Lichenization is assumed to be a very ancient mode of fungal nutrition, but fossil records are rare. Here we describe two fragments of exceptionally preserved, probably charred, lichen thalli with internal stratification. Cyanolichenomycites devonicus has a cyanobacterial and Chlorolichenomycites salopensis a unicellular, presumably green algal photobiont. Fruiting bodies are missing. Cyanolichenomycites devonicus forms asexual spores in a pycnidium.
All specimens were examined with scanning electron microscopy techniques. The fossils were extracted by maceration. Extant lichens and free-living cyanobacteria were either experimentally charcoalified for comparison or conventionally prepared.
Based on their septate hyphal structure, both specimens are tentatively interpreted as representatives of the Pezizomycotina (Ascomycota). Their presence in 415 million yr (Myr) old rocks from the Welsh Borderland predates existing Late Cretaceous records of pycnidial conidiomata by some 325 Myr and Triassic records of lichens with broadly similar organization by some 195 Myr.
These fossils represent the oldest known record of lichens with symbionts and anatomy as typically found in morphologically advanced taxa today. The latter does not apply to Winfrenatia reticulata, the enigmatic crustose lichen fossil from the Lower Devonian, nor to presumed lichen-like organisms such as the Cambrian Farghera robusta or to the Lower Devonian Spongiophyton minutissimum.
If you can't find a tool you're looking for, please click the link at the top of the page to "Go to old article view". Alternatively, view our Knowledge Base articles for additional help. Your feedback is important to us, so please let us know if you have comments or ideas for improvement.
Lichens are the symbiotic phenotype of lichen-forming fungi, which associate with a photoautotrophic partner whence they derive fixed carbon in an ecologically obligate, mutualistic symbiosis (Honegger, 1991). Lichen-forming fungi and their photoautotrophic partners are an impressive example of the innovative force of mutualistic symbiosis. Approximately 85% of extant lichen-forming fungi are symbiotic with minute green algae (Chlorophyta), c. 10% with cyanobacteria (most of them being diazotrophic, thus providing the fungal partner with fixed carbon (C) and nitrogen (N)), and c. 4% simultaneously with both (for a review, see Honegger, 2009). Lichen-forming fungi are a taxonomically diverse group of nutritional specialists, with c. 20% of all extant fungal species being lichenized (Kirk et al., 2008). Ninety-nine per cent of lichen-forming fungi are ascomycetes of the subphylum Pezizomycotina, in which phylogenetic multilocus analyses imply multiple gains and losses of lichenization (Lutzoni et al., 2001, 2004; Nelsen et al., 2009). Less than 1% of species belong to the Agaricomycotina among the Basidiomycota.
Extant lichens are morphologically diverse, the majority of species forming crustose, leprose, microfilamentous or microglobose thalli with no internal stratification. Only one out of four species of lichen-forming fungi produce internally stratified, morphologically complex leaf-like, strand-shaped, tubular or shrubby thalli which rise above the substratum. Such thalli are the product of an extraordinary hyphal polymorphism, the main building blocks being hydrophilic, conglutinate, tissue-like zones, usually as a peripheral cortex, and loosely interwoven hyphal meshworks with hydrophobic wall surfaces which form a gas-filled internal medullary layer. These 3D lichen thalli are the most highly advanced vegetative structures in the fungal kingdom (for reviews, see Honegger, 1991, 2001). The photobiont cell population, either green algae or cyanobacteria, is kept at the periphery of the medullary layer. The algal or cyanobacterial cell turnover is controlled, in an as yet unknown manner, by the fungal partner (Honegger, 1987, 1991). These lichen-forming fungi compete for space above ground, facilitate gas exchange, provide water and dissolved mineral nutrients, and secure the adequate illumination of their photoautotrophic partner. Every photobiont cell is actively positioned in the algal layer by means of growth processes within the contacting fungal hyphae (Honegger, 1986, 1991).
Lichenization is probably a very ancient mode of fungal nutrition. Marine cyanobacterial mats, which were invaded, but obviously not damaged by hyphae of saprotrophic or mutualistic fungi, were detected in phosphorites of the c. 600 million yr (Myr) old Doushantuo Formation in south China (Yuan et al., 2005). However, as concluded from geochemical signals, cyanobacteria seem to have colonized palaeosol surfaces in terrestrial ecosystems as early as the Archaean (2.6–2.7 billion yr ago; Watanabe et al., 2000). Live or dead, cyanobacterial and algal mats (Tomescu & Rothwell, 2006; Tomescu et al., 2006) served as a source of fixed C and, in the case of diazotrophic cyanobacteria, of fixed N to land-dwelling fungi long before the advent of embryophytes (Retallack, 2009). Switches from saprophytic or parasitic to mutualistic interactions with cyanobacteria and algae occurred independently in unrelated groups of fungi, and lichenization was repeatedly lost (Lutzoni et al., 2001, 2004; Nelsen et al., 2009).
The problems related to the calibration of fungal time trees are discussed by various authors (Blackwell, 2000; Padovan et al., 2005; Taylor & Berbee, 2006; Pulquério & Nichols, 2007; Avise, 2009; Blair, 2009; Lücking et al., 2009; Berbee & Taylor, 2010). According to molecular clock data sets, the predominantly lichenized Lecanoromycetes diverged early in the evolution of the subphylum Pezizomycotina; estimates range from c. 850 to 250 Myr ago (Lücking et al., 2009). Unfortunately, fossil records are very scarce, the main reasons for this being the lack of biodegradation-resistant compounds in lichen thalli, but also the lack of experts capable of detecting fossil lichens. The Cambrian-Ordovician terrestrial Farghera robusta shares its external morphology with extant liverworts or certain lichens (Retallack, 2009), but no cellular details can be resolved. By contrast, the anatomically well-preserved Winfrenatia reticulata, a peculiar, epiphytic, cyanobacterial lichen-like consortium from a hot spring ecosystem, was detected in the Lower Devonian Rhynie Chert (Taylor et al., 1995a, 1997; Karatygin et al., 2009). This crustose-reticulate fossil bears no morphological similarity to extant lichens. Based on light microscopy of thin sections, its fungal partner was proposed to belong to the Mucormycotina (referred to as Zygomycota; Taylor et al., 1997), a phylum with no extant lichenized members. The Lower Devonian Spongiophyton minutissimum, hypothesized as a fossil lichen, has a compact, nonstratified thallus structure and bears no anatomical similarity to extant lichens, and neither a fungal nor an algal or cyanobacterial partner was properly identified in scanning electron microscopy (SEM) micrographs (Taylor et al., 2004). However, its C isotope composition (δ13C) resembles that of extant lichens (Jahren et al., 2003).
When did the first foliose or fruticose lichens emerge with morphologically advanced, internally stratified thalli that are anatomically similar to extant taxa? Phylogenetic timing based on a data set of three loci suggested that the Parmeliaceae core, ancestor of the main group of extant, foliose macrolichens, evolved c. 100 Myr ago (Amo de Paz et al., 2011). However, fragments of structurally well-preserved, dorsiventrally organized lichens with peripheral cortex and internal stratification similar to extant Parmelia spp. were detected in the Upper Triassic (Carnian; c. 220 Myr ago) of Lower Franconia in Germany (Ziegler, 2002). The external morphology of the Middle Jurassic Daohugouthallus ciliiferus resembles that of fruticose reindeer lichens, but neither fungal nor algal or cyanobacterial partners could be resolved (Wang et al., 2010c). Minute, squamulose thalli resembling extant Phyllopsora spp. (Bacidiaceae), but also foliose and fruticose fossils anatomically similar to extant genera among the Parmeliaceae (Lecanorales) were found in Baltic and Dominican amber (Miocene, c. 40–60 Myr ago) (Mägdefrau, 1957; Poinar et al., 2000; Rikkinen & Poinar, 2002, 2008).
Here we report on two fossil fragments from Lower Devonian (Lochkovian) strata of the Welsh Borderland (Edwards, 1996; Edwards & Richardson, 2004), representing dorsiventrally organized, internally stratified lichen thalli comparable to extant species but some 195 Myr older than the Triassic fossils. Neither was preserved in situ: after charcoalification by wildfires the fragments were transported by wind and water before ending up in fluvial sediments. Both fossils were recovered from macerates of grey siltstone.
Materials and Methods
Extraction and SEM of fossils
Specimens were extracted from a pale grey, unconsolidated siltstone by gentle disaggregation in water, followed by treatment with concentrated hydrochloric acid for 24 h, diluted with water (1 d) and then subjected to commercial strength (40%) hydro-fluoric acid (3 d) to dissolve the matrix. The resulting residues were washed repeatedly with water until no longer acidic and then washed through 250-μm polyester mesh. The dissolved mineral matter passed through the mesh and the organic debris, which resembles coarse tea leaves, was retained. This was transferred to a Petri dish, allowed to dry and then viewed with a stereomicroscope to select specimens for detailed examination in a scanning electron microscope. Specimens were mounted on sticky carbon discs on aluminium stubs and splutter-coated with gold-palladium, before viewing in an FEI (Philips, Eindhoven, the Netherlands) XL30 ESEM FEG scanning electron microscope at 20 kV. After initial photography, the specimens were turned over and imaged again and finally broken apart with a razor blade to reveal as much information as possible and re-imaged.
Specimens for transmission electron microscopy (TEM) and SEM analyses were freshly collected and processed either within a few days or after storage, in the desiccated state, at −20°C in a deep-freezer where they retained their full viability. Taxa investigated were: Lobaria virens (With.) J.R. Laundon: Forêt de Huelgoat, Brittany, France; Peltigera aphthosa (L.) Willd.: Klöntal, Kt. Glarus, Switzerland; Peltigera canina (L.) Willd.: Murgtal, Kt. St. Gallen, Switzerland; Sticta sylvatica (Huds.) Ach.: Forêt de Huelgoat, Brittany, France; Xanthoria parietina (L.) Th. Fr.: Botanical Garden of the University of Zürich, Switzerland.
In order to be able to compare the anatomy of extant lichens with that of the fossils, which are presumed to be charcoalified (Edwards & Axe, 2004; Glasspool et al., 2006), extant cyanobacterial lichens with Nostoc photobiont (Leptogium lichenoides (L.) Zahlbr. and Peltigera canina (L.) Willd.) and free-living colonies of Nostoc commune Vaucher ex Bornet & Flahault were freshly collected in the Botanical Garden of the University of Zurich, transported to Cardiff in the desiccated state and made into charcoal within 3 d of collection. The samples were partially dissected with a razor blade into suitable fragments and then placed on small pieces of aluminium foil which were folded into flat parcels to exclude air. They were then placed in a furnace preheated to 400°C for 5 min, removed and allowed to cool.
SEM of extant specimens
Conidiophores and conidia in mucilage-filled pycnidia or the mycobiont–photobiont interface in cyanobacterial and green algal lichens, respectively, were visualized by chemically removing the mucilage from either dissected pycnidia or of cytoplasmic remains from dissected thallus fragments before chemical fixation, dehydration and critical point drying (Honegger, 1984a, 1985; for details, see Supporting Information Methods S1).
SEM of charcoalified, extant specimens
After charcoalification (see Charcoalifying experiments), specimens were mounted on sticky carbon discs on aluminium stubs and splutter-coated with gold-palladium, before viewing in an FEI (Philips) XL30 ESEM FEG scanning electron microscope at 20 kV. After initial photography, the specimens were fractured with a razor blade to reveal fresh surfaces and re-imaged.
TEM of extant specimens
Ultrathin sections (details in Methods S3) of conventionally prepared pycnidia or thallus fragments (methods in Honegger & Brunner, 1981; Honegger, 1984a; details in Methods S2a) or of cryofixed, freeze-substituted thalline fragments, respectively (methods in Honegger et al., 1996; details in Methods S2b), were examined in a Hitachi (Tokyo, Japan) HS 8 transmission electron microscope at 50 kV. Replicas of gelatinous sheaths of symbiotic Nostoc sp. from cephalodia of Peltigera aphthosa were made according to the protocol for algal cell wall isolation and purification (Honegger, 1982; Brunner & Honegger, 1985; details in Methods S4).
At the SEM level, both specimens, although very fragmentary, are surprisingly well preserved, probably as a result of charring by wildfires and hence a lack of compression. Neither specimen contains mature fruiting bodies.
An internally stratified cyanobacterial lichen with pycnidial conidioma
Fungi, Ascomycota, Pezizomycotina
Cyanolichenomycites Honegger, Edwards et Axe nov. gen.
Generic diagnosis : lichen-forming ascomycete with septate hyphae. Dorsiventrally organized, heteromerous thallus with a conglutinate, several cell layers thick cortex at the upper thallus surface, a loosely interwoven medullary layer and a population of cyanobacterial colonies resembling extant Nostoc spp. in a distinct photobiont layer underneath the cortex. Mycobiont–photobiont interface with characteristic intragelatinous fungal protrusions in the thick sheaths of the cyanobacterial colonies.
Type species : Cyanolichenomycites devonicus Honegger, Edwards et Axe, nov. spec. Vegetative part (anatomy best visible in the fracture plane at the periphery of the fertile part): dorsiventrally organized lobule, c. 0.1 mm thick, with distinct photobiont layer. Cyanobacterial photobiont cells retained as amorphous pyrite, their gelatinous sheaths with characteristic constrictions being well preserved. Fertile part semi-globose, with immersed pycnidial conidioma. Conidiophores conical, with phialidic conidium formation.
Holotype : sputter-coated specimen on SEM grid, sample Nr. HD496/01; illustrations in Figs 1 (a,b), 2(a–e); see description below.
Repository : National Museum of Wales, Cardiff, UK.
Etymology : the generic name is a combination of the cyanobacterial photobiont preference (Greek kyaneos: blue-green; Cyanobacteria: photosynthetic, gram-positive bacteria, formerly termed blue-green algae) of a lichen-forming fungus (Greek leichen: lichen; Greek mykes: fungus), its fossil state being indicated by the latin suffix -ites. The species epithet refers to the geologic time (Devonian).
Locality: Stream section north of Brown Clee Hill, Shropshire, Welsh Borderland, UK.
Stratigraphy : Mid micrornatus-newportensis Sporomorph Assemblage Biozone, in the lower part of the Ditton Group, Lochkovian (Lower Devonian).
Age : Early Devonian, c. 415 Myr
The pycnidium (from Greek: pycnós: tight), that is, the semi-globose part of Cyanolichenomycites devonicus (Figs 1a, 2a), is superficially damaged and thus lacks its original surface structure, especially its opening (ostiole). Pycnidial conidiomata, usually bottle-shaped, are the site of asexual spore (conidium) formation. Their inner wall is lined by conidiophores (specialized, conidia-bearing hyphae). The pycnidial cavity of extant samples is filled with hygroscopic mucilage. Hundreds of conidia ooze out of the apical opening (ostiole; Fig. 1c) after having been released into the mucilage of the fully hydrated pycnidium (for details on pycnidial conidiomata in extant lichens, see Notes S1).
The pycnidial conidioma of C. devonicus comprises a central cavity which was probably mucilage-filled. The conical cells contained in the cavity (Fig. 1b) are conidiophores which produce conidia in a similar manner to extant taxa (Fig. 1f–h). No mature, detached conidia are seen, all conidiophores being in the process of phialidic conidium formation (Fig. 1b), a widespread mode of conidiogenesis among extant nonlichenized ascomycetes and also in the predominantly lichenized Lecanoromycetes (Fig. 1f–h; Honegger, 1984a).
The pycnidial conidioma of C. devonicus was either surrounded by several thallus lobules like the one that is preserved or part of a larger thallus, as in extant taxa (Fig. 1c–h). The lateral lobule has kept its original surface layer, a cortex whose surface is built up by globose, conglutinate cells and a smooth covering resembling a fossil plant cuticle (Fig. 2b). Vertical fractures along the pycnidial periphery show several layers of cells with conglutinate walls forming a thick peripheral cortex and a loosely interwoven meshwork of hyphae, the latter resembling the medullary layer of extant, dorsiventrally organized lichens (Fig. 2e). Below the cortex are Nostoc-like cyanobacterial colonies (Fig. 2c–e). From most of these only the gelatinous sheaths are preserved. However, a few of the presumed cyanobacterial cells are retained in their characteristic shape and position within their sheath (Figs 2c–e). Two cyanobacterial sister cells are preserved after cell division (Fig. 2c,d). A larger cyanobacterial cell might be a heterocyst (Fig. 2c). As concluded from elemental analysis (energy dispersive X-ray analysis, EDAX), the remains of these cyanobacterial cells are composed of amorphous pyrite. In the backscatter mode the pyritized cyanobacterial cells yield a bright signal and differ distinctly from the sheath and the surrounding fungal hyphae (Fig. 2d). Extant cyanobacterial lichens with Nostoc photobiont and free-living Nostoc commune colonies were subjected to mild charring conditions. All of them lost their cyanobacterial cells but retained their gelatinous sheaths (Fig. 3a–f).
As in extant cyanobacterial lichens with internally stratified thallus and Nostoc sp. as photobiont (Fig. 2f–i), the fungal hyphae grow into and between the gelatinous sheaths of the cyanobacterial colonies. In fossil (Fig. 2e) and extant samples (Figs 2f–h, 3b,c), including nonlichenized Nostoc commune (Fig. 3e,f), characteristic holes are seen in the cyanobacterial sheaths at the constrictions between almost mature, adjacent cells. The microfibrillar polysaccharides of the mucilaginous sheaths of cyanobacterial colonies (Fig. 3g,h) are continuously synthesized at the cyanobacterial cell surface and thus reflect the size and developmental stage of the cyanobacterial cells. Successive developmental stages of an extant cyanobacterial photobiont (Nostoc sp.) are seen in transmission electron micrographs of ultrathin sections (Fig. 2i). Structurally the vegetative part of this fossil cyanobacterial lichen resembles extant cyanobacterial Peltigerales, but the conidia-bearing part cannot be linked to any extant clade of lichen-forming ascomycetes.
An internally stratified, presumed green algal lichen
Fungi, Ascomycota, Pezizomycotina
Chlorolichenomycites, Honegger, Edwards et Axe nov. gen.
Generic diagnosis : lichen-forming ascomycete with septate hyphae. Dorsiventrally organized, heteromerous thallus with loosely interwoven medullary layer and a population of globose green algal cells in a photobiont layer underneath the one cell layer thick, conglutinate cortex at the upper thallus surface. Lower thallus surface formed by loosely interwoven hyphae.
Type species: Chlorolichenomycites salopensis Honegger, Edwards et Axe, nov. spec.
Thallus 60–110 μm thick. Conglutinate cortex at the upper thalline surface comprising one cell layer, c.21 μm thick, with smooth surface. Most of the photobiont cells confined to the photobiont layer below the cortex, but some occurring between the medullary hyphae or at the lower thallus surface, respectively. Mycobiont–photobont interface characterized by simple wall-to-wall appositions. Photobiont cells globose, with 16–21 μm diameter, pyritized contents in framboidal form and a thin, degradable cell wall.
Holotype : sputter-coated specimen on SEM grid, sample Nr. HD466/02; illustrations in Fig. 4(a–e); see description below.
Repository : National Museum of Wales, Cardiff, UK
Etymology: the generic name is a combination of the green algal photobiont preference (Greek chloros: green; Chlorophyta: green algae) of a lichen-forming fungus (Greek leichen: lichen; Greek mykes: fungus), its fossil state being indicated by the latin suffix -ites. The species name refers to the type locality, Salop being another name for Shropshire, a county in the West Midlands of England, which borders Wales to the west.
Locality : stream section to the north of Brown Clee Hill, Shropshire, UK
Stratigraphy : Mid micrornatus-newportensis Sporomorph Assemblage Biozone in the lower part of the Ditton Group, Lochkovian (Lower Devonian).
Age : Early Devonian, c. 415 Myr
The well-preserved fragment of C. salopensis (Fig. 4a–e) bears strong similarity to dorsiventrally organized thalli of extant lichen-forming ascomycetes with green algal photobionts (Fig. 4f,g). The globose, presumed photobiont cells, whose pyritized contents are retained in framboidal form (raspberry-like, spherical assembly of microcrystals; Fig. 4c–e), correspond in size and shape to representatives of the genus Trebouxia (Trebouxiophyceae, Chlorophyta), the most common and widespread lichen photobionts which associate with > 50% of extant lichen-forming fungal species. Pyritization probably occurred under anoxic conditions in the sediment. The walls of the fossil green algal cells are poorly preserved, which indicates that their structure and composition were distinctly different from the relatively thick walls of the adjacent fungal hyphae. Only very rarely is the delicate algal wall around the framboidal cell contents preserved (Fig. 4e). Extant Trebouxia spp. reveal cell diameters between 10 and 20 (± 4) μm (Fig. 4f,g), the fossil photobiont cells being in the same range. The occurrence of photobiont cells of different diameters in close vicinity (Fig. 4d,e) is a characteristic feature of trebouxioid green algal photobionts in extant lichen thalli.
Simple wall-to-wall apposition at the mycobiont-photobiont interface is commonly found in numerous extant taxa (Honegger, 1984b, 1991). Chlorolichenomycites salopensis, the fossil lichen-forming ascomycete, seems to have lacked the capacity to precisely position each cell of the photoautotrophic partner by means of haustorial complexes and growth processes at the contact site as evident in Lecanoromycetes of extant macrolichens (Honegger, 1986, 2009).
As septa were found in C. devonicus and C. salopensis we tentatively interpret them as lichenized representatives of the Pezizomycotina. Both fossils lack sexual reproductive structures; therefore it is not possible to assign them to a fungal class, despite the anatomical similarity with extant Lecanoromycetes. Estimates based on phylogenetic analyses of three DNA regions place the divergence between the ancestors of the Dothideomycetes and Lecanoromycetes, the latter comprising the majority of extant lichen-forming ascomycetes, in the Carboniferous (c.350 Myr ago; Amo de Paz et al., 2011). The discovery of Early Devonian cyanobacterial and presumed green algal lichens with internally stratified thalli indicates either that this split might have occurred earlier in the Palaeozoic (e.g. at 850–480 Myr ago, as proposed by Padovan et al., 2005; for a review, see Lücking et al., 2009), or that morphologically advanced lichens already existed before this divergence.
Fossil pycnidial conidiomata (pycnidia)
Pycnidia are very common and widespread among extant lichen-forming Arthoniomycetes and Lecanoromycetes; they are filled with water-soluble, hygroscopic mucilage which, in humid weather, swells and oozes out of the ostiole, together with masses of detached conida (Fig. 1c). The majority of Lecanoromycetes produce microconidia (Fig. 1f–h) which serve as gametes in spermatization processes. Macroconidia for vegetative dispersal occur in few lichen-forming Arthoniomycetes (Vobis, 1980; Vobis & Hawksworth, 1981; Honegger, 1984a). Micro- and macroconidia differ in their dimensions and functions, but the process of conidiospore formation is the same.
The oldest conidia-bearing (nonlichenized) ascomycete, Palaeopyrenomycites devonicus, was found within the peripheral parts of the stems and rhizomes of the Early Devonian lycopod Asteroxylon mackei from the Rhynie Chert. This (or a co-occurring) ascomycete produces, beside sexual reproductive stages in perithecia, arthroconidia on hyphomycete-type conidiophores at the surface of the plant host (Taylor et al., 2005). A report on pycnidial conidiomata of the saprophytic or plant pathogenic Melanosphaerites devonicus on plant fossils from the Upper Devonian/Carboniferous Boundary (c. 350 Myr ago) of Bear Island (Grüss, 1928) remains very doubtful, the dimensions, structure and mode of spore release being characteristic of Chytridiomycota (details in Notes S2); these are well documented in the Lower Devonian (Taylor et al., 1992). Three ascomycete species with characteristic pycnidial conidiomata were found on cycadeoidalean cones in the Mid to Late Cretaceous of Japan (c. 90 Myr old; Watanabe et al., 1999) and on leaf surfaces in Dominican and Mexican amber of the Upper Oligocene to Lower Miocene (referred to as Coelomycetes, an obsolete term for pycnidia-producing, sterile fungi; Poinar, 2003).
The conglutinate fungal cortex of lichen thalli
A tissue-like zone, built up by fused (conglutinate) hyphae, is a characteristic feature of lichen thalli with internal stratification. Hyphae switch from one cell to the next from a filamentous growth pattern, as seen in the medullary and photobiont layers, to tissue-like growth, as typically seen in peripheral cortical layers. The hydrophilic cortex passively absorbs water and dissolved nutrients; it is brittle and opaque in the desiccated state, but elastic and translucent when fully hydrated. Thus the cortex plays important roles in water relations and light transmission to the photobiont layer and provides the thallus with mechanical stability. In many extant taxa the peripheral fungal cortex is brightly coloured as a result of secondary metabolite deposition. Upon extracellular crystallization most of these polyphenolics are water-insoluble, absorb UV light and transmit longer wavelengths to the thalline interior, thus acting as potent sunscreens (for reviews, see Honegger, 2001, 2009).
The upper cortex is well preserved in both fossils; it might have fulfilled the same functions as in extant taxa. A lower cortical layer is missing. It might have been either not differentiated or lost. Extant lichen-forming fungi with internally stratified thalli differentiate a cortex on: (1) on the upper surface only; examples are the placodioid taxa which fix themselves tightly to the substratum, or squamulose taxa (e.g. Cladonia spp., Lecanorales) or foliose Peltigera spp. (Peltigerales) among the Lecanoromycetes; these morphotypes are common and widespread among extant rock and soil inhabiting lichens; (2) on the lower surface only; examples are Peltula spp. (Lichinales) among the Lichinomycetes; or (3) on the whole thallus surface, as typically seen in numerous foliose and fruticose Lecanoromycetes. Many structurally similar fragments, albeit with no distinctly recognizable remains of photobiont cells, were found in numerous samples from the same collecting site, and also in much older samples from the Upper Silurian (D. Edwards, work in progress).
Cyanobacterial and algal photobionts
Extant Nostoc commune is a cosmopolitan species which tolerates temperature extremes, desiccation and high solar irradiation. Related Nostoc strains of clade II sensu O'Brien et al. (2005) are symbionts of a wide range of lichen-forming fungi and of the Glomeromycete Geosiphon pyriforme, the only known extant fungus with a cyanobacterial endosymbiosis (Redecker & Raab, 2006); these fungi receive fixed C and N from their cyanobacterial partner. Nostoc strains associate with a wide range of plants from hornworts and liverworts to ‘giant rhubarbs’ (Gunnera spp.), which benefit from fixed N as released by their diazotrophic cyanobacterial partner (Rai et al., 2002; O'Brien et al., 2005; Papaefthimiou et al., 2008). Part of the ecological success of extant, free-living and symbiotic Nostoc spp. is attributable to their massive, hygroscopic extracellular glycan sheath and their ability to synthesize, in response to UV irradiation, efficient ‘sunscreen pigments’ such as the lipid-soluble, heterocyclic indole-alkaloid scytonemin and/or water-soluble mycosporine-like amino acids, which also serve as osmolytes (Garcia-Pichel & Castenholz, 1991; Büdel et al., 1997; Wright et al., 2005; Fleming & Castenholz, 2007; Soule et al., 2007, 2009; Wang et al., 2010b); both types of photoprotective pigments are secreted and accumulate in the cyanobacterial glycan sheath.
The cell shape and overall morphology of the cyanobacterial photobiont of Cyanolichenomycites devonicus strongly resembles extant representatives of the genus Nostoc Vaucher ex Bornet et Flahault, but a strong morphological similarity with marine, Precambrian Palaeonostoc spp. (Sastri et al., 1972; Nautiyal, 1988) is also evident. The extremely slow evolutionary rates of cyanobacteria are discussed by Schopf (1994). Lichen thalli apparently provided an important environment for the genetic diversification of Nostoc strains at a global scale, as concluded from analyses of the gene cluster encoding the enzyme complex for microcystin production (Kaasalainen et al., 2012).
The excellent preservation of the gelatinous sheath of the fossil, Nostoc-like cyanobacterial photobiont of Cyanolichenomycites devonicus (Fig. 2c–e) and its retention in extant symbiotic and free-living Nostoc spp. during the charcoalifying process (Fig. 3b,c,e–h) are particularly interesting. This extracellular glycan sheath facilitates the formation of extensive colonies. It can dry out completely, thus giving the colony a brittle, papery appearance, but absorbs huge amounts of water upon rehydration, swells dramatically and becomes viscous-slimy in the fully hydrated state. It binds heavy metals and protects the cyanobacterial colony from microbial degradation (Helm et al., 2000). The gelatinous sheath of Nostoc spp. is built up by a dense meshwork of a fibrillar glucan (Honegger, 1982; Fig. 3e–h) with unique chemical composition. The glycan sheath of Nostoc commune comprises a 1-4-linked xylogalactoglucan backbone with D-ribose and nosturonic acid (Helm et al., 2000). The present structural observations on the glycan sheaths of fossil and extant Nostoc spp. hopefully facilitate the recognition of this type of cyanobacteria and their interactions with fungi and plants in fossil records.
Three different types of cell wall were found among the green algal photobionts of extant lichen-forming fungi. Most common and widespread are cellulosic walls whose structure and composition resemble those of the walls of parenchymatous plant cells. This type of cell wall is a characteristic of Trebouxia and Trentepohlia photobionts (Honegger, 1984b; Brunner & Honegger, 1985). Photobionts of the genera Coccomyxa and Elliptochloris have thin walls with a characteristic, trilaminar outermost wall layer with a membrane-like appearance at TEM level. The trilaminar wall layer comprises algaenan, an enzymatically nondegradable, sporopollenin-like biopolymer (Honegger & Brunner, 1981; Brunner & Honegger, 1985) which is synthesized also by numerous nonsymbiotic green algal representatives of the Trebouxiophyceae (Kodner et al., 2009), the oil shell producing Botryococcus braunii included (Audino et al., 2002). In marked contrast to the cellulosic mother cell walls of Trebouxia spp., which are enzymatically degraded after autospore formation, the nondegradable, trilaminar wall layer of Coccomyxa and Elliptochloris spp. is retained, in the free-living state as well as within the lichen thallus. The third wall type among unicellular, green algal lichen photobionts is structurally similar to the second, but lacks nondegradable material in the trilaminar outermost wall layer. This wall type was found in the genera Dictyochloropsis and Myrmecia (Brunner & Honegger, 1985). As algaenan-containing algal walls fossilize (see Botryococcus braunii), the unicellular, presumed green algal photobiont of Chlorolichenomycites had a cell wall of either the first or the last wall type as described above.
Lichens: common and widespread in Devonian cryptogamic ground covers?
How abundant were lichens in Lower Devonian ecosystems? It seems highly likely that various types of lichens colonized rock or soil surfaces in suitable climatic areas. Extant terrestrial vegetation zones in Arctic, Antarctic, Alpine, desert and steppe ecosystems are lichen-dominated wherever the competitive pressure of tracheophytes is low or absent as a result of harsh environmental conditions. Lichens, together with free-living fungi, cyanobacteria, algae and bryophytes, are integral elements of the ecologically immensely important soil crust communities (Belnap et al., 2001); these amount to > 12% of terrestrial ecosystems, their dimensions and development being monitored with remote sensing technologies (Karnieli et al., 2001; Weber et al., 2010). The global impact of photoautotrophic cryptogamic ground covers on CO2 and N2 fixation is currently being explored (Elbert et al., 2012). Lichenized and nonlichenized fungi, together with free-living cyanobacteria and algae, might have formed extensive, species-rich terrestrial rock surface and soil crust communities before and at the advent of plants (Heckman et al., 2001; Tomescu et al., 2006; Retallack, 2009). Their abundance might be indicated by the large quantities of nonvascular plant fragments present in Silurian and Lower Devonian macerates (Edwards, 1982, 1986; Wellman, 1995).
Lichens were not the only fungal symbiosis with photoautotrophs in Early Devonian ground covers. The detection of three plant genes involved in mycorrhiza formation events, whose functions are largely conserved from the common ancestor of land plants to extant bryophytes, pteridophytes, gymnosperms and angiosperms (Wang et al., 2010a), supports the hypothesis that mutualistic fungal associations played a crucial role during terrestrialization of land plants (Taylor & Osborn, 1996; Selosse & Le Tacon, 1998). Fossil spores of Glomeromycota (i.e. arbuscular mycorrhizal fungi (AMF)), the key players in these interactions, are known from the Ordovician (Redecker et al., 2000), and beautifully preserved arbuscules, the characteristic haustoria of AMF, were found in axes of Lower Devonian Rhyniales (Rhynie Cherts; Taylor et al., 1995b). Fossil records of bryophyte–AMF associations are lacking, but extant representatives of Glomeromycota associate with various liverworts (Marchantiophyta) and hornworts (Anthocerophyta). As the fungal symbionts of extant, basal thalloid liverworts do not belong to the Glomeromycota, but to the even more ancient Mucormycotina (Endogone spp.), the latter were hypothesized to have played a prime role during terrestrialization of land plants (Bidartondo et al., 2011). Did Mucormycotina form lichens in the Early Devonian not only in aquatic ecosystems, as in Winfrenatia reticulata (Taylor et al., 1995a, 1997; Karatygin et al., 2009), but also in terrestrial soil crust communities?
As a consequence of their striking structural similarity to extant taxa, C. devonicus and C. salopenis were relatively easily recognized as lichens. However, there are numerous microbial, fungal and lichen fossils with structures, shapes and anatomies unknown among extant taxa, many of them being poorly preserved; examples are the extensive microbial crusts which covered Australian palaeosols in the Mid Cambrian (Retallack, 2011). Did fungi ever form stems of several metres height or length, as postulated for the enigmatic, Early Devonian Prototaxites (Hueber, 2001; Selosse, 2002)? If so, what was their C source: detritus or photosynthates derived from photoautotrophic symbionts? What might these symbionts have been: cyanobacteria, algae or both? How could mechanical stability be generated and water relations controlled in such impressive structures? Lots of very interesting questions remain to be answered by palaeolichenologists, palaeomycologists and palaeomicrobiologists.
This research was financed by a Leverhulme Trust grant to D.E. and by a Swiss National Science Foundation grant (Nr. 3100A0-116597) to R.H., which are gratefully acknowledged. Our sincere thanks are due to Peter Fisher for SEM expertise, to the librarian Martin Spinnler for assistance and to Marc-André Selosse and four anonymous referees for their helpful comments on an earlier version of this manuscript.