Stable isotope abundance analysis

Five 1m² plots were selected at each site; each plot included fully MH orchids and four to five autotrophic reference plant species. Flower stalks of seven fully MH orchids, leaves of four to five autotrophic reference plants and soil samples from the organic layer were taken from each of the five plots at each site. The reference plants collected at the respective sites are listed in Supplementary Data Table S2. In addition, on site B, fruit bodies of a litter-decaying SAP fungus (Marasmius sp.) were found and collected in five replicates.

Samples were dried at 105 °C, ground to a fine powder and stored in a desiccator with silica gel until analysed. Relative N and C isotope abundances of the samples were measured using a dual-element analysis mode with an elemental analyser coupled to a continuous flow isotope ratio mass spectrometer as described in Bidartondo et al. (2004). Measured abundances are denoted as δ values that were calculated according to the given equation δ¹⁵N or δ¹³C=(R_sample/R_standard−1)×1000 [‰], where R_sample and R_standard are the ratios of heavy isotope to light isotope of the samples and the respective standard. Standard gases (N₂ and CO₂) were calibrated with respect to international standards by using the reference substances N1 and N2 for N isotopes and ANU sucrose, NBS 18 and NBS 19 for C isotopes, provided by the International Atomic Energy Agency (Vienna, Austria). Reproducibility and accuracy of the isotope abundance measurements were routinely controlled by measuring the test substance acetanilide (Gebauer and Schulze, 1991). At least six test substances with varying sample weight were routinely analysed in each batch of 50 samples. The maximum variation of δ¹³C and δ¹⁵N within and between batches was always <0·2 ‰. To compare isotope abundances of orchids and reference plants from different sites, the data were normalized. Enrichment factors (ε) were calculated per plot: ε=δ_S−δ_REF, with S as a single δ¹³C or δ¹⁵N value of an adult orchid and REF as the mean value of non-orchid reference plants from the respective plot (Preiss and Gebauer, 2008). The original δ¹³C and δ¹⁵N values of orchids, the respective reference plants, fungi and soil samples are available in Supplementary Data Table S2. Total N concentrations in leaf, stem, fungus and soil samples were calculated from sample weights and peak areas using a six-point calibration curve per sample run based on acetanilide measurements (Gebauer and Schulze, 1991). Acetanilide has a constant N concentration of 10·36%. Enrichment factors and total N concentrations of orchids and reference plants were tested for normal distribution. Enrichment factors ε¹³C and ε¹⁵N were not normally distributed and therefore these data were tested for statistical differences using the Kruskal–Wallis non-parametric test followed by a post-hoc Mann–Whitney U-test with an adjusted significance level according to Holm (1979). The autotrophic reference plants were treated as one group after confirming insignificant differences among the data of each species. Total N concentrations were normally distributed and thus the Student’s t-test was used to test total N concentrations in orchids and reference plants for statistical differences.

Stable isotope natural abundance and total N concentrations

Comparisons of enrichment factors ε¹³C (H=77; d.f.=7; P<0·001) and ε¹⁵N (H=77; d.f.=7; P<0·001) among the seven orchid species and the set of reference plants revealed highly significant differences in our data set. All study orchids were enriched by 7·9±0·2 ‰ (G. appendiculata) to 12·1±0·5 ‰ (G. flabilabella) in ¹³C and by 4·7±0·6 ‰ (G. flabilabella) to 8·8±2·7 ‰ (L. thalassica) in ¹⁵N in comparison with autotrophic reference plants growing at the same sites (Fig. 8). Based on post-hoc tests, this enrichment in ¹³C and ¹⁵N was highly significant for all orchid species (in all cases U=0; P<0·001).

Open in a separate window

Fig. 8.

Mean enrichment factors ε¹³C and ε¹⁵N±1s.d. as calculated for six MH orchid species associated with SAP non-Rhizoctonia fungi (triangles), one mycoheterotrophic orchid species associated with a fungus forming ECM (filled square), one litter-decaying SAP fungus (open circle) (n=5 each) and of photosynthetic reference plants (Ref, n=90, see Supplementary Data Table S2) collected together with each of the orchids at four sites in the Xitou Experimental Forest, Nantou County, Taiwan. Please note that the orchids associated with SAP non-Rhizoctonia fungi fall into two significantly distinct groups, Galeola falconeri, Cyrtosia javanica, Gastrodia flabilabella (filled triangles) and Gastrodia fontinalis, G. nantoensis and G. appendiculata (open triangles). Both of these groups are significantly different from Lecanorchis thalassica, the orchid associated with an ECM fungus. Mean ε values of reference plants are zero by definition. The box represents±1s.d. for the reference plants.

The orchids themselves fall into three distinct groups. The ECM-associated L. thalassica was significantly more enriched in ¹⁵N than all other orchids associated with SAP fungi (U=0 to U=3; P<0·001 to P=0·003; mean difference 3·7 ‰) and significantly less enriched in ¹³C than the group composed of G. falconeri, C. javanica and G. flabilabella (U=0; P<0·001; mean difference 3·0 ‰). However, L. thalassica was not significantly distinguishable in its ¹³C enrichment from the orchid group composed of G. nantoensis, G. appendiculata and G. fontinalis (U=25; P=0·295). The group composed of G. falconeri, C. javanica and G. flabilabella was more enriched in ¹³C than the group composed of G. nantoensis, G. appendiculata and G. fontinalis (U=0; P<0·001; mean difference 3·5 ‰). However, both of these groups were not significantly different in their ¹⁵N enrichments (U=107; P=0·836).

The investigated litter-decaying SAP fungus Marasmius sp. was enriched by 9·4±0·5 ‰ in ¹³C and by 4·3±0·2 ‰ in ¹⁵N compared with autotrophic reference plants (Fig. 8) and thus had similar enrichments in ¹³C and ¹⁵N to the orchid group composed of G. nantoensis, G. appendiculata and G. fontinalis.

Orchid flower stems had a slightly higher mean total N concentration (2·88±0·50mmol g d.wt^–1; n=35) than reference plant leaves (2·67±0·67mmol g d. wt^–1; n=90; Supplementary Data Table S2). However, this slight difference was not significant (t=1·671; P=0·097; d.f.=123).

Carbon and nitrogen sources and total nitrogen concentrations

All seven of the MH study orchids turned out to be significantly enriched in the heavy isotopes ¹³C and ¹⁵N in comparison with autotrophic plants growing in identical habitats. This finding confirms that fungi also known to be enriched in ¹³C and ¹⁵N (Gebauer and Dietrich, 1993; Gleixner et al., 1993) serve as their C and N source. Though leafless, some putative MH orchids have green flowering stems and thus a low photosynthetic activity. One leafless orchid with frequently green flowering stems and green seed capsules is Corallorhiza trifida. This species is associated with an ECM fungus and is less enriched in ¹³C and ¹⁵N than fully MH orchids also associated with ECM fungi (Zimmer et al., 2008). Thus, C. trifida has been classified as being partially MH. All study orchids of this investigation provided no indications for chlorophyllous flowering stems and had isotope signatures typical of various types of fully MH plants. According to their pattern in ¹³C and ¹⁵N enrichment, the MH orchids fall into three groups. (1) The ECM-associated L. thalassica is the species most enriched in ¹⁵N. This pattern is in agreement with current knowledge (Hynson et al., 2013) and can be traced back to the fact that ECM fungi are more enriched in ¹⁵N than sympatrically growing wood- or litter-decaying SAP fungi (Gebauer and Taylor, 1999). (2) The two orchid species associated with wood-decaying Meripilaceae, Galeola falconeri and Cyrtosia javanica, belong to the group most enriched in ¹³C. The specifically high ¹³C enrichment of this group of MH orchids is expected due to the fact that wood is more enriched in ¹³C than leaf tissue (Gebauer and Schulze, 1991; Cernusak et al., 2009) and confirms previous findings of similar patterns in MH orchids associated with wood-decaying fungi (Ogura-Tsujita et al., 2009; Martos et al., 2009; Hynson et al., 2013). The third MH orchid belonging to this group is the Hydropus-associated Gastrodia flabilabella, suggesting that this Hydropus sp. should be a wood-decaying fungus. (3) The third MH orchid cluster characterized by a significantly lower ¹⁵N enrichment than (1) and a significantly lower ¹³C enrichment than (2) is composed of the three Gastrodia spp. growing sympatrically in a bamboo forest, G. appendiculata, G. fontinalis and G. nantoensis. The isotopic pattern of this cluster of MH orchids provides strong evidence that their mycorrhizal associates, Mycena of three different types and Gymnopus, are all litter-decaying fungi. This evidence is further strengthened by the rather similar isotopic pattern found for the fruit bodies of the litter-decaying fungus Marasmius sp. This is the first report we are aware of that indicates a significantly different isotopic pattern for MH orchids associated with either wood-decaying or litter-decaying fungi. The only hint towards an isotopic distinction between MH orchids associated with wood- or litter-decaying fungi was previously given by Martos et al. (2009), who investigated the isotopic compositions of the MH orchids Gastrodia similis associated with wood-decaying Resinicium from La Réunion and Wullschlaegelia aphylla mycorrhizal orchids with litter-decaying Gymnopus and Mycena from Guadeloupe. Unfortunately, comparisons of the data from the study of Martos et al. with each other and with our investigation based on normalized enrichment factors is not possible due to a lack of stable isotope data for forest ground plants growing in micro-environments identical to the MH orchids.

Total N concentrations in MH orchids have been reported to be significantly higher than in the majority of co-occurring autotrophic plants (Gebauer and Meyer, 2003; Liebel et al., 2010; Liebel and Gebauer, 2011; Sommer et al., 2012). The most likely reason for these unusually high total N concentrations in MH orchid tissues is their N gain through the fungal source. Fungi are known to have considerably higher total N concentrations in their tissue than autotrophic plants growing in the same environment (Gebauer and Dietrich, 1993; Gebauer and Taylor, 1999). Total N concentrations found in the flower stems of the MH orchid species investigated here (2·88±0·50mmol g d.wt^–1) are rather similar to the total N concentrations reported for other fully or partially MH orchids (Gebauer and Meyer, 2003; Liebel et al., 2010; Liebel and Gebauer, 2011; Sommer et al., 2012). The reference plant leaf total N concentrations investigated here (2·67±0·67mmol g d.wt^–1) were only slightly lower than N concentrations in the flower stems of the MH orchids and statistically not distinguishable. Total N concentrations in the leaves of the understorey plants investigated here from four different forests in Taiwan are about twice as high as leaf total N concentrations in autotrophic forest ground plants from temperate Central Europe (Gebauer et al., 1988; Gebauer and Meyer, 2003) and Mediterranean Italy (Liebel et al., 2010), and three times higher than leaf total N concentrations in autotrophic forest ground plants from severely N-limited regions, such as SW Australia (Sommer et al., 2012) or boreal Norway (Liebel and Gebauer, 2011). The most likely reason for the high total N concentrations in the leaves of autotrophic forest ground plants in Taiwan is a high mineral N availability due to high soil N mineralization rates. Thus, N limitation as a driver for the switch from autotrophy towards mycoheterotrophy is unlikely for the MH orchids investigated here. More likely as a driving factor for the development of mycoheterotrophy is light limitation for photosynthesis on the forest ground. As reported for other habitats of MH plants (Bidartondo et al., 2004; Zimmer et al., 2007; Preiss et al., 2010) only 2–6% of incoming light reached the ground of our study forests and thus severely limited the photosynthetic capacity of forest ground vegetation.

In conclusion, our data provide further evidence for the importance of associations with non-Rhizoctonia SAP fungi among fully MH orchids from sub-tropical Asia. Abundant fallen litter and wood in the warm and humid forests of Taiwan obviously supplies ideal substrates for the continuous growth of SAP fungi. For the MH orchids studied here, as seeds germinate, they can construct efficient mycorrhizal interactions preferably with wood- or litter-decaying SAP fungal partners, but sympatrically also with ECM fungi, in order to thrive in deeply shaded forest understoreys with low light conditions.