Comparative data on termite diet humification

This is a literature-based comparative dataset on the diet humification of termites (Isoptera) and associated traits. Termites are major decomposers in the tropics and feed along a humification gradient, i.e. degree of mineralization of the food substrate, ranging from fresh plant matter to highly mineralized carbon dispersed in the soil mineral matrix. The data were gathered in order to investigate evolutionary relationships underlying the diversification of trophic niche in these animals.<br>We compiled nitrogen isotopic ratios from 195 species, measured by 19 studies on 17 sites located between 40° N and 19° S, spanning five biogeographic regions. Complementary data on food substrate types, mandible type, body mass and phylogenetic relations are provided for 65 species. We searched for data on termite mean δ15N (‰) using “termite” and “stable isotope” as key words in Google Scholar, along with reference lists of identified studies and citations of such studies. Our search included studies indexed until September 30, 2019. We did not consider carbon stable isotope ratios, as these are uninformative regarding diet humification (although they can inform about other aspects of trophic niche, e.g. preference for C3 or C4 plants). We used measurements of pseudergates in species with linear development or true workers (called altogether workers hereafter, which comprise the bulk of termite colonies. When isotope data were presented in figures, we extracted them using WebPlotDigitizer 4.1 (Rohatgi 2018). In all compiled studies, the ratio of nitrogen stable isotopes in termite tissue (15N/14N) was measured by mass spectrometry and expressed as per mil (δ15N). In a few cases, species with dimorphic workers had each form measured separately, so we used data on the most abundant form. In parallel, we gathered a subset of the compiled studies which provided δ15N of termite colonies paired with that of their food substrate (wood, soil or food stored in the nest) (n = 24 colonies distributed among 13 species) to determine the degree to which termite measurements actually reflect that of their food.<br>We also gathered data on food substrate types according to seven commonly recognized substrates: lichens, litter (including grass and leaf litter), wood (sound and poorly decayed woody substrates), rotten wood (highly decayed, friable wood), fungus comb, soil (humus and mineral soil matrix) and nest (nest substrate produced by termites). For each species, substrates were scored as predominant in the diet (1) or not (0), based on the compiled studies and the authors’ field experience.<br>Worker mandible type was inferred from the feeding classification reported by stable isotope studies, whereas mean worker (or pseudergate) wet body mass (mg) was obtained with the aid of a recent data compilation (Pequeno et al. 2017), coupled to some complementary studies and novel measurements.<br>A phylogenetic tree including 106 of the 195 species was prepared from well-supported chronograms based on mitochondrial genomes (Bourguignon et al. 2015, 2017) and studies on specific clades (Bourguignon et al. 2016; Liao et al. 2018). Minor clades that could not be resolved using published phylogenies were treated as polytomies. Node ages were obtained for most nodes from figures using WebPlotDigitizer 4.1, and undated nodes were assigned evenly distributed ages between dated nodes (or between dated nodes and terminals) using Phylocom 4.2 (Webb, Ackerly &amp; Kembel 2008).<br>"Termite trophic niche (Table 1)" is a .csv file containing the raw comparative data on termite species compiled from the literature for the trophic scaling analysis. The table is organized so that each line represents one species, and each column represents one variable. There are 196 rows (the first row contains column names, followed by 195 rows of data by species) and 29 columns or variables. Columns are ordered from left to right as follows:species: valid species namename_in_reference: name of the species as reported in the original data source (if that name was outdated by the time of this tabulation)code: species name abbreviation, for conveniencesubfamily: taxonomic subfamilyfamily: taxonomic familybody_mass_mg: mean wet body mass of worker (mg)body_mass_type: method used to estimate body mass (direct measurement of worker wet mass, prediction of wet mass from dry worker mass, or ratio between biomass and density of species in a given area)mandible_type: worker mandible type (grinding or pounding) (Eggleton 2011)feeding_group: feeding group (I, II, III or IV) (Donovan et al. 2001)n_delta15N: sample size (number of colonies) of nitrogen isotope measurements, pooled over all studies on the speciessite_1: site where first measurement on nitrogen isotopic ratio of the termite was obtaineddelta15N_1: nitrogen isotopic ratio in site 1 (δ15N, per mil)baseline_1: basal resource used for calibration of termite δ15N in site 1baseline_delta15N_1: termite δ15N measured in site 1site_2, site_3, site_4 and associated columns: same as before for any other sites where the species was assessed for δ15Nreference_delta15N: data source for δ15N measurementsreference_mass: data source for body mass measurements<br>"Termite trophic niche (Table 2)" is a .csv file containing colony-level data used to test for termite tissue δ15N as a proxy for food substrate δ15N, while controlling for body mass. The table is organized so that each line represents one colony, and each column represents one variable. There are 34 rows (the first row contains column names, followed by 33 rows of data by colony) and 13 columns or variables. Columns are ordered from left to right as follows:id: colony identityspecies: valid species namegenus: valid genus namefamily: valid family namesubfamily: valid subfamily namecolony: number of colony within speciesmandible.type: worker mandible type (grinding or pounding) (Eggleton 2011)body.mass: mean wet body mass of worker (mg)delta15N.termite: termite tissue δ15Nresource: resource substrate typedelta15N.resource: δ15N of resource substrate were the termite colony was sampleddelta15N.nest: δ15N of termite nest materialreference: data source for δ15N measurements (sources for body mass as in Table S1)<br>"Termite phylogenetic tree" is a .txt file containing the dated termite phylogenetic tree with 106 species, in Newick format.