Achnanthes sensu stricto belongs with genera of the Mastogloiales rather than with other monoraphid diatoms (Bacillariophyta)

Abstract

Despite variation in their protoplast organization and wall structure, monoraphid diatoms have traditionally been grouped as a single family or order, intermediate between the araphid and biraphid diatoms. However, the predominantly marine or brackish species of Achnanthes sensu stricto share protoplast and frustule features with representatives of the Mastogloiales rather than with other monoraphid diatoms. Meanwhile, studies of morphogenesis in Achnanthes have revealed that cells become monoraphid by filling in one raphe system during valve formation, indicating that the monoraphid condition is derived rather than primitive. Evidence from light and electron microscopy is presented to support the transfer of Achnanthes to the Mastogloiales, and an emended description of the order is given. It is concluded that the Achnanthales sensu Round et al. is a paraphyletic group and that the closest relatives of the various monoraphid genera must be sought among other raphid diatoms.

Keywords:

• Achnanthes
• Achnanthidium
• Aneumastus
• Craspedostauros
• diatom systematics
• Mastogloia
• monoraphid

Introduction

In their treatment of the diatom genera, Round et al. (1990) grouped all monoraphid diatoms in a single order, the Achnanthales, within the raphid diatoms, subclass Bacillariophycidae, whereas Hustedt (1931–1959) treated monoraphid diatoms as one of his four groups of pennate diatoms, Araphideae, Raphidioideae, Monoraphideae and Biraphideae. Simonsen (1979) questioned the traditional position of the Achnanthaceae as a precursor of the Naviculaceae, arguing, with reference to recapitulation and regressive evolution, that heterovalvate (monoraphid) diatoms were derived from isovalvate biraphid taxa rather than the reverse. He also cited the occurrence of raphe traces on rapheless valves as being reduced raphe slits rather than the beginning of a raphe system. The systematic arrangement of Round et al. (1990) reflects the evidence that raphe slits are initiated in both valves but subsequently filled in on one valve (Mann, 1982; Boyle et al., 1984; Mayama & Kobayasi, 1989). Thus, rather than being regarded as an intermediate on the way to the true raphid diatoms (Hustedt, 1931–1959), the monoraphid taxa are treated as a derived group within them (Round et al., 1990). However, in a recent cladistic analysis of some naviculoid diatoms, Cox & Williams (2006) found that representatives of three monoraphid genera (Achnanthes Bory, Achnanthidium Kützing and Cocconeis Ehrenberg) did not group together, and that Achnanthes was part of a monophyletic group with Mastogloia Thwaites ex W. Smith, Aneumastus Mann & Stickle and Craspedostauros E.J. Cox.

Mann (in Round et al., 1990) created a new order, the Mastogloiales, for Mastogloia and Aneumastus, defined as having: two fore and aft chloroplasts that are H-shaped in girdle view, the plates under the valves being indented under the raphe and connected by a central pyrenoid; areolae that are occluded by cribra or volae, but not hymenes; girdle bands that are open porous bands and valvocopulae that are loculate. Mastogloia is a large genus, found in brackish and marine habitats, and traditionally characterized by the possession of partectal chambers on its valvocopulae. Because of its morphological variability, Hustedt (1933) divided it into a number of sections, some members of which have been studied by Stephens & Gibson (1979 1980a b c ) and Paddock & Kemp (1990). On the other hand, Aneumastus is a much smaller genus, created by Mann & Stickle (Round et al., 1990) for Navicula tuscula Ehrenberg and closely related species (Lange-Bertalot, 2001), sharing chloroplast type and some aspects of valve morphology with Mastogloia, but without loculate valvocopulae. Cox (1999a ) suggested that Craspedostauros belonged in the Mastogloiales rather than with any stauroneid diatoms, because it also possessed the same type of chloroplasts, cribrate areolae and porous girdle bands. This was confirmed by a phylogenetic analysis of stauros and non-stauros bearing diatoms (Cox & Williams, 2000), which placed Craspedostauros in the same clade as Mastogloia and Aneumastus, and which excluded taxa from the Cymbellales and Naviculales (Naviculaceae, Plagiotropidaceae, Stauroneidaceae). The more recent analysis of a larger number of naviculoid diatoms (Cox & Williams, 2006) retained this relationship, but also placed Achnanthes brevipes C.A. Agardh within the same clade as these genera.

The evidence supporting the position of A. brevipes as a member of the Mastogloiales includes the possession of two lobed or H-shaped plastids with a central pyrenoid, arranged fore and aft in the cell, as well as cribrate pores in both the valves and girdle bands. Achnanthes brevipes is a synonym of Achnanthes adnata Bory (Toyoda et al., 2005), the type of Achnanthes sensu stricto (see Round et al., 1990), which largely comprises robust, marine or brackish species (see also Novarino, 1992; Toyoda et al., 2003), in contrast to the smaller, predominantly freshwater achnanthoid diatoms. Other monoraphid genera have contrasting chloroplast types and arrangement (usually only one chloroplast per cell), a variety of pore types and, usually, non-porous girdle bands. The internal raphe fissures are also usually non-coaxial in the latter group, whereas in Achnanthes they are coaxial, simple or hooked to the same side. Echoing Cleve's (1895–1896) conclusion that the Achnantheae “represents rather a facies belonging to widely different types than a family of allied species”, Round et al. (1990) commented that Achnanthes and Achnanthidium are “possibly distantly related”, although they noted similarities in the flexure of the frustule and stalk formation. However, flexed valves occur in several distantly related taxa, such as Gephyria Arnott, Rhoicosphenia Grunow, Campylopyxis Medlin and Rhoikoneis Grunow, and stalks are also produced by other diatom genera, such as Licmophora C.A. Agardh, Climacosphenia Ehrenberg, Pseudohimanthidium Hustedt & Krasske, Cymbella C.A. Agardh, Gomphonema Ehrenberg and Didymosphenia M. Schmidt. As with some of their other supra-generic taxa, it is unclear why Round et al. (1990), having recognized the disparities within this group, retained a monoraphid order unless they were simply unsure where to place the various genera, if the order as a whole was abandoned.

More recently a number of other genera have been recognized within the freshwater achnanthoid diatoms (Bukhtiyarova & Round, 1996; Round & Bukhtiyarova, 1996; Round & Basson, 1997), based on variation in striation and pore type. This paper will discuss the morphology and structure of Achnanthes sensu stricto, with particular reference to A. brevipes, Achnanthes longipes C.A. Agardh and similar taxa, and the systematic position of the genus. It will argue for Achnanthes to be transferred to the Mastogloiales and for the systematics of the other monoraphid diatoms to be reassessed.

Materials and methods

Specimens of Achnanthes sensu stricto, Aneumastus and Mastogloia species were examined from fresh and cleaned samples, and from slides held in The Natural History Museum, London (BM) (Table 1). Because of the diversity within Mastogloia, comparisons focussed on specimens of the generitype, Mastogloia dansei (Thwaites) Thwaites ex W. Smith and other species in the Lanceolatae group (Stephens & Gibson, 1980a ).

Table 1.  Samples examined with locality and collection information (where available)

Designation	Sample type	Locality	Date collected	Other details
BM 12891	Slide	Java		Cleve & Möller 147
BM 18457	Slide	Callau, Peru		Collected by Montagne
BM 18469	Slide	Euganeen		Collected by Meneghini
BM 24351	Slide	Pentland Hills, Lothian, Scotland	July 1854	W. Smith material
BM 24333	Slide	River Tamar, Devon, England	1843	W. Smith material
BM 26358	Slide	Belgium		Van Heurck Types du Synopsis No. 47
BM 32616	Slide	Santafiora, Italy		J.D. Möller slide
BM 99748	Slide	Seascale, Cumbria, England	5 April 1972
E77	SEM stub	Schluensee, Holstein, Germany	15 May 1984
EJC 11	SEM stub	River Thames, Greenwich, England	17 December 1992
EJC 74	SEM stub	Port Phillip Bay, Victoria, Australia	May 1993	Material from R. Wetherbee
EJC 75	SEM stub	½km south of Cape Columbine,  Cape Province, South Africa	6 September 1979
EJC 154	SEM stub	Kinloch Rannoch, Perthshire, Scotland	24 July 2002	Cultured material, isolated from  scraping under-surface of bridge
EJC 157	SEM stub	Langebaan beach, Cape Province,  South Africa	6 September 1979
EJC 159	SEM stub	Mouth of Steenbras River,  Cape Province, South Africa	13 September 1979
EJC 170	SEM stub	Lewes, Sussex, England	1851	W. Smith material of  Mastogloia dansei

Some material was isolated and grown in culture in liquid medium in plastic petri dishes in culture cabinets at about 15°C under a 16:8 h light: dark regime (45 µmol photons m⁻² s⁻¹). Clones were maintained in Woods Hole MBL freshwater medium or in f/2 seawater medium (Stein, 1973) according to their freshwater or marine origin.

After being washed in distilled water, samples were prepared for light and scanning electron microscopy (LM and SEM, respectively) by heating gently in 50% nitric acid, washing well in distilled water and mounting in Naphrax or Zrax diatom mountant (for LM) or on glass coverslips attached to aluminium stubs, subsequently coated with gold-palladium for SEM. Light micrographs were taken using Zeiss microscopes and Agfa Ortho, Kodak T max or 100 Delta pro film. SEM images were taken using a Philips XL30 SEG (NHM) or a Nanolab 7 (Max-Planck-Institute for Limnology, Plön, Germany) scanning electron microscope and Ilford HP4 or Kodak Tri-X 400 pro, or Agfa Pan film, respectively.

Terminology

Achnanthes will be used throughout this paper in its narrow sense only, for the predominantly marine and brackish taxa with cribrate areolae closely related to A. brevipes.

The two valves of a monoraphid diatom frustule have been referred to in various ways: raphid and rapheless, R and P (for pseudoraphe) (Round et al., 1990), lower and upper (Hendey, 1964), or hypovalve and epivalve (because of their consistent relationship to the substratum and curvature of the frustule) (Hendey, 1958; Giffen, 1973). However, the convention is that epi- and hypo- refer to the older and younger valves of each frustule respectively (Ross et al., 1979), rather than to their relationship to the substratum. For simplicity, the notation of R (for raphe) and P (for pseudoraphe) will be used in this paper to discriminate between the two valves of an Achnanthes frustule. The use of pseudoraphe valve is also appropriate for Achnanthes and other monoraphid diatoms because a raphe slit is formed when this valve is initiated, but is subsequently filled-in during morphogenesis. This is in contrast to araphid diatoms in which a raphe slit is never formed.

Several terms have been applied to the solid silica between the rows of pores that form the striae. In the older literature, costa was often used, but Ross et al. (1979) discriminated between interstriae and costae, suggesting that costae should be reserved for later-thickened interstriae. Subsequently, with particular reference to pennate diatoms, Cox & Ross (1981) introduced the term virga as a replacement for interstria. They argued that the latter emphasized position rather than structure and that the structure of pennate diatoms with a single row of areolae in their striae resembled basket-work. Thus virga was suggested to replace interstria and vimines for the crossbars between the virgae. These terms are also more appropriate from a morphogenetic perspective (Cox, 2002). The use of costa will be restricted to those cases where the virgae have become secondarily thickened.

Pores within the striae of raphid diatoms can be variously occluded (Mann, 1981) with three types of occlusion usually being recognized: hymenes, volae and cribra. Hymenate occlusions cover the entire pore but are perforated by small, circular holes, 5–10 nm in diameter. Volate occlusions (sensu Mann) are formed of an unperforated, thin flap of silica that is attached to part of the pore wall, often leaving a curved slit-like opening. However, it should be noted that Mann's interpretation of volae differs from that of Ross & Sims (1972) (Cox, 2004), and also includes flaps of silica attached at more than one point, as in Placoneis Mereschkowsky (Round et al., 1990). The perforations in cribrate pores may be reticulate or regular, but are much larger than in hymenes, >30 nm rather than 5–10 nm in diameter, and often elongate around the margins of the velum, where the cribra are attached by a small number of struts to the pore wall. Some species of Achnanthes also have markedly larger poroids near the apices of the P valve, which have been termed pseudo-ocelli (Hendey, 1964) or orbiculi (McIntire & Reimer, 1974). Because they are occluded by a cribrum, like the stria areolae, it is probably more appropriate to call them orbiculi and avoid the term pseudocellus (cf. Ross et al., 1979), which refers to a “field of areolae decreasing in size from those on the main part of the valve”. The latter is more often a feature of centric than of pennate diatoms.

Striae are usually uniseriate, i.e. with a single row of areolae between adjacent pairs of virgae. Biseriate striae, e.g. in Gomphonema, are formed when developing vimines are slightly offset and branch before joining with their neighbours (cf. Cox, 1999b : figs 67, 68, 70–73, 77–81). If alternate virgae are thickened to form costae, this will create the appearance of biseriate striae in LM, but these are not homologous with the biseriate striae of Gomphonema because they have a different ontogeny.

Results

General morphology and structure of Achnanthes species

The most distinctive characteristics of Achnanthes, and traditionally the defining characters of the genus, are the possession of a genuflexed frustule and its monoraphid condition (Figs 1–11). The concavely flexed valve is raphid (R), the convex valve raphe-less (P), and cells may occur stacked up on each other, attached to a substratum by a stalk produced from the lowermost cell (Fig. 1). Live cells usually possess two lobed or slightly H-shaped chloroplasts with a central rounded pyrenoid, the plastids lying fore and aft in the cells, one on either side of the central nucleus (Figs 1, 3). Some species have many smaller chloroplasts distributed around the periphery of the cells (Fig. 2).

Figs. 1–3. Light microscopical drawings of live diatoms. Fig. 1. Several Achnanthes brevipes cells forming a stack attached by a mucilage stalk produced from the lowermost cell.Fig. 2. Achnanthes longipes in girdle view, showing the presence of many small chloroplasts per cell. Fig. 3. Achnanthes brevipes in valve view, showing the presence of two fore and aft chloroplasts. Scale bar: 10 µm.

Figures 4–14. Light micrographs of cleaned frustules of Achnanthes spp. All valve views arranged with primary side of valve to the left.Fig. 4. Girdle view of a half frustule of A. brevipes. Note the concave valve surface and the pores in the girdle bands. (BM 18469).Fig. 5. Girdle view of single valve of Achnanthes javanica Grunow (BM 12891). Note the concave valve surface and double rows of areolae between pairs of costae. Figs 6, 7. A. brevipes showing R & P valves, respectively (BM 18469, A. subsessilis var. multiarticulata). Figs 8, 9. A. brevipes, R & P valves, respectively (BM 18457, Achnanthes pachypus Kützing). Note the lateral sternum (P valve) and the apical orbiculi (arrowheads) on the secondary side of the valve. Figs 10, 11. A. javanica, R & P valves, respectively (BM 12891). Note that the sternum on the P valve is extremely lateral. Figs 12, 13. A. coarctata, R & P valves, respectively (EJC 154). Note the strongly lateral sternum in the P valve.Fig. 14. A. longipes, R. valve. Note the biseriate rows of areolae between pairs of costae (BM 12891). Scale bars: 10 µm. Figs 4, 6, 7, scale bar on Fig. 4; Figs 5, 8–14, scale bar onFig. 10.

Most species have linear or elliptical valves with blunt apices (Figs 6–14), although a few have slightly constricted valves (Figs 12, 13), some with cuneate apices (Fig. 14). The striae are usually transverse to slightly radiate, comprised of distinct pores and readily resolved with LM (Figs 4–14). SEM reveals that the pores are occluded by external cribra (Figs 15–17, 19, 21–25, 28–35). Although most species have uniseriate striae (Figs 4, 6–13, 15–35), in others, two (or three) rows of pores lie between each pair of transverse costae (Figs 5, 14). Several girdle bands are associated with each valve (Fig. 4); the bands are split rings, perforated by one to several rows of cribrate pores (Figs 26, 27, 32, 33).

Figures 15–27. SEMs of Achnanthes spp. Figs 15–18. External views of R valves. Fig. 15. A. coarctata (EJC 154). Note the more lanceolate valve outline with a central constriction. The polar raphe fissures are laterally deflected. Fig. 16. A. brevipes (EJC 159) showing the more or less straight raphe slits with expanded central endings and laterally deflected polar endings. Fig. 17. A. groenlandica Cleve (EJC 75). Note the strong inflexion of the valve face and the large areolae comprising the striae. The raphe slit extends onto the valve mantle at the poles with very little deflection. Fig. 18. A. groenlandica (EJC 75). Note the more widely spaced areolae within each stria compared with Fig. 17. Figs 19–22. Internal central raphe endings. Fig. 19. A. brevipes (EJC 159). The apparent slight deflection of the central raphe endings over the slightly thickened central ‘nodule’. Fig. 20. A. coarctata (EJC 154). The central raphe endings form tight hooks. Fig. 21. A. groenlandica (EJC 75). Figures 15-27. Note the tight hook-like endings of the central raphe fissures and the transverse thickening across the centre of the valve. The valvocopula appears to curve over the internal edge of the valve. Fig. 22. A. cf. yaquinensis (EJC 157). The central raphe endings are slightly expanded and appear slightly deflected due to the curvature of the internal valve surface and the transverse thickening at the centre. Note that the costae are incomplete near the centre of the valve and partly obscured by the transverse thickening. Fig. 23. Internal polar raphe ending of A. cf. yaquinensis (EJC 157). Note the costae are not complete between the most apical striae. The polar raphe fissure terminates in a small helictoglossa.Fig. 24. Internal view of whole P valve of A. brevipes (EJC 159). Note the polar orbiculi adjacent to the sternum. Fig. 25. External view of whole P valve of A. cf. yaquinensis (EJC 157). Note the lateral sternum and large apical orbiculi on the secondary side of the valve. Figs 26, 27. Internal polar raphe endings of A. coarctata (EJC 154). Both apices of a single valve showing the polar helictoglossae and split valvocopula curving over the slight pseudosepta. Scale bars represent 2 µm.

Figs. 28–35. SEMs of Achnanthes spp. Fig. 28. A. cf. groenlandica (EJC 75). Note that the sternum is only slightly lateral and there is no polar orbiculus. Fig. 29. A. cf. groenlandica (EJC 75). Girdle view of external apex of frustule; note different sizes of areolae and complexity of cribra between R and P valves. The raphe fissure can be seen extending over the valve mantle on the R valve. Fig. 30. External view of whole P valve of A. brevipes (EJC 159). Note the lateral sternum and that the polar orbiculi have lost their cribra. Figs 31–34. External views of valve apices. Fig. 31. A. groenlandica (EJC 157). Girdle view showing raphe extending onto mantle and papillae along valve margins. Fig. 32. A. cf. yaquinensis (EJC 157). Note the large cribrate orbiculus and marginal ridge on the non-sternum side of the valve. A row of small papillae is visible on the mantle edge. Fig. 33. A. coarctata (EJC 154). Note the unusual pore at the valve apex and the slight ridge around the valve face-mantle junction. Fig. 34. A. brevipes (EJC 11). Girdle apex and spine; note papillae along valve margin. Fig. 35. Internal view of A. coarctata, showing details of cribra (EJC 154). Note the variation in areola size and cribrum complexity. Scale bars: 2 µm.

The R valve bears a central raphe system. Externally the central raphe fissures are usually slightly expanded (Figs 6, 8, 10, 12, 14–18), the polar fissures often abruptly deflected to one side (Figs 6, 10, 14–17) but sometimes only slightly deflected and extending onto the valve mantle at the apex (Fig. 18). Internally the central endings are straight (Fig. 19) or hooked to one side (Figs 20–22), the polar endings terminated by helictoglossae (Figs 23, 26, 27). Some species have a stauros-like thickening at the centre of the valves (Figs 21, 22). Later developed costae may separate the striae internally (Figs 22, 23).

The sternum on the P valve is narrow but distinct, varying in position from central to laterally displaced (towards the primary side of the valve) (Figs 7, 9, 11, 13, 24, 25, 28, 30), sometimes slightly curved near the apices (Fig. 24). The stria pores on the P valve may be markedly larger than on the R valve (Fig. 29) and the stria density may also be lower. Some species have a large lateral orbiculus at each apex on the primary (Figs 24, 30) or secondary side of the valve (Figs 9, 25, 32). Spines or ridges may be present at the face-mantle junction of the P valve (Figs 25, 32–34), and small papillate projections of silica are sometimes found on valve margins (Figs 31, 32, 34).

Features of other genera of the Mastogloiales

Live specimens of M. dansei, Mastogloia elliptica (C.A. Agardh) Cleve, Mastogloia grevillei W. Smith, Mastogloia smithii Thwaites, Aneumastus tusculus (Ehrenberg) Mann & Stickle and Aneumastus pseudotusculus (Hustedt) Cox & Williams have two fore and aft, H-shaped chloroplasts with a central pyrenoid (Fig. 36; Cox, 1996: fig 25a–c, e, f). Cleaned frustules of Mastogloia reveal the presence of loculate valvocopulae along each side of the girdle (Figs 39, 42, 44), while isolated valves show the presence of raphe slits that are more or less sinuous on the outer surface, but straight internally (Figs 37, 38, 40, 41, 43). The striae are coarsely areolate (Figs 37, 38, 40, 41, 43), uniseriate in most of these species but biseriate in M. grevillei (Figs 40, 41). Central areas are usually small and round.

Figs. 36–49. Light micrographs of other members of the Mastogloiales. Fig. 36. Live cell of Aneumastus (Schluensee, Holstein, Germany) showing fore and aft chloroplasts. Figs 37–39. Cleaned material of Mastogloia dansei from the type slide (BM 24333). Figs 40–42. Cleaned material of Mastogloia grevillei from the type slide (BM 24351). Figs 43, 44. Cleaned material of Mastogloia lacustris (BM 26358). Figs 45–47. Cleaned valves of Aneumastus tusculus (BM 32616). Figs 48, 49. Cleaned valves of Craspedostauros britannicus E.J. Cox (BM 99748). Scale bars: 10 µm. Figs 37–47, scale bar onFig. 37. Figs 48, 49, scale bar onFig. 48.

SEM confirms that the external raphe fissures are undulate (Figs 50, 51, 53), with slightly expanded central endings (Figs 50, 53) while both polar fissures are deflected towards one side of the valve (secondary), curving over the valve mantle and ending near the valve margin (Figs 50, 51). Internally the raphe slits are straight throughout, and the central fissures are undeflected and narrow (Figs 52, 54, 56). The polar fissures are often hidden by a slight pseudoseptum over which the valvocopula curves (Figs 52, 54, 55). Areolae are occluded by cribra, but these are usually only visible internally (Figs 54–57) because they lie below the external valve surface. In cross-section they are seen to be approximately midway between the valve surface and the innermost projection of the costae (Fig. 57). The loculi on the valvocopula lie along the central portion of the cell (Figs 52, 54), but the external openings are nearer the valve apices (Figs 51–54). There are a few, small, internal openings on the abvalvar side of the valvocopula (Figs 52, 54). The advalvar edge of the valvocopula is strongly undulate, matching the topography of the costae over which it sits (Fig. 53; cf. Novarino, 1990).

Figs. 50–57. SEMs of other members of the Mastogloiales. Figs 50–57. Mastogloia dansei (EJC 170). Fig. 50. External view of whole frustule. Note the sinuate raphe path and the unilaterally deflected polar raphe fissures. Fig. 51. Frustule apex showing the valvocopulae with external openings towards the apex and the non-porous copula of the epitheca. The polar raphe fissure can be seen continuing over the apex, ending near the valve margin.Fig. 52. Internal view of valve plus valvocopula. Note that the partecta occupy slightly more than half the valve length, and are situated centrally. A few small internal openings are visible on the valvocopula (arrowheads) and the boundaries between adjacent partectal chambers can be detected. Part of a non-porous copula is also visible. Fig. 53. External view of a frustule with a break on one side revealing the undulate advalvar margin of the valvocopula where it overlaps the valve margin (fragment of valve visible near the apex). Fig. 54. Detail of one end of valve in Fig. 52. Note the break in the valvocopula over a pseudoseptum at the valve apex. Fig. 55. Internal view of valve apex showing the pseudoseptum overlain on one side by the broken valvocopula. Fig. 56. Internal central raphe endings. Cribrate areolae are visible between the thickened virgae. Fig. 57. Break in valve showing the position of the cribra below the external valve surface. The virgae are markedly thickened internally and the external areola openings are partially occluded. Scale bars: 2 µm, except Fig. 50 where scale bar represents 10 µm.

Like the above Mastogloia spp., cleaned valves of Aneumastus reveal slightly sinuous external raphe fissures, with central endings deflected slightly to the primary side, and polar endings deflected to the secondary side and continuing over the valve mantle (Figs 45–47, 58–62). The internal raphe slits are straight, opening in a slight rib, particularly nearer the apices (Figs 65, 66). The central fissures are straight and unexpanded (Fig. 64), while the polar endings are terminated by helictoglossae, beyond which the rib extends a little (Figs 65, 66). The central area tends to be transversely expanded but does not reach the margin (Figs 59, 60, 62, 63).

Figs. 58–66. SEMs of other members of the Mastogloiales. Aneumastus spp. (E 77). Fig. 58. External views of entire frustules of A. tusculus showing the undulate raphe fissures, with unilaterally deflected polar fissures, and change in striation from uniseriate to biseriate striae near the valve margin. Figs 59, 60. External views of entire frustules of A. pseudotusculus. Note the striae are uniseriate throughout. Fig. 61. Detail of external valve apex of A. tusculus. Note the valvocopula (vc) has a single row of larger openings and the adjacent copula (c) has two rows of pores. Fig. 62. Detail of external central area of A. tusculus. Note the irregular external areola openings where the striae are uniseriate and the more regular external openings in the biseriate section near the margin. Fig. 63. Internal view of valve centre of A. tusculus. Note the cribrate areolae and the change in the arrangement towards the valve margin. Fig. 64. Internal view of valve apex of A. tusculus. Note very slight thickening of the valve margin around the apex and that the raphe rib extends slightly beyond the helictoglossa. The internal face of the valvocopula is perforated by many small areolae. Fig. 65. Internal view of valve apex of A. pseudotusculus. Note the development of a slight pseudoseptum and the regular arrangement of areolae throughout the valve compared to A. tusculus (cf. Fig. 64). Fig. 66. Apex of an A. tuscula frustule which has broken revealing the hollow nature of the valvocopula (arrowhead). Note the many very fine pores on its inner surface and the fewer larger pores on its outer surface. Scale bars: 2 µm.

Figures 67–70. SEMs of other members of the Mastogloiales. Craspedostauros spp. Fig. 67. External valve centre of C. australis E.J. Cox (EJC 74). Note the cribrate areolae and double rows of pores in the girdle bands. Fig. 68. External view of apex of C. capensis E.J. Cox (EJC 75). The areolae decrease in size towards the valve margin. Fig. 69. Internal view of valve centre of C. australis (EJC 74). Note the small helictoglossae at the ends of the central raphe fissures and the thickened transverse stauros. Fig. 70. Internal view of valve apex of C. australis (EJC 74). Note the thickening of the raphe sternum continues beyond the helictoglossa. Scale bars: 2 µm.

Striae are composed of cribrate areolae (Figs 64–66), the cribra rarely visible in external view because they are below the valve surface (Figs 62, 63). The areolae near the axial area may be larger (sometimes transversely expanded) and more strongly defined than those nearer the valve margin (Figs 59, 61, 62, 63). In A. tusculus the striae become biseriate near the margin (Figs 58, 61, 62), whereas in A. pseudotusculus they are uniseriate throughout (Figs 59, 60, 63–65). The distal parts of the striae of A. tusculus lack internal transverse thickenings between the areolae (Figs 64, 65). Girdle bands are split rings with rows of pores (Figs 61, 62). The valvocopula is hollow (Fig. 66) with a series of external openings along its length (Figs 61, 62) and many smaller pores on its inner surface (Fig. 66).

Live specimens of Craspedostauros also have two fore and aft, H-shaped chloroplasts with a central pyrenoid (Cox, 1999a : figs 7–10). Valves are usually less robust than those of Mastogloia and Aneumastus, often bluntly linear, slightly constricted at the centre with a narrow stauros (Figs 48, 49; Cox, 1999a : figs 11, 12, 15–18, 20–21). The raphe is central and usually straight, curving abruptly to one side at the poles (Fig. 48). Cells have many girdle bands with two rows of pores (Fig. 49; Cox, 1999a : figs 14, 19, 22).

The external raphe fissures are slightly expanded at the centre, more or less straight or very slightly deflected (Figs 67, 68; Cox, 1999a : figs 26, 29, 30). At the poles the fissures are abruptly deflected and curve over the valve apices (Fig. 68; Cox, 1999a : figs 35–38). Internally the fissures are terminated in helictoglossae at the poles and the centre (Figs 69, 70; Cox, 1999a : figs 40–44, 46, 47). SEM reveals that the areolae are cribrate, rather squarish in outline and variable in size (Figs 67–70; Cox, 1999a : figs 31–34, 39). In the genus as a whole, areolae can vary from approx. 0.15 µm in diameter with only four or five small apertures, to approx. 0.3 µm in diameter with more than a dozen small apertures. Areolae in girdle bands are also occluded by cribra (Cox, 1999a : figs 42, 45, 51, 52).

Discussion

Similarities with other members of the Mastogloiales

As is clear from the cladistic analysis of Cox & Williams (2006), A. brevipes belongs within the same clade as the Mastogloiales (Mastogloia, Aneumastus, Craspedostauros) sharing protoplast and frustule features. The possession of double H-shaped chloroplasts with a central spherical pyrenoid, arranged fore and aft in the cells, seems to be largely confined to members of the Mastogloiales, although Climaconeis Grunow, Biremis D.G. Mann & E.J. Cox and Scolioneis D.G. Mann have similar chloroplast morphology and arrangement (Cox, 1979 1990, unpubl.). However, the latter genera contrast with the Mastogloialean genera in their frustule construction, particularly with respect to stria structure. Within the Mastogloiales, both the striae and girdle bands contain cribrate pores. Climaconeis has pores occluded by hymenes (Cox, 1979; Reid & Williams, 2002), Biremis possesses more complex, chambered valves (Cox, 1990; Round et al., 1990), while Scolioneis has areolate striae occluded by hymenes (Round et al., 1990). Among the other raphid orders, chloroplast morphology and arrangement, as well as areola and girdle band structure are often diagnostic, e.g. Cymbellales, Lyrellales, Bacillariales, Rhopalodiales and Surirellales. Greater diversity in both protoplast and frustule features is found within the Naviculales, but this probably indicates the need for further investigation of their interrelationships and possible taxonomic revision.

Achnanthes brevipes differs from Mastogloia, Aneumastus and Craspedostauros in being heterovalvate (monoraphid) and having straight external central raphe endings, whereas the latter are isovalvate and have deflected external central raphe endings. Boyle et al. (1984) showed that the P valve of Achnanthes coarctata (Brébisson) Grunow is initially raphid, and SEMs of forming valves of A. coarctata also show traces of a raphe slit (Cox, unpubl.). Therefore the monoraphid (heterovalvate) condition of A. brevipes is clearly derived from the biraphid condition, rather than being a fundamental difference between Achnanthes and the Mastogloiales. Central external raphe fissure paths often vary within genera, and some species of Achnanthes have slightly deflected central raphe endings. Features of the raphe fissures, which variously separate all three genera and A. brevipes, also vary within the Mastogloialean genera. Such features were less informative than structural characters, such as pore construction, in other analyses (Cox & Williams, 2000; Cox & Reid, 2004) and are probably rarely informative above the genus level.

Striae may be uniseriate or, at least partially, biseriate in Aneumastus, Achnanthes and Mastogloia, the rows of areolae being separated by transapical costae that develop on the internal surface of some of the virgae during morphogenesis. Thus the apparently biseriate condition of A. longipes is probably derived from a uniseriate condition rather than being fundamentally different. The possession of biseriate striae has been used to characterize Mastogloia grevillei W. Smith, but cells with a biseriate M. grevillei-like valve and a uniseriate M. dansei-like valve have been recorded (as Mastogloia elliptica var. dansei (Thwaites) Cleve (Stoermer, 1967). Thus, the transition from uniseriate to biseriate striae can be effected without any genotypic change. R valves of Achnanthes often have a transverse fascia that may be thickened and more stauros-like, reminiscent of the stauros in Craspedostauros. Interestingly P valves never seem to have a fascia but they may have markedly larger areolae than R valves.

Spaulding et al. (2003) described a new species of Aneumastus, A. aksaraiensis Spaulding, Akbulut & Kociolek, although they expressed some uncertainty about its allocation to this genus. They argued that it differed from other Aneumastus spp. in the shape of the chloroplasts, the presence of biseriate rather than uniseriate striae, the presence of pseudosepta at the poles and a valvocopula that does not extend into the valve interior. However, it can be argued that the differences are of degree rather than kind. The chloroplasts of A. aksaraiensis are fore and aft, H-shaped in valve view with a central pyrenoid, so the only difference here is any lack of lobing in girdle view. Striae in A. tusculus are biseriate near the valve margins, while the biseriate striae of A. aksaraiensis are terminated by single larger areolae near the raphe slits. The degree of internal thickening around these larger areolae sets them off somewhat from the biseriate parts, but the biseriate sections of striae in A. tusculus are similarly set off from the uniseriate rows of areolae. The valve mantle is often slightly thicker at the poles than along the sides of the valve in other Aneumastus, and markedly so in Aneumastus balticus Lange-Bertalot and Aneumastus minor (Hustedt) Lange-Bertalot (Lange-Bertalot, 2001). Thus, the development into what would be recognized as a pseudoseptum in A. aksaraiensis could again be viewed as one extreme of a continuum. Round et al. (1990) described the valvocopula of Aneumastus as modified to fit around the thickened valve margin. Spaulding et al. (2003) interpret this as extending into the valve interior, but fail to illustrate the valvocopula in their species.

Dissimilarities with other monoraphid genera

Since Round et al. (1990) re-established Achnanthidium, eight additional genera (Karayevia Round & Bukhtiyarova, Kolbesia Round & Bukhtiyarova, Lemnicola Round & Basson, Pauliella Round & Basson, Planothidium Round & Bukhtiyarova, Pogoneis Round & Basson, Psammothidium Bukhtiyarova & Round and Rossithidium Round & Bukhtiyarova) have been described for groups of freshwater achnanthoid diatoms (Bukhtiyarova & Round, 1996; Round & Bukhtiyarova, 1996; Round & Basson, 1997). Nevertheless, Round & Bukhtiyarova (1996) admit that some taxa remain unsatisfactorily allocated to Achnanthidium. The discrimination of these genera rests largely on SEM data, particularly areola shape and arrangement, as well as some features of the raphe system.

Although Bukhtiyarova & Round (1996) describe Psammothidium as having cribra occluding their areolae, their micrographs do not convincingly show this. It is more probable that the occlusions of all the freshwater achnanthoid diatoms are either volate or hymenate, unlike Achnanthes with its cribrate areolae. Lemnicola has biseriate and Planothidium bi- to multiseriate striae. Pogoneis has cross-lineate striae, while the areolae of Karayevia and Kolbesia are transversely elongate (Bukhtiyarova & Round, 1996; Round & Bukhtiyarova, 1996; Round & Basson, 1997). In Kolbesia each stria on the R valve appears to consist of a single, extremely elongate areola (Round & Bukhtiyarova, 1996). Unlike Achnanthes but like Achnanthidium, Lemnicola, Pauliella, Planothidium and Psammothidium have non-coaxial internal central raphe endings (Bukhtiyarova & Round, 1996; Round & Bukhtiyarova, 1996; Round & Basson, 1997). Karayevia, Kolbesia, Pogoneis and Rossithidium have coaxial internal central raphe endings like Achnanthes, but their external polar raphe fissures are not strongly deflected nor do they continue over the valve apices (Round & Bukhtiyarova, 1996; Round & Basson, 1997). Kolbesia has hooked external polar raphe fissures rather like Navicula (Round & Bukhtiyarova, 1996), while the external polar raphe fissures of Lemnicola are simply deflected (Round & Basson, 1997), but to opposite sides.

In addition to the achnanthoid genera, three cocconeid genera are also included in the Achnanthales by Round et al. (1990): Cocconeis, Anorthoneis Grunow and Campyloneis Grunow. They have broadly elliptical to almost circular valves that are often very shallow. There are often more marked differences between the areolation on the R and P valves than in the achnanthoid genera. Unlike Achnanthes, both Cocconeis and Anorthoneis have areolae occluded by hymenes. However, Campyloneis has cribrate areolae on the P valve (Round et al., 1990; de Stefano et al., 2003), internal septa and two chloroplasts (Hendey, 1964) rather than the single C-shaped chloroplast seen in Cocconeis and Anorthoneis. In all three genera, the internal central raphe fissures are non-coaxial, unlike Achnanthes. More recently, two additional cocconeid genera have been recognized, Psammococconeis Garcia (2001) and Amphicocconeis de Stefano & Marino (2003). Unlike the other cocconeid diatoms but like Achnanthes, both have coaxial raphe slits but, like Cocconeis, Anorthoneis and the R valves of Camplyoneis, their areolae are occluded by hymenes.

Evolution of the monoraphid condition

With the exception of Achnanthidium and Achnanthes spp., which can attach by stalks, monoraphid genera usually grow adnate to a substratum, attached by the R valve. Even in the stalked forms, the stalks are secreted from the R valve only. Therefore this attached mode of existence may offer an explanation for the development of heterovalvy and the loss of a functional raphe system on one valve. If one assumes that the raphid valve became the means of attachment rather than motility, and polysaccharide secretion was restricted to that valve only, cellular activity associated with polysaccharide secretion would then became localized near the functional raphe system and could allow the other to be lost. Andrews (1981) suggested that the loss of a functional raphe system was probably in response to an environmental pressure, but it could equally be the outcome of an increased tendency to attach and be sessile. Mann (pers. comm.) has suggested that it could be a response to parasitism; infilling the exposed raphe slits would eliminate them as a mode of entry for any parasite. The occurrence of monoraphy in a number of morphologically distinct genera, accompanied by different structural and protoplast features, supports the argument that it has arisen on a number of occasions (cf. Cox & Williams, 2000) and that the Achnanthales sensu Round et al. (1990) is a paraphyletic group requiring taxonomic revision. Andrews (1981) recognized that the monoraphid diatoms were a morphologic rather than phylogenetic category and, therefore, that not all its members need share common ancestry. Hardly any of the taxa discussed in this paper have been included in molecular studies of diatom phylogeny, but the limited data available (Medlin & Kaczmarska, 2004; Sorhannus, 2004) support the separation of Achnanthes (sensu stricto) from other monoraphid diatoms, and the hypothesis that the monoraphid condition arose on more than one occasion.

Ontogenetic evidence confirms that both valves are initially raphid, one becoming infilled during its development (Mayama & Kobayasi, 1989; Cox, unpubl.) and there are records of vestigial raphe slits in several monoraphid and araphid diatoms (Le Cohu & Maillard, 1983; Lange-Bertalot & Le Cohu, 1985; Kociolek & Rhode, 1998). The presence of a vestigial raphe system was used to support the transfer of several species from Asterionella Hassall to Actinella Lewis (Kociolek & Rhode, 1998). Diadesmis gallica W. Smith provides another example of the loss of a functional raphe system, which, in this case, can progress to the infilling of both raphe systems and the development of linking spines (Granetti, 1977; Cox, 2006). While cladistic analysis using both protoplast and frustule data supports Achnanthes being a member of the Mastogloiales (Cox & Williams, 2006), the relatives of the other monoraphid taxa must be sought among other naviculoid diatoms, including the fossil members.

Mastogloiales D.G. Mann emend. E.J. Cox

Cells with two fore and aft chloroplasts that are usually more or less H-shaped in girdle and valve view, the plates under the valves being indented below the raphe and connected by a central pyrenoid. Striae comprising areolae that are occluded by cribra, but not hymenes. Virgae sometimes thickened to form well-defined but narrow costae, pairs of costae sometimes enclosing more than one row of pores so that the striae appear biseriate or multiseriate in LM. A stauros or fascia may interrupt the striae at the centre of the valve. External raphe fissures may be sinuous but internal fissures are straight, occasionally slightly hooked, and the central internal fissures are not unilaterally deflected. Girdle bands are open porous bands, with one to several rows of cribrate pores and valvocopulae that are sometimes hollow or loculate. Cells may be genuflexed in girdle view with functional raphe slits on one valve only (the concave valve).

Genera included: Mastogloia, Aneumastus, Craspedostauros, Achnanthes.

Acknowledgements

Thanks are due to Peter York, Alex Ball and Chris Jones for their technical support with both light and electron microscopy. I also thank David Williams and Pat Sims for their constructive comments on drafts of this manuscript and their willingness to discuss this work. Two reviewers provided very helpful comments and editorial feedback on the paper, which were much appreciated.