L. azorica (Magnoliophyta, Magnoliopsida, Laurales, Lauraceae), the Azorean laurel, is a dioecious evergreen tree that grows up to 15 m height. The leaves (up to 15 cm long and 8 cm wide) are alternate, simple, entire, elliptic, oblong or obovate, acute and aromatic. Young twigs and leaves are brown-tomentose, becoming glabrous. Flowers are yellowish-green; perianth 4-lobed, segments ca. 4 mm. The fruits are fleshy, ellipsoid, up to 2 cm, black (when ripe) (Fig. 1) (Franco 1971, Schaefer 2005).

Endemic to the Azores, this species is common in submontane laurel forests and Juniperus-Ilex montane forests. Scattered or locally common in Picconia-Morella lowland forests, Juniperus montane woodlands and Pittosporum exotic forests (Elias et al. in press). It can also be found, rare or scattered, in Cryptomeria plantation forests. It occurs mainly between 100 and 900 m altitude, in all Azorean Islands.

Before Portuguese occupation and the process of forest cut for wood consumption and agriculture development, L. azorica was probably one of the most common tree species in the Azores. In fact, submontane laurel forests could have occupied more than 40% of the islands' surface and this species is also very frequent in montane forests that were the dominant vegetation between 600 and 900 m altitude in Faial, Pico, São Jorge, Terceira and São Miguel (Elias et al. 2016). However, landscape transformation by the Portuguese settlers affected mostly the laurel forests and L. azorica now occurs mostly in the remnant Juniperus-Ilex montane forests.

Data related to arthropods were obtained within the scope of the BALA project (Biodiversity of Arthropods in the Laurisilva of the Azores) that started in 1999 (Borges et al. 2005, Ribeiro et al. 2005). Laurus trees were selected and sampled in six Azorean Islands covering the western (Flores), central (Faial, Pico, Terceira) and eastern (São Miguel, Santa Maria) groups. The location map and an extensive description of Azorean Archipelago and vegetation can be found in Nunes et al. (2015). Experimental design can be found in detail in Borges et al. (2005) and Ribeiro et al. (2005). In summary, 100 transects of 150 m were sampled in seven Azorean Islands and 20 native forest reserves. In each transect, the three most frequent tree species were selected for sampling. We adopted a stratified sampling design: minimum of four transects for each forest fragment, but a different number of transects amongst Islands to reflect the area of forest reserves in the Islands. Considering the availability of Laurus trees, one forest fragment was monitored in Faial, two in Flores, five in Pico, six in São Miguel, three in Santa Maria and thirty-five in Terceira. For comparison between Islands, we standardised data to the number of samples available.

Arthropods were sampled using a beating tray. We used a modified beating tray, which consisted of an inverted cloth funnel pyramid of 1 m wide and 60 cm deep. A plastic bag was placed at the tip where arthropods, leaves and small branches were collected (Ribeiro et al. 2005). A beating tray is an efficient method to assess species diversity, abundance and distribution. Species were sorted and identified using a Leica M5 stereomicroscope, specific literature and a reference collection on the Azorean terrestrial arthropod biodiversity. When identification was not possible, we kept a morphospecies identifier to a given taxon. The specimens were deposited in the Entomological Collection Dalberto Teixeira Pombo at the University of the Azores. Each species was assigned to one of the three colonising statuses according to its distribution in the Azorean Archipelago (Borges et al. 2010): endemic (species restricted to the Azores), native non-endemic (species that arrived naturally to the Archipelago, but are also present elsewhere) and introduced (species accidentally or deliberately introduced by man). Species details can be consulted in Borges et al. (2016).

Two kinds of data were assessed for this inventory: literature records (Suppl. material 1) and herbarium records (Suppl. materials 2, 3, 4, 5). Most of the herbarium records were collected using standard collection protocols on native vegetation areas (Gabriel and Bates 2005, Borges et al. 2018) and others were collected for herborisation purposes. Samples were obtained from relevés with 30 cm or 10 cm-side, placed at different heights on the trees, allowing the estimation of cover and richness of species. Taxonomy for lichens follows Aptroot et al. (2010), while for bryophytes, taxonomy follows Hodgetts et al. (2020). All data are included in the Azores Bioportal (http://azoresbioportal.uac.pt/) for the general public.

More details on lichen and plants species sampling can be found in our previous work (Elias et al. 2019, Gabriel et al. 2019, Gabriel 2000). Vascular plants data, used in this study, result from the data collected by RBE and are listed in Suppl. material 2. The bryophytes and lichen dataset, used in this study, are listed in Suppl. materials 3, 4, 5 and are deposited in the Cryptogamic Collection of the Herbarium of the University of the Azores (AZU) (Angra do Heroísmo).

We described and compared the structure and the composition of species communities of L. azorica in different islands. The present analysis design follows the analysis plan of our previous study on taxa associated with Juniperus brevifolia (see Nunes et al. (2015) for more details). Since the sampling effort was different between islands (different number of transects available with Laurus), we standardised the data for comparison, but we used raw data when comparing different groups inside the same island.

Therefore, for species composition, we compared Islands for their species richness, abundance, functional groups and feeding modes and, for species community structure, we investigated patterns of rarity with species abundance distribution models and with species community similarity analyses.

Species composition

We collected a total of 10,313 specimens, corresponding to 94 species and morphospecies, 50 families, 13 orders and three classes (Table 1 and Suppl. material 6).

Species richness

Considering species richness of the different islands and in the Archipelago as a whole, we found that endemic, native and introduced species were evenly distributed and no difference was observed amongst the three groups regarding their number of species (Fig. 2, Table 2).

At Island level, we found that over 94 species observed at the Archipelago level and 71 species were collected in Terceira (Table 2). Standardising the sampling effort between Islands, we found that Santa Maria (SMR) was the most diverse Island, whereas São Miguel (SMG) was the least. The difference between SMR and SMG was significant (Suppl. material 7). The same pattern was observed within endemic and introduced species (Suppl. material 7). Regarding native species, a more complex pattern was observed: SMR was still the most diverse Island, but Flores (FLO) was the least and the difference was significant between the two Islands (Suppl. material 7).

Species collected belong to thirteen orders. We found that most species were spiders (Araneae) at the Archipelago level and in almost every Island. This finding was true for the overall species as well as for the three colonising groups. Lepidoptera and Hemiptera were the second most rich groups (Fig. 3).

Abundance

Considering the Archipelago as a whole, most specimens were from native species (about 50%), endemic species accounted for 44% of collected individuals, whereas introduced species account for only 6% of the collected individuals (Fig. 4, Table 2). The same pattern is observed for Islands: introduced species comprised the least abundant group, while endemic and native represented about 90% of the total abundance and near 98% in Faial. In most Islands, native species was the most abundant group, except in FLO and in SMG where endemic species were most abundant (Fig. 4, Table 2). Species from the three colonising groups were significantly different (Table 2).

The six most abundant species have 50% of the total abundance: the native Trioza laurisilvae Hodkinson, 1990 (Hemiptera) (n = 2675); the endemic spider Gibbaranea occidentalis Wunderlich, 1989 (Araneae) (n = 662); the endemic moth Argyresthia atlanticella Rebel, 1940 (Lepidoptera) (n = 510); the native spider Lathys dentichelis (Simon, 1883) (Araneae) (n = 498), the endemic spider Savigniorrhipis acoreensis Wunderlich, 1992 (Araneae) (n = 34) and the endemic tree hopper Cixius azoterceirae Remane & Asche, 1979 (Hemiptera) (n = 426) (see details in Suppl. material 6). The top ten most abundant species include three more Hemiptera and a Blattaria all native or endemic.

Faial Island shows the highest number of specimens and Flores the lowest (Suppl. material 8). In addition, native species were more abundant in FAI than in the other Islands, but the difference was significant only between FAI and FLO (Suppl. material 9). The same pattern was observed when considering the total abundance (Suppl. material 9). Within endemic species, the significant difference was between SMG and SMR (Suppl. material 9) and within introduced species, the significant difference was between FAI and SMR (Suppl. material 9).

Considering the Archipelago as a whole (Fig. 5), Hemiptera were the most abundant group over all species (4958 specimens representing 48%) and also within native species (3316 specimens representing 65%), whereas Araneae were the most abundant group within introduced and endemic species, accounting respectively, for 64% and 39% (Fig. 5). Hemiptera were the second most abundant group (36%) within the endemic species, while they represented less than 1% of specimens within the introduced species. Amongst the introduced species, lepidopterans and booklice were the second and third most abundant species, representing 14% and 13%, respectively. Opiliones, Pseudoscorpiones, Julida, Coleoptera, Microcoryphia, Trichoptera and Thysanoptera altogether accounted for less than 1% for the total abundance and also within the different colonisation status groups, except Julida which represents 6% of the introduced species (Fig. 5).

At Island level, Hemiptera and Araneae account for more than 75% of species abundances in all species, endemic or native species groups. A different pattern was observed within introduced species; spiders, lepidopterans and booklice were the most abundant groups (Fig. 5). Although these patterns were globally observed in the different Islands, we found some differences. For example, in Flores, most of the introduced species were booklice and millipedes, while in Pico, they were spiders and millipedes (Fig. 5).

Functional groups and feeding modes

Herbivores and predators represent about 80% of total species abundance and richness at the Archipelago level, as well as at the Island level. Herbivores were represented by 41 species (6159 specimens) and predators by 37 species (3225 specimens) (Fig. 6).

A comparison between Islands showed that abundance of herbivores per unit sample was higher in FAI and the difference was significant between FAI and FLO (Suppl. material 10). No differences were observed between abundances of predators in the different Islands, except between PIC and SMG. Saprophytes were more abundant in SMR and the difference was significant between SMR and FAI (Suppl. material 10). Fungivore species was the poorest group present only in three Islands SMR, SMG and Terceira (TER) and representing less than 1% species and abundance (two species and seven individuals, Suppl. material 8).

Crossing functional and colonising groups, we found that most herbivores were endemics and native species, while most predators were introduced species. The distribution pattern was observed both for abundance and species richness (Suppl. materials 11, 12).

Species fall into four different feeding modes: external digestion and sucking; chewing and cutting; piercing and sucking; and siphoning. Very few species, 5% of the total species abundance (six species and 570 individuals), were siphoning species amongst them, five species were endemic species and one introduced species. Most individuals exhibited piercing and sucking feeding mode representing about 50% of the overall species abundance (5075 individuals). However, the chewing and cutting group was the most diverse group represented by 35 species (37% and 1581 individuals) (Fig. 7, Suppl. material 13).

Crossing analysis between feeding mode and colonising groups revealed a pattern similar to the functional group. Most endemic and native species were herbivores with piercing and sucking feeding mode and most introduced species were predators with external digestion and sucking (Suppl. materials 14, 15).

Species community structure

In this study, we considered rare species as those represented by seven individuals or less (see Gaston 1997). Thus, 45% of the overall species collected were considered rare. At the Archipelago level, about 79% of introduced species and 45% of native species were rare, while only 17% of endemic species were considered rare (Table 3). These proportions are highly variable between the different Islands, but the common pattern is that most of introduced species are rare, ranging from 70% in SMR to 100% in FAI (Table 3).

Overall, 13% of species were represented by only one individual (singleton). At the Archipelago level, about 29% of introduced species and 10% of native species were represented by only one individual, whereas only 3% of endemic species were singletons. The proportions of singletons species were variable between the different Islands and the higher proportions were found within introduced species ranging from 25% in FLO Island to 80% in FAI Island (Suppl. material 16). Suppl. material 16 also provides proportions of doubletons and tripletons species.

Species abundance distribution patterns

Preston’s abundances frequency distribution, also called octaves (Preston 1948), revealed interesting species distribution shapes.

The overall assemblage at the Archipelago level, showed a poisson log-normal distribution shape (PLN) (A) with abundances distributed into 12 octaves and 45% of species falling in the first three octaves 0, 1, 2 and, therefore, considered rare species (Gaston 1997). This calculation method agrees with the manual method that we described in the section above. The native Hemiptera species Trioza laurisilvae Hodkinson falls in the octave 12 with 2675 individuals (Fig. 8).

Considering the three colonising groups, native species assemblage also showed the PLN shape Fig. 8C) and species abundances were distributed into 12 octaves. This method showed that about 45% of native species were rare species, falling into octaves 0, 1 and 2. The SAD model for endemic species (Fig. 8B) showed a log-normal shape (LN) with species distributed into 10 octaves and 17% of rare species. The Introduced group curve was a log-series (LS)-shape with species distributed into seven octaves and about 54% of rare species (Fig. 8D). Analysing species abundance distribution of the different Islands, we found variable patterns (Suppl. material 17). When considering the total species in Islands, gambin models shapes showed a PLN-shape in almost every Island, except in Faial which showed a LS-shape. An analysis of species separated in the different colonising groups revealed more details. SADs shapes of endemic species showed a PLN-shape in TER and SMR, but more LS-like in the other Islands. Conforming with native species at the Archipelago level, native species in Islands showed a LN-shape model, except in Faial where the model was more LS-shaped. SADs of introduced species showed a LS-shape in each Island, except in FLO where four species were collected and the four fall into different octaves, therefore presenting a quite flat gambin model.

Analysis of gambin α-values are consistent with the shapes of the different models. Considering the Archipelago level, endemic species group showed the highest α-value (7.55) followed by native (2.32) and introduced (1.15) species. The same pattern was observed in the different Islands, except in FLO, where the introduced species group has the highest α-value (7.08), followed by native (3.52) and endemic (2.33) species (Fig. 9, Table 4).

Species community similarity

We found high similarity between communities of Terceira (TER) and Pico (PIC) in one hand and, on the other hand, high dissimilarity between FLO and FAI (lowest (PIC-TER) and highest (FLO-FAI) Bray-Curtis index values) (Table 5). This result was not explicitly consistent with the number of shared species between the two Islands. Considering all species and native species, the highest number of shared species was between TER and SMG (46 and 17 species, respectively, for all species and native species). Within endemic species, the highest number of shared species was between TER and PIC (18 species) and between PIC and SMG (18 species). Within introduced species, the highest number of shared species was between TER and SMR (nine species). The Non-metric Multidimensional Scaling (NMDS) ordination shows Terceira species communities at the centre of the graphic with the other Islands displayed around Terceira (Fig. 10).

NMDS ordination also revealed that the different Islands were closer or more distant depending on whether we considered all species or species separated in different colonising statuses (Fig. 10A, B, C and D). Stress values of the NMDS ordination for the four investigated groups (all species, endemic, native and introduced) are very similar (respectively 0.182, 0.194, 0.138 and 0.175) (Fig. 10).

Considering all species: NMDS ordination (Fig. 10A) showed that species communities of PIC and TER were closer which is consistent with the lower dissimilarity index (value 0.50 in Table 5) observed between the two Islands. The graphical representation showed a high dissimilarity between FLO, SMG and SMR which is also confirmed by the high values of the Bray-Curtis index. Although two sampling points of FAI were close to FLO and SMG, the other points were scattered apart and the Bray-Curtis index values were very high ranging from 0.78 to 0.866 (Table 5).

Within endemic species, NMDS ordination (Fig. 10B) showed a clear distinct pattern indicating independent communities with low shared species amongst the different Islands. TER and SMR were the closest (index value = 0.560) and FAI and SMG the most distant (index value =0.91) (Table 5). NMDS ordinations of native species showed that FAI was apart from the other Islands (Fig. 10C). Bray-Curtis index values confirmed the ordination showing high values of indices ranging from 0.756 to 0.858 (Table 5). Contrasting the results of endemic and native species, ordination of introduced species shows high overlap between Islands (Fig. 10D). Islands sampling points were scattered and ravelled indicating the homogeneity of introduced species in the Archipelago (Fig. 10D).

ANOSIM sustained the structure observed in the different groups. Dissimilarities between the different Islands were all significant (P < 0.001) and R² values indicated the highest dissimilarity within endemic species (R² = 0.80), intermediate dissimilarity when considering all species (R² = 0.66) or native species (R² = 0.42) and very low dissimilarity within introduced species (R² = 0.36).

Vascular plants

There are only four vascular plants, epiphytes of L. azorica and they are all ferns: Hymenophyllum tunbrigense (L.) Sm.; Vandenboschia speciosa (Willd.) G.Kunkel; Elaphoglossum semicylindricum (T.E.Bowdich) Benl; Polypodium macaronesicum subsp. azoricum (Vasc.) Rumsey, Carine & Robba (see Suppl. material 2 for colonisation status and more details on their classification).

Bryophytes

We found that L. azorica was the substrate for both liverworts and mosses. Liverworts were represented by 57 species (three orders Jungermanniales, Metzgeriales and Porellales, comprising 18 families) (Suppl. material 3). Three species belong to the order Metzgeriales, while 28 species belong to Order Porellales (49%) and 26 to Order Jungermanniales (46%). Liverwort species may be found growing on living bark (epiphytic) or leaves (epiphyllic) and on decaying wood (epixylic). An important group of epiphyllous liverworts may be found growing on L. azorica leaves, including for instance: Frullania microphylla, Drepanolejeunea hamatifolia, Cololejeunea microscopica, Lejeunea lamacerina, Myriocoleopsis minutissima, Metzgeria furcata and Colura calyptrifolia. The Vulnerable Cololejeunea azorica, may also colonise L. azorica leaves. A relatively small number of liverwort species has been found growing on decaying L. azorica, including Plagiochila bifaria, Frullania tamarisci and Telaranea europaea.

The 35 moss species belong to five orders (Bryales, Dicranales, Hookeriales, Hypnales and Orthotrichales). Most of species belong to order Hypnales (19 species; 54%), whereas Dicranales accounts for about a quarter of the species (nine species; 26%) (Suppl. material 4). Most of the species are epiphytes, including the Endangered Daltonia lindigiana, Echinodium renauldii and Thamnobryum rudolphianum. It is interesting to note that Hypnum uncinulatum (Least Concern) and Andoa berthelotiana (Vulnerable) have been found growing on very old L. azorica leaves.

Lichens

The lichen community was composed of 32 species sorted into four classes (Arthoniomycetes, Dothideomycetes, Eurotiomycetes and Lecanoromycetes) and 13 identified orders (Arthoniales, Caliciales, Lecanorales, Monoblastiales, Ostropales, Peltigerales, Pertusariales, Pleosporales, Pyrenulales, Strigulales, Teloschistales, Trypetheliales, Verrucariales) (Suppl. material 5). Amongst lichens, four species are epiphyllous and the others are epiphytic.

Communities species composition

Colonising status

Arthropods species communities on L. azorica were dominated by native and endemic species. The ten most abundant species are all endemic or native (Suppl. material 6). The two groups were present in high proportion at the Archipelago level, but also at Island level, ranging from 87% (in S. Maria - SMR) to 98% (in Faial - FAI) of the total abundances. However, in terms of species richness, the number of introduced species was lower, but the difference between the three colonising groups was not significant. These findings support the results of a previous study on arthropods species associated with Juniperus brevifolia canopies. Introduced species were represented by very few specimens (4%) and species (30%). The observed pattern might be explained by some characteristics of the L. azorica environment: (i) native forest: native species communities found in L. azorica benefit from the stability offered by native forest, whereas introduced species are hampered because they depend on disturbance-related factors (Borges et al. 2006, Borges et al. 2008, Florencio et al. 2013); (ii) habitat accessibility - the structural complexity of canopy might constrain the establishment of introduced species; (iii) niche saturation and species competition - the high diversity and abundance of canopy communities, as well as the dominance of predators species (e.g. Araneae, we developed this part below) on L. azorica contributed to saturation of ecological niches and, therefore, no places were left for introduced species and (iv) climate austerity - forest canopies, especially Azorean native forest canopies located at high elevations, are prone to climatic hazards including wind and rapid climate conditions turnovers. Some of these hypotheses have been developed in a previous study on J. brevifolia canopy (Nunes et al. 2015).

These results support those of a recent study in Azorean native forest, but whose samples were obtained with SLAM traps. Considering four dominant orders (Araneae, Coleoptera, Lepidoptera and Psocoptera), very few specimens of introduced species (6%) were collected, whereas the number of species was not different for endemic or native species groups (Tsafack et al. 2021b). A previous study, targeting both canopy and soil species of Azorean native forests, also found few specimens of introduced species (11%) compared to endemic and native species abundance (Gaspar et al. 2008). However, instead of no difference in species richness between the three biogeographical groups as we found, Gaspar et al. (2008) suggested the presence of high number of introduced species (34%) which is similar to our findings. The fact that the Gaspar et al. (2008) study included many plants and also soil samples might explain this discrepancy.

Taxonomic composition

Amongst the thirteen orders collected in this study, Hemiptera was the most abundant group and Araneae was the most diverse group. Together, Hemiptera and Araneae assemble more than 75% of abundances and more than 50% of number of species collected. However, within introduced species, Araneae was both the most abundant and the most diverse group. The ten most abundant species are mostly composed by Araneae and Hemiptera along with a moth and a cockroach (Suppl. material 6). The most abundant species Trioza laurisilvae Hodkinson, 1990 (Hemiptera) is considered a specialist of L. azorica (see Rego et al. 2019).

Our results support previous findings (Gaspar et al. 2008) which suggested that Araneae and Hemiptera were the most important groups in terms of number of species and abundance. However, contrary to our results, Hemiptera was the most diverse group and Araneae the most abundant group. In addition, Gaspar et al. (2008) results suggest that Coleoptera was the most diverse group in the canopy, whereas in the present study, very few beetles species (> 5%) were found on L. azorica. The differences might be explained by the fact that the current study is focused on the L. azorica tree, whereas Gaspar et al. (2008) included all trees in the native forests. The prevalence of Araneae and Hemiptera species was also observed on J. brevifolia canopies (Nunes et al. 2015).

Functional and feeding groups

Our results revealed the dominance of herbivores and predators species representing up to 80% of both number of species and number of specimens. The proportion of functional groups is consistent with the taxonomic composition that we previously developed. In fact, most of Hemiptera species being herbivores and all spider species being predators, the proportion of functional groups observed was then foreseeable. This is similar to functional groups observed on J. brevifolia (Nunes et al. 2015), on Erica azorica (Ribeiro et al. 2005) and on other endemic tree canopies of Azorean Islands (Gaspar et al. 2008, Rego et al. 2019). Saprophytes accounted for about 5% of both abundance and number of species. Fungivores were represented by very few individuals (seven individuals) belonging to two species observed in three Islands (SMG, SMR and TER).

We found that this general pattern (dominance of herbivores tailed by predators and few saprophytes) was a common pattern observed in the different Islands. Moreover, the distribution of functional groups within the colonising groups was similar with some exceptions for the introduced taxa. The two fungivores species belong to the introduced taxa. We observed a dominance of saprophytes specimens in Flores (FLO) (> 50% of abundance).

Regarding species feeding mode, species communities found on L. azorica were mostly piercing and sucking species corresponding to species of the order Hemiptera and, therefore, representing about 50% of the overall species abundance (5075 individuals). Distribution patterns of species according to their feeding mode was closely related to their functional groups. Most endemic and native species were herbivores with piercing and sucking feeding mode, whereas most introduced species were predators with external digestion and sucking (Rigal et al. 2018).

Community species structure

Investigations on rarity and similarity patterns show that community structure at the Archipelago level contrast most of Islands communities’ structures. Some Islands, like Terceira (TER) and Santa Maria (SMR), showed high similarity with the study at the Archipelago level. However, for the other Islands, similarities with Archipelago level seem to depend on the colonising status of the species.

Rarity patterns

At the Archipelago level as well as for all Islands, a gambin model log-series (LS) shape fitted introduced species distribution, indicating that most of the introduced species were rare. LS-shape models are characteristic of simple species communities with dominance of rare species and very few represented by high number of individuals (Ulrich et al. 2016, Tsafack et al. 2021a). This finding is consistent with our investigations on species composition, where we found that, at Archipelago level, 79% of introduced species were rare. In some Islands like Faial, all introduced species (100%) were considered rare.

On the other side, gambin models fitting endemic and native species distributions fitted log-normal (LN) and poisson log-normal (PLN) shapes. LN distributions are generally considered describing stable communities (May 1975).

The present study suggests that introduced species communities are not yet well established on the L. azorica canopy which seems to be a stable refuge for indigenous (endemic and native) species communities.

Similarity patterns

Patterns of similarity between different Islands mainly rely on the colonising status, whether species were endemic, native or introduced species. Considering all species, no clear pattern emerged, communities of the six Islands overlapped, showing high similarities in species assemblages.

Contrary to communities collected on Juniperus brevifolia (Ribeiro et al. 2005, Nunes et al. 2015), these were not observed as a clustering, mirroring the geographical distance between islands. At best, two Islands of the central group (PIC and TER) were closer, while the two Islands of the eastern group (SMG and SMR) were clearly distant. This stochastic overlap pattern was emphasised when we considered native and introduced species. This study highlights the high similarity on native species assemblages between Islands and, moreover, between introduced species. High similarity was also observed within introduced species collected on J. brevifolia (Nunes et al. 2015). However, as expected, endemic species assemblages offer a sharper distinct clustering between Islands. However, this pattern also does not rely on the geographical distance between Islands.

All species of vascular plants collected on L. azorica (four species) were epiphytes species and none was hemiparasites, contrasting the hemiparasite found on J. brevifolia (Arceuthobium azoricum Wiens & Hawksw) (Nunes et al. 2015). One species was common to L. azorica and J. brevifolia (Hymenophyllum tunbrigense (L.) Sm).

L. azorica shelters a diverse community of lichen and bryophytes, but we found that species diversity was lower than on Juniperus brevifolia community. This might be explained by the crinkled structure of J. brevifolia which offers diverse micro-habitats in all parts of the tree (Nunes et al. 2015). Contrary to J. brevifolia, L. azorica trunk and branches are rather smooth without cracks that favour retention of water and other nutrients which might serve as resources for arthropods or plants species. Furthermore, the pH values are quite different between the two phorophyte species and this is an important factor for discriminating different bryophyte communites (Gabriel and Bates 2005).

It is worth stressing that L. azorica supports a rich epiphyllous community, a feature which is characteristic of the Macaronesian mature forests, but very rare in other temperate habitats. More than 20 species have been found growing on L. azorica' leaves, adding a whole new layer of life to the Azorean forests.

About a fifth of the bryophytes, found associated with L. azorica, are IUCN conservation concern' species: seven mosses (three species, endangered; four species, vulnerable) and 12 liverworts (seven species endangered; five species vulnerable). Amongst the identified threats to the conservation of native ecosystems and species, both the habitat (Hodgetts et al. 2019) and climate change (Patiño et al. 2016) are important issues to avoid biodiversity erosion. Thus, the presence of epiphyllous bryophytes, which are relatively easy to recognise in the field, should be monitored as an indicator of habitat changes in the Azorean forests.

Our study identifies the contribution of the endemic tree L. azorica in supporting arthropods and plant species communities in native forest fragments. Although L. azorica seems to support poor communities compared to J. brevifolia, we found that profiles of species distribution provide clear insights on overall species distribution in native forest. Canopy community distribution confirms the results obtained in a previous study which suggest the stability of native and endemic species communities over introduced species community in native forests fragments (Tsafack et al. 2021b).

At the Azorean scale, the study warns again generalisations, suggesting that most Islands present a particular species distribution pattern without geographical correlation and that conservation programmes should be adapted to each Island.