Electrophysiology and Co-ordinated Behavioural Responses in the Colonial Bryozoan Membranipora Membranacea (L.)

Membranipora membranacea belongs to the largest class of living Bryozoa (= Ecto-procta), the Gymnolaemata, and to the dominant order, the Cheilostomata. It forms extensive flat colonies on the fronds of various kelps (Laminaria spp.). The zooids are elongate and rectangular, each measuring about 800 × 300 × 200 μm. They are arranged quincuncially, or as bricks in a wall, and each is thus contiguous to six others. All adjacent zooids are inter-connected by a pair of communication organs (pore plates) in the dividing wall.

Bryozoa are filter feeders depending on an extensible lophophore of slender ciliated tentacles for the collection of food particles. This apparatus is delicate and presumably highly vulnerable to damage. The main defensive mechanism is the rapid withdrawal of the lophophore into the zooecium and (in cheilostomes only) the subsequent closure of the operculum (see Fig. 1). In Membranipora and other anascan cheilostomes, extension of the lophophore is associated with eversion of the tentacle sheath. This follows the increase in hydrostatic pressure caused by the downward pull of the parietal muscles on the frontal membrane.

Until recently very little was known about the nervous systems of Bryozoa and there was some disagreement between authors (Nitsche, 1868; Gerwerzhagen, 1913, Marcus, 1926; Graupner, 1930; Bronstein 1937). Moreover there was no work of note on the Cheilostomata. The basis of present knowledge is the recent detailed studies by Lutaud (1969, 1971, 1973) on the anascan cheilostome Electra pilosa (L.).

Each zooid contains a circumpharyngeal nerve ring with a well developed ganglion on the dorsal (anal) side. From this ring at least four nerves run up into each tentacle (Lutaud, 1973; Thorpe & Laverack unpublished) and other nerves connect with various organs. Gordon (in press) states that in Cryptosula pallasiana there are six nerves running up each tentacle.

Marcus (1926) in a major study concluded that there was no nervous connexion between zooids and that no zooid responded to a stimulus applied to a neighbour. There has been no work of substance to contradict this, although Bronstein (1937) reported colonial responses in Bowerbankia and interzooidal connecting nerves in the stolons of this species. Recent work using the electron microscope has failed to confirm the existence of any nerves in the stolons of this species (personal communication - D. P. Gordon). Hiller (1939) wrote a short note claiming to have observed interzooidal nervous connexions in Electra pilosa. Reviewers have regarded this evidence as insufficient for them to disagree with the conclusions of Marcus (1926) (Hyman, 1959; Brien, 1960; Bullock & Horridge, 1965; Ryland, 1970).

In 1969, however, Lutaud, using methylene blue staining, produced evidence of what appears to be a colonial nervous system in E. pilosa (Fig. 2). She described in each zooid a nerve which runs around the basal border of the interzooidal walls and connects via the pore plates with a similar nerve in each adjacent zooid. This network is connected to the ganglion of each zooid via two nerves running along the tentacle sheath. Ryland (1970) concluded that the physiology and role of this nervous system required investigation; and such was the main purpose of our work.

The species Membranipora membranacea was chosen because it is readily available locally throughout the year and the large, flat colonies are convenient for experimentation. The closely related Electra pilosa used by Lutaud (1969), Marcus (1926) and Hiller (1939) was rejected because of the small size of the colonies compared to those of Membranipora, although much of our work was subsequently repeated on this species.

The colonies were collected by pulling up fronds of Laminaria from the sublittoral zone. New stocks had to be obtained every two or three days because survival in the laboratory was poor even in running sea water.

For these experiments parts of colonies of Membranipora on Laminaria were cut off and mounted, using pins, on a sheet of cork glued to the bottom of a dish of cold sea water. The sea water in the dish was changed at frequent intervals to maintain a temperature of about 10 °C. Every effort was made to keep the period of emersion as short as possible when transferring colonies to the dishes and also to avoid stimulating them by vibrations during experiments.

Mechanical stimulation was given using a glass micropipette, of tip diameter about 20 μm, mounted on a micromanipulator.

Electrical stimulation was by means of a silver stimulating electrode of tip diameter about 100 μm. A Tektronix type 161 pulse generator was used in conjunction with a Tektronix type 162 waveform generator to give square-wave pulses, the amplitude and duration of which could be adjusted as required. After placing the electrode against the surface membrane of a zooid, time was given for the colony to recover from possible effects of the mechanical disturbance before any electrical stimulus was given.

To find out what effect they had on the transmission of the colonial response, incisions were made in the colony, using a new scalpel blade. After cutting, the animals were left to recover before experimentation was commenced.

The most instructive technique for measuring lophophore retraction times would probably be high speed cinephotomicrography, but as suitable equipment was not available an alternative method was devised.

Because of the microscopic size of the zooids, any apparatus using physical means to monitor movement was unsuitable. An optical method was, therefore, employed. Zooids were so positioned that, when the lophophore was extended, a light beam to a photodiode (Texas Instruments type H-38 NPN photo duo-diode) was partially obscured. The resulting change in light intensity altered the resistance of the photo-diode, which produced a variation in the voltage drop across it. This voltage change was amplified and then monitored on a pen recorder and oscilloscope.

The apparatus was built around a binocular microscope with a moving stage. The photodiode was positioned at the focal point of one eyepiece, which was about 1·5 cm above the lens. (Modified from Campbell, 1972.)

Strips of Membranipora on Laminaria were cut to about 3 cm × 0·2 cm. These were mounted in a glass bottomed Perspex cell filled with cold sea water. The animals were held in position by placing the ends of the strips into slots cut in the Perspex. These were so positioned that the light beam to the objective lens was parallel to the surface of the colony and at right angles to any extended lophophores. Extended lophophores were located using the free eyepiece (which did not have the photodiode attached to it).

This apparatus was subsequently modified by the addition of a stimulating electrode. The input to this was also fed into the pen recorder, thus making it possible to measure the time elapsing between stimulus and response (reaction time).

Standard polythene suction electrodes, modified from McFarlane (1969), of tip diameter about 100 μm were positioned on the frontal membranes of individual 2ooids. Recordings, amplified by differential pre-amplifiers and monitored on oscilloscopes (Tektronix types 564B and 561B), were made of both spontaneous and stimulated activity from various points on the colony. Electrical stimuli were given using suction electrodes.

It was found that mechanical stimulation of an extended lophophore of Membranipora resulted solely in the retraction of that lophophore. If, however, not the lophophore but the frontal membrane was stimulated, the result was the immediate rapid withdrawal of all the extended lophophores within some distance of the zooid stimulated. The response was the same to electrical stimuli.

The shape of the area covered by the response is very variable but is often approximately as shown in the accompanying diagram (Fig. 3), being greatest in a direction parallel to the longitudinal axis of the zooids.

The area affected increases with the strength of the stimulus up to a maximum of about 10 cm by 5 cm. This was never exceeded in the colonies we studied.

The time taken by a colony to re-extend its lophophores following a single stimulus depends on the extent of the response, but is usually not more than about 2 min. After a colonial response it was always the zooids furthest from the point of stimulation which re-extended their lophophores most quickly, and those nearest which reappeared last. Those on the edge of the area of a response often withdrew only momentarily and reappeared almost immediately.

The sensitivity of the animals to mechanical stimulation depended also upon the part of the surface membrane touched. However, it was difficult to investigate this because of the extremely small size of the zooids. The tentacle sheath, the area of the frontal membrane immediately surrounding this and the operculum appeared to be the most sensitive, with no great variation in sensitivity over the rest of the membrane.

Another factor governing the size of the colonial response was the velocity of the stimulus. A rapid mechanical stimulus produced a greater response from the colony than a slow one, although the degree of membrane deformation was about the same in each case. If the stimulator was left in contact with the membrane, the animals reextended their lophophores after a short time.

In the case of electrical stimulation, the threshold values for the duration and voltage of the stimulus were found to vary considerably, depending on the state of the colony. Similarly the response to stimuli of fixed voltage and duration varied between stimuli. The threshold voltage was typically about 1·5 volts at 10 msec.

Whether a decrease or an increase was obtained in the area of a colonial response was found to depend upon the size and frequency of successive stimuli. The results were comparable for both mechanical and electrical stimuli.

If stimuli just above threshold at short interstimulus intervals (e.g. one or two seconds) were used, a marked initial increase in response was shown. The response to the second stimulus was far greater than that to the first, and was maximal or near maximal. Subsequent stimuli produced very little increase in the area of response (Fig. 4a). If stimuli well above threshold were used, a near maximal response was obtained to the initial stimulus, and subsequent stimuli had little or no further effects (Fig. 4b). If the stimuli were continued the response diminished and eventually most of the polypides re-emerged.

With long interstimulus intervals (e.g. 2 min) and stimuli well above threshold there was no increase in response area and duration following the second stimulus (Fig. 5) On the contrary, the area and the duration of the response following successive stimuli were found to decrease. With stimuli just above threshold at long interstimulus intervals, the response to each stimulus was about the same as to the one before it.

As stated earlier, after a burst of stimuli well above threshold at long interstimulus intervals, the colony ceases to respond. If, however, the point of stimulation is moved from the zooid previously stimulated to one of its immediate neighbours there is a sudden partial return of excitability with the response increasing to about half of that obtained to the initial stimulus. With successive stimuli this response also diminishes (Fig. 6).

In the case of either electrical or mechanical stimulation, a cut across the colony will prevent the spread of a colonial response. The response will spread to a limited extent around the end of a cut (Fig. 7).

Using the light beam apparatus it was found that the time for the complete retraction of the lophophore in Membranipora membranacea was usually 60–80 msec. This is the normal withdrawal time for the ‘escape’ response. Other withdrawals at slower speeds were recorded, taking anything up to 500 msec. The animals are also capable of stopping the withdrawal before completion and of changing the speed of a withdrawal whilst it is taking place. The great majority of recorded withdrawals were, however, of the very rapid type.

Extension rates are far more variable, frequently not at an overall constant speed and may last from about 200 msec up to a few seconds.

With Electra pilosa the results obtained were very similar, although the fast withdrawals were slightly slower than those of Membranipora.

Often when several lophophores were in the field of view they would all be retracted apparently in synchrony. The time taken for this colonial response was the same (approximately) as that for a single zooid. Again this was also found to be the case with Electra.

The latency between stimulus and response in Membranipora is about 20–30 msec. This increases by about 10 msec cm⁻¹ with increasing distance of the zooid from the point of stimulation.

Two types of electrical pulses were recorded from Membranipora. These we have named ‘type one’ (T1) and ‘type two’ (T2). Very similar pulses can also be recorded from Electra pilosa.

T1 pulses are of very short duration (about 3 msec) and are biphasic with an amplitude of about 10 μV (Fig. 8). They occur occasionally singly, but more often in bursts, during which the frequency may reach a maximum well in excess of two hundred pulses per second.

Simultaneous recordings at various points on the colony show that T1 pulses are conducted between zooids over an area greater than that covered by the colonial response. There is a short delay between the arrival of T1 pulses at different recording electrodes. From this the conduction velocity can be shown to be about 100 cm sec⁻¹ in a direction parallel to the longitudinal axis of the zooids and about 50 cm sec⁻¹ at right angles to this. Similar figures for conduction velocities are obtained using the delay between stimulation and the arrival of the first Ti pulse at electrodes a known distance from the point of stimulation.

In response to an electrical stimulus large numbers of T1 pulses are produced at high frequency (Fig. 9). The variation of frequency with time is shown in Figs. 10 and 11. The T1 frequency is initially high; this frequency then increases slightly (up to 250 pulses sec⁻¹) followed by a decrease to a tonic level. There is subsequently a relatively sudden drop in frequency with the discharge eventually ceasing entirely. The bursts with the greatest total duration are those with the highest initial T1 frequency.

There is a significant decrease in both overall and peak frequency (Fig. 12) of bursts with increasing distance from their point of origin. This effect is most pronounced with short bursts. With successive stimuli the discharge is continued for a markedly decreasing length of time (Fig. 13).

With a stimulus interval of 2 min the T1 bursts in response to each stimulus terminates before the next stimulus arrives and the number of T1 pulses decreases with successive stimuli. In an experiment in which repetitive stimuli were given (2 msec, 25 V) at 630 msec intervals, the T1 response to the first shock was not completed by the time the second shock arrived. The peak T1 pulse frequency after the second stimulus increased by twenty-five per cent over the peak T1 pulse frequency following the first shock. There was a further small increase after the third shock. The peak Ti frequency after the fourth shock was the same as that after the third but the fifth, sixth and seventh shocks produced a steadily declining response.

T2 pulses (Fig. 14) are of much greater duration (100–120 msec) than Ti pulses and consist of two distinct parts. The first part is a biphasic pulse of large amplitude (up to 200μV) and short duration (5–10 msec); the second is not biphasic and is of small amplitude (10–15 μV), but long duration (about 100 msec).

It can be shown by observation that a T2 pulse is recorded whenever the lophophore is retracted. It is also recorded when, as often happens, the retracted lophophore is pulled further down inside the tentacle sheath.

T2 pulses can normally be seen to accompany high frequency T1 activity, but occasional ‘spontaneous’ T2 pulses may occur when the number of T1 pulses recorded is few or none.

The possibility that the second part of the T2 pulse is an artifact caused by movement of the frontal membrane of the zooid under the electrode tip was investigated. It was established that the amount of movement necessary to produce an artifact of the same amplitude and duration as the recorded pulse was far larger than any movement that could have been produced by the animal.

The present results show that there is a colonial retraction response in both Membranipora membranacea and Electra pilosa. The conclusion of Marcus (1926) that no zooid responds to a stimulus applied to a neighbour holds only in the case of stimuli applied to a single lophophore.

Stimulation of the surface membrane does produce a colonial response. There are several possible methods by which this response may be coordinated, but the most obvious and in our view most likely is that there is a colonial nervous system, probably as described in Electra by Lutaud (1969). This links contiguous zooids by through-conducting nerve pathways.

An alternative hypothesis is that hydrostatic pressure changes caused by retraction of the lophophore are transmitted between zooids through holes or thin membranes in the rosette (pore) plates in the interzooidal walls. Using a scanning electron microscope (Thorpe and Laverack, unpublished) no apparent open pores were observed in the rosette plates or in any other parts of the interzooidal walls. Similarly transmission electron-micrographs of pore plates (Thorpe and Laverack, unpublished) do not show any open pores or possibly flexible membranes. As described by Banta (1969) the rosette plates are rigid and calcified with balls of cells completely occluding the pores. No hydrostatic pressure sensors have been described in Bryozoa and there is no evidence to suggest their presence. Finally, the involvement of pressure changes appears to be disproved by the fact that a response can be initiated by stimulating the membrane of a marginal zooid still lacking a functional polypide.

Alternative hypotheses involving the mechanical stimulation of neighbours by a retracting lophophore do not explain either the failure of the response to cross cuts in the colony or how such a response can traverse areas of retracted zooids. Only the presence of a colonial nervous or alternatively a neuroid system could adequately explain the limitation of a response to a particular area.

The presence of a neuroid system cannot be discounted, but Bullock & Horridge (1965) have pointed out the extreme difficulty in determining unequivocally whether nervous or non-nervous pathways conduct recorded electrical activity. Many epithelia have structural modifications between adjacent cell walls but on its own this property is not enough to prove that such epithelia conduct spiking activity. As yet, there is very little data available on the fine structure of epithelia in Membranipora. There is good evidence for a colonial nervous system, however (Lutaud, 1969), and the results reported here accord closely with the expected properties of that system. In the absence of further data, we suggest that the T1 system we have recorded from corresponds to the histologically demonstrated nerve plexus. Our arguments in favour of this view are set out in the following paragraphs.

The shape of the area responding to a stimulus applied to one zooid (Fig. 3) is consistent with predictions based on Lutaud’s (1969) anatomical studies, if bidirectional polarization is assumed, with the greatest polarization along the longitudinal axis of the zooids. This assumption is also consistent with the variations in conduction velocity of T1 pulses in different directions across the colony. The large variation in shape of the area of response is probably due to local variations in the sensitivity of different parts of the colony affected by any one stimulus.

The non-linear increases in the area of the colony responding to successive stimuli are not easily explained in terms of a simple nerve network. If the ‘all or none’ nature of neuronal conduction is accepted, then most simple models which have been proposed for nerve nets (Pantin 1935; Ramsay 1952) and even the complex computer model of Josephson (1964) will not account for the observed results in Membranipora.

Some of our results are, however, very similar to those obtained by Horridge (1957, 1968) working on the spread of colonial excitation in Madreporarian and Alcyonarian corals. He concluded that such responses could be explained by multiple firing from the point of excitation in response to a single stimulus. The small increase in the response to the second and subsequent stimuli, which Horridge found in Porites and we have found in Membranipora, occurs, Horridge suggests, because a rapid burst of impulses follows the initial stimulus and subsequently the probability of repetitive firing in the nerve net diminishes.

Horridge explained ‘facilitation ‘, during which there is a more than linear increase in response to successive stimuli, in terms of repetitive firing at other places on the nerve net as well as at the point of stimulation. In Membranipora, the increase in response to stimuli just above threshold can be explained more simply as the summation of the two stimuli (or conversly a lowering of threshold in response to the first one).

The partial return of the colonial response on changing the point of stimulation from one zooid to one of its immediate neighbours is compatible with Horridge’s (1968) model. Decreases in the area and duration of colonial responses are caused by a lessening of repetitive T1 pulse firing from the zooid stirnulated; thus there should be a substantial increase in response if another zooid is stimulated. This is, in practice, found to be the case.

The effect of cuts across the colony in Membranipora resembles their effect on epithelial conduction in echinoderms. Kinosita (1941), Smith (1950), Bullock and Horridge (1965) all indicated that in certain asteroids and echinoids epithelial transmission occurs in straight lines, and it is a matter for debate as to how cuts are circumnavigated, if at all. In both the Echinodermata and the Bryozoa the ability of nervous activity to travel around the ends of cuts may be taken to indicate the presence of a nerve net of some kind. In Bryozoa such a nervous system has been described by Lutaud (1969).

The contraction times for the lophophore retractor muscle in Membranipora are themselves of interest since the peak contraction rate is in excess of twenty times their own length per second. This most exceptional rapidity is less surprising, however, when the extremely small size of the animal is taken into account. It would not be surprising if these muscles showed interesting adaptations to contraction at high velocity under very light load. From electromicrographs (Gordon, in press; Thorpe and Laverack, unpublished) the tissue appears to be smooth muscle.

The latency of the contraction response is apparently about 25 msec for zooids at a negligible distance from the point of stimulation. For zooids further away this is increased by the time taken for the stimulus to be transmitted to the animal. This is approximately 10 msec cm⁻¹ (or 100 cm sec⁻¹).

The properties of the T1 system lead us to believe it is nervous for the following reasons:

• (i) T1 pulses are of very short duration: 3 msec.

• (ii) they are conducted across the colony, arguing for a system which links all, or many, zooids together.

• (iii) the conduction velocity (100 cm sec⁻¹) is the same as that calculated for the speed of transmission across the colony, as measured by lophophore retraction.

• (iv) they are generated after electrical and mechanical stimulation.

• (v) they reach high frequencies (occasionally > 200 sec⁻¹) and occur in bursts of varying length.

We suggest that T1 pulses are nerve pulses travelling through the colonial nervous system. They are conducted across the colony at the same rate (100 cm sec^𢈒1) as the spread of lophophore retractions. They are produced in response to stimulation and apparently influence lophophore retraction. The reduction in the number of T1 pulses following successive stimuli corresponds with the reduction of the colonial response. They are produced in a burst (i.e. multiple firing) in response to a single stimulus as predicted from the model proposed by Horridge (1968). The fit with this model is not complete however since the number of pulses in a burst does not decrease linearly with distance from the point of stimulation and T1 pulses can be recorded outside the area of the colonial response. Clearly, therefore, Horridge’s model is not adequate to explain our results.

It seems probable that in Membranipora the spread of response is limited in some way by the reduction in peak frequency and increase in total duration of a burst of Ti pulses as it travels across the colony.

The T2 pulses are probably of a compound nature. The initial phase is rapid (5–10 msec) and of large amplitude (up to 200 μV) whilst the latter stage is comparatively slow (about 100 msec). In our opinion these two responses are due to the motor innervation and subsequent contraction of the lophophore retractor muscle. T2 pulses occur whenever a lophophore is retracted and their latency is similar to that of lophophore retraction. The retraction time of the lophophore is somewhat faster than the second phase of a T2. This difference can, however, be accounted for if allowance is made for the extra time taken to retract the lophophore further down inside the zooecium after it has passed through the orifice, and continuation of electrical activity in the muscle after the cessation of visible contraction.

The initial phase of a T2 occurs only before a muscle contraction and has such a short duration that it would seem to be a result of nervous activity. This activity could be the firing of a motor nerve innervating the lophophore retractor muscle, probably running from the suprapharyngeal ganglion. Its very large amplitude strongly suggests that some kind of giant fibre system is involved. A structure which could be a giant axon is described by Lutaud (1971) in the main tentacle sheath nerve in Electra.

There are several unusual features shown in the electrophysiological responses of Membranipora (and Electra). The high frequency of firing in the colonial nervous system (200 +sec⁻¹) is unusual but not unique. The very long periods (up to 10 sec) for which these small nerves will continue to fire at high frequency are most surprising. This suggests that the axons of the colonial nervous system have an exceptional degree of tolerance to changes in internal ionic concentration.

The activity of an individual zooid of Membranipora consists mainly of extension and retraction of the lophophore. This movement, however, is now seen to be dependent on the integration of two distinct nervous systems.

Each zooid has a well-developed ganglion and its own discrete nervous system. At the same time the zooid is connected to its neighbours and is influenced by them via the colonial nervous system. The contraction of the lophophore retractor muscle is controlled, probably via a giant axon, by the ganglion. Whether the lophophore is extended or withdrawn depends upon the input to the ganglion of T1 pulses from the colonial nervous system and also information from the zooidal sensory nervous system.

In addition it appears that the physiological state of the ganglion can change with time. Responses of individual zooids are therefore governed by both internal and external conditions.

The great regularity of the T1 pulses in a burst suggests that each ganglion has cells capable of acting as pacemakers. Since no extra pulses are found during a burst it is also probable that the firing of the colonial nervous system by one ganglion inhibits the production of any other pulses of lower frequency from other ganglia. Therefore the same pulses within the colonial nervous system may simultaneously produce both excitatory and inhibitory effects on other ganglia.

We wish to thank Dr J. S. Ryland for reading and criticizing the manuscript, Drs J. L. S. Cobb and I. D. McFarlane, Mr P. R. Balch and the staff of the Gatty Marine Laboratory for their assistance.

G.A.B.S. is supported by a Science Research Council Research Studentship.