Organismal elemental composition can vary within and among species due to differential allocation of limiting resources to maintenance, growth, defense and reproduction (Hessen, Elser, Sterner, & Urabe, 2013; Sterner & Elser, 2002). This variation in elemental composition (i.e. elemental phenotype) is a driver of ecosystem processes such as primary production, secondary production and nutrient cycling (e.g. Elser & Urabe, 1999; Hawlena, Strickland, Bradford, & Schmitz, 2012; Leroux & Schmitz, 2015). Ecosystem ecologists have described the patterns of intra and interspecific variation in elemental composition and are now uncovering mechanisms responsible for this variation (El-Sabaawi et al., 2012; González, Fariña, Kay, Pinto, & Marquet, 2011; Sardans et al., 2016; Vanni & McIntyre, 2016). Phylogeny, ontogeny, trophic level, body size and environmental conditions (e.g. climate) have been demonstrated as key sources of variation explaining the elemental signature of organisms, but evidence from multiple study systems is equivocal. What is more, the majority of studies demonstrating these links are from small-scale experiments or field observations at local extents (Elser et al., 2007). Recent syntheses of data from local studies (e.g. Borer et al., 2013; Martiny et al., 2013) and theory (Leroux et al., 2017; Marleau, Guichard, & Loreau, 2015) have shown how the elemental composition of organisms can vary along geographical and spatial gradients. Consequently, novel insights on the mechanisms for variation in organismal elemental composition may emerge when comparing patterns at multiple spatial extents (Kaspari & Powers, 2016). In this issue of Functional Ecology, Gonzalez et al. (2018) collect an impressive dataset of detritus-based communities inhabiting tank bromeliads from 5 sites spanning >40° latitude across South and Central America to investigate multiple potential drivers of variation in consumer elemental composition. Specifically, they test for evidence of phylogenetic, trophic group and body size signatures across taxonomic resolutions at local (i.e. site) and continental (i.e. all sites) extents. Phylogeny (i.e. node distance) was not or only weakly related to variation in consumer elemental composition (e.g. N at local extent) across geographical extents and taxonomic resolutions. However, taxonomy (i.e. order, family) explained a substantial amount of the interspecific variation in consumer elemental content at local and continental extents. This novel two-tiered approach to detect evolutionary signals in elemental composition uncovered results that are consistent with other studies showing a strong influence of taxonomic resolution on consumer elemental composition (Fagan et al., 2002; González et al., 2011; Hendrixson, Sterner, & Kay, 2007; Woods, Fagan, Elser, & Harrison, 2004). Based on the Growth Rate Hypothesis (GRH), we expect small organisms to grow faster and to have higher nutrient content, particularly P, than large organisms to fuel RNA and protein synthesis (Elser, Dobberfuhl, MacKay, & Schampel, 1996; Hessen et al., 2013; Sterner & Elser, 2002). Although Gonzalez et al. (2018) is not a direct test of the GRH, they did observe the highest elemental contents (N, P) and lowest C:nutrient ratios in detrital consumers with small body sizes. There is a well-known elemental bottleneck at the detritivore and herbivore trophic levels as these organisms feed on resources (i.e. autotrophs, detritus) with relatively high C:nutrient content compared to their own body composition (Boersma et al., 2008; Leroux, Hawlena, & Schmitz, 2012; Leroux & Schmitz, 2015). Carnivores, however, feed on a high-quality resource (i.e. lower C:nutrient ratio) relative to detritivores (or herbivores). Gonzalez et al. (2018) found that organisms at high trophic levels had higher N content and lower C:N ratio than their prey. But carnivores had lower P than their prey and shredders had higher N than filter feeders, gatherers and scrapers. Future work should attempt to assess the role of energy limitation and other potential drivers of the differing results for C, N and P distribution across trophic levels. Finally, the authors observed large variation in P, N:P ratio and C:P ratio within species across geographic space and geography impacted the magnitude but not the qualitative relationships observed. A number of features make this study a strong test of the potential sources of variation in consumer elemental variation. First, the study design with multiple sites spanning a large latitudinal gradient offers a unique macroecological perspective which could shed light on general drivers and context dependencies in elemental composition with implications for ecosystem processes. This study is part of the growing field of spatial ecological stoichiometry (Collins et al., 2017; Kaspari & Powers, 2016; Leroux et al., 2017; Marleau et al., 2015; Prater et al., 2017; Sitters, Atkinson, Guelzow, Kelly, & Sullivan, 2015), which has the potential to improve our predictions of the causes and consequences of variation in elemental composition and the fluxes of matter among elemental pools under global changes. Second, the attention to testing multiple competing and interacting factors (i.e. phylogeny, taxonomy, trophic group, body size) provides a very thorough treatment of potential drivers of organismal elemental composition. Third, the majority of large-scale spatial organismal stoichiometry is done for autotrophs (e.g. Borer et al., 2013; Martiny et al., 2013; Sardans et al., 2016), so this study fills an important gap in our understanding of ecological, evolutionary and geographical correlates for variation in consumer elemental composition. While Gonzalez et al. (2018) investigate the role of invertebrate trophic groups, a next step may be to explicitly build quantitative food webs to investigate the patterns and causes of variation in elemental composition along food chains (Boersma et al., 2008; Leroux et al., 2017). Two ways to overcome the immense challenge of collecting food web data across large spatial extents are to study relatively simple ecosystems (e.g. boreal forest, Leroux et al., 2017) or to make use of distributed experimental networks (e.g. NutNet, Borer et al., 2014). Ecological stoichiometry has a long history in aquatic ecosystems, but we know that energy and nutrient flows differ substantially in aquatic and terrestrial ecosystems (Elser et al., 2007; Shurin, Gruner, & Hillebrand, 2006). Future empirical work should conduct cross-ecosystem comparisons to uncover general causes of variation in organismal elemental composition and elemental fluxes and the consequences of this variation for ecosystem processes (Hessen et al., 2013). Mathematical models integrating food web and ecosystem theories (Loreau, 2010) can contribute to advancing this research agenda. I am grateful to J. Marleau and C. Fox for providing constructive feedback on an earlier draft of this commentary.