Archaeosporites rhyniensis gen. et sp. nov. (Glomeromycota, Archaeosporaceae) from the Lower Devonian Rhynie chert: a fungal lineage morphologically unchanged for more than 400 million years

Abstract

Background and Aims

Structurally preserved arbuscular mycorrhizas from the Lower Devonian Rhynie chert represent core fossil evidence of the evolutionary history of mycorrhizal systems. Moreover, Rhynie chert fossils of glomeromycotan propagules suggest that this lineage of arbuscular fungi was morphologically diverse by the Early Devonian; however, only a small fraction of this diversity has been formally described and critically evaluated.

Methods

Thin sections, previously prepared by grinding wafers of chert from the Rhynie beds, were studied by transmitted light microscopy. Fossils corresponding to the description of Archaeospora spp. occurred in 29 slides, and were measured, photographed and compared with modern-day species in that genus.

Key Results

Sessile propagules <85 µm in diameter, some still attached to a sporiferous saccule, were found in early land plant axes and the chert matrix; they developed, in a similar manner to extant Archaeospora, laterally or centrally within the saccule neck. Microscopic examination and comparison with extant fungi showed that, morphologically, the fossils share the characters used to circumscribe the genus Archaeospora (Glomeromycota; Archaeosporales; Archaeosporaceae).

Conclusions

The fossils can be assigned with confidence to the extant family Archaeosporaceae, but because molecular analysis is necessary to place organisms in these taxa to present-day genera and species, they are placed in a newly proposed fossil taxon, Archaeosporites rhyniensis.

INTRODUCTION

Members of the phylum Glomeromycota (Schüßler et al., 2001; Tedersoo et al., 2018; Naranjo-Ortiz and Gabaldón, 2019), the fungal lineage sometimes also referred to as Glomeromycotina (Spatafora et al., 2016), form intimate partnerships with plants, known as arbuscular mycorrhizas. Research on the evolutionary history of these fungi includes morphological, molecular and genetic data (Redecker and Raab, 2006; Redecker et al., 2013; Kamel et al., 2017; Feijen et al., 2018); moreover, a well-preserved fossil record has contributed greatly to our understanding of these organisms in past environments (Taylor et al., 2015; Strullu-Derrien et al., 2018). The Early Devonian Rhynie chert from Aberdeenshire, Scotland, which originated in a continental palaeoecosystem that became petrified through inundation with silica-rich water from hot springs (Trewin and Kerp, 2017; Garwood et al., 2020), figures prominently in this context. It yields exquisitely preserved fossils of glomeromycotan arbuscular fungi in early land plants (Taylor et al., 1995, 2005) that provide core fossil evidence of the early evolutionary history of mycorrhizal systems (Parniske, 2008; Martin et al., 2017; Brundrett et al., 2018).

The arbuscular mycorrhizal fungi (AMF) comprise the glomeromycotan arbuscular fungi (GAF) and mucoromycotinan arbuscular fungi (MAF) or ‘fine root endophytes’, such as Planticonsortium tenue (Walker et al., 2018a). Until molecular analysis was introduced for AMF classification at the turn of the 21st century (Kramadibrata et al., 2000; Morton and Redecker, 2001), AMF in general, and GAF in particular, have traditionally been defined by mode of spore formation and characteristic features, such as spore-wall architecture and ornamentation (Walker et al., 2018b). Conversely, present-day species descriptions rely on a combination of morphological and molecular evidence, particularly as some groups, a particular example of which would be Archaeospora, cannot be separated to species without a phylogenetic analysis from study of the rRNA gene (Schüßler and Walker, 2019). Information on spore development and spore wall architecture is well preserved in numerous fossil glomeromycotan propagules from the Rhynie chert that are located within intact or partially degraded land plant axes and in the chert matrix (Kidston and Lang, 1921; Taylor et al., 1995; Dotzler et al., 2006, 2009; Karatygin et al., 2006; Krings et al., 2015, 2017a, b; Harper et al., 2017). The variety of glomeromycotan spores described and illustrated to date from the Rhynie chert shows that GAF were morphologically diverse, at the ordinal systematic level, by the Early Devonian. Moreover, they have pushed back the evolutionary origin of the main categories of spores in the Glomeromycota (i.e. glomoid, gigasporoid, acaulosporoid) to a time before the evolution of true roots (Dotzler et al., 2009). Many of the Rhynie chert spores are so similar to extant fungi that, were it not for the lack of molecular evidence, they could readily be assigned to extant lineages (Brundrett et al., 2018).

Among the extant lineages, the Archaeosporaceae is one of the earliest-diverging clades in the Glomeromycota (Schüßler et al., 2001; www.amf-phylogeny.com). The most obvious characteristic of members of this family is the production of small (~23–110 µm in diameter), colourless, more or less globose acaulosporoid propagules, which are formed within the neck or stalk of a colourless, thin-walled, terminal balloon-like structure (Fig. 1A), termed the sporiferous saccule (Walker et al., 1984). The saccule usually collapses, and normally the spore becomes detached during extraction from the soil, leaving little evidence of its developmental origins (Hafeel, 2004), though collapsed or remnant saccules may remain attached (Fig. 1B).

Fig. 1.

Extant Archaeospora trappei acaulospore development. (A) Sporiferous saccule (s) budding (arrow) prior to spore formation; cytoplasm is transferring into the enlarging bud. (B) Developed acaulospore (a) still attached to the saccule (s), showing wall components (C1, C2); W5983.

To determine whether fossil Archaeosporaceae-like spore–saccule complexes occur in the Rhynie chert, we re-examined all slides used by Dotzler et al. (2009), as well as similar additional material. We found evidence of the frequent occurrence of Archaeospora-like fossils in the matrix, in decaying plant remains, and in largely intact axes of Rhynie chert land plants, occasionally alongside much larger spore–saccule complexes of the unnamed type described by Dotzler et al. (2009). This paper formally describes the new form as a new fossil genus and species for which the name Archaeosporites rhyniensis is proposed; its discovery contributes to our understanding of morphological diversity of glomeromycotan fungi in early terrestrial ecosystems.

MATERIALS AND METHODS

Geological settings

The Rhynie chert Lagerstätte, located in the northern part of the Rhynie outlier of Aberdeenshire, Scotland, includes series of chert lenses that have preserved even the finest details of early terrestrial plants, animals and a variety of microorganisms, especially fungi (Rice et al., 2002; Trewin and Kerp, 2017; Garwood et al., 2020). The chert lenses are interpreted as having accumulated on an alluvial plain associated with ephemeral ponds and lakes in a geothermal wetland (Channing and Edwards, 2009). The chert was formed by rapid infusion of hot, alkaline, silica-rich water that percolated to the surface and inundated the ecosystem, bringing life to an abrupt halt, and providing a remarkable fossil record of the life forms existing at the time (Powell et al., 2000; Garwood et al., 2020). The Rhynie chert biota has been regarded as early (but not earliest) Pragian to earliest Emsian in age based on spore assemblages (Wellman, 2006, 2017; Wellman et al., 2006). An age estimate based on high-precision U-Pb dating of zircon and titanite from hydrothermally altered andesite indicates an absolute age of 411.5 ± 1.3 Ma for the Rhynie chert biota (Parry et al., 2011), while another age constraint using ⁴⁰Ar/³⁹Ar in K-feldspar from a quartz-feldspar vein that is part of the hydrothermal system responsible for the formation of the Rhynie chert yields a mean age (recalculated to be U-Pb-comparable) of the fossilized biota of 407.1 ± 2.2 Ma (Mark et al., 2011). However, the andesite cannot be fixed with certainty in the stratigraphic sequence and is certainly older than the hydrothermal alteration. As a result, the date estimate in Mark et al. (2011) likely gives a more accurate age of the hydrothermal system, and hence the age of the Rhynie chert biota. An absolute age of 411.5 ± 1.3 Ma is very close to the Lochkovian/Pragian boundary (410.8 ± 2.8 Ma), while the age suggested by Mark et al. (2011) would correspond approximately to the Pragian/Emsian boundary (407.6 ± 2.6 Ma), which concurs with the dispersed spore biostratigraphy by Wellman (2006, 2017) and Wellman et al. (2006).

Fossil material

Fossils used in this study were identified in thin sections prepared from several different chert blocks by cementing wafers of the chert to glass slides and then grinding the rock slices until the sections were thin enough to transmit light. Slides are deposited in the collection of the Forschungsstelle für Paläobotanik am Geologisch-Paläontologischen Institut, Westfälische Wilhelms-Universität, Münster, Germany, under accession numbers P3951–3958, 3967, 3978, 3980, 3981 and 3999, and in the Bayerische Staatssammlung für Paläontologie und Geologie (SNSB-BSPG) in Munich, Germany, under accession numbers SNSB-BSPG 1964 XX 625, 633; 1965 I 330; 2013 XIV 6; 2016 XII 6, 12, 13, 18, 19, 24, 26, 31, 34 and 37–40.

All fossils were photographed with brightfield illumination through a Leica DFC-480 camera on a Leica compound microscope, and processed in Adobe Photoshop CS6, or with Canon EOS 60D or 6D cameras through a Zeiss Axioskop or a Zeiss Photomicroscope III. Depth of field limitations required the use of image-stacking software (heliconsoft.com) to produce single images with the best possible focus depth. The images were displayed at 38 × 28 cm on a 1920 × 1200-pixel monitor and the dimensions determined with calibrated software calipers (iconico.com/caliper), each pixel representing 0.16 µm on the ×1000 images and 0.4 µm for the ×400 photographs. Because of the nature of the Rhynie chert thin sections, the measurements may be underestimates. Because they were fossilized in their natural position (in situ), it is not possible to be certain that the sections are exactly through the ‘great circle meridian’ (equator) of the specimens. Moreover, post mortem dehydration by silica may have caused some shrinkage (e.g. Figs 7 and 8A in Trewin and Fayers, 2015). For extant species, it is possible to focus through the specimens to find the best optical section so that the measurements are always at the equator.

Extant fungal material and methods

Acaulospores from extant members of the Archaeosporaceae were extracted from pot-culture substrate by swirling a 15- to 30-mL sample in water and, after allowing settling for a few seconds, decanting the supernatant through a 45-μm mesh sieve. For a set of samples, the flow through the 45-μm sieving was studied to check whether small spores with diameters below 40–45 μm were present in the cultures. No such spores could be detected. The contents were examined under a dissecting microscope with incident light. Specimens were selected with fine forceps and mounted on microscope slides in polyvinyl alcohol lactoglycerol, with or without the addition of Melzer’s reagent (5:1 v/v) and examined, after polymerization, by transmitted light microscopy (Zeiss Axioskop and Photomicroscope III), under brightfield or Nomarski differential interference contrast. Some specimens were measured directly with a calibrated eyepiece graticule, each division representing 1.6 (×40 objective) or 0.6 (×100 objective) µm. Staining with methyl blue, observations and subsequent descriptions followed established methods (Walker, 1983; Kramadibrata et al., 2000).

Voucher specimens examined.

Voucher specimens of extant species, mounted on microscope slides, are numbered (e.g. W5340) in the records of C. Walker. With the exception of Archaeospora spainiae, the holotype of which was loaned by Z + ZT (Universität Zürich), these are held in the herbarium of the Royal Botanic Garden Edinburgh (E). Herbarium acronyms follow Index Herbariorum (http://sweetgum.nybg.org/science/ih/). These included A. ecuadoriana, A. myriocarpa, A. schenckii, A. spainiae and A. trappei. For detailed information on voucher specimens, refer to Supplementary Data Table S1. The only other species currently in the genus, A. europaea, was published after this work was completed (Oehl et al., 2019).

Statistical methods

All statistical analysis was done in R (R Core Team, 2018). The data were first analysed by comparing the longest and shortest dimensions of the propagules across species. Data failed to meet the assumptions of multivariate normality and homogeneity of covariance matrices. Non-parametric inference of the multivariate data was made using the non-parametric comparison of multivariate samples (npmv) package (Burchett et al., 2017). Propagule volumes were calculated [as an ellipsoid (oblate spheroid) body, v = 4/3 π • b² • c] for all spores. These data also did not meet the assumptions of normality and homoscedasticity, and were analysed with the Kruskal–Wallis rank sum test.

TERMINOLOGY USED TO DESCRIBE PROPAGULE AND SACCULE MORPHOLOGY

The descriptive term ‘acaulosporoid spore’ was coined for propagules formed by species of Acaulospora, and translates to ‘spore without a stalk’ (Gerdemann and Trappe, 1974). Two species, Acaulospora laevis (the type species for the genus) and Acaulospora elegans, were initially included. Later authors added more species, some of which are now placed in different genera, orders and families. One of these, Acaulospora trappei, was transferred to the genus Archaeospora based primarily on analysis of the β-tubulin gene (Morton and Redecker, 2001). Later molecular analysis showed that, although having superficially similar propagule development, the Acaulosporaceae and Archaeosporaceae were widely separated phylogenetically at the ordinal level (Walker et al., 2007a). Another, more closely related, genus, Ambispora (Archaeosporales, Ambisporaceae), also produces acaulosporoid spores (Walker et al., 2007a).

The spores of Acaulospora laevis were originally described as developing laterally on the ‘neck’ or ‘stalk’ of the saccule (Gerdemann and Trappe, 1974), but they actually develop internally, within the saccule (Fig. 16 in Mosse, 1970). The particular position of the spore within the saccule neck has been used to apply the terms ‘acaulosporoid’, ‘entrophosporoid’ and ‘ambisporoid’, depending on whether they are produced laterally in a balloon-like protrusion, centrally in the saccule neck, or laterally in a pedicel-like process (Fig. 2A_1,2–C_1,2). Some develop, in the manner described for Archaeospora trappei, laterally in a ‘bud’ (Fig. 2A_1,2), i.e. in the acaulosporoid manner as described for members of the genus Acaulospora, while others form centrally by expansion of the saccule neck (Fig. 2C_1,2), i.e. the entrophosporoid manner described for Entrophospora infrequens (Ames and Schneider, 1979). Still others develop a protrusion (Fig. 2B_1,2), termed a ‘pedicel’ by Morton and Redecker (2001), which is similar to some members of Ambispora; such development could thus be called ‘ambisporoid’. The pedicel is a protrusion of the wall of a spore developing de novo within the saccule stalk, rather than continuous with a subtending hypha from which a chlamydospore develops, as in glomoid spores. All three of these patterns can be found within a single species of extant Archaeospora (e.g. Archaeospora ecuadoriana), and hence merely represent topological variations with no diagnostic value (Schüßler and Walker, 2019). Acaulosporoid propagules have erroneously been termed ‘azygospores’ in the past (e.g. Schenck and Smith, 1982). To distinguish this kind of propagule from the widely differing categories of spores in the Glomeromycota (Walker et al., 2018b), the more prescriptive terminology, ‘acaulospore’, will be used herein.

Fig. 2.

Three types of acaulospore development in the neck of sporiferous saccule in extant and fossil specimens, each with corresponding graphical representation. Extant specimen, Archaeospora ecuadoriana, W5340 (modified from Fig. 2a–c in Schüßler and Walker, 2019): (A_1,2) acaulosporoid; (B_1,2) ambisporoid; (C_1,2) entrophosporoid. Fossil specimens, Archaeosporites rhyniensis gen. et sp. nov.: (D_1,2) acaulosporoid; SNSB-BSPG 2016 XII 12 (D₁ is the holotype); (E_1,2) ambisporoid; SNSB-BSPG 2016 XII 24; (F_1,2) entrophosporoid; SNSB-BSPG 2016 XII 12.

RESULTS

Twenty-nine different Rhynie chert thin sections have yielded a total of >250 Archaeospora-like acaulospores (Figs 2A_1,2–C_1,2, 3A–O, 4A–E, G–K, 5A and 6B), 37 of which were physically connected to a saccule (Figs 2D_1,2–F_1,2, 3A–O, 4A, C, E, K and 6B). Specimens (with or without attached saccules) occur within plant tissue (decaying or otherwise) and in the surrounding chert matrix. They often co-occur with considerably larger spore–saccule complexes of the type initially described by Dotzler et al. (2009) in (partially) degraded axis portions of the early land plant sporophyte Aglaophyton majus. There seems to be no particular correlation as to whether the acaulospores formed within plant tissue or in the matrix (Table 1); however, intact saccules are more common within degraded plant axes.

Fig. 3.

Intraspecific morphological variability of Archaeosporites rhyniensis gen. et sp. nov.; all figures to same scale. (A) SNSB-BSPG 2016 XII 12. (B, C) SNSB-BSPG 2016 XII 13. (D–H) SNSB-BSPG 2016 XII 24. (I) SNSB-BSPG 2016 XII 31. (J, K) SNSB-BSPG 2016 XII 38. (L) SNSB-BSPG 2016 XII 39. (M, N) P3980. (O) P3981.

Fig. 4.

Archaeosporites rhyniensis gen. et sp. nov. compared with extant species of Archaeosporaceae. (A and B, C and D, E and F) Acaulosporoid spores remaining attached to an easily recognizable empty saccule (s). (A) P3980. (B) Archaeospora ecuadoriana; W5340. (C) SNSB-BSPG 2016 XII 6. (D) Archaeospora myriocarpa; W5878. (E) P3980. (F) Archaeospora trappei; W4610. (G, H) Saccules (s) lacking acaulospore development. (G) SNSB-BSPG 1965 I 350. (H) Archaeospora trappei; W4743. (I, J) Acaulospores with congealed contents. (I) SNSB-BSPG 2016 XII 6. (J) Archaeospora myriocarpa; W5878. (K, L) Entrophosporoid spores still attached to a collapsed saccule (s). (K) SNSB-BSPG 2016 XII 26. (L) Archaeospora schenckii; W4389.

Fig. 5.

Comparison of acaulospore wall structure (C1–3) between fossil and extant specimens. (A) Archaeosporites rhyniensis gen. et sp. nov. with adherent debris on the outermost wall component (C1) and apparent third innermost wall component (C3); P3981. (B) Extant Archaeospora spainiae; W6437.

Fig. 6.

Size comparison of Rhynie chert spore–saccule complexes. (A) Example of large spore–saccule complex described by Dotzler et al. (2009); SNSB-BSPG 2013 XIV 6. (B) Archaeosporites rhyniensis gen. et sp. nov. spore–saccule complex; P3980. Arrowhead indicates saccule.

Table 1.

Distribution of intact spore–saccule complexes, detached acaulospores and detached saccules of Archaeosporites rhyniensis in chert matrix or within plant remains, or which cannot be determined as either in thin sections prepared from Rhynie chert block SNSB-BSPG 2016 XII

SNSB-BSPG 2016 XII slide no.	Chert matrix		Plant remains		Cannot be determined
	Spore or saccule	Spore and saccule	Spore or saccule	Spore and saccule	Spore or saccule	Spore and saccule
12	13	0	6	11	10	0
13	13	8	9	2	11	4
18	11	2	0	1	0	2
19	5	5	0	0	0	1
24	7	5	2	8	4	2
26	13	2	0	0	1	0
34	5	1	6	0	2	0
37	5	3	0	0	1	0
38	11	1	0	0	0	0
39	4	1	3	1	4	0
40	6	0	3	0	0	1
Total	93	28	29	23	33	10

Saccules

Specimens with saccules were not all in good condition. Measurements were therefore made only from those that were definitely identifiable (e.g. Figs 3A–O and 4A, C, E, G, I, K). The fossil saccules appear as long-necked, balloon-shaped structures that, as measured from the distal point of the saccule to the centre of the acaulospore attachment at the saccule neck, are 31–132 µm long × 27–75 µm at the widest point (mean 83 × 41 µm), but can be globose (Fig. 2E₁, F₁), ovoid (Fig. 3B, K) or ellipsoid (Fig. 3E, G, H). Saccules are bound by a single, smooth wall component (C1), ~0.5–1 µm thick, have no obvious structural features such as laminations or ornamentation, and lack contents. In general, the saccules are far more difficult to recognize than the acaulospores; saccules have been informally termed ‘ghost saccules’ because they are easily overlooked (e.g. Fig. 6B). Moreover, even when spore–saccule complexes are intact within the confines of plant axes, they are usually only visible at high magnifications (i.e. ×400 or ×1000 total magnification) and with careful screening. Due to their small size and overall shape, if the saccules are disconnected from acaulospores (Fig. 4G) they can easily be mistaken for vesicles of other mycorrhizal fungi; and if, in the matrix, saccules are typically collapsed or deformed (Fig. 4K), they may be rendered unrecognizable. Therefore, only saccules of complete spore–saccule complexes were measured.

Acaulospores

The acaulospores range from globose to subglobose, 24–83 × 24–78 (mean 51 × 44; n = 248) µm. The colour of the acaulospores in nature cannot be determined, but in the fossil specimens the saccule is grey and the spore colourless or yellow to brown (Fig. 3A–O). For those wall components that can be attributed with a good degree of certainty, the wall structure is composed of two or three components (Fig. 5A). The very thin, flexible nature of the saccule wall, which is an integral, surrounding part of the propagule, is considered as wall 1, with a single component (C1) that is 0.5–2.5 µm thick. Only the outer (saccule) wall (C1) and a single wall (C2) can be seen in most instances. Component 2 is similar to a ‘semi-rigid wall’ sensu  Morton (1990), ~0.5–3 µm thick (measurements rounded to the nearest 0.5 µm). Some specimens have a granular mucilage-like outer coating to which fine substrate particles adhere, lending the appearance of a thick outer wall. Other specimens have a third, innermost component (C3), which may represent a ‘germinal wall’ (Spain, 2003; Hafeel, 2004), but we do not rule out that it may be an artefact of microscopy due to refringence. Such differences may reflect different stages of development, or may be artefacts of preservation.

The fossils indicate that the acaulospores develop in the neck of the sporiferous saccule in any of three ways, i.e. acaulosporoid (Fig. 2D₁), ambisporoid (Fig. 2E₁) and entrophosporoid (Fig. 2F₁). However, the precise details are only visible in specimens that are well preserved and ideally sectioned. The distance between the saccule base and the centre of the acaulospore attachment is variable between acaulospore positions, e.g. acaulosporoid (45–60 µm), ambisporoid (57–65 µm) and entrophosporoid (13–15 µm). The attachment site in the acaulosporoid type, or ‘lateral bud’, is 5.5 µm wide; the bud is ~1–1.5 µm high. The pedicle in the ambisporoid type is approximately ≥10 µm by 3–4 µm long and lacks a septum. Finally, the distance between the spore attachment point and saccule neck in entrophosporoid specimens ranges from 10 to 13 µm.

Other features

In contrast to the saccules, some of the acaulospores have internal contents that range from amorphous matter (Fig. 4I) to fine particles, which may be located in the centre of the spore or along wall component C3 (Fig. 3O). None of the specimens is presented in such a way that the scar (cicatrix) where the spore detached from the saccule could be observed. The walls lacked ornamentation and were not obviously laminated (Fig. 5A). There is no evidence of dimorphic spore production (i.e. acaulospores and glomoid spores) on the parental hyphae of the spore–saccule complexes.

SYSTEMATIC PALAEOMYCOLOGY

Fungi, Glomeromycota, Glomeromycotina, Archaeosporales, Archaeosporaceae

Fossil genus: Archaeosporites gen. nov. C. Walker, C.J. Harper et M. Krings

Figs 2A _1,2 –C _1,2, 3A–O, 4A–E, G–K, 5A and 6B

Index Fungorum: IF 556488

Etymology: The name underscores the similarity to the extant genus Archaeospora; the suffix -ites (Greek: to indicate a close connection) is used to designate a fossil taxon as proposed by Pirozynski and Weresub (1979).

Type species:  Archaeosporites rhyniensis sp. nov. C. Walker, C.J. Harper et M. Krings (hic designatus)

Diagnosis: Fossil propagules (acaulospores), sessile, occurring singly, forming laterally or centrally within neck of sporiferous saccule; saccule <150 µm long, formed by blastic development of hyphal tip, with thin wall of one component; acaulospores globose to ellipsoid, <100 µm in diameter; wall of two main components, outer component (saccule wall) overlying structural wall; outer wall component sometimes overlaid by small adherent particles.

Fossil species:  Archaeosporites rhyniensis sp. nov. C. Walker, C. J. Harper et M. Krings (Fig. 2D₁)

Index Fungorum: IF 556489

Holotype: A saccule with a laterally produced acaulospore present in thin section SNSB-BSPG 2016 XII 12 (Fig. 2D₁) housed in the palaeobotanical collection of the SNSB-Bayerische Staatssammlung für Paläontologie und Geologie, Munich, Germany.

Diagnosis: Spores developing within neck of balloon-shaped sporiferous saccule, formed by expansion of hyphal tip that is lost as the spore becomes detached; saccule approx. 41–78 × 30–78 μm (n = 8), smooth, delicate, lacking protrusions, with a single wall approx. 1 μm thick; spores (including remains of saccule wall) 24–83 × 24–78 (mean 51 × 44; n = 248), globose to subglobose, occasionally ellipsoid, broadly ellipsoid, obovoid or ovoid, with two or three wall components, (C1–C3); C1, saccule wall, thin, flexible, up to 1 μm thick, sometimes lacking due to breakdown in substrate; main spore wall (C2) up to 3 μm thick; a very fine wall component (C3; germinal wall?), less than 1 μm thick, may be present.

Specimens examined: United Kingdom, Scotland, Aberdeenshire, Rhynie chert Lagerstätte. Besides the type specimen [mandated in Article 8.5 of the relevant International Code for Nomenclature (Turland et al., 2018) as a single specimen], more than 50 specimens are present in the type slide. Only eight of these specimens had recognizable saccules attached, although some had at least some evidence of the remains of the saccule surrounding the acaulospore.

Etymology: Referring to the provenance of the fossil, the Rhynie chert Lagerstätte.

Type locality: Rhynie, Aberdeenshire, Scotland, National Grid Reference NJ 494276 (57°20′09.97″ N, 002°50′31.83″ W).

Age: Early Devonian; Pragian–earliest Emsian, 411.5 ± 1.3 Ma (Parry et al., 2011), 407.1 ± 2.2 Ma (Mark et al., 2011).

DISCUSSION

The Early Devonian Rhynie chert has provided astonishingly detailed insights into the early evolution of arbuscular mycorrhizal symbioses and the diversity of GAF in early terrestrial ecosystems, due to the exquisite preservation of fungal hyphae, vesicles, arbuscules and spores that closely resemble present-day arbuscular mycorrhizas in both sporophytes and gametophytes of the early land plant Aglaophyton majus (Kidston and Lang, 1921; Remy et al., 1994; Taylor et al., 1995, 2005), and structures suggestive of the presence of similar symbioses also in other early land plants (reviewed by Krings et al., 2017a; Brundrett et al., 2018; Strullu-Derrien et al., 2018; Walker et al., 2018b). Evidence of GAF diversity in the Rhynie palaeoecosystem primarily occurs in the form of different types of propagules that are indistinguishable from, or closely reminiscent of, propagules produced by members of extant genera such as Glomus, Rhizophagus, Sclerocystis, Gigaspora and Scutellospora (Kidston and Lang, 1921; Taylor et al., 1995; Dotzler et al., 2006, 2009; Karatygin et al., 2006; Krings et al., 2015, 2017b; Harper et al., 2017; Walker et al., 2018b). Moreover, some of the enigmatic mantled fungal reproductive units from the Rhynie chert, e.g. Helmutella devonica, Windipila spinifera and W. wimmervoecksii, may also be members of the Glomeromycota (Krings and Taylor, 2014; Krings and Harper, 2017, 2020), albeit with hyphal investment morphologies not seen in present-day taxa, such as Corymbiforme globiferum (basionym Glomus globiferum) and Glomus mortonii (Koske and Walker, 1986; Bentivenga and Hetrick, 1991; Wu and Sylvia, 1993; Błaszkowski, 1995). Taken together, these fossils constitute the largest body of fossil evidence of GAF gathered to date from any ancient ecosystem. However, systematic screening of Rhynie chert thin sections has recently yielded additional types of propagules that have not been recognized previously, thus suggesting that the fossils described to date represent only a fraction of the diversity that was actually present in the Rhynie palaeoecosystem.

Affinities

The spore–saccule complexes detailed in this study are morphologically different from any type of fungal propagule previously described from the Rhynie chert. Dotzler et al. (2009) described an acaulospore-forming fungus from the Rhynie chert similar to present-day Ambispora spp. The superficial similarity of overall form (i.e. development of an acaulospore within a sporiferous saccule) may lead investigators to speculate that these two fungi might represent grossly differently-sized members of the same genus. Similar consideration of the fungi now shown to belong to three different families (Archaeospora, Ambispora and Acaulospora) originally placed all these in the same genus, namely Acaulospora (e.g. Archaeospora trappei, Ambispora nicolsonii and all current members of the genus Acaulospora were initially placed in the one genus). However, biochemical and molecular markers allowed the current placing to be made in separate genera. Although it is not possible to use these techniques on fossils, the size difference and wall structure details can be used to conclude that the two fossil acaulosporoid species mentioned above are similarly different. The organism described by Dotzler et al. (2009), with acaulospores 270–427 × 236–473 µm (Fig. 6A) and saccules 200–450 (–700) × 200–450 µm long (Fig. 6A), is very different from the form described in this study, which has acaulospores on average 52 × 46 µm and saccules 83 × 41 µm in size (Fig. 6B). Moreover, the wall architecture of the new form is the same as that of modern Archaeospora spp., whereas the acaulospores described by Dotzler et al. (2009) correspond to wall architectures seen in present-day Ambisporaceae (see below). Finally, the new spore–saccule complexes have been found in both the chert matrix and land plant axes in varying states of preservation, whereas Dotzler et al. (2009) report their specimens only from within partially degraded plant axes.

Development of acaulospores from a sporiferous saccule is known to occur in three different present-day families of Glomeromycota, i.e. the Acaulosporaceae (Acaulosporales), Ambisporaceae (Archaeosporales) and Archaeosporaceae (Archaeosporales) (Morton and Benny, 1990; Morton and Redecker, 2001; Walker et al., 2007b). The fossils detailed in this study differ from Acaulosporaceae and Ambisporaceae primarily in the dimensions of the acaulospores, which may reach diameters of >380 µm in Acaulosporaceae (Schultz et al., 1999) and >390 µm in Ambisporaceae (Spain et al., 2006), but correspond to the acaulospores produced by members of the Archaeosporaceae, which attain diameters of only 21–114 µm (Spain et al., 2006). All morphological variation in the fossils is well within the range of intraspecific variability in modern Archaeospora spp. (e.g. Schüßler and Walker, 2019). Members of the Acaulosporaceae, Ambisporaceae and Archaeosporaceae may produce two kinds of spores, i.e. acaulospores and glomoid spores (e.g. Spain, 2003; Walker et al., 2007b; Taylor et al., 2014); however, no evidence of glomoid spores has been found to date in the fossils, and thus this trait cannot yet be compared. Based on overall morphology of the spore–saccule complex, the fossils correspond well with the original descriptions of Archaeospora trappei (Figs 1A, B and 4F, H) (basionym Acaulospora trappei) (Ames and Linderman, 1976; Morton and Redecker, 2001), A. schenckii (Fig. 4L) (basionym Entrophospora schenckii) (Sieverding and Toro, 1987; Schüßler and Walker, 2010), A. myriocarpa (Fig. 4D, J) (basionym Acaulospora myriocarpa) (Schenck et al., 1986; Oehl et al., 2011), A. ecuadoriana (Figs 2A_1,2–C_1,2 and 4B) (Schüßler and Walker, 2019) and A. spainiae (Fig. 5B) (basionym Palaeospora spainiae) (Oehl et al., 2015; Schüßler and Walker, 2019).

The number and nature of wall components and layers constituting the individual wall groups are used as important taxonomic characters in the Glomeromycota (Walker, 1983; Walker and Vestberg, 1998). The wall structure and morphological characteristics of extant and fossil material are comparable by light microscopy. The wall of the fossil acaulospores described here consists of three components (C1–3; Fig. 5A) corresponding to that of extant members of the Archaeosporaceae (Morton and Redecker, 2001; Fig. 5B). The number of distinguishable wall components is always three or more in the Acaulosporaceae and two to four in Ambisporaceae (Mosse, 1970; Walker et al., 2007b), and in the fossils described by Dotzler et al. (2009) there are three wall groups, which consist of six components. The outermost wall component (C1), if preserved, is a thin layer (≥1 µm thick) of similar thickness to many of the modern Archaeospora spp.; some of the fossil specimens have fine substrate particles on C1 of the saccule and acaulospore wall, which has been described as a distinct wall for A. spainiae (Oehl et al., 2015, p. 94). As in most modern Archaeospora spp., the fossil acaulospore wall proper (C2) is the most prominent wall component (Fig. 5A); it appears to occur in the form of a ‘semi-rigid wall’ sensu  Morton (1990). The innermost wall component (C3) is not present in all of the fossil specimens, and it may represent a germinal wall as described by Morton (1990), which is also present in A. myriocarpa (Schenck et al., 1986, pp. 113–114), but we do not rule out that it could be just the boundary between the plasma membrane and the rigid C2. It is impossible to determine if the fossil acaulospores are in their original coloration or if the colour of the fossils is the result of post-inundation changes and fossilization. Extant members of the Archaeosporaceae are colourless to translucent white except when moribund or dead (e.g. Ames and Linderman, 1976; Morton and Redecker, 2001). It is also impossible to know if the Rhynie organisms had pigmented acaulospores in life, but some fossil specimens appear brown (Fig. 4I). There are no coloured living acaulospores of this size in either Archaeospora or other genera with smooth acaulospores, but moribund or dead spores of extant fungi in the family can become brown (Fig. 4J). It seems likely that, given the lack of pigmentation in modern species of Archaeospora, Archaeosporites rhyniensis propagules were also colourless.

Acaulospores in extant Archaeosporaceae develop within the neck of the saccule in one of three ways (i.e. acaulosporoid, entrophosporoid and ambisporoid), a feature also observed in the fossils, even within the same thin section (Fig. 2D_1,2–F_1,2). This characteristic is described for A. myriocarpa, A. schenckii or A. ecuadoriana (Fig. 2A_1,2–C_1,2), and Schüßler and Walker (2019) note that it remains unclear whether all Archaeospora species produce acaulospores in all three topologies.

Saccules are delicate structures and, consequently, are less persistent than the acaulospores, which have a more complex and thickened wall structure (Fig. 4K, L). Saccules of Archaeosporites rhyniensis range from 27 to 132 µm long, which fits well within the range of saccule length in modern Archaeospora spp., i.e. 22–179 µm. This is compared with the much larger saccules of the Acaulosporaceae (up to 360 µm long) (Błaszkowski, 1989; Palenzuela et al., 2013), Ambisporaceae (up to 380 µm) (Spain et al., 2006; Walker, 2008) and fossil saccules described by Dotzler et al. (2009), which reach between 200 and 700 µm in length. The saccules of the fossils described in this study do not possess lateral hyphal outgrowths, which is a typical feature of members of Acaulospora (Gerdemann and Trappe, 1974); rather, they are simple, smooth, balloon-shaped structures, precisely as illustrated in Fig. 1A of Ames and Linderman (1976) for Archaeospora trappei.

Specimens of the extant Archaeospora schenckii, A. trappei and A. myriocarpa were measured and compared statistically with the Rhynie chert fossils, which had smaller shortest and longest dimensions (F = 520.7; d.f. = 3.6, 346.8; P < 0.001) and a larger range, but fell within the overall limit given in the protologue descriptions for all species (including A. spaniae) of 23–114 × 21–96 µm (Fig. 7). When converted to approximate volumetric measurements, the analysis results were the same, i.e. the Rhynie acaulospores are significantly smaller (H = 484.22, d.f. = 3, P < 0.001) than those of extant Archaeospora spp. Although one has to consider a possible shrinkage factor resulting from water loss during silicification (Fleischmann et al., 2007), the differences in dimensions are still statistically so significant that we feel confident in formally describing a new fossil genus and species.

Fig. 7.

Boxplot of calculated acaulospore volumes for Archaeospora myriocarpa, A. schenckii, A. trappei and Archaeosporites rhyniensis. Solid horizontal lines within the boxes represent the median value, the portion of the box below the median represents the lower quartile, and the portion of the box above the median represents the upper quartile. The horizontal line above the box represents the maximum, the horizontal line below the box represents the minimum, and the circles represent outliers. The dagger (†) indicates that the taxon is extinct.

Based on morphology, the Rhynie chert fossils described herein can satisfactorily be placed in Archaeospora (Table 2), but given the need for molecular evidence in modern-day taxonomy, we therefore propose to name them in the fossil genus and species Archaeosporites rhyniensis gen. et sp. nov. The new genus and a species are thus defined on morphology from these fossils, but with the understanding that the many specimens referred to herein could represent several morphologically similar taxa (probably at the level of species) that cannot be distinguished based on fossils.

Table 2.

Summary of key features used to compare affinities among and between extant and fossil kinds of ‘acaulospore from a sporiferous saccule’-producing fungi. Note that expansion includes any minimum or maximum measurement in any orientation (diameter, length, width, etc.). Superscript letters correspond only to references within that particular row

Family, specimen or taxon	Acaulospores		Saccules		Wall organizatione   (number of wall groups and components/layers)	Acaulospore development in saccule neck topologyf	Reference(s)
	minimum expansion (µm)a	maximum expansion (µm)b	minumum expansion (µm)c	maximum expansion (µm)d
Acaulosporaceae	(11–)44	380(–520)	48	≤360	3 groups up to 7 layers	Acaulosporoid, entrophosporoid	Mosse, 1970 f; Gerdemann and Trappe, 1974b; Błaszkowski, 1989a, 1990b; Schultz et al., 1999b; Walker et al., 2007be; Palenzuela et al., 2013a,c; Schüßler and Walker, 2019f
Ambisporaceae	87	>390	125	≤380	2–3 groups 3–4 components	Ambisporoid	Spain et al., 2006 b,d; Walker et al., 2007be; Oehl et al., 2012a; Palenzuela et al., 2013c; Schüßler and Walker, 2019f
Archaeosporaceae	21	114	22	179	3 groups 3 components	Acaulosporoid, entrophosporoid and ambisporoid	Spain et al., 2006 a,b; Schüßler and Walker, 2019c–f
Unnamed Rhynie chert spore–saccule complex†	236	473	200	450(–700)	3 groups 6 components	Ambisporoid	Dotzler et al., 2009 a–f
Archaeosporites rhyniensis †	24	83	27	132	3 groups 3 components	Acaulosporoid, entrophosporoid and ambisporoid	This studya–f

Ecological roles of Archaeosporites and Archaeospora species

Based on morphological criteria, the Rhynie chert fossils described in this study are so similar to the extant Archaeospora that they can be placed in the Archaeosporaceae. By analogy with the modern species, the fossil organism may have been symbiotic with plants and formed arbuscular mycorrhizas. However, there is currently no evidence that can be used to link mycelia and arbuscules to Archaeosporites rhyniensis. Furthermore, it is impossible to know whether the land plant axes that contained the spore–saccule complexes of A. rhyniensis were alive at the time of colonization and fungal sporulation, or whether the acaulospores merely formed within litter layers comprising dead and decaying plant parts. In extant species of Archaeospora, the propagules are most abundant in the substrate (which may be decaying plant material, soil, or substrate in which potted host plants are grown). For example, in a sample from pot culture (Voucher W6496-5, Attempt 1289-5 of C. Walker, of Archaeospora trappei with Plantago lanceolata) there was not a single propagule in 10 cm of roots (cleared and stained with trypan blue). This extraction produced large numbers of acaulospores, 107 of which were preserved on microscope slides (W6496-1 to 6496-4). Only eight of these acaulospores still had remnants of the saccule attached, and only a single developing saccule was found (C. Walker, Gloucester, UK, unpubl. res.). Propagule formation in roots is mentioned in the original species descriptions of A. trappei, A. myriocarpa and A. ecuadoriana (Ames and Linderman, 1976; Schenck et al., 1986; Schüßler and Walker, 2019), but not for A. spainiae (Oehl et al., 2015); however, data on the occurrence or distribution of propagules within the soil or plants are not provided. The propagules of A. rhyniensis were found within plant tissue and in the surrounding matrix, and there is no evidence to suggest that one substrate was favoured for spore production (Table 1). It appears that, on average, the delicate saccules survive better when located within plant axes. Preservation of the spore–saccule complexes within the protective confines of plant tissue may have shielded the fragile structures and fine structural details against the destructive forces during fossilization in a similar manner to that suggested in other studies of fossil microorganisms (e.g. Krings and Taylor, 2012, 2015).

Members of the Archaeosporaceae today form mycorrhizal associations with plants (Table 1 in Morton and Redecker, 2001). However, in comparison with other mycorrhizal fungi (e.g. Glomus spp.), there are relatively few experiments with plants inoculated with Archaeospora spp. and on the impact of Archaeospora mycorrhizas on plant fitness. Those investigations that did include Archaeospora spp. found a positive impact of the presence of the fungus on phosphorus uptake and herbivory tolerance in individual host plants (Helgason et al., 2002; Bennett and Bever, 2007). Long-term ecological studies over a period of 32 years suggest that alternating mycorrhizal symbioses with Archaeospora and members of the Glomeraceae in certain conifers increases the fitness of the individual plant, surrounding forest and the microbial community (Lu et al., 2019). In extant geothermal wetlands such as Yellowstone National Park, which has been regarded as a modern analogue to the Rhynie chert ecosystem (Channing and Edwards, 2009; Garwood et al., 2020), Archaeospora spp. was abundantly recorded for several study sites (Appoloni et al., 2008). The diversity of mycorrhizal fungi can be used as a proxy of plant fitness in extant ecosystems (Van der Heijden et al., 2007; Hempel et al., 2013). Bearing in mind that the Rhynie chert has yielded a considerable diversity of fossils of glomeromycotan fungi from several of the major lineages within this phylum (reviewed by Krings et al., 2017a; Walker et al., 2018b), and although there is no evidence of mycorrhizal associations between Archaeosporites rhyniensis and the plants in the Rhynie ecosystem, it seems reasonable to suggest that the availability of several different lineages of GAF may have been beneficial to the performance of the land plants in this early terrestrial ecosystem.

Conclusions

One of the intriguing palaeomycological facets of the Early Devonian Rhynie chert is the abundance and morphological diversity of arbuscular mycorrhizal fungi, especially because the early land plants that lived in this hot-spring palaeoecosystem all lacked true roots. However, it becomes increasingly clear that there is still considerable undocumented diversity in the form of new specimens and hitherto unknown or unrecognized features of mycorrhizal fungi in the Rhynie chert. Archaeosporites rhyniensis described in this study is one excellent example of this diversity. This fossil is so morphologically similar to some extant Glomeromycota that, were it not for the need for molecular evidence in modern day taxonomy, it may well be placed in a modern genus, and thus suggests that the genus Archaeospora perpetuated for considerably more than 400 million years. Fossils such as Archaeosporites rhyniensis, albeit incomplete (i.e. not connected to the parental mycorrhiza in the plant), not only increase the number of calibration points that can be used in aligning molecular clock estimates with fossil evidence in the discussion of the evolution of mycorrhizal systems, but also reiterate questions pertaining specifically to the Rhynie chert mycorrhizas. For example, Strullu-Derrien et al. (2014) suggest that Rhynie chert land plants did not exclusively form mycorrhizas with Glomeromycota, but also with members of the Endogonales (Mucoromycotina), based on the presence of certain types of intercellular hyphae, hyphal segments interpreted as intracellular coils, and small propagules resembling glomoid spores in Horneophyton lignieri. Although this idea has found its way into the general discussion of the origin and evolution of mycorrhizal symbioses (e.g. Feijen et al., 2018; Hoysted et al., 2018; Bonfante and Venice, 2020), the fossil evidence is equivocal. This is in contrast with the Rhynie chert fossils of Glomeromycota, which are often exquisitely preserved and therefore can be assigned to modern families, orders or genera with confidence (Brundrett et al., 2018). The morphology of mucoromycotinan colonization in plants today—from ectomycorrhizas formed by Endogone lactiflua (Walker, 1985) to arbuscular mycorrhizas with Planticonsortium tenue (Walker et al., 2018a) and similar, unidentified fungi (Beck et al., 2005)—has been illustrated in great detail, and there is also increasing literature evaluating the characteristics and evolution of these relationships based on molecular and genetic data (e.g. Chang et al., 2019; Hoysted et al., 2019; Rimington et al., 2019, 2020). As a result, there is now a good base for seeking similarly conclusive evidence from the fossil record. There is considerable interest in the mycorrhizal component of modern ecosystems and the roles these symbioses play in ecosystem functioning, while we are only beginning to assess such interactions in the fossil record. This paper not only adds to the knowledge of Glomeromycota in early palaeoecosystems, but can also serve to stimulate interest in these organisms, and encourage other researchers to prepare and study suitable specimens from their collections.

SUPPLEMENTARY DATA

Supplementary data are available online at https://academic.oup.com/aob and consist of the following. Table S1: list of voucher specimens of extant fungal specimens used in this study, including species, locality data, and culture information.

FUNDING

This work was supported in part by funds from the United States National Science Foundation (EAR-0542170, and DEB-1441604 subcontract S1696A-A to M.K.) and the Deutsche Forschungsgemeinschaft (DFG grant Ke 584/13-2 to H.K. and M.K.).