Fungal species diversity and occurrence
A total of 1325 leaves (CF: 275, FA: 275, FS: 275, MT: 275 and MA: 225) were collected and 236 fungal taxa recorded (CF: 48, FA: 46, FS: 64, MT: 42 and MA: 36). One hundred and fifteen taxa (CF: 27, FA: 27, FS: 28, MT: 19 and MA: 14) were isolated into culture and identified to species level, using both morphology and phylogeny (Tennakoon et al. 2019a, b, 2020, 2021b).The percentage occurrences of fungal species andthe Shannon diversity index (H) for each host species are presented in Tables 1–6.The time for fungal communities to reach the peak of species diversity was 28 days for Ficus ampelas (H = 2.89), F. septica (H = 3.33), Macaranga tanarius (H = 3.17) and Morus australis (H = 2.82), but 35 days forCeltis formosana (H = 3.27) (Table 1). For each host species, the Shannon diversity index ranged from 0.5 to2.01 at the beginning of decomposition, but the index steadily increased and and reached the peak of about 2.8–3.3 at 28–35 days (Table 1). After that, fungal diversity and the total number of fungal species started to decline. The average diversity of fungi on each host species in the experimental period ranged between 1.51 and 2.61, viz. Celtis formosana (H = 2.38), Ficus ampelas (H = 2.27), F. septica (H = 2.61), Macaranga tanarius (H = 2.28) and Morus australis (H = 1.52). Species richness also increased from the start of decomposition (1–10 taxa), reached a maximum at middle stages (25–35 taxa) and steadily declined in the final stages (5–15 taxa) (Fig. 3). All identified species belonged to Ascomycota. No Basidiomycota species were observed during the decomposition process in this experiment.TheNMDS analysis indicated that the fungal community composition was significantly different among the selected tree species (Fig. 4).
Fungal community composition in Ficus ampelas
A total of 275 Ficus ampelas leaves were examined during thedecay process and the percentage occurrences were calculated (Table 2). On day 63, leaves were highly skeletonized comprising vascular tissues with attached remnants of non-vascular tissues. Fourty-six fungal taxa were recorded and 27 species were successfully isolated into cultures and identified to species level (Tennakoon et al. 2019a, b, 2020, 2021b). This included a new genus (Longihyalospora), nine new species (Acrocalymma ampeli, Cercophora fici, Colletotrichum fici, Longihyalospora ampeli, Micropeltis fici, M. ficinae, Neoanthostomella fici, Neofusicoccum moracearum and Phaeosphaeria ampeli) (Tennakoon et al. 2019a,b, 2020, 2021b) and eleven new host records (Beltrania rhombica, Ceramothyrium longivolcaniforme,Coniella quercicola, Diaporthe limonicola, D. pseudophoenicicola, Discosia querci, Lasiodiplodia thailandica, Pseudopestalotiopsis camelliae-sinensis, Torula fici, Wiesneriomyces laurinus and Yunnanomyces pandanicola).The findings have been published by Tennakoon et al. (2019b, 2020, 2021b) and are used for comparison here. An additional 21 species were identified to genus level, and belonged to Aspergillus,Backusella, Cylindrocladium, Fusarium, Mortierella, Mycosphaerella, Penicillium, Periconia, Phoma, Phyllosticta, Rhizopus, Volutella and Zygosporium (Table 2). Five taxa were unidentified Ascomycete species (due to lack of morphology data).
The most dominant species (POC>10%) onFicus ampelas leaves wereDiaporthe limonicola (40%), followed by Neoanthostomella fici (37.82%), Acrocalymma ampeli (36.73%), Micropeltis fici (36.36%), Longihyalospora ampeli (30.18%), Ceramothyrium longivolcaniforme (18.91%), Beltrania rhombica (18.55%), Pseudopestalotiopsis camelliae-sinensis (16.73%), Neofusicoccum moracearum (15.27%), Torula fici (14.91%), Diaporthe pseudophoenicicola (11.64%), Discosia brasiliensis (10.91%) and Yunnanomyces pandanicola (10.55%) (Fig. 5).
The percentage occurrences of fungal communities using Hierarchical Cluster Analyses (HCA), showed presence of at least four distinct communities during succession: Stage I (day 0–3), Stage II (day 7–21), Stage III (day 28–35) and Stage IV (day 42–63) (Fig. 5). In Stage I (0–7 days) the species number was low with low percentage occurrences. The dominant species at this stage were Ceramothyrium longivolcaniforme, Diaporthe limonicola, Longihyalospora ampeli, Micropeltis fici and M. ficinae. The dominant species of Stage II (7–21 days) were Acrocalymma ampeli, Beltrania rhombica, Diaporthe limonicola, Longihyalospora ampeli, Micropeltis fici, Neoanthostomella fici and Yunnanomyces pandanicola. The highest species diversity was noted during Stage III (28–35 days) and the dominant species were Acrocalymma ampeli, Beltrania rhombica, Diaporthe limonicola, D. pseudophoenicicola, Mycosphaerella sp., Neoanthostomella fici, Neofusicoccum moracearum, Pseudopestalotiopsis camelliae-sinensis, Torula fici and Wiesneriomyces laurinus. In Stage IV (42–63 days), the species diversity and number of species declined. The community was dominated by only a few species with relatively high percentage occurrence. Dominant species were Cercophora fici, Fusarium sp. 4, Mortierella sp., Neoanthostomella fici, Penicillium sp. 1, Pseudopestalotiopsis camelliae-sinensis and Wiesneriomyces laurinus.
The highest species diversity as indicated by Shannon diversity index was at day 28 (H = 2.89) and lowest at day 63 (H = 1.50). The average diversity of fungi on Ficus ampelas leaves during the experimental period was 2.27. The species diversity changes during the experimental period are shown in Table 1.
Fungal community composition in Ficus septica
A total of 275 Ficus septica leaves were examined during thedecay process and the percentage occurrenceswere calculated(Table 3). On day 63, leaves were highly skeletonized comprising vascular tissues with attached remnants of non-vascular tissues. Sixty-four species were recorded and 28 species were successfully isolated into cultures and identified to species level.These includea new family (Cylindrihyalosporaceae), a new genus (Cylindrihyalospora), 16 new species (Bertiella fici, Colletotrichum fici-septicae, Conidiocarpus fici-septicae, Coniella fici, Cylindrihyalospora fici, Diaporthe fici-septicae, Diplodia fici-septicae, Discosia ficeae, Leptodiscella sexualis, Microthyrium fici-septicae, Muyocopron ficina, Mycoleptodiscus alishanense, Neophyllachora fici, Ophioceras ficina, Pseudocercospora fici-septicae andStictis fici) and twelve new host records (Arthrinium malaysianum, Conidiocarpus betle, Lasiodiplodia thailandica, L. theobromae, Neopestalotiopsis phangngaensis, Parawiesneriomyces chiayiense, Periconia alishanica, Pestalotiopsis portugallica, Robillarda roystoneae, Sirastachys castanedae, Stachybotrys microspore and Torula fici). All the new findings have been published by Tennakoon et al. (2021b) and are used for comparison here. Another 27 species were identified to genus level, which comprised Acremonium, Appendiculella, Arthrinium, Aspergillus, Camarographium, Colletotrichum, Cylindrocladium, Fusarium, Mucor, Penicillium, Phoma, Phyllosticta, Pleurostoma, Pseudocercospora, Rhizopus, Syncephalastrum and Torula (Table 3).Two species were identified as Botryosphaeriaceaespecies and three species as coelomycete species. Four taxa were unidentified Ascomycetes (due to lack of morphology data).
The dominant species (POC>10%) onFicus septica leaves wereConiella fici (36.36%), followed by Colletotrichum fici-septicae (36%), Lasiodiplodia thailandica (35.27%), Diaporthe fici-septicae (29.45%), Diplodia fici-septicae (28.36%), Conidiocarpus fici-septicae (22.18%), Leptodiscella sexualis (22.18%), Sirastachys castanedae (21.82%), Conidiocarpus betle (21.81%), Neophyllachora fici (20%), Ophioceras ficina (20%), Lasiodiplodia theobromae (17.09%), Bertiella fici (16.36%), Pestalotiopsis portugallica (16%), Parawiesneriomyces chiayiense (16%), Arthrinium malaysianum (15.63%), Microthyrium fici-septicae (14.54%), Muyocopron ficina (14.18%), Stachybotrys microspora (13.82%), Aspergillus sp. (13.45%), Cylindrihyalospora fici (12.73%) and Robillarda roystoneae (12.73%) (Fig. 6).
The percentage occurrences of fungal communities using Hierarchical Cluster Analyses (HCA), showed presence of at least four distinct communities during succession: Stage I (day 0–7), Stage II (day 14–21), Stage III (day 35–42) and Stage IV (day 49–63) (Fig. 6). During Stage I (Day 0–7), fungal communities comprised low speciesnumber with low percentage occurrence. The dominant species at this stage were Appendiculella sp., Colletotrichum fici-septicae, Conidiocarpus betle, C. fici-septicae, Cylindrocladium sp. 1, Microthyrium fici-septicae, Mycoleptodiscus alishanense, Neophyllachora fici, Phyllosticta sp., Rhizopus sp. and Stachybotrys microspora (Fig. 6). The dominant species of Stage II (14–21 days) were Colletotrichum fici-septicae, Conidiocarpus betle, C. fici-septicae, Coniella fici, Cylindrihyalospora fici, Diaporthe fici-septicae, Diplodia fici-septicae, Discosia ficeae, Leptodiscella sexualis, Lasiodiplodia thailandica, Muyocopron ficina, Pseudocercospora fici-septicae and Torula fici (Fig. 6). The highest species diversity was present during Stage III (35–42 days) and dominant species were Arthrinium malaysianum, Aspergillus sp., Bertiella fici, Colletotrichum fici-septicae, Conidiocarpus betle, Coniella fici, Cylindrihyalospora fici, Diaporthe fici-septicae, Diplodia fici-septicae, Lasiodiplodia thailandica, L. theobromae, Leptodiscella sexualis, Muyocopron ficina, Ophioceras ficina, Periconia alishanica, Pestalotiopsis portugallica, Robillarda roystoneae, Sirastachys castanedae, Stachybotrys microspora and Stictis fici (Fig. 6). At Stage IV (49–63 days), the species diversity and number of species declined. The community was dominated by only a few species with relatively high percentage occurrence. Dominant species were Arthrinium malaysianum, Bertiella fici, Fusarium sp., Lasiodiplodia thailandica, L. theobromae, Mucor sp., Parawiesneriomyces chiayiense, Penicillium sp. and Syncephalastrum sp. (Fig. 6).
Shannon diversity indices showed that the species diversity was highest at day 28 (H = 3.33) and lowest at day 0 (H = 1.56) (Table 1). The average diversity of fungi on Ficus septicaleaves in the experimental period was 2.61. The species diversity change during the experimental period is shown in Table 1.
Fungal community composition in Celtis formosana
A total of 275 Celtis formosana leaves were examined during thedecay process and the percentage occurrences were calculated(Table 4). On day 63, leaves were highly skeletonized comprising vascular tissues with attached remnants of non-vascular tissues. Fourty-eight fungal taxa were recorded and 27 species were successfully isolated into culture and identified up to species level. These include eight new species (Arxiella celtidis, Colletotrichum celtidis, Diaporthe celtidis, Discosia celtidis, Memnoniella celtidis, Muyocopron celtidis, Periconia celtidis and Pseudoneottiospora cannabina) and 19 new host records (Arthrinium hydei, Bartalinia robillardoides, Coniella quercicola, Dematiocladium celtidicola, Diaporthe millettiae, Dimorphiseta acuta, Dinemasporium parastrigosum, Lasiodiplodia theobromae, Muyocopron dipterocarpi, M. lithocarpi,Neopestalotiopsis asiatica, Ophioceras chiangdaoense, Pestalotiopsis dracaena, P. papuana, P. trachycarpicola, Phyllosticta capitalensis, Pseudorobillarda phragmitis, Sirastachys pandanicola, Strigula multiformis).All the new findings have been published by Tennakoon et al. (2021b) and are used for comparison here. Another 16 taxa were identified to genus level and belong to Appendiculella, Aspergillus, Asteridiella, Cladosporium,Cylindrocladium, Fusarium, Idriella, Mucor, Penicillium, Phoma, Trichoderma and Zygosporium (Table 4). Two taxa were unidentified Ascomycete species (due to lack of morphology data) and another two were Coelomycetes.
The dominant species (POC>10%) onCeltis formosana leaves wereMuyocopron celtidis (37.82%), followed by Strigula multiformis (37.45%) Arxiella celtidis (34.91%), Coniella quercicola (33.45%), Pestalotiopsis dracaenea (24%), Sirastachys pandanicola (23.64%), Dimorphiseta acuta (22.91%), Discosia celtidis (21.09%), Muyocopron dipterocarpi (17.09%), Memnoniella celtidis (16.73%), Lasiodiplodia theobromae (16.36%), Phyllosticta capitalensis (16.36%), Periconia celtidis (16%), Dinemasporium parastrigosum (15.27%), Pestalotiopsis trachycarpicola (15.27%), Diaporthe millettiae (14.91%), Coniella quercicola (14.91%), Bartalinia robillardoides (14.55%), Dematiocladium celtidicola (14.18%), Pseudorobillarda phragmitis (14.18%), Pestalotiopsis papuana (13.45%), Diaporthe celtidis (12.73%) and Arthrinium hydei (10.18%) (Fig. 7).
The percentage occurrences of fungal communities using Hierarchical Cluster Analyses (HCA), showed presence of at least four distinct communities during succession: Stage I (day 0–7), Stage II (day 7–14), Stage III (day 21–42) and Stage IV (day 49–63) (Fig. 7). The Stage I (0–7 days) communities were low in species number and had a low percentage occurrence. The dominant species at this stage were Appendiculella sp., Arxiella celtidis, Aspergillus sp., Colletotrichum celtidis, Cylindrocladium sp. 1, Diaporthe celtidis, Phyllosticta capitalensis and Strigula multiformis (Fig. 7). The dominant species of Stage II (7–14 days) were Arxiella celtidis, Coniella quercicola, Diaporthe millettiae, Dimorphiseta acuta, Fusarium sp. 1, Ophioceras chiangdaoense, Pestalotiopsis dracaenea, P. trachycarpicola, Phyllosticta capitalensis, Pseudorobillarda phragmitis and Strigula multiformis (Fig. 7). The highest species diversity was present during Stage III (21–42 days), dominant species were Arthrinium hydei, Arxiella celtidis, Bartalinia robillardoides, Coniella quercicola, Dematiocladium celtidicola, Diaporthe millettiae, Dimorphiseta acuta, Discosia celtidis, Lasiodiplodia theobromae, Muyocopron celtidis, M. dipterocarpi, M. lithocarpi, Periconia celtidis, Pestalotiopsis dracaenea, Pestalotiopsis papuana, Pestalotiopsis trachycarpicola, Pseudorobillarda phragmitis, Sirastachys pandanicola and Strigula multiformis (Fig. 7). In Stage IV (49–63 days), the species diversity and number of species declined. The community was dominated by a few species with relatively high percentage occurrence. Dominant species were Aspergillus sp., Dinemasporium parastrigosum, Discosia celtidis, Fusarium sp., Lasiodiplodia theobromae, Memnoniella celtidis, Muyocopron celtidis, Muyocopron dipterocarpi, Periconiaceltidis and Pseudoneottiospora cannabina (Fig. 7).
Shannon diversity indices showed that the species diversity was highest at day 35 (H = 3.27) and lowest at day 0 (H= 0.53). The average diversity of fungi on Celtis formosana leaves in the experimental period was 2.38. The species diversity change during the experimental period is shown in Table 1.
Fungal community composition in Morus australis
A total of 225 Morus australis leaves were examined during thedecay process and the percentage occurrences are listed in Table 5. On day 49, leaves were highly skeletonized comprising vascular tissues with attached remnants of non-vascular tissues. Thirty-six fungal species were recorded and 14 species were successfully isolated into cultures and identified to species level.These include six new species (Arthrinium mori, Memnoniella alishanensis, M. mori,Periconia alishanica, Phaeodothis mori, Pseudopithomyces mori) and 11 new host records (Alternaria burnsii, Arthrinium paraphaeospermum, A. rasikravindrae, Cladosporium tenuissimum,Gilmaniella bambusae, Pestalotiopsis formosana, P. neolitseae, P. parva, Pseudopithomyces sacchari, Pseudorobillarda phragmitis, Spegazzinia musae). All the new findings have been published by Tennakoon et al. (2020, 2021b) and are used for comparison here. Another 17 species were identified to genus level and belong toAppendiculella, Aspergillus, Cercospora, Cladosporium, Fusarium, Gilmaniella, Mucor, Neopestalotiopsis, Penicillium, Syncephalastrum and Trichoderma (Table 5). Two taxa were unidentified Ascomycetes (due to lack of morphology data).
The dominant species (POC>10%) ofMorus australis leaves were Arthrinium mori (42.22%), followed by Pestalotiopsis formosana (36.89%), Spegazzinia musae (28.89%), Phaeodothis mori (26.67%), Pseudopithomyces mori (18.67%), Gilmaniella bambusae (17.33%), Pestalotiopsis neolitseae (14.67%), P. parva (13.33%), Memnoniella mori (12.89%) and Alternaria burnsii (11.56%) (Fig. 8).
The percentage occurrences of fungal communities using Hierarchical Cluster Analyses (HCA) showed presence of at least four distinct communities during succession: Stage I (day 0–7), Stage II (day 7–14), Stage III (day 21–42) and Stage IV (day 42–49) (Fig. 8). The Stage I (0–7 days), fungal communities were low in number and had a low percentage occurrence. The dominant species at this stage were Alternaria burnsii, Appendiculella sp., Cercospora sp., Pestalotiopsis formosana, P. parva and Pseudopithomyces mori (Fig. 8). The dominant species of Stage II (Day 14) were Arthrinium mori, Cercospora sp., Fusarium sp., Pestalotiopsis formosana, Phaeodothis mori and Spegazzinia musae (Fig. 8). The highest species diversity was present during Stage III (21–42 days) and dominant species were Arthrinium mori, A. paraphaeospermum, A. rasikravindrae, Cladosporium sp. 1, C. tenuissimum, Fusarium sp. 1, Gilmaniella sp., Memnoniella mori, Penicillium sp., Pestalotiopsis formosana, P. neolitseae, P. parva, Phaeodothis mori, Pseudopithomyces mori, Pseudorobillarda phragmitis and Spegazzinia musae (Fig. 8). In Stage IV (Day 49), the species diversity and number of species declined. The community was dominated by a few species with relatively high percentage occurrence. The dominant species were Arthrinium mori (Fig. 8).
Shannon diversity indices showed that the species diversity was highest at day 28 (H = 2.82) and lowest at day 0 (H = 1.27) (Fig. 4). The average of diversity of fungi on Morus australis leaves in the experimental period was 1.52. The species diversity change during the experimental period is shown in Table 1.
Fungal community composition in Macaranga tanarius
A total of 275 Macaranga tanarius leaves were examined during thedecay process and the percentage occurrences were calculated(Table 6). On day 63, leaves were highly skeletonized comprising vascular tissues with attached remnants of non-vascular tissues. Fourty-two fungal species were recorded and 19 species were successfully isolated into cultures and identified to species level. These include one new family (Oblongihyalosporaceae), two new genera (Neodictyosporium and Oblongihyalospora), eight new species (Diaporthosporella macarangae, Leptospora macarangae, Oblongihyalospora macarange, Parawiesneriomyces chiayiense, Periconia alishanica, Memnoniella alishanensis, Neodictyosporium macarangae and Nigrospora macarangae) and eleven new host records (Alternaria burnsii,A. pseudoeichhorniae, Arthrinium sacchari, Cladosporium tenuissimum, Dictyocheirospora garethjonesii, Hermatomyces biconisporus, Memnoniella echinata, Periconia byssoides, P. celtidis, Pseudopithomyces chartarum and Stachybotrys aloeticola). All the new findings have been published by Tennakoon et al. (2021b) and are used for comparison here. Another 17 species were identified to genus level and belong to Aspergillus, Asteridiella, Cladosporium, Hermatomyces, Idriella, Meliola, Penicillium, Phoma and Phyllosticta (Table 6). Six taxa were unidentified Ascomycetes (due to lack of morphology data).
The dominant species (POC>10%) ofMacaranga tanarius leaves wereDiaporthosporella macarangae (77.45%), followed by Leptospora macarangae (33.09%), Periconia byssoides (30.18%), Pseudopithomyces chartarum (29.46%), Stachybotrys aloeticola (22.18%), Periconia alishanica (17.82%), Penicillium sp. 1 (17.09%),Alternaria pseudoeichhorniae (14.91%), Aspergillus sp. 1 (12.73%), Alternaria burnsii (12.36%), Memnoniella alishanensis (11.64%),Hermatomyces biconisporus (11.27%), Meliola sp. 1 (11.27%), Parawiesneriomyces chiayiense (11.27%) and Neodictyosporium macarangae (10.91%) (Fig. 9).
The percentage occurrences of fungal communities using Hierarchical Cluster Analyses (HCA) showed presence of at least four distinct communities during succession: Stage I (day 0–3), Stage II (day 7–14), Stage III (day 21–42) and Stage IV (day 56–63) (Fig. 9). The Stage I (0–3 days) fungal communities were low in species number and had a low percentage occurrence. The dominant species at this stage were Diaporthosporella macarangae, Meliola spp. and Phyllosticta sp. (Fig. 9). The dominant species of Stage II (7–14 days) were Alternaria burnsii, Diaporthosporella macarangae, Leptospora macarangae, Memnoniella alishanensis, Oblongihyalospora macarange, Penicillium sp. 1, Periconia byssoides and Pseudopithomyces chartarum (Fig. 9).The highest species diversity was present during Stage III (21–42 days), dominant species were Alternaria pseudoeichhorniae, Arthrinium sacchari, Aspergillus sp. 1, Cladosporium spp., Diaporthosporella macarangae, Hermatomyces biconisporus, Leptospora macarangae, Neodictyosporium macarangae, Nigrospora macarangae, Parawiesneriomyces chiayiense, Periconia alishanica, P. byssoides, P. celtidis, Phoma sp. 2, Pseudopithomyces chartarum and Stachybotrys aloeticola (Fig. 9). In Stage IV (56–63 days), the species diversity and number of species declined. The community was dominated by a few species with relatively high percentage occurrence. Dominant species were Cladosporium spp., Diaporthosporella macarangae, Leptospora macarangae, Parawiesneriomyces chiayiense, Penicillium sp. 1, Periconia alishanica, P. byssoides, Pseudopithomyces chartarum and Stachybotrys aloeticola (Fig. 9).
Shannon diversity indices showed that the species diversity was highest at day 28 (H = 3.17) and lowest at day 0 (H = 1.55). The average of diversity of fungi on Macaranga tanarius leaves in the experimental period was 2.28. The species diversity changes during the experimental period are shown in Table 1.
Trophic modes of taxa in fungal communities
The fungi identified in this succession study were classified by trophic modes using the fungal traits ver. 1.2 online database (https://docs.google.com/spreadsheets/d/1cxImJWMYVTr6uIQXcTLwK1YNNzQvKJJifzzNpKCM6O0/edit#gid=33668129) (Põlme et al. 2020). Eight trophic modes were observed, viz. foliar epiphytes, endophytes, lichenized spp., litter saprotrophs, plant pathogens, soil saprotrophs, sooty molds and unspecified saprotrophs (Fig. 10). Most identified fungi were litter saprotrophs (61.1%) and this trophic mode was recorded throughout the decomposition process. Other trophic modes weresoil saprotrophs (14.4%), plant pathogens (10.4%), foliar endophytes (3.9%), sooty molds (1.7%), foliar epiphytes (1.3%) and lichenized species (0.8%). Fungal species with many trophic modes (e.g., Colletotrichum spp., Pestalotiopsis spp.) were enumerated in each one.
Host phylogeny and fungal composition
The findings of degree of separation (R value) of fungal community composition within selected tree species is shown in Fig. 11. Statistically, a high R value implies a strong variation in fungal composition between the selected host species, whereas a low R value indicates a minor difference in fungal composition. Accoding to the results, highest R value recorded in between Celtis formosana and Macaranga tanarius (R = 0.88, P < 0.001), while lowest R value recorded in between Ficus septica and Morus australis (R = 0.62, P < 0.001). However, host phylogeny weakly correlated with fungal composition during the decomposition process (Fig. 11). Even phylogenetically closely related tree species have higher R value, than phylogenetically distant host species. For example, phylogenetically close Ficus ampelas and F. septica (same genus) has high R value (R = 0.78, P < 0.001) than phylogenetically distant Ficus septica and Celtis formosana (R = 0.65, P< 0.001). In addition, Ficus ampelas and F. septica R value is much higher than Ficus septica and Morus australis (R = 0.62, P< 0.001) as well. In adition, the percentage of overlapping species in other tree species pairs also indicates that host phylogeny is associated to a lesser degree with fungal diversity in this study (Fig. 12). For example, FA is more phylogenetically close to MA (members of the same family), than to either CF (different family) or MT (different order). However, the number of overlapping species between FA-MA is less than FA-CF and FA-MT (Fig. 12). In general, the expectation would be that the number of overlapping species between phylogenetically close tree species should be higher than those that are more phylogenetically distant. Another example is that FS is phylogenetically closer to MA than CF and MT, but result indicates that there are high number of overlap species in FS-CF and FS-MT than FS-MA (Fig. 12). Therefore, based on our results we conclude that host phylogeny has a lower association with fungal composition than expected.
Leaf litter chemistry
Total N content differed among tree species during decomposition, but it increased in all hosts as decomposition progressed (Fig. 13). The highest initial total N content was recorded in Morus australis (2.11%) and Celtis formosana (1.73%). The initial total N contents of the other three host species (Ficus ampelas, F. septica and Macaranga tanarius) varied between 1.19–1.27% (Fig. 13). Total N content and fungal composition of leaf litter were significantly correlated (r2 = 0.61, P = 0.001, Fig. 14).In contrast, the total C content of all leaf litter decreased with decomposition (Fig. 13), and there was a significant correlation with fungal composition according to the goodness-of-fit-statistics (r2) analysis (r2 = 0.40, P = 0.001, Fig. 14). The highest initial total C was observed in Macaranga tanarius (49.58%) leaf litter and the lowest in Morus australis (37.86%).
As a result of N enrichment, the C/N ratio decreased in the leaves of all host species overtime (Fig. 13). The highest initial C/N ratio was revealed in Macaranga tanarius (38.94) leaf litter and lowest in Morus australis (17.94). The initial C/N ratio of the other three host species (Ficus ampelas, F. septica and Celtis formosana) varied between values of 22.87–32.75 (Fig. 13). The C/N ratio and fungal composition of leaf litter were significantly correlated (r2 = 0.54, P = 0.001, Fig. 14). Lignin content also significantly corresponded with fungal composition (P = 0.001, r2 = 0.25), but it was much lower when compared to Lignin: N (r2 = 0.47, P = 0.001) ratios. It is worthy to note that lignin content increased with decomposition time and accumulated in the final stages of decay (Fig. 13).
Host species heterogeneity, leaf litter chemistry and sampling times shape fungal community succession
Variation partitioning analysis of factors demonstrated that host speciesheterogeneity (31%) and leaf litter chemistry (24%) accounted for the large variation in fungal community composition in the five hosts. Sampling time explained a smaller proportion of variance (5%) (Fig. 14). Goodness-of-fit-statistics (r2) also revealed significant correlations between fungal composition and host speciesheterogeneity (r2 = 0.65, P = 0.001).However, this was considerably lower when comparing host phylogeny (r2 = 0.21, P = 0.001) and host families (r2 = 0.11, P = 0.001). Both variation partitioning and goodness-of-fit-statistics indicate that leaf litter chemistry is significantlycorrelated to fungal community composition (Fig. 14). In particular, total N shows the strongest correlation (r2 = 0.61, P = 0.001), while lignin content shows a much lower correlation with fungal composition (r2 = 0.25, P = 0.001).The significant correlation of host species can be due to leaf litter quality, either physical or chemical characteristics of the particular tree species leaves. These include leaf toughness, particle size, leaf surface properties (physical characteristics) and initial C or N content, C/N ratio, soluble sugars, polyphenols, waxes, cellulose, hemicellulose, and lignin content (chemical characteristics) are subjected to leaf litter quality of the tree species and ultimately on fungal communities (Promputtha et al. 2002; Bothwell et al. 2014; Purahong et al. 2016; Ge et al. 2017; Osono 2017; Zhao et al. 2020).