Summary: The name Arthromitus has been applied collectively to conspicuous filamentous bacteria found in the hindguts of termites and other arthropods. First observed by Joseph Leidy in 1849, the identity of these filaments has remained contentious. While Margulis and colleagues declared them to be a life stage of Bacillus cereus , others have assumed them to belong to the same lineage as the segmented filamentous bacteria (SFB) from vertebrate guts, a group that has garnered much attention due to their unique ability to specifically modulate their host's immune response. Both SFB and Arthromitus filaments from arthropod guts were grouped under provisional name ‘ Candidatus Arthromitus’ by Snel and colleagues as they share a striking similarity in terms of their morphology and close contact to the host gut wall. While SFB form a distinct lineage within the family Clostridiaceae , the identity of the filaments from arthropod guts remains elusive. Using whole‐genome amplification of single filaments capillary picked from termite guts and fluorescence in situ hybridization of 16S rRNA with group‐specific oligonucleotide probes, we show that they represent a monophyletic lineage within the family Lachnospiraceae distinct from that of SFB. Therefore, ‘ Candidatus Arthromitus’ can no longer be used for both groups. Given the historic precedence, we propose to reserve this name for the filaments that were originally described by Leidy. For the SFB from vertebrate guts, we propose the provisional name ‘ Candidatus Savagella’ in honour of the American gut microbiologist Dwayne C. Savage, who was the first to describe that important bacterial group.

Termites are important contributors to carbon and nitrogen cycling in tropical ecosystems. Higher termites digest lignocellulose in various stages of humification with the help of an entirely prokaryotic microbiota housed in their compartmented intestinal tract. Previous studies revealed fundamental differences in community structure between compartments, but the functional roles of individual lineages in symbiotic digestion are mostly unknown. Here, we conducted a highly resolved analysis of the gut microbiota in six species of higher termites that feed on plant material at different levels of humification. Combining amplicon sequencing and metagenomics, we assessed similarities in community structure and functional potential between the major hindgut compartments (P1, P3, and P4). Cluster analysis of the relative abundances of orthologous gene clusters (COGs) revealed high similarities among wood- and litter-feeding termites and strong differences to humivorous species. However, abundance estimates of bacterial phyla based on 16S rRNA genes greatly differed from those based on protein-coding genes. Community structure and functional potential of the microbiota in individual gut compartments are clearly driven by the digestive strategy of the host. The metagenomics libraries obtained in this study provide the basis for future studies that elucidate the fundamental differences in the symbiont-mediated breakdown of lignocellulose and humus by termites of different feeding groups. The high proportion of uncultured bacterial lineages in all samples calls for a reference-independent approach for the correct taxonomic assignment of protein-coding genes.