by a tentacle of the ctenophore Pleurobrachia elicits a reversal of beat direction and increase in beat frequency of comb plates in rows adjacent to the catching tentacle (Tamm and Moss 1985). These ciliary motor responses were elicited in intact animals by repetitive electrical stimulation of a tentacle or the midsubtentacular body surface with a suction electrode. An isolated split-comb row preparation allowed stable intracellular recording from comb plate cells during electrically stimulated motor responses of the comb plates, which were imaged by high-speed video microscopy. During normal beating in the absence of electrical stimulation, comb plate cells showed no changes in the resting membrane potential, which was typically about -60 mV. Trains of electrical impulses (5/s, 5 ms duration, at 5-15 V) delivered by an extracellular suction electrode elicited summing facilitating synaptic potentials which gave rise to graded regenerative responses. High K+ artificial seawater caused progressive depolarization of the polster cells which led to volleys of action potentials. Current injection (depolarizing or release from hyperpolarizing current) also elicited regenerative responses; the rate of rise and the peak amplitude were graded with intensity of stimulus current beyond a threshold value of about -40 mV. Increasing levels of subthreshold depolarization were correlated with increasing rates of beating in the normal direction. Action potentials were accompanied by laydown (upward curvature of nonbeating plates), reversed beating at high frequency, and intermediate beat patterns. TEA increased the summed depolarization elicited by pulse train stimulation, as well as the size and duration of the action potentials. TEA-enhanced single action potentials evoked a sudden arrest, laydown and brief bout of reversed beating. Dual electrode impalements showed that cells in the same comb plate ridge experienced similar but not identical electrical activity, even though all of their cilia beat synchronously. The large number of cells making up a comb plate, their highly asymmetric shape, and their complex innervation and electrical characteristics present interesting features of bioelectric control not found in other cilia.