Mycological Progress

, Volume 5, Issue 3, pp 178–184

Germination shields in Scutellospora (Glomeromycota: Diversisporales, Gigasporaceae) from the 400 million-year-old Rhynie chert


  • Nora Dotzler
    • Bayerische Staatssammlung für Paläontologie und Geologie und GeoBio-Center LMU
    • Department Biologie I und GeoBio-Center LMU, Bereich Biodiversitätsforschung: MykologieLudwig-Maximilians-Universität München
    • Bayerische Staatssammlung für Paläontologie und Geologie und GeoBio-Center LMU
    • Department of Ecology and Evolutionary Biology, and Natural History Museum and Biodiversity Research CenterThe University of Kansas
  • Thomas N. Taylor
    • Department of Ecology and Evolutionary Biology, and Natural History Museum and Biodiversity Research CenterThe University of Kansas
  • Reinhard Agerer
    • Department Biologie I und GeoBio-Center LMU, Bereich Biodiversitätsforschung: MykologieLudwig-Maximilians-Universität München
Original Article

DOI: 10.1007/s11557-006-0511-z

Cite this article as:
Dotzler, N., Krings, M., Taylor, T.N. et al. Mycol Progress (2006) 5: 178. doi:10.1007/s11557-006-0511-z


Glomeromycotan spores from the Lower Devonian Rhynie chert provide the first evidence for germination shields in fossil fungi and demonstrate that this complex mode of germination was in place in some fungi at least 400 million years ago. Moreover, they represent the first direct marker relative to the precise systematic position of an Early Devonian endomycorrhizal fungus. In extant fungi, germination shields occur exclusively in the genus Scutellospora (Glomeromycota: Diversisporales, Gigasporaceae). These structures are regarded as a derived feature within the phylum Glomeromycota, and hence their presence in the Rhynie chert suggests that major diversification within this group of fungi occurred before the Early Devonian.


Arbuscular mycorrhizaEvolutionGerminationPragian (Early Devonian)Spore wall


The Early Devonian Rhynie chert has provided a wealth of insights into the diversity of fungal life some 400 million years ago. As a result, members of the Chytridiomycota, Zygomycota, Glomeromycota and Ascomycota are today known in great detail from this palaeoecosystem; many of the Rhynie chert fungi have also been demonstrated in various interactions with other organisms (summarised in Taylor et al. 2004 and Taylor and Krings 2005). Among these interactions are several examples of arbuscular mycorrhizae (e.g. Remy et al. 1994; Taylor et al. 1995, 2005). The fungal partners consist of aseptate hyphae that enter the rhizomatous axes of various land plants and extend through the intercellular system of the outer cortex. The formation of intracellular arbuscules within a well-defined region of the cortex (termed the ‘mycorrhizal arbuscule zone’) substantiates these fungi as endomycorrhizal. The mycorrhizal fungi from the Rhynie chert have been assigned to the Glomeromycota based on structural similarities to extant representatives in this phylum of fungi and corresponding mode of mycorrhiza formation (cf. Taylor et al. 1995; Helgason and Fitter 2005); however, to date, none has been classified in greater detail and related to a particular modern taxon.

Glomeromycota is a monophyletic phylum that includes the arbuscular mycorrhizal (AM) fungi (Schüßler et al. 2001; Helgason et al. 2003; Corradi et al. 2004). It is estimated that more than 80% of vascular plants today live in symbiosis with these fungi (Smith and Read 1997). Molecular clock estimates based on amino acid sequences suggest that the Glomeromycota separated from other fungal groups ~1,200 myBP (Heckman et al. 2001); more conservative estimates place the divergence at about 600–700 myBP (Berbee and Taylor 2001). Despite the proposed antiquity of the Glomeromycota, early fossil representatives are rare. Thick-walled spores suggested to be those of Glomeromycota have been reported from Precambrian sediments (Pirozynski and Dalpé 1989 and references therein). Similar spores are known from the Ordovician of North America (Redecker et al. 2000, 2002). In none of these accounts, however, is there any information regarding associations with plants. Glomeromycotan spores in tissues of late Palaeozoic land plants have been described from the Upper Devonian of Canada (Stubblefield and Banks 1983) and the Carboniferous of North America (Wagner and Taylor 1982). Moreover, hyphae or hyphae-like structures, aggregated in cortical tissues of the underground parts of several Carboniferous plants have variously been interpreted as arbuscules (e.g. Weiss 1904; Osborn 1909; Halket 1930; Agashe and Tilak 1970); however, most of these reports have later been questioned and the structures re-interpreted as non-fungal (coalesced cell contents) or non-mycorrhizal (cf. Stubblefield and Taylor 1988). The earliest persuasive evidence for glomeromycotan mycorrhizae in seed plants occurs in the form of non-septate hyphae, vesicles, arbuscules and clamydospores in silicified roots of the Triassic cycad Antarcticycas schopfii Smoot, T.N. Taylor and Delevoryas (Stubblefield et al. 1987; Phipps and Taylor 1996). Thus, the Rhynie chert mycorrhizae represent the earliest fossil evidence for Glomeromycota in symbiosis with land plants (Remy et al. 1994; Redecker 2002).

Also present in the Rhynie chert are different types of fungal spores. However, these remains have been largely neglected as a source of information about the diversity of fungi in this palaeoecosystem. In this study, we describe glomeromycotan spores from partially degraded axes of the Rhynie chert land plant Asteroxylon mackiei Kidst. and W.H. Lang that display a complex mode of germination involving the formation of a germination shield. The fossil spores can be directly related to a particular modern taxon because, in extant fungi, this mode of germination is restricted to species in the genus Scutellospora C. Walker and Sanders (Glomeromycota: Diversisporales, Gigasporaceae).

Materials and methods

The Rhynie chert Lagerstätte, an in situ silicified Early Devonian palaeoecosystem, is located in the northern part of the Rhynie outlier of Lower Old Red Sandstone in Aberdeenshire, Scotland. The cherts occur in the upper part of the Dryden Flags Formation, in the so-called Rhynie Block, a few hundred metres northwest of the village of Rhynie. The Lagerstätte consists of at least 10 fossiliferous beds containing lacustrine shales and cherts that are interpreted as a series of ephemeral fresh water pools within a hot springs environment. The chert-bearing formation is Pragian in age and has been radiometrically dated to 396±12 Ma. Detailed information about the geology and palaeontology of the Rhynie chert Lagerstätte can be found in Trewin and Rice (2004).

Spores were identified in petrographic thin-sections prepared by cementing a thin wafer of the chert to a glass slide and then grinding the rock slice with silicon carbide powder until it becomes sufficiently thin for examination in transmitted light (cf. Hass and Rowe 1999). Slides are deposited in the Bayerische Staatssammlung für Paläontologie und Geologie, Munich (Germany), under accession numbers BSPG 1964 XX 625 and 631, and BSPG 1964 XX 31.003, 31.005 and 31.006. For comparison, spores of Scutellospora castanea C. Walker (BEG 1: produced on onion, in neutral soil) were fixed and embedded in PVLG (Polyvinyllactoglycerol) according to a procedure outlined in Walker (1979) and studied in transmitted light.


Spores with germination shield occur in cortical tissues of partially degraded axes of the early lycophyte Asteroxylon mackiei; a total of 12 specimens of this type of spore have been discovered to date. Spores are globose to subglobose, 260–350 μm in diameter and possess a non-ornamented surface (Fig. 1a). The spore wall is subdivided into two wall groups. The outer wall group is well preserved, up to 18 μm thick and two- or three-layered. A distinct dark layer (~3 to 4 μm thick; cf. arrow in Fig. 1a) occurs on the inner surface of the outer wall group. This layer either represents the inmost part of the outer wall group or the outmost layer of the inner wall group. The original thickness and composition of the inner wall group are difficult to estimate. A translucent region, up to 30 μm thick, occurs between the dark layer and outer surface of the spore lumen. It is not entirely clear, however, whether this region represents the original thickness of the inner wall group or is the result of shrinkage of both the inner wall group layers and spore lumen during fossilisation. The germination shield extends along the inner surface of the dark layer (Fig. 1a–e); the original location of the shield is difficult to reconstruct due to the lack of details about the composition of the inner wall group. The germination shield is round or oval in outline, ~140 μm in diameter and up to 15 μm high. It is distinctly lobed, with each of the lobes 25–33 μm wide, or displays a complex infolding along the margins (appearing in section to consist of compartments, cf. Fig. 1a,b). In one of the spores, the connection of the germination shield to the spore lumen is apparent (Fig. 1a–c). Another specimen shows what we interpret as a germ tube that is formed by the germination shield and penetrates the outer wall group (Fig. 1e). Unfortunately, none of the spores with germination shield displays a subtending hypha or suspensor-like base. However, in one of the A. mackiei axes, a structurally similar but somewhat smaller (i.e. 190×150 μm in diameter) spore without germination shield occurs that is attached to a slightly bulbous subtending hypha (Fig. 1f), which is up to 18 μm in diameter and remotely resembles the characteristic suspensor-like base seen in extant Scutellospora species (e.g. Fig. 1g).
Fig. 1

Scutellosporites devonicus Dotzler et al. from the Rhynie chert (af) and spores of the extant Scutellospora castanea C. Walker (gk). a Section through a fossil spore with germination shield. Arrow indicates the distinct dark layer that either represents the inmost portion of the outer wall group or outmost layer of the inner wall group. Bar=50 μm. b Detail of a, focusing on the germination shield. Bar=35 μm. c Germination shield in near median longitudinal section. Bar=35 μm. d Germination shield in oblique surface view, showing lobes/infoldings along the margin. Bar=20 μm. e Germ tube penetrating the outer wall group. Bar=30 μm. f Slightly bulbous base of a smaller glomeromycotan spore in Asteroxylon mackiei. Bar=20 μm. g Same as f, but from the extant S. castanea. Bar=30 μm. h Spore of S. castanea with germination shield. Bar=50 μm. i Detail of h, focusing on the germination shield. Bar=30 μm. j Germination shield of a second S. castanea spore in optical longitudinal section. Bar=30 μm. k Germination shield in oblique surface view. Arrows indicate the margins of the shield. The infoldings are visible as narrow dark lines. Bar=30 μm


The fossil spores are similar to spores produced by species in the extant genus Scutellospora (Glomeromycota: Diversisporales, Gigasporaceae). Scutellospora consists of some 20 species of arbuscular mycorrhizal fungi, all of which produce large spores (between 120 and 640 μm in diameter) with multi-layered walls. Germination includes the formation of a germination shield, which is a specialised structure that distinguishes Scutellospora from the closely related genus Gigaspora Gerd. and Trappe (Walker and Sanders 1986) and all other members of the Glomeromycota. There are several other genera within the Glomeromycota, e.g. Acaulospora Gerd. and Trappe (Diversisporales, Acaulosporaceae) and Pacispora Oehl and Sieverd. (Glomerales, Glomeraceae), in which spore germination also includes the formation of a specialised structure between two layers of the spore wall (e.g. Stürmer 1998; Stürmer and Morton 1999; Oehl and Sieverding 2004). However, this structure, termed the ‘germination orb’, is more delicate, usually smaller (e.g. 14–26×20–38 μm in P. franciscana Oehl and Sieverd., cf. Oehl and Sieverding 2004), less complex and clearly distinguishable morphologically from the germination shields produced by the Rhynie chert spores and extant Scutellospora. It is still being debated whether germination orbs and germination shields are heterologous structures or synapomorphies of the Diversisporales lineage.

For comparison of the fossils with extant representatives of Scutellospora, specimens of Scutellospora castanea C. Walker were analysed (Fig. 1h–k). A complete description of S. castanea is provided in Walker et al. (1993) and we restrict our discussion to a brief characterization of the germination shield: In S. castanea, this structure is oval in outline (Fig. 1k) and occurs on the inner spore wall group. It is up to 210 μm long, 185 μm wide, 10–15 μm high and characterised by a complex infolding along the margins (appearing in optical section to consist of compartments, cf. Fig. 1h–j). Optical longitudinal sections through shields of S. castanea are virtually indistinguishable from sections through the fossil germination shields (compare Fig. 1a,b with Fig. 1i,j).

Based on the striking similarities in germination shield morphology between the fossils and S. castanea, as well as other species of Scutellospora detailed in the literature (e.g. Koske and Walker 1986; Walker and Sanders 1986; Walker and Diederichs 1989; Walker et al. 1998; Herrera-Peraza et al. 2001), we interpret the fossils as belonging to an early member of the genus Scutellospora. However, the fossil spores only provide an incomplete picture of this fungus. For example, none of the spores with germination shields display a subtending hypha or specialised base. Because extant Scutellospora spores are always borne on a characteristic bulbous, suspensor-like base (Fig. 1g), documentation of this feature would strengthen the proposed affinities of the fossil spores. A somewhat smaller fossil spore (lacking germination shield), which co-occurs with the large spores with germination shield, displays a slightly bulbous base (Fig. 1f). We cannot establish at present whether this spore belongs to the fungus that produced the spores with germination shield. As a result, we refrain from including the fossil spores with germination shield in Scutellospora, but rather introduce a new genus, for which the name Scutellosporites is proposed.


Glomeromycota C. Walker and A. Schüßler

Diversisporales C. Walker and A. Schüßler

Gigasporaceae J.B. Morton and Benny

Scutellosporites Dotzler, M. Krings, T.N. Taylor and Agerer, gen. nov.

Derivation of generic name. The name underscores the similarity to the extant genus Scutellospora; the ending -ites is used to designate a fossil taxon.

Generic diagnosis. Spores globose to subglobose, up to 350 μm in diameter, with non-ornamented surface; spore wall composed of two wall groups; outer wall group >15 μm thick, two- or three-layered; distinct dark layer present on inner surface of outer wall group; germination by means of germination shield extending along inner surface of dark layer; shield round or oval, >100 μm long and >10 μm high, distinctly lobed or with infolded margins.

Type species. Scutellosporites devonicus Dotzler et al.

Scutellosporites devonicus Dotzler, M. Krings, T.N. Taylor and Agerer, spec. nov. Fig. 1a–f

Specific diagnosis. As for the genus

Derivation of specific epithet. Indicating the geologic age of the fossil.

Holotype. BSPG 1964 XX 631 (Fig. 1a in this paper)

Type locality. Rhynie, Aberdeenshire, Scotland, National Grid Reference NJ 494276

Age and stratigraphic position. Early Devonian (Pragian, ~400 myBP)

Remark. Glomeromycotan spores that resemble Scutellosporites devonicus have been described from degraded tissues of various Rhynie chert plants by Kidston and Lang (1921) as Palaeomyces gordoni Kidst. and W.H. Lang. However, in none of these spores is a germination shield obvious. Because it is most likely that there existed more than one taxon of mycorrhiza-forming glomeromycotan fungi in the Rhynie chert, we refrain from assigning the spores with germination shields specifically to P. gordoni.


One of the remarkable discoveries in the Early Devonian Rhynie chert is the presence of arbuscular mycorrhizae that are strikingly similar to mycorrhizae today and were produced by the same group of fungi, i.e. members of the Glomeromycota (Remy et al. 1994; Taylor et al. 1995; Helgason and Fitter 2005). Despite the detailed analyses that have been carried out on these ancient mycorrhizae, an exact systematic placement of the fungal partners has not been possible to date, due primarily to the fact that diagnostic features necessary in establishing the affinities of a glomeromycotan fungus (e.g. spore wall structure and colour, auxilliary cells) could not be determined with the fossils.

The prominent germination shields described in this study correspond to germination shields produced by the extant Gigasporaceae genus Scutellospora, and thus represent the first direct diagnostic marker that can be used to determine the systematic position of one of the Rhynie chert mycorrhizal fungi. In extant Glomeromycota, prominent and well-recognizable germination shields are known to occur exclusively in Scutellospora. Similar pre-germination structures (germination orbs) found in genera such as Pacispora and Acaulospora are much more delicate and become rarely visible, even in broken specimens or after specific preparations of the inner wall (Spain 1992; Oehl and Sieverding 2004). Members of Scutellospora display a complex mode of germination, in which, before germ tube formation, a germination shield is developed between two layers of the spore wall. The position of the germination shield varies between species of Scutellospora and may occur between the individual layers of the inner wall group (e.g. in S. scutata C. Walker and Dieder., cf. Walker and Diederichs 1989) or on the surface of the inmost wall layer (e.g. in S. castanea, cf. Walker et al. 1993). At maturity, the germination shield produces one to several germ tubes that penetrate the outer portion of the spore wall (Walker and Sanders 1986). A satisfactory interpretation with regard to the nature and function of the germination shields has not been published to date. One interpretation is that they are either sexual or parasexual, or perhaps asexual vestiges of some previously sexual structure (C. Walker, personal communication).

Because germination shields represent complex structures that consistently occur between distinct layers of the spore wall, this feature is regarded as derived within the Glomeromycota (Bentivenga and Morton 1996). As a result, the presence of spores with germination shield in the Rhynie chert suggests that major diversification within this group of fungi occurred before the Early Devonian. It is interesting to note, however, that the derived state of the germination shield in the family Gigasporaceae (i.e. Gigaspora and Scutellospora) has been questioned based on molecular studies (Simon et al. 1993; Redecker 2002). These authors hypothesise that Gigaspora is an advanced rather than a plesiomorphic genus; species in Gigaspora form a very narrow clade compared to the large variation within Scutellospora (Schwarzott et al. 2001). It has also been suggested that Scutellospora may be paraphyletic (Redecker 2002). Berbee and Taylor (2001) estimate the divergence time between Gigaspora and two species in the genus Glomus Tul. and C. Tul. at approximately 300 myBP based on a nucleotide substitution rate of 1.26%. Although Scutellospora was not included in their data set, the occurrence of spores with germination shields in deposits that are 100 million years older than the estimated divergence of Gigaspora from other Glomeromycota supports the hypothesis that the germination shield is an ancestral feature within the Gigasporaceae. If in fact Gigaspora is advanced, the mode of germination involving a germination shield was lost during the evolution of this genus (Redecker 2002). In addition, the complex system of spore walls composed of one to several wall groups seen in Scutellospora was also lost because members of Gigaspora display a much simpler wall organisation (Walker and Sanders 1986). The inner wall group in Scutellosporites devonicus is not preserved, and thus its original thickness is difficult to estimate. However, we suggest that the inner wall group was quite massive because one of the spores (the holotype specimen) clearly shows that the germination shield does not occur close to the surface of the spore lumen, but rather appears stalked (Fig. 1a). This suggests that the shield had to pass through a massive inner wall group before extending along the inner surface of the dark layer. The fact that the surface boundary lines of both the spore lumen and erect portion of the shield (‘stalk’) are not wrinkled or otherwise unnaturally distorted (cf. Fig. 1a–c) indicates that the erect portion of the shield does not represent a preservational artefact. As a result, this feature substantiates that the inner wall group was of considerable thickness because the ‘stalk’ of the germination shield could not be explained if the inner wall group were only a few micrometre thick. In extant representatives of Scutellospora the inner wall group is usually only 0.6–2 μm thick (INVAM homepage). However, for a few species, up to 18 μm thick inner wall groups have been recorded, but these are based on material mounted in PVLG, which results in expansion of the wall (e.g. from 2 to 15 μm within a few minutes in S. spinosissima C. Walker and Cuenca, cf. Walker et al. 1998). The considerable thickness of the inner wall group of S. devonicus in comparison to that seen in extant Scutellospora suggests that perhaps the inner wall group was gradually reduced and eventually lost in Gigaspora.

Time estimates for the appearance of individual lineages and taxa within the Glomeromycota are typically based on molecular and genetic studies of modern taxa. In many cases the more general results from these studies are supported by the fossil record (e.g. Simon et al. 1993; Helgason and Fitter 2005; Taylor and Krings 2005). At a finer scale of resolution, however, the fossil record has to date mostly failed in producing suitable evidence in support for or against hypotheses based on molecular data, due primarily to the inherent incompleteness of the fossil record. As a result, Scutellosporites devonicus from the Rhynie chert is an important discovery because it displays the first direct marker that can be used to establish the precise systematic position of an Early Devonian mycorrhizal fungus. As the molecular phylogeny of the Glomeromycota is continuously refined, it will be interesting to see how the characters attributed to S. devonicus fit character states based on molecular data.


This study was supported in part by funds from the Alexander von Humboldt-Foundation (V-3.FLF-DEU/1064359 to M.K.) and the National Science Foundation (EAR-0542170 to T.N.T. and M.K.). We thank A. Schüßler and C. Walker for providing valuable information that contributed to this study and two anonymous reviewers for their insightful comments and suggestions.

Copyright information

© German Mycological Society and Springer 2006