Carbon sources for the Palaeozoic giant fungus Prototaxites inferred from modern analogues

Abstract

A wide range of carbon isotope values in the Devonian fossil Prototaxites has been interpreted to support heterotrophy and the classification of Prototaxites as a giant fungus. This inference remains controversial because of the huge size of Prototaxites relative to co-occurring terrestrial vegetation and the lack of existing fungal analogues that display equally broad isotopic ranges. Here, we show wide isotopic variability in the modern saprotrophic fungus Arrhenia obscurata collected adjacent to shallow meltwater pools of a sparsely vegetated glacial succession in the Washington Cascades, USA. Soils collected specifically around the edges of these pools were up to 5‰ higher in δ¹³C than adjacent soils consistent with C₃ origin. Microbial sources of primary production appear to cause these high δ¹³C values, and the environment may be analogous to that of the Early Devonian landscapes, where Prototaxites individuals with extreme isotopic variance were found. Carbon isotopes are also compared in Prototaxites, Devonian terrestrial vascular plants, and Devonian algal-derived lake sediments. Prototaxites isotopic values show little correspondence with those of contemporaneous tracheophytes, providing further evidence that non-vascular land plants or aquatic microbes were important contributors to its carbon sources. Thus, a saprotrophic fungal identity is supported for Prototaxites, which may have relied on deposits of algal-derived organic matter in floodplain environments that were less dominated by vascular plants than a straight reading of the macrofossil record might suggest.

1. Introduction

Late Silurian to Late Devonian Prototaxites has puzzled scientists since its discovery (Dawson 1859). At up to 8 m tall (Hueber 2001), Prototaxites towered over the landscape, particularly before similar heights were first achieved by vascular plants in the Middle Devonian (Stein et al. 2007). Whether it should be classified as a land plant, alga, lichen or fungus has remained contentious for 150 years (Carruthers 1872; Church 1919; Jonker 1979; Hueber 2001). Despite allochthonous marine specimens (Schweitzer 1983), Prototaxites is well represented in non-marine deposits (Griffing et al. 2000; Hotton et al. 2001; Hillier et al. 2008) and is a land organism with the most recent anatomical work supporting a fungal interpretation (Hueber 2001). In subsequent geochemical research, the carbon isotope (δ¹³C) ratios of Prototaxites specimens were found to vary 11‰ within an individual locality and up to 13‰ overall (Boyce et al. 2007). The samples separated into two groups, a ¹³C-enriched group between −16‰ and −19‰ and a ¹³C-depleted group between −27‰ and −29‰. This variability led the authors to suggest a heterotrophic origin for Prototaxites carbon—thus supporting a fungal interpretation—and to further suggest that the ¹³C-enriched Prototaxites specimens assimilated carbon from non-vascular autotrophs possessing a carbon-concentrating mechanism, such as terrestrial microbial soil crusts or some hornworts (Smith & Griffiths 1996; Evans & Belnap 1999). However, these authors did not identify any present-day fungal analogues for this large range of carbon isotope values.

Modern analogues are most likely to be found in environments that partially resemble the Early Silurian and Devonian landscapes where Prototaxites occurred, floodplains characterized by sparse and patchy vegetation (Griffing et al. 2000; Hotton et al. 2001; Hillier et al. 2008). Early successional systems developing after glacial retreat have some of the desired characteristics, and the simplified soil and vegetation structure lends itself to comparing carbon isotope patterns among the individual components. Here, we compared carbon isotope values among plants, ectomycorrhizal (symbiotic) fungi, saprotrophic fungi and soils from a sparsely vegetated glacial foreland at Lyman Glacier, Washington, USA as a possible analogue to Early Devonian landscapes. One of the saprotrophic genera collected, Arrhenia, is commonly associated with algae and mosses (Redhead et al. 2002), and thus appears probably to derive carbon from non-vascular plant sources. To constrain the possible isotopic fractionation between saprotrophic fungi and source carbon, we compiled results from culture studies with saprotrophic fungi.

2. Material and methods

(a). Literature data compilation

Data on δ¹³C values of terrestrial plants and Prototaxites from the Devonian were compiled from the following published sources: Beerling et al. (2002); Boyce et al. (2003, 2007), Jahren et al. (2003); Fletcher et al. (2004) and Peters-Kottig et al. (2006). Complete details are given in appendix A. To better constrain the potential isotopic relationship between fungi and source carbon in Prototaxites, literature values were also compiled from field and culture studies of modern fungi growing on complex substrates such as litter and wood.

(b). Modern data

(i). Site description

The study area is the foreland of Lyman Glacier. Lyman Glacier is at 48°10′ N, 120°53′ W, at an elevation of 1800 m in the Cascade Mountains of Washington, USA (figure 1). Sites were located 100, 300, 450, 600 and 750 m from the glacial terminus. The corresponding ages since glacial retreat were approximately 20, 40, 50, 60 and 70 years. The timing of glacial retreat has been determined from historical photographs and observations (Jumpponen et al. 1998, 2002). Individual sites were relatively heterogeneous and included both drier patches with vascular vegetation and more poorly drained areas of ephemeral water accumulation.

Figure 1.

Aerial photo of the Lyman Glacier foreland. The glacier is at the left (south), the terminal moraine is at the right (north) with a stream outlet at the upper right. The sampling transect was along the western shore of the large lake. Image taken July 2006 for the US Geological Survey, High Resolution State Orthoimagery for Washington, 2008. Scale bar, 0–1000 m.

(ii). Sample collection

At each site, current-year foliage was collected from five plants of each of eight species if present. Tree and shrub species were generally less than 1 m in height. Species collected were Luzula piperi (Juncaceae), Saxifraga ferruginea (Saxifragaceae), Luetkea pectinata (Rosaceae), Epilobium latifolium (Onagraceae), Abies lasiocarpa (Pinaceae), Salix phylicifolia (Salicaceae), Cassiope mertensiana (Ericaceae) and Phyllodoce empetriformis (Ericaceae).

Fungal sporocarps were collected along the transect and the distance from the glacial terminus was recorded to the nearest 5 m. Fungi were identified to genus or species and classified as ectomycorrhizal (Laccaria montana, Inocybe lacera and Cortinarius tenebricus) or saprotrophic (Arrhenia obscurata and Galerina spp.). All fungi were basidiomycetes. At Lyman Glacier, A. obscurata commonly fruited in wet spots among mosses or on wet soil, often with algae on the soil surface (J. Trappe 1988–1999, personal communication). Five samples of relatively well-drained surface soils (top 5 cm) from each site were also collected. The sites were originally visited in late August 1999 for collections of soils, foliage and fungal sporocarps. During a later collection in early September 2009, soils were specifically collected around the edges of eight shallow meltwater pools 700 m from the glacial terminus (five samples per pool; an example is shown in figure 2). Pools 1–4 had no or very limited surrounding vegetation, whereas pools 5–8 were surrounded by early successional vegetation such as mosses and Carex (sedges). Additional details on site description and sample collection are in Hobbie et al. (2005).

Figure 2.

Soil was sampled in September 2009 at the Lyman Glacier foreland around 700 m from the glacial terminus, around shallow meltwater pools where A. obscurata commonly fruited. Photo: Ari Jumpponen.

(iii). Sample processing and isotopic analyses

Fungal and foliar samples were analysed for δ¹³C on a Finnigan Delta-Plus isotope ratio mass spectrometer linked to a Carlo Erba NC2500 elemental analyser (Finnigan MAT GmbH, Bremen, Germany) located at the US Environmental Protection Agency, Corvallis, OR, USA. The internal standards for isotopic and concentration measurements were acetanilide and pine needles (NIST 1575). The average difference of duplicate samples was 0.2‰ for δ¹³C. Soils from 1999 were analysed for δ¹³C at the Max Planck Institute for Biogeochemistry in Jena, Germany using a Finnigan Delta-Plus with acetanilide (n = 9, s.d. = 0.1‰) as the working standard. Soils from 2009 were analysed at the University of New Hampshire Stable Isotope Laboratory, with NIST 1515 (peach leaves), Underhill Bs (mineral soil), Underhill Oa (organic soil) and tuna as working standards.

3. Results

Saprotrophic fungi were first collected at 270 m from the terminus of Lyman Glacier, whereas ectomycorrhizal fungi were first collected at 340 m. Means and standard deviations for taxa of saprotrophic fungi are: A. obscurata −20.7 ± 2.6‰, n = 34; and Galerina −22.1 ± 1.0‰, n = 20. Means and standard deviations for taxa of ectomycorrhizal fungi are: C. tenebricus −25.9 ± 0.8, n = 4; I. lacera −25.8 ± 0.6, n = 6; and L. montana −25.9 ± 0.6‰, n = 35. The saprotrophic Arrhenia had some extremely ¹³C-enriched values and a large overall range from −14‰ to −25‰ (figure 3). Overall, saprotrophic fungi at Lyman Glacier averaged 4.7‰ enriched in ¹³C relative to co-occurring ectomycorrhizal fungi, with individual Arrhenia more than 10‰ enriched in ¹³C relative to co-occurring ectomycorrhizal fungi. Of plants that could have supplied sugars to ectomycorrhizal fungi, the conifer A. lasiocarpa averaged −26.6‰ and the deciduous willow S. phylicifolia averaged −27.6‰. There was about a 5‰ range in tracheophyte δ¹³C overall, with highest values for the evergreen shrub Cassiope mertensia (−24.9 ± 0.3‰ at 450 m) and lowest for the forb L. pectinata (−29.8 ± 0.2‰ at 150 m) (complete data by location and plant species are given in appendix B). Overall, plant δ¹³C averaged −27.2‰, samples of the more well-drained soil occupied by these plants averaged −25.6 ± 0.7‰ (s.d.) (range −25.4‰ to −26.0‰), and soils surrounding shallow meltwater pools averaged −24.3 ± 1.6‰ (s.d.) (range −21.3‰ to −26.1‰). The two soil types differed significantly in average δ¹³C (t-test: p < 0.001, d.f. = 60) and in variability of δ¹³C (F-test: p < 0.001). Soil, plant and fungal δ¹³C values for the Lyman forelands are shown in figure 3.

Figure 3.

Carbon isotopes in ectomycorrhizal fungi, saprotrophic fungi, plants and soil at the Lyman Glacier foreland. Plants are shown as green squares, mineral soil collected in 1999 as upside-down pink triangles and mineral soil collected around meltwater pools in 2009 as upside-down black triangles, with standard errors indicated by error bars. Plants are means of eight species. Saprotrophic fungi are indicated by filled symbols: Arrhenia, red triangles; Galerina, blue circles. Ectomycorrhizal fungi are indicated by clear symbols: Cortinarius, triangles; Inocybe, squares; Laccaria, circles. Full data for plant species and soils are given in appendix B.

Carbon isotope patterns in Devonian Prototaxites are shown in figure 4 and are distinct relative to those of Devonian fossil plants. In particular, Prototaxites is distinguished from other fossils in the wide range of δ¹³C values, with three samples clustering between −19.0‰ and −15.8‰ and the remaining eight samples clustering between −28.9‰ and −26.6‰. Data compiled from the literature indicate that modern saprotrophic fungi are about 3‰, enriched in ¹³C relative to their substrates (table 1), but algal-derived lake sediments from Scotland (−31‰ to −34‰, Stephenson et al. 2006) are the only Devonian samples so far found to be more depleted in ¹³C than this second Prototaxites cluster.

Figure 4.

Carbon isotopes in terrestrial plant fossils and Prototaxites from the Devonian. Taxa represented are: Prototaxites (filled circles), Spongiophyton (upright clear triangles), Psilophyton (clear squares), Leclercqia (filled squares), Callixylon (upside-down clear triangles) and Archeopteris (clear diamonds). Other taxa are as indicated directly in the figure. Fossil details are in appendix A. The three indicated Rhynie fossils are Aglaophyton major, Asteroxylon mackiei and Rhynia gwynne-vaughnii.

Table 1.

¹³C enrichment of fungi relative to substrates. ε_F−S = δ¹³C_Fungi − δ¹³C_Substrate. (Standard errors are given when available. Chaetomium globosum is an ascomycete, all other fungi are basidiomycetes.)

substrate	organism	εF−S (‰)	reference
field studies
wood	decay fungi	3.1	Gleixner et al. (1993)
wood cellulose	decay fungi	1.8	Gleixner et al. (1993)
litter	Termitomyces (termite fungi)	3–4	Tayasu (1998)
wood	decay fungi (20 taxa)	3.1 ± 0.2	Kohzu et al. (1999)
wood	decay fungi (five taxa)	3.5 ± 0.3	Hobbie et al. (2001)
culture studies
Fagus wood	Trametes versicolor	3.5 ± 0.5	Kohzu et al. (1999)
Fagus wood	decay fungi (five taxa)	3.3 ± 0.4	Kohzu et al. (2005)
Fagus wood	Chaetomium globosum	1.6 ± 0.2	Kohzu et al. (2005)

4. Discussion

The carbon isotope measurements in saprotrophic fungi at Lyman Glacier are remarkable for their wide range. Because no crassulacean acid metabolism plants, C₄ plants or hornworts are present at Lyman Glacier and the δ¹³C of the C₃ plants collected ranged from −30‰ to −25‰, alternate, non-tracheophyte sources of carbon must account for saprotrophic fungi of δ¹³C greater than −20‰. Analysis of nearby ectomycorrhizal fungi eliminates carbon within the plant-mycorrhizal symbiosis as a viable source for ¹³C-enriched Arrhenia. Ectomycorrhizal (symbiotic) fungi incorporate transferred plant sugars, whereas saprotrophic fungi generally incorporate carbon derived from sugar polymers such as cellulose and hemicellulose (Hobbie 2005). In a global synthesis, saprotrophic fungi were enriched by 1.1–3.5‰ in ¹³C relative to co-occurring ectomycorrhizal fungi (2.3 ± 0.7‰ s.d., mean of 23 studies (Mayor et al. 2009); global average, −22.9‰ for saprotrophic fungi and −25.3‰ for ectomycorrhizal fungi). By contrast, saprotrophic fungi at Lyman Glacier averaged 4.7‰ enriched in ¹³C relative to co-occurring ectomycorrhizal fungi, with Arrhenia and Galerina averaging, respectively, 5.2‰ and 3.8‰ enriched in ¹³C relative to co-occurring ectomycorrhizal fungi; the saprotrophic Arrhenia also varied more in δ¹³C (s.d. of 2.6‰) than Galerina or the ectomycorrhizal fungi (s.d. of 1‰ or less). Together, the isotopic values of the saprotrophic fungus Arrhenia, co-occurring ectomycorrhizal fungi and vascular plants indicate that the more ¹³C-enriched Arrhenia require a carbon source other than vascular land plants. Fungi typically have a 3‰ enrichment in ¹³C relative to their carbon sources (table 1), indicating that Arrhenia ranging from −26‰ to −14‰ are likely to have consumed carbon sources that varied between −29‰ and −17‰.

The glacial foreland at Lyman Glacier is pocked with shallow meltwater pools (figure 2). If the vascular plants at Lyman Glacier are insufficiently enriched in ¹³C to account for some δ¹³C values for Arrhenia, then aquatic photosynthesis in these pools is a probable source of ¹³C-enriched carbon. Phytoplankton can vary widely in δ¹³C depending on environmental conditions and can be quite enriched in ¹³C if they possess a carbon-concentrating mechanism (Laws et al. 2002). The Arrhenia samples fruited around the fringes of these pools, suggesting that organic deposits derived from algae or cyanobacteria in these pools are probable additional sources for Arrhenia. The δ¹³C values of the soils immediately surrounding some of these pools were as high as −21‰ (figure 3). Given the measured values for well-drained soils and C₃ vegetation of between −28‰ and −25‰, higher δ¹³C values for most of the soils surrounding pools do not indicate an exclusively terrestrial C₃ origin, but rather some contribution of aquatically fixed carbon in these areas of ephemeral water cover. Similar ¹³C-enriched values to those measured here have been reported previously for microbes from eutrophic lakes (Bontes et al. 2005; Xu et al. 2007) and hot springs (Jahnke et al. 2004). δ¹³C values of −17‰ to −24‰ have also been reported in cyanobacteria or cyanobacterial lichens from rock outcrops and savannah soils, with the higher δ¹³C values under wet conditions when the diffusion of CO₂ through water and CO₂-concentrating mechanisms were most likely to be important for cyanobacterial photosynthesis (Ziegler & Lüttge 1998).

The carbon isotopic range of saprotrophic fungi from this barren glacial foreland approximate the range of values reported from Devonian Prototaxites and this locality may provide a modern analogue of an Early Devonian landscape with sparse vascular vegetation. The ¹³C-enriched values seen in some Prototaxites fossils (figure 4) require consumption of an autotrophic substrate with a carbon-concentrating mechanism (Boyce et al. 2007). In addition to the previously suggested sources of microbial soil crusts and hornworts, the Lyman Glacier ecosystem indicates that microbial photosynthesis from ephemeral terrestrial pools is an additional substrate possibility for ¹³C-enriched Prototaxites. In Lower Devonian sediments of the Gaspé Peninsula, vascular plants were found in facies indicative of moist environments including levees and swampy areas. By contrast, Prototaxites was prevalent in channel fills that contained material transported from more proximal areas of the floodplain (Griffing et al. 2000; Hotton et al. 2001) that were likely to have been more sparsely infiltrated by vascular plants. Rooting structures from the Lower Devonian of the Anglo-Welsh Basin have recently been interpreted as those of Prototaxites involved in the consumption of exopolymeric substances, dissolved organic carbon in sediment pore fluids, and cyanobacterial and algal wall material deposited in ephemeral pools caused by seasonal flooding (Hillier et al. 2008). Our analysis of the modern analogue Arrhenia supports such an interpretation.

The ¹³C-depleted and ¹³C-enriched populations of Prototaxites had been interpreted, respectively, as consistent with a C₃ tracheophyte substrate and as requiring a non-tracheophyte alternative with a carbon-concentrating mechanism (Boyce et al. 2007). However, the approximately 3‰ enrichment of modern saprotrophic fungi in ¹³C relative to their substrates (table 1) suggests that non-tracheophyte alternatives may also contribute to the carbon sources of the more ¹³C-depleted Prototaxites samples as well. The more ¹³C-depleted Prototaxites samples correspond poorly with the actual isotopic values of fossils of vascular plants and their close relatives, both locally (Boyce et al. 2007) and globally (Peters-Kottig et al. 2006) (figure 4), as the plant fossils are consistently too enriched in ¹³C to be the exclusive carbon source for Prototaxites. In the Late Devonian, isotopic values of local Callixylon (progymnosperm wood) and Prototaxites overlap, but the 3‰ offset expected with the consumption of a woody substrate would require plants to be more ¹³C-depleted than observed. Similarly, in the Early Devonian, Prototaxites did not overlap in δ¹³C with local vascular plant fossils and had minimal overlap with a larger compilation of Early Devonian vascular plants (figure 4).

One recent report suggested that Prototaxites fossils represent rolled up liverwort mats, with the broad range in Prototaxites δ¹³C indicating variable degrees of mixotrophy in early liverworts (Graham et al. 2010). Although δ¹³C patterns in Prototaxites cannot distinguish between heterotrophic contributions from liverworts or fungi, the evidence strongly suggests that liverwort mats are unlikely to form the anatomical features reported from Prototaxites. For example, the layering of Prototaxites trunks forms concentric rings rather than spiral rings in transverse section, anatomical structures frequently pass through multiple layers, no sediment is entrained between layers and the tissue is composed of tubes regularly oriented along the long axis of the trunk. These observations are consistent with the growth increments in a coherent organism but are not consistent with a rolled up mat (see Hueber (2001) for a comprehensive review of Prototaxites anatomy).

Both bryophytes and aquatic algae without carbon-concentrating mechanisms can be ¹³C-depleted relative to C₃ vascular plants (Raven et al. 2000; Jahren et al. 2003; Fletcher et al. 2004). For example, algal-derived lake sediments from the Middle Devonian, ranging from −31‰ to −34‰, are much lower in δ¹³C than any Devonian tracheophyte fossils (Stephenson et al. 2006). Because their preservational potential is low, the diversity of bryophyte-grade land plants in the Devonian is likely to have been higher than the macrofossil record would suggest (Hotton et al. 2001; Boyce 2010). The most ¹³C-depleted Prototaxites specimens suggest that bryophytes and algae could have provided an important contribution to Prototaxites carbon sources throughout the Devonian, including after the advent of the first vascular plant forests. Given that neither ¹³C-enriched nor ¹³C-depleted populations of Prototaxites are consistent with the exclusive consumption of vascular plants, we conclude that bryophytic and algal alternatives to vascular plants were even more important as carbon sources for Prototaxites than previously hypothesized. By linking carbon isotope patterns in both ancient Prototaxites and modern Arrhenia to that of aquatic-derived photosynthesis, our results strengthen the suggestion by Hillier et al. (2008) that sedimentary deposits of cyanobacterial or algal-derived organic matter from ephemeral water bodies were probably important substrates for the subterranean hyphal network of Prototaxites.

Appendix A

Carbon isotope signatures of plants and Prototaxites from the Devonian given in figure 4. The estimated age is also given.

fossil taxon	age (Ma)	error (Ma)	sample type	δ13C (‰)	s.e. (‰)	experimental error (‰)	notes	reference
Aglaophyton major	410		axes	−25.7				Boyce et al. (2003)
Archaeopteris jacksonii	380	2	leaves and axes	−28.1	0.5		n = 2	Beerling et al. (2002)
A. macilenta	380	1	leaves and axes	−25.4				Beerling et al. (2002)
Asteroxylon mackiei	410		axes	−24.9				Boyce et al. (2003)
Bitelaria dubjanskii	399.5		cuticles	−24.1				Peters-Kottig et al. (2006)
Callixylon newberryi	376		wood	−27.7		0.1		Boyce et al. (2007)
C. newberryi	376		wood	−27.4		0.2		Boyce et al. (2007)
Crenaticaulis verruculosus	399.5		coalified tissue	−24.9				Peters-Kottig et al. (2006)
Eospermatoperis erianus	384	1	axes	−24.4				Beerling et al. (2002)
Leclercqia complexa	384		cuticles	−20.9				Peters-Kottig et al. (2006)
L. complexa	387	2	leaves and axes	−21.3	0.09		n = 3	Beerling et al. (2002)
Pertica varia	399.5		cuticles	−27.3				Peters-Kottig et al. (2006)
Prototaxites loganii	400			−15.8		0.1		Boyce et al. (2007)
P. loganii	396			−28.4		0.5		Boyce et al. (2007)
P. loganii	396			−28.7		0.1		Boyce et al. (2007)
P. loganii	396			−26.6		0		Boyce et al. (2007)
P. loganii	396			−26.6		0		Boyce et al. (2007)
P. loganii	396			−19		0.1		Boyce et al. (2007)
P. loganii	396			−15.7		0		Boyce et al. (2007)
P. southworthii	376			−28.9		0.1		Boyce et al. (2007)
P. southworthii	376			−27.2		1.0		Boyce et al. (2007)
Prototaxites spp.	392			−26.6		0		Boyce et al. (2007)
Prototaxites spp.	392			−27.4		0.5		Boyce et al. (2007)
Psilophyton forbesii	400	2	axes	−25.4				Beerling et al. (2002)
Psilophyton forbesii	396	2	axes	−26.3				Beerling et al. (2002)
Psilophyton princeps	396		axes	−23.6		1.4		Boyce et al. (2007)
Renalia hueberi	399.5		cuticles	−26.2				Peters-Kottig et al. (2006)
Rhacophyton ceratangium	365	3	axes	−24				Beerling et al. (2002)
Rhynia gwynne-vaughnii	410		axes	−24.2				Boyce et al. (2003)
coal (of cf. Sawdonia)	400			−23.5		0.4		Boyce et al. (2007)
Spongiophyton minutissimum	405		cuticle	−23.3			n = 6	Fletcher et al. (2004)
S. minutissimum	405		bulk	−24.6			n = 12	Fletcher et al. (2004)
S. minutissimum	405		thalli	−24.2a			n = 96	Jahren et al. (2003)
S. nanum	386		cuticle	−21.4			n = 2	Fletcher et al. (2004)
Tetraxylopteris schmidtii	384		cuticles	−22.6				Peters-Kottig et al. (2006)

^aRange −25.8‰ to −21.1‰.

Appendix B

Carbon isotope signatures for eight plant species along the Lyman Glacier foreland, ±s.e. (n) and mineral soil (0–5 cm depth), as shown in figure 3. (Plant mean ± s.e. is of species averages at each location.)

taxa	distance from glacial terminus (m)
	δ13C (‰)
	150	300	450	600	750
Abies lasiocarpa	−26.4 ± 0.3 (7)	−28.2 ± 0.4 (7)	−25.1 ± 0.5 (5)	−26.6 ± 0.3 (5)	−25.6 ± 0.4 (4)
Cassiope mertensiana	—	—	−24.9 ± 0.3 (4)	−25.9 ± 0.5 (5)	−25.2 ± 0.1 (5)
Epilobium latifolium	−28.1 ± 0.4 (2)	−28.4 ± 0.6 (5)	−28.5 ± 1.0 (2)	−27.5 ± 0.2 (5)	−26.9 ± 0.4 (5)
Luetkia pectinata	−29.8 ± 0.2 (4)	−28.7 ± 0.5 (4)	−27.7 ± 0.2 (4)	−28.8 ± 0.3 (6)	−27.2 ± 0.5 (6)
Luzula piperi	−28.3 ± 0.2 (5)	−26.8 ± 0.4 (5)	−27.1 ± 0.2 (6)	−28.9 ± 0.5 (4)	−27.0 ± 0.6 (4)
Phyllodoce empetriformis	−28.6 ± 1.9 (2)	−28.0 ± 0.5 (6)	−26.1 ± 0.3 (6)	−26.4 ± 0.3 (5)	−25.8 ± 0.5 (5)
Salix phylicifolia	—	−28.7 ± 0.3 (5)	−27.2 ± 0.4 (5)	−27.0 ± 0.2 (9)	−26.5 ± 0.3 (5)
Saxifraga ferruginea	−26.7 ± 0.6 (5)	−27.0 ± 1.0 (4)	—	−25.4 ± 0.3 (4)	−26.5 ± 0.7 (4)
plant mean	−28.0 ± 0.5 (6)	−28.0 ± 0.3 (7)	−26.7 ± 0.5 (7)	−27.1 ± 0.5 (8)	−26.3 ± 0.3 (8)
mineral soil (1999)a	−26.0 ± 0.4 (5)b	−25.6 ± 0.3 (5)	−25.5 ± 0.1 (4)	−25.5 ± 0.3 (5)	−25.4 ± 0.5 (3)
mineral soil (2009)c					−24.3 ± 0.3 (40)

^a1999 soils were sampled from upland areas not adjacent to water.

^bSampled at 100 m.

^c2009 soils sampled adjacent to eight pools at 700 m; values for soils at individual pools were −23.8 ± 0.3‰, −23.5 ± 0.6‰, −21.3 ± 0.5‰, −26.1 ± 0.2‰, −24.3 ± 0.3‰, −25.2 ± 0.3‰, −25.5 ± 0.3‰ and −24.7 ± 0.2‰.