The influence of diet and host specificity on the fecal microbiome of three adult dragonfly species, Pseudothemis zonata, Orthetrum lineostigma, and Orthetrum melania, was investigated. The fecal bacterial communities were analyzed using 16S rRNA gene sequencing, and stable isotope analysis was used to investigate their food sources. The results showed significant differences in the composition of fecal bacterial communities among the three species, with host specificity potentially playing a more important role than diet. The dominant phyla in the fecal bacterial communities of all three species were Firmicutes, Proteobacteria, and Bacteroidetes. The operational taxonomic units (OTUs), Shannon index, and phylogenetic diversity index were not significantly different among the three species, indicating that there were no major differences in the diversity of the fecal bacterial communities. The stable isotope analysis showed that the food sources were similar among the three species, being primarily small insects found near the aquatic habitats. However, the fecal bacterial communities of two closely related species, O. lineostigma and O. melania, were different despite their similar food sources. In contrast, the fecal bacterial communities of O. lineostigma and P. zonata were similar, despite the different food sources of these two species. Our findings suggest that host specificity and diet can influence the composition of the intestinal microbiome in these insects, but the degree of influence may depend on the specific host and environmental conditions.

We investigated the floral resources of the honeybee (Apis mellifera) by pollen DNA metabarcoding as the first step in honeybee conservation. We attempted to identify as many floral resources and as wide resource availability as possible in areas and compared the richness of floral resources between forests and national parks. We found that both forest and national parks can sufficiently support honeybee floral resources; however, national parks were more stable habitats with more complex vegetation structure than those of forests. Also, honeybee's preferences for floral resources were determined using niche breadth, niche overlap, and network analysis among the floral resources in forest areas. Quercus and Actinidia were the most frequently detected taxa of pollen resources in both forests and national parks by pollen metabarcoding. Even though we anticipated that more various pollen resource taxa would be found in national parks than forests, no differences in the number pollen resource taxa were detected between them. According to the niche overlap analysis result, the honeybee preferred the pollen resources in both forests and national parks, and the co-occurrence pattern analysis of pollen resource taxa in both forests and national parks showed a simple network. Altogether, our findings suggested that DNA metabarcoding of pollen collected by the honeybee can provide useful information on the flowering plants in a given area. As the use of floral pollen resources varies spatially and temporally, periodic monitoring of pollen resources can identify which pollen is required to maintain the health of honeybee colonies.