The Physiology of Contractile Vacuoles: I. Osmotic Relations

Considerable discussion has centred around the contractile vacuoles of Protozoa, and both structure and functions are discussed in Lloyd’s review of the subject (1928). Of the various suggestions which have been made as to what is their function, the only two probable ones are (i) that they are organs for the excretion of waste matter (other than water), and (ii) that they control the internal osmotic pressure. These two possibilities are not mutually inconsistent. The contractile vacuole may have more than one function in the same organism, or different functions in different species, but it is very unlikely that the mechanism is not fundamentally the same in all species. There is as yet no convincing positive evidence for the theory that they are excretory, and this is not surprising in view of their small size; but this possibility need not affect an examination of the second alternative, which is the purpose of this paper. The osmotic control theory was suggested by the supposed occurrence of contractile vacuoles in fresh-water Protozoa, but not in marine or parasitic forms. Actually they do occur in a number of marine Ciliates, but this does not invalidate the argument. A marine form might have an internal osmotic pressure either greater than or the same as that of the external medium, and so might or might not have a contractile vacuole. A fresh-water form, however, must have its internal osmotic pressure greater than that of the surrounding medium, and so would have to have a contractile vacuole, unless it could maintain its internal concentration in some other way.

If the contractile vacuole is a controller of the internal osmotic pressure one would expect it to be affected by large changes in the osmotic pressure of the medium. Various workers have reported a loss of the contractile vacuole from fresh-water forms which were placed in sea water, and a reappearance of it when the organisms were replaced in fresh water (e.g. Zülzer, 1910), but in such cases it can always be objected that the stoppage of the vacuole was due to a toxic effect of the unnatural environment. Hogue (1923) found that contractile vacuoles appeared in a marine amoeba which was placed in fresh water—a piece of evidence the importance of which has not been sufficiently stressed. A more detailed analysis of the relations between the rate of vacuolar output and the osmotic pressure of the external medium requires the measurement of the durations and of the ultimate diameters of the vacuoles. (By “duration” of a vacuole is meant the time interval between the systole of that vacuole and of the one before it ; by “ultimate diameter “the diameter just before systole, which is the maximum diameter.) In some of the earlier work on contractile vacuoles the diameters were unfortunately not measured. Adolf (1926), however, measured both diameters and durations for Amoeba proteus, and found that there was no significant decrease in the rate of output when the amoebae were placed in solutions of NaCl and of other salts of concentrations up to M/20; and that there was no direct relation between the surface area of the organism and the rate of vacuolar output. He puts this forward as evidence against the osmotic control theory. When the amoeba is in equilibrium with its environment, the rate of output of water from the contractile vacuole must equal the rate of intake of water through the body surface. If the external osmotic pressure were raised above the internal osmotic pressure—as may have been the case in Adolfs experiments—then the contractile vacuole might be expected to stop. There are four possible explanations of Adolfs results :

• (1) that the surface of the amoeba is freely permeable to the solutes used;

• (2) that the internal osmotic pressure of the amoeba is high, and that the increase in external osmotic pressure was too small to have any effect ;

• (3) that the internal osmotic pressure was raised by shrinkage of the body, and that the contractile vacuole was maintaining it above that of the new external medium ;

• (4) that the contractile vacuole was drawing on the internal water supply of the amoeba without there being any entry of water to replace the water evacuated.

(1) and (2) are both unlikely. As regards (4), Chalkey (1929) has shown that, in solutions of non-electrolytes of approximately the same osmotic pressures as were used by Adolf, amoebae go on shrinking for about 2 hours. There is no proof, in the absence of volume measurements, that Adolf s amoebae had reached equilibrium with the new outside medium in respect of body volume, and that water was passing into them from the outside.

In the experiments about to be described the body volume was measured in order to meet the above objections ; and since more significance could be attached to an increase (if such were to occur) than to a decrease in rate of output, hypotonic as well as hypertonic media were used.

The species used were chosen from among the Peritrich Ciliates, which have the following advantages:

• (i)

  The contractile vacuole contracts very frequently (once in 30−60 sec. at 15° C.) in fresh-water species, and is present and contracts fairly frequently (once in 1−20 min. at 15° C.) in marine species.

• (ii)

  The contractile vacuole is always in the same place.

• (iii)

  The organisms are sessile—a fact which enables them to be used in conjunction with the apparatus about to be described.

• (iv)

  The organisms are nearly perfectly symmetrical about an axis of rotation ; this enables estimations of their volume and surface area to be made.

• (v)

  The ultimate vacuolar diameter and the rate of output of these forms remain constant, within the limits of experimental error, under constant conditions.

The species used were :

Fresh-water forms (see Kent, 1880−1):

Rhabdostyla brevipes¹ (C. and L.). This species was sessile on the larvae² of the mosquito Aedes geniculatus, and was obtained from a rot-hole in a beech tree. Individuals of Rhabdostyla brevipes lived healthily in the laboratory, in water from rot-holes, until the skins of the larvae on which they were growing were cast off. They are, however, difficult material to use, as the mosquito larvae are liable to wriggle. The adjustment of the cover-slip is critical, and long experiments are frequently ended prematurely by the movement of the larva.

Marine forms (see Hamburger and Von Buddenbrock, 1911):

Vorticella marina Greef, Zoothamnium niveum Ehrbg., Zoothamnium marinum Mereschk., Cothurnia innata O.F.M., Cothumia³ sp. ? socialis Gruber, Cothumia curvula Entz., Cothumia ingenita⁴ O.F.M.

All the marine forms were found on Cladophora in the Drake’s Island tank of the Plymouth Aquarium. The Cladophora came originally from intertidal rockpools in Plymouth Sound. Most of the work on marine forms was done in Plymouth, but a few experiments were done in London in winter time. It was found that material would arrive from Plymouth in a healthy state so long as the weather was cool, and would remain healthy in the laboratory in London for a week if it was kept below 15° C. No material was used which did not appear to be in good condition.

The Protozoa were kept in a continuous flow of fluid while under observation by means of the apparatus shown in Fig. 1. Fluid flows from the capillary opening of the supply unit on to the slide at the edge of the cover-slip, at a rate of a drop in 1−2 sec. It flows under the cover-slip, out through the blotting paper at the other side, into the funnel, and away. In all experiments in which room temperature could not be maintained as low as 15−16° C., the temperature of the fluid supplied was controlled by means of the water-jacket. When the room temperature was not excessively high, the temperature of the fluid under the cover-slip (as determined by a thermocouple) was found to be nearly the same as that of the water-jacket. Accordingly in the majority of experiments a thermocouple junction was not used, since it adds considerably to the difficulties of manipulation, and the water-jacket was adjusted so as to give a temperature under the cover-slip of 15−16° C. A number of supply units of this apparatus were fitted up. By changing the supply unit a rapid change of the external medium of the organism can be brought about. Such a change of units can be effected in a few seconds. The constant flow of fluid ensures that there can be no lack of oxygen nor accumulation of carbon dioxide. The fluids used for fresh-water organisms were mixtures of seawater and London tap water at pH 7.9−8.0, and for the marine forms mixtures of sea water and London or Plymouth tap water at pH 7.9−8.2. The pH was taken with indicators, and was corrected where necessary with NaOH.

The Protozoa were in no way compressed during the experiments. The material on which they were growing supported the cover-slip, and particles or free-swimming Ciliates often passed in between the cover-slip or slide and the organism under observation. The contractile vacuoles of all species were, just before systole, perfectly circular in outline from whatever direction they were viewed. They are therefore held to have been spherical at that time, although they were frequently of irregular shape during the earlier part of diastole. They emptied themselves completely to the exterior, and there were no signs of any “canals” such as occur in Paramecium. The time of systole was estimated generally within a second. The ultimate diameter was measured with a Watson screw micrometer eyepiece. For this purpose the lines of the micrometer were kept continually set on the vacuole, and the scale was read when systole occurred. The scale was then returned to zero or thereabouts until the vacuole next appeared. A $16$ in. objective was used. The ultimate volume of the contractile vacuole was calculated as and the average rate of output for the duration of each vacuole was obtained by dividing the ultimate volume of that vacuole by its duration. The mean rate of output for a longer period of time was obtained by dividing the total output by that time; this eliminates practically all error in the measurement of time, but the error involved in measuring the vacuolar diameter remains. The rate of output is extraordinarily constant under constant conditions, although there is occasionally a sharp deviation for a single cycle of the vacuole. The standard error of the mean rate of output was estimated as ⁠, where d = difference between mean rate of output and rate of output for any one vacuolar duration, and n = number of these rates of output observed. It was generally for fresh-water species 5 −10 per cent., and for marine species 5−15 per cent, of the mean rate of output. The number of readings was in some cases too small for the standard error of the mean to have much statistical value. For measurements of the body volume only two, three, or four readings could be obtained under a given set of conditions. The maximum deviation from the mean was generally between 5 and 10 per cent, of the total body volume.

Each organism was kept in a constant flow of fluid (fresh water for fresh-water forms, sea water for marine forms) under control experimental conditions for an hour before observations began. Measurements for estimation of the body volume were taken at intervals throughout the experiment. While these measurements were being taken, or sometimes for purposes of rest for the observer, there were intervals during which observations of the contractile vacuole ceased. In the tables of results (p. 369) the actual time is given during which the organisms were kept under any given set of conditions, inclusive of such intervals, but exclusive of the hour for acclimatisation at the beginning of the experiments. The organisms used (with the exception of Vorticella marina) are practically perfectly symmetrical about an axis of rotation. An accurately reproduced side view therefore contains all the data necessary for an estimation of the body volume. Each of these estimations was obtained as follows :

1. During the experiment a small drawing of the side view of the organism was made, and measurements of a number of its salient dimensions (e.g. total length, breadth at various levels, depth of rim) were taken.

2. From the above, after the experiment was over, a large-scale drawing was constructed, with a linear magnification of about × 3000.

3. The figure so obtained is divided into two halves—mirror images of each other—by the central axis. The distance (ȳ) of the centre of gravity of one of these halves from the central axis was found by means of a geometrical construction and calculations (see Henrici and Turner, 1903). The measurement of areas which this method entails was done with a planimeter.

4. The body volume (v) was calculated from

   

   where A = area represented by one-half of the scale drawing described in (2). The ciliary disc, in respect of which the Protozoa used are asymmetrical, and the contractile fibre present in the stalk of some species, were considered sufficiently small to be neglected entirely. The transparent sheath which surrounds the contractile fibre and which forms the bulk of the stalk is a secretion and is dead material. No allowance was made for the gullet. It is important that individuals chosen for observation should lie accurately in the same optical plane, since otherwise the length measurements are inaccurate.

The contractile vacuole of Rhabdostyla brevipes was found to have a duration of 30 − 60 sec., and an ultimate diameter of 7 − 11 microns, at 15 ° C. The average rate of output under these conditions was n-6 cubic microns per second. Individuals were subjected successively to (1) tap water, (2) a known mixture of tap water and sea water, (3) tap water. Fig. 2 illustrates a typical experiment. Transference of the organisms to (2) led to an immediate decrease in the rate of output and in the ultimate diameter until these reached a new steady value. Also in many cases a decrease in body volume occurred which was very noticeable, although the body volume could not be measured accurately for this species owing to the fact that an individual does not remain for long in the same optical plane. In Fig. 5 is shown the relation of concentration of sea water in the medium with mean rate of output. In calculating the latter no readings were included which were taken immediately after a change of medium and before the rate of output had settled down to a steady value. In spite of individual variations it is clear that there was a falling off of rate of output with increasing concentration of sea water, until in about 12 per cent, sea water the rate of output was zero. The ultimate diameter was also decreased. Both mean rate of output and mean ultimate diameter returned in most cases to their original values when the organism was replaced in tap water.

In the higher concentrations (10 − 15 Per cent, sea water) the cilia stopped and the organisms were contorted by shrinkage, but in all cases after the organisms had been returned to fresh water they appeared perfectly healthy and normal.

The contractile vacuole of marine forms generally had a duration of about 10 − 15 min. (though for Cothurnia curvula it was about 3 − 5 min.), and an ultimate diameter of 10 − 20 microns. The rates of output ranged from about 0.5 (for Cothurnia curvula) to 10 cubic microns per second (for Cothurnia ingenita and Zoothamnium niveum).

In the series of experiments described below (see also Tables I and II, pp. 369, 370), various marine Peritricha were subjected successively to (1) 100 per cent, sea water, (2) hypotonic sea water of known dilution, (3) 100 per cent, sea water. In general it may be stated that dilution of the sea water led to an increase in body volume and in rate of output. A typical experiment is illustrated graphically in Fig. 3. This series of experiments may be summarised briefly as follows:

• (1) In 100 per cent, sea water the body volume, rate of output, and ultimate diameter remained constant.

• (2) In hypotonic sea water the body volume increased rapidly and immediately, and in most cases remained constant at a new high level. In a few experiments there was a falling off in the body volume after the initial increase (Fig. 4). In many experiments in which very dilute sea water (approximately 5 per cent, or less) was used, the organisms swelled up until they were globular; and then clear drops of fluid raised the pellicle up in blisters, which often swelled and became nearly spherical. In a few cases the protoplasm flowed out into the blisters. The rate of output increased rapidly and immediately, and then either remained constant at a new high level, or decreased at first but subsequently became constant at a level which was still considerably higher than the original (100 per cent, sea water) level. In one case the organism (Cothurnia sp. ? socialist was still maintaining a steady rate of output 1600 min. after it had been transferred to 25 per cent, sea water. In the case of some individuals which had been transferred to 75 per cent, sea water (or stronger) the rate of output rose temporarily and then fell approximately to its original level. When blisters had been formed, in very dilute sea water, the contractile vacuole generally slowed down and stopped. The ultimate diameter was liable to sharp variations after a change of medium, but subsequently became constant. In moderately hypotonic (50-100 per cent.) sea water it was generally less, and in more dilute sea water generally more, than what it had been originally (in too per cent, sea water). In very dilute sea water the contractile vacuole sometimes failed to empty itself completely.

  Fig. 4.

  

  View largeDownload slide

  The effect of dilute sea water on the body volume and rate of output of Cothurnia curvula; a case in which there was a falling off in body volume while the organism was still in the hypotonic medium. N.B. For $1212$ per cent, sea water the average rates of output for groups of three vacuolar durations have been plotted.

  Fig. 4.

  

  View largeDownload slide

  The effect of dilute sea water on the body volume and rate of output of Cothurnia curvula; a case in which there was a falling off in body volume while the organism was still in the hypotonic medium. N.B. For $1212$ per cent, sea water the average rates of output for groups of three vacuolar durations have been plotted.

• (3) In 100 per cent, sea water the body volume returned immediately and rapidly to its original value, or less. In those cases in which the hypotonic sea water was very’ dilute the pellicle wrinkled when the organism was replaced in 100 per cent, sea water, and the wrinkles remained for some time (e.g. 15 − 30 min.). It is therefore probable that volume measurements obtained under these conditions are too high. The rate of output returned approximately to its original value immediately and rapidly. It was generally more variable than formerly. The ultimate diameter became variable, but returned approximately to its original value. The relation of concentration of sea water with body volume and rate of output are shown in Figs. 6 and 7 respectively. Readings taken over a period of 5 − 10 min. after the change of medium were discarded, so that no readings were included which were taken before a steady level was reached. The maximum increase in rate of output was × 70, in $1212$ per cent, sea water. In greater dilutions the rate of output fell off.

In another series of experiments on marine Peritricha the organisms (Cothurnia curvula and Cothurnia ingenita) were subjected by successive steps to more and more dilute mixtures of sea water and Plymouth tap water. The relation between rate of output and concentration of sea water (Fig. 7, curves B and C) is similar to that found in the experiments described above. In one experiment (Fig. 8) the organism was taken in steps down to 1 per cent, sea water, and then back to 100 per cent, sea water by the same steps in the reverse order. The body volume and rate of output were much lower on the return journey than they had been on the same steps on the outward journey. In two experiments the organism was subjected to Wembury stream water (pH not corrected) and then to Plymouth tap water (pH not corrected). In Wembury stream water a fairly steady rate of output was maintained for a long period (40 hours in one experiment, 15 hours in the other), but in Plymouth tap water the pellicle was raised up in blisters and the contractile vacuole stopped.

In both these series of experiments on marine Peritricha the cilia stopped beating when the organism was placed in the more dilute mixtures (e.g. sea water 25 per cent, or less), although there was considerable individual variation in this respect. Also sometimes they started beating again in the diluted sea water, but at other times they remained stopped until some time after the organism had been replaced in 100 per cent, sea water. Except in some cases in which blisters had been formed, the organisms appeared to be perfectly healthy at the end of the experiments, and the cilia beat again normally. On several occasions individuals divided soon after experiments had been performed on them.

In all these experiments, both on fresh-water and on marine forms, the total number of systoles observed was over 4000.

The increase in the body volume of marine Peritricha, consequent on treatment with hypotonic sea water, may be explained in two ways :

1. Osmotic swelling due to a cell membrane which is relatively impermeable to salts, and yet freely permeable to water, or

2. Ionisation of the cell proteins due to the reduction in salt concentration.

If (1) is true, the cell membrane must be relatively impermeable to salts. If (2) is true, the cell membrane need only be impermeable to proteins, and may be freely permeable to salts. Evidence is as yet inconclusive as to which is the right explanation ; (2) may possibly play a small part, even if (1) accounts for most of the swelling. There is strong evidence that in many different kinds of animal cells the cell boundary is semi-permeable with regard to salts and water (Lucké and McCutcheon, 1932). Some preliminary experiments with marine Peritricha on the effect of ammonia, an alkaline substance likely to penetrate the cell and there to influence the ionisation of the cell proteins in the same way as would a reduction in the salt concentration, have indicated that there is no change in volume while the organism is alive. Again, when an individual of Cothurnia curvula was treated with a mixture of a glycerol solution and sea water such that the salt concentration was reduced to one-sixth while the osmotic pressure remained unaltered, there was no change in body volume or in rate of output. Thus there is evidence, though as yet incomplete, for believing that the changes in volume observed were due to the fact that the cell surface of these Protozoa is semi-permeable with regard to salts. If this is so, information can be deduced concerning the osmotic pressure of the vacuolar fluid.

The osmotic pressure of the vacuolar fluid has never been measured, but it can be inferred that it is probably near that of pure water, at least for fresh-water species, unless excretory matter is present. For no more salts can leave the organism than enter it, unless its salt content is to be depleted (which could not go on indefinitely) ; and no appreciable amount of salts can enter a fresh-water organism from fresh water except occasionally in the food. On the other hand, the internal osmotic pressure of fresh-water Protozoa, though low, must be greater than that of the surrounding fresh water, so that the contractile vacuole must be separating water from the internal solutes of the organism. And in marine forms, if the cell membrane is relatively impermeable to salts, the same argument can be applied, namely, that since no more salts can leave the organism than enter it, the contractile vacuole must be separating fluid of very low osmotic pressure from an internal solution of osmotic pressure not less than that of sea water. Assuming the semi-permeability of the cell membrane with regard to salts, the contractile vacuole must in both cases be doing work, and it must therefore be regarded as an active mechanism involving the expenditure of energy. Its operation will raise the internal osmotic pressure of the organism until a steady state is reached, which will depend partly on the rate of inflow of water, and hence on the permeability of the cell membrane to water. The magnitude of the difference between the internal and the external osmotic pressures will depend on the rate of vacuolar output and on the permeability of the cell membrane to water, and may be insignificant if the’ latter is great as compared with the former.

The secretory theory of diastole, as outlined above, is entirely contradictory to any suggestion that the contractile vacuole grows larger by osmotic uptake of water from the surrounding protoplasm. For osmotic uptake there would have to be inside the vacuole a quantity of solute such that at the greatest volume of the vacuole the osmotic concentration of the vacuolar fluid was not less than that of the surrounding protoplasm. At the beginning of diastole, when the volume of the vacuole is much less, the concentration of the vacuolar fluid would have to be correspondingly greater. In marine Peritricha, whose internal osmotic pressure must be not less than that of sea water, the initial concentration of the vacuolar fluid would have to be extremely great, and it seems unlikely that such a concentration is actually produced. And for Amoeba proteus, a fresh-water form, Adolf has shown that the relation between vacuolar volume and time is linear during the period of a single diastole. Such a relation is inconsistent with simple osmotic uptake. In view of these objections to the theory of osmotic uptake of water by the vacuole, the validity of the secretory theory is assumed in the discussion which follows.

The factors which are likely to affect the secretory activity of the vacuole are of two types : (a) those dependent on the concentration of the sea water outside at the time in question, and (b) those governed by the state of activity of the organism or of the vacuolar mechanism itself, e.g. general health and condition of organism, possibly food reserves, age, etc. It was observed that in marine Peritricha there was considerable individual variation in the rate of output among members of the same species, and that this could not be attributed to size. Specimens which had been sent from Plymouth to London in cold weather had a high rate of output, while those which had been sent up in hot weather had an extraordinarily low rate of output, and the vacuolar duration was as high as half an hour. Those that had been kept for any length of time in the laboratory in London also had a low rate of output. Hogue (1923) observed that old cultures of amoeba developed a low rate of output, and this has been confirmed by the present writer. It is therefore probable that the rate of output is considerably influenced by the state of the organism. Small differences in condition might account for the differences found by Adolf in the rates of output of amoebae.

The observed increases in the body volume of Peritricha fall short of those which would take place if the cell membrane were perfectly semi-permeable, and if the cell contents were no more than a dilute solution of salts. Such a falling short could be ascribed to salt loss, to the presence of osmotically inactive substances within the cell, or to volume control by the contractile vacuole. It is very unlikely that the pellicle, which is extremely delicate, could exert any significant pressure. Cole (1932) has shown that the inward pressure of the cell membrane of the egg of Arbacia is very small, and such pressure may therefore safely be ignored. To what extent the other factors are operative cannot be discussed until the results of further experiments are available, but it seems probable, from the nature of the curve relating body volume with concentration of external medium (Fig. 6), that in very dilute media salts escape. This would explain the return to a body volume smaller than the original (p. 374), and also the falling off in body volume while the organism was still in the hypotonic medium (Fig. 4). Whether the falling off in rate of output which was often observed while the organism was still in the hypotonic medium (Table II) is to be ascribed to a loss of salts, or to the dying away of a stimulus set up by the change of medium, or to fatigue of the vacuolar mechanism, is uncertain.

From the curves given we find:

The contractile vacuole has therefore all the potentialities required not only for a maintainer but also for a regulator of the internal osmotic pressure. Whether it is effective will depend on the precise adjustment of the mechanism.

There are two possible functions which might be served by osmotic control: (a) Maintenance of the internal osmotic pressure at a level higher than the external osmotic pressure, even though the internal osmotic pressure is influenced by small changes in the external osmotic pressure; (b) Regulation of the internal osmotic pressure so as to keep it constant irrespective of small changes in the external osmotic pressure. For marine forms the curve relating rate of output with concentration of external medium is flat between 100 and 75 per cent, sea water. No significant change in output, such as would be required for “regulation,” occurs between these values. Whether any “maintenance” occurs is unknown, but it is unlikely that the internal osmotic pressure of marine forms is much above that of the surrounding sea water. Although the contractile vacuole of marine species might be regarded as a relic of an ancestral fresh-water habitat, it is possible that in maintenance it performs a useful function. Unless the cell membrane is entirely impermeable to salts, these will enter, although perhaps very slowly; a Donnan equilibrium will thereby be set up owing to the presence of indiffusible proteins inside the cell, so that the internal osmotic pressure will be raised slightly above that of the outside sea water. It is possible that the contractile vacuole might be required to relieve the resulting tension on the pellicle. Against this, as against any other suggestion of a function for contractile vacuoles in marine Protozoa, may be brought the objection that many marine Protozoa successfully do without them.

Adaptation of Peritricha of marine origin to fresh water is probably possible. Zoothamnium spp. and Vorticella marina were completely incapacitated in very dilute sea water, being liable to excessive swelling, but in many cases individuals of Cothurnia spp. were but little inconvenienced, and in some individuals of Cothurnia ingénita and Cothurnia curvula the cilia continued to beat, although rather sporadically, in Wembury stream water. At other times, however, no individuals of Cothurnia spp. would successfully endure even 10 per cent, sea water. All individuals of the same batch behaved alike in this respect. It seems probable that successful adaptation was effected by loss of salt. The rate of vacuolar output falls off in very low concentrations of sea water, and therefore it is unlikely that in such concentrations regulation takes place, although the contractile vacuole may have been preventing excessive swelling by maintenance. The experiments described above support Hartog’s (1899) suggestion, advocated by Lloyd (1928) in his review, that the contractile vacuole prevents the organism from swelling excessively. Lloyd points out that contractile vacuoles occur in fresh-water organisms (including algae as zoospores or as adults) which are devoid of rigid cell walls, but not in those forms or stages in which rigid cell walls are present. It might be questioned whether such cellulose walls could withstand the great pressure which would be developed by a small difference in concentration between the inside and the outside. The very small size of such cells would make it more possible, but a knowledge of the strength of the cell walls is required to settle this problem.

I am grateful to Dr C. F. A. Pantin for suggesting this work to me, and to Dr J. Gray for much helpful advice and criticism. I am also indebted to Prof. H. G. Jackson and Dr E. J. Allen for laboratory facilities, to Dr E. A. Spaul for much encouragement, and to Prof. Sugden and Dr R. G. Cooke for advice on the physical and mathematical aspects of the work. I was granted the use of the London University table while working at the Plymouth laboratory.