The stalk of Carchesium and Vorticella coils by the action of a contractile organelle. The organelle lies within a thread of cytoplasm which is encased in a complex extracellular tube. Study with the light microscope and the electron microscope suggests that the structure of the tube and the course of the organelle determine the form of the coiling. The contractile organelle contains a system of interconnected membranous tubules and the cytoplasm around it also contains membranous saccules. Both tubules and saccules extend along the length of the stalk.
Vorticella and Carchesium have stalks which are capable of coiling helically. The coiling is brought about by the action of a contractile organelle which differs markedly from muscle in its physiological properties. This paper is concerned with the structure of the stalks and seeks to explain some aspects of the coiling in structural terms. Experiments on the activation of the stalks after glycerination are described elsewhere (Amos, 1971, and in preparation).
Fauré-Fremiet (1905) examined the stalks of both genera with the light microscope and found them to be similar. They consisted of a cylindrical sheath containing the contractile thread, or spasmoneme, which pursued a helical course throughout the length of the sheath. Longitudinal fibres lay on the inner surface of the sheath on the opposite side to the spasmoneme. Fauré-Fremiet regarded these fibres, the bâtonnets, as stiffeners which, because of their position, ensured that the sheath would bend into a helix, rather than merely collapse, when the spasmoneme contracted. Randall & Hopkins (1962) confirmed his description with the electron microscope in the case of Vorticella, but presented a different account of Carchesium. In Carchesium, they reported an axial spasmoneme separated equally on all sides from the sheath by an annulus containing many tubular fibres. This structure, which seems ill suited to helical coiling, is visible in electron micrographs of a related peritrich, Zoothamnium, obtained by Fauré-Fremiet, Favard & Carasso (1963). Since the Carchesium studied here agrees with Fauré-Fremiet’s description of the genus, it seems likely that the organisms examined with the electron microscope by Randall and Hopkins were not Carchesium but Zoothamnium. The present account of the Carchesium stalk is therefore believed to be the first extensive one, though the spasmoneme was described by Carasso & Favard (1966).
Many studies have been made of the coiling of Vorticella and it has recently been analysed with a high speed camera by Jones, Jahn & Fonseca (1970), so observations here are chiefly confined to the type of deformation involved, and its relation to the structure of the stalk.
MATERIAL AND METHODS
Carchesium was collected from a pond near Kenn in north Somerset. The species (see Fig. 3) resembled Carchesium polypinum L., as shown by Stein (1854), but the published descriptions of this species are inadequate for identification. All the organisms used came from the same clone. They were cultured in a soil extract prepared by autoclaving 80 g of loam in 1 1. of glass-distilled water, filtering and autoclaving once more. One small drop of fresh milk was added to 25 ml of soil extract in a Petri dish before inoculation. Subcultures were made weekly. Vorticella convallaria L. was obtained from the Culture Collection of Algae and Protozoa in Cambridge. It was cultured in a salt solution nutrified with the Glaxo preparation ‘Complan ‘. The salt solution contained 1·7 mM NaCl, 1·2 mM NaHCO3, 0·054 mM KC1 and 0-054 mM CaCl2. It was sterilized and poured over a sterile solid layer of 3 % agar, made up in the same solution, in a Petri dish. One small drop of an autoclaved 2 % suspension of Complan in water was added and the culture was inoculated by inserting a glass slip from a previous culture, to which the organisms had attached themselves. Subcultures were made monthly.
Light microscopy was carried out with a Zeiss Standard WL microscope equipped for phase contrast and Nomarski’s differential interference contrast. Photographs were taken with electronic flash.
The methods used for electron microscopy were largely conventional. The organisms were fixed between o and 30 C in a solution containing 2·5 % glutaraldehyde and 2·7 mM sodium cacodylate buffer adjusted to pH 7·4 at 2 °C by means of 0·1 N HC1. After exposure to 2 changes of this medium for a total period of 2 h at 0·3 °C the specimens were washed for 2 h in 4 changes of a similar medium, but with 6 % sucrose instead of glutaraldehyde. They were then postfixed in 1 % osmium tetroxide with 6 % sucrose and 50 mM veronal acetate buffer. The buffer was made up to give a pH of 7·4 at 2 °C. The fixed material was washed for 1 h in the same medium but without osmium tetroxide and with only 4% sucrose. In 2 subsequent washes, each 15 min long, the sucrose concentration was reduced to 2 % and then to zero. No experiment was made to test the value of the graded reduction in sucrose concentration, but it is believed that it might have lessened the osmotic stress on the fixed material. Dehydration in ethanol, treatment with propylene oxide and embedding in Araldite were carried out according to Luft’s method (1961). Sections were cut with glass and diamond knives, and grey sections were picked up on grids with carbon-coated Celloidin films. They were stained with a saturated solution of uranyl acetate in 50 % ethanol for 30–50 min and also with the lead citrate solution of Reynolds (1963) for 3–5 min. The electron microscope was a Philips EM 200 operated at an accelerating voltage of 60 kV.
Fauré-Fremiet (1905) has described the formation of the stalk in Vorticella, which begins with the attachment of the body to a submerged surface. A tube of extracellular material is secreted which apparently elongates by the addition of new material at one end. At the same time, an outpushing of the cytoplasm of the body enters the tube. In Carchesium, but not in Vorticella, the body subsequently divides many times and the stalk becomes bifurcated after each division, so that a colony of many individuals or zooids is formed. The spasmoneme of each zooid is restricted to its own branch, and there is no cytoplasmic continuity between zooids.
The size and structure of the stalk vary with its position in the Carchesium colony. In the species examined here, the stalk of a newly attached zooid is 7–10 μm in diameter, but as it grows in length it increases in diameter to 10–18 μm, so that the final form is a slender cone tapered towards the base. The main trunk of the colony loses its ability to coil after thirty or so zooids have been produced. It swells somewhat and becomes filled with a refractive granular material. The chief features of a contractile region are shown in Fig. 1. Inside the sheath runs the cytoplasmic strand in a lefthanded helix, with a pitch of about 140 μm in the main trunk and 60–100 μm in the branches. The spasmoneme, which is embedded in the strand, has the form of a ribbon running helically, with the outer surface flattened against the inside of the sheath. The other surface of the spasmoneme, facing inwards, is covered with a layer of cytoplasm containing many spherical mitochondria. The cytoplasmic strand has a circular cross-section in which the spasmoneme lies eccentrically, the cytoplasmic layer being 4 μm deep on the inner surface of the spasmoneme but so thin on the outer surface that it is barely detectable with the light microscope. The transverse section of the spasmoneme is elliptical when the stalk is extended, with the minor axis lying radially. In the spasmoneme of a branch of a colony the major axis is about 4 μm and the minor 1·5 μm. The minor axis increases when the stalk coils, so that the crosssection of the spasmoneme becomes almost circular when the organelle is fully contracted (Fig. 5). The bâtonnets lie longitudinally on the inner surface of the sheath. They are straight and parallel to one another but staggered in such a way that a helical palisade is formed (see Figs. 1, 6) which runs around the stalk on the opposite side to the spasmoneme.
The scopula, the junction between stalk and zooid, is a highly organized region, in which the cytoplasmic strand becomes continuous with the cytoplasm of the body of the zooid (Fig. 7). The spasmoneme is continuous with longitudinal myonemes in the body. Protruding from the base of the body of the zooid are rod-shaped structures, the scopula organelles, which occur quite generally in this position in peritrichs. Their fine structure is somewhat similar to that of the basal bodies of cilia (Randall & Hopkins, 1962; Fauré-Fremiet, Favard & Carasso, 1962).
The base of the main stalk, which is normally attached to the substrate, is sealed by a transverse disk to which the cytoplasmic strand is attached by about 15 elongated processes which splay out from a swelling at its tip. They will be referred to as rootlets, since they seem from their position to anchor the spasmoneme. Rootlets are also present at the end of the cytoplasmic strand in each side-branch of the colony (Fig. 8). The scopula, the holdfast, and the special modifications of aged stalks have not been examined further.
The structure of the regions of the stalk which are capable of coiling will now be described. It must be noted that all the observations with the electron microscope were made on coiled stalks, since contraction of the spasmoneme invariably occurred before fixation. The following account refers to Carchesium. The small differences in structure in Vorticella are indicated later.
The sheath is an extremely thin membrane which is thrown into folds during coiling. In the electron microscope, sections perpendicular to its surface show 2 dark lines with a clear space between them, the total thickness of the 3 layers being 12 nm. The inner surface is covered with a less-dense layer about 20 nm thick, which appears to be highly ordered. Oblique sections of this layer show regular parallel lines 9 nm apart, which probably represent longitudinal fibrils (Fig. 17). Transverse sections of fibrils are seen rarely and never distinctly, but this may be due to the close proximity of the fibrils to one another and their small diameter.
The space between sheath and cytoplasm is optically structureless, but, as Randall & Hopkins (1962) discovered in Vorticella, it contains a dense mesh work of fibrils of a characteristic type. In Carchesium this fibrillar matrix occupies two-thirds of the volume of the stalk. Each fibril is so thin that the diameter cannot be measured precisely in sections: it is less than 3 nm. Nevertheless, the fibrils in a transverse section of the stalk run straight, some as far as 2 μm (Fig. 11), which implies an enormous rigidity, especially since the matrix as a whole is deformed by coiling before fixation. The predominant orientation of the fibrils is longitudinal, but they often intersect at right angles, forming T-shaped or rectangular patterns. It seems from simple observation, without statistical analysis, that the frequency of intersections at right angles is too high to be accounted for by chance, so that a large proportion of the matrix fibrils must be linked to form a rectangular lattice. As Randall and Hopkins discovered, the fibrils have electron-dense beads spaced regularly along their length. In Carchesium the beads are about 8 nm in diameter. The mean of 10 measurements of the distance between the centres of adjacent beads, involving a total of 54 spacings, was 31·9 nm (S.D. 3·0).
These fibrils are not easily seen in life but after the disintegration of the cytoplasmic strand they become plain under phase contrast (Fig. 6), presumably because the contrast between them and their surroundings is enhanced by loss of refractive material from the matrix. Each one is about 15 μm long and 0·2 μm in diameter. Because of their helical arrangement, a transverse section of the stalk contains a crescentic array of transverse sections of bâtonnets (Fig. 11). The sections grow smaller at one side of the crescent but not at the other, which indicates that the bâtonnets taper to a point at one end, as shown in Fig. 1.
The electron microscope resolves each bâtonnet into a loose bundle of about thirty subfibres. The number is not clear because the subfibres are matted together and appear to be partly fused. They have various diameters from 15 to less than 3 nm, and there is some indication that the subfibres are in turn compounds of even finer fibres. Glancing longitudinal sections of the stalk (Fig. 10) show that most of the subfibres run longitudinally but some run at various angles across the 2μm gap between bâtonnets, forming an irregular interlaced network. The apparent discontinuity of the subfibres may simply be due to the fact that they run out of the plane of section. The bundles have transverse, regularly spaced bands of higher density about 4 nm thick. It has not been possible to decide whether the banding is due to dense regions within the subfibres or to bridges between them, but since the bands sometimes cross the whole width of the bâtonnet the dense regions of adjacent subfibres may perhaps be held in register by some form of transverse linkage. The mean of 10 measurements of the spacing between bands, including a total of 52 spaces, was 27·1 ±2·0 nm. The measurement suggests that the subfibres are compounds of the beaded fibres which occur in the matrix, the dark bands corresponding to rows of beads.
There are no traces of endoplasmic reticulum or Golgi cisternae in the cytoplasm of the stalk. The plasma membrane presents the usual trilaminar appearance with a total thickness of about 7·5 nm. In coiling it is thrown into extensive folds, as is the sheath. There is a large space between the fibrillar matrix and plasma membrane in material prepared for the electron microscope. This space is an artifact arising from shrinkage of the cytoplasm by about 30% during preparation. In life the region where the space would be can be seen to contain cytoplasm with mitochondria (compare Fig. 5 with Fig. 9).
The plasma membrane has the form shown in Fig. 9 in transverse sections of the stalk. Two distinct regions of the membrane, on either side of the strand, are denser than elsewhere. This appearance is due to a pair of flattened saccules lying on the inner surface of the plasma membrane. One of these saccules is shown at higher magnification in Fig. 18. Their constant position in the transverse section suggests that the saccules have the form of a pair of ribbons running throughout the length of the cytoplasmic strand and twisting around it in parallel with the spasmoneme. Some transverse sections appear to show more than two saccules, but this may perhaps be accounted for by a branching of the ribbons. The saccules are approximately 1·1 μm wide. No structure has been seen inside them and their contents are not stained by uranyl acetate or lead citrate. In some sections the separation of the saccular membranes is fairly constant, varying only from 8 to 12 nm for considerable distances along the cross-section and increasing in some regions to about 100 nm. In other sections the membranes are 0·5 μm apart. In view of the extensive shrinkage of the cytoplasm, mentioned above, it is not at present possible to decide which appearance corresponds to the state in life, if either of them does. But since the mitochondria are often equally well preserved in both types of section the variation in width may perhaps be genuine and not simply due to variation in preservation. The saccular membrane has the same appearance and thickness as the plasma membrane, from which it is separated by a gap of about 5 nm which does not vary appreciably. This gap is presumably a zone of adhesion between the saccular and plasma membranes, since they do not come apart from one another even when the saccules are highly dilated.
Isolated sinuous microtubules with an external diameter of about 15 μm have been observed, distributed apparently at random in the cytoplasm. Also, structures superficially like the basal bodies of cilia are common. They consist of parallel tubules arranged in a circle to form a cylinder 430 nm high and 124 nm in diameter. From the frequency with which they are sectioned it can be deduced that there are about 200 in a segment of stalk 50 μm long. The cylinders are often, though not always, oriented perpendicularly to the plasma membrane and the tubules. There are 9 single tubules in the bodies, with a tenth placed eccentrically within the cylinder, each tubule having an external diameter of approximately 22 nm (Fig. 19).
Only one section has been obtained of the rootlets which emerge from the end of the cytoplasmic strand in each branch of the colony. It appears from this section (Fig. 22) that there is a cylinder of 9 single tubules which are only 13 nm in diameter at the base of each rootlet and that there are densely staining bodies inside the cylinder of tubules. The rootlets themselves are cylindrical structures 0·16 μm in diameter which protrude from the cytoplasmic strand into the fibrillar matrix. They contain 9 tubules which have a similar diameter to the basal ones and which end at different levels. Both the basal and the distal tubules, which are probably continuous with each other, are embedded in a densely staining matrix which has the appearance of a continuation of the spasmoneme. A membrane 10 nm thick with a dense outer layer surrounds the tubules and this membrane is in turn clad in a diffuse amorphous coat, spaced 20 nm from the surface of the membrane but apparently connected to it by many fine filaments.
The total mitochondrial area in 5 transverse sections of the stalk was 5·2% of the cytoplasmic area. This may be taken to mean that about 5% of the cytoplasmic volume is mitochondrial, provided that differential shrinkage of mitochondria and cytoplasm does not occur during preparation. The mitochondria are spherical. In the terminal branches of actively growing colonies their diameter is about 1·5 μm, but in the main stalk they are smaller and more numerous. As Fauré-Fremiet et al. (1962) have observed in the mitochondria of the peritrich Epistylis, the outer membrane is stained less with uranyl acetate and lead citrate than the inner, and the stroma stains much more intensely than the contents of the sinuous villi.
The spasmoneme contains a mass of filaments with tubules running through it (see Figs. 9, 12–15 and 23). It is not separated from the rest of the cytoplasm by a membrane. No substructure has been seen in the filaments, which are only about 2 nm in diameter. They are roughly parallel to one another, about 3 nm apart and their crosssections are locally arranged in straight lines in some transverse sections of the spasmoneme (Fig. 12). The orientation of the filaments in relation to the spasmoneme is roughly longitudinal.
The tubules within the spasmoneme have been described by Favard & Carasso (1965). There are about 150 tubules in a transverse section of the spasmoneme from a branch and 400 in that of a main stalk. The diameter varies from one tubule to another in the range 38–70 nm, and their distance apart is about 250 nm. Branching of the tubules has not often been observed but it may be common, since the transverse sections often occur in groups of two or three and sometimes have two or three lobes. It seems likely that all the tubules are joined together to form a continuous network. Each tubule is limited by a membrane 1·7–5·2 nm thick in which no structure has yet been seen with ordinary fixed material, but a trilaminar appearance was obtained with glycerinated preparations. Wisps of an amorphous substance are often visible inside the tubules.
Although the stalk of Vorticella is much smaller, its structure is similar to that of Carchesium. It is between 100 and 300 μm long in the species examined, varying greatly from one specimen to another. The total diameter is 2·9 μm and that of the spasmoneme 1·2μm. Some features of the fine structure are shown in the electron micrograph of Fig. 16. The sheath, matrix and bâtonnets can be identified, as with Carchesium. Within the spasmoneme there is a similar arrangement of tubules and filaments. The tubules have roughly the same diameter (60 nm) as in Carchesium and the same separation, and since the spasmoneme is smaller there are only about 60 in the transverse section.
The course of the tubules is interesting in view of a possible analogy with the membranous systems of striated muscle. Favard and Carasso have concluded from their studies of Carchesium and Vorticella that all the tubules at a given level run parallel to one another at a small angle to the longitudinal axis of the stalk, each ending blindly at both ends near opposite surfaces of the spasmoneme. The evidence for this view is that longitudinal sections of the spasmoneme, such as that shown in Fig. 15, contain tubules running obliquely. Observations on Vorticella suggest that this may not be the correct interpretation. When a series of sections at different depths is taken, the angle between the tubules and the longitudinal axis appears to reverse in sign as one proceeds from the near to the far surface of the spasmoneme, suggesting instead a helical organization (see Figs. 13–15). Measurements on sections of this type reveal that tubules near to the surface of the spasmoneme are at about 30° to the longitudinal axis of the organelle. If they are continuous along the length of the spasmoneme, the tubules must pass around it in a helix. Since Favard and Carasso observed continuity for as much as 8 μm, it is extremely probable that they do. There is no evidence that the tubules ever rise to the surface or communicate with the plasma membrane.
Sections of the junction between the spasmoneme and the myonemes of the body show that the myonemes are structurally similar to the spasmoneme and continuous with it. Also, the tubules appear to pass from the spasmoneme into the myonemes of the body through the junction (see Fig. 23). The fate of the tubules within the body has not yet been discovered.
The coiling of the stalks
From observations made with phase contrast Sugi (1961) calculated a reduction in volume of the spasmoneme of Carchesium of between 24 and 38% during contraction, but even with the clearer image formed by the Nomarski interference contrast microscope the error in the measurement of volume is larger than 40 %. No consistent change in volume was observed in the present work.
The overall effect of the coiling is that the distance from the body to the base of the stalk is reduced by approximately 90%. Also, the body is rotated. To measure the rotation small colonies of Carchesium were viewed from above while contracting in response to vibrations. In colonies with two or three individuals it was observed that the zooids revolved around one another during the uncoiling of their common stalk. The rotation was anticlockwise to an observer looking down the stalk from the zooids towards the point of attachment. When measurements, each judged to the nearest 90°., were made on 20 colonies with an average number of about 7 turns, the average rotation per turn of the coiled stalk was found to be 51°, but the amount varied between colonies from 22 to 90°.
An abnormal type of coiling was often observed in Carchesium stalks from which the zooid had been detached by compression of the living organisms between slide and coverslip. The break invariably occurred at the scopula, and the stalk immediately began to coil near this region. A zone of helical bending about 150 μm long passed down the stalk, while at the rear of the zone the spasmoneme broke into fragments and the stalk became straight. The spasmoneme produces local bending under these conditions.
The stalk in these peritrichs is clearly adapted for pulling the body towards the point of attachment. The significance of the movement is unknown. It may serve as a defence against predation or mechanical injury. Ideally the coiling should not involve rotation, particularly in Carchesium, where the branches of the colony are in danger of twisting around one another, but a small amount of rotation does occur.
It is instructive to consider what pattern of strain must be produced on the surface of the stalk if there is to be no rotation, and to compare this pattern with the structure of the stalk. The pattern was investigated by means of a model. A square lattice was drawn on the surface of a cylinder of plastic foam, with one axis parallel to the axis of the cylinder and the other circumferential. A helix of brass wire, left-handed like the coiled stalk, was threaded through an axial hole in the cylinder in order to deform it into a helix. The cylinder was provided with disks at each end by which it could be clamped to the wire. It was possible to rotate the disks on the wire and clamp them so that any desired degree of torsion was imparted to the cylinder (see Fig. 21). The strain in the surface was made visible by the distortion of the lattice of squares. To represent the pattern of strain in 2 dimensions the distorted squares were measured and drawn on a diagram (Fig. 2) in which the centres of the bases were drawn equidistant from one another on a series of equally spaced parallel lines. The diagram represents the surface of one complete turn: if extended vertically it would repeat.
An obvious feature of the pattern is a line of maximum compression which runs around the cylinder once for each turn, along the helix of smallest radius. The greatest extension occurs on another helix, passing through the outermost points of the surface. The diagram suggests a reason for the helical organization of the extended stalk, namely that the line of maximum compression must correspond to a helix on the extended surface. This is presumably guaranteed by the attachment of the palisade and perhaps also of the spasmoneme to the rest of the stalk along a helix. To coil without rotation of the body the stalk must bend in a definite proportion to the shortening of the spasmoneme. When the spasmoneme is fully contracted, the stalk must form neither more nor less than one turn for each turn of the extended palisade. The relation between bending and contraction is determined by many factors, including the resistance to bending of the bâtonnets. If they were slightly less stiff, the stalk would presumably bend more for a given degree of shortening of the spasmoneme, and would consequently form more turns than there are in the extended palisade. The slight torsion which occurs in Carchesium may perhaps arise in this way. The coiling results from the interaction of mechanically dissimilar parts. It is probably fortuitous that similar amounts of torsion were observed during the stretching of solid metal springs (see Jones et al. 1970).
Sugi (1960) observed that coiling in living Carchesium in response to local electrical stimulation was sometimes restricted to the parts of the stalk nearest to the electrode. Also, local coiling in stalks severed from the zooid has been described here. Both observations suggest that the spasmoneme is attached to the stalk all along its length, so local contraction can produce local bending. An alternative which seems less likely is that the spasmoneme can bend actively in a restricted region. If the spasmoneme is indeed attached along its length, the matrix surrounding the cytoplasmic strand, or at least its peripheral zone, must be solid enough for attachment to be made. This may be so in spite of the low refractive index of the matrix. Possibly the rectangular arrangement of the matrix fibres produces a high ratio of strength to density.
Unfortunately little is known of the chemical nature of the extracellular material in the stalks of peritrichs. Sulphur-containing proteins appear to be present, PAS-positive substances and nucleic acids absent (Randall & Hopkins, 1962; Levine, 1960; Fauré-Fremiet, 1941). On structural grounds, several proteins must be present in Carchesium and Vorticella.
Contraction of the spasmoneme
During coiling the spasmoneme of Carchesium contracts to approximately one third of its extended length. The degree of contraction in Vorticella is more difficult to measure because of the smaller size of the stalk, but Jones et al. (1970) reported that the spasmoneme shortened by only 10% of its extended length in this genus.
The extended spasmoneme in both genera is positively biréfringent with respect to its length, but when it contracts it becomes optically isotropic (Schmidt, 1940; Amos, 1971). This suggests that contraction involves a disordering or folding of structures which were arranged longitudinally in the extended organelle. The structures in question can be neither the filaments nor the tubules observed with the electron microscope: their orientation is chiefly longitudinal even in maximum contraction. The birefringence must arise either from a substance which, like resilin (see Elliott, Huxley & Weis-Fogh, 1965) has no structure visible in the electron microscope, or from a structural change within the filaments or tubules.
Membranous structures have been observed near to bundles of filaments in the myonemes of other ciliates. In Spirostomum (Ettienne, 1970) and Stentor (Bannister & Tatchell, 1968) the membrane frequently appears to consist of closed vesicles. Favard & Carasso (1965) report that calcium is stored inside the tubules in the spasmoneme of Carchesium and Vorticella, Their chief evidence is the precipitation of crystals inside the tubules after treatment with oxalate. Similar evidence has been adduced for the presence of calcium in the sarcoplasmic reticulum of striated muscle. The spasmoneme of Vorticella is activated by calcium (Amos, 1971). If, after excitation, the tubules released calcium ions, the whole organelle could be activated rapidly by diffusion, since no part of the spasmoneme is more than 100 nm from a tubule. The tubules appear to form a closed system, more similar to the sarcoplasmic reticulum than to the transverse tubules of a striated muscle fibre, with which Favard and Carasso compared them. Clearly they do not provide a route for expelling calcium from the cell. The observation of Jones et al. (1970) that the spasmoneme can be activated independently of the myonemes of the body is interesting in view of the continuity between these structures. It suggests that the tubules are not responsible for the longitudinal spread of activation along the stalk, which has been demonstrated by Sugi (1960). Possibly the longitudinal saccules have a conducting function.
It is possible that the rootlets have a common origin in development with scopula organelles. The region where they occur may be regarded as a part of the scopula which is separated from the rest during the growth of the cytoplasmic strand. The cylinders in the cytoplasmic strand may perhaps also arise from scopula organelles. Some of the organelles are situated almost inside growing strands (Fig. 7). From this position they may perhaps be carried down into the cytoplasmic strand as it issues from the scopula, and may subsequently be transformed into the cylinders. Such a theory implies a rapid multiplication of scopula organelles during growth, since the cylinders in the strand are so numerous.
Like their origin, the function of the microtubular structures is a matter of speculation. The rootlets presumably anchor the cytoplasmic strand in the matrix, and the scopula organelles may have a similar mechanical function, though the scopula organelles of Epistylis, studied by Fauré-Fremiet et al. (1962), appear to have some control over the deposition of the extracellular material of the stalk.
I thank Professor T. Weis-Fogh for his encouragement and the provision of facilities for this work, and Dr A. V. Grimstone for his guidance and critical reading of the manuscript. The work was supported by King’s College, Cambridge and by a grant from the Department of Zoology, Cambridge.
ABBREVIATIONS ON PLATES
cylinder composed of microtubules
fibrils lining the sheath
membranous component of the sheath
myoneme within the body
tubules within the spasmoneme