Raw data associated with: Gibert & Wieczynski. 2021. Constraints and variation in food web link-species space, Biology Letters The repository contains: 1.Raw data 2.Food web metadata 3.Food web structural data 1)The raw data (all .txt files) Text files containing the binary adjacency matrix for the food webs analyzed in the paper. Each adjacency matrix contains a 1 whenever a species in a column eats a species in a row, and a zero otherwise. 2) Food web metadata (Food_web_metadata.csv) Comma-separated spreadsheet containing all the metadata for each food web, including original citation, lat/long, ecosystem type, country or origin and other variables of interest. 3) Food web structural data (struct_data_NEW.csv) Comma-separated spreadsheet that contains some of the most common structural descriptors of food webs, including: TL (max trophic level) mean TL (average trophic level) Omnv (average degree of omnivory as calculated by R package cheddar) Basal (the proportion of basal species, i.e., species that do not consumer other species) Int (the proportion of intermediate species, i.e., species that are neither basal or top) Top (the proportion of top species, i,e, the proportion of species that are consumed by no other species) Links (the total number of feeding interactions) Gen (the average "generality" in the food web, a measure of variation in the number of feeding interactions of predators) Vul (the average "vulnerability" in the food web, a measure of variation in how many species a species is consumed by in the food web) Taxa (the number of species)