Asymmetric control mechanism yields food web complexity

We compiled a large dataset of empirical food webs mainly from several sources listed below:(i) [https://www.web-of-life.es/map.php?type=7](https://www.web-of-life.es/map.php?type=7)(ii) Brose, U., Archambault, P., Barnes, A.D. et al. 2019. Predator traits determine food-web architecture across ecosystems. Nat Ecol Evol 3, 919–927. (URL: [https://doi.org/10.1038/s41559-019-0899-x](https://doi.org/10.1038/s41559-019-0899-x))(iii) Cohen, J. E., Briand, F., & Newman, C. M. (2012). Community food webs: data and theory (Vol. 20). Springer Science & Business Media. (URL: [https://digitalcommons.rockefeller.edu/cohen_joel_laboratory/1/](https://digitalcommons.rockefeller.edu/cohen_joel_laboratory/1/))

This dataset includes 277 empirical food webs across the globe, with each trophic linking matrix listed in the file "Linking matrices" (.txt). We also specified the name of every species ID represented in each matrix (see the file "SpeciesID&SpeciesType" with .txt). In "Details for real food webs.xlsx", we described the geographic location, latitude, longitude and the corresponding reference for each empirical food web.

Using these empirical food webs, we conducted a Principal Component Analysis (PCA) for their multiple structural properties, including species richness (N), connectance (C), omnivory (O) and the maximal food chain length (FCLmax).