Observations on the Ultrastructure of the Choanoflagellate Codosiga Botrytis (Ehr.) Saville-Kent with Special Reference to the Flagellar Apparatus

J. Cell Sci. 17, 191-219 (1975) 191 Printed in Great Britain

OBSERVATIONS ON THE ULTRASTRUCTURE OF THE CHOANOFLAGELLATE CODOSIGA BOTRYTIS (EHR.) SAVILLE-KENT WITH SPECIAL REFERENCE TO THE FLAGELLAR APPARATUS

D. J. HIBBERD Culture Centre of Algae and Protozoa, 36 Storey's Way, Cambridge, CB3 oDT, England

SUMMARY The ultrastructure of the choanoflagellate Codosiga botrytis is described with particular reference to the flagellar appendages, the flagellar rootlet system, the transition zone, the basal body and accessory centrioles, and the stalk. The controversial early reports of flagellar appendages in this species have been confirmed and they have been detected in 2 further species, Salpingoeca frequentissima and Monosiga sp. The appendages consist of a delicate bilateral vane 2 fim wide on either side of the axis, composed of extremely fine overlapping or interwoven fibrils. The flagellar root system consists of a large number of radiating microtubules associated with bands of electron-dense material near the basal body; striated roots are absent. The microtubules extend from several separate foci, those in any one group originating near a composite electron-dense band, and for a distance of 300 nm from the basal body they are separated by blocks of interstitial material. The flagellar basal body forms one of a diplosome pair of centrioles. The triplet microtubules of the accessory centriole are embedded in amorphous electron-dense material and the whole is enveloped in a sheath of similar appearance. The existence of a third centriole close to the diplosome pair is also reported. The relatively complex structure of the flagellar transitional zone is described. The stalk is composed of a core of circular lacunae, which may or may not contain finger-like protoplasmic extensions of the posterior end of the cell, surrounded by a continuation of the sheath material which encloses the remainder of the protoplast. In the stalk only there is a further closely sheathing layer about 15 nm thick which is regularly striated, the spacing of the striations in shadowcast material and sections being about 3 times that measured by negative staining. The structure of choanoflagellates differs widely from that of the algal class Chrysophyceae, the group in which they are included in some classifications, and from the remainder of the algae; they do not appear to have a place in either the algae or the plant kingdom. The structure of Codosiga botrytis is briefly compared with that of sponge choanocytes and collared cells in the Metazoa and some of the possible phylogenetic implications of this are indicated.

INTRODUCTION The choanoflagellates, or collared flagellates, are a clearly circumscribed group of unicellular, marine, brackish or freshwater monads, the most distinctive characters of which are a single anteriorly directed flagellum surrounded by a ring of tentacles (the collar). The protoplast is colourless and is enclosed in a periplast which may be either 192 D.J.Hibberd a relatively simple, close-fitting sheath, or an elaborate basket-like lorica composed of ribs of silica (Leadbeater, 1972 a). The taxonomic position and phylogenetic relation- ships of choanoflagellates have been the subject of much interest and speculation and these organisms still have a firm place in classifications of both the plant and animal kingdoms. In the algae they form a Subclass in the Chrysophyceae (golden-brown flagellates) in the system of Bourrelly (1968) and the Class Craspedophyceae in the Division Chromophyta (Chadefaud, i960; Christensen, 1962, 1966); in the Protozoa they are included in the Order Choanoflagellida in the Class Zoomastigophorea (Honigberg et al. 1964). The controversy concerning the phylogeny of these organisms stems from the discovery by James-Clark (1868) of cells in sponges with an apparently identical structure, and arguments concerning the relationship of the sponges to choanoflagellates continue to the present day. Phylogenetic speculation has recently taken a new turn with the demonstration of collared cells not only in sponges but in a wide range of metazoan groups (reviewed by Norrevang & Wingstrand, 1970, and Lyons, 1973) and it has been suggested (Norrevang & Wingstrand, 1970) that the choanocyte-like cell is a fundamental cell type in metazoans, probably derived phylogenetically from some flagellate ancestor. Notwithstanding the general interest of choanoflagellates and the desirability of detailed comparison with choanocytes in other groups and with the algae with which they are said to be related, detailed information on their generaljinternal structure is lacking. Recent work has mostly been concerned with examination of the external morphology of species important in the marine nanoplankton (Throndsen, 1970; Leadbeater, 1972a, b; Thomsen, 1973; Leadbeater & Morton, 1974a), though some of these papers include sections which show the general internal structure of the species concerned. The only studies using modern methods concerned specifically with internal morphology are those of Laval (1971) and Leadbeater & Morton (19746), the latter dealing mainly with the process of ingestion. The present communication provides an account of the internal and external ultrastructure of Codosiga botrytis (Ehr.) Saville-Kent. This is a common and widely occurring freshwater representative of the group and there are several early accounts of its morphology as seen in the light microscope (Saville-Kent, 1882; James-Clark, 1867, 1868; Fisch, 1885; France^ 1897; Lapage, 1925; de Saedeleer, 1927, 1929) summarized by Hollande (1952). Codosiga botrytis is the type species of the Genus and therefore is of particular interest since the taxonomy of the whole group, particularly the forms lacking a complex lorica, is much confused. It was also of great interest to examine this particular species to clarify the structure of the lateral flagellar appendages first described by Petersen (1929) and Vlk (1938) from light microscopy of material prepared by Loffler's method for bacterial flagella and later by Petersen & Hansen (1954) from electron microscopy of mordanted and fuchsin-stained material. Owing to the relatively poor resolution of this method, the detailed structure of these putative appendages could not be determined and there has therefore been much speculation as to their true nature, because of the phylogenetic importance of flagellar appendages in the algae, and also because of their apparent absence from all other choanoflagellate Infrastructure of Codostga botrytis 193 species examined subsequently. The increased resolution obtained by shadowcasting whole mounts with chromium rather than with gold/palladium alloy and the use of negative staining in the present study has enabled this long-standing problem to be largely resolved.

MATERIALS AND METHODS The present lack of information on the internal structure of choanoflagellates can almost certainly be attributed to lack of sufficient quantities of material for embedding, since these organisms, whilst common in nature, rarely occur in quantity, and only 2 species, both marine, have so far been brought into culture (Gold, Pfister & Liguori, 1970; Leadbeater & Morton, 19746). Certain species, however, live attached to other free-living animals and plants and it is therefore possible to obtain material indirectly by collecting these host organisms in bulk. The majority of the material used for the present study was a net plankton sample from Grasmere in the English Lake District collected on 20 June 1973, which consisted mainly of the diatom Asterionellaformosa var. formota Hassall on which the Codostga grew in relatively large quantity. The material was posted to Cambridge and was processed on the day following collection. Supplementary observations on the flagella of Monosiga sp. and Salpingoeca frequentissima (Zach.) Lemm. were made on samples similarly collected from Blelham Tarn on 20 November 1973 and 30 April 1974 respectively. For light microscopy, material was examined alive using phase-contrast illumination. Material for shadowcasting was prepared by fixing a suspension of the plankton, concentrated by centri- fugation, on a coated grid by exposure to the vapour of 2 % osmium tetroxide for 30 s-i min. Excess liquid was removed from the grid after allowing the material to settle, and the preparation was then allowed to dry before shadowcasting with chromium. Material for negative staining was fixed in osmium tetroxide vapour in the same way, after which a solution of 2 % uranyl acetate in distilled water was applied for 15-30 s, all but a thin film being removed with filter paper before drying. Material for sectioning was prepared by centrifuging the Asterionella bearing the choanoflagellate into a pellet which was then fixed for 17 min in a mixture containing equal parts of 2 % glutaraldehyde and 1 % osmium tetroxide in 0-05 M cacodylate buffer, pH 7, the fixatives being chilled and mixed immediately before fixation. The pellet was then rinsed in 3 changes of buffer over a period of 30 min, gradually increasing the strength of the buffer to o-i M, after which it was treated with 2 % osmium tetroxide in o-i M cacodylate buffer, pH 7, for 2 h, dehydrated in an ethanol series and embedded in Epon. All stages of fixation and dehydration to 95 % ethanol were carried out at o°~4 °C. Sections were cut with a diamond knife on a Reichert OmU2 ultramicrotome, stained in uranyl acetate and Reynold's lead citrate and examined on an AEI EM801 electron microscope. Observations are based on a total of 420 electron micrographs.

Identity of material Codosiga botrytis and Salpingoeca frequentissima are well denned taxa and the material used here clearly identifies with previous descriptions of these species. In the absence of any distinctive feature and in view of the present confused state of the taxonomy of choanoflagellates in general, it has not been possible to identify positively the species of Monosiga used for observations of negatively stained flagellar appendages.

OBSERVATIONS General morphology The structure of C. botrytis as seen in the light microscope is well documented and new data are included here only to form a basis for the observations from electron microscopy and to demonstrate the identity of the material with previously published descriptions. 13 CEL 17 i94 D.J.Hibberd The majority of cells of C. botrytis in the present collection occurred singly (Figs. 2, 3, 5), with only a small proportion in pairs borne on a common stalk (Fig. 4). Clusters of up to 5 cells which are typical for this species were present in later collec- tions from the same locality. The cell body is roughly pyriform with a truncate anterior end (Figs. 3, 5) which sometimes appears slightly depressed. The majority of cells are 8-10 fim long and 5-7 fim wide. A single flagellum arises from the middle of the anterior end of the cells (Figs. 3-6). Its average length is 30 fim (range 23-46 fim) and it bears a distal hair point approximately 3 fim long (Fig. 6). The flagellum either beats rapidly in an approximate sine wave (Fig. 3) or slowly, stiffly and jerkily. When at rest it is fully extended and held rigidly (Fig. 5). The flagellum is surrounded by a funnel-shaped collar approximately 11 fim long (Figs. 3,5) which is made up of 35-40 equally spaced cylindrical tentacles, 150-200 nm in diameter (Figs. 6, 11), which are inserted into the side of the cell slightly below the truncate anterior end (Figs. 5, 8, 15). The tentacles narrow to a diameter of 100-125 nm at the point of attachment to the cell (Figs. 8, 15). Although their contents are in direct contact with the densely granular cytoplasm of the cell body, they appear to contain only amorphous or fibrous-reticulate material (Figs. 11, 15, 16) which in transverse sections sometimes appears to form very faint lacunae or circular profiles. The contents of the tentacles only very occasionally faintly appear to be arranged longitudinally. In cells flattened by coverslip pressure, the collar contracts and its distal opening becomes narrower than its base (Fig. 4). The posterior end of the cell body is drawn out into a thin stalk by means of which it is attached to the cells of Asterionella (Fig. 3); the posterior ends of cells occurring in pairs fuse into a single stalk at a distance of approximately 1-5 /tm from the cell body (Fig. 4). The stalk is about 9 fim long in the majority of cells or cell pairs (range 5-10/im) though in a few it appears to be absent, with the Codosiga cells completely sessile on the diatoms. The diameter of the stalk is approximately 800 nm as measured from shadowcast material (Figs. 36,38). The detailed structure of the stalk is described below. Cells occurring in pairs are held together not only by the stalk but by a lateral linkage of the cell bodies towards their posterior ends by a cytoplasmic bridge, which is just visible in the light microscope (Fig. 4). Favourably oriented shadowcast preparations (Fig. 7) show that the bridge contains a median block of dense material which in sections (Figs. 8, 9) appears as a partition midway along the bridge. This measures approximately 280 x 60 nm and contains transverse pores or less-dense areas about 20 nm in diameter (Fig. 9). The protoplasts are each enclosed in a delicate close-fitting sheath (Figs. 8, 10) which is hardly visible in the light microscope and which appears amorphous in shadowcast preparations (Fig. 7). Sections show that the sheath continues beyond the anterior end of the cell body to surround the tentacles (Fig. 11). Sheath material also covers the anterior end of the cell and further material extends anteriorly in a shallow dish-like curve from the point of flagellar insertion (Figs. 8, 15). The sheath material is composed of 2 layers; the inner, which always appears to be present, is only moder- ately electron-opaque (Figs. 8, 10), whereas the second layer which lies immediately against it is very electron-opaque (Fig. 8) and appears to be a secondary development infrastructure of Codosiga botrytis 195 since it is formed more extensively on the outer faces of the cell pairs which from their shapes appear to have originated from a relatively recent division (Fig. 8). The structure of the sheath material covering the stalk is described below. With the exception of the flagellar apparatus and the stalk, the internal structure of the cell is relatively simple. The most conspicuous organelle is the spherical nucleus which measures approximately 2 /*m in diameter (Figs. 4, 8, 10). It lies centrally, in the anterior third of the cell and contains a single, dense, spherical nucleolus up to 1 /tm in diameter (Figs. 4, 8,10). A single Golgi body, composed of 5-6 rather inflated cisternae, lies immediately anterior to the nucleus and posterior to the flagellar apparatus (Figs. 8, 15). Profiles of vesicles 75-100 nm in diameter occur in relatively large numbers in the region surrounding the Golgi body (Fig. 15). The mitochondria which mostly occur in the cytoplasm surrounding the nucleus (Fig. 13), have flat plate-like cisternae (Fig. 15). Two or three large spherical vesicles almost always occur in the tapering posterior part of the cell (Figs. 4, 5) though they always appear collapsed in sections (Fig. 8). Since they have not been seen to discharge externally, they are possibly food vesicles. However, although they do contain granular material, no recognizable contents, such as more or less intact bacteria, have been seen within them. Small vesicles containing amorphous material and dense bodies 200 nm in diameter are generally distributed throughout the cell (Figs. 8, 10). The remainder of the cell volume is densely packed with ribosomes (Figs. 8, 15).

Flagellar apparatus Flagellar appendages. Shadowcast preparations of the flagellum of Codosiga botrytis reveal the presence of a delicate bilateral wing-like process or vane (Figs. 12-14), which extends along the flagellum from inside the collar, usually to a point approximately two-thirds along its length where it is abruptly truncated. It is possible, however, that it normally extends further than this and that the distal part has been lost, since it appears extremely delicate, often being completely absent or present only as remnants, and even the most nearly intact specimens show signs of disintegration (Figs. 12, 14). The vane could not be detected in sectional material and therefore the exact point at which it originates at the proximal end of the flagellum is not known. The width of the vane on each side of the flagellar axis is approximately 2 /tm, giving an overall width of axis plus vane of about 4-5 fiva. The vane appears to be composed of 2 sets of interwoven or overlapping fine fibrils each of which lies at an angle of about 65°-70° to the axis, and which therefore cross at an angle of about 4Oc-5O° (Fig. 14). The vane is not homogeneous, however, since there appears to be very little or no overlapping of fibrils in a zone 500 nm wide on either side of the flagellum (Fig. 14) and the fibrils in this region appear more loosely arranged. In addition, the vane is ridged longitudinally, with the ridges on opposite sides of the vane spaced at the same distance from the flagellar axis (Fig. 14). The arrangement of the ridges in other specimens follows the same general pattern as that illustrated in Fig. 14, a narrow distal zone about 200 nm wide followed by a broad zone about three times this width being particularly evident, but their distances from either the edge of the vane or from the flagellar axis do not correspond exactly. 13-2 IQ.6 D. J. Hibberd Owing to their extreme fineness, the diameter of the vane fibrils cannot be determined accurately from shadowcast preparations and attempts to negatively stain them in Codosiga were unsuccessful. However, a similar vane was also found in a species of Monosiga, also epiphytic on Asterionella, and although it is lost or disintegrates in

Fig. i. A 3-dimensional reconstruction of the flagellar rootlet system and transition zone in Codosiga botrytis. a, A-tubule; ax, axosome; b, B-tubule; c, C-tubule; ca, composite arc; cf, central filament; »', interstitial material; mt, microtubule; tp, transverse plate. this species much more readily than in C. botrytis, negative staining was successful (Fig. 13). This gives a fibril diameter of 3-5 nm but owing to the state of disintegration of the vane it has not been possible to determine how, if at all, the fibrils interweave. A vane of similar overall dimensions to that in C. botrytis has also been found in Salpingoeca frequentissima but in this case no fibrillar substructure could be detected. Flagellar axis, basal body and transition zone. Cross-sections of the flagellum show Ultrastructure of Codosiga botrytis 197 the usual 9 peripheral doublet microtubules and 2 separate central microtubules (Fig. 31). The flagellum is approximately 255 nm in diameter but the flagellar membrane often appears to be extended laterally into one or two low ridges (Fig. 31). The whiplash point of the flagellum has not been transected but 3 or 4 microtubules can be seen in negatively stained preparations of this region. The basal body is constructed in the normal way as a cylinder of 9 tilted triplet microtubules (Figs. 20, 24). The point at which the triplet structure ends has not been determined from serial tangential sections of the anterior end of the cell but in longitudinal sections there is always an abrupt change of density about at the level at which the cell membrane begins to form a cone (Figs. 21, 22, 27). This change in appearance is almost certainly due to a diminution in the total number of microtubules and it therefore seems likely that the triplet microtubules change to doublets at this point by loss of the C-tubule. At this level delicate strands appear to link the distal end of the basal body to the cell membrane (arrows in Fig. 22). The interval between the termination of the C-tubules and the origin of the central tubules of the axoneme, the transition zone, has a relatively complex structure. At the distal end of the cone of protoplasm from which the flagellum arises, the cell membrane, which continues from this point as the flagellar membrane, is tightly con- stricted around the axoneme so that the overall diameter of the flagellumi s only 210 run. Immediately above this level, the flagellarmembran e is laterally inflated for a distance of 100 nm so that the overall maximum width of the flagellum becomes approximately 450 nm (Figs. 21, 22, 27, 30). At the level of constriction, a single partition, the transverse plate, crosses the lumen. This has a central lenticular thickening, the axosome, which measures 75 x 22-5 nm (Figs. 21, 27). The lumen of the transition zone proximal to the transverse plate contains flocculent material of relatively low electron density (Figs. 21, 27) which appears to extend from the axoneme for a distance of about 15 nm before it spreads laterally to fill the lumen (Fig. 27). The two central tubules of the axoneme appear to originate about 150 nm above the axosome and are linked to it by a thin strand of electron-opaque material (Figs. 27-30). The transverse section through this part of the transition zone illustrated in Fig. 30 shows that spokes run radially from this central strand to the centres of the doublet fibres, and that fainter strands extend between the spokes and also interconnect the doublets in a circle. The doublet fibres are also linked by short connexions between the A- and B-tubules. The profile in Fig. 30 probably lies immediately above the transverse plate, since a part of the flagellarmembran e close to the doublets is included in its thickness; the doublets are connected to this by short arms of dense material which extend outwards from their centres. In the majority of sections passing through or close to the flagellar basal body, the profile of a second basal body, or centriole lying approximately at right angles to the flagellar basal body can be seen (Figs. 15, 18, 21, 24, 27). The accessory centriole is approximately 350 nm long and transverse sections show 9 tilted triplet microtubules (Fig. 28) exactly as in the basal body. In this case, however, the tubules are embedded in amorphous electron-dense material and the whole is enveloped in a sheath of similar material approximately 40 nm thick separated from it by an electron-transparent 198 D. J. Hibberd space of similar width (Fig. 28). An exact transverse section through one extreme end of this centriole has been obtained only once and shows that the profiles of the C-tubules are very weak at this point and that the sheath material is missing (Fig. 29). From their structure and mutual orientation, the flagellar base and accessory centriole are interpreted as a diplosome pair, one of which has become a basal body. Examination of a number of series of sections through this region of the cell has consistently revealed the presence of a further centriole close to the diplosome pair but in a slightly more posterior position. This is illustrated in serial longitudinal section in Figs. 21-23 and in serial transverse section in Figs. 24-26. This third centriole has a similar length to the accessory centriole but it does not appear to be enveloped in amorphous electron-dense material. Microtubular rootlet system. A complex system of radiating microtubules and bands of electron-dense material occurs around the basal body in C. botrytis and this can be seen even in low-power longitudinal and tangential sections (Figs. 8, 11). Longitudinal sections at higher magnifications through or near the basal body show a stack of about 5 microtubules about 160 nm high (Fig. 18) which appears in either longitudinal or transverse section, and sometimes appears longitudinally and transversely sectioned on opposite sides of the same basal body (Fig. 17). These microtubules appear to originate near a band of dense material about 200 nm long and 65 nm wide which lies close and parallel to the basal body (Figs. 15, 17, 21), though the resolution of the micrographs is not sufficient to determine whether their ends lie against the outer face of the band or a short distance away from it. The proximal end of the band is level with the proximal end of the basal body and suitably oriented sections show that it has a compound structure, being composed of 2 outer layers 20 nm thick and 2 very thin layers equally spaced between them (Fig. 17). This band appears in tangential longitudinal section in Fig. 23. From their origin near the composite band to a distance of approximately 300 nm from the basal body the microtubules are separated by blocks of amorphous electron- dense material arranged in vertical rows (Figs. 15, 17, 21). Three of these rows are arrowed in Fig. 17 but the exact number is difficult to determine in longitudinal section. At a greater distance from the basal body than the bands of interstitial material the microtubules lie closer together, the microtubule bundle tapering from a thickness of 200 nm near the basal body to approximately 120 nm where it underlies the plasmalemma. Tangential sections through these structures are relatively rare but more informa- tive. They reveal that the microtubules are not evenly distributed round the basal body but extend from 5 or more separate foci (Figs. 19, 20). The groups of microtubules are of unequal size and are apparently arranged asymmetrically around the basal body (Fig. 19). In all of the relatively few tangential sections through this structure, one of the groups has been found to be larger than the rest, with the microtubules radiating through an angle of almost 180°. The arrangement of the dense material associated with the microtubules can be seen most clearly in this largest group, though that in the others appears to be identical. In tangential sections the thickness and lamellate structure of the central composite Ultrastructure of Codosiga botrytis 199 band appear the same as in longitudinal sections but in this plane the band appears perfectly semicircular in shape, with a radius of 75 nm (Fig. 19). In addition, there appears to be an electron-dense spot at the geometrical centre of the band (Fig. 19), though not enough examples have been seen to determine whether this is a general feature. The amorphous material separating the microtubules also appears as a series of arcs and these have the same centre as the composite arc. There are 3 main arcs spaced at approximately equal intervals of 45 nm from the compound band, the middle one being the densest and thickest, and the outermost the least dense and usually incomplete or not very clear. In addition, tangential sections reveal the presence of less well defined arcs in median positions between the 3 main arcs and between the innermost of these and the compound arc. It is possible, however, that the latter intermediate arc is not a true structure but merely an appearance resulting from the close proximity of the ends of the microtubules if they originate in this position and not immediately against the face of the central compound arc. After leaving this zone the radiating microtubules run in a straight line under the plasmalemma of the anterior end of the cell to a distance of at least 1-2 fim from the basal body (Fig. 20). Those from adjacent complexes therefore cross over (Fig. 20) and probably intermesh shortly after leaving the outermost band of interstitial material, though the exact manner in which they do so has not been determined. About 30-40 microtubules can be seen in tangential sections such as that illustrated in Figs. 19 and 20, and assuming that these are stacked 4-5 deep there is probably a total of 120-200 microtubules in the whole root system. The microtubules curve under the shoulder and down the sides of the cell for some considerable distance, though the point at which they terminate is not known. About 130-160 microtubules have been counted in transverse sections of the cell at the level of the tentacles and so the majority if not all of them are probably still present at this level. These microtubules are evenly distributed around the periphery of the cell in ones, twos and threes and are not arranged regularly in arcs of 4 as has been described for Salpingoecapelagica Laval (Laval, 1971).

Stalk The stalk of Codosiga botrytis has been found to have an unexpectedly complex structure. The central part, at least at the proximal end, consists of finger-like protoplasmic extensions of the tapering posterior end of the cell (Figs. 8, 32). These fingers appear to contain only fibrillar-reticulate material (Fig. 32) and so contrast with the majority of the cell cytoplasm which is densely packed with ribosomes. Transverse sections of the stalk at a greater distance from the cell show a variable number of more or less circular lacunae surrounded by the stalk sheath (Figs. 33, 37). Profiles with 5, 7 and 9 lacunae have been seen and it is possible that this number is related to the number of cells borne on the stalk. The lacunae do not appear to contain the protoplasmic fingers when seen in transverse section and it is evident from shadowcast preparations that the latter may extend only into the most proximal region of the stalk. Occasional preparations have been seen, however (Fig. 36), in which it is equally clear that they may also extend for varying distances inside the sheath and it is therefore probable that the lacunae represent the empty spaces from which the protoplasmic 200 D. J. Hibberd fingers have been withdrawn. It has not been possible to determine whether this withdrawal occurs naturally or is simply a preparation artifact. The fibrillar sheath material of the cell body continues down the whole length of the stalk and there also the very dense outer layer may or may not be present. In addition, there is a denser and more compact sheathing layer lying closely against the lacunar part of the stalk (Fig. 34). In longitudinal section this layer appears about 15 run thick and is composed of subunits with an average spacing of 32 nm (Fig. 34). These appear as rectangular or L-shaped electron-dense blocks about 24 nm wide separated by electron-transparent spaces about one-third of this width in vertical longitudinal section (Fig. 34.) The zones have similar dimensions in vertical tangential section (Figs. 34, 37) where it can also be seen that they are pitched at an angle of about 200 to the horizontal. It can thus be seen that it is this layer which is responsible for the striated appearance of the stalk in shadowcast preparations (Figs. 35, 36, 38), where the average spacing of the striations is also about 32 nm, and where they are also tilted at a shallow angle to the horizontal. The stalk also appears striated in negatively stained material (Fig. 39). In this case however, the average spacing of the striations is only 10-13-5 nm and the material appears as alternating light and dark bands, two complete sets of which are equivalent to one complete set of the units visible in vertical longitudinal or vertical tangential section. The reason for this difference is not clear but it must be due to a more subtle arrangement of the striated material than has been resolved by means of sections.

DISCUSSION The basic arrangement of organelles in C. botrytis is similar to that determined in the electron microscope for this species by Fjerdingstad (1961a) and for a number of other choanoflagellates (Laval, 1971; Leadbeater, 1972a, b; Leadbeater & Manton, 1974; Leadbeater & Morton, 1974a, b). All species so far described have a more or less central, spherical nucleus with a single nucleolus, a single Golgi body lying between the anterior end of the nucleus and the flagellar basal body, a single apically inserted flagellum with a relatively long hair point, a circular collar of tentacles surrounding the flagellum, mitochondria with flattened cristae, and cytoplasm very densely packed with ribosomes. The most important new observations made here are on the structure of the flagellar apparatus, particularly the flagellar vane, the microtubular rootlet system and the transition zone, and on the stalk.

Flagellar vane It has been gratifying largely to confirm the early observations of Petersen & Hansen (1954) of lateral structures of a basically fibrillar nature on the flagellum of C. botrytis, since the 2-/im-broad dense zone on both sides of the flagellum described by them has sometimes been regarded as a possible artifact or simply as mucilage. Petersen & Hansen themselves were uncertain whether the zone was composed of hairs which crossed on drying, whether it was an envelope with a fibrous reticulate structure, or whether it was in one plane or projected on all sides of the flagellum. It is now virtually Ultrastructure of Codosiga botrytis 201 certain that the zone is a flat vane with an organized structure composed of adhering fine filaments of a very narrow diameter, though it has not yet been possible to confirm this by means of sections, and it is also not yet certain whether the fibrils are interwoven or whether they merely overlap. The vane certainly does not resemble any of the varied flagellar appendages described for motile cells in the algae (See Manton, 1965, for a review; Hibberd, Greenwood & Griffiths (1971); Hibberd & Leedale, 1972), and is completely unlike the tripartite tubular flagellar hairs characteristic of the Chrysophyceae and other algal classes (Bouck, 1971, 1972). It most closely resembles the flagellar appendages described for the choanocytes of certain sponges and this will be discussed more fully below. Except for the light-microscope observations of Petersen (1925, 1929) on Salping- oeca sp. and the present observations on Monosiga sp. and Salpingoeca frequentissima, there are no reports of flagellar appendages in other choanoflagellate species. With respect to their apparent absence from marine species, including Codosiga gracilis (Kent), de Saedeleer, Leadbeater & Morton (19746) have pointed out that it is possible that the vane could be destroyed during preparation owing to some complication resulting from the saline environment, and it is also apparent from the present study that the vane is much more delicate in some other freshwater species than it is in C. botrytis. The possibility still remains that the vane is restricted to freshwater forms but this explanation seems much more unlikely in view of the constant structure and occurrence of flagellar appendages in other groups.

Microtubular rootlets Previous information on the structure of flagellar roots in choanoflagellates is scant. Fjerdingstad (1961a) in his study of C. botrytis reported a number of microtubules connected with the basal body, though the structures in this region appear mostly not to have been preserved, and Leadbeater & Morton (1974a) describe microtubules radiating from a ring around the flagellar base in Acanthoeca spectabilis Ellis but no further details are given. Laval (1971) described microtubules radiating along curved paths from the basal body in Salpingoeca pelagica, these being inserted on a compound osmiophilic ring. The outermost part of this structure can be seen from Laval's micrographs to be composed of 3 bands of dense material, though whether this ieally forms a ring around the flagellar base is not entirely clear and the resolution of the micrographs is not high enough to allow detailed comparison with the present study. More recently Leadbeater & Morton (19746) have described a system of radiating microtubules around the basal body in Codosiga gracilis and in this case have illustrated very clearly that the microtubules originate near the innermost of 4 concentric rings of osmiophilic material surrounding the flagellar basal body and not from several separate foci as in C. botrytis. The outer 3 rings in C. gracilis are, as in C. botrytis, blocks of dense-staining material inserted between the microtubules. The flagellar rootlet system in Codosiga botrytis therefore resembles that in C. gracilis and Sal- pingoecapelagica in its general form, but is so far unique in that the points of origin of the microtubules do not form a complete ring around the basal body. One result of this arrangement is that a greater number of microtubules can be incorporated in the 202 D. J. Hibberd system than would be possible if they were inserted on a circular structure at the same distance from the flagellar base as the composite arcs. The number of tubules involved in the structure in C. botrytis must, in any case, be larger than in C. gracilis or Sal- pingotca pelagica since they are stacked 4-5 deep near the basal body, compared with the 2 or 3 layers in C. gracilis (Leadbeater & Morton, 19746) and the apparent single layer in S. pelagica (Laval, 1971). Comparison of the root system in these 3 species suggests that the number of microtubules involved may be proportional to cell size. Microtubules radiating from flagellar basal bodies are common in a wide range of organisms (Pitelka, 1969, 1974) but this particular arrangement appears to be restricted to and characteristic of the choanoflagellates. The compound arc in Codosiga probably belongs to the range of structures collectively described as dense plaques by Wolfe (1972), structures against which both mitotic and non-mitotic microtubules abut and probably originate in a variety of positions in a variety of cell types (see Wolfe, 1972, for a review). They have only rarely been found near to basal bodies, however, and even in these cases the resemblance between the plaque structures and the system in Codosiga is not very strong, as comparison with the striated fibre linking the basal bodies in the green flagellate Chlamydomonas described by Ringo (1967) and the laminated cap in the ciliate Nassula described by Tucker (1970) will illustrate. However, in the light of the general resemblance referred to, it would clearly be of the greatest interest to elucidate the process of cell division in a choanoflagellate. Striated root fibrils, attached to the proximal end of basal bodies and penetrating deeply within the cell are a conspicuous component of the flagellar apparatus in a very wide range of flagellated and ciliated cells (Pitelka, 1969, 1974; Wolfe, 1972) and their absence in C. botrytis, C. gracilis (Leadbeater & Morton, 19746) and apparently also in Salpingoeca pelagica (Laval, 1971) is therefore noteworthy. Since striated root fibrils are commonly supposed to provide anchorage and support for flagella, any such function in C. botrytis must be provided entirely by the microtubular rootlet system, and this is one possible reason for its relatively elaborate structure.

Transition zone The structure of the basal body and transition zone in C. botrytis agrees in general with that described for a wide range of flagellate cells, and appears to be a modified form of Pitelka's (1974) 'Type II' basal body in which the C-tubules terminate at the level of deflexion of the cell membrane (the base of the cone into which the flagellum is inserted in Codosiga), the doublet microtubules continuing upwards for 200 nm or more (150 nm in Codosiga) before a transverse plate appears and the central axonemal tubules begin. In Pitelka's 'Type I' basal body the transverse plate occurs near the site of termination of the C-tubules and is approximately level with the cell membrane. Unusual features of the transition zone in C. botrytis are the conical projection of the cell and the laterally dilated flagellar membrane above it, the relatively long distance between the transverse plate and the central pair of tubules and the thin strand linking these structures. Since C. botrytis has a normal 9 + 2 axoneme, it is a clear exception to the general rule established by Barnes (1961) which has since largely been sub- stantiated (Wolfe, 1972), that all cilia or derivations of cilia which possess a diplosomal Ultrastructwe of Codosiga botrytis 203 basal structure also possess a 9 + 0 fibril pattern. This indicates that C. botrytis possibly derives from a biflagellate progenitor, the second flagellum having become completely reduced.

Stalk The striated structure of the stalk in C. botrytis was not resolved by Fjerdingstad (1961 a) in his study of this species, but Leadbeater & Morton (19746) have recently described a diagonal pattern, with a spacing of 25 nm, in shadowcast stalks of Cgradlis. This spacing is of the same order as that measured here but it is not known whether there is a similar reduction in this interval when measured from negatively stained mateiial. More detailed studies of the stalk aie clearly required in order to determine more fully its structure and the function of the various components.

Tentacles The tentacles of C. botrytis appear to contain only amorphous or fibrous material with no granular component, and thus differ from those in Acanthoeca spectabilis for which Leadbeater & Morton (1974a) describe central dense-staining granules and fine fibrillar or tubular structures, and those in C. gracilis which contain ribosomes and indistinct fibrils or tubules, the latter penetrating deep into the cytoplasm (Leadbeater & Morton, 19746). Since so few species have been sectioned, it is not yet clear whether genuine differences between species exist or whether the appearance of the contents varies with preparation. It does seem certain, however, that the tentacles do not contain microtubules as normally defined, and those described as running from the base of the tentacles in C. botrytis by Fjerdingstad (1961 a) can now be seen probably to represent part of the microtubular rootlet system. The narrowing of the tentacles at their point of attachment to the cell body in C. botrytis has also been illustrated for Acanthoeca spectabilis (Leadbeater & Morton, 1974a) and it is possible that this will be found to be a general featuie in choanoflagellates.

Sheath The membranous sheath in C. botrytis is similar to the 2-layered sheath described by Leadbeater (1972 a) for Monosiga ovata Kent but it is more substantial than the extremely delicate one in C. gracilis (Leadbeater & Morton, 19746). All these sheaths, however, have a similar relatively loose, fibrous-reticulate composition and are clearly different from the firmer envelopes with a more condensed structure in Salpingoecapelagica (Laval, i97i)and S. frequentissima (personal observations). This difference supports the present distinction (Bourrelly, 1968) between the family Monosigaceae, which includes species with mucilaginous envelopes often invisible in the light microscope, and the Salpingoecaceae, which comprises species with firm smooth envelopes.

Cytoplasmic bridge The bridge described by Fjerdingstad (1961 a) as interconnecting only the envelopes of cell pairs in C. botrytis, is almost certainly homologous with the cytoplasmic link 204 D. jf. Hibberd described here. There are no other reports of a similar structure and its function and origin are unknown, though it is almost certainly a remnant of the division process.

Comparison with the algae, sponge choanocytes and metazoan collared cells A full discussion of the possible taxonomic and phylogenetic implications of the new facts on the structure of C. botrytis is not within the scope of the present communication, but it may be useful to draw attention to some of the major issues on which they bear and to some of the more obvious gaps in our knowledge of the groups concerned. The question of the inclusion of choanoflagellates within the algae, either in the Chrysophyceae or as a completely separate class, has been reviewed by Leadbeater (1972 a) who concluded that the evidence for this is tenuous and that the choanoflagellates should probably be placed with the zooflagellates within the animal kingdom. More recently, Leadbeater & Manton (1974) have suggested that mitochondrial substructure alone is sufficient to rule out any close affinity between the choanoflagellates and all the classes of algae included in the division Chromophyta by Christensen (1962, 1966), and that they should be deleted from the plant kingdom. The present results on the structure of the flagellar apparatus strongly support these points of view and on the basis of the structure of the flagellar appendages alone it now seems unlikely that choanoflagellates are even remotely related to the Chryso- phyceae or any of the other classes of algae, though the position of a small number of very poorly known pigmented forms with an apparently similar organization to choanoflagellates still remains enigmatic. With regard to the possible relationship between choanoflagellates and 9ponges, the similarity in the ultrastructure of the collar and the general arrangement of organelles in choanoflagellates and sponge choanocytes has recently been pointed out by Laval (1971). The only notable difference in ultrastructure is the apparent absence of a nucleolus from the nucleus of sponge choanocytes (Rasmont, 1959; Brill, 1973). In the light of the new information on Codosiga interesting new comparisons on the structure of the flagellar apparatus are revealed. Most noteworthy is the similarity between the vane in Codosiga and the flagellar appendages described for sponge choanocytes by Afzelius (1961a, b), Feige (1966, 1969) and Brill (1973), which resemble it in width, division into zones parallel to the flagellar axis and in the apparently meshed structure and narrow diameter of its constituent fibres. This would seem to add considerable weight to the argument that the 2 groups are phylogenetically related, but it must be pointed out that a full comparison cannot yet be made as it has been impossible to detect the vane of Codosiga in sections and that in sponge choanocytes has not been studied by means of shadowcast whole mounts or negative staining. Indeed, a significant difference between the vane in the 2 groups appears to be the relative massiveness and durability of the vane in choanocytes compared with the delicacy and the apparent ease with which the structure is lost in choanoflagellates. The flagellar root systems in the 2 groups may also be organized in a similar way since Fjerdingstad (1961 b) and Garonne (1969) have described microtubules radiating from the flagellar basal body in sponge choanocytes, though these are apparently not associated with dense interstitial material and Afzelius (1961a), Fjerdingstad (19616) Ultrastructure of Codosiga botrytis 205 and Brill (1973) have demonstrated that the flagellar membrane in choanocytes is extended into projecting ridges similar to those described for C. botrytis by Fjerding- stad (1961 a) and in the present paper. Lack of detailed information on sponge choanocytes limits further comparison but in view of the similarities pointed out above and the still enigmatic position of the sponges in the animal kingdom, further comparative investigations are greatly desirable. On the basis of the discovery of collared cells in a variety of Metazoa, Norrevang & Wingstrand (1970) have proposed that not only the sponges but all metazoans may have evolved from a choanoflagellate ancestor, and Lyons (1973) has suggested that collared cells may form a distinct cell type in forms as diverse as choanoflagellates and man. However, the collared cells of the coral described by Lyons (1973) and those of the bracheolaria larva of the echinoderm described by Norrevang & Wingstrand (1970) and those described in the papers cited by these authors, in most cases possess a conspicuous cross-banded flagellar root extending from the basal body deep into the cytoplasm, but they do not appear to possess a complex microtubular root system or a flagellar vane. In view of the conservativeness of flagellar structure in general it is possible that differences of this nature are phylogenetically significant and it is by no means clear that they ' hardly concern the choanocyte pattern as such' (Norrevang & Wingstrand, 1970), since it is also possible that a microvillar ring has evolved more than once in response to the need to ingest external particles. The fact that sponge choanocytes differ from metazoan collared cells in the same ways as choanoflagellates is also relevant to discussions concerning their phylogeny and affinities with the remainder of the Metazoa.

Thanks are due to Dr J. W. G. Lund of the Windermere Laboratory of the Freshwater Biological Association for providing the material used in this investigation and to Dr B. S. C. Leadbeater for access to unpublished information and for commenting on the manuscript.

REFERENCES AFZELIUS, B. A. (1961a). Some problems of ciliary structure and ciliary function. In Macro- molecular Structure and Biological Function, Proc. 1st IUB/IUBS Symposium, vol. 2 (ed. T. W. Goodwin & O. Lindberg), pp. 557-567. New York and London: Academic Press. AFZELIUS, B. A. (19616). Flimmer flagellum of the sponge. Nature, Lond. 191, 1318-1319. BARNES, B. G. (1961). Ciliated secretory cells in the pars distalis of the mouse hypophysis. J. Ultrastruct. Res. 5, 453-467. BOUCK, G. B. (1971). The structure, origin, isolation, and composition of the tubular mastigonemes of the Ochromonas flagellum. J. Cell Biol. 50, 362-384. BOUCK, G. B. (1972). Architecture and assembly of mastigonemes. Adv. Cell molec. Biol. 2, 237-271. BOURRELLY, P. (1968). Les Algues d'Eau douce, Tome II: Les Algues Jaunes et Brunes. Paris: Boubee et Cie. BRILL, B. (1973). Untersuchungen zur Ultrastruktur der Choanocytes von Ephydatia fluviatilis L. Z. Zellforsch. mikrosk. Anat. 144, 231-245. CHADEFAUD, M. (i960). Les veg^taux non vasculaires (Cryptogamie). In Traits de Botanique, Systdmatique, vol. 3 (ed. M. Chadefaud & L. Emberger), pp. 1-1018. Paris: Massot et Cie. CHRISTENSEN, T. (1962). Alger. In Botanik. 2, Systematisk Botanik, vol. 2 (ed. T. W. BOcher, M. Lange & T. Sorensen), pp. 1-178. Copenhagen: Munksgaard. CHRISTENSEN, T. (1966). Alger. In Botanik, 2. Systematisk Botanik, vol. 2 (ed. T. W. BOcher, M. Lange & T. Sorensen), pp. 1-180. Copenhagen: Munksgaard. 206 D. J. Hibberd

FEIGE, W. (1966). Ober breitfliigelige Anhange an der Choanocyten-geissel von Spongilliden. Naturwissenschaften 53, 617-618. FEIGE, W. (1969). Die Feinstruktur der Epithelien von Ephydatia fluviatilis. Zool.Jb. (Anat.) 86, 177-237. FISCH, C. (1885). Untersuchungen uber einige Flagellaten und verwandte Organismen. Z. iviss. Zool. 42, 47-125. FJERDINGSTAD, E. J. (1961a). Ultxastructure of the collar of the choanoflagellate Codonosiga botrytis (Ehrenb.). Z. Zellforsch. mikrosk. Anat. 54, 499-510. FJERDINGSTAD, E. J. (19616). Ultrastructure of choanocyte collars in Spongilla laaistris (L.). Z. Zellforsch. mikrosk. Anat. 53, 645-657. FRANCE, R. H. (1897). Der Orgamsmus der Craspedomonaden. Budapest: K6nigliche Ungarische Naturwissenschaftliche Gesellschaft. GARONNE, R. (1969). Une formation paracristalline d'ARN intranucleaire dans les choanocytes de l'^ponge Haliclona rosea O.S. (Demosponge, Haploscleride). C. r. hebd. Sdanc. Acad. Set., Paris 269, 2219-2221. GOLD, K., PFISTER, R. M. & LIGUORI, V. R. (1970). Axenic cultivation and electron microscopy of two species of choanoflagellida. J. Protozool. 17, 210-212. HIBBERD, D. J. & LEEDALE, G. F. (1972). Observations on the cytology and ultrastructure of the new algal class, Eustigmatophyceae. Ann. Bot. 36, 49-71. HIBBERD, D. J., GREENWOOD, A. D. & GRIFFITHS, H. B. (1971). Observations on the ultrastructure of the flagella and periplast in the Cryptophyceae. Br. phycol.J. 6, 61-72. HOLLANDE, A. (1952). Ordre des Choanoflagelles ou Craspedomonadines. In Traitd de Zoologie, Anatomie, Systematique, Biologie (ed. P. P. Grass6), pp. 579-598. Paris: Masson et Cie. HONIGBERG, B. M., BALAMUTH, W., BOVEE, E. C, CORLISS, J. O., GOJDICS, M., HALL, R. P., KUDO, R. R., LEVINE, N. D., LOEBLICH, A. R., WEISER, J. & WENRICH, D. H. (1964). A revised classification of the Phylum Protozoa. J. Protosool. 11, 7-20. JAMES-CLARK, H. (1867). On the spongiae ciliatae as infusoria flagellata; or, observations on the structure, animality and relationship of Leucosolenia botryoides, Bowerbank. Mem. Boston Soc. nat. Hist. 1, 305-340. JAMES-CLARK, H. (1868). On the spongiae ciliatae as infusoria flagellata; or, observations on the structure, animality, and relationship of Leucosolenia botryoides, Bowerbank. Ann. Mag. nat. Hist., $th Ser. 1, 133-142; 188—215; 250-264. LAPAGE, G. (1925). Notes on the choanoflagellate Codosiga botrytis Ehrbg. Q.jfl microsc. Sci. 69, 47I-5°8- LAVAL, M. (1971). Ultrastructure et mode de nutrition du choanoflagell^ Salpingoeca pelagica, sp. nov. Comparison avec les choanocytes des spongiaires. Protistologica 7, 325-336. LEADBEATER, B. S. C. (1972a). Fine-structural observations on some marine choanoflagellates from the coast of Norway. J. mar. biol. Ass. U.K., 52, 67-79. LEADBEATER, B. S. C. (19726). Ultrastructural observations on some marine choanoflagellates from the coast of Denmark. Br. phycol.J. 7, 195-211. LEADBEATER, B. S. C. & MANTON, I. (1974). Preliminary observations on the chemistry and biology of the lorica of a collared flagellate (Stephanoeca diplocostata Ellis). J. mar. biol. Ass. U.K. 54, 269-276. LEADBEATER, B. S. C. & MORTON, C. (1974a). A light and electron microscope study of the choanoflagellate Acanthoeca spectabilis Ellis and A. brevipoda Ellis. Arch. Mikrobiol. 95, 279- 292. LEADBEATER, B. S. C. & MORTON, C. (19746). A microscopical study of a marine species of Codosiga James-Clark (Choanoflagellata) with special reference to the ingestion of bacteria. Biol. J. Linn. Soc. 6 (in Press). LYONS, K. M. (1973). Evolutionary implications of collar cell ectoderm in a coral planula. Nature, Land. 245, 50-51. MANTON, I. (1965). Some phyletic implications of flagellarstructur e in plants. Adv. bot. Res. 2, i-34- NORREVANG, A. & WINGSTRAND, K. G. (1970). On the occurrence and structure of choanocyte- like cells in some echinoderms. Acta zool., Stockh. 51, 249-270. PETERSEN, J. B. (1925). Om svingtraadene hos flagellater. Nuova Notarisia 36, 191-193. PETERSEN, J. B. (1929). Beitr3ge zur Kenntnis der Flagellaten geisseln. Bot. Tidsskr. 40, 373-389. Ultrastructure of Codosiga botrytis 207 PETERSEN, J. B. & HANSEN, J. B. (1954). Electron microscope observations on Codonosiga botrytis (Ehr.) James-Clark. Bot. Tidsskr. 51, 281-291. PITELKA, D. R. (1969). Fibrillar systems in Protozoa. In Research in Protozoology, vol. 3 (ed. Tze-Tuan Chen), pp. 279—388. Oxford: Pergamon. PITELKA, D. R. (1974). Basal bodies and root structures. In Cilia and Flagella (ed. M. A. Sleigh), pp. 437-469. London and New York: Academic Press. RASMONT, R. (1959). L'ultrastructure des choanocytes d'^ponges. Amtls Set. nat. (Zool.) 12, 253-262. RINGO, D. L. (1967). Flagellar motion and fine structure of the flagellar apparatus in Chlamy- dommas.J. Cell Biol. 33, 543-571. SAEDELEER, H. DE (1927). Notes de Protistologie. I. Craspedomonadines: Material syst6matique. Annls Soc. r. zool. Belg. 58, 117-147. SAEDELEER, H. DE (1929). Notes de Protistologie. II. Craspedomonadines: Morphologie et physiologic Reel Jnst. zool. Torley-Rousseau 2, 241-287. SAVILLE-KENT, W. (1882). A Manual of the Infusoria, vol. 1, pp. 1-472. London: D. Bogue. THOMSEN, H. A. (1973). Studies on marine choanoflagellates. I. Silicified choanoflagellates of the Isefjord (Denmark). Ophelia 12, 1-26. THRONDSEN, J. (1970). Marine planktonic acanthoecaceans (Craspedophyceae) from Arctic waters. Nytt Mag. Bot. 17, 103-111. TUCKER, J. B. (1970). Morphogenesis of a large microtubular organelle and its association with basal bodies in the ciliate Nassula. J. Cell Set. 6, 385-429. VLK, W. (1938). t)ber den Bau der Geissel. Arch. Protistenk. 90, 448-488. WOLFE, J. (1972). Basal body fine structure and chemistry. Adv. Cell molec. Biol. 2, 151-192.

(Received 8 July 1974) 208 D. J. Hibberd

Figs. 2-5. Phase-contrast light microscopy of Codosiga botrytis. Fig. 2. A living colony of Asterionella with 2 attached cells (arrows), x 300. Fig. 3. A living cell showing the pyriform shape and truncate anterior end, the beat of the flagellum, the collar, the stalk and the foot attached to an Asterionella cell, x1000. Fig. 4. A living cell-pair flattened by coverslip pressure showing the central nucleus (n), the parabasal Golgi body (jg), posterior vacuoles, the cytoplasmic bridge ibr), and the narrowed and partially retracted collar, x 1500. Fig. 5. A living cell showing the lateral attachment of the collar, the posterior vacuoles and the stiffly held resting flagellum. x 1000. Figs. 6, 7. Shadowcast whole cells of Codosiga botrytis. Fig. 6. The electron-microscope equivalent of Fig. 5, showing the distal hair point of the flagellum and the tentacles of which the collar is composed, x 2500. Fig. 7. Part of a cell-pair showing the amorphous lorica and the cytoplasmic bridge with a dense median thickening, x 7500. Ultrastructure of Codosiga botrytis 209

C E h 17 2io D.J.Hibberd

Figs. 8—II. General morphology of Codosiga botrytis. Fig. 8. Longitudinal section through a cell-pair showing the spherical nucleus {n) containing a single nucleolus, the parabasal Golgi body {g), mitochondrial profiles (»«), posterior vesicles (v), the cytoplasmic bridge (br), the laterally attached tentacles (t) and the close-fitting sheath (s). x ioooo. Fig. 9. Part of the cell-pair in Fig. 8. at higher magnification to show the structure of the median partition of the cytoplasmic bridge, x 50000. Fig. 10. Transverse section at the level of the nucleus. Labels as in Fig. 8. x 20000. Fig. 11. Tangential section of the anterior end of a cell showing the ring of equally spaced tentacles closely invested by the sheath, and the rootlet system (arrow) surrounding the flagellar basal body. Labels as in Fig. 8. x 20000. Ultrastructure of Codosiga botrytis 211

14-2 212 D.J.Hibberd

Fig. 12. Chromium-shadowcast preparation of the flagellum of Codosiga botrytis to show the general appearance of the flagellar vane, x ioooo. Fig. 13. Negatively stained preparation of part of the partially disintegrated flagellar vane in Monosiga sp. showing the extremely narrow diameter of the component fibrils, x 200000. Fig. 14. Chromium-shadowcast preparation of the flagellar vane of Codosiga botrytis showing the arrangement of the fibrils. The double arrows indicate the directions in which the fibrils lie and the small arrows indicate the more obvious longitudinal ridges. The bracket indicates the part of the vane on either side of the flagellum in which the fibrils do not appear to overlap in the same way as in the remainder, x 25000. Ultrastructure of Codosiga botrytis 213 2i4 D.J.Hibberd

Figs. 15-20. Detailed structure of the anterior end of Codosiga botrytis. Fig. 15. Longitudinal section passing through the flagellum (/) the basal body (bb) and one tentacle (i). cz, accessory centriole; g, Golgi body; m, mitochondria; mt, microtubule. x 40000. Fig. 16. Part of a transverse section between the point of insertion of the tentacles and the anterior end of the cell showing the peripheral arrangement of the microtubules and the contents of a tentacle, x 50000. Fig. 17. Longitudinal section of the flagellar rootlet system showing microtubules in longitudinal and transverse section, the composite arc (ca) and 3 vertical bands of interstitial material (arrowed), x 75000. Fig. 18. Section passing obliquely through the basal body and accessory centriole {cz) and transversely through a 5-deep stack of flagellar root microtubules. x 60000. Fig. 19. Tangential section of the flagellar rootlet system illustrated in Fig. 11 at higher magnification showing the 5 sets of radiating microtubules surrounding the flagellar basal body. In the laigest of these the origin of the microtubules at or near the composite arc can be seen. The 3 main arcs of interstitial material are arrowed. ca, composite arc. For further explanation see text, x 75 000. Fig. 20. Section passing tangentially through the basal body and rootlet system showing more clearly the path of the microtubules. Two points at which tubules originating from adjacent foci appear to overlap are arrowed, x 50000. Ultrastructure of Codosiga botrytis 215 15

mt . 2i 6 D.J.Hibberd

Figs. 21-23. Serial longitudinal sections through the basal body and rootlet system showing the accessory centriole of the diplosome pair (c2) and a further centriole (C3) in oblique section. The level at which the C-tubule of the basal body is apparently lost is arrowed in Figs. 21 and 22. A composite arc (ca) appears in tangential section in Fig. 23. The small arrows in Fig. 22 indicate the delicate strands linking the basal body to the cell membrane, x 50000. Figs. 24-26. Successively deeper transverse sections through the anterior end of a cell showing the flagellar basal body (bb), the accessory centriole of the diplosome pair in oblique section (c2) and a third centriole (C3) in longitudinal section, x 50000. Fig. 27. Median longitudinal section through the basal body, transition zone and proximal part of the flagellum. C2, accessory centriole. See text for further explanation, x 60000. Figs. 28, 29. Longitudinal sections from the same series through the anterior end of the cell passing transversely through the accessory centriole. Fig. 28. The main part of the centriole with the triplet microtubules embedded in amorphous material and the sheath of similar material surrounding the whole, x 50000. Fig. 29. An extreme end of the centriole showing the absence of the sheath and the very weak profiles of the C-tubules. x 50000. Fig. 30. Transverse section through the flagellar transition zone immediately above the transverse plate, x 75000. Fig. 31. Transverse section through the axoneme showing the 9 + 2 structure and the projecting ridges of the flagellar membrane, x 75000. Ultrastructure of Codosiga botrytis 217 21 22 iT'^tL t 23 2i 8 D.J.Hibberd

Figs. 32-39. Structure of the stalk in Codosiga botrytis. Fig. 32. Longitudinal section through the extreme posterior end of the cell showing profiles of 4 protoplasmic fingers, x 40000. Fig. 33. A transverse section at an unknown level showing the core of lacunae surrounded by the 2 layers of the sheath material, x 50000. Fig. 34. Oblique longitudinal section showing the striated layer in vertical section (upper part) and tangential section (lower part), x 75000. Fig. 35. Shadowcast whole mount of the posterior end of the cell and the proximal part of the stalk. The cell protoplasm appears to terminate before the striarions begin, x 25000. Fig. 36. Shadowcast whole mount of a stalk containing protoplasmic fingers of varying length, x 25 000. Fig. 37. Oblique longitudinal section of the distal part of the stalk and the foot which is still attached to a cell of Asterionella. The striated layer can be seen in tangential section in the upper part of the figure, x 40000. Fig. 38. Shadowcast whole mount of a stalk showing striarions with a periodicity of about 34 nm. x 50000. Fig. 39. Negatively stained stalk showing striarions with a periodicity of about 15 nm. x 75000. Ultrastructure of Codosiga botrytis 219