Application of neontological taxonomic concepts to Late Eocene coralline algae (Rhodophyta) of the Austrian Molasse Zone

Traditionally, different diagnostic characters have been used in the identification of fossil and Recent coralline algal genera. The taxonomy of fossil coralline algae has focused on well calcified features such as basal filaments and conceptacle perforation. In contrast, the taxonomy of Recent material uses a combination of several features with a low fossilization potential, such as epithallial cells and structures of sexual reproductive organs. In the studied material of the Late Eocene Austrian Molasse Zone Lithoporella, Neogoniolithon, Spongites, Phymatolithon and Sporolithon are identified and described applying features of neontological taxonomic concepts. These features are: (1) the arrangement of basal filaments; (2) the occurrence of cell fusions; (3) the relative length of subepithellial initials; (4) the conceptacle perforation; (5) the orientation of filaments around the conceptacle pore; and (6) the type of conceptacle roof formation. Some of these features were thought to be unpresentable in fossil material until recently. The fossilization potential of diagnostic features and the identification of the documented genera and species are discussed in detail. Moreover, a checklist for the description of fossil taxa is provided.


INTRODUCTION
Until recently, several diagnostic characters used in present day coralline red algal taxonomy were thought to be unpreservable in fossil material. Wray (1977) and Poignant (1984) therefore concluded that fossil and Recent coralline algae have to be classified in different ways. The taxonomy of fossil coralline algae has usually focused on calcified characters with a high fossilization potential such as the arrangement of basal filaments, the perforation of asexual conceptacles, and the occurrence of trichocytes (Wray, 1977). The taxonomy of Recent coralline algae, however, uses additional characters including cell connections, the shape of epithallial cells, the length of subepithallial initials, and the formation of sexual reproductive organs (see Figure 2) (Woelkerling, 1988;Braga et al., 1993). Consequently, some genera and most species described from fossil material cannot be compared to any Recent taxon and, conversely, some Recent genera are not recognized in the fossil record. The unification of taxonomy of Recent and fossil corallines is crucial to the understanding of their phylogeny, palaeoecology and palaeobiogeography. Braga et al. (1993) demonstrated, using the genus Spongites Kiitzing (1841) as an example, that key features of the taxonomy of present day corallines such as cell connections, epithallial cells and subepithallial initials are indeed preservable and can occasionally be recognized in fossil coralline algae. Since then, several studies have dealt with the identification of fossil taxa using diagnostic characters used in present day taxonomy (Bassi, 1995a, 199513;Braga & Aguirre, 1995;Aguirre et al., 1996;Basso et a/., 1996Basso et a/., , 1997. Moreover, Braga et al. (1993) presented an identification key for fossil coralline algae that utilizes both traditional and neontological features. This key demonstrated that traditional characters still have to be used in certain cases in the identification of several fossil genera. This is due to the fact that several key features used in Recent taxonomy, such as the formation of sexual reproductive organs, have not yet been observed in fossil material.
This study aims to evaluate the identification key of Braga et al. (1993)  taxonomy in the identification of fossil material. Using five genera it tries to bring us one step further in unifying fossil and present day coralline algal taxonomy. Moreover, it intends to provide a modern documentation of coralline algae of the study area to serve as a base for further actualistic approaches. We provide a checklist for the description of coralline algal species with all features known from fossil material (Table 2) and discuss the preservation potential of taxonomic features.

MATERIAL AND METHODS
The studied material comes from the Priabonian (Late Eocene) red algal limestones ('Lithothamnienkalk') of the Austrian Molasse Zone (Fig. I) (Aberer, 1958;Malzer, 1981;Wagner et al., 1986;Wagner, 1996). These red algal limestones are up to 80 m thick and were deposited on a mixed carbonate-siliciclastic ramp. They are known from deep wells of the Rohol AG, Vienna (Austria) only (Wagner, 1980(Wagner, , 1996. A detailed study of the palaeoecology and facies of the red algal limestones is in preparation by the authors.
Two hundred palaeontological thin sections and several scanning electron microscope samples from 10 deep wells (

SYSTEMATIC PALAEONTOLOGY Except for Lithoporella melobesioides (Foslie) Foslie (1 909) and
Phymatolithon sp., the described species could not be referred to any described Recent or fossil taxa. The designation of fossil species suffers from the problem that the vast majority of species are poorly described and type specimens are partially missing. Although there have been several recent efforts to revise and redescribe original material (e.g. Moussavian & Kuss, 1990;Piller, 1994;Rasser & Piller, 1994;Braga & Aguirre, 1995;Basso et al., 1997), most fossil species cannot be identified with confidence.
Only uniporate conceptacles without preserved reproductive organs were found in the studied material. It is not known if these are tetra/bisporangial (asexual) or gametangial (sexual) conceptacles. If they are gametangial conceptacles, which are always uniporate, the described genus may have multiporate tetra/bisporangial conceptacles which are not preserved in the studied material. Owing to the regular coaxial thallus, the described genus in this case would belong to Mesophyllum (although the coaxial core is no longer diagnostic for Mesophyllum Lemoine (1928), we apply the same arguments as for Neogoniolithon). If, however, the identified uniporate conceptacles are gametangial conceptacles of Mesophyllum, they would have to be formed from filaments which enclose the chamber . Owing to the filament structure in the walls of the identified conceptacles, this kind of formation can be excluded. We therefore refer this species to Neogoniolithon. Besides Phymatolithon sp., Neogoniolithon sp. is one of the most abundant species in the studied material. It forms coralline algal bindstones in association with the latter and occurs fragmented in most samples. Measured sample: MOL110.
Genus Spongites Kiitzing, 1841 Diagnosis. Non-endophytic thallus which lacks haustoria; thallus organization dimerous or monomerous; dimerous thallus portions lacking palisade cells; filaments around the conceptacle pore canals subparallel to the roof surface (Penrose & Woelkerling, 1992). Braga et a/. (1993) additionally mentioned the presence of non-coaxial core filaments to separate Spongites from Neogoniolithon. Remarks. Hydrolithon (Foslie) Foslie (1 909) was considered to be congeneric with Spongites by Woelkerling (1988). Penrose & Woelkerling (1992), however, showed that both genera can be separated by the filament arrangement in the conceptacle roof. As it was shown by Braga et a/. (1993) and by the current study, this character can be applied to fossil material as well. The separation from Neogoniolithon is discussed above. Peripheral filaments: some parts of the thallus show growth rhythms with a thickness of 7-10 cell rows; cell length 6-1 1 pm (M = 10, SD = 2), diameter 7-9pm (M = 8, SD = 1). Some cells of contiguous filaments are joined by cell fusions. Subepithallial initials and epithallial cells not preserved.
Tetra/bisporangial conceptacles usually raised one half above the thallus surface (Plate 2, fig. 3). Conceptacle height 170-200 pm, diameter 350-600 pm; roof thickness 55-90 pm. Pore diameter 11-13 pm. Length of cells in conceptacle roofs 7-9 pm (M = 8, SD = l), diameter 5-10pm (M = 8, SD = 2). Type of conceptacle roof formation unknown, sexual conceptacles and carposporangia unknown. Remarks. In accordance with the traditional generic concepts of Wray (l977), this species would belong to Lithothamnion. However, this genus cannot be identified using present day concepts owing to the unpreserved epithallial cells and subepithallial initials. Owing to the non-coaxial core filaments, it can be referred to either Lithothamnion or Phymatolithon (see also remarks of Phymatolithon).

Genus Phymatolithon Foslie, 1898
Diagnosis. Plants lacking an arborescent growth form and haustoria. Thallus monomerous, core filaments non-coaxial, epithallial cells rounded or flattened, but not flared, subepithallial initials as short or shorter than underlying cells (Braga et al., 1993;Wilks & Woelkerling, 1994). Irvine & Chamberlain (1994) additionally define Phymatolithon by the 'Phymatolithon-type' surface view of epithallial cells. Remarks. Following Braga et al. (1993) this genus is indistinguishable from Leptophytum Adey (1966) in fossil material. This fact is reflected by the generic differentiation of Chamberlain & Keats (1 994), who focus on growth form, surface view of epithallial cells, and whether the conceptacle initiation is 'shallow' or 'deep'. However, the cited authors also mention that the differentiation between Lepthophytum and Phymatolithon is provisional. As Wilks & Woelkerling (1994) concluded that Leptophytum is not a distinct genus in the current state of research, it is not considered here. The occurrence of rounded epithallial cells and short subepithallial initials allows a distinct separation of Phymatolithon from other genera of this subfamily.  fig. 4) with a thallus thickness of usually 150pm. Sometimes warty growth forms occur with 1.3-1.4 mm long and 1.5-0.8 mm thick protuberances. Thallus monomerous.
Tetra/bisporangial conceptacles are multiporate, without a rim; old conceptacles may be buried within the thallus. Conceptacles distinctively raised above the thallus surface with a floor usually ten cell layers below the thallus surface (less frequently only five) (Plate 2, figs 5, 6). Height 100-160pm, diameter 2 1 M 6 0 pm. Thickness of roof 45-70 pm, conceptacle pore diameter up to 27 pm. Length of cells in conceptacle roof 6-12pm (M = 9, SD = 3), diameter 5-8pm (M = 7, SD = 1). Conceptacle roof formed by filaments interspersed between sporangia (Plate 2, fig. 7). Sexual conceptacles and carposporangia unknown. Remarks. As traditional concepts for fossil corallines do not take into account the occurrence of epithallial cells and subepithallial initials, this genus would, in the past, have been identified as Lithothamnion. This species is the same as described by Rasser (1994) as Lithothamnion sp. Growth form, anatomy, and conceptacle size are close to those of Lithothamnion crispithallus Johnson (1957), which, following this work, should now belong to Phymatolithon. A new combination would, however, require a study of the original material and this has not been included in the present study. Phymatolithon sp. is the most abundant species in the studied material, forming coralline algal bindstones together with Neogoniolithon sp. Moreover, it is the dominant coral encruster. Measured samples: MOLl2 and MOL29 1.
Family Sporolithaceae Verheij, 1993 cell fusions and secondary pit connections occur; tetra/bisporangia formed between filaments, on one or more stalk cells, apical plug at tetra/bisporangial apex (Verheij, 1993). Remarks. Verheij (1993) separated the family Sporolithaceae from the family Corallinaceae. Because of the preservation of calcified sori and paraphyses this family can easily be identified in fossil material.
Peripheral portion 85-200 pm thick, mostly 100 pm. Cell shape rectangular; no growth rhythms and trichocytes occur; cell layers irregular; cell fusions occur. Cell length 25-30 pm (M = 28, SD = 2), diameter 10-16pm (M = 13, SD = 2). Some thalli show preserved epithallial cells which are characterized by brightish cell layers on the outermost thallus surfaces in thin section (Plate 3, fig. 2). Scanning electron microscope samples show that the epithallium is recrystallized and cells are replaced by large calcite crystals (Plate 3, figs 5, 6). Cell length approx. 7 pm, diameter approx. 13 pm. Owing to the poor state of preservation, the shape of epithallial cells cannot be identified.
Epithallium not preserved. Tetra/bisporangia arranged in sori with up to 35 tetra/ bisporangia each. Sori completely raised above the thallus surface. Sori do not arise from a layer of elongated cells. Old sori are not flaked off, but buried in the thallus. Conceptacle height 63-70pm, diameter 3-5 pm. One to nine filaments interspersed between the tetra/bisporangia. Number of cells in the paraphyses unclear; they seem to vary between three and five. Sexual conceptacles and carposporangia unknown. Remarks. This species cannot be compared with any previous species of Sporolithon. Only a single thallus of Sporolithon sp. 2 was found in the studied material (MOL220).

Primigenous and core filaments
Except for the occurrence of palisade cells which helps to identify Lithoporella (Plate I, fig. I), the importance of the basal filament organization for the generic identification has decreased during the last decade. This is mainly true for the occurrence of coaxial and non-coaxial core filaments in monomerous thalli. The decreasing importance mainly affects the definitions of Neogoniolithon and Mesophyllum, which have traditionally been characterized by a coaxial core (e.g., Woelkerling, 1988). Some workers do not accept the validity of this feature for the identification of both Neogoniolithon (see Penrose, 1992) and Mesophyllum (see , although the type species of both genera show distinct coaxial cores (Woelkerling, 1988). This character is therefore still used in Recent (e.g., Irvine & Chamberlain, 1993) and fossil (Braga et al., 1993;Bassi, 1995a) taxonomy (Table 1).

Trichocytes
Trichocyte (Fig. 2D) occurrence and arrangement have long been used to delimit genera within the Mastophoroideae, predominantly within the Spongites complex (Woelkerling, 1985). Chamberlain (1983) and Jones & Woelkerling (1984), however, showed that trichocyte occurrence varies within species and is influenced by environmental conditions. Trichocyte occurrences in the peripheral filaments are still used by Braga et al. (1993) to define Neogoniolithon and Spongites.

Subepithallial initials
The length of subepithallial initials with respect to the length of peripheral or postigenous cells subtending them is one of the most important characters used to separate genera within the Melobesioideae (Wilks & Woelkerling, 1994 (Table I). As they are usually calcified, this feature can also be applied to fossil material (Braga et al., 1993). The recognition of subepithallial initials, however, depends on the preservation of the overlying epithallial cells (Fig. 3).

Epithallial cells
The shape of the epithallial cells is an important feature used to identify Lithothamnion and Sporolithon (Woelkerling, 1988). Both of these taxa are characterized by flared, but not rounded, cell walls (Table 1). According to Braga et al. (1993)   this species are, however, recrystallized and replaced by large calcite crystals (Plate 3, figs 5, 6). Therefore, the shape of the cells cannot be recognized. Epithallial cells in Phymatolithon sp. are slightly better preserved. In our material they can easily be identified and separated from subepithallial initials by large crystals forming the cell walls in the scanning electron micrograph (Plate 2, fig. 8; Plate 3, figs 5, 6; Fig. 3) and by their transparency in thin sections (Plate 3, fig. 2). Owing to the poor preservation, the shape of the cells cannot be recognized with confidence. However, unlike the epithallial cells described and figured by Braga et al. (1993) and Aguirre et al. (1996), they do not appear to be flat (Fig. 3). In the current study, epithallial cells helped to identify the underlying vegetative initials and thus indirectly allowed the designation of the genus Phymatolithon.

Conceptacles
Conceptacle perforation is a well known traditional feature used for the identification of both Recent and fossil genera (Wray, 1977) and subfamilies (Woelkerling, 1988;Bosence, 1990;Braga et al., 1993). The orientation of conceptacle roof filaments around conceptacle pores is an important character in separating genera within the Mastophoroideae in present day taxonomy (Penrose & Woelkerling, 1992). Both Neogoniolithon and Spongites are characterized by filaments which are arranged subparallel to the roof surface (Braga et al., 1993;Penrose & Woelkerling, 1992) (Table 1). This study proves that this feature can be well preserved in fossil material and enabled us to present the oldest record of Neogoniolithon.
Several fossil tetra/bisporangia have been described from multiporate conceptacles (e.g., Conti, 1947;Johnson, 1957; Mastrorilli, 1973;Lemoine, 1977;Bosence, 1983). Bosence (1 983) presents both sexual and asexual conceptacles preserved in one thallus of Lithophyllum. Preserved gametangia and tetra/ bisporangia borne in uniporate conceptacles are, however, unknown. This fact forced palaeontologists to interpret uniporate conceptacles as tetra/bisporangial only on the basis of the lack of multiporate conceptacles in the same thallus. In the current study we could indirectly exclude the gametangial nature of uniporate conceptacles in Neogoniolithon sp. Our designation of Spongites is, however, based on the lack of multiporate conceptacles. Consequently, we cannot exclude the possibility that both species we referred to Spongites have multiporate conceptacles which were not found owing to the low abundance of specimens in the studied material.

a a
The kind of conceptacle roof formation is another important taxonomic feature. Roofs of multiporate conceptacles are formed by elongations of filaments interspersed between sporangia. After the release of sporangia these filaments in the conceptacle chambers are secondarily decalcified (Woelkerling, 1988) and thus usually not preserved in the fossil record.
Interspersed filaments in conceptacles of Phymatolithon sp. are preserved in the studied material (Plate 2, fig. 7) as the tetra/ bisporangia are not released and the conceptacle is buried within the thallus. The current study and the recognition of conceptacle primordia in fossil material (Aguirre et al., 1996) suggest that even more reproductive characters are potentially preservable. Bosence (1983)

CONCEPTACLES
Further taxonomic studies of fossil coralline algae will have to focus on this topic.

Species identification
Owing to poor descriptions and different valuations of diagnostic criteria, the status of most fossil species is unclear and only a few Recent species have been traced back to the fossil (e.g. Braga & Aguirre, 1995;Basso et al., 1996Basso et al., , 1997. The main problem is that traditional identifications of fossil species are usually restricted to characters such as cell and conceptacle dimensions (e.g. Lemoine, 1939;Conti, 1947;Mastrorilli, 1967;Bucur & Filipescu, 1994). However, these features alone are not used the identification of Recent species in fossil material. Table  2 summarizes all the diagnostic features which are known from fossil algae. The application of this checklist for the description of fossil taxa is a minor requirement in the recognition of Recent taxa in fossil material. The identification of Recent species focuses on the combinations of several characters. Chamberlain (1994) separated species of Spongites by cell dimensions, occurrence of trichocytes, dimensions of filaments around conceptacle pores, growth form and colour. Species of Phymatolithon are separated by growth form, the amount to which tetra/bisporangial conceptacles are raised above the thallus surface, the occurrence of rimmed conceptacles (Chamberlain, 1994), whether old conceptacles are flaked off or buried in the thallus, colour (Irvine & Chamberlain, 1994), position of core filaments, shape of cells interspersed between tetra/bisporangial conceptacles, the occurrence of vegetative cells beneath the floor of tetra/bisporangial conceptacles and the thickness of conceptacle roofs (Wilks & Woelkerling, 1994). The identification of species in Sporolithon focuses on the number of cells that sori are raised above the thallus surface, the number of cells in paraphyses, the occurrence of a basal layer of elongated cells below tetra/bisporangia, the dimensions of tetra/bisporangia and whether old conceptacles are flaked off or not (Verheij, 1993;Townsend et al., 1995).
We can show that all calcified features used to describe present day species are observable in well preserved fossil material. Growth forms can be recognized in two-dimensional sections (e.g., Plate 2, fig. 4) and the amount to which conceptacles are raised above the thallus surface can even be recognized if conceptacles are buried within the thallus (e.g. Plate 2, fig. 3). In some cases usually uncalcified features such as the filaments interspersed between sporangia are preserved (Plate 2, fig. 5). Reproductive features which are diagnostic for Sporolithon can easily be observed in scanning electron microscopy samples (Plate 3, fig. 3). Only the occurrence of vegetative cells beneath the floor of conceptacles has not yet been observed and the colour of thalli is obviously not applicable.
The present study proves that most features used in the present day taxonomy of coralline algae can be applied to fossil taxa. The identification of certain fossil genera, however, still has to take into account traditional features which are no longer accepted in Recent taxonomy. This is because several phycologists tend to focus on taxonomic characters which are unknown in the fossil record. Our study also shows that the identification key of Braga et al. (1993) is a useful tool for the identification of fossil corallines which, however, has to be updated according to the latest published studies on present day taxonomy.
The most important features used to identify Recent species are easily observable in fossil coralline algae. Nevertheless, they are rarely applied to the identification of fossil taxa, even in modern studies. Therefore, we provide a checklist (Table 2) including all known features preserved in fossil material. Further documentations of fossil species will have to focus on these characters to trace back Recent species to the fossil record. This is crucial for the understanding of coralline algal phylogeny, palaeoecology and palaeobiogeography.