Trends, methods, and biases in bee functional trait research
To survey the literature on bee functional traits, we searched for publications on the ISI Web of Science, the SciELO database, and Google Scholar using combinations of the terms “bees” and “traits” or “functional traits,” as well as in the cited literature of relevant papers. We included studies that used methods that conformed to our definition of a functional trait approach to an ecological question, i.e., which analyzed multiple functional traits comparatively across a sample of multiple bee species in a given environmental context. Correspondingly, we excluded taxonomic revisions and the large body of studies that report or describe bee traits for single species, which provide useful trait data but which do not use functional traits as a comparative framework. We extracted metadata from each of these publications, including information on the focal traits measured, the authors’ definitions of these traits, and the sources of trait data (Supplementary Table 2). All reviewed publications are listed under the Data Sources section.
We found and analyzed 97 papers assessing bee functional traits comparatively across species, published between 2006 and February 2023 (Figure 2a; Supplementary Table 2). Half of the studies examined bees in agroecosystems (48 studies; 49.5%); the remainder were divided between natural landscapes (18 studies; 18.6%), urban landscapes (14 studies; 14.4%), and comparisons across multiple landscape types (17 studies; 17.5%); (Figure 2b). The focal topics of these studies were highly variable, but with a particular emphasis on landscape change (e.g., urbanization, habitat fragmentation, land management); (Figure 2c). These studies sampled bees from 32 different countries, but were overwhelmingly conducted in North America (36 studies; 37.1%) and Europe (39 studies; 40.2%), with only nine studies (9.2%) conducted in Africa and Asia combined, and only 13 (13.4%) in Central and South America (Figure 3).
On average, each study quantified 4.37 functional traits (range = 2 - 10; std. error = 0.18) across a sample of 118.9 bee species (range = 5 - 638; std. error = 11.09). The most commonly studied functional traits were body size (in 85 or 87.6% of studies), nesting biology (either nesting location, nesting ability, or both; in 72 or 74.2% of studies), diet breadth (66 studies; 68.0%), and sociality (64 studies; 66.0%). Also common were measurements of tongue length (28 studies; 28.9%), seasonality (29 studies; 29.9%), and voltinism (12 studies; 12.4%). Nearly a quarter of the studies (22 studies; 22.6%) assessed other, less common functional traits, including measures of pilosity (hairiness), foraging range, colony size, native vs. exotic status, reproductive strategy (parasitism), and the use of different pollen-carrying structures (Figure 2d). Studies sourced their trait data from the literature and published data records (80 studies; 82.5%), from their own measurements and observations (52 studies; 53.6%), and, less commonly, as estimations from allometric equations (10 studies; 10.3%). Below, we discuss measurement strategies, terminology, and possible quantification pitfalls for the most commonly assessed traits in these studies.