Osmotic Relations in Molluscan Contractile Tissues: II. Isolated Gill Preparations from Lamellibranchs (Mytilus Edulis L., Ostrea Edulis L., Anodonta Cygnea L.)

The effect of osmotic pressure on ciliary activity was studied by Frédericq (1922) who subjected Mytilus cilia to both hypo-and hypertonic sea water; he found that ciliary movement persisted 30 min. after immersion in 49–167% sea water, ^† but ceased in 37% and 185%; after 16 hr. there was still movement in 49–126% but none in 167%. Gray (1922) made a few remarks on the effect of osmotic pressure, noting that ‘… the stoppage of cilia in hypertonic solutions is brought about in an entirely different way to the stoppage in an acid solution, and it is therefore not surprising to find that the stoppage in hypertonic solutions is not influenced by the presence of hydroxyl ions…’. Wells, Ledingham & Gregory (1940) studied the action of hypotonic sea water on the cilia of various species of animals, including M. edulis. They found that sudden salinity changes produced ‘Shock Reactions’ followed by accommodation. They also noted that the different types of cilia on the Mytilus gill varied in sensitivity to osmotic pressure, and that disintegration of the gill occurred at salinities where most of the cilia were still beating; the latter observation may be of ecological significance, as disintegration may prevent the efficient working of the gill as much as immobilization of the cilia may do.

The writer’s experiments were undertaken (i) to extend the above observations into the hypertonic range, (ii) to determine if possible how far the effects of salinity change depend on the initial concentration, on the final concentration, or on the extent of the change, and (iii) to obtain comparative data on lamellibranchs of different ecological salinity relations.

Single filaments were isolated by needles from the inner gill-lamellae and supported by glass threads on the underside of a slide over a capsule holding 40 ml. sea water; the activity of the various ciliated tracts was observed under a microscope with a $12$ in. objective. To change the bathing solution the slide was transferred to another capsule containing the fresh medium. Each experiment was conducted on filaments from a single gill and activity was rated normal (+ + +), reduced (+ +), weak (+), nil (o), or hyperactive (+ + + +). This method is subjective to some extent, in that in observing the accommodated beat after many hours it was difficult accurately to estimate slight changes from the initial beat ; it was easy, however, to assess the relative frequency between cilia types after treatments since these observations were all made within a few minutes of each other. To obtain a more quantitative expression of the shock reaction in later experiments the times were noted at which the different cilia ceased and resumed beating.

The frontal, latero-frontal, food-groove and short abfrontal cilia retained their normal activity for at least 24 hr. after isolation in natural sea water; the long ab-frontals were usually weaker after an hour and only feebly active in 24 hr. (though often quite vigorous if the capsule was shaken). The laterals were variable but often survived at their normal level of activity for 2–8 hr. and occasionally short lengths of the tracts appeared unaffected after 24 hr.—this clearly disagrees with Gray (1928, p. 99) who states that the laterals come to rest in sea water, a ‘.. medium containing more magnesium than is present in the blood’. Analyses of Mytilus blood have since shown that the magnesium content is 97% that of sea water (Krogh, 1939, p. 56; M. edulis) and 96·5% (Robertson, 1950, personal communication ; M. gaUoprovincialis’), so that in this respect sea water appears to be a suitable bathing medium and Gray’s results must be re-interpreted on some other basis. (It will be shown later that lateral cilia can withstand 150% sea water for several hours, though usually at a reduced level of activity.)

Survival of the different tracts was as satisfactory in artificial as in natural sea water; individual filaments showed slightly better survival in either natural or artificial sea water during the same experiment with no obvious ‘preference’ over the course of the investigation.

On sudden immersion in artificial sea water of more than 250% all cilia ceased beating within 1 min. and failed to recover; with less hypertonic solutions the different tracts showed greater or lesser degrees of salinity tolerance. The long-term results are expressed in Fig. 1, which disregards the temporary effects of change and shows the extent to which the cilia finally accommodated to the hypertonic medium and the limiting concentration in which only weakly active cilia (+) were found. By reading downwards the differentiating effect of any one concentration may be readily seen. In the case of the latero-frontals the individual cilia appeared fully active in 150%, but the metachronal rhythm was often somewhat disturbed; a similar disturbance occurred with the laterals in all hypertonic media. (Gray, 1928, p. 128, remarks that ‘The regulation which gives rise to the metachronal rhythm may quite likely be (and even probably is) independent of mechanical activity…’.)

The most resistant cilia are therefore those of the frontal, short abfrontal and food-groove tracts; the long abfrontals and latero-frontals are roughly equivalent, while the laterals are decidedly the most sensitive to an increase in osmotic pressure. This series conforms well with that described by Wells et al. (1940, p. 380) who found the same hierarchy of sensitivities on dilution.

As was the case with heart preparations some types of cilia showed a shock reaction on first meeting a change of medium. In those concentrations in which full accommodation was later achieved, the cilia usually suffered no shock reaction or obvious reduction in activity (the long abfrontals are the chief exception to this, but they are an unsatisfactory type to study quantitatively since they are, even more than the laterals, so susceptible to mechanical stimuli) ; in the higher concentrations all types of cilia except short abfrontals slowed down before accommodating to lower levels of activity ; this slowing down culminated in a temporary inhibition at the highest concentrations compatible with even weak activity.

Having established that some types of cilia survived well in 150% sea water, the way was clear to conduct an examination of some of the fundamental factors producing the dilution effects. As was mentioned above, a few workers have described the effects of, and limits of survival in, hypotonic sea water, but their experiments were begun in 100 % sea water. By running parallel experiments on filaments from the same gill, commencing with some filaments accommodated in 150% as well as others in 100% artificial sea water, then diluting to various concentrations, it was thought possible to decide if the effective factor was either (1) the final concentration, or (2) the arithmetic difference between initial and final concentrations, or (3) the ratio final/initial concentration.

For example, a typical experiment included the preparations shown in Table 1 :

The most critical results were obtained on the shock reaction, as the time of onset and duration of inhibition could be determined with a stop-watch. After some practice, observations could be begun within 4 sec. of changing the slide to the fresh medium.

Frontal cilia subjected to a drop of too to 30% usually showed an inhibition lasting 3–4 min. The 150 to 80% treatment never caused inhibition, while 150 to 30 % caused inhibition for periods of 4–14 min. ; the preparations with an inhibition time most resembling the 100 to 30% were 150 to 45 % and 150 to 40%. Similarly, 100–40% did not cause inhibition, whereas 150 to 40% did.

Latero-frontals were inhibited for much longer periods and, in the case of large drops in concentration, often failed to recover; it was clear that neither 150 to 80% nor 150 to 30% represented situations analogous to 100 to 30%. On the other hand, 150 to 75% had approximately the same effect as 100 to 50%, while 150 to 50% and 150 to 100% were quite distinct.

Short abfrontals were not stopped by either 100 to 30% or 150 to 30% ; 100 to 25% did not cause inhibition, but 150 to 25% did for 10–25 min.

These observations show that while exact correspondence of inhibition times was not always obtained there was sufficiently close agreement to conclude that neither final concentration nor arithmetic difference between initial and final concentrations is of itself the effective factor determining the shock reaction behaviour; instead, the ratio ‘final/initial’ concentrations is decisive. The same conclusion appears to apply to the behaviour of the cilia after accommodation to the new medium. The activities of the cilia at 2–4 hr. after immersion in the hypotonic sea water were most differentiated in the extreme cases tested, namely the 100 to 30% drop; no preparation diluted from 150% matched the former exactly in all activity ratings, but from a number of experiments it was clear that the moss-parallel treatment was 150 to 45%, that 150 to 80% was much more mild and 150 to 30% more severe in their effects; in other words, concentration changes approximating to the same ‘final/initial ‘ratio led to quantitatively similar effects.

Wells et al. (1940, p. 383) noted that in some cases (the short abfrontal cilia of Mytilus as well as polychaete muscles) ‘… the inhibition evoked by downward change is preceded by a phase of excitation’ ; the phenomenon occurred after great changes (100 to 20%). It was here found present also in the frontal and food-groove cilia after the changes 100 to 50% or lower. Changes between 150 to 100% and 150 to 30%, inclusive, usually caused increased activity in these same three tracts for periods up to 4 hr., though in some of the extreme changes this phase was suppressed or was perhaps less than 4 sec. and so overlooked. The phenomenon is related, therefore, to a change in concentration rather than the final medium encountered. The effect does not seem to be related only to downward changes as it may occur in all three tracts after an increase in concentration; this compares with the observation of Wells & Ledingham (1940) who state that Nereis diversicolor body wall is at first excited on returning from hypotonic to normal sea water.

With extreme treatment numbers of ciliated cells became freed from the gill and even whole patches of the epithelium loosened. This was presumably due to a dissolution of the intercellular matrix (see Discussion), the ciliary activity helping to shake the cells free (particularly noticeable with latero-frontals). Such disintegration did not occur in 100% sea water after 24 hr. but was marked in hypertonic media; it was noted by Wells et al. in sea water less than 35%, so that it is a phenomenon found in concentrations at both ends of the physiological range; furthermore, in the present work on two changes of media the curious fact emerged that the extent of disintegration was related to the ratio of concentrations just as was ciliary activity (above). No attempt was made to establish an absolute scale of disintegration, but individual filaments during the same experiment could be easily compared.

Despite a wealth of publication on feeding in oysters by ciliary currents very little has appeared on the physiology of the cilia themselves. Tomita (1934a) showed for O. circumpicta that mechanical (ciliary) activity of pieces of the gill was dependent on the concentration of the sea water medium; and also (19346) that there was little effect on ciliary activity on increasing the hydrogen-ion concentration until pH 5-5 (approx.) when a sudden decline occurred.

The gills of Ostrea have much interfilamentar tissue making the isolation of single filaments impracticable. There is a further disadvantage with this animal in that the gills are heterorhabdic and plicate, the folding making it difficult to examine cilia lining the water-pores. Preparations were made using pieces from all four gills since they are similar in their structure and ciliation (see Atkins, 1937, type B (1 b)), and no difference was found in either their ease of preparation or their reactions to various media. The right valve of the shell and the right mantle lobe were removed and the gills divided into sections as in Anodonta^* one lamella of each section was cut off, usually in small strips since the interfilamentar tissue of the grooves was seldom strong enough to hold all the filaments together during this operation. The sections of lamellae left in situ, however, were in about half of the attempts reasonably entire; they were cut off dorsally and mounted as for Anodonta, but using sea water. Unfortunately the preparation was fragile, tending easily to part along the grooves on stretching or, even if mounted satisfactorily, to undergo concertina-like movements by means of longitudinal muscles and so obliterate the water-pores or cover them and the grooves by folding. In most parts of the gill the apical and ordinary filaments are firmly united by a mass of tissue forming them into the shape of the plical folds, whereas the principal and two transitional filaments are chiefly united to this mass by small interfilamentar junctions and only sparsely by more substantial horizontal septa. This mass of tissue in the ridges of the plicae precluded clear observation on the cilia within the water-pores between apical and ordinary filaments so that attention was directed to the principal filament at the bottom of each groove. Providing these filaments had not been split or tom from their neighbours (transitionals) during mounting and were not tom subsequently as a result of muscular contractions, it was possible to observe lateral and latero-frontal^† cilia along the principal filaments. Frontal^‡ cilia were observed by the same device described for Anodonta though they were, perforce, those of the apical filaments; in some cases it was possible to observe the frontals of the ordinary filaments, but this could not be relied upon. On the whole the preparation, although the best obtained after many attempts, was not a very satisfactory one and only served to determine the range of salinity the three tracts could withstand; shock reactions were not studied since the tissue holds so much solution that concentration changes alongside the cilia might be slowed down and not be constant between preparations.

All three types of cilia were capable of maintaining normal activity in natural sea water for at least 24 hr. (in one case frontals showed undiminished activity 44 hr. after mounting in the capsule).

After sudden transfer to dilute sea water the frontals and latero-frontals survived as well and as long in 80 and 60% as their controls in 100% ; in 40% survival was as long but often at a reduced level of activity, while all three types failed to recover in 20 and 10% sea water after 1–2 min. Laterals were the most sensitive cilia, their activity diminishing even in 80% while they were usually inhibited in 40% ; as in Mytilus these cilia were not very regular in their behaviour, mechanical disturbances setting them beatmg from an inactive state. A clue to this inconstancy was seen in scanning the whole preparation when it was evident that in the same piece of gill laterals of one principal filament might be beatmg while those of another were at rest. It was usual for the laterals from end to end of any one filament either to beat or not in all the species investigated, i.e. the discrepancy was between entire filaments and so was not detected in Mytilus using single filaments.

For experiments with hypertonic sea water, further controls were run in 100% artificial sea water; they gave results equivalent to those in natural sea water. All cilia failed to recover in 200 % sea water except for very occasional quivering among the frontals after about an hour. Frontals survived in 125 % as well as in 100% and in 150% with reduced vigour. Latero-frontals, and especially laterals, were affected by even 125%.

Swelling, followed by disintegration, occurred with all preparations placed in 20 and 10% ; disintegration alone occurred in 200% and, to a slight extent, in 150% ; these hypertonic media usually causing the tissue to become distinctly granular in appearance and less transparent. This was probably the result of dehydration of the cells with consequent disturbance in the physical state of the proteins and other colloids of the protoplasm. The swelling in hypotonic solution also resulted in constriction or even obliteration of the water-pores and other spaces between the filaments so that the laterals and latero-frontals were obstructed mechanically; as this swelling was most marked in sea water of less than 40% however, it is not considered that it affected the results for these tracts since frontals, shown to be the most hardy, had always ceased to beat in 20% without any indication of similar mechanical hindrance.

The salinity relations of Ostrea cilia appear from these results to resemble those of Mytilus cilia, except that the oyster cilia are less resistant to salinity change. This difference falls in fine with that already found in the heart, and with the ecological relationships of the two species (Pilgrim, 1953).

In contrast to the numerous published researches on other aspects of fresh-water mussel physiology, the study of their gill-cilia has been but little investigated. Ringer & Buxton (1885), in the course of the former’s now classical studies on the importance of various ions to tissue viability, noted that gill-cilia of Anodon (= Anodonta) ceased to beat in distilled water within 24 hr., the cells having swollen and separated from the gill surface. Segerdahl (1922) found that cilia on isolated gill fragments kept beating up to 5 days in tap water and up to 9 days in 0·3 % NaCl, though these were extreme cases. Zweibaum (1925) made tissue cultures from Anodonta gill, using diluted Frog-Ringer as a culture medium; the optimum concentration was R/3 (Δ = 0·15° C.) which maintained the preparations up to 63 days, while in Δ = 0·o5°C. and Δ = o·38°C. survivals averaged 16 days and 13 days respectively.

Owing to the complex eulamellibranchiate gill structure of Anodonta with the interfilamentar junctions stronger than the interlamellar it was not practicable to use single filaments as was done with Mytilus, so a preparation was devised as follows :

The animal was removed complete from its shell, pinned down, lying on one side, and the uppermost mantle lobe cut away to expose the gills of that side. The outer gill was pinned back and not used in these experiments since it has, in the female particularly, very thick interlamellar septa, while its use as a marsupium in the breeding season would have led to large numbers of embryos and glochidia being included in the preparations. The inner gill alone was used ; it was divided dorso-ventrally with scissors at intervals of 1–1·5 cm; each section in turn was now reflected to expose the dorsal margin of the ascending lamella which, for a considerable distance, is free of attachment to the visceral mass; this margin (rendered free where attached) was held up in broad-ended forceps while the interlamellar tissue was snipped a little at a time with fine scissors. Each section would contain two to four septa and towards the free (ventral) edge of the gill the lamellae could often be pulled apart once the initial cutting was completed dorsally ; they usually parted along the food-groove yielding two similar pieces of tissue. The ventral 1–2 cm. only of each section was used and the piece was placed, frontal surface down, on a slide between two ridges of vaseline (or, better, ‘Hard Grade’ Vacuum Grease) with the ventral margin parallel to the ridges (Fig. 3). During the mounting the tissue was kept moist with the bathing medium. A firm glass thread, about 0-3 mm. diameter, was now placed to span between the grease ridges and parallel with the filaments to hold one side of the preparation in position ; a second thread was similarly placed to hold the opposite side, but before pressing this thread well into the grease the tissue was stretched longitudinally with needles in order to open out the interfilamentar water-pores in the gill. The thread was then fixed in position and counteracted the tendency for the longitudinal gill-muscles to contract and so obliterate the pores. This proved to be the crucial part of the technique; it was further essential to have one or preferably both of the glass threads at least 2 mm. from the edge of the tissue (see below).

The slide was then inverted over a capsule containing a suitable medium. Examination with the microscope gave a view of the outer surface of the gill ; the terminal cilia could be seen at the ventral ends of each (half-) filament bordering the food-groove (itself slightly damaged and not clearly visible); within most of the water-pores were seen the latero-frontal and lateral cilia, the latter with a particularly striking metachronal rhythm. The frontal cilia were not easily detected over the bulk of the preparation as they lay parallel to the line of sight; further, their activity was very probably impeded by their being pressed against the slide, and by accumulations of mucus exuded following this mechanical stimulation. Instead, they were visible where the free 2 mm. of the tissue curled over the glass threads; by focusing it was possible to see, at their several levels, the frontal tracts of three to four filaments in profile; in this position the frontal cilia were not subject to pressure against the slide nor to sheets of entangling mucus.

For a bathing medium 4% sea water was used; tap water, although the medium applied without adverse effect to the whole animal in the aquarium, failed to maintain normal ciliary activity for the same length of time ; this was understood to imply that the medium bathing, the isolated gill tissue enters the vessels normally containing blood, so that it is necessary to bathe with a solution imitating the blood in ionic strength. Shock reactions were not investigated for the same reason as in Ostrea ; the water spaces are even more extensive in Anodonta gill.

In 4 % sea water or in artificial Anodonta blood, the whole preparation remained healthy in appearance with all tracts of cilia observed beating vigorously for at least 24 hr. ; in two instances up to 40 and 49 hr. respectively.

In 2 % sea water the laterals sometimes, and the other three types always, maintained normal activity; in 1% latero-frontals did not always beat as vigorously as in 4%. Similarly, in the case of artificial blood comparably diluted, there was the same falling-off of activity in laterals and latero-frontals. Frontals and terminals sometimes beat normally in (London) tap water but generally all activity was quickly reduced and laterals inhibited.

A few frontal and terminal cilia maintained a weak beat up to 2 hr. in distilled water and in M/4oo-NaHCO₃, but in the majority of cases all cilia ceased beating after 1–2 min. in these fluids. Although the distilled water was rather acid it is not thought this was the main contributory factor to inhibit the beat since M/400-NaHCO₃ failed to support even weak activity for more than a few minutes.

All cilia were apparently unaffected by 8–12% sea water, terminals and often others beat actively in 16–20%, while frontals sometimes beat feebly in 32% and terminals in 40%. Terminal cilia in hypertonic media tended to stand out straight from the surface and beat very vigorously, rather than bend in the direction of beat. The reaction to comparable strengths of artificial blood was in each tract equivalent in respect to survival and activity..

Slight swelling occurred in 1 % sea water but no disintegration within 24 hr. ; in tap water swelling was more pronounced and may have interfered with the lateral cilia as the water-pores narrowed. Disintegration was most striking, following extensive swelling, in distilled water, the tissue falling to pieces if shaken, in an hour or less. Raising the pH by using M/400-NaHCO₃ failed to prevent this deterioration or even to retard it noticeably. In hypertonic media there was disintegration without swelling, becoming evident in 16% sea water and more extensive as concentration increased. It is interesting to note that disintegration was present in hypertonic artificial blood to about the same extent as in the corresponding sea water concentrations ; swelling in diluted media was, however, a little less marked in the former.

The frontal cilia (all three species), short abfrontals (Mytilus) and terminals (Anodonta) are excellent for experimentation and withstand considerable degrees of concentration and dilution of the media. The latero-frontals and especially laterals, of all three, however, do not survive very long in the isolated gill even in their normal media; it is suggested that unequivocal results can be obtained from these tracts only with extreme caution. Some disadvantages of these two tracts are pointed out in earlier sections; in Ostrea and Anodonta it was found also that a length of lateral ciliated tract beating normally, with its precise metachronal rhythm, might suddenly cease, the cilia remaining stationary at the end of the effective beat; this was seen on numerous occasions in 100 and 4% sea water respectively and always all cilia appeared to cease together, but resumption some seconds later progressed slowly along the tract. There was no obvious reason for the cessation, and a sharp blow on the bench did not produce it. Lucas (1931) found laterals and latero-frontals in several species to behave in this way both in vivo and in vitro; he stated there was no nerve supply down the gill filaments so the central nervous system cannot be responsible for the effect. Some degree of intercellular co-ordination, however, must be present in tracts with such highly organized rhythm, and it is presumably upon this mechanism that there acts the stimulus for cessation.

The differing sensitivities of various cilia types found by Wells et al. (1940) on dilution in Mytilus are shown to hold for hypertonic sea water in Mytilus, and for both hypo-and hypertonic media in Ostrea and Anodonta.

In the previous paper (Pilgrim, 1953), it was pointed out that a correlation appeared to exist between blood concentration and the whole metabolism of these animals, particularly their heart rate. A curious exception to this correlation is in the rate of beat of the cilia; in all three species the gill cilia of corresponding types beat at approximately the same rate in their proper media. It is instructive to compare the ecological ranges and the sea water concentrations in which heart preparations beat well (see Table 2, Pilgrim, 1953), with corresponding values for frontal cilia (the most reliable ones occurring in the three species used). The ranges, in percentage sea water, are, respectively, Mytilus 45–100, 40–160, 40–150; Ostrea 70–100, 70–120, 40–125; Anodonta fresh water–c. 14, 2–c. 24, 1–12. The frontal cilia of Mytilus and of Ostrea are, like their hearts, capable of normal activity over salinity ranges greater than those to which the whole animals are subjected in natural conditions. Further, the heart and cilia tolerances within each of these two species are similar; this is to be expected in poikilosmotic animals. Anodonta heart will not beat when isolated in fresh (tap) water, and this is consistent with the homoiosmotic nature of the animal; the frontal cilia, however, show a reduced activity when isolated in tap water. In the case of the three heart muscles it is reasonable to postulate that rate of beat is proportional to the concentration of one or more ions in the immediate environment, i.e. blood and pericardial fluid, which are almost identical, and together vary with changes in outer environment. Much the same ionic media bathe both sides of the ciliated cells on the gills of the marine species over a wide range of salinities, but in Anodonta these cells are, in vivo, constantly subjected to pond water on one side and, on the other, to the blood which is 10-20 times as concentrated. Despite their very dilute bathing media the cells can evidently maintain a sufficiently high concentration of the necessary ion(s) to support vigorous ciliary activity. Probably they are much more dependent on blood concentrations and are ionically somewhat insulated from the fresh water; that they are not completely unaffected by external media is shown by their inability to function in distilled water, or in sea water above 32%, in which conditions the cells must be unable to maintain their internal ionic constitution in the face of such extreme concentration gradients. The activity of the cilia over wide ranges of salinity lends further support to Needham’s suggestion (1930) that ecological distribution is related to factors in the life history of the species rather than to the behaviour of the adult.

Disintegration is shown to occur in Mytilus (below 35% and above 150% sea water), Ostrea (below 30% and above 150%) and Anodonta (in tap water and above 12% sea water); in each species swelling accompanies or precedes the effect in hypotonic but not in hypertonic media. A different explanation is suggested in the two situations:

(a) Gray (1931) showed that the stability of the intercellular matrix of Mytilus gill was dependent very largely on the magnesium ion but also on the other major cations ; on dilution, then, there are probably insufficient of the requisite ion(s) to stabilize the matrix which softens and frees the cells ; they free themselves all the more easily if still beating. Conditions for stability are markedly different at different pH (Gray), but this was held constant here. In Anodonta blood and natural waters magnesium is scarce and possibly calcium is the more critical ion, yet disintegration was approximately equal in 1 % sea water and in $14$Anodonta ‘blood’; calcium is about 18 × as concentrated in the latter.

(b) In hypertonic media disintegration was usually accompanied by a granular appearance in the tissue; the latter is attributed to dehydration of the cells with consequent denaturation of the protoplasmic proteins. Disintegration itself, it is here suggested, is due to the cells shrinking and loosening from the matrix and from each other; if still beating they assist in freeing themselves.

The added loosening effect of the still-active cells gave rise to rather anomalous results on occasion; disintegration was more apparent in 150% sea water (Mytilus) than in 200%, presumably the greater ciliary activity freeing the larger number of cells.

1. Isolated preparations are described from the gill epithelium of three species of Lamellibranchs ; sea water was found to be a suitable medium for Mytilus and Ostrea and 4% sea water for Anodonta.

2. Sudden changes in concentration of the media led to ‘Shock Reactions’ in several types of cilia. The extent of the shock reaction following dilution was shown to be related to the ratio of the concentrations of the media before and after the change, rather than to either of these values separately.

3. In all three species the range of salinity compatible with mechanical activity varied with the type of cilium, being widest in the frontals, narrowest in the laterals, and intermediate in the other tracts.

4. Disintegration of the epithelia is found in all three species in both hyper-and hypotonic media, with accompanying swelling in the latter.

The work described in this and the previous paper was done in the Department of Zoology, University College, London.

I wish to record my gratitude to Prof. D. M. S. Watson, F.R.S., for providing me with laboratory facilities and for his interest in my work. I am indebted to Dr G. P. Wells for his encouragement and supervision at all stages of the investigation, as well as for his suggestions and criticisms. My sincere thanks are also due to the New Zealand Government for a National Research Scholarship, without which I could not have undertaken the work.