- The Paleontological Society
A large, morphologically heterogeneous population of acanthomorphic acritarchs from the early Neoproterozoic Wynniatt Formation, Victoria Island, northwestern Canada, is ascribed to two form-genera, Tappania and Germinosphaera, but just a single natural taxon, Tappania. Analysis of Tappania morphology shows it to have been an actively growing, benthic, multicellular organism capable of substantial differentiation. Most notably, its septate, branching, filamentous processes were capable of secondary fusion, a synapomorphy of the “higher fungi.” Combined with phylogenetic, taphonomic and functional morphologic evidence, such “hyphal fusion” identifies Tappania reliably, if not conclusively, as a fungus, probably a sister group to the “higher fungi,” but more derived than the zygomycetes.
The presence of Tappania in the Mesoproterozoic Roper Group of Australia extends the record of putative fungi to 1430 Ma. Along with other Proterozoic acritarchs exhibiting fungus-like characteristics (e.g., Trachyhystrichosphaera, Shuiyousphaeridium, Dictyosphaera, Foliomorpha), there is a case to be made for an extended and relatively diverse record of Proterozoic fungi.
The key to understanding the early evolution of eukaryotes lies in the Proterozoic fossil record, especially microfossils. Some of these early fossils are demonstrably cyanobacteria, and a handful of others can be recognized as members of extant protistan groups. The vast majority, however, fall into the category of “acritarch,” which is to say, “small microfossils of unknown and probably varied biological affinities consisting of a central cavity enclosed by a wall of single or multiple layers and of chiefly organic composition” (Evitt 1963: p. 300). As an informally established group, this definition is certainly not binding, and it has transmogrified over time such that most acritarchs are now widely understood to be “the vegetative and reproductive walls of unicellular protists” (Knoll 1994), predominantly phytoplankton (Vidal and Moczydłowska-Vidal 1997), and with no absolute requirement for organic composition (e.g., Xiao and Knoll 1999). This may well be a useful generalization for the dinocyst-like acritarchs that dominate the Paleozoic record (e.g., Talyzina et al. 2000), but it is premature to assume that it holds in the Proterozoic. A significant number of Proterozoic acritarchs, for example, are demonstrably benthic (Butterfield 1997, 2001), and I have recently shown a diverse Neoproterozoic assemblage to be the disarticulated remains of a single vaucheriacean metaphyte (Butterfield 2004). At a deeper level, Cavalier-Smith (2001) has even challenged the assertion that most Proterozoic acritarchs are eukaryotic (but see Javaux et al. 2003 and Butterfield 2004).
Most acanthomorphic, and many spheromorphic, acritarchs of Proterozoic age are unusually large relative to their Paleozoic counterparts (Knoll and Butterfield 1989; Mendelson and Schopf 1992). Combined with a widespread absence of excystment structures and evidence for adventitious vegetative growth (e.g., Germinosphaera, Trachyhystrichosphaera, Tawuia, Tappania; [Butterfield et al. 1994; Kumar 2001; Javaux et al. 2001]), such a habit points to a multinucleate, even multicellular, grade of organization. In the terminal Proterozoic Doushantuo biota, for example, both Cerionopora and “Unnamed form A” (Zhang et al. 1998) are represented by large spheroids filled with multiple 15–50 μm cell-like structures, possibly spores within a sporangium or, alternatively, the constituent cells of a metazoan embryo (cf. Megaclonophycus [Xiao and Knoll 2000: Fig. 9.7–9.12]). Co-occurring Megahystrichosphaeridium combines marked morphological variability and near-macroscopic dimensions (150–700 μm diameter) with a tendency to express patterns of surficial polygons (Xiao and Knoll 1999; Zhou et al. 2001), not unreasonably interpreted as the boundaries between constituent cells (cf. Eovolvox [Kázmierczak 1975: Pls. 18.1, 20.5]); likewise Tianzhushania/Megasphaera (Yin et al. 2004) and the single, 1000 μm diameter specimen of Zhang et al.'s (1998) “Unnamed form B.” Differential preservation of the outer envelopes of spheroidal Wengania, a putative seaweed, offers a further multicellular source for Doushantuo acritarchs (e.g., Zhang et al. 1998: Figs. 3.3, 3.7, 14, 15).
Nor are Proterozoic acritarchs necessarily the remains of photoautotrophs. Some have proven to be heterotrophic amoebozoans/cercozoans (Porter and Knoll 2000; Porter et al. 2003; Pawlowski et al. 2003) and others metazoans (Xiao and Knoll 2000). Moreover, the established presence of Ediacaran-age metazoans demands the co-occurrence of their heterotrophic sister groups, the choanoflagellates, mesomycetozoans (Mendoza et al. 2002), and/or Fungi, all of which are liable to occur as acritarch-like microfossils.
None of this is to suggest that all acritarchs are misinterpreted as unicellular phytoplankton; only that such an assumption cannot help but obscure a much wider fossil disparity, including alternate grades of organization, habit, and metabolism. A return to the original definition of acritarchs as “problematica” offers a new and potentially rich source of paleobiological data, likely to include the stems of various Phanerozoic crown groups. The challenge, of course, lies in extracting biologically meaningful information from these relatively simple fossils. Here I examine a large, exceptionally well preserved population of early Neoproterozoic acritarchs that proves to be neither unicellular nor planktic; intriguingly, it shows a character combination that is synapomorphic of the extant higher fungi.
GEOLOGICAL AND ECOLOGICAL SETTING
The fossils are from the Wynniatt Formation, a 500-m-thick succession of essentially flat-lying platform carbonates and subordinate shales exposed in the Minto Inlier (Shaler Supergroup), Victoria Island, northwestern Canada, (Rainbird et al. 1996; Butterfield and Rainbird 1998). The age of the Wynniatt Formation is constrained radiometrically by the 723 +4/−2 Ma Franklin diabase intrusions, and 1077 ± 4 Ma detrital zircons in the underlying Nelson Head Formation (Rainbird et al. 1996). Stratigraphic relationships and estimates based on a distinctive strontium isotope curve (Asmerom et al. 1991) further constrain the time of deposition to between 800 and 900 Ma, i.e., the early Neoproterozoic.
Wynniatt shales are pervasively fossiliferous; however, the fossils described here come from essentially a single stratigraphic horizon near the top of “Member 2” (Rainbird et al. 1996: Table 1), a rusty-weathering gray to black shale 30–145 meters thick with conspicuous mudcracks at both the base and top. Underlying, overlying, and occasionally interfingering, stromatolitic and oolitic carbonates further attest to the shallow-water setting. Similar facies and fossils are recorded for at least 150 km along strike on Victoria Island (Butterfield and Rainbird 1998), and may continue some 1000 km to the southwest, correlating with the Rusty Shale of the Mackenzie Mountains Supergroup (Rainbird et al. 1996) and/or early Neoproterozoic strata in the southern Franklin Mountains (Samuelsson and Butterfield 2001). The macrofossils Chuaria and Tawuia are conspicuous in both the Wynniatt and Rusty Shale assemblages (see Hofmann and Rainbird 1995).
Four field samples from two field localities have been used in this study: 88-KL-2 is from the northeastern part of the Minto Inlier, near the outlet of the gorge emptying Kilian Lake (72°12′N, 111°43′W)—specifically, a 30-cm-thick shale horizon sandwiched between rippled and cross-laminated dolomite at the gradational contact between Member 2 shales and Member 3 carbonates (Rainbird et al. 1996: table 1). Samples 89-VI-21, 89-VI-22 and 89-VI-23 were collected from a section on Minto Inlet, approximately 150 to the southwest of 88-KL-2 (71°26′N, 116°01′W), from a horizon about 20 m below the Member 2/ Member 3 contact.
All four field samples are finely laminated medium to dark gray shales, and preserve a similar microfossil assemblage. Abundant spheroidal and filamentous structures represent more or less in situ benthic photosynthesizers, probably cyanobacteria (Butterfield and Rainbird 1998), and support sedimentological evidence for a shallow-water depositional environment (Butterfield and Chandler 1992). The samples also share a number of eukaryotic form-taxa, including Leiosphaeridia, Chuaria, Tawuia, Trachyhystrichosphaera, Germinosphaera, Tappania, and a number of figured but as yet undescribed forms (see Butterfield and Rainbird 1998).
Approximately 100 g of each sample (88-KL-2, 89-VI-21, 89-VI-22 and 89-VI-23) was dissolved in 30% HF with minimal agitation, and the resulting slurries passed through a 62-μm mesh sieve. Screenings were recovered in aqueous suspension and organic-walled microfossils picked by pipette under a transmitted-light stereomicroscope (50–100× magnification). Isolated fossils were permanently mounted on glass slides with a heat-setting resin (Petropoxy 154) and photographed with a compound microscope and digital camera. Localized topographic variation within most specimens has been removed by superimposing images taken at different focal depths. Some specimens were prepared for SEM, but this technique yielded no additional information beyond that available from transmitted-light microscopy.
The fossils described here constitute a highly variable, bimodal continuum of forms. Those of the principal mode are based on a central vesicle bearing a variable number of irregularly distributed processes and occasional larger-scale outgrowths. The central vesicle ranges from spheroidal to elongate, and from 30 μm (Fig. 1E) to over 400 μm (Fig. 2D) in transverse dimension (x̄ = 137.4 ± 46.8 μm, n = 50). Processes are typically heteromorphic and range from 0.3 μm (Fig. 1G) to >4 μm (Fig. 1J,K,O) in diameter. In some instances, simple cylindrical processes may be distributed relatively uniformly over the vesicle surface (Fig. 1C); in others, they occur as isolated knoblike buds (Figs. 1I, 3A) or elongate filamentous extensions (Fig. 1D). In most cases, however, the processes are further distinguished by distal branching (Figs. 1A,B,F,J,M,N, 2A, 4, 5) and a capacity to form closed loops through secondary fusion. This fusion appears to be relatively indiscriminate and gives rise to a wide range of expression: occasionally the processes return directly to the vesicle to form simple loops (Fig. 4A); in other cases they have fused either with themselves (Fig. 4F) or, more commonly, with other processes (Fig. 4E,G), resulting in a distally interconnected network (Figs. 2A, 5). Multiple layers of process networks are also developed, sometimes to the extent of obscuring the central vesicle (Fig. 6A,B,E,F). Such variability, combined with a recurrence of unfused buds—on both the vesicle (Fig. 3B) and processes (e.g., Fig. 4B–D)—attests to the actively growing habit of these structures.
All of the processes are hollow and many communicate freely with the vesicle lumen (Figs. 1H,J,K,O, 4A,B,D). Others, however, are separated from the vesicle by a basal, or near-basal, locally darkened, cross-wall (Fig. 4E– G). Remarkably, many processes are further divided by multiple internal septa, such that the processes are represented by series of isolated compartments (Fig. 4B–H). These cells vary considerably in length, but most are conspicuously longer than wide—the shortest appear to be associated with process initiation (Fig. 4B) and/or process branching (Fig. 4C– F,H). Process branches are further characterized by the presence of oblique and/or intersecting septa (Fig. 4C–F,H).
In some specimens the transition from the vesicle to process is conspicuously gradual, such that it is not always clear where the vesicle ends and a process begins (e.g., Fig. 1L). Recognition of these broad-based “presumptive” processes is often facilitated by the superposition of the process and vesicle walls, which results in discretely defined areas of doubled opacity. The effect is similar to the doubling of wall thickness during collapse-induced folding, but in this case it is an expression of real topographic differentiation of the vesicle wall. Indeed, “typical” collapse-induced folding is conspicuously lacking in this population, direct evidence of a relatively plastic “histology” (Butterfield 2003). The pseudo-folds document a highly variable range of vegetative outgrowths, from nascent processes (Fig. 1L), to linear and arcuate ridges (Figs. 1K,N,O, 2B), to major outpocketings that rival the parent vesicle in size (Figs. 1A, 2D).
As with the “normal” processes, these more irregular outgrowths were capable of secondary growth and fusion (e.g., Figs. 2D, 3E, 4H), and in some instances it is possible to reconstruct the course of their development. In CAMSM X.41243, for example, the large elongated outgrowth (Figs. 2D, 3E; recognized here by its doubled opacity, but also observed as a separate, basally attached, outgrowth during preparation) appears to have been initiated by the fusion of two primary processes originating from the primary vesicle. This secondary vesicle then expanded to produce an elongated sac-like structure, but with a proximal portion developing a process-like outgrowth that grew in parallel with the main body of the secondary vesicle, and subsequently rejoining it to produce the large rectangular foramen (Figs. 2D, 3E). The smaller, more distal foramen (Fig. 2D, arrow) formed through a similar course of differential outgrowth and fusion.
Other specimens show different stages and/or habits of secondary vesicle formation. Local inflation of a process junction (Fig. 4D, lower right), for example, yields a structure topologically equivalent to the outgrowth in Figure 2D, conceivably representing the earliest phase of differentiation. An ontogenetically more advanced phase appears to be represented in CAMSM X.43216 (Figs. 1M, 4H), where a seemingly fully differentiated secondary vesicle is suspended within the fused process network of its host vesicle.
In isolation, at least some of the secondary vesicles produced by these fossils would appear as smooth-walled, more or less irregularly shaped vesicles bearing variable numbers of irregularly distributed, sometimes branched, sometimes septate, usually equatorially arranged processes (e.g., Figs. 1M [4H], 2D [3E]). Interestingly, just such forms co-occur in abundance with the fossils described above (Fig. 2C, 7A–Q). These Germinosphaera-like vesicles range from 18 to 150 μm in transverse dimension (x̄ = 89.6 ± 33.2 μm, n = 55), show little evidence of collapse-induced folding, and bear from one to fifteen, hollow, mostly equatorially positioned processes. Processes diameter varies from 2 to 10 μm (x̄ = 5.3 ± 2.3 μm, n = 20), with proximal portions commonly expanded to merge gradually with the central vesicle (Fig. 7F,J,P). Process length varies from small knoblike buds (Figs. 3B, 7B) to tubular extensions more than 130 μm long (Fig. 7A); except for the buds, all of the processes are open ended, suggesting some sort of prior connection. Some of the processes branch (Fig. 7D,K,L,P), and at least one specimen features an unambiguous loop produced by process fusion (Figs. 3C, 7Q). A single specimen has two processes with proximal, outwardly curving septa (Figs. 3D, 7J).
These fossils present a daunting array of shapes and forms, but most can be accommodated in two (loosely defined) form-genera, Tappania Yin, 1997, and Germinosphaera Mikhailova, 1986.
Tappania was originally described from the Meso-Neoproterozoic Ruyang group of North China (Yin 1997) as spherical to subspherical vesicles bearing a variable number of irregularly distributed, short, hollow, heteromorphic, tubular to conical processes that communicate freely with the vesicle lumen, usually accompanied by a prominent, neck-like, distally closed extension. Such a habit is common in the Wynniatt assemblage, with some specimens all but indistinguishable from the type material (e.g., Fig. 1H,O). At this level, the only substantial difference is the more regular expression of neck-like extensions in Ruyang Tappania plana, though these are not characteristic of co-occurring (and possibly conspecific) T. tubata (see Yin 1997). Other differences between the Ruyang and Wynniatt populations are readily interpreted as artifacts of taphonomy, most probably a consequence of the more rigorous palynological techniques used to extract the Ruyang fossils. Partial and complete loss of distal process networks is commonly encountered in the Wynniatt population (e.g., Figs. 1H–I,L 6B–D,G,H) with the break often associated with the thickened basal septum. The end-product of such breakage yields short simple processes with thickened termini, a conspicuous feature of the Ruyang population (Yin 1997)—in other words, the processes of Ruyang Tappania were likely to have been septate and considerably more elaborate than as presently expressed. The attenuated form of the Tappania type material has also been recognized by Javaux et al. (2001) who identified T. plana in the 1450 Ma Roper Group of northern Australia, despite its conspicuously longer and occasionally branched processes. Roper Group Tappania do express the neck-like extensions seen in the Ruyang assemblage, but, unlike the type material, these vesicles typically lack processes, and their neck-like extensions occasionally show distal openings (Javaux et al. 2003: Fig. 1.8; E. Javaux personal communication 2004); interestingly, Wynniatt Tappania co-occur with smooth-walled vesicles bearing both open and closed outgrowths (e.g., Butterfield and Rainbird 1998: Figs. 3G, 4C,I,J). More importantly, at least one specimen of Tappania in the Roper population is now recognized as having basally septate processes (Harvard University Paleobotanical Collection palynological slide GG1 340.2.2+25; E. Javaux and A. H. Knoll personal communication 2004).
Germinosphaera was originally described from the Neoproterozoic Dashkin Suite, Siberia (Mikhailova 1986) as spheroidal vesicles bearing one or two hollow tubular extensions that communicate freely with the vesicle lumen, a diagnosis that was later emended to include vesicles with one to six open-ended, occasionally branched processes restricted to a single equatorial plane (see Butterfield et al. 1994: p. 36). The Wynniatt assemblage includes many such forms, but also specimens with up to 15 processes (e.g., Fig. 7C,J,P). Insofar as these more elaborate forms are otherwise indistinguishable from Germinsphaera s.s., and that earlier descriptions were based on small populations, there is a strong case for further expanding the Germinosphaera diagnosis (as opposed to establishing a new form-genus). By the same token, the occurrence of occasional specimens with basally septate (Figs. 3D, 7J) or terminally fused (Figs. 3C, 7Q) processes might reasonably be considered members of a relatively continuous, intraspecific, distribution.
Wynniatt Tappania and Germinosphaera share a conspicuous range of characteristics, including a similarly plastic wall construction, vegetative growth, a capacity for secondary fusion, intermittently septate processes, and a strikingly similar habit of process branching (compare, e.g., Fig. 7D,P with Figs. 1J,K, 4B). Indeed, the only substantive difference between these two forms is the relative restriction of processes in Germinosphaera to a single equatorial plane (but see Fig. 7H,I). Interestingly, Roper Group Tappania tend to have a Germinosphaera-like distribution of equatorially arranged, often conspicuously asymmetrical processes (e.g., Javaux et al. 2001: Fig. 1C; E. Javaux personal communication 2004). Combined with the facts that some Wynniatt Tappania bear contiguous Germinosphaera-like secondary vesicles (Figs. 1M, 2D, 3E) and that the processes of most isolated Germinosphaera are terminally incomplete, such a habit presents a strong case for identifying these two morphological modes as “organ taxa” of a rather more complex organism.
The identification of two named genera in a single organism presents a taxonomic dilemma. Although Germinosphaera has priority by date of publication, it clearly fails as a name for describing the rich morphology of the Tappania component. The reason is simple—Germinosphaera is a relatively simple form-genus likely to occur in a disparate range of plant protists and fungi (Butterfield et al. 1994: p. 38). By contrast, the morphology of Tappania is sufficiently distinctive to identify it as a real biological entity. In other words, Tappania is a natural taxon and, as such, is the preferred name to describe the whole organism—in the same way, for example, that Leiosphaeridia 1958 can be legitimately supplanted by Jacutianema 1980 when shown merely to be a constituent “organ” (see Butterfield 2004). Such usage is also supported by the ICBN, at least when dealing with the higher fungi, which similarly contend with supernumerary form- and organ-taxa: “In non lichen-forming ascomycetous and basidiomycetous fungi with mitotic asexual morphs (anamorphs) as well as a meiotic sexual morph (teleomorph), the correct name covering the holomorph (i.e., the species in all its morphs) is the earliest legitimate name typified by an element representing the teleomorph…” (Article 59.1); “Irrespective of priority, names with a teleomorphic type take precedence over names with an anamorphic type when both types are judged to belong to the same holomorphic taxon” (Article 59.4). This is not to suggest that the Tappania phase of Tappania is necessarily a fungal teleomorph but, like a teleomorph, it does represent the more complex and potentially more diagnostic phase of a multiphasic organism.
As in other studies (e.g., Butterfield 2000, 2004; Yin et al. 2004), recognition of heteromorphic growth series, differential taphonomy, and disparate intraspecific morphology in Tappania has depended on a combination of exceptional preservation and unusually large populations. The payoff, however, is a much more accurate reflection of true paleodiversity and its secular trends. If Proterozoic form-taxa do not reflect true biological species—as increasingly appears to be the case—then they fail even to provide a measure of morphological diversification (contra Knoll 1994: p. 6746). The number of potential form-taxa deriving from single populations of, e.g., Tappania or Jacutianema (Butterfield 2004) is more than enough to obscure, possibly obliterate, any useful evolutionary signal.
The more interesting question is what Tappania represents biologically. In the first instance it is worth recognizing it as a eukaryote. The combination of large size, complex morphology and capacity to remold and secondarily fuse its cell walls reflects a cytokinetic and cytoskeletal sophistication limited entirely to a eukaryotic grade of organization (Cavalier-Smith 2001; Javaux et al. 2001, 2003). Moreover, its highly variable and commonly asymmetric morphology identifies Tappania as both benthic and metabolically active (unlike certain superficially similar dinocysts which have an entirely distinct habit and mode of formation [Lentin et al. 1994; Kokinos and Anderson 1995; M. Head personal communication 2003]). Most remarkably, the presence of multiple cross-walls within the actively growing processes identifies these structures as multicellular. Tappania was a benthic, metabolically active, multicellular eukaryote.
Its ability to form a coherent system of differentiated multicellular processes and to differentiate morphologically distinct secondary vesicles further identifies Tappania as developmentally complex with an organism-level control over cell fate (Michod 1997; Butterfield 2000). The presence of relatively closely spaced septa in some newly erupted processes of Tappania (e.g., Fig. 4B), combined with a greater separation in more “mature” processes, suggests that both terminal and intercalary cells contributed to growth, presumably via intercommunicating, semipermeable cell walls (Heinlein 2002). Such “grade of organization” detail is important in its own right, but should also help to resolve the phylogenetic relationships of these fossils (see Butterfield 2000).
The most conspicuous feature of Wynniatt Tappania is the fusion of its filament-like processes. In extant organisms, it is a phenomenon best known from the “higher fungi” (Dikaryomycotina = ascomycetes + basidiomycetes) whose vegetative hyphae are capable of both intra-organismal and inter-organismal fusion (Gregory 1984). The underlying mechanisms of fungal anastomosis have yet to be fully resolved, but they clearly entail both genetic and epigenetic control over hyphal tip growth, intercellular signaling, targeting of hydrolytic enzymes, secondary fusion of the plasma membrane, and construction of a new, contiguous, cell wall (e.g., McCabe et al. 1999; Xiang and Morris 1999; Glass et al. 2000; Hickey et al. 2002; Galagan et al. 2003). Whatever its phylogenetic affiliations, Tappania had clearly mastered an equivalent level of morphogenetic control.
Is it possible that Tappania is, in fact, an early fungus? Certainly hyphal anastomosis is a conspicuous synapomorphy of the “higher fungi” (Gregory 1984), and at least some filamentous fungi are capable of both apical and intercalary growth (Burnett 1976: p. 88). Moreover, the oblique and intersecting septa seen in some Tappania processes (Fig. 4C–F) have possible parallels in the crosiers and clamp connections of extant ascomycetes and basidiomycetes. At a larger scale, Wynniatt Tappania also show an intriguing similarity with the ascocarps of gymnoascacean plectomycetes (Ascomycota), in which ascal clusters are surrounded by a loose network of thicker-walled hyphae, with or without differentiated “appendages” (Currah 1985). There is, however, no evidence of multiple asci or ascospores within Tappania vesicles, and the absence of any dehiscence structures (except, perhaps, the distally open neck-like extensions in some Roper Group Tappania [Javaux et al. 2001, 2003]) rules out the possibility that they represent single asci. The fact that plectomycetes are both relatively derived (Heckman et al. 2001) and primitively terrestrial (Kohlmeyer 1986) also presents problems for this interpretation. This does not, of course, rule out the possibility of a fungal relationship for Tappania: Kohlmeyer (1986) argues convincingly for both the Fungi and Ascomycota having marine origins, and molecular-clock analyses have suggested that “higher fungi” extend back to the Mesoproterozoic (Heckman et al. 2001; Hedges et al. 2004).
Hyphal fusion is a synapomorphy of the higher fungi, but it is also reported to occur in the stipes of laminaralean brown algae (= kelp) (Gregory 1984). As in Tappania, these algal “hyphae” are associated with obliquely oriented septa (Fritsch 1945: p. 227), as well as with large, irregularly shaped cells (“allelocysts,” “cellules multiclaves”) (Fritsch 1945: p. 235, Fig. 84H–J). Fritsch (1945: p. 235), however, expresses some concern over the interpretation of laminaralean “allelocysts/cellules multiclaves,” and there is a possibility that some of these fungal-like attributes may be those of parasitic fungi, now known to infect kelp (cf. Schatz 1983). Interpretation of Tappania as a kelp is also compromised by fundamental disparities at the organismal level, if not the purportedly late (Mesozoic) appearance of photosynthetic heterokonts (Medlin et al. 1997).
Another approach to understanding Tappania is to consider its functional morphology—what exactly were its developmentally complex, metabolically expensive processes for? Unusually elongated processes, for example, might be interpreted as the hyphae of an osmotrophic mycelial saprobe (e.g., Figs. 1D, 7A), particularly when seen to penetrate associated acritarchs (Fig. 1D). Most processes, however, tend not to extend outward but grow back on themselves to form a complex of variously sized loops. In this case, the elaborated processes might be interpreted as an evolutionary response to predation pressure, the additional girth offering a size refuge from phagotrophic predators. This is certainly an expected strategy for planktic unicells (Butterfield 1997, 2001) but is less likely for benthic organisms which are not constrained by issues of buoyancy. More reasonably, the network of Tappania processes might be interpreted as enhancing survival in the presence of micropredatory, myzotrophic (pierce and evacuate-type feeding) protozoans, the multiple septa usefully isolating the central cytoplasm from the extremities. If the processes were adaptations against micropredation, however, then they should presumably have covered the whole of the vesicle in fairly regular fashion. This was clearly not the case in Tappania, its most conspicuous feature being the marked irregularity of its processes and the considerable exposure of unprotected surfaces, particularly in the Germinosphaera phase.
More speculatively, the irregularly looped processes of Tappania might be compared with the specialized hyphal traps of certain extant predatory fungi (cf. Barron 1981), most notably the nematode-trapping hyphomycetes (anamorphic fungi), Dactylella and Arthrobotrys (identified as ascomycetes on the basis of 18S rDNA; Ahrén et al. 1998). Trap formation in these nematophagous fungi is induced by the presence of prey, giving rise to localized development of adhesive nets and/or knobs, strikingly similar to the process networks— and, indeed, the shorter, simple processes and buds—seen in Wynniatt Tappania (Fig. 8; cf. Figs. 3A,B, 4A–D,F). At one level, of course, the comparison of Tappania with nematophagous fungi is entirely superficial: looplike traps in modern fungi are limited to terrestrial mycelia, and there is little likelihood of pre-Ediacaran nematodes (Butterfield 2004). But that's not to say that Tappania wasn't heterotrophic, or that its netlike and knoblike processes couldn't have been used for capturing motile prey. Testate amoebae, for example, are important prey items of extant predatory fungi (Barron 1981).
Insofar as chloroplasts originated as cyanobacteria, the positive identification of middle to late Mesoproterozoic red algae (Butterfield 2000) sets a minimum date for the evolution of phagotrophic heterotrophy, a habit requiring a relatively sophisticated combination of sensory, capture and ingestion/digestion systems. The trapping and (osmotrophic) digestion behaviors suggested here for Tappania are not fundamentally more complicated than those found in unicellular phagotrophs. Indeed, such a strategy might be expected given the antiquity of osmotrophic heterotrophy (Martin et al. 2003), the antiquity of marine fungi (Kohlmeyer 1986), the inherent advantages of multicellularity (Butterfield 2000), and the absence of more effective (e.g., metazoan) predators. If Tappania did indeed make a living trapping motile prey, its presence in the Mesoproterozoic Roper Group (Javaux et al. 2001) nearly triples the known age range of multicellular eukaryotic predators (cf. Bengtson and Zhao 1992; Hua et al. 2003); notably, it would also preclude predation, per se, as the proximal cause of the Cambrian explosion (cf. Stanley 1973; Butterfield 1997, 2001, 2004).
OTHER PUTATIVE PRE-DEVONIAN FUNGI
Unlike the unwalled cells of animals and phagotrophic protists, Fungi and fungal-like protists have relatively robust cell walls and thus a reasonable preservation potential (Butterfield 2003). Even so, the search for early fossil fungi has been decidedly unilluminating. Proterozoic candidates such as Eomycetopsis (Schopf 1968), and the “phycomycetes” (subsequently Caudosphaera) reported by Timoféev (1970) exhibit no convincing fungal synapomorphies and find equal or better comparisons with the cyanobacteria and plant protists, respectively. Darby (1974) speculated that some specimens of Gunflint Huroniospora might represent unicellular fungi on the basis of apparent budding. A single elongate structure containing eight poorly defined bodies from the Neoproterozoic Skillogalee Dolomite is conceivably, but not demonstrably, interpreted as an ascus with contained ascospores (Schopf and Barghoorn 1969). More recently, Hermann (1979) has proposed a number of putative fungal taxa from the late Mesoproterozoic Lakhanda Suite—Mycosphaeroides, Mucorites, Eosaccharomyces, Aimia, Majasphaeridium—but again without convincing synapomorphies. Likewise Burzin's (1993) claim that the nonseptate branched clavate filaments of Ediacaran Vendomyces are the remains of chytridiomycetes.
In the early Paleozoic, Redecker et al. (2000) have argued that mid-Ordovician fossils are the remains of glomalean fungi. Unfortunately, the morphological simplicity of these fossils fails to distinguish them from Vendomyces, various plant protists (e.g., Protosiphon), or indeed the infective hyphal phase of the mesomycetozoan fish parasite Ichthyophonous (cf. Spanggaard et al. 1995). Sherwood-Pike and Gray (1985) have made a stronger case for identifying ascomycetous fungi in Silurian-age (Ludlow) mudstones from Gotland, but the earliest uncontroversial fossil fungi, based on diagnostic morphology, behavior and/or host response, are those from the early Devonian Rhynie Chert (Taylor and Taylor 1997; Taylor et al. 1999).
Claims of early macroscopic fungi are even more dubious. Retallack (1994), for example, has argued that most Ediacaran macrofossils were lichens (photosymbiont-bearing ascomycetes), on the basis of their peculiar taphonomy, whereas Peterson et al. (2003) suggest they might be better viewed as stem-group fungi on the basis of inferred heterotrophy and immobility. Neither of these possibilities, however, is supported by specific characters or character analysis. The only compelling case for a macroscopic, pre-Mesozoic fungus is the gigantic stumplike Devonian fossil Prototaxites, which appears to be constructed from a variety of funguslike “hyphae” (Hueber 2001).
Other Early Acritarchs as Possible Fungi
Many acritarchs are so classified because they are effectively featureless, others because the features they express have yet to be recognized in extant groups. For those in the second category, the lack of phylogenetic resolution may well derive from a search image weighted excessively in favor of unicellular plant protists. Meso-/Neoproterozoic Trachyhystrichosphaera, for example, has long defied taxonomic placement, despite its distinctive morphology and widespread occurrence (see Butterfield et al. 1994: pp. 44–46). Like Tappania, it is characterized by (1) a conspicuously large and variable central vesicle; (2) a conspicuously variable number and distribution of heteromorphic, usually open-ended processes; (3) absence of excystment structures; and (4) a tendency to develop outgrowths from the central vesicle, sometimes to the extent of forming secondary vesicles (Fig. 9). Indeed, the comparison is sufficiently close to suggest that the two taxa share a common body plan. Both Tappania and Trachyhystrichosphaera were actively growing, benthic organisms (see Butterfield 1997, 2001) and probably of a multicellular/multinucleate grade of organization. Also like Tappania, there is no a priori reason to assume that Trachyhystrichosphaera was photoautotrophic.
In the Mesoproterozoic Ruyang Group, Tappania co-occurs with a variety of other acritarch taxa, including Shuiyousphaeridium, Dictyosphaera, and Foliomorpha (Yan and Zhu 1992; Yin 1997; Xiao et al. 1997). Shuiyousphaeridium is a large (110–250 μm) spheroid with a conspicuously reticulate/polygonal surface-ornamentation, unevenly covered with short, hollow processes, either simple or furcated, and commonly interconnected by lateral branches or membranous material (Yin 1997). Combined with the presence of a medial split excystment structure and a lack of communication between the processes and vesicle lumen, these characteristics are consistent with Shuiyousphaeridium being an unusually large dinocyst (cf. Kokinos and Anderson 1995). It might also, however, be compared to the ascocarps of various extant pyrenomycetes, especially the primitively marine Halosphaeriales and Lulworthiales (e.g., Jones et al. 1986; Nakagiri and Tubaki 1986; Kohlmeyer 1986; Kohlmeyer et al. 2000). In this case, the reticulate wall structure would correspond to the thickened, polygonal cells that form the (multicellular) ascocarp, the laterally fused processes to the hypha-like “appendages” commonly adorning these structures (e.g., in the terrestrial powdery mildews [Erisiphales], but also in marine pyrenomycetes [Kohlmeyer and Volkmann-Kohlmeyer 2003]), and the medial split to the dehiscence structure that releases asci/ascospores. The absence of included asci, ascospores, and pseudoparenchyma in these fossils is readily explained by differential degradation, marine ascomycetes having single-walled asci and nonmelanized hyaline ascospores (vs. the double-walled, melanized condition of their terrestrial counterparts [Kohlmeyer 1986]). Interpretation of Shuiyousphaeridum as a marine pyrenomycete, or even a fungus, is certainly not established, and it remains properly identified as an acritarch. Even so, there is good reason to recognize it as multicellular, and no a priori reason to assume that it was photoautotrophic.
Ruyang Dictyosphaera also occur as large spheroids with a pronounced reticulate/polygonal patterning of the vesicle wall, the only obvious distinction with Shuiyousphaeridium being the absence of processes—most likely a product of ontogenetic or taphonomic variation within a single biological species (Xiao et al. 1997; Kaufman and Xiao 2003). Ruyang Foliomorpha is similarly large, spheroidal and patterned (Yan and Zhu 1992: Fig. 3.4–3.7), but in addition expresses a single broad process that Yan and Zhu (1992) interpret as the stipe of a benthic coenocytic thallus. Insofar as the reticulate vesicle of Shuiyousphaeridium/Dictyosphaera/Folimorpha might represent a multicellular ascocarp, this single larger process might reasonably be compared to the tubular release structure produced by the ascocarps of various marine fungi; e.g., Halosphaeriopsis (cf. Jones et al. 1986: Fig. 19.3d) or Abyssomyces (cf. Kohlmeyer and Volkmann-Kolhlmeyer 2003: Figs. 1, 2).
The point is not trivial. Kaufman and Xiao (2003), for example, have recently argued that the conspicuously light carbon isotope values found in individual specimens of Ruyang Dictyosphaera specimens can be used to infer exceptionally high PCO2 in the Mesoproterozoic. It is an interesting result, but it is also based entirely on the assumption that Dictyosphaera was an (unusually large) unicellular, planktic, photoautotroph. As a probable benthic organism, however, it is more likely to have sampled local carbon reservoirs than mixed-layer CO2; as a probable multicellular organism, it is impossible to make any easy assumptions about surface area and diffusion; as a possible heterotroph its isotopic signature would merely be a reflection of its preferred substrate.
As the sister group of animals + choanozoans (Cavalier-Smith and Chao 2003), fungi were unquestionably present by at least the Ediacaran, and the identification of middle Neoproterozoic testate amoebae indicates that the Opistikonta (animals + choanozoans + fungi) had diverged from its Amoebozoa sister group by at least 750 Ma (Porter et al. 2003). Recent molecular-clock analyses have estimated fungal-metazoan divergence at ca. 1500 Ma (Hedges et al. 2004), not inconsistent with the identification of Meso- and Neoproterozoic Tappania as a fungus. Molecular clocks and molecular phylogenies, however, are fraught with analytical artifacts and do little to constrain alternative hypotheses. Martin et al. (2003), for example, have presented an intriguing case for rerooting the eukaryotic tree such that the Fungi represent the ancestral, presumably Archean, stalk.
Problematic fossils showing some, but not all, characters of extant crown groups provide a unique view of early evolutionary development (Budd and Jensen 2000). In the present case, Tappania would seem to fall somewhere between the “higher” and “lower” fungi; i.e., between those forms with regularly anastomosing, septate hyphae, and those without. Unfortunately, the “lower fungi” are recognized largely by their lack of diagnostic characters and have proven to be deeply polyphyletic (Redecker 2002). Even so, molecular analyses clearly identify the zygomycetes (non-anastomosing, nonseptate hyphae) as the immediate sister group of the Dikaryomcotina (Liu and Hall 2004), and the development of comparable hyphae in Ichthyophonus (Spanggaard et al. 1995) suggests that the hyphal condition could be primitive for the Opisthokonta. The presence of apical and intercalary hyphal growth in both the zygomycetes (Burnett 1976: p. 88) and Tappania (see above) stands as a potentially positive characteristic for placing Tappania between the dikaryomycetes and zygomycetes.
Assignment of a fossil to a particular stem group of course precludes the presence of any crown-group synapomorphies, and no derived dikaryomycete features have yet been identified in Tappania. This is not the case, however, for Ruyang Shuiyousphaeridium/Dictyosphaeridium/Foliomorpha, which I have argued might represent the multicellular fruiting bodies (ascoma) of a relatively derived (subphylum Pezizomycotina = euascomycetes) ascomycetous fungus. If correct, then the fungal crown would have been firmly established by at least 1 Ga. Interestingly, the most primitive ascomycetes (subphylum Taphrinomycotina [Landvik et al. 2001; Liu and Hall 2004]), lack differentiated ascoma, leaving open the possibility that Tappania, too, might represent a crown-group ascomycete.
The case for identifying Tappania as fungus is strong, but not conclusive. The presence of “hyphal fusion” and obliquely oriented septa in the stipes of modern kelp points to the possibility of convergence, an issue that has also figured in the assessment of Prototaxites. In the Paleozoic case, microanatomical (Hueber 2001) and δ13C (Boyce et al. 2003) data have tipped the balance clearly in favor of a fungal relationship. Isotopic analysis of multiple Tappania specimens (see Kaufman and Xiao 2003) might provide a comparable test of its phylogenetic affinities. In the meantime, Tappania remains a “probable” fungus, though one that is supported by more positive morphological data than any other pre-Devonian candidate. Its confirmation would extend the known range of the Fungi back to at least 1430 Ma (see Javaux et al. 2001), and displace 1200 Ma Bangiomorpha as the oldest taxonomically resolved eukaryote on record (cf. Butterfield 2000).
Whatever its phylogenetic relationships, Tappania provides an important measure of the morphogenetic “tool kit” available to early multicellular eukaryotes. Its biphasic morphology and irregularly distributed anastomosing processes reflect a sophisticated genetic and epigenetic developmental program, including a marked sensitivity to environmental stimuli (cf. McCabe et al. 1999; Xiang and Morris 1999; Glass et al. 2000; Hickey et al. 2002; Galagan et al. 2003). Moreover, the differentiation of multicellular processes/hyphae from an apparently coenocytic central vesicle identifies a capacity to form “symplastic domains”—isolated groups of communicating cells that function as developmental or physiological units—possibly via the selective trafficking of macromolecules by dynamic plasmodesmata and/or pit connections (Heinlein 2002). In other words, Tappania represents a complex multicellular grade of organization, seemingly more sophisticated than that expressed by (more or less) contemporaneous Bangiomorpha.
The Fungi are by far the most metabolically diverse eukaryotes (Martin et al. 2003), and the ascomycetes their most taxonomically and ecologically diverse phylum (Liu and Hall 2004). Both are renowned for their roles as saprobes, parasites, mutualists, predators, and exploiters of extreme environments (Barron 1981; Taylor et al. 1999; Blackwell 2000; Gorbushina et al. 2003; Liu and Hall 2004). Much of this diversity clearly arose as a coevolutionary response to the radiation of terrestrial plants and animals in the early Paleozoic, but it now appears that the Fungi also contributed significantly to pre-tracheophyte/pre-metazoan ecosystems. Such recognition should help to reduce the backlog of problematic Proterozoic fossils and shed important new light on early ecological structures and biological evolution, not least their expropriation by motile multicellular predators at the onset of the Phanerozoic.
Fieldwork for this study was funded in part by the National Geographic Society. I thank R. H. Rainbird (Geological Survey of Canada), G. M. Young (Western Ontario), A. H. Knoll (Harvard) and K. Swett (Iowa) for additional logistic support and field work. A. D. M. Rayner (Bath) and M. Head (Cambridge) offered, respectively a mycologist's and a palynologist's opinion on these fossils. D. Grazhdankin translated Russian literature, and A. H. Knoll reported on Tappania specimens in the Harvard University Paleobotanical Collection. I am particularly grateful to S. Porter (Santa Barbara) for her incisive review, and to E. Javaux (University of Liège) for discussion of the Roper Group Tappania. This is Cambridge Earth Science contribution 7813.
- Accepted 26 April 2004.