Abstract
Aim Microbial communities often vary spatially in how they assemble and knowledge is lacking about which factors determine the biogeography of host‐associated microbiomes. Our aim is to assess the relative importance of spatial, environmental and host‐associated factors on microbial community composition of an important defoliating insect. Location Boreal forests in eastern Canada (Quebec, Newfoundland and Labrador). Taxon The eastern spruce budworm ( Choristoneura fumiferana ) and its associated bacterial communities. Methods We characterized bacterial communities associated with spruce budworm larvae using bacterial 16S rRNA gene amplicon sequencing. We quantified how much of the variation in these bacterial communities could be explained by (1) environmental conditions, (2) the microbiome of foliage the larvae were eating, (3) host tree species and (4) spatial structure as quantified using Moran's Eigenvector Maps (MEMs). Results Budworm larval microbiomes varied significantly among sites and between host tree species. Larvae bacterial community structure was strongly correlated with the structure of bacterial communities taken from paired foliage samples. Spatial structure, foliage bacterial communities and host tree species collectively explained almost one‐sixth of the variation in budworm bacterial communities while environmental conditions did not explain variation on their own. Main conclusions Lepidopteran microbiomes primarily originate from the foliage diet. However, subtle differences in microbial communities between larvae and foliage suggest that some bacteria establish and grow in the budworm microbiome, and that dispersal of bacteria from sources other than foliage as well as differences in environmental filtering between larval bodies and foliage play a role in the assembly of the budworm microbiome. While spatial location and spatial structure were also important drivers of spruce budworm bacterial community composition, none of the environmental variables we measured could explain the variation among sites, and identifying the drivers of this spatial variation remains an open question that will need to be addressed by future studies.