The Arthropod Fauna of Oak ( Quercus spp., Fagaceae) Canopies in Norway

: (1) We document the invertebrate fauna collected from 24 oak canopies in east and west Norway as a contribution to the Norwegian Biodiversity Information Centre’s ‘The Norwegian Taxonomy Initiative’. (2) A snap-shot inventory of the canopies red listed species found reﬂects how poor the Norwegian insect fauna is still studied. Moreover, the implication of the IUCN red list criteria for newly described or newly observed species is discussed.

Emitting insecticides into the forest canopy to sample invertebrates has opened up a new area of forest biodiversity research. Originally developed in the tropics, canopy fogging techniques are now being used increasingly in temperate forests to increase the knowledge of European arboreal fauna [11,14,[24][25][26][27][28][29][30][31][32][33][34][35][36][37][38]. Stork and colleagues [34] discuss the efficiency of fogging as a method for sampling arthropods from the canopies. A larger spectrum of species is sampled compared with any other single method. This makes fogging a useful method for arthropod snapshot inventories. The major disadvantage is that external and internal feeders are underrepresented (phloem feeders, leaf miners and wood borers), non-obligate occasional by-passers (tourists) will be captured and that the method is sensitive to wind and precipitation [39,40].
This study presents empirical data and analyses of oak canopy invertebrate data from a survey of 24 oak canopies in Norway. We proposed the following hypotheses: 1) there are large geographical differences in species composition and 2) trees on cultivated lands (Berge and Mule Varde) have a different species composition than forest trees. Both hypotheses are related to climatic differences on macro-(H1) and microlevels (H2) (e.g., [41]) as well as the geography of Norway, where oaks are distributed along the coast, usually with scattered populations [42,43]. H2 is founded on the generally more uniform structure of managed lands and lack of a multi-layered canopy of such forest stands [44]. The project was granted by the Norwegian Biodiversity Information Centre as a part of the Norwegian Taxonomy Initiative to search for the hidden life and new species in Norway.

The Oaks
Quercus robur and Q. petraea have a sympatric distribution and often hybridize [45], though Q. robur is claimed to be more widespread [42,43]. Thus, we have not distinguished between the two species of oak or their hybrids in this study.

Site Descriptions
The study was carried out at six sites in southern Norway in June-July 2011 and 2012 ( Figure 1, Table 1). All sites were continuous oak-dominated forests, except Berge (site 1) and Mule Varde (site 5), which had oak trees scattered on managed land. Four oaks were treated at each site. The sites were carefully selected to represent a gradient from the inner fjords of West Norway, via known biodiversity hotspots inland Vestfold and Telemark to the coastal areas of SE Norway [8,[46][47][48] aligned with the hypotheses.
Site 1 (Berge) is a protected landscape area and classified as IUCN category V [49]. It contains the largest assemblage of old and pruned oak trees in the country. This and the proximity to a lake with specialized swamp vegetation and several old buildings are the main reasons for its protection status [47].  Site 1 (Berge) is a protected landscape area and classified as IUCN category V [49]. It contains the largest assemblage of old and pruned oak trees in the country. This and the proximity to a lake with specialized swamp vegetation and several old buildings are the main reasons for its protection status [47].
Site 2 (Skeianeset) is a steep slope facing south and has according to one of the highest concentrations of hollow, previously pruned oaks in Norway [46]. The area is characterized by having an unusually high proportion of red-listed species of plants, bryophytes and fungi and is considered to be one of the most important deciduous forests in West Norway [46]. Site 3 (Steinknapp) is a nature reserve that is known to harbor many rare and threatened species (e.g., [48]). Its importance for biological diversity also explains its status as a nature reserve (IUCN category IA). Most likely, large parts of this area were clear-cut in the past as really old oaks are sparsely present and the more or less continuous oak forest is rather homogenous. The oaks treated in this study were just outside of the reserve.  Site 2 (Skeianeset) is a steep slope facing south and has according to one of the highest concentrations of hollow, previously pruned oaks in Norway [46]. The area is characterized by having an unusually high proportion of red-listed species of plants, bryophytes and fungi and is considered to be one of the most important deciduous forests in West Norway [46]. Site 3 (Steinknapp) is a nature reserve that is known to harbor many rare and threatened species (e.g., [48]). Its importance for biological diversity also explains its status as a nature reserve (IUCN category IA). Most likely, large parts of this area were clear-cut in the past as really old oaks are sparsely present and the more or less continuous oak forest is rather homogenous. The oaks treated in this study were just outside of the reserve.
Site 4 (Djupedal) is also a nature reserve protected according to the IUCN IA criteria. In contrast to the nearby site 3, there are several giant oaks in this area and the forest is characterized as old growth. Moreover, the forest is more closed and heterogeneous than at site 3. Site 6 (Skjaersjø) is a mixed deciduous forest with larger areas of conifer woods intermixed.

Data Collection
The trees were chosen to represent 'typical' trees in the areas. This implies that after traversing the site, the chosen trees were not at the edges, not standalone trees except for at Berge and Mule Varde where most trees were standalone. Furthermore, the biggest and smallest trees were also avoided. Arthropod sampling was performed by emitting a 1% concentration of natural pyrethrum, Py-Sekt, into the canopy using a Golden Eagle 2610E fogger for approximately 10 minutes in the period between 1 AM and 3 AM on a windless and dry night. Py-sekt contains 1-5% piperonyl butoxide and 0-1% pyrethrum [50]. It breaks down quickly in direct sunlight and is, therefore, relatively safe to use in natural environments [51]. The available space for arthropods will obviously vary both according to the breadth and height of the crown, but for practical reasons we preferred to collect knocked-down invertebrates from a fixed area. Twenty large butterfly nets (18 with Ø50 cm and 2 with Ø100 cm, mesh size from 0.3-0.5 mm) were mounted on the ground or on the lower branches beneath the crown to collect the knocked-down invertebrates, i.e., 5.11 m 2 of the area beneath each tree was sampled. As so, the proportion of the crown projection area covered will vary slightly between individual trees but is assumed not to affect the qualitative data. The nets remained on the ground for approximately one hour after fogging before the collected material was transferred to 80% ethanol. The material was then sorted and shipped to the co-authors of this paper for identification, with the exception of Lepidoptera and cecidomyiid midges, which remain unidentified.
Most of the material is stored in the Natural History Museum at the University of Oslo and the Norwegian Institute of Bioeconomy Research's entomological collection. The phorid flies are at the Zoological Museum at Cambridge University, England, and a part of sciarid material, including the holotype of Bradysia quercina Menzel and Köhler, 2014, is deposited at the Senckenberg Deutsches Entomologisches Institut, Müncheberg, Germany.

Species Records
Species designated as new records for Norway or the Nordic countries at the time of identification were based on the individual expert's consideration, but also on published records in Fauna Europaea [52] and records in the Norwegian Biodiversity Information Centre's species record database accessed throughout the preparation of this manuscript at www.artsdatabanken.no. Specimens fully identified to species level were included in the analyses and counted in addition to unidentified species with only one species collected in the respective higher taxon. Uncertain species identifications, i.e., denoted with confer (cf.) or near, were included when the species were not already identified with certainty from other specimens. In cases where the identity of the species was clear, yet undefined (i.e., denoted as sp., sp. 1, etc.), the species beyond the number of identified species were counted. When more unidentified species within the same genus were found, i.e., spp., they were not included in the counts except for counting 1 when no other species in that genus was found. Abundances of common species of spiders and collembolans were sometimes indicated as 'few', 'some' and 'many', and were thence given dummy numbers 5, 10 and 20, respectively.

Data Analyses
Rarefaction curves extrapolated to three times the sample size, i.e., 72 trees were carried out with EstimateS, version 9.1.0 [53]. The extrapolation relies on statistical sampling methods rather than modeling. Here, the bias corrected form of Chao1 is the asymptotic richness estimator for individual-based abundance data [54]. We chose to extrapolate because rarefaction curves of insect assemblages are usually steep and do not converge unless Diversity 2021, 13, 332 5 of 31 a massive sampling effort is conducted. However, extrapolation beyond three times the sample size is not recommended [53] because the variance increases with the extrapolation.
Whittaker's β was calculated as a measure of species turnover along the sampling gradient. It is insensitive to species richness and is calculated as follows: where S = total number of species, α max = highest number of species in any one locality and N = the number of localities [55]. It ranges from zero (no turnover) to 100 (every locality has a unique set of species). These calculations were performed to complement multivariate analysis using detrended correspondence analysis (DCA) with Canoco, version 4.56 [56] to relate species composition and site characteristics along the sampling gradient. The aim was to investigate whether the species composition within a site differed from the composition of species at the other sites and relate that to environmental characteristics. DCA assumes unimodal species responses to environmental factors in contrast to principal components analysis, or its detrended equivalent, where linear responses are assumed [57]. Therefore, over a longer geographic gradient with different climatic or other underlying environmental factors, DCA is to be preferred. The multivariate analysis was performed on untransformed species abundances with downweighing rare species.

Faunistics
Combined, more than 6800 specimens were identified to 722 species. Ninety-two species (12.7%) were new to the Norwegian fauna upon sampling (Table A1), 61 Diptera, 24 Hymenoptera, one Thysanoptera and six oribatid mites. Of these, the following 21 species (2.9%) were new to science: 16 phorid flies (13 described in [25]), one sciarid midge [27], one chironomid midge [58], one aphelinid wasp [59] and two oribatid mites awaiting description. Additionally, of the 92 new Norwegian records, 15 were found in the Nordic countries for the first time (Table A1). Diptera was the most species-rich order of invertebrates with 334 species (46.3%), followed by Hymenoptera with 117 (16.2%) and Coleoptera with 84 (11.6%). Additionally, Diptera was represented with the highest number of specimens with 1339 (19.5%), followed by Hemiptera with 1108 (16.1%) and Coleoptera with 821 (12.0%). Collembola and Araneae were not included in the specimen calculations as their abundances were ranked for the common species. These figures correspond well with other inventories from canopies.
Amongst the sites, the six most species-rich orders were represented in stable proportions with respect to the number of species present ( Figure 2), with Diptera being the clearly most species rich at all the sites (29% in Skjaersjø to 47% in Berge). The proportion of specimens for the six most abundant orders, however, showed a varied pattern in that Isopoda constituted 23% of the specimens collected at Djupedal, Hemiptera almost 45% at Mule Varde and Coleoptera 25% at Skjaersjø (Figure 2). Moreover, the number of collected species ranged from 166 in Berge to 370 in Steinknapp, and the number of specimens collected was 4.6 times higher in Steinknapp (2440) than at Berge (536) ( Table 2). Steinknapp contained 1.8 times as many species as the second most species-rich site, Djupedal (just a few kilometers away). Although species new to science were found in all the localities, 14 of the 21 new species were found in Steinknapp (25 specimens) with five species as the second highest number in any of the other localities (Skjaersjø, 37 specimens). In addition, 45 species new to Norway (134 specimens) were found in Steinknapp, followed by 20 species (60 specimens) in Djupedal. localities, 14 of the 21 new species were found in Steinknapp (25 specimens) with five species as the second highest number in any of the other localities (Skjae rsjø, 37 specimens). In addition, 45 species new to Norway (134 specimens) were found in Steinknapp, followed by 20 species (60 specimens) in Djupedal. Even though 50.6% of the species (358 species) were represented by singletons and 56.1% (397 species) were found in only one tree (uniques), the turnover along the entire sampling gradient (all 24 trees) was as low as β = 13.34. Rejecting H1, this means that the species communities along the gradient are comparably similar. Between-site turnover showed the same with β = 18.27. Within the sites, however, turnover was higher (Table 2), ranging from 31.01 (Berge) to 44.30 (Skjae rsjø). Thus, despite the high turnover within each site (Table 2), the shift in species composition throughout the sampling gradient was comparably lower, indicating that a similar set of species appear in low numbers in geographically disjunct locations. This separation of sites is also reflected in the DCA ordination diagram (Figure 3), as the two sites on cultivated land (Berge and Mule Varde) were nicely grouped separately from the other sites indicating similar within-site composition of species but different from each other (except tree 11 from Steinknapp), and thus supporting H2. At the opposite side of the gradient, the Djupedal site also indicates a similar species composition within the site, but different from the other sites. The strong explanatory powers of the DCA axes one and two (Eigenvalues = 0.51 and 0.32, respectively), as well as the long gradient (3.98 SD), corroborate this.  Even though 50.6% of the species (358 species) were represented by singletons and 56.1% (397 species) were found in only one tree (uniques), the turnover along the entire sampling gradient (all 24 trees) was as low as β = 13.34. Rejecting H1, this means that the species communities along the gradient are comparably similar. Between-site turnover showed the same with β = 18.27. Within the sites, however, turnover was higher (Table 2), ranging from 31.01 (Berge) to 44.30 (Skjaersjø). Thus, despite the high turnover within each site (Table 2), the shift in species composition throughout the sampling gradient was comparably lower, indicating that a similar set of species appear in low numbers in geographically disjunct locations.
This separation of sites is also reflected in the DCA ordination diagram (Figure 3), as the two sites on cultivated land (Berge and Mule Varde) were nicely grouped separately from the other sites indicating similar within-site composition of species but different from each other (except tree 11 from Steinknapp), and thus supporting H2. At the opposite side of the gradient, the Djupedal site also indicates a similar species composition within the site, but different from the other sites. The strong explanatory powers of the DCA axes one and two (Eigenvalues = 0.51 and 0.32, respectively), as well as the long gradient (3.98 SD), corroborate this.

Acari
Two oribatid mites new to science were found. Damaeus sp. n. was abundant, with 51 specimens and was present at all the sites except Djupedal, while Phthiracarus sp. n. was found with five geographically disjunct specimens (Table A1). In addition, the following four oribatid species were recorded from the Nordic countries for the first time: Liacarus (Dorycranosus) splendens (Coggi, 1898) with one specimen from Steinknapp, Oribatella (Oribatella) quadricornuta (Michael, 1880) with 14 specimens from Steinknapp, Phauloppia nemoralis (Berlese, 1916) with one specimen from Skeianeset and two from Steinknapp, and Xenillus (Xenillus) discrepans Grandjean, 1936 with 14 specimens from Skeianeset, three from Mule Varde and one from Skjaersjø, respectively.
Among the arboreal species of oribatid mites inhabiting the oak canopies, we can include the following species living in the growths of mosses and lichens therein: Camisia (C. species were also common in the treetops, feeding on the lichen thalluses: Phauloppia lucorum (C. L. Koch, 1841) and Phauloppia nemoralis (Berlese, 1916). The following oribatid species, preferring decaying wood, were also frequent in tree canopies: Caleremaeus monilipes (Michael, 1882), Carabodes (C.) rugosior Berlese, 1916 and Euphthiracarus (E.) cribrarius (Berlese, 1904). Arboreal species are usually bigger (length of body 600-1000 µm), dark brown or black, with a heavily sclerotized cuticle and a thick layer of waxy cerotegument on the body surface, protecting them from desiccation. Forest litter and soil species, on the other hand, are characteristically smaller, lighter in color, with a weaker sclerotized cuticle and a thinner layer of cerotegument (families Tectocepheidae, Oppiidae, Suctobelbidae, Brachychthoniidae, etc.). They were not found in the tree canopies.

Isopoda
Trachelipus ratzeburgii (Brandt, 1833) is categorized as near threatened on the Norwegian red list [61]. It appeared with 16 specimens in Djupedal and three in Skjaersjø (Table 3).

Collembola
Being scavengers for most, springtails are common in trees [62]. All of the 23 species found in the oak canopies can be considered as common species, with Entomobrya nivalis (Linnaeus, 1758) as the most abundant species in this study by far. This species, together with E. albocincta (Templeton, 1835), E. corticalis (Nicolet, 1842), E. marginata (Tullberg, 1871) and Sminthurinus alpinus Gisin, 1953, are known arboreal species being associated with the lichens growing on bark.

Psocodea
Twenty-four species of the order Psocodea were collected from the oak canopies, all belonging to families formerly referred to as the paraphyletic «order Psocoptera» [65,66]. Most Psocodea feed on algae, microfungi and lichens, or decomposing stages of these, as well as pollen. Most of the foliage-living species are associated with either conifers or broadleaved trees, whereas bark-living species (on trunks as well as branches and twigs) are less discriminate. For most Psocodea, the character of the foodstuff itself, which may be dependent on physical factors such as moisture, light and exposure, is probably more important than the tree species. No Psocodea species was found at all the sites, but Reuterella helvimacula (Enderlein, 1901), Valenzuela flavidus (Stephens, 1836) and Mesopsocus unipunctatus (Müller, 1764) were the most common species (see Table A1). Almost all of the collected species are arboreal on a variety of tree species; Lachesilla quercus (Kolbe, 1880) has been believed to be confined to oak [6], but may also be found on other tree species, and outside the distribution of oak. Its apparent association with oak may rather be an expression of its preference [67,68] for dead leaves lingering on the tree, as commonly found on oaks, or on cut-off branches on the ground. Valenzuela flavidus and Graphopsocus cruciatus (Linnaeus, 1768) are associated with foliage of various deciduous trees [6,69].

Thysanoptera
Five specimens of Poecilothrips albopictus Uzel, 1895 were found at the two sites in Drangedal and in Larvik. This species was taken for the first time in Norway and its distribution indicates that it is fairly common. The biology of Thysanoptera is generally Diversity 2021, 13, 332 9 of 31 poorly known and it cannot be claimed that any of the 14 species in this study are associated with oaks-they are more likely to be associated with substrates offered by the tree, such as fungal spores, algae, etc.

Diptera
This was by far the most species rich group, with 334 species collected, 18 species new to science, 7 species new to the Nordic fauna and an additional 52 species caught in Norway for the first time (Table A1). Phoridae was the family with the largest number of specimens collected (212 specimens), followed by Ceratopogonidae (203) and Chironomidae (123). Phoridae was also the most species rich family by far, with 76 species, of which 16 species were new to science (all of them in the genus Megaselia); in addition, four species were new to the Nordic countries and 23 were new to Norway [25]. Borophaga agilis (Meigen, 1830) was reported new to Norway in [25], but was later found to have been reported in [70]. Sciaridae was the second most species-rich group, with 43 species (one species new to science and eleven new records for Norway) [27,71], followed by Chironomidae with 42 species (one species new to science [58], and two new to Norway). In addition, the following families were represented by new records: Limoniidae and Lauxaniidae (one new to the Nordic countries and one new to Norway, respectively), Ceratopogonidae (one new to the Nordic countries) and Fanniidae (one new to Norway).
The ecology of Diptera is mostly poorly known, and the abundant families in this study, e.g., Phoridae, Ceratopogonidae and Chironomidae, are usually neglected in general faunistic surveys. Only adults were identified, while habitat requirements are a characteristic of the larvae of most species in these families. Nonetheless, most of the species in the sciarid genera Bradysia, Corynoptera and Scatopsciara in this study (see Table A1) might have a connection with oak trees beyond accidental visits, as they are mentioned as deciduous forest species in the literature [27,72]. Other species of Sciaridae are also mentioned as deciduous forest associates (see Table A1). Additionally, Phyllodromia melanocephala (Fabricius, 1794) (Empididae) and Systenus bipartitus (Loew, 1850) (Dolichopodidae) are species known to inhabit deciduous forests. The first was one of the most common species, with 77 specimens collected and from all the sites.
Many species of Diptera are known to be trunk dwelling, fungivores or associated with rotting wood, habitats that are present abundantly in old oak trees. A rather high proportion of the collected species, where ecological information is available, can be assigned to either of these categories, most of them with few specimens. One exception was Forcipomyia titillans (Winnertz, 1852), a rotting matter associate [73], which was found with 22 individuals.
Other individual species accounts worth mentioning are those being abundant at all the sites or aggregated at any one site. Culicoides impunctatus Goetghebuer, 1920 (Ceratopogonidae) is a haematophagous parasite on vertebrates and is also known to aggregate close to the breeding sites, which are humid areas, preferably peat bogs [74]. It was abundant in Steinknapp and Skjaersjø in particular, with 36 and 30 specimens collected, respectively. Phora edentata Schmitz, 1920 (Phoridae), a species new to Norway, was fairly abundant at most of the sites, which indicates that it is a rather common species. Two other species, Rhagio lineola Fabricius, 1794 (Rhagionidae) and Lyciella platycephala (Loew, 1847) (Lauxaniidae) were abundant in most sites. Both of these species are common and occupy many habitats. Twelve specimens of Anapausis helvetica Haenni, 1984 (Scatopsidae) were collected from Mule Varde and not from elsewhere. This species is rarely collected, but present knowledge may indicate an association with open areas, farmlands and parks [75]. Platypalpus ecalceatus (Zetterstedt, 1838) (Hybotidae) was collected with 13 individuals and only in Djupedal. This species is most likely a predator, as are nearly all Empidoidea (Terje Jonassen, pers. comm), but we cannot readily explain why it appears aggregated at only one site. We can see a similar pattern for two other Empidoidea, the dolichopodids Chrysotimus flaviventris (von Roser, 1840) and Dolichopus plumipes (Scopoli, 1763), being represented with 21 and 66 specimens in the Drangedal samples, respectively, and almost absent from all the other sites (see Table A1). Ten specimens of Megaselia robertsoni Disney, 2008 (Phoridae), a species new to Norway, were found only at Steinknapp.

Hymenoptera
A total of 117 species of Hymenoptera were collected, with one species new to science, four species new to the Nordic countries and 21 additional species new to Norway (Table A1). Many of the specimens could only be identified to genera or 'near to' designated species. Thus, we cannot rule out that there are additional undescribed species in this material. Of the two suborders, Symphyta and Apocrita were represented only by Apocrita. Of the 118 species, 12 Aculeata, i.e., nine Formicidae and three Crabronidae, were found, with the remaining 106 species all belonging to the 'Parasitica infraorder'. Ceraphronoidea with 22 species (68 specimens); Chalcidoidea, 55 species (160); Cynipoidea, nine species (31); Diaprioidea, 11 species (15); Platygastroidea, 21 species (56). The Ichneumonoidea superfamily was not processed, only one species of Gelis sp. (1) has been added to the list. Ants in the mound building Formica rufa group, namely F. polyctena (Förster, 1850) were, not surprisingly, the most abundant species. They were all collected in Drangedal and from all the treated trees at Djupedal. None of the remaining species were abundant in any of the sites, but 30 specimens of Tamarixia pubescens (Nees, 1834) (Eulophidae), a new species to the Nordic fauna, were collected and taken at all the sites. This is a parasitoid of psyllids known to parasitize Trioza remota Förster, 1848 [76], which, as nymph, is an oak obligate. T. remota was, however, not found in this study. Seladerma tarsale (Walker, 1833) (Pteromalidae) was also rather common with 24 specimens, whereof 14 were collected in Steinknapp. This species is a primary parasitoid of Agromyzidae flies [77]. No Agromyzidae were present in the material, however.
The representation of species shows a well-defined association with oak-galls.
Several of the following species are associated with oak or oak habitats: the curculionid Archarius pyrrhoceras (Marsham, 1802), Coeliodes rana (Fabricius, 1787), Orchestes quercus (Linnaeus, 1758), the already-mentioned cantharid M. seriepunctatus, the ciid Cis vestitus (Mellié, 1848), the melandryid Conopalpus testaceus (Olivier, 1790), the chrysomelid Cryptocephalus labiatus (Linnaeus, 1761) and the cerambycid Leiopus linnei Wallin, Nylander and Kvamme, 2009 [10,36,78]. Furthermore, many species are known to be arboreal (see Table A1) but being rare in this material was common for most of them. A common, arboreal species was Otiorhynchus singularis (Linnaeus, 1767) (Curculionidae), which is a species found almost everywhere. Thirty-one specimens were found at all the sites but Skjaersjø. Another weevil, Strophosoma capitatum (De Geer, 1775), a common herbivore on broadleaf trees, was found with 86 specimens at all but the two sites in Western Norway. The predacious Cantharidae Malthodes guttifer Kiesenwetter, 1852 was collected at all the sites, except for Berge, with a total of 61 specimens. This is a common species associated with shrubs and often found climbing trees [10]. Eleven specimens of Orchesia micans (Panzer, 1793) (Melandryidae) were taken in Skjaersjø, its only appearance in the study. It has a close association with polypore fungi in the genus Inonotus [79]. The throscid Trixagus dermestoides (Linnaeus, 1767) was found with 11 specimens, ten of them from Steinknapp. This species is known as a generalist pollen and mold feeder (e.g., [80]), with habitats plentiful in oaks.

Species Accumulation
The number of invertebrate species collected was 722 and with an overall turnover of 13.34, suggesting a rather homogenous species pool along the sampling gradient, thus rejecting H1. Despite the apparent homogeneity, there is a logarithmic relationship between the number of specimens collected and the number of species found (Figure 4), suggesting that a much more profound sampling effort needs to be performed before the accumulation curve starts to converge. A steep species accumulation curve is to be expected, as the sample size was low and there was a high number of singletons and uniques.

Invertebrate Samples
The number of collected specimens in this study was very low compared w material collected from a comparable study of 24 pine trees over a geographic g from west to east Norway, and where nearly 30,000 specimens were collected us same methodology [38]. One explanation is fairly obvious, as the weather in bo pling periods (June/July 2011/2012) was generally cool and wet. The monthly temp in 2011 was, on average, slightly higher than the normal temperature (ranging fr °C below (Kvam, June) to +1.7 °C above (Kvam, July)), but the precipitation range 104% (Kvam, July) to 270% (Drangedal, July) of the normal [81,82]. For the year 2 monthly temperature was lower than the normal temperature (from −1.8 °C (Dra June) to −0.2 °C (Kvam, July) below), and these months were also generally wett the normal (from 69% (Kvam, June) to 169% (Kvam, July)) [83,84]. Other reasons low catch may be related to the structure and complexity of the oak canopy co with the more open canopy of, for example, pine, in that a larger proportion of th tebrates remain in the tree-either stuck in the dense foliage or on the branches [4

Invertebrate Samples
The number of collected specimens in this study was very low compared with the material collected from a comparable study of 24 pine trees over a geographic gradient from west to east Norway, and where nearly 30,000 specimens were collected using the same methodology [38]. One explanation is fairly obvious, as the weather in both sampling periods (June/July 2011/2012) was generally cool and wet. The monthly temperature in 2011 was, on average, slightly higher than the normal temperature (ranging from −0.1 • C below (Kvam, June) to +1.7 • C above (Kvam, July)), but the precipitation ranged from 104% (Kvam, July) to 270% (Drangedal, July) of the normal [81,82]. For the year 2012, the monthly temperature was lower than the normal temperature (from −1.8 • C (Drangedal, June) to −0.2 • C (Kvam, July) below), and these months were also generally wetter than the normal (from 69% (Kvam, June) to 169% (Kvam, July)) [83,84]. Other reasons for the low catch may be related to the structure and complexity of the oak canopy compared with the more open canopy of, for example, pine, in that a larger proportion of the invertebrates remain in the tree-either stuck in the dense foliage or on the branches [40].

Faunistics
Despite the fact that the ecology is unknown for many species (see Table A1), a large proportion of the species found in this study must be assumed to be occasional visitors (i.e., the oak canopy is not their primary habitat). As oaks offer a wide selection of sites to rest, swarm and feed, an abundance of generalists is to be assumed, as well as opportunists taking advantage of the secondary habitats in the trees, for example, the ant Camponotus ligniperda (Latreille, 1802) living in dead parts of the tree or the numerous species associated with deposited leaf litter or soils. Yet, a few other species are likely to be accidental visitors from the surroundings, e.g., species associated with grasses and Calluna (see Table A1). The presence of the marine chironomid Halocladius variablis (Staeger, 1839) in Steinknapp is surprising, as the distance to the ocean is about 30 km. Its presence in Skeianeset and Mule Varde makes sense, however, as both sites are close to the sea.
Even though neither the psyllid Trioza nor agromyzid flies were found as adults, we must believe them to be present, as parasitoids of both were common-Tamarixia pubescens (Eulophidae) and Seladerma tarsale (Pteromalidae), respectively. Both host groups are known to live on oaks [85,86]. Another fact to note is that no species of the egg parasitoid family Mymaridae (Chalcidoidea) were collected. Mymaridae are among the smallest insects in the world and, regarding the number of species and specimens collected, it is inconceivable that Mymaridae species would not be present in larger numbers as well.
Unfortunately, due to their size and fragility, they are likely to remain in the canopy foliage after fogging.
Correspondence in the presence of species over a broader selection of the literature shows that 80 of the species collected in this study were also present in other European studies on oak canopy or oak tree faunas [2,6,9,10,36,60,63,64,69,78,87,88].

Conservation and Distribution of Invertebrates
Some paradoxes arise when comparing the number of red-listed species with the number of species new to science or new occurrences. Only five red-listed species were found, while the number of new occurrences, including new species, were 92 altogether, most of them with very few specimens. This demonstrates how poorly known the Norwegian arboreal invertebrate fauna still is. One of the criteria for inclusion on the Red List is that a species should be known to reproduce for more than 10 years in the period 1800-2015 [61]. Moreover, rarity is not a criterion for inclusion as such, but reduced population sizes, reduced habitats or reduced distributions are. Thus, the value of the red list category for a species is based on the changes in the intermediate-term development of its population and no new species or species observations will qualify for considerations into the list, but it should incentivize the monitoring of those species. Inasmuch, a new species does not necessarily have to be rare, it may just have been overlooked. Several new species or occurrences were widespread and with intermediate numbers, e.g., Damaeus n. sp. (50 specimens, five localities), Xenillus (Xenillus) discrepans (18 specimens, three localities), Tamarixia pubescens (30 specimens, all localities), Megaselia ignobilis (19 specimens, four localities) and Phora edentata (40 specimens, four localities) (Table A1). Canopy specialists may well have been overlooked, as some are, apparently, rarely collected using conventional techniques and the obvious inaccessibility to the canopy complicates sampling.
Oaks used to be evenly distributed within its distributional range in Norway, and fragmentation was caused by overexploitation and a colder climate in the beginning of the sub-Atlantic era [89]. The rejection of H1 can be a response to a historically continuous distribution of oaks by the remaining relic populations of invertebrates. Additionally, compared with the more diverse forest sites, the poorer community of plants, homogeneous canopy structure [44] and different microclimate [41,90] in the actively managed sites, Mule preparing Figure 1. Finally, we would like to thank the landowners in question for giving permission to use their land.

Conflicts of Interest:
The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results.