Invertebrate Assemblages on Biscogniauxia Sporocarps on Oak Dead Wood: An Observation Aided by Squirrels

: Dead wood is an important habitat for both fungi and insects, two enormously diverse groups that contribute to forest biodiversity. Unlike the myriad of studies on fungus–insect relationships, insect communities on ascomycete sporocarps are less explored, particularly for those in hidden habitats such as underneath bark. Here, I present my observations of insect community dynamics on Biscogniauxia spp. on oak dead wood from the early anamorphic stage to matured teleomorph stage, aided by the debarking behaviour of squirrels probably targeting on these fungi. In total, 38 insect taxa were observed on Biscogniauxia spp. from March to November. The community composition was signiﬁcantly correlated with the presence/absence of Biscogniauxia spp. Additionally, Librodor ( Glischrochilus ) ipsoides , Laemophloeus submonilis , and Neuroctenus castaneus were frequently recorded and closely associated with Biscogniauxia spp. along its change from anamorph to teleomorph. L. submonilis was positively associated with both the anamorph and teleomorph stages. L. ipsoides and N. castaneus were positively associated with only the teleomorph but not with the anamorph stage. N. castaneus reproduced and was found on Biscogniauxia spp. from June to November. These results suggest that sporocarps of Biscogniauxia spp. are important to these insect taxa, depending on their developmental stage.


Introduction
Dead wood is an essential component of biodiversity in forest ecosystems [1,2]. Fungi, in particular, is a major group of saproxylic communities; they have a large impact on saproxylic communities due to their unique wood decay abilities [3], and their fruit bodies and spores are important to the diet of a variety of organisms, including protozoa [4], invertebrates [5], and vertebrates [6,7]. A better understanding of the relationships between fungi and saproxylic communities is critical to clarifying the mechanisms that maintain biodiversity in forest ecosystems.
In terms of their diversity and function, insects are another major group present within saproxylic communities [1]. Insects have intimate relationships with fungi as fungivores, vectors of fungal propagules, and foragers of wood decomposed by fungi [8]. Numerous studies have investigated the insect communities present on fungal fruit bodies, the majority of which are basidiomycetes [5,9,10]. Host specificity [11][12][13][14], the evolution of host use [15], and spore dispersal [16][17][18][19][20] have been intensively studied for a variety of fungus-insect relationships. Furthermore, topics in general ecology (e.g., coexisting patterns on patchy resources [21,22]) and applied ecology (e.g., effects of forest management on ecological communities [23,24]) have also been investigated using the insect-basidiocarp system. However, studies on the relationships between insect communities on ascomycete fruit bodies, a sister taxon of basidiomycetes which also produces macroscopic fruit bodies, are quite limited, with examples of symbiotic associations in ambrosia beetles [25], woodwasps [26], and fungus-growing termites [27] and insect pathogens, such as in the genera Beauveria [28], Metarhizium [29], and Ophiocordyceps [30].
Quercus serrata is a deciduous oak that dominates low-elevation rural forests throughout Japan. In recent decades, a serious dieback of Q. serrata (oak wilt disease) resulted in huge amount of dead wood [31,32]. Therefore, evaluations of the biotic communities associated with Q. serrata dead wood is necessary to understand and predict biodiversity in areas affected by oak wilt disease. The symbiotic association between a pathogenic fungus, Raffaerea quercivora, and ambrosia beetles, Platypus querucivorus, which carry the propagules of the fungus in their mycangia, is well studied [33,34]. In addition, since Q. serrata logs are traditionally used for cultivating shiitake mushrooms (Lentinula edodes), several ascomycete species that negatively affect the yield of shiitake mushrooms are known to occur on Q. serrata bed-logs [35]. The most famous ascomycete species is in the genus Trichoderma (including species formerly denoted as Hypocrea), which is an antagonistic and/or mycoparasitic taxon and causes serious damage to shiitake cultivation [36]. Trichoderma spp. has an intimate relationship with gall midge belonging to the genus Camptomyia on shiitake bed-logs [37]. Despite recent advances in the studies of microscopic and pathogenic ascomycetes associated with insects, relationships between insects and other taxa in macroscopic ascomycetes in genera such as Biscogniauxia, Daldinia, Diatrype, Graphostroma, and Hypoxylon, which fruit on Q. serrata dead wood, are largely unknown.
To investigate successive changes of fungal and insect communities and their interactions during decomposition of Q. serrata dead wood, I started a multi-year survey of 32 experimentally cut logs of Q. serrata in August 2015. In March 2016, I observed that a squirrel or squirrels frequently visited and intensively debarked the logs, where colonies of anamorphic ascomycetes were found to appear on the sapwood. Subsequently, a variety of insects were found on the ascomycetes throughout their changes from the anamorph to the teleomorph stage. In the present study, I describe the insect assemblages that were observed on the Q. serrata logs and their relationships with the different sexual stages of ascomycetes during a growing season.

Experimental Setup
The present study was conducted in a secondary forest dominated by Q. serrata and Pinus densiflora in Kami town, Miyagi, Japan (38 • 37.2 N, 140 • 48.6 E). In 2016, the mean annual temperature at the nearest meteorological station at Kawatabi (38 • 44.6 N, 140 • 45.6 E) was 11.6 • C (3.2 • C in February to 28.7 • C in August), and the annual precipitation was 1537.5 mm. Snow covers the ground from November to March (Japan Meteorological Agency, available online: https://www.jma.go.jp/jma/indexe.html (accessed on 21 August 2021)). In August 2015, two Q. serrata trees were felled and cut into 32 logs with a length of 1 m each (diameter 3.4-24.5 cm). The logs were laid on the ground approximately 1 m from each other. I began observations of the logs' surfaces (top and bottom) in March 2016, when the ground was still covered with snow, but the tops of the logs were visible.
On 10 March 2016, at 6:30 a.m., I observed that a squirrel (Sciurus lis) approached the logs and tore off the bark using its teeth. I found Geniculosporium type anamorph of ascomycete on the sapwood surface where the bark had been removed by the squirrel (Figure 1a). The surface of the anamorph was scratched overall (Figure 1a

Data Collection
Debarked, anamorph, and teleomorph areas were measured every month (except for April and May) by placing 1 cm grid squares on the log surface and counting the number of grid squares that had debarkation, anamorphs, and teleomorphs.
Insects that were observed on the log surface were recorded every week (except for April and May) as binary data (presence/absence on each log). Insects were identified with reference to the keys and nomenclature of Kurosawa et al. [38], Ueno et al. [39], and Hayashi et al. [40] for Coleoptera; Ishikawa et al. [41] for Hemiptera; and Terayama et al. [42] for Hymenoptera (ants). Identification to the species level was difficult, so several species were identified at genus, family, and order level. Taxa that occurred on ≥20 logs were recorded as dominant taxa.

Data Analysis
All statistical analyses were conducted using R ver. 4.0.5 [43]. A generalised linear model (GLM) was applied to explain the species richness of invertebrates. The diameter of the logs, position (top/bottom) on the logs, and anamorph and teleomorph areas on the log surface were set as explanatory variables. A binomial distribution error was assumed, and a logit link function was used. The log-transformed surface area of the logs was set as an offset term.
The relationship between invertebrate community composition and environmental factors was visualised using non-metric multi-dimensional scaling (NMDS) with the vegan package [44]. Similarities of the invertebrate communities across the logs were calculated using the Raup-Crick similarity index (vegdist command), and this matrix was used to develop the NMDS ordination plot (metaMDS command). The significance of the difference in community composition between the top and bottom positions of the logs was determined using permutation multivariate analysis of variance [45] with 10,000 permutations (adonis command). In addition, community variance between samples (calculated

Data Collection
Debarked, anamorph, and teleomorph areas were measured every month (except for April and May) by placing 1 cm grid squares on the log surface and counting the number of grid squares that had debarkation, anamorphs, and teleomorphs.
Insects that were observed on the log surface were recorded every week (except for April and May) as binary data (presence/absence on each log). Insects were identified with reference to the keys and nomenclature of Kurosawa et al. [38], Ueno et al. [39], and Hayashi et al. [40] for Coleoptera; Ishikawa et al. [41] for Hemiptera; and Terayama et al. [42] for Hymenoptera (ants). Identification to the species level was difficult, so several species were identified at genus, family, and order level. Taxa that occurred on ≥20 logs were recorded as dominant taxa.

Data Analysis
All statistical analyses were conducted using R ver. 4.0.5 [43]. A generalised linear model (GLM) was applied to explain the species richness of invertebrates. The diameter of the logs, position (top/bottom) on the logs, and anamorph and teleomorph areas on the log surface were set as explanatory variables. A binomial distribution error was assumed, and a logit link function was used. The log-transformed surface area of the logs was set as an offset term.
The relationship between invertebrate community composition and environmental factors was visualised using non-metric multi-dimensional scaling (NMDS) with the vegan package [44]. Similarities of the invertebrate communities across the logs were calculated using the Raup-Crick similarity index (vegdist command), and this matrix was used to develop the NMDS ordination plot (metaMDS command). The significance of the difference in community composition between the top and bottom positions of the logs was determined using permutation multivariate analysis of variance [45] with 10,000 permutations (adonis command). In addition, community variance between samples (calculated using the betadisper command) was compared between the top and bottom positions using an analysis of variance (anova command). Finally, the significance of the effects of the environmental variables on the invertebrate communities was determined using the envfit command; sampling month, diameter, anamorph percentage, and teleomorph percentage were set as environmental variables.
A set of GLMs were applied to explain the occurrence of the three dominant invertebrate species. The environmental variables that were detected to be significant using envfit (diameter, anamorph percentage, and teleomorph percentage) were set as explanatory variables. Binomial distribution errors were assumed. Logit link functions were used.

Results
The mean percentage of debarked area per log was 10% in March, increased slightly from June to July, and remained constant thereafter (Figure 2a). The tops of the logs were debarked more than bottoms of the logs. The anamorph area occupied almost all parts of the debarked area in March but decreased to zero by the end of July (Figure 2b). In contrast, the teleomorph area was seldom observed until June but greatly increased in July and occupied almost all parts of the debarked area until November (Figure 2c).
were set as environmental variables.
A set of GLMs were applied to explain the occurrence o brate species. The environmental variables that were detect vfit (diameter, anamorph percentage, and teleomorph percen variables. Binomial distribution errors were assumed. Logit

Results
The mean percentage of debarked area per log was 10% from June to July, and remained constant thereafter (Figure debarked more than bottoms of the logs. The anamorph are the debarked area in March but decreased to zero by the en trast, the teleomorph area was seldom observed until June and occupied almost all parts of the debarked area until Nov In total, 38 taxa of insects were recorded (Table S1). Cole and consisted of 28 taxa, followed by 4 taxa of Hemiptera a chaeognatha, Hymenoptera, and Psocodea included one tax In total, 38 taxa of insects were recorded (Table S1). Coleoptera was the largest group and consisted of 28 taxa, followed by 4 taxa of Hemiptera and 3 taxa of Lepidoptera. Archaeognatha, Hymenoptera, and Psocodea included one taxon each. The GLM indicated that insect diversity was negatively correlated with log diameter and positively correlated with teleomorph area ( Table 1). The bottom sides of the logs had lower insect richness than the top sides.
The observed insect communities were significantly correlated with sampling month, log diameter, and anamorph area ( Figure 3); however, the position on the logs (top/bottom) was not correlated with the insect communities. Among the recorded insect taxa, Librodor (Glischrochilus) ipsoides, Laemophloeus submonilis, and Neuroctenus castaneus were observed on ≥20 logs and were thus assigned as dominant taxa; these species were not recorded in March but were frequently recorded in June and July, particularly on the top sides of the logs (Figure 4). Although L. submonilis and L. ipsoides disappeared by August and September, respectively, N. castaneus remained until November. Additionally, N. castaneus reproduced on the surface of Biscogniauxia teleomorphs (Figure 4c). The GLM based on the data from July suggested that the occurrence of L. submonilis was positively associated with the anamorph area and teleomorph areas ( Table 2), whereas the occurrences of L. ipsoides and N. castaneus were positively associated with log diameter and teleomorph area but not with the anamorph area. The observed insect communities were significantly correlated with sampling month, log diameter, and anamorph area ( Figure 3); however, the position on the logs (top/bottom) was not correlated with the insect communities. Among the recorded insect taxa, Librodor (Glischrochilus) ipsoides, Laemophloeus submonilis, and Neuroctenus castaneus were observed on ≥20 logs and were thus assigned as dominant taxa; these species were not recorded in March but were frequently recorded in June and July, particularly on the top sides of the logs (Figure 4). Although L. submonilis and L. ipsoides disappeared by August and September, respectively, N. castaneus remained until November. Additionally, N. castaneus reproduced on the surface of Biscogniauxia teleomorphs (Figure 4c). The GLM based on the data from July suggested that the occurrence of L. submonilis was positively associated with the anamorph area and teleomorph areas (Table 2), whereas the occurrences of L. ipsoides and N. castaneus were positively associated with log diameter and teleomorph area but not with the anamorph area.

Discussion
In the present study, the debarking behaviour of squirrels improved the visibility of insect communities on Biscogniauxia sporocarps, as their initial anamorphic stages are usually hidden under the bark. Fungivorous saproxylic insects with flat morphology often inhabit invisible microhabitats, such as the space between bark and wood, which reduces their visibility considerably [1,46]. Debarking by squirrels targeting Biscogniauxia sporocarps enabled the author to observe the sporocarps starting from the very early stages of their development. Squirrels very accurately found Biscogniauxia sporocarps hidden under the bark. It is well-known that squirrels can detect hypogeous sporocarps by smell [47,48]. Xylariaceous ascomycetes effuse a variety of volatiles, depending on the developmental stage of their sporocarps [49], which may be detected by squirrels. McKeever [47] also reported that fungi constitute the largest portion of squirrel diets, particularly in the

Discussion
In the present study, the debarking behaviour of squirrels improved the visibility of insect communities on Biscogniauxia sporocarps, as their initial anamorphic stages are usually hidden under the bark. Fungivorous saproxylic insects with flat morphology often inhabit invisible microhabitats, such as the space between bark and wood, which reduces their visibility considerably [1,46]. Debarking by squirrels targeting Biscogniauxia sporocarps enabled the author to observe the sporocarps starting from the very early stages of their development. Squirrels very accurately found Biscogniauxia sporocarps hidden under the bark. It is well-known that squirrels can detect hypogeous sporocarps by smell [47,48]. Xylariaceous ascomycetes effuse a variety of volatiles, depending on the developmental stage of their sporocarps [49], which may be detected by squirrels. McKeever [47] also reported that fungi constitute the largest portion of squirrel diets, particularly in the summer and early autumn, and represent the second largest portion of their diets in the early spring; however, the majority of the fungi eaten by squirrels may be basidiomycetes [48]. I also observed a squirrel eating old (fruiting from the previous year) Sarcomyxa serotina (basid- Although it is unclear whether squirrel fungivore of xylariaceous ascomycetes is common, I found the same type of debarking on freshly felled oak logs approximately 8 km away from the study site in 2018 and again at the study site in February 2021 ( Figure S1, the personal observation of Yu Fukasawa). Currah et al. [6] investigated the stomach contents of squirrels and flying squirrels in North America and found fragments of Xylariaceae and Diatripaceae ascocarps. However, again, basidiomycetes sporcarps constituted the majority of their gut contents. Xylariaceous ascomycetes may constitute a supplemental diet during the early spring when other fresh food sources are unavailable. Fungal spores contain a high percentage of nitrogen, but their digestibility is very low in squirrels [50].
Among the recorded insect taxa, almost all taxa except for Niponius osorioceps and Pheidole fervida belonged to families that include species reported to be fungivores [1,51,52]. More specifically, species belonging to Aradidae, Cucujidae, Nitidulidae, Anthribidae, Monotomidae, Silvanidae, Corylophidae, and Biphyllidae are known as ascomycete eaters [51,53]. I even observed Lepidopteran larvae of unknown identity grazing the surface of Biscogniauxia stromata. Tineoidea moths are known to have fungivorous habits, although they feed on basidiomycotan wood decay fungi [22,54]. Powell [55] reported that Lepidopteran species belonging to Pyralidae feed on the stromata of Hypoxylon occidentale (Xylariaceae, Ascomycota). As shown through NMDS (Figure 3), the presence of Biscogniauxia spp. may strongly attract fungivorous insects and affect their community structure. Lee et al. [56] reported that stem canker caused by Annulohypoxylon truncatum (Xylariaceae) on oak stem significantly increases invertebrate diversity in Korea, irrespective of the presence or absence of sap flow.
In the present study, I identified three dominant insect species: Librodor ipsoides (Nitidulidae, Coleoptera); L. submonilis (Cucujidae, Coleoptera); and N. castaneus (Aradidae, Hemiptera)-they each had significant relationships with the occurrence of Bisocogniauxia spp. (Table 2). Although the gut contents of these species were not surveyed in the present study, it is highly likely that these three species were all fungivores and had intimate associations with Biscogniauxia spp. during their life cycle. Nitidulidae is a wellknown fungivorous Coleoptera occurring in not only fungal fruit bodies but also in the sap flows of damaged trees, fermenting fruits, and pathological plant tissues--such species probably have yeasts or pathogenic fungi as a normal and essential part of their diet [51]. Specifically, species in the genus Librodor are known as 'sap beetles' that forage and breed in fermented sap flow [57](referred as a genus Glischrochilus). L. ipsoides has also been found in the sap on Quercus acutissima trees in Japan, although its abundance is quite low compared with that of other sap beetles [58]. On the other hand, L. ipsoides has been found in the fruit bodies of a basidiomycete Cryptoporus volvatus on Pinus densiflora in Korea [59], but it was not found in C. volvatus in Japan [22]. In the present study, I newly found L. ipsoides on Biscogniauxia spp. fruit bodies at a relatively high frequency (>30%; Figure 4), indicating that the stromata and/or conidia of Biscogniauxia spp. are an important habitat of this species. Lawrence [60] reported observations of Prometopia sexmaculata (Nitidulidae) breeding in the stromata of Hypoxylon (Xylariaceae) on oak. Cucujidae are also known as Coleopteran ascomycete eaters. Species in the genus Laemophloeus have been frequently reported in association with ascomata, such as Daldinia, Tubercularia, Hypoxylon, and Biscogniauxia (=Nummularia) [51]. Therefore, it is not surprising that I found L. submonilis on Biscogniauxia spp. (B. maritima and B. plana) in the present study, even though it is a newly found association in Japanese species.
In Aradidae (Hemiptera), most species are fungivores [53,61] that feed on fungal hyphae, using their piercing-sucking mouthparts to suck the cell contents, and have adapted gut systems [62], but little is known about the fungal host association of most species, particularly for ascomycetes [53]. Most of the Aradidae species occur preferentially on dead wood during the early stages of decay, probably due to the presence of their dietary fungal species [53]. As such, outbreaks of some Aradidae species are closely associated with large dieback events, such as forest fires and pests. Aradus lugubris can appear immediately after forest fires, when they feed on Daldinia loculata (Lév.) Sacc. (Xylariaceae), a known fire-related ascomycete in boreal forests [63]. Similarly, large-scale wind throws and bark beetle outbreaks in a Norway spruce forest in the Bavarian Forest in Germany, where dead spruce snags were intensively colonised by a basidiomycete Fomitopsis pinicola [64], provided a suitable habitat for Aradus obtectus [53]. In Japan, Q. serrata trees have recently suffered from oak wilt disease [65]. However, relationships between oak wilt disease and the occurrence of Biscogniauxia spp. are unclear. Fukasawa et al. [66] compared latent fungal communities within Q. serrata trunks in stands with or without the prevalence of oak wilt disease and did not detect Biscogniauxia spp.
Parental care has been observed in Aradidae, during which the male safeguards the egg mass for several weeks, and parental care may be extended to the nymphal stage [53]. In the present study, I also observed that adults appeared to guard the mass of nymphs ( Figure 4). In addition, I observed many nymphs spraying (probably some chemicals) simultaneously from their tail when they were shaded by my hands.
To summarise, I observed the insect communities on saproxylic Biscogniauxia spp. on oak dead wood during the transition from the anamorphic to teleomorphic stages; this was aided by debarking by squirrels. The presence of Biscogniauxia spp. significantly affected the insect communities. More specifically, the population dynamics of two Coleoptera and one Hemiptera dominant species revealed that Biscogniauxia spp. represents an important habitat in their life cycles.
Supplementary Materials: The following are available online at https://www.mdpi.com/article/ 10.3390/f12081124/s1. Figure S1: Another evidence of debarking of Quercus serrata log by squirrel targeting on ascomycete anamorph, Table S1: Insect taxa recorded in the present study.
Funding: This research received no external funding.