Studies on scale morphogenesis in the golgi apparatus of Pyramimonas tetrarhynchus (prasinophyceae)

Pyramimonas tetrarhynchus Schmarda is the type species of Pyramimonas, a genus of green flagellates with about 60 described species (listed in Belcher, 1966) generally placed in the class Prasinophyceae. The subject of a number of ultrastructural investigations, its general cellular organization is relatively well understood (Manton, 1968; Swale & Belcher, 1968; Belcher, 1969; Swale, 1973; Walne & Moestrup, manuscript in preparation).

As with so many other members of the Prasinophyceae, the cells are covered by non-mineralized scales, 3 different layers on the cell body and 2 layers on the flagella plus the hair scales. The formation of scales in the cisternae of the Golgi apparatus was demonstrated originally in the prasinophycean organism, Halosphaera, by Manton, Oates & Parke (1963). Subsequent studies of other genera also implicated dictyosomes in scale formation (Manton, Rayns, Ettl & Parke, 1965; Manton & Ettl, 1965; Manton, 1966b). That scale formation may occur in the Golgi apparatus is thus well documented; however, with regard to the precise details of scale morphogenesis in the Prasinophyceae, little is known.

Gunning & Steer (1975) addressed the question of cisternal individuality within a dictyosome. Their interest was based, among other things, on the fact that in the prymnesiophyte (haptophyte) (for clarification of terminology see Hibberd, 1976) Chrysochromulina chiton, each cisterna produces a particular type of scale (Manton, 1967a, b). Hence, a given dictyosome consists of cisternae that contain a single scale type. The scales are relatively large, and apparently only one scale is formed in each cisterna. Conversely, with regard to smaller scales, either a few (e.g. Prymnesium, Manton, 1966a, fig. 4) or many (e.g. Pyramimonas amylifera, Manton, 1966b) may be formed together in the same cisterna.

In contrast to the above situations, the formation of several different scale types together in the same cisterna has been reported infrequently and is in need of further study. Thus, we decided to investigate this phenomenon in greater detail when we obtained abundant suitable material in connexion with our other investigations on eyespots and photoresponse systems in these organisms. The morphology of some of the scale types is a species characteristic, and the related dictyosomal processes concerning their formation are thus presumed to be a consequence of genetic control. We here report on the ultrastructural details of scale morphogenesis in the Golgi apparatus of Pyramimonas tetrarhynchus.

The culture of Pyramimonas tetrarhynchus used in this study was isolated on 20 April 1969, by Prof. Tyge Christensen, after he found it in large numbers among Cladophora fracta, which occurred as floating masses in the village pond in Ejby between Herlev and Glostrup on the island of Sjælland, Denmark. Unialgal cultures were maintained in the algal collection at the Institut for Sporeplanter, in soil water medium with CaCO₃ added (Starr, 1964) at 15 °C, approx. 2000 lx and a 16:8 L/D diurnal cycle.

For electron microscopy cells were fixed with 2% glutaraldehyde in soil water medium at about 4 °C for 45 min to 1 h. They were then centrifuged to a pellet, rinsed several times with cold soil water medium over a 90-min period, and postfixed with 2% OsO₄ in 0·1 M sodium cacodylate buffer, pH 7·0, overnight in a refrigerator. After rinsing in buffer, cells were dehydrated in a graded ethanol series followed by propylene oxide, and were embedded in Spurr’s plastic. Thin sections were cut with a diamond knife on a Reichert OmU2 ultramicrotome, mounted on 200-mesh copper grids with carbon or Formvar/carbon films, stained sequentially in uranyl acetate and lead citrate, and viewed with a JEM-T8 electron microscope.

With reference to new details of cellular fine structure, see Walne & Moestrup (in preparation), in which previous studies are also discussed. For purposes of general orientation here, it suffices to reiterate that the cells of P. tetrarhynchus are roughly obovoid, bluntly pointed or smoothly rounded posteriorly, and quadriflagellate, the 4 flagella emanating from the base of an anterior invagination, the flagellar pit. The single, large, cup-shaped chloroplast has 4 anterior lobes, each of which is divided into 2 additional lobes at the level of the flagellar pit (Figs. 1, 2). In this region the cell appears rectangular in transverse section (Fig. 1), as each side is surrounded peripherally by 2 lobes of the chloroplast, interior to which lie the nucleus, the scale reservoir that presumably functions in the release of scales to the cell surface, and 2 contractile vacuoles in a precise arrangement with respect to each other and to the other organelles. At a slightly lower level, a dictyosome lies in each of the 4 corners.

No functional differences among the 4 dictyosomes are suggested from this study, although without definitive biochemical characterization all statements concerning function are only tentative and presumptive, based solely on ultrastructural morphology. Similarly, no apparent functional differences were ascertained in the much smaller species, P. orientalis (Moestrup & Thomsen, 1974), which contains only 2 dictyosomes. In our study, scales were produced in all dictyosomes; numerous scale profiles are seen at low magnification in the 4 dictyosomes shown in Fig. 2.

Each dictyosome, subtended at the forming face by endoplasmic reticulum (ER), consists of about 18 cisternae (Figs. 3, 4, 9). This ER is part of an extensive system that extends from the nuclear envelope along the inner side of the chloroplast (it can be discerned in Fig. 2) and that also surrounds other parts of the cell, e.g. scale reservoir and part of the flagellar pit. It is more completely described in Walne & Moestrup (in preparation).

The peripheral regions of the dictyosomes are surrounded by numerous vesicles, which are also found between the forming face and the subtending ER. Such vesicles can be seen in many of the illustrations but are especially clear in tangential sections of dictyosomes (Figs. 5, 6) and resemble the so-called transitional vesicles (Gunning & Steer, 1975), which are presumed to abscise from the ER and subsequently to coalesce and form new cisternae at the forming face. That portion of the ER which faces the dictyosome may be rough (e.g. Fig. 9), but transitional vesicles apparently are abscised from ribosome-free regions. The corresponding portion facing the chloroplast is generally smooth, except in the regions between the chloroplast lobes (e.g. Fig. 4). Some transitional vesicles are covered by a fuzzy coat and are thus ‘coated vesicles’. There are no discernible morphological differences between the vesicles at the forming face and those at the dictyosomal peripheries. Some of the latter are free and some in contact with the cisternal margins (Figs. 3, 4, 9), and in face view of a dictyosome, the vesicles are seen to be distributed along its entire periphery (Figs. 13, 14).

At the maturing face, scales are released in vesicles. The 6 types have been described by previous workers, but so as to clarify our presentation brief descriptions are repeated here. Innermost toward the cell membrane is a layer of very small scales, the underlayer scales, which, on the cell body, are diamond-shaped; on the flagella, their form is sometimes difficult to discern, but typically they are pentagonal (Figs. 32, 34). In section, the 2 types of underlayer scales are often difficult to distinguish from one another, the best characteristic being the comparatively prominent central hub in the flagellar scales.

On the flagella, the small pentagonal scales are covered by 9 rows of larger scales, in this paper termed limulus scales. Each limulus scale consists of a plate or disk-divided by radii into 7 segments with cobweb-like markings (Fig. 31). Emanating from one margin are 3 fingerlike projections. In addition, the flagella bear a smaller number of hair scales.

On the cell body the underlayer diamond-shaped scales are covered by 2 layers of large, morphologically different quadrangular scales (e.g. Figs. 32, 33). The intermediate layer (IBS, intermediate body scales) consists of box-shaped scales with square bases, in which 8 radii extend from the centre to the margin. At these junctions emanate 8 vertical struts, which are attached above to an upper rim. Both the upper and lower rims of the IBS scales manifest a somewhat scalloped configuration (Fig. 33). Outer body scales (OBS) have similar square bases but only 4 radii, which are parallel to the sides and form a cross. From the centre a prominent column extends and is attached to the corners of the base by 4 arches. Both scale types are often ornamented with small knobs and spines (Figs. 11, 33).

From Manton’s work we know that the large body scales are formed in the Golgi cisternae with their dorsal sides oriented toward the forming face; this also applies to the flagellar limulus scales. In Fig. 34 it can be seen that the 3 projecting ‘fingers’ of the limulus scales are not straight but rather bend toward the flagellar surface, in such a way that dorsal and ventral sides of the scales can be distinguished. A corresponding intradictyosomal orientation, with the dorsal side toward the forming face, can be seen in the scale shown in Fig. 30. In the paper on Mesostigma, Manton & Ettl (1965) stated, however, that there is no such fixed intradictyosomal orientation toward the forming face during production of the small, underlayer scales, which are morphologically similar to the underlayer scales of P. tetrarhynchus. Apparently, a similar situation occurs in the latter also, where during formation the underlayer scales are not always oriented with their dorsal sides toward the forming face but rather toward any part of the inner surfaces of the cisternal membranes. Hence, in Figs. 4 and 12 to the right, underlayer scales can be seen with their dorsal sides away from the forming face, whereas in Fig. 6 in the centre, an underlayer scale has its dorsal side toward the forming face; in both cases, note that the orientation of the scales is with respect to the inner surface of the cisternal membrane.

Figs. 3–6 illustrate intracisternal localization of scale formation. At least 4 and probably all 6 scale types can be formed within the same cisterna. Since it is extremely difficult to distinguish the 2 small scale types during their formation, we assume tentatively that formation of both types occurs in a single cisterna. In Fig. 14 to the upper right, both underlayer types are seen in face view. An additional problem concerns the hair scales, which occur in such relatively small numbers that we have not been successful thus far in identifying them together with all of the other scale types in one and the same cisterna.

Scale formation does not occur randomly in the cisternae. The small, underlayer scales (body or flagellar) are formed only peripherally where the larger scales are not seen. In tangential sections of peripheral regions of dictyosomes (Figs. 5, 6), underlayer scales are seen in rows or in less-ordered arrangement on the inner surfaces of cisternal membranes. In transverse sections (Figs. 3, 4) or surface views (Figs. 13, 14) of cisternae, the small scales are seen to be restricted to the peripheral regions, just internal to ‘coated vesicles’.

The 3 larger types of scales, IBS, OBS and limulus, apparently are formed indiscriminately in the more central portions of the cisternae but always with their dorsal sides oriented toward the forming face of the dictyosome. For illustration of this point, the reader is referred to Figs. 3 and 4, in which the cisternae are numbered from 1 to 18, starting at the forming face. An underlayer scale, an OBS and a limulus scale are formed adjacent to each other in cisterna 17 in Fig. 3. In Fig. 4, cisterna 15: limulus, OBS, limulus, OBS; cisterna 16: limulus, IBS, OBS, underlayer; cisterna 17: underlayer, limulus, IBS, underlayer.

As stated previously, hair scales are not often seen in the cisternae. When they are, often 5 are formed together in nearly parallel array (Fig. 29). They can be distinguished from the fingerlike projections of the limulus scales by their beaded appearance. In Figs. 13 and 14 they are seen to be localized largely to the periphery but always interior to the underlayer scales. In Fig. 3, cisterna 13, right, some hair scales can be discerned in the middle of the lumen, approximately equidistant from the upper and lower cisternal surfaces. In P. orientalis hair scales are generally more centrally located in the cisternae (Moestrup & Thomsen, 1974). Our relatively few observations of them, however, permit no conclusions at this time.

In the following, development of the individual scale types will be considered, with the exception of the small, underlayer scales, which have not been investigated in detail. They are seen in various developmental stages in Figs. 13 and 14.

Due to the rather complex nature of the problem, data from transverse and surface sections of the cisternae are treated separately, since they contribute to an understanding of the problem in different ways.

Transverse sections of dictyosomes are seen in Figs. 7–12, and from comparison with Figs. 3 and 4 it can be concluded that early developmental stages of the larger scales are not clearly discernible before the cisternae reach about position 9, when dense figures with some structural detail can be seen in surface sections of the cisternae. In positions 12 and 13 (Fig. 3) incipient separation of the previously adpressed cisternal membranes allows space in the lumen for further scale development and elaboration. By position 13, 3-dimensional development of IBS and OBS is well under way; and by position 17 their elaboration is completed and scales are fully formed.

Consecutive developmental stages of an OBS are seen in Figs. 7–9, where in Fig. 7 the 2 ‘V’s represent 2 very immature scales in positions 12 and 13; in Fig. 8 a further-developed scale is seen in position 143 and in Fig. 9 nearly mature scales occur in positions 13 and 17. Condensation (or compression) of scales occurs between cisternal positions 13 and 17, so that scale contours become more clearly defined (Fig. 9, see also Figs. 10, 12). At position 18 the intimate relationship between a scale and the cisternal membrane is loosened, and mature scales are transported from the dictyosomes in vesicles derived from cisternal membranes (Figs. 3, 12).

A comparable series showing IBS development can be seen in Fig. 11, where in position 13 scale elaboration is about half finished; in position 16 it is nearly completed, and in position a scale is fully developed. In this case scale condensation occurs between positions 16 and 17.

Figs. 16–28 show the 2 large types of body scales from the earliest discernible developmental stages in surface views of the cisternae (cf. Figs. 13, 14).

The earliest stages of OBS that we have been able to identify are seen in Fig. 16 and consist of a dense quadrangular centre from which extend 8 alternately thick and thin wavy arms. This stage can be discerned positionally between cisternae 9 and 13 (cf. Figs, 3, 7). The next stage is seen in the serial sections shown in Figs. 17 and 18, where the arms are surrounded by an incompletely formed quadrangular structure, delineated along its exterior surface by a membrane (see especially Fig. 18). Each side of the quadrangle apparently consists of 4 units. It is possible (e.g. see Figs. 16 and 21, upper right) that at an even earlier stage the tip of each arm is invested with 2 of these units, bent more or less like an arrowhead. If so, then in the stage shown in Figs. 17 and 18, the 2 bent apical units straighten out and later join together with other such units from the other arms to form a square. The 4 thick arms, still wavy, extend to and form the corners of the basal quadrangle; the 4 thin arms, bent in a similar fashion, extend into the mid-region of each side of the base; the surrounding membrane invaginates at the points of contact with the arms. This stage corresponds to about cisternal position 13.

Subsequently, scale elaboration is initiated in the third dimension. In Fig. 20 to the left, 2 such scales are seen in about cisternae 14 or 15, and one scale to the right in about cisternae 13 or 14 (cf. Fig. 19). During elaboration the 4 thick arms, closely surrounded by the cisternal membrane, extend upward to form a central column (cf. Figs. 8, 9). The 4 thin arms remain in the basal plane, continually in contact with the corners of the quadrangular centre (Fig. 19), from which the column ultimately develops. The wall of the column itself consists of 4 thinner columns (Fig. 14), 2 of which are seen in longitudinal section in Fig. 10. At the base, they develop in contact with the corners of the quadrangular centre. During the course of elongation, the 4 thin arms straighten out to form a cruciate figure (seen best in Fig. 24) in contact with the periphery of the box-like scale, which gradually becomes more regularly quadrangular. By the end of elaboration the thick arms project from the top of the column, and the scales are condensed.

The mode of IBS formation is rather different from that of the OBS, although there are some points of similarity. In the earliest recognizable stages (Fig. 22), 8 thin arms radiate uniformly from an indistinct centre in 8 directions, although in a single plane. At a very early stage the 8 arms are joined by even thinner elements to a cobweb-like structure (Figs. 21, 23), somewhat reminiscent of the flagellar limulus scales (in certain cases some irregularly distributed connexions of a transitory nature are also observed in OBS). As the sides of the IBS develop, there appears first a characteristic stellate figure (Figs. 20, 21), bordered on the exterior margins by a membrane, the two positioned in such a way that the 8 indentations in the outline of the star correspond to the sites where the 8 thin arms appear to be in contact with the membrane (Fig. 23). Gradually, the sides of the scales develop completely and are eventually enclosed by a membrane on both sides, interior and exterior (Fig. 23, bottom; cf. Fig. 11). More explicit details of formation of the 8 struts and the upper and lower rims have been difficult to ascertain, and a complete description of their development is not yet possible. Certain details, however, have been clarified sufficiently to be presented here.

1. During elongation, each strut consists of 2 units (Figs. 25, 26) that become fused together in the mature scale.

2. During or just after completion of the elongation process, each side of the upper rim is covered by about 4 elements (Fig. 28).

3. During elongation the lower rim is very indistinct (Fig. 26, cf. Fig. 11).

4. In a nearly mature scale such as shown in Fig. 27, the dual nature of the struts can still be discerned, but they are now joined to the adjacent struts (upper left in the scale shown).

5. In the earliest stages (e.g. Fig. 21), some dense structures are attached to the apices of each of the thin arms and are reminiscent of the units in the upper rim (Fig. 28) and also of the units that form the sides of the OBS.

On the basis of these observations, morphogenesis of the IBS can be explained in the following way.

At the distal end of each of the 8 thin arms are fastened 2 units, corresponding to those that form the sides of the OBS. The free or distal ends of these units extend during development of the surrounding membrane (from the stellate figure), and each pair forms a strut (Fig. 25), the 2 units of which are completely fused in the mature scale and which are bent distally to form a ‘V’. These will eventually join together with adjacent units at the upper rim, as the individual components extend and/or upon synthesis of additional materials. The formation of the lower rim is not as well documented but appears to occur by the deposition of new material and/or material of the cobweb structure that disappears in the mature scale. When the scale reaches cisternal position 16 and 17, a condensation occurs, similar to that in the OBS. Additionally, the 8 arms straighten out or extend; however, in contrast to the situation in OBS, irregularities in the formative process often result in disruption of the characteristic pattern (cf. scales in face view in Fig. 32).

In the limulus scales, symmetry is based on the number 7, in contrast to 8 in the larger body scales. Early stages are seen in many places in the figures (e.g. Fig. 13). The original cobweb configuration is retained in these mature scales, and the 3 fingerlike projections which are bent irregularly at the outset (Fig. 21, centre) straighten out and condense during development, as each ‘finger’ is surrounded by the cisternal membrane like the fingers of a glove (Fig. 31, and to some extent in Fig. 30).

It commonly happens in many scale-bearing organisms, that the scales fall off during processing for electron microscopy. In the employed fixation of P. tetra-rhynchus the scales were most often retained in their natural positions, and this has provided an opportunity for us to estimate the numbers of scales that cover the cell, based on sections such as those shown in Figs. 32–35. Despite the possibility of error and the consequent uncertainty of our numbers, such measurements are nevertheless of interest, especially since none comparable has been carried out before.

With regard to the flagella, Manton et al. (1963) showed that the limulus scales are attached in an imbricate fashion in 9 longitudinal rows along the flagellar surface. The underlayer scales are also arranged in longitudinal or slightly helical rows (Fig. 34). In 2 transverse sections (Fig. 35) 24 underlayer scales subtending the layer of 9 limulus scales can be counted on each flagellum, indicating that the underlayer scales are arranged in 24 rows. This is confirmed by counts of numerous other transverse sections (including also fig. 12 in Manton, 1968).

The number of flagellar underlayer scales can now be calculated, since the flagellar length is reported to be about 30–35 μm (Swale & Belcher, 1968; Belcher, 1969). The distance between scale centres measures 0·065 μm; thus they number about 369 μm (1:0·065 × 24). Hence, one flagellum approx. 35 μm long is covered by about 12 915 underlayer scales; and about 50∞o such scales would cover all 4 flagella.

The number of limulus scales is difficult to calculate from sections; however, the distance between 2 scale centres measures about 0·2 μm, which matches exactly a corresponding distance in the direct preparation of the Pyramimonas stage of Halo-sphaera shown in Manton et al. (1963), and which may therefore be considered reasonably accurate. Thus, limulus scales number about 45/μm (1:0· 2 × 9); about 1575/one flagellum 35 μm long; and on the 4 flagella about 6000.

The number of hair scales is low, but our data are insufficient to permit even preliminary calculations as to their total numbers.

Figs. 32 and 33 are especially useful in the calculation of numbers of body scales.

The distance between 2 underlayer scale centres (in both directions) is about 0· 063 /tm, making approx. 16/μm (1:0·063 = 15· 87), and a total of about 250/μm ².

Calculations for IBS and OBS scales are identical, in that the centre-to-centre distance is 0·42 μm, giving numbers of about 2·38/μm, and a total each of about 5· 7/μm ².

The cellular surface area can be calculated in the following way. On the basis of measurements of 50 cells, the average length is reported to be 22/μm; the average width 15 μm; the flagellar pit 5 μm deep and 3 μm in diameter at the base (cf. Belcher, 1969, fig. 2). The surface area is most easily calculated by considering the cell as a cone on which is superimposed a cylinder or a quadrangular box with an indentation as shown in Fig. 36. Using the above numbers, we calculate the following:

Surface of the cone: π× 7·5 × 13·3 μm² = 313·5 μm². Surface of the cylinder/box: A figure of about 600 μm² may be considered realistic, based on the following:

If the cell is considered as a cylinder: 2π× 7·5 × 11 μm² =518 μm².

If it is considered to be a box: 4× 11 × 15 μm² = 660 μm².

End surface of box: 225-16 μm² = 209 μm².

Sides of flagellar pit: 2π × 2 × 5 μm² = 62·9 μm².

Total surface area is thus approx. 1200 μm².

Then, the total number of underlayer scales on the body is about 300000 (1200 × 250); of OBS and IBS scales there are about 6800 (1200 × 5·7) each.

Thus, a single cell of the indicated dimensions must consequently be covered by about 350000 underlayer scales, plus about 20000 of the larger scales (i.e. 6000–7000 each of IBS, OBS and limulus), together with a small but undetermined number of hair scales.

This demonstration that a number of morphologically different scale types can be formed in one and the same cisterna raises a number of questions for which there are no immediate answers. It was known previously (Ovtracht & Thiéry, 1972) that in the snail Helix pomatia, dictyosomes in certain cells in the so-called multifid gland produce 2 different chemical compounds (presumably a protein and a glycoprotein) that are localized in different parts of the same cisterna. This can be compared to the intracisternal differentiation in P. tetrarhynchus, where the small underlayer scales are formed peripherally, the larger scales more centrally. Unfortunately at present there are no data on the chemical composition of the scales, and thus we cannot say whether the different scale types are of the same or of different composition. The only such investigation is of Platymonas, where the theca, which is presumably homologous with the underlayer scales in Pyramimonas, consists of a pectin-like polysaccharide (Lewin, 1958; Gooday, 1971; Manton, Oates & Gooday, 1973).

An understanding of scale formation in Pyramimonas is also complicated by the apparently indiscriminate production of several scale types both in the centres and at the peripheries of the cisternae. As stated in the Introduction, Gunning & Steer (1975) have discussed the question of the individuality of a single cisterna, based among other considerations on investigations of the prymnesiophyte Chrysochromulina chiton, where the different scale types are formed successively in different but adjacent cisternae. Presumably in this and similar situations, each cisterna is induced to form a particular type of scale, whereas in P. tetrarhynchus, there is apparently an identical inducement of all cisternae, with the inducement to form different scale types occurring intracisternally at different sites in each cisterna.

The production of several different scale types within a single cisterna has been reported thus far infrequently and with reservations. It can, however, be seen in several places in the literature: fig. 19 in Manton (1968) – also P. tetrarhynchus – shows a cisterna with a limulus scale and a large body scale close together; fig. 19 in Norris & Pearson (1975) – P. parkeae – shows a limulus scale and 3 large scales in one cisterna; fig. 26 in Moestrup & Thomsen (1974) – P. orientalis – shows underlayer scales and lace scales (similar to OBS) in one cisterna, and hair scales and lace scales in the adjacent cisterna. Two scale types are also formed in one cisterna in the prymnesiophyte Chrysochromulina microcylindra (Leadbeater, 1972). The details in the latter are, however, quite different and possibly may not be comparable to the situation in P. tetrarhynchus, since the one scale type is formed in cisternae that lie just outside of and at right angles to the stack of cisternae in the dictyosome proper. Often, however, there are connexions between the 2 groups of cisternae. That intracisternal differentiation may also occur in other species was indicated earlier by Manton (1966b) who noticed in P. amylifera that ‘the small scales are commonly peripheral in position’.

At present, the factors governing the inducement of scale formation can only be surmised, but the fact that the scale types often are species-specific suggests a genetic control, which might occur via the coated vesicles and the transitional vesicles. Both kinds of vesicles derive ultimately by blebbing or abscission from both nuclear envelope and its ER extensions, and quite possibly already at abscission the vesicles carry the inducement factor(s), perhaps in the coatings and/or contents. The information could be transmitted subsequently to different sites in the dictyosomes, since the vesicles surround them and fuse variously, both at the forming face to form new cisternae and apparently also at different levels along the periphery of the dictyosome, whose cisternae often manifest increasing diameters with increasing distance from the forming face.

Spatial elaboration during scale morphogenesis occurs in some organisms via synthesis of scales in pre-formed moulds of the same form as the finished scales, e.g. in the prasinophyte Mesostigma (Manton & EttI, 1965); in the prymnesiophyte Coccolithus huxleyi (Klaveness, 1972); and in the chrysophyte Synura petersenii (Schnepf & Deichgräber, 1969).

In contrast, scale formation cannot occur in this manner in P. tetrarhynchus, since empty moulds are never observed. The incipient, intracisternal 8-armed figures that are transformed ultimately to the large body scales are already evident early on when the cisternae are only flattened sacs. Later, as the dorsal sides of the cisternae expand, the scales develop 3-dimensi0nally. From our micrographs it appears that the cisternal membrane may be involved as a template in the determination of scale form and elaboration, since in its development it appears to pull (or press) the basal scale units into position, e.g. the 4 arches in OBS, the 8 struts in IBS. The cobweb structure in the IBS and limulus scales may function in the determination of scale morphology and symmetry, by maintaining a particular spatial relationship between the 8 spokes (7 in limulus scales) during scale development. After the final geometric form is attained, the cobweb material, which may have served as a pattern, disappears (only in IBS) and may become incorporated, along with newly synthesized material, in the final stages of scale formation.

The extension of the 4 arches in the OBS is believed to occur concomitantly with the extension of the dorsal side of the cisternal membrane, in that the 4 thick arms of the OBS are attached to that membrane and attain their final height as the membrane extends. Additionally, smaller morphological details such as the small thorns on the scales appear to be formed in conjunction with the membranes, which in nearly mature scales follow their contours very precisely (e.g. Fig. 11). The converse situation, that the scale grows and pulls the membrane with it, is not supported by the illustrations. If the foregoing explanation is correct, then presumably the genetic information concerning scale form and elaboration lies in the cisternal membranes, though some degree of self-assembly may be expected if the scales are biochemically different. The difference between this mode of formation and that mentioned above where an empty mould is formed first, is thus not so great, as membranes are involved ultimately in the determination of form in both cases.

Growth curves and information on whether scale formation occurs throughout the cell cycle or only at certain stages are requisite to calculations of the rates of scale formation. Once the exact generation time of P. tetrarhynchus is known, the time required for the passage of one cisterna through the dictyosome can be calculated, using sections such as those shown in Figs. 13 and 14, in which the number of large scales formed in a single cisterna is estimated to be about 16. If scales are formed continuously during a 24-h cell cycle, for example, then a scale would be formed in about 1·5 h. If, however, scales were formed only at certain times during the cycle, then the rate of scale production would be higher.

Future studies on P. tetrarhynchus should include growth curves, determination of the occurrence of scale formation in the cell cycle, the development of techniques for the isolation of the scales in pure fractions and their subsequent chemical characterization. Determination of the chemical composition of the different scale types in P. tetrarhynchus is requisite to further understanding of the intracellular mechanisms of scale morphogenesis.

We thank Professor Tyge Christensen for providing us with cultures of his isolate of Pyramimonas tetrarhynchus. P.L.W. gratefully acknowledges Fulbright-Hays and A.A.U.W. Senior Postdoctoral Research support, N.S.F. grant No. BMS 75-19782, and support from the Institut for Sporeplanter, University of Copenhagen. We thank Dr Jytte R. Nilsson, University of Copenhagen, and Dr Russell A. Porcella, Division of Medical and Health Sciences, Oak Ridge Associated Universities, for helpful discussions, Dr D. J. Hibberd, University of Cambridge, for commenting on the manuscript, and Dr J. Lewinsky, University of Copenhagen, for preparing the drawing.