Fig. 4: Herbarium records exhibit less taxonomic, phylogenetic, and functional bias than iNaturalist observations.

Taxonomic bias (a) is represented as the ratio of herbarium to iNaturalist records for each species of plant, accompanied by bar plots above and below which correspond to the number of herbarium and iNaturalist records for each species respectively. Phylogenetic (b) and functional (c) bias is represented as the number of herbarium and iNaturalist records per plant arranged around the phylogenetic tree or functional dendrogram. To enhance visualization, we took the square root of the number of records. Finally, we tested for bias in the distribution of the number of records per species at the tips of both the phylogenetic tree and functional dendrogram using Pagel’s λ, which varies between 0 and 1 with 0 indicating no bias (number of records per sample randomly distributed across the tree) signal and 1 indicating high bias (number of records per sample are highly correlated with the phylogenetic or functional structures). We assessed significance using likelihood ratio tests and reported p-values.