The Foraminiferal Toothplate, A Review

The concept of the foraminiferal toothplate and the way in which this has changed through time is laid out. Historical developments are contrasted with more recent information. Ultrastructure and lamellarity are used to define and differentiate the toothplate from other internal anatomical elements. The function of the toothplate in the test remains an enigma.


The context
Recent developments in ultrastructural and morphological studies of smaller benthic foraminifera have caused a watershed in foraminiferal taxonomy. The routine use of the SEM to observe sections, half-sections, etched specimens and casts have substantially increased the amount of available information. Concurrently, a plethora of new names has been proposed to describe the features observed. This has resulted in a bewildering array of terms, the definitions of which are seldom precisely stated and, if they are defined at all then they tend to denote different things to different authors.
Glossaries have been assembled (Loeblich & Tappan, 1964, 1987Hottinger et al., 1991) in an attempt to unify usage by defining terms unambiguously, but because these glossaries are based on selections lifted from the literature in general rather than on primary observations, their quality and authority is mixed. The internal structure commonly known as the toothplate is a case in hand. Because of the likely importance of the toothplate in the understanding of the morphology of the foraminiferal test (with all the ramifications this entails), an attempt will be made here to analyse the toothplate historically, morphologically as well as structurally. Because of the changes, subtle or substantial, over the years as to the concept of the toothplate, other structures with which it has been equated will be analysed as well in order to clarify its position.
Historical overview Internal structures were observed and commented upon in a variety of pleurostomellid taxa, and this from a very early date onwards. Seguenza (1859) described in extenso a tube present between foramen and aperture in Ellipsoidina. This was later confirmed and expanded upon by Brady (1868;see Fig. 1). Surprisingly, Brady (1884)

failed to link
Pleurostomella with Ellipsoidina and did not refer again to internal structures. Silvestri (op. var. 1900Silvestri (op. var. -1904 also painstakingly documented the presence and form of the 'tronchi tubulari' in Ellipsoglandulina, Ellipsopleurostomella and Ellipsopolymorphivza (Figs. 2 & 3). In the same vein, the description of Ellipsoidella pleurostonielloides by Heron-Allen & Earland (1910) is surprising because it clearly documents the presence of an internal structure, but no attention is really given to it.
Equally surprising, earlier authors failed entirely to report internal structures in what are now commonly known as the buliminids. Williamson (1858) was unaware of any internal structures in the buliminids, despite the trouble he obviously took of embedding specimens in Canada balsam. He did comment on the peculiarity of the aperture, i.e. the fact that one lip disappears behind the other. Although Carpenter (1862), assisted by Parker and Jones, was the first to seriously study the internal structures of an appreciable number of taxa, as far as the buliminids were concerned he only confirmed the statements made earlier by Williamson, and h e too apparently did not notice the presence of toothplates. Heron-Allen & Earland in their various publications also never referred to (and hence noticed?) the presence of anything inside the lumen of the buliminids.
This lack of attention then continued for more than half a century. The first mention of the toothplate as a structure by Cushman was when he indicated the presence of an internal tube when describing Siphogenerina (Cushman, 1913). Although it was alluded to in passing in his first major classification proposal (1927), Cushman never devoted much attention to this structure. Even in the revisional study of the Virgulininae he payed no particular attention to the possible internal structures of the taxa (Cushman, 1937). Although mentioned, the presence or shape of the internal spiral tube in Virgulina and Bolivina was not used in the descriptions of the individual species nor were figures provided to illustrate this feature.  Seguenza (1859) and Brady (1868) of Ellipsoidina ellipsoides. Brady retraced some of Seguenza's original drawings and added his own new observations. The specimens drawn are housed in the BM(NH) and show the great accuracy of Brady's drawings.
The mentioning of the toothplate in passing by Cushman was followed by most researchers following him in time. Galloway (1933) and Glaessner (1945) went barely beyond the remarks made by Cushman although Galloway used "the presence of a tooth in the aperture which extends down onto the preceding chamber as an undulating column" to differentiate the Bulimininae from the Turrilininae. He also implemented the figures of the pleurostomellids, including in the descriptions of the genera the presence of 'a grooved calcareous ribbon' or 'a calcareous column' (Galloway, 1933).
The first serious study of the toothplate is by the hand of Hoglund (1947). Through the observation of specimens immersed in clearing oils, he was able to accurately draw what he termed the tongues in Bulimina, Globobulimina (Fig. 4) and what was to become Stainfortkia (Fig. 5). However, Hoglund did not go beyond the recording of the features of

Fig. 2.
A reproduction of the type drawings of Ellipsopolymorphina by Silvestri (1901), with the very prominent toothplate running between the successive foramina. the toothplate, he did not venture any suggestions as to its possible function nor did he attempt to use it for taxonomic purposes. Contrary to later authors, he refrained from using terms as toothplate or tongue in the description of Robertina and Robertinoides, both aragonitic forms (Fig. 6).
The major breakthrough as far as the use of the toothplate in foraminiferology is concerned happened when Hofker turned his attention to the internal organisation of foraminifera. In 1951, Hofker defined the toothplate as: "a peculiar inner structure, a more or less developed, often contorted plate, running from a former, now septa1 foramen to the next one, through a chamber. This plate is often attached with one of its sides to the axial wall of a chamber, and shows an often folded free border on the opposite side, which in many cases is ornamented by typical denticulations or other structures." (Hofker, 1951a, In a single stroke, he also extended the range of occurrence of the toothplate from the buliminids (Fig. 7) to a very wide spectre of taxa, including amongst others the rotaliids, elphidiids, conorbinids, robertinids, globorotaliids, &c (Fig. 8). With the vast majority of his subsequent publications, Hofker presented a thorough comparative study in which the toothplate played a major role.
Interestingly, Smout (1954Smout ( , 1955 ignored the occurrence of toothplates in the rotaliids as reported by Hofker to the point of not discussing it. But in a subsequent publication p. 353). (1904), of Ellipsopleurostomel2a pleurostomella. Note especially the careful analysis of the shape of the toothplate as seen from frontal and lateral viewpoints. (Smout, 1956), the toothplate as a structure was discussed in the light of Hofker's studies. Nevertheless, the author shied away from fully assessing the relation between toothplate and other internal structures, including canals.

Fig. 3. Another reproduction of type drawings by Silvestr
In a letter to L. Hottinger, referred to by Hottinger (1967) (3 t f . 273b

Fig. 5.
A reproduction of the type figures of Stainforthin concava (Hoglund, 1947 Dies ist, im Kurzen, eine 'Definition' der Zahnplatte. Wie ich schon des iifteren erwalent habe, ist die Zahnpldtte auch schon in vie 1 en a gg I u tinier t e n For am i n i f e r e n, den V a I vu 1 in i d ae, ausgebildet. Sie findet sich schon in der Boliz~iriri liassica vor." Taking his cue from Hofker, Reiss (1963) re-evaluated the toothplate, differentiating elements of the aperture or the septum from the toothplate proper (ihid. p. 28 et "9.). Like H o f k e r , Reiss w a s o r i g i n a l l y of t h e o p i n i o n t h a t t h e toothplate as a structure occurs i n a variety of taxa far extending beyond the buliminids. In the same vein, Reiss linked the occurrence of toothplates with the construction of canals in low trochospiral forms (Reiss & Merling, 1958;Reiss, 1963).
Restricting the interpretation of the toothplate to the plate encountered in buliminids, N0rvang (1966,1968) attempted repeatedly to understand the morphology and the function of the toothplate through a refined technique, avoiding some of the difficulties inherent in the preparation and observation of classical thin sections.
In a major overhaul of the classifications, Loeblich & Tappan (1964) restricted the use of the term toothplate to internal apertural modifications applicable to the buliminids only: "Problematical also is the question as to whether all socalled tooth plate foraminifera are closely related, as postulated by Hofker. If related to a physiologic function of the animal, a toothplate may have developed at more than one time, just as similar test form, chamber arrangement, or apertural character may appear in agglutinated, porcelaneous, or hyaline lineages. In the Treatise classification, apertural tooth plate development is regarded as an advanced apertural feature which developed independently in various lines. Thus, the entosolenian tube in the Glandulinidae, the internal siphon in the Pleurostomellidae, and the tooth plates of the Buliminidae, Bolivinitidae, and Caucasinidae, or the internal partitions of the Ceratobuliminidae and Robertinidae are regarded as convergent rather than divergent features." A few years later, the SEM became available as an observation tool (OatIey, 1966) allowing for the first time highly detailed observations at high resolutions.

Recent developments
The seminal study of the foraminifera1 test wall by Towe & Cifelli (1967) demonstrated the value of the electron microscope for research attempting to analyse morphological elements and their mutual relations. The careful analyses of lamellarity patterns by Hansen and coworkers helped to resolve many standing problems.
The study by Hansen & Reiss (1971) of the rotalian wall structures gave up the use of the term toothplate in a nonbulimine context, referring to platelike structures in i.a. Anitnoilia, Pseudovofalia, Pararofalia as foraminal plates and coverplates. In this, they supported the proposals by Loeblich & Tappan (1964): "At the umbilical border of the partly resorbed foramen, near the axial chamber wall (previous coil), the inner lining forms a plate-like extension, folded along an axis running towards the spiral side. This plate is glued to the umbilical wall of the adjacent coil at its end near the spiral wall and appears in section as a hook-like structure. This plate is referred to here as the 'foraminal plate' and is present in all chambers, including the last formed one. It is in fact part of the 'toothplate' of Hofker (1951), Reiss & Merling (1958), Reiss (1963), and of Uji6 (1965), as well as part of the 'axial plate' as well as the 'apertural lip' as shown on P1. 21 in Cifelli (1962). In addition to forming the septal flap and the foraminal plate, the inner lining extends from the foraminal plate into the previous chamber where it forms an umbilical 'coverplate', glued to the axial wall (previous coil), to the lip, as well as to the foraminal plate -below the fold-of that chamber. The coverplate leaves a communication open to the chamber interior corresponding to part of the posterior labial aperture. ... The coverplate is in fact part of the 'toothplate' of Hofker; Reiss & Merling; Reiss;and Uji6 (opxit.), and corresponds to part of the 'axial plate' of Cifelli (1962). Chamber wall, septal flap, foraminal plate, and coverplate in a previous chamber, are formed by the inner lining as one continuous structure." (p. 332) "Whether or not the folded foraminal plate of the rotaliaceans is homologous with a true toothplate, like in the Buliminacea, remains questionable. It is certainly not characteristic of a11 rotaliaceans." (Hansen & Reiss, 1971, p. 342).
Buliminid toothplates were for the first time observed in the SEM by Glacon & Sigal (1974), confirming the diversity in shape shown earlier by Hoglund andHofker. Scott (1977, 1978) reported on the toothplates in Bolivinita, extending the preliminaries by Glaqon & Sigal and confirming some of their reserves as to the interpretations made by Hofker. Haynes (1981) apparently restricted the use of the term toothplate also to a buliminid context, but the structure is mentioned as occuring in larger rotaliids where they would form the secondary chambers. For example: "Similar arcuate equatorial chambers occur in the complex rotaliaceans such as the Miogypsinidae and Lepidocyclinidae but appear to be modifications of the toothplate".
Here, Haynes echoes to some extent the terminology used by Hansen & Reiss (1972) when they described the formation of chamberlets in the asterigerinids as being the result of a toothplate cutting across the main lumen.
In a study of the toothplate of Bulinriiia, Verhallen (1986) attempted to resurrect the ideas of Hofker and surmised that the toothplate in Bulimina is folded onto itself so as to form a kind of axial canal "If we assume that all successive central pillars are interconnected by bridge structures, they form a compound internal canal, which apparently runs throughout the entire test, more or less along the test axis, but in a complex, staggered spatial pattern" (Verhallen, 1986, p. 375).
He further postulates that this canal contains ectoplasm, assisting in the uptake of oxygen.
Using the techniques first developed by Hansen, Revets (1989) analysed the lamellarity of the buliminid toothplate, reporting it to be made up of inner lining only. He defined the toothplate as "a piece of inner lining (is) drawn out from under the axial-most part of the aperture into the chamber lumen and attached to the septum, close to the foramen. At least one edge of this plate leaves the chamber wall, becoming the free border." In various subsequent publications, Revets applied the diversity in toothplate morphology within the Buliminacea for taxonomic purposes.
Miiller-Merz (1980) rotaliid foraminifera essentially adopted the position of Hansen & Reiss (1971), rejecting the term toothplate and using umbilical-, foraminal-a n d coverplate to denote structures in the taxa investigated. Contrary to the studies by Hansen & Reiss, no etched sections were prepared but broken specimens and Araldite casts for SEM observation and thin sections for light microscopy were used. Muller-Merz (1980) distinguished a n umbilical flap from a coverplate by pointing out that the former is a primary closure of the umbilical-most part of a chamber, while the latter is its secondary analogue, i.e. not present in the final chamber.
Dealing with Pararotalia, Neorotalia a n d C a l c a r i n a , Hottinger et a/. (1991) returned to the position held by Hofker a n d t h e early Reiss a n d reinstated t h e t e r m   Verhallen. Hottinger et al. (1991) define the toothplate as "a contorted plate running from an intercameral foramen to a n aperture, attached to both; folded in a single, double, or spiral fold; folds or tongues Explanation of Plate 1 Fig. 1,2,4 (Cushman, 1926), tcpotypes. 6. Habitus, showing the early coiled chambers, with most later chambers in rectilinear arrangement (200 pm). 7. Section overview, with the toothplates arranged in opposition, demonstrating the biserial coiling (200 pm). 8. Close-up of two successive toothplates (t) (50 pm). 9. Close-up of the attachment site (as) of the toothplate (t). Note also the peculiar shape of the toothplate (20 pm).

Revets
with free, often serrated distal ends and distally protruding into the aperture; a toothplate separates partly or entirely the main chamber lumen from an axial space (adapertural depression) in post-embryonic stages; interconnected toothplates produce a primary canal; in contrast to a n umbilical plate, the toothplate is never associated with a foliar or stellar chamberlet and it does protrude with a free edge distally and adaxially to the aperture".
As a result, Hottinger et al. (1991) recognise a toothplate in Pararotalia and Neorotalia but deem it absent in Rotalia, Ammonia, Cuvillierina and Calcarina. that the toothplate is a part of the chamber wall running from foramen to aperture, folded onto itself so that it forms a primary canal. But according to Hottinger et al. (1991) the canal may remain open in the direction of the lumen of the adjacent chamber, while Verhallen (1986) claimed that the tube formed in the buliminids is isolated from the chamber lumina. The toothplate always protrudes in the aperture. The Hansen position is less clear-cut. His apparent rejection of the toothplate as occurring in the rotaliaceans did not restrict its possible presence to the buliminids. It is explicitly mentioned in the context of the robertinids, where the feature identified a s t h e toothplate is bilamellar (Hansen, 1979) while in the asterigerinids, the toothplate is identified as doubled-up inner lining (Hansen & Reiss, 1972). The Revets position is the most restrictive of the three. Confining the toothplate to a plate of inner lining running from aperture to foramen, it is effectively limited to the buliminids sensu lato. The photographs provided by Hansen & Reiss of etched sections of rotaliids confirm this point of view: the so-called toothplates (according to the reinterpretation of Hottinger e f al., 1991) a r e a t least bilamellar, and they seem to receive secondary lamination (see Fig. 10, feature w).
Resolving the problem of the nature of the toothplate can only be achieved through the combination of morphological and structural observations. The lamellar nature of the different internal elements allows the recognition of their mutual relationships, while morphology can be used to trace developments of the structures in the different taxa.

Position and shape of the toothplate
A characteristic feature of the toothplate, and recognised as such from early on, is the fact that it is partly visible from outside the test as a protrusion into the aperture. However, this in itself is not sufficient to indicate the presence of a toothplate. The most obvious examples are of course the various kinds of "teeth" found in the aperture of many miliolids. It is well established that these protrusions are indeed protrusions, without any progression into the chamber lumen, let alone coming near the foramen. There are very good reasons to believe that the miliolid "teeth" are constructed by the organism after the building of the chamber to which they belong (Arnold, 1964;Angell, 1980). In the case of the larger, complex miliolids, structures d o occur which run along the entire chamber length, but there is no relation with apertural modifications along the lines seen in the classical toothplate-bearing forms (see e.g.

Hamaoui & Fourcade, 1973 on Xkapydionina).
A necessary condition for a structure to be a toothplate is that it must progress through the chamber lumen towards the foramen. This effectively rules out the milioloid "tooth" but does not dispose of the structure present in a number of agglutinated foraminifera. A toothplate-like structure has been recognised in Tritaxia, Martinottiella and in Clazdina where it forms a fairly massive plate, but clearly troughshaped. The apparent 'toothplate' in Egpwlloides scaher is apparently agglutinated t h r o u g h o u t (Haynes, 1973). Interestingly, an in-depth analysis by Coleman (1980) showed the plates in Clavuliiza to be made u p of calcite alone, without allochtonous particles embedded, while the tooth visible in the aperture is clearly agglutinated. At the s e p t u m , it fuses o n t h e t o p of the foramina1 tooth. Strikingly, the plate is continuous with a calcareous nonperforate layer coating t h e s e p t u m (Coleman, 1980). Technically speaking, the s t r u c t u r e can be seen a s homologous to the toothplate and continuous with a septa1 flap. However, it seems i n a p p r o p r i a t e to b o r r o w terminology used to describe features occurring in a different subordo, and, as stated by Loeblich & Tappan (zjide supra), convergence a n d analogy a r e rife within the foraminifera, a fact to which toothplate-like structures are apparently no exception.
In fully spiral, high trochospiral taxa (i.e. with chambers Explanation of Plate 2 Fig. 1-3. Sagrina pulclzrlla d'Orbigny, 1839, topotypes. 1. Habitus (100 pm). 2. Overview of an opened specimen, already showing the complex nature of the toothplates (100 pm). 3. Close-up of a toothplate (t) in an earlier chamber, illustrating the complex folding of the plate, the presence of internal spines and the positioning in the foramen (f) (20 pm). Fig. 4-6. Staiilforthia concuz~o (Hoglund, 1947), Laeso, Kattegat, Denmark. 4. Habitus (200 pm). 5. Close-up of the aperture, apparently very bulimine in outline (25 pm). 6. View of a dissected specimen, with the very prominent 'toothplate'. Note however the attachment site is well clear from the foramen (50 pm). Fig. 7 Revets in a non-rectilinear arrangement), the asymmetry of the test allows the unambiguous localisation of the toothplate. The upper part of the toothplate is located at the axial side in the asymmetric aperture (Pl. 1, fig. 2). It thereby defines a sulcus, a depression present between itself and the axial apertural border. This sulcus is in fact the internal surface of the trough which is the toothplate proper. The attached border of the toothplate is continuous with the axial border of the aperture and, lower down, with the lateral axial chamber wall. The upper part of the free border is very often capped by the apertural lip which has turned into the apertural opening, and it is this upper part which is visible in the aperture. The trough, that is, the toothplate, is contained between the free and attached border (Pl. 1, fig.  3). As the toothplate descends through the lumen towards the foramen, the free border nears the attached border, mainly because the plate as a whole narrows. The precise mode of attachment is quite diverse, sufficiently so to warrant separate treatment. The description just given is the basic architecture of the toothplate encountered in the buliminids. As far as the majority of buliminid taxa are concerned, differences in shape are due to the varying proportions of the elements making up the structure. For example, the toothplate in Uvigerina is at its top clearly curved, as can be ascertained by looking into the aperture, but while it descends through the apertural neck, the free border stretches out to stand almost perpendicular and straight on the lateral chamber wall just before it fuses with the foraminal phialine lip, while in most bolivinitids the free border flares out into the lumen (Revets, in press).
Those taxa with at least some chambers in rectilinear series show a different toothplate shape, which can nevertheless be traced back to the basic model. The toothplate has become entirely free to the point where it is bilaterally symmetric and is present in the centre of the chamber. The plate is 'buckled' somewhat above the foramen, where it flexes from concave to convex just before it fuses with the foramen opposite the site from where it originates under the aperture (Pl. 1, figs. 7-9 ). There are taxa intermediate between the basic model and the just described one, so there can be little doubt that the structure is a real toothplate. Note incidently that this is basically the same architecture as the one encountered in Clavulina (ut supra).
However, in a not to be neglected number of buliminid taxa, the toothplate has become a highly convoluted structure. From the relatively intricate toothplate in Fijiella and Mimosina, the pinnacle of complexity has been reached in Sagrina, where the plate, contorted in various directions, even bears long, fine spines (Pl. 2, figs. 2, 3 ). These morphologies d o not support the proposal of a single possible function for the toothplate at all, but rather intimate that a uniform explanation might well be impossible.
In low trochospiral taxa (the rotaliids) the toothplate, as recognised by Hottinger et al. (1991) and thereby echoing Hofker (opvar.) and Reiss (op. var.), rises up from the septa1 flap, attached with one border to the umbilical chamber wall and butts into the inner-most corner of the aperture, where it protrudes its free, serrated edge. While traversing the lumen it is folded along two axes: parallel and normal to the coiling axis. Thus it partially delimits an intratoothplate space, presumably the equivalent of the sulcus in the buliminid toothplate. Because successive toothplates attach to each other in an adaxial position, a continuous spiral canal is formed which in the taxa investigated remains open to the adjacent chambers.

The attachment site
Regardless of whether the toothplate runs up or down the chamber lumen, it has to make contact with either the inside of the apertural wall or the outside of the foraminal wall. Because the toothplate as a structure belongs to a chamber, i.e. it is constructed within a chamber and thus occurs in every single one (contrary to a coverplate), Revets (1989) claimed that the toothplate descends into the lumen towards the foramen. It is thus a structure belonging to the apertural complex, an essentially internal feature underhanging the aperture. The precise mode of attachment in the foraminal region is quite variable, and obviously a function of the complexity of the foraminal morphology. The toothplate may fuse with the upper edge of the toothplate protruding in the foramen without any further developments. It may however also spill over onto the lateral part of the foramen, constructing something akin to a bridge between foraminal toothplate and foraminal border. Depending on the taxon, the emphasis of attachment may shift from foraminal toothplate to lateral foraminal wall attachment (Fig. 11). In case of the latter, the attachment always seems to be on the axial side of the foramen. Another kind of attachment is where the toothplate fuses with the foraminal lip. In some taxa, the toothplate is reported to attach to the septum without contacting the foramen at all; the cassidulinids ( Fig. 12) being the best example (Nomura, 1983a & I b). Fig. 1-3. Neorotalia calcar (d'orbigny, 1826), Gulf of Elat. 1. Overview of an partially sectioned specimen, polished and etched. Note that the final chamber is intact (200 pm). 2. Oblique close-up of the last chambers. Note the open umbilical-most part in the final chamber, and the closed equivalent in the previous chamber (100 pm). 3. Detailed view of the umbilical-most part of the final and part of the penultimate chamber. The foraminal plate (fp) is present in the final chamber and continues into the previous chamber to form there the coverplate (cp) by fusing with the previous foraminal plate and thus sealing off the umbilical camera1 opening. The canal thus formed receives extra lamellae, as it is topologically outside (s). Note also the changes in lamellarity due to this deposition (50 pm) a: aperture, f: foramen, 1 lip, fp: foraminal plate, cp: coverplate, i: inner lining, 0: outer lamella, s: secondary lamination. In the Buliminacea the situation is simple to the extent that no other internal structures are present: there can be no confusion between toothplate and, say coverplate, foraminal plate &c. It is therefore pointless to use a different term to denote the toothplate: 'central pillar', 'tongue', or 'sipho' are superfluous and serve only to obscure its true nature from the student.

Explanation of Plate 3
Unfortunately, the situation is not as simple in the Rotaliacea. Hottinger et al. (1991) equate the toothplate in Pararotalia and Neorotalia with the foraminal plate and coverplate as understood by Hansen & Reiss (1971). But the observation of a foraminal plate and coverplate in the earlier stages of Calcarina spengleri by Hansen & Reiss (1971) is rejected by Hottinger et al. (1991): the genus is described as lacking toothplates, coverplates and foraminal plates altogether. The latter authors do distinguish the toothplate from coverplates in Ammonia while Rotalia and Cuvillierina are deemed to possess umbilical plates but no toothplates. They consider these genera do not possess a spiral umbilical canal (but see Haynes & Whittaker, 1990), while this canal has been shown in Pararotalia and Neorotalia. It seems thus that the toothplate in low trochospiral forms is linked to the presence of a canal. The question this raises is: d o toothplates occur in Challengerella, Asterorotalia, Cavarotalia and Pseudorotalia? According to Miiller-Merz (1980), they do not. The internal structures delimiting the canals are the perfect homologues of the foraminal-and coverplate of Ammonia, except that the sealing off by the coverplate is incomplete, allowing access from the chamber to the partly sealed-off umbilical end of the chamber, those lumina forming the canal.

Limitations of morphology
Except for the contributions by Hansen & Reiss (1971) and Revets (1989), all discussions and changes have been based on morphological interpretations only. Most authors have attempted to come to grips with the different anatomical elements by studying their relative positions and occurrences within the tests. This necessarily results in a restricted view of the structures because a good deal of knowledge has to be inferred. By observing only shape and superficial relations inside chambers, it is not possible to decide how a structure is built, i.e. is it drawn up from the septum, does it descends into the lumen, is it glued against recognising them for what they are, i.e. inner lining, outer lamella or secondary outer lamellae, the mutual relations between the internal structures and to the remainder of the test can be resolved. As a result, some conclusions can also be drawn as to the role or function of these features (i.e. functional surface of Hansen shown by the deposition of secondary lamination, thus proving the feature to remain in contact with the outside world, topologically speaking; pure inner lining, indicating the strictly internal nature of the structure). A good example is the recognition of a septal flap. According to the precise definition (Hansen & Lykke Andersen, 1976), a septal flap is formed when the inner lining of the chamber wall continues inside the chamber to cover its septum, i.e. the wall of the previous chamber forming the floor of the current chamber. It is thus defined as that part of the septum that exhibits an ultrastructural pattern of inner lining, outer lamella, inner lining. Obviously, the chamberwall a t the back of a chamber forming the limit of an interlocular space cannot be denoted as a septal flap, since it is bilamellar, or, later in ontogeny, will bear more outer lamellae due to secondary lamination being deposited. These distinctions can only be made properly through the preparation of adequately positioned etched sections.
More importantly, i t allows a much more precise dissecting of the apertural complex, and a proper assessment of the various anatomical features recognised by different authors over the years. Etched sections should reveal, for example, what the precise nature of the umbilical flap (Miiller-Merz, 1980) or umbilical plate (Hottinger et al., 1991) is with respect to the other chamber elements. It is virtually certain that the umbilical flap is bilamellar and perfectly continuous with the remainder of the chamber wall. If this is indeed the case, then there is every reason to demote the term in the c;ense of a discrete structural element, since it would denote nothing more but the chamberwall at the umbilical-most part of the chamber. The term would be misleading since it purports to define something which is in fact nothing but a direction or position, not a distinct anatomical feature.

The apertural complex
In the Buliminacea, etched sections show that the toothplate is made from inner lining (PI. 1,figs. 4,5 ). The Buliminacea d o not possess a septa1 flap, hence the toothplate is necessarily drawn from the wall underlying the aperture and progresses through the chamber lumen towards the foramen. The apertural lip is no more than a continuation of the chamber wall and thus bilamellar. Consequently, the foraminal region is trilamellar at the attachment site of the toothplate. It is thus also very likely that the toothplate is constructed contemporaneously with the remainder of the chamber. The toothplate is attached to the lateral wall deeper down in the chamber by 'gluing' the free border against the wall, but it does not extend into the previous chamber at all. Inside earlier chambers the lamellarity remains as it was. This observation proves that a closed tube is indeed absent, as the inside of such a tube would be topologically outside world, and should thus receive secondary lamination. Preliminary observations on Ammonia show that the foraminal plate and the coverplate are originally bilamellar (Hansen & Reiss, 1971, pl. 3, fig. 4) composed of inner lining covered by outer lamella and that the outer lamella receives secondary lamination for as long as it remains accessible from the umbilicus (ibid., p. 333, pl. 4, fig. 3-4). This seemed to be the case in Pseudorotalia (ibid., pl. 7, figs. 2, 3, 5 ) as well as in Neorotalia (ibid., pl. 9, figs. 3-5. See also fig. 10).
New etched sections vindicate this interpretation. In Neorotalia calcar, the final chamber is partially open on the umbilical side, allowing free access to the chamber lumen. However, in the penultimate chamber, this access is closed off by a true bilamellar coverplate with at its beginning a foraminal plate which is rather prominent. As the coverplate butts into the foraminal plate of the prepenultimate chamber, it covers its outside by a secondary lamella, so that this foraminal plate cum coverplate is trilamellar. Thus, purely structurally, Neorotalia calcar is the equivalent of Ammonia. The only structural difference is the presence of a canal in Neorotalia cnlcar, absent in Ammonia. Interestingly, the canal in Neorotalia receives secondary lamination in its inside (PI. 3, fig. 3), thus supporting the hypothesised scenario of Revets (1989) as to the structural changes during ontogeny in tubes or canals. These observations contradict the analysis put forward by Hottinger e t a l . (1991) and confirm the interpretations of Hansen & Reiss (1971).
The internal structures in rotaliids are not equivalent to toothplates, rather they all seem to conform to the foramina1 plate coverplate concept. Clearly, a major divide exists between the buliminid toothplate and what is referred to as toothplate in the rotaliids.

A definition
As the concept of the toothplate was originally created for the internal structure encountered in buliminids and its description consequently limited to what one encounters in these taxa it seems to be nothing but good practice to adhere to the original intentions, but further clarified by the newly obtained information. The case to d o SO is particularly strong since more than sufficient grounds have been found to prove that subsequent extensions of the concept are incorrect and thus unwarranted.
The toothplate is a wholly internal structure without direct contact to the outside world. The upper part of the toothplate is capped by the apertural lip, effectively shielding it from the outside. Interestingly, this recalls the situation encountered in Clavulina, where the upper part of the "toothplate" is clearly agglutinated, but the internal part is wholly calcitic. Structurally, it is composed of a single piece of inner lining and remains so troughout ontogeny. The attempt to reinstate the ideas of Hofker by Verhallen (OP. c i t . ) cannot be supported. Careful reading of the proposal furthermore reveals that the ideas put forward are hypothetical, and not based on actual observation of the closed tube. It is basically a curved plate, trough-like, which runs from aperture to foramen. In principle it has one so-called attached border, which is the lateral contact with the chamber wall. In forms where chambers become arranged in rectilinear series, this attached border becomes free as well. The toothplate attaches to the foramen itself, depending on the taxon, with the foraminal lip, foraminal toothplate top or both, without progressing into the foramen or spreading over the septum. It remains an open trough, except perhaps at the bottom, where it ends blindly where the curving free border reaches the lateral chamber wall.

A reappraisal of previous points of view
Restricting the definition of the toothplate ( u t s u p r a ) effectively enforces the reappraisal of the structures that have previously been called 'toothplate' but do not conform any longer. Lamellarity patterns show that there are essentially no differences between the rotaliid 'toothplates' and the foraminal plates and coverplates as defined by Hansen & Reiss (1971). As mentioned above, in many of the forms with 'toothplates' these structures are connected to the presence of a spiral-umbilical canal. It is of more than passing interest to pause at the illustration of a rudimentary spiral canal in Ammonia provided by Billmann et al. (1980, Revets pl. 2, fig. 4), especially since the internal structures of Ammonia continue to be regarded explicitly as foraminal a n d coverplates. As shown, the similarities between A m m o n i a a n d neorotalian structures are such that n o reasons remain to denote these structures by different names. The only possible way forward is to apply the terms foraminal plates and coverplates to the internal structures in rotaliids. The term toothplate should be allowed to lapse in the rotalian context and be reserved for exclusive use in the buliminids.

Outstanding problems
Despite the definitions given and the distinctions drawn, a large number of difficult and potentially serious problems remain, both within the buliminids as within the rotaliids.
Because of the form of the toothplate, especially the formation of a dead-end, the function or role remains as enigmatic a s e v e r . The restrictive definition of the toothplate a n d its consequent limitation to buliminids effectively removes as well the temptation to link its presence to canals &c.
In a variety of taxa, the nature of the toothplate is either unclear or else highly unusual. Paradoxically, in view of the extensive early attention devoted to them, the toothplate structure in the Pleurostomellidae is not clear, especially its relation to the foramen. Nevertheless, preliminary results do show that the plates are monolamellar inner lining, and hence fall within the definition of the toothplate. The internal structures in Tubulogenerina are highly variable in shape and the relation to classical buliminid structures is at present utterly unclear. From the excellent illustrations given by Gibson (1987Gibson ( , especially pl. 4, figs. 3-5, pl. 5, figs. 2, 5, 6: 1989 and Gibson & al. (1991) some indications may be gleaned for a different interpretation altogether. The often complex and multiple structures appear to be bilamellar and it seems that a septa1 flap is present. If these potentially controversial interpretations w e r e to be confirmed t h r o u g h closer scrutiny, Tubulogerierina may well have to be moved to the edges of the buliminid taxa. The observation of t h e internal structures in Stainforthia (PI. 2, and Francesita show these taxa to be closer to cassidulinids, structurally speaking (Revets, in press).
The internal organisation of Siphouvigeriria (PI. 2, figs. 7-9) is highly unusual and presents serious problems especially because of the similarities with Stilostomella, another very peculiar taxon.

CONCLUSIONS
Progress in the debate surrounding the toothplate has only been possible through the use of a variety of different techniques, allowing the gathering of as much primary information as possible. The supplementing of purely morphological observations by lamellarity proved to have made a crucial impact in the elucidation of the nature of the structures supposed to be toothplates. By involving lamellarity it has become possible to dissect the various plates and reveal their true nature. This resulted in a n unambiguous definition of the toothplate, considerably restricting its taxal distribution. Unfortunately, the results obtained effectively refute all currently existing hypotheses as to the function of the toothplate.