PMC

[A] displays pigment contents (in μg cm^-2) as box-whisker diagrams separated into the four different sample groups "natural mountain" (NM), "natural valley" (NV), "phytometer mountain" (PM) and "phytometer valley" (PV). [B and C] display relations of flavonoids and chlorophylls to OTU richness. Values differed significantly between phytometer and natural trees (KS test, D = 0.9, p < 0.01). The pigment concentrations were significantly higher in natural mountain than in natural valley samples (D = 1, p = 0.01), but did not differ within the phytometer samples (D = 0.55, p = 0.43).

Diversity indices for the full [A] and rarefied [B] data set. A general trend of higher fungal diversity is displayed for natural samples (NV = natural valley, NM = natural mountain) at least for alpha and Shannon diversity [compare ]. The leaf mycobiome of natural valley samples (NV) displayed the highest diversity values. All four indicators were significantly higher than those of natural mountain samples (NM). For both data sets, the factor elevation obtained insufficient statistical support for phytometer samples (PV = phytometer valley, PM = phytometer mountain) but resulted in statistical significance within natural trees.

[A-C] for the full data set and [D-F] for a smaller data set after rarefying (randomised downsampling) to 9500 reads per sample. [A and D] display read abundances of all natural and phytometer samples after data curation (removal of rare and unassigned OTUs). [B and E] display the interrelations between read abundance and OTU richness. Phytometer samples had lower richness values than natural samples irrespective of sequencing depth. [C and F] display randomised species accumulation curves allowing the comparison of OTU richness at an identical sequencing effort. For both data sets the leaf mycobiome of phytometer trees revealed poorer richness than of natural trees.

[A] shows a map of Germany with the origin of the phytometer trees ("p"-labelled circle, Baumschule Hans Reinke GmbH in Rellingen near Hamburg, Germany) and the location of the major investigation area ("i"-labelled circle, Marktschellenberg, Bavaria, Germany). [B] Leaves of naturally grown trees were sampled in two locations at different altitudes of the investigation area ("v"-labelled valley site and "m"-labelled mountain site). This image of the Untersberg has been released by its author TomK32 into the public domain of German Wikipedia (https://commons.wikimedia.org/wiki/File:Untersberg-Westseite_von_Bischofswiesen_aus.jpg, last accessed 04/2016). The phytometer sites were established there, too. [C and D] The phytometer trees were planted in each altitude and two data logger modules (d1 and d2) were installed.

[A] displays results from non-metric multidimensional scaling (NMDS) showing clearly separated phytometer and natural samples and a strong dispersion of the phytometer samples. [B] Principal coordinate analysis (PCO) confirmed this bisection into phytometer and natural fungal assemblages. It additionally displays those environmental parameters, which had a significant influence on fungal composition (chl.avg = chlorophyll content, flav.avg = flavonoid content). The higher dispersion of phytometer samples persisted in PCO and points towards a larger heterogeneity of the corresponding leaf mycobiome. [C] displays taxonomic composition of leaf-inhabiting fungi from phytometer trees, natural samples and the entire beech mycobiome for the nine most abundant orders (see text for the type of abundance measuring). All OTUs without taxonomic representation at the order level in the reference data set ("unidentified fungi") as well as less abundant orders (e.g. Botryosphaeriales, Diversisporales or Malasseziales) are combined into "Others".