Skip to main content
NCBI home page
Nature Communications logoLink to Nature Communications
. 2026 Sep 4;17:10519. doi: 10.1038/s41467-026-77477-1

Global spectrum of leaf nitrogen allocation driven by evolution and environment

Erqian Cui 1,2, Songbo Tang 1,2, Jianyang Xia 1,2,✉
PMCID: PMC13638996  PMID: 42834096

Abstract

Terrestrial carbon uptake under climate change is significantly constrained by nitrogen (N) availability. While plants can alleviate this limitation through flexible leaf N allocation, the global patterns of such partitioning remain poorly understood. Here, by analyzing a global dataset of 520 seed plant species spanning 98 families, we find that plants averagely allocate 38.7% of leaf N to photosynthetic components (including Rubisco, bioenergetics, and light‑capture machinery) and 13.1% to cell walls. The leaf N allocation pattern exhibits a distinct evolutionary shift: early-diverging plants preferentially invest N in cell walls, while recently evolved species allocate more N to photosynthesis. This pattern is fundamentally constrained by evolutionary history and further shaped by recent environmental factors, ultimately leading to a trade-off in leaf N partitioning between photosynthetic components and cell walls. Such a trade-off refines the leaf economics spectrum, demonstrating that conservative-strategy species favor structural investment (cell walls), whereas acquisitive-strategy species prioritize metabolic gain (photosynthesis). These findings suggest that incorporating lineage-specific N allocation strategies can improve projections of global vegetation responses to nutrient limitation and climate change.

Subject terms: Carbon cycle, Ecosystem services, Phylogenetics

Plants need nitrogen to absorb CO2, but how this nutrient is allocated among leaf functions remains poorly understood. Analyzing 520 plant species, this study finds a fundamental trade-off in leaf N partitioning between photosynthesis and cell walls, a pattern constrained by deep evolutionary history and modified by environmental factors.

Introduction

Modern Earth system models increasingly incorporate nitrogen (N) limitation to improve predictions of plant growth and carbon sequestration1,2. However, most models assume a fixed coupling between leaf N content and photosynthetic capacity, typically scaling the maximum carboxylation rate directly to leaf N3. This simplification overlooks the dynamic nature of N allocation within leaves, where substantial variation exists across plant species and functional types. At the physiological level, N supports a wide array of biomolecules, including proteins, phospholipids, nucleic acids, chlorophyll, and alkaloids4. Within leaves, N is partitioned between the photosynthetic apparatus, comprising Rubisco (PR), bioenergetic components (PB), and light-capture proteins (PL), and structural components such as cell walls (PCW)5–7. While photosynthetic N drives carbon assimilation, N investment in cell walls promotes leaf toughness and longevity, enhancing structural defense8. The balance between these functions varies widely among species and reflects key adaptive trade-offs9. Thus, deciphering plant N allocation within leaves is crucial for understanding plant adaptive strategies and improving trait-based vegetation models, particularly under global trends of declining N availability and increasing environmental stress10.

The observed N allocation patterns in leaves reveal substantial differences in N utilization across plant functional groups and taxa. For example, herbaceous plants allocate a larger fraction of leaf N to Rubisco compared with woody species, supporting higher photosynthetic rates in shorter-lived leaves11. Gymnosperms, such as Picea, allocate a lower fraction of leaf N to photosynthetic components than angiosperms, but invest more N in protective compounds (e.g., structural proteins)12. Similar patterns are observed in plant invasions: invasive species often exhibit increased N allocation to photosynthetic components and reduced investment in cell walls relative to native species, enhancing growth and competitiveness13,14. Together, these findings suggest a general trade-off in N partitioning, with species ranging from high photosynthetic investment and short leaf lifespan to high structural investment and longevity. However, how leaf N allocation patterns shape the broader spectrum of coordinated plant resource economy strategies remains poorly understood.

Seed plants have evolved a wide array of nutrient acquisition and utilization strategies over hundreds of millions of years. However, how evolutionary history has shaped nutrient-use strategies remains unclear. Early-diverging lineages emerged in harsh environments characterized by desiccation, extreme temperatures, intense UV radiation, and biotic pressures such as herbivores and pathogens15. To cope with these stresses, ancestral plants may have prioritized investment into cell wall compounds that supported both osmoregulatory function and structural defense16. As these ancient taxa gradually transformed the terrestrial environment by altering hydrothermal conditions, atmospheric composition, and soil nutrient availability, they paved the way for the rise of more resource-acquisitive strategies17,18. In this context, more recently diverged lineages may have shifted toward greater investment in photosynthetic machinery to capitalize on improved resource availability and accelerate growth and reproduction. Indeed, photosynthetic capacity has been shown to increase along the phylogenetic continuum, peaking in angiosperms19. Therefore, we hypothesize that leaf N allocation strategies have shifted over evolutionary time, from cell wall-oriented (serving both structural defense and osmoregulation) to photosynthesis-oriented investments. Furthermore, mycorrhizal symbioses have played a central role in plant adaptation by enhancing nutrient uptake under stress. Plants associated with different mycorrhizal types utilize distinct N acquisition pathways and efficiencies20. This directly influences N availability within the plant, a primary driver of internal N allocation between the photosynthetic apparatus and structural components. Meanwhile, recent evidence indicates that plant resource-use strategies are closely linked to mycorrhizal types: arbuscular mycorrhizal (AM)-associated plants generally follow an “acquisitive” strategy, while ectomycorrhizal (EcM)-associated plants follow a “conservative” strategy21. These strategies are linked to leaf N partitioning. Therefore, plants with different mycorrhizal types may exhibit distinct leaf N allocation patterns.

A growing number of studies now quantify N allocation to key leaf processes, including carboxylation, electron transport, light absorption, and cell wall synthesis. Together, these data enable a global synthesis of the leaf N functional spectrum across terrestrial plant species. Here, we compiled a dataset encompassing 520 species across 98 families, integrating measurements of N partitioning among major functional components within leaves. We first assessed phylogenetic signals in these N fractions and identified strong evolutionary conservatism across lineages. Given that plant traits often carry imprints of past climatic conditions during their evolutionary histories22, we then integrated site-level environmental data to quantify the relative contributions of environmental factors and evolutionary history to the observed variation in leaf N partitioning. Motivated by the widely recognized leaf economic spectrum, we investigated the relationship between leaf N allocation and plant ecological strategies. Finally, we identified a global spectrum of plant N allocation and clarified how N allocation corresponds to variation in ecological strategies among plant functional groups.

Results

Phylogenetic patterns of plant leaf N allocation

We summarized N functional fractions in leaves of the 520 studied species using phylogenetic trees (Fig. 1a). The integrated database contained 92 herbaceous species and 428 woody species. Mycorrhizal association types among woody species were highly phylogenetically conserved, with most EcM species belonging to the Fagales and Pinales, whereas AM species occurred across many divergent orders. As shown in the phylogenetic tree, leaf N allocation to the different leaf functions varied by nearly two orders of magnitude among the studied species. For example, the allocation of leaf N to the cell walls exhibited the greatest interspecific variation, ranging from 0.013 in Fagopyrum dibotrys to 0.460 in Castanopsis hystrix. Globally, 0.387 of leaf N was allocated to photosynthetic components (Rubisco: 0.185; electron transport: 0.048; light capture: 0.154), while 0.131 was invested in cell wall structure (Fig. 1b).

Fig. 1. Phylogenetic patterns of plant leaf N allocation.

Fig. 1

Open in a new tab

a Phylogenetic tree of the study species, displaying patterns of leaf N allocation. b Probability density distribution of global leaf N fractions. Values represent mean ± standard deviation (SD), derived from [n] population-level observations. Fractions of leaf N partitioning to Rubisco enzyme, bioenergetics (i.e., electron transport), light capture and cell wall structures are denoted as PR, PB, PL and PCW, respectively. Plant life forms and mycorrhizal types are differentiated by color-coding in the figure. Herb refers to herbaceous plants. Herb-AM denotes herbaceous plants associated with arbuscular mycorrhizal fungi. W-AM and W-EcM denote woody plants associated with arbuscular mycorrhizal and ectomycorrhizal fungi, respectively. Families containing a large number of species are labeled at the periphery of the phylogenetic tree. The number of samples per species is represented by circle size. The botanical elements in this figure are from Openclipart (https://openclipart.org), a Public Domain (CC0 1.0) repository.

We first tested whether phylogenetically related plant species share similar leaf N allocation patterns. As shown in Supplementary Table 1, leaf N fractions exhibited significant phylogenetic signals (all P  <  0.05). Pagel’s λ values for fractions in Rubisco, electron transport, light absorption, and cell walls were 0.864, 0.704, 0.848, and 0.348, respectively (with values closer to 1 indicating stronger evidence of phylogenetic conservatism). Notably, all leaf N fractions exhibited a common statistical pattern: significantly positive λ (> 0), but low and non-significant K values and Moran’s I. Based on our assessment of heterogeneity in phylogenetic signal through node-specific variation analysis (Supplementary Fig. 2), we found that the signal is strongly conserved within specific deep or major clades but weak or absent among recently diverged lineages. This pattern indicates that while deep evolutionary history imposes fundamental constraints on potential leaf N allocation strategies, recent allocation divergence is likely driven by environmental factors at fine taxonomic levels (e.g., at the species or within-genus level).

Constraints and drivers shaping plant leaf N allocation

We further investigated the relative roles of environmental drivers and evolutionary history in regulating leaf N functional allocation using random forest models (Fig. 2a–d). Overall, both environmental factors and deep evolutionary history jointly shaped the observed allocation patterns. Specifically, total leaf N content emerged as the most important predictor. Key environmental drivers included temperature, atmospheric N deposition, soil moisture, and total soil N. In addition, divergence time imposed a significant phylogenetic constraint on allocation strategies. Furthermore, according to weighted linear regression, seed plants evolved to invest more leaf N into the photosynthetic apparatus from evolutionarily ancient to recently diverged taxa (Fig. 2e–h and Supplementary Table 2, all P < 0.05). In comparison, ancient taxa such as Pinaceae exhibited higher leaf N investment in cell walls, while herbaceous plants that have recently diverged, such as Brassicaceae, tended to allocate less leaf N to the cell walls. Together, these attribution and phylogenetic analyses collectively demonstrate that while deep evolutionary history imposes fundamental constraints on allocation possibilities, recent divergences are primarily driven by environmental filtering.

Fig. 2. The relative importance of environmental factors and evolutionary history on leaf N allocation.

Fig. 2

Open in a new tab

a–d Relative contributions of individual drivers to variations in leaf N fractions based on random forest models. MAT: mean annual temperature; MAP: mean annual precipitation; Ndep: total atmospheric N deposition; SPEI: standardized precipitation evapotranspiration index; STN: soil total N; SMO: soil moisture; LN: leaf N concentration; Phy: family-level divergence time. e–h Changes in leaf N fractions (PR, PB, PL and PCW) across divergence time at the population level. Each point corresponds to the mean value of a single species at one sampling site (defined as a population). Analyses were performed using least‑squares regressions weighted by the sample size of each population. Statistical significance was assessed using two‑sided t‑tests for the regression slopes. Shaded zones denote 95% confidence intervals. Regression coefficients are provided in Supplementary Table 2.

Trade-off between leaf nitrogen functional fractions

In pairwise analyses, leaf N functional fractions showed a fundamental trade-off, with significant correlations under the SMA model. Investment in the photosynthetic apparatus declined significantly as allocation to cell walls increased. The negative allometric relationships were consistently strong, with scaling exponents of −0.629 (Rubisco), −0.763 (bioenergetics), and −1.543 (light‑capture components) (Supplementary Fig. 3). The negative correlations among leaf N fractions were also captured by Phy.SMA regression (Supplementary Table 3), confirming that these trade‑offs are robust to phylogenetic correction. This indicates that the observed allocation patterns reflect intrinsic physiological trade‑offs rather than a mere consequence of evolutionary history. Consistent with the random forest analysis, these results suggest that evolutionary history constrains leaf N allocation strategies across environmental gradients.

The fundamental trade-off between photosynthetic and structural allocation defines a bivariate functional spectrum of leaf N, from species that invest more in cell wall proteins to those that prioritize photosynthetic capacity at the expense of structural defense (Fig. 3). Major taxa differed significantly in their occupancy of leaf N functional space. Herbaceous plants were typically confined to the side of higher photosynthetic N investment, whereas ectomycorrhizal woody plants tended to occupy the side with more N invested in cell walls. In contrast, the footprint of AM woody plants spanned the entire leaf N functional spectrum.

Fig. 3. The bivariate functional spectrum of leaf nitrogen.

Fig. 3

Open in a new tab

a–c Relationships between leaf N allocation to photosynthetic apparatus (Rubisco, electron transport, light absorption) and cell wall structures. Colored ellipses represent 95% confidence regions for each group, estimated via 2D kernel density estimation. Coefficients for SMA and Phy.SMA models of these relationships are presented in Supplementary Table 3. d Conceptual diagram illustrating the ecological trade-off in plant leaf N allocation strategies. The fraction of leaf N partitioning to total photosynthetic apparatus is denoted as PP. Each point in a-d corresponds to the mean value of a single species at one sampling site. The error bars represent the standard error of the mean (mean ± SE) calculated across all measurements within each species-site combination.

Plant leaf N allocation strategy refines the leaf economics spectrum

We quantified the relationships between proportional leaf N allocation to key functional fractions and core leaf economic traits (Nm, SLA, and Asat). The proportion of N allocated to photosynthetic components was positively correlated with leaf economic traits, whereas that allocated to cell walls showed a consistent negative correlation (Fig. 4a, P < 0.05). Integrating this with the classic leaf economics spectrum (LES) theory, our findings demonstrate that species with an acquisitive strategy preferentially allocate leaf N to photosynthetic components, while species with a conservative strategy favor allocation to structural cell walls.

Fig. 4. Relationships between leaf N fractions and leaf economic traits.

Fig. 4

Open in a new tab

a Correlation matrix between leaf N fractions and leaf economic traits. Nm: leaf N concentration; SLA: specific leaf area; Asat: leaf light-saturated net photosynthetic rate. b–e Comparison of leaf N allocation between acquisitive and conservative strategy groups (herb vs. woody, W-AM vs. W-EcM). Herb: herbaceous. W-AM and W-EcM denote woody plants associated with arbuscular mycorrhizal and ectomycorrhizal fungi, respectively.

Based on leaf economic traits, two key contrasts represent the acquisitive-conservative strategy spectrum: herbaceous vs. woody plants, and AM vs. EcM woody plants (Supplementary Fig. 4). Correspondingly, we found that leaf N allocation patterns differed significantly between these two pairs of groups. As shown in Fig. 4b, c, herbaceous plants had the highest fractional leaf N allocation to photosynthesis (i.e., 0.227 of leaf N for Rubisco, 0.066 of leaf N for bioenergetics, 0.149 of leaf N for light‑capture components). In comparison, woody plants invested more leaf N in cell walls (all P < 0.05, Supplementary Fig. 5a). Additionally, the woody species associated with distinct mycorrhizal types differ markedly in their N use strategy (Fig. 4d, e). The allocation of leaf N in the cell walls of EcM woody plants was highest at 0.194, which was 1.7 times higher than in AM woody plants and 2.4 times higher than in herbaceous plants. For photosynthetic functions, EcM woody plants tended to allocate more N into Rubisco, while AM woody plants invested more N into light capture (all P < 0.05, Supplementary Fig. 5b). Our findings refine the leaf economics spectrum by demonstrating that conservative-strategy species prioritize structural durability through investment in cell walls, whereas acquisitive-strategy species invest heavily in metabolic gain (photosynthesis).

Discussion

Environment and evolution jointly regulate leaf N allocation strategies

Plants inherently face trade-offs in allocating resources between investment in carbon assimilation for growth and investment in cell walls for multifaceted functions, such as osmoregulation and defense against herbivores/pathogens23,24. Our analysis reveals that leaf N partitioning between photosynthetic components and cell‑wall fractions varies substantially across species, and this interspecific variation follows a distinct evolutionary trajectory. Early-diverging lineages predominantly invest N in cell walls, whereas later-diverging lineages shift this investment toward photosynthetic components. This evolutionary shift in N allocation likely reflects a fundamental transition in selective pressures during the colonization of land by plants. The allocation strategy serves two core functions: first, the cell walls play a critical role in maintaining osmoregulation by regulating cell turgor pressure and stabilizing water balance, thereby helping plants cope with water stress during their transition to land; second, the cell walls act as a physical barrier, providing structural defense against herbivores and pathogens, with its components such as lignin and cellulose directly contributing to tissue resistance and durability. Plant behaviors are typically shaped by both evolutionary history and local environmental pressures25,26. Accordingly, our random forest analysis shows that leaf N allocation patterns are jointly determined by environmental factors and deep evolutionary history, with atmospheric N deposition, temperature, soil total N, and soil moisture as the primary environmental drivers (Fig. 2). Thus, the coordination between photosynthetic and structural N use reflects not only evolutionary trajectories but also adaptive strategies that enable plants to acclimate to a changing environment (Fig. 5).

Fig. 5. Conceptual model of plant N allocation strategies shaped by evolution and environment.

Fig. 5

Open in a new tab

Fractions of leaf N partitioning to total photosynthetic apparatus and cell wall structures are denoted as PP and PCW, respectively. AM and EcM species denote plants associated with arbuscular mycorrhizal and ectomycorrhizal fungi, respectively. The leaf economic spectrum was represented by leaf N concentration (Nm), specific leaf area (SLA), and light-saturated net photosynthetic rate (Asat). Higher values indicate an acquisitive strategy; lower values, a conservative strategy. The botanical elements in this figure are from Openclipart (https://openclipart.org), a Public Domain (CC0 1.0) repository.

Meanwhile, under different environmental conditions, plants dynamically adjust N allocation to optimize local survival27. Therefore, a trade-off in leaf N allocation between photosynthesis and cell wall also exists within plant species28,29. By integrating evidence from global manipulative experiments, we found that changes in atmospheric CO2 concentration and light conditions drive a trade-off in leaf N allocation between photosynthetic components and cell walls (Supplementary Fig. 6). For example, under elevated atmospheric CO2, plants significantly reduced N allocation to carboxylation enzymes while increasing the proportion allocated to cell walls. Under low-light conditions, plants markedly raised the proportion of leaf N to light capture components and decreased the proportion allocated to cell walls. Such a trade-off likely arises from two underlying mechanisms. First, the biosynthesis of both chloroplasts and cell walls requires substantial N, leading to direct competition between cell expansion and photosynthetic development30. Second, a higher N allocation to cell walls, along with thicker mesophyll cell walls, can restrict CO2 diffusion to chloroplasts, thereby reducing the N requirements for photosynthesis7,9. Our findings reinforce the concept that resistance is costly, diverting resources from growth24.

Leaf N allocation reveals a clear evolutionary trajectory: early‑diverging lineages prioritize N investment in structural cell walls, while more recently diverged lineages shift investment toward photosynthetic components. The underlying mechanisms can be summarized as follows: while deep evolutionary history imposes fundamental phylogenetic constraints on allocation strategies, recent divergences are primarily driven by environmental filtering. This N partitioning pattern directly refines the leaf economics spectrum, demonstrating that conservative-strategy species favor structural durability (cell walls), whereas acquisitive-strategy species invest heavily in metabolic gain (photosynthesis). These strategies are clearly distinct among functional groups: herbaceous plants and AM woody plants typically employ an acquisitive strategy with high photosynthetic N investment, while EcM woody plants generally adopt a conservative strategy, directing more N to structural cell walls.

Tree mycorrhizal association types control leaf N allocation patterns

Almost all seed plants form symbiotic relationships with mycorrhizal fungi, which enhance nutrient uptake and improve tolerance to environmental stress31,32. Mycorrhizae are primarily recognized for their role in phosphorus (P) acquisition, as P is considerably less mobile and more strongly sorbed to soil particles than N33. Their influence on plant N relations is nonetheless well documented, operating through both direct N transfer from fungal hyphae and indirect effects on soil N mobilization and root exploration. Recent studies have shown that species associated with different mycorrhizal types exhibit distinct N acquisition strategies21,31, yet whether these differences extend to internal N partitioning among functional pools remains largely unexplored. Our study shows that mycorrhizal type also modulates how N is functionally allocated within leaves (Fig. 3). EcM plant species tend to follow a more conservative N-use strategy, investing more N in structural defenses such as cell walls, consistent with their slower growth and longer resource payback times. In contrast, AM species adopt an acquisitive N-use strategy, allocating more N to the photosynthetic apparatus to support fast growth over shorter time scales. These findings are supported by mycorrhizal inoculation experiments, which have shown that AM fungi can shift N allocation toward photosynthesis at the expense of structural components34,35. Thus, adjustments in the allocation of N within leaves may be a key mechanism by which mycorrhizas facilitate plant adaptation to environmental changes.

We also found distinct differences in N allocation potential between the two major mycorrhizal groups. EcM species clustered more tightly along the bivariate functional spectrum of leaf N, whereas AM species exhibited more diverse allocation patterns (Fig. 4). This difference may help explain the dominance of AM trees in species-rich forests36. Notably, the greatest divergence between AM and EcM species occurs in N allocation to light-capture components (Supplementary Fig. 5b). This pattern may stem from the prevalence of EcM symbioses in conifer-dominated forests with lower canopy diversity and reduced light competition, diminishing the need to invest heavily in light capture32. Our dataset also includes Fabaceae species, which are capable of symbiotic N fixation. This functional group may adopt fundamentally different N allocation strategies, as their access to atmospheric N allows them to bypass soil N limitation. With greater total N availability, they can sustain photosynthetic and structural functions with a smaller proportional N investment. To test this expectation, we compared N allocation patterns between Fabaceae and non‑Fabaceae species in our dataset. Overall, Fabaceae species exhibit slightly lower proportional N investments than non‑Fabaceae species across most functional pools, with a statistically significant difference observed only in the light‑capture fraction (P < 0.05, Supplementary Fig. 5c). This pattern may reflect the relatively higher total N status of Fabaceae, which reduces the need for preferential N allocation to functional components. However, differences in species provenance or environmental backgrounds between the two groups may have partially confounded this pattern. Future studies under controlled nutrient conditions are needed to directly examine the linkage between belowground N fixation capacity and aboveground N allocation strategies. Together with our findings on mycorrhizal associations, these results highlight the importance of symbiotic partnerships, whether with fungi or N‑fixing bacteria, in shaping leaf N allocation patterns. Understanding these microbial interactions provides insight into plant functional diversity and adaptive responses across ecosystems37.

Leaf N allocation deepens the understanding of the leaf economics spectrum

Our findings underscore that leaf N allocation trade-offs represent a fundamental physiological mechanism underpinning the classic LES theory. While the LES robustly describes the covariation between leaf traits such as leaf N concentration, photosynthetic rate, and specific leaf area, it primarily reflects the consequence of strategic resource investment38,39. Our analysis demonstrates that this spectrum is partly driven by the internal partitioning of leaf N between competing functional pools (Supplementary Fig. 7a). Acquisitive-strategy species, characterized by high photosynthetic rates and specific leaf area, achieve their strategy not merely through higher total leaf N, but by preferentially allocating a greater proportion of this N to compounds directly powering photosynthesis, primarily Rubisco. Conversely, conservative-strategy species invest a larger proportion of their lower total N into structural components (e.g., cell wall proteins), thereby enhancing tissue toughness and longevity at the expense of immediate photosynthetic gain. This N partitioning trade-off thus provides the mechanistic link that explains the tight coupling among photosynthetic capacity, leaf N concentration, specific leaf area, and leaf lifespan observed globally.

We also quantified the relationships between area-based and mass-based leaf N fractions. The positive correlation among photosynthetic component concentrations aligned with the pattern observed for their proportions. However, the trade-off between photosynthetic and cell wall component concentrations was relatively weak (Supplementary Fig. 7b, c). This pattern can be explained by two fundamental principles. First, photosynthetic components are synergistic, requiring coordinated investment to optimize function, resulting in strong positive correlations regardless of whether concentrations or proportions are considered. Second, the structural constraints impose a minimum investment threshold for cell walls to maintain leaf quality and defense. Once this baseline requirement is satisfied, additional N becomes available for flexible allocation toward photosynthetic components. Consequently, when expressed as concentrations, both pools may increase concurrently as total leaf N rises, thereby masking the fundamental allocation trade-off. In contrast, proportions inherently reflect allocation priorities within a fixed total N pool, making the trade-off between photosynthetic gain and structural investment more pronounced.

Limitations and future directions

In conclusion, our analysis reveals a clear evolutionary trajectory in leaf N allocation: from gymnosperms to angiosperms, N investment shifts from structural cell walls toward photosynthetic components. This pattern arises from a dual-layered mechanism: deep evolutionary history imposes phylogenetic constraints, whereas environmental factors drive divergence at fine taxonomic levels (Fig. 5). These findings refine the LES theory by explicitly linking N partitioning to ecological strategy: conservative species prioritize structural durability (cell walls), while acquisitive species maximize photosynthetic gain. Given limited data availability, some of our conclusions are based on a relatively small number of species and therefore need to be validated with future high-resolution, large-scale datasets on N allocation across diverse lineages and biomes. In the current literature, leaf N allocation is estimated primarily through two complementary approaches. While the allocation to cell walls relies on direct chemical measurement, the photosynthetic fractions are partly estimated from trait-based models that incorporate co-measured physiological data. These established approaches provide complementary and empirically grounded estimates; nonetheless, the methodological distinctions between direct measurement and modeling may introduce a degree of uncertainty into the quantitative comparisons.

Leaf N content and its functional allocation are fundamental controls of photosynthetic capacity and N-use efficiency40,41. Our findings reveal substantial interspecific variation in N partitioning, uncovering a global spectrum of N use strategies that underpin plant productivity. Substantial evidence shows that C4 plants achieve higher N-use efficiency than C3 plants, an advantage attributed to the C4 CO2‑concentrating mechanism, which suppresses photorespiration and boosts Rubisco catalytic efficiency, thereby reducing the N cost of Rubisco42. The N thus saved can be reallocated to other photosynthetic components, resulting in higher photosynthetic N-use efficiency43. However, the limited representation of C4 species in our dataset precluded robust statistical comparisons. Future studies with expanded C4 sampling are urgently needed to test whether these known physiological differences in N-use efficiency are reflected in distinct allocation patterns between photosynthetic and structural components across these two functional types. Our dataset includes a small number of agricultural and medicinal species, whose N utilization may be subject to anthropogenic interference. This may introduce a confounding influence on the phylogenetic analyses. Nevertheless, their inclusion provides valuable insights into the nutrient utilization patterns of these important species.

In the current generation of Earth system models, core photosynthetic parameters, such as the maximum carboxylation rate (Vc,max) and maximum electron transport rate (Jmax), are typically scaled to total leaf N44. To capture spatial heterogeneity, these models adopt a range of N-use schemes, including plant functional type-specific allocation45, or climate-driven allocation of leaf N1. While these approaches improve global model performance, they often fail to capture trait variation across lineages or local ecological contexts. Emerging evidence indeed suggests that trait-environment relationships alone are insufficient to predict trait variation46. Our results underscore that phylogeny, alongside mycorrhizal association types, plays a critical role in shaping plant N allocation strategies. Incorporating these evolutionary and symbiotic constraints into model structures will be essential for improving projections of global vegetation responses to nutrient limitation and climate change. Overall, advancing beyond total leaf N to explicitly represent within-leaf N allocation offers a key step toward more mechanistic and predictive trait-based Earth system models.

Methods

Data collection

To explore the N functional allocation strategy of plant leaves, we compiled a global database of leaf N allocation among multiple functions from published literature using Web of Science (http://apps.webofknowledge.com), Google Scholar (http://scholar.google.com), and the National Knowledge Infrastructure Database (http://cnki.net). Here, we mainly focus on the allocation of leaf N to a range of photosynthetic and defense-related functions, including ribulose-1,5-bisphosphate carboxylase-oxygenase enzyme (Rubisco, PR), bioenergetics (i.e., electron transport, PB), light absorption by chlorophyll (PL), and cell wall structure (PCW). We used the combination of keywords ((TS = (leaf)) OR TS = (foliar)) AND (((TS = (nitrogen allocation)) OR TS = (nitrogen fraction)) OR TS = (nitrogen partitioning)). For each study, the data were directly obtained from the text, tables, and supplementary materials of the literature or acquired from graphics using GetData Graph Digitizer v2.26. To minimize publication bias across different data sources, we considered only studies that included at least two types of leaf N components. Considering the assumption of independent observations, we calculated the site-specific mean value for a given species (defined as a population). Along with the mean value, we also recorded the corresponding sample size, variance, and sampling year for each population. To maximize the preservation of raw data variability and explicitly incorporate intraspecific variation, we therefore conducted all core analyses at this population level. After systematic screening, a total of observations for 520 seed plant species spanning 98 families were included in our database. However, as not all original studies measured every N fraction, the number of species with valid data varies across different N fractions. Consequently, the sample species are 516 for photosynthetic carboxylation (PR), 396 for photosynthetic bioenergetics (PB), 378 for light-harvesting complexes (PL), and 123 for cell walls (PCW). The final dataset was derived from 157 literature sources, spanning 146 sampling locations (Supplementary Fig. 1).

In the current literature, leaf N allocation is estimated primarily through two complementary approaches: the direct extraction method and trait-based modeling. The chemical extraction method is mainly used to extract N in Rubisco, chlorophyll for light capture, the cytochrome f complex involved in electron transport, and the cell walls. Trait-based modeling integrates a suite of established ecological and physiological mechanisms to estimate how total leaf N is partitioned among key photosynthetic pools: carboxylation (e.g., Rubisco), bioenergetics (e.g., electron transport components), and light-harvesting complexes. These allocations are inferred from combined measurements of gas exchange and chlorophyll content. Detailed extraction procedures and calculation formulas are provided in Supplementary Note 1.

In addition to leaf N allocation data, we also compiled a site-level information database containing geographic coordinates, soil properties (STN: soil total N; SMO: soil moisture), and climatic variables (MAP: mean annual precipitation; MAT: mean annual temperature; SPEI: standardized precipitation evapotranspiration index; Ndep: N deposition) for each observation. Environmental data were initially extracted from the original publications. Missing temperature and precipitation data were extracted from the CRU dataset47 using longitude and latitude information. Similarly, missing soil total N was extracted from soilGrid250m48, while soil moisture data were extracted from the Global Land Evaporation Amsterdam Model (GLEAM; https://www.gleam.eu/). The SPEI was calculated from potential evapotranspiration (PET) and MAP data using the spei function in the SPEI R package, while the required MAP and PET data were obtained from the CRU dataset. The N deposition data were obtained from a global atmospheric N deposition dataset49.

Species and phylogeny

To explore the effects of evolutionary history on plant N use strategies, we conducted phylogenetic analyses on the collected species. All species names were first confirmed through The Plant List database (http://www.theplantlist.org) using the R package ‘Taxonstand’50. Then the phylogenetic trees for the 520 species were constructed with the ‘V.PhyloMaker2’ package in R51. To test the phylogenetic signals for each leaf N fraction among these species, we estimated Pagel’s λ, Blomberg’s K and Moran’s I using phylosig in the R package ‘phytools’52. Values of λ range from 0 to 1, while λ  =  0 indicates no phylogenetic dependence, where λ  =  1 implies that variation in leaf N fractions completely depends on phylogeny53. Blomberg’s K evaluates whether trait similarity among related species is greater than expected under Brownian motion; K > 1 indicates stronger phylogenetic conservatism than expected, while K < 1 suggests convergent evolution. Moran’s I measures phylogenetic autocorrelation, where I > 0 indicates that closely related species tend to have similar trait values (positive autocorrelation), and I < 0 suggests dissimilarity among relatives (overdispersion). For all indices, significance was assessed against the null hypothesis of no phylogenetic signal (λ  = 0, K = 1, or I = 0) using likelihood ratio tests or permutation tests (P-values).

Furthermore, to investigate the evolutionary trend of N allocation among multiple functions in plant leaves, we quantified the variation of leaf N fractions over divergence time. The divergence times of the studied 98 families are based on the ages estimated under the relaxed calibration strategy with the complete fossil set54. Subsequently, we employed linear regression models to examine the relationships between leaf N fractions and divergence time, using each population's sample size as the weighting factor. Four of the 98 families in our study were omitted from the relevant analyses as their divergence times could not be matched.

Statistical analyses

To explore differences in N allocation among major evolutionary clades, we divided the species into four types: Herbaceous (Herb), Woody-AM, Woody-EcM, or Woody-others. The mycorrhizal types of woody plants are obtained from the FungalRoot database55 and the FRED 3.0 database56. If a specific mycorrhizal association was not available for a particular species, we assigned the species to the mycorrhizal group of its genus32. Woody-AM and Woody-EcM are woody plants that are associated with arbuscular mycorrhizal and ectomycorrhizal fungi, respectively. The term ‘Woody-others’ here refers to the woody plants with other types of mycorrhizas, including orchid mycorrhizas (ORM), ericoid mycorrhizas (ERM), nonmycorrhizal (NM) roots, or inconsistent mycorrhizas (AM–NM). Consider the non-normal distribution of some grouped dataset, we conducted a non-parametric test (wilcox.test) using the ‘stats’ package in R (3.6.1) to compare the differences in N allocation strategies among different life forms and mycorrhizal types.

Considering the allometric relationships among leaf N fractions (Supplementary Fig. 2), we employed standardized major axis (SMA) regressions to examine the relationships between N allocation to photosynthetic apparatus versus cell walls using the ‘smatr’ package in R (3.6.1). To account for the influence of evolutionary history, we conducted phylogenetic independent contrasts (PICs) on the fraction data. We then applied SMA regression to these contrasts (hereafter “Phy.SMA”). Subsequently, to explicitly evaluate the influence of phylogenetic history, we compared the slopes from Phy.SMA and SMA models. To visualize the occurrence probability of given combinations of leaf N fractions in a bivariate space, we constructed contour plots from bivariate kernel density distributions. The 2D kernel density estimation was conducted using the geom_density_2d function, and the results were presented with contours. This method can be useful for dealing with overplotting among different plant groups.

Variable selection and model construction

In this study, we employed a Random Forest (RF) regression approach to quantify the relative contributions of key drivers to leaf N allocation patterns. This method was chosen not for predictive purposes, but specifically for its robustness in assessing variable importance and capturing non-linear relationships in complex ecological datasets. Drivers were organized into three categories: (i) Climate: N deposition (Ndep), mean annual precipitation (MAP, mm), and mean annual temperature (MAT, °C); (ii) Soil: total soil N (STN) and soil moisture (SMO); and (iii) Plant traits: total leaf N content and divergence time at the family level. All analyses were implemented using the randomForest package in R57. Model performance in explaining the variance in leaf N allocation was evaluated using the coefficient of determination (R²) and the P-value, while the relative importance of each predictor was quantified via the percentage increase in mean squared error (%IncMSE). This approach allows us to decompose and rank the influence of multiple drivers on the observed allocation patterns.

Effect size calculation

Our database also includes manipulative experiments investigating leaf N allocation variation. In the analyses presented above, we used only control group data to establish baseline patterns. Furthermore, we used paired data from these manipulative experiments (treatment vs. control groups) to elucidate how plants adjust leaf N allocation in response to environmental changes. The log-transformed response ratio (LnRR) was calculated as the effect size to quantify the direction and magnitude of these adjustments across different experimental treatments (e.g., elevated CO₂, N addition, and light availability). An LnRR close to zero indicates little effect relative to the control. We used the rma.mv() function from the ‘metafor’ package58 to fit mixed-effects, meta-analytic models. These models were used to estimate the overall effect size and its 95% confidence intervals, while incorporating a random effects structure to account for non-independence among effect sizes (e.g., within studies).

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.

Supplementary information

Supplementary Information (2MB, docx)
Reporting Summary (75.3KB, pdf)
Transparent Peer Review file (1.6MB, pdf)

Source data

Source Data (455.8KB, xlsx)

Acknowledgements

We gratefully acknowledge the following data sources: the CRU dataset for temperature, precipitation, and potential evapotranspiration; SoilGrids250m for soil total nitrogen; GLEAM for soil moisture; and Tian et al. for climatic, atmospheric CO2, and nitrogen deposition data49. We also extend our thanks to the authors of all original data source publications for making their data available.

Author contributions

E.C. and J.X. devised and conducted the analysis. E.C. collected the dataset and drafted the first manuscript. S.T. and J.X. provided critical suggestions on the method and results. All authors contributed to the writing and revising of the manuscript.

Peer review

Peer review information

Nature Communications thanks Congcong Liu and the other, anonymous, reviewer(s) for their contribution to the peer review of this work. A peer review file is available.

Funding

This work was financially supported by the National Natural Science Foundation of China (32325033, 42301066), the Fundamental and Interdisciplinary Disciplines Breakthrough Plan of the Ministry of Education of China (JYB2025XDXM904).

Data availability

The processed data underlying all figures and tables in this manuscript are provided as a Source Data file accompanying the paper. The raw datasets generated in this study, including leaf nitrogen fractions for 520 seed plant species spanning 98 families, have been deposited in Zenodo (https://doi.org/10.5281/zenodo.21624103)59. A complete list of paper source references is provided in the Supplementary Information. Source Data is provided as a Source data file. Source data are provided with this paper.

Code availability

The codes used in this study are available are available via Zenodo at https://doi.org/10.5281/zenodo.2162410359.

Competing interests

The authors declare no competing interests.

Footnotes

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary information

The online version contains supplementary material available at https://doi.org/10.1038/s41467-026-77477-1.

References

  • 1.Ali, A. A. et al. A global scale mechanistic model of photosynthetic capacity (LUNA V1. 0). Geosci. Model Dev.9, 587–606 (2016). [Google Scholar]
  • 2.Stocker, B. D. et al. Terrestrial nitrogen cycling in Earth system models revisited. N. Phytol.210, 1165–1168 (2016). [DOI] [PubMed] [Google Scholar]
  • 3.Luo, X. et al. Global variation in the fraction of leaf nitrogen allocated to photosynthesis. Nat. Commun.12, 4866 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4.Gaudinier, A. et al. Transcriptional regulation of nitrogen-associated metabolism and growth. Nature563, 259–264 (2018). [DOI] [PubMed] [Google Scholar]
  • 5.Ghimire, B. et al. A global trait-based approach to estimate leaf nitrogen functional allocation from observations. Ecol. Appl.27, 1421–1434 (2017). [DOI] [PubMed] [Google Scholar]
  • 6.Onoda, Y. et al. Physiological and structural tradeoffs underlying the leaf economics spectrum. N. Phytol.214, 1447–1463 (2017). [DOI] [PubMed] [Google Scholar]
  • 7.Xue, W. et al. Cell wall thickness has phylogenetically consistent effects on the photosynthetic nitrogen-use efficiency of terrestrial plants. Plant Cell Environ.46, 2323–2336 (2023). [DOI] [PubMed] [Google Scholar]
  • 8.Flexas, J. et al. Cell wall thickness and composition are involved in photosynthetic limitation. J. Exp. Bot.72, 3971–3986 (2021). [DOI] [PubMed] [Google Scholar]
  • 9.Evans, J. R. & Clarke, V. C. The nitrogen cost of photosynthesis. J. Exp. Bot.70, 7–15 (2019). [DOI] [PubMed] [Google Scholar]
  • 10.Mason, R. E. et al. Evidence, causes, and consequences of declining nitrogen availability in terrestrial ecosystems. Science376, eabh3767 (2022). [DOI] [PubMed] [Google Scholar]
  • 11.Hikosaka, K. & Shigeno, A. The role of Rubisco and cell walls in the interspecific variation in photosynthetic capacity. Oecologia160, 443–451 (2009). [DOI] [PubMed] [Google Scholar]
  • 12.Wang, J. et al. Variation, coordination, and trade-offs between needle structures and photosynthetic-related traits across five Picea species: consequences on plant growth. BMC Plant Biol.22, 242 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 13.Feng, Y. L. et al. Evolutionary tradeoffs for nitrogen allocation to photosynthesis versus cell walls in an invasive plant. Proc. Natl. Acad. Sci. USA106, 1853–1856 (2009). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14.Qing, H., Cai, Y., Xiao, Y., Yao, Y. & An, S. Leaf nitrogen partition between photosynthesis and structural defense in invasive and native tall form Spartina alterniflora populations: effects of nitrogen treatments. Biol. Invasions14, 2039–2048 (2012). [Google Scholar]
  • 15.Benton, M. J., Wilf, P. & Sauquet, H. The Angiosperm Terrestrial Revolution and the origins of modern biodiversity. N. Phytol.233, 2017–2035 (2022). [DOI] [PubMed] [Google Scholar]
  • 16.Veromann-Jürgenson, L. L., Tosens, T., Laanisto, L. & Niinemets, Ü Extremely thick cell walls and low mesophyll conductance: welcome to the world of ancient living!. J. Exp. Bot.68, 1639–1653 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17.Dahl, T. W. & Arens, S. K. The impacts of land plant evolution on Earth’s climate and oxygenation state-an interdisciplinary review. Chem. Geol.547, 119665 (2020). [Google Scholar]
  • 18.Kenrick, P. & Crane, P. R. The origin and early evolution of plants on land. Nature389, 33–39 (1997). [Google Scholar]
  • 19.Gago, J. et al. Photosynthesis optimized across land plant phylogeny. Trends Plant Sci.24, 947–958 (2019). [DOI] [PubMed] [Google Scholar]
  • 20.Bonfante, P. & Genre, A. Mechanisms underlying beneficial plant-fungus interactions in mycorrhizal symbiosis. Nat. Commun.1, 48 (2010). [DOI] [PubMed] [Google Scholar]
  • 21.Deng, M. et al. Tree mycorrhizal association types control biodiversity-productivity relationship in a subtropical forest. Sci. Adv.9, eadd4468 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 22.Famiglietti, C. A., Worden, M., Anderegg, L. D. & Konings, A. G. Impacts of climate timescale on the stability of trait-environment relationships. N. Phytol.241, 2423–2434 (2024). [DOI] [PubMed] [Google Scholar]
  • 23.Cosgrove, D. J. Growth of the plant cell wall. Nat. Rev. Mol. Cell Biol.6, 850–861 (2005). [DOI] [PubMed] [Google Scholar]
  • 24.Monson, R. K., Trowbridge, A. M., Lindroth, R. L. & Lerdau, M. T. Coordinated resource allocation to plant growth-defense tradeoffs. N. Phytol.233, 1051–1066 (2022). [DOI] [PubMed] [Google Scholar]
  • 25.Reich, P. B. et al. The evolution of plant functional variation: traits, spectra, and strategies. Int. J. Plant Sci.164, S143–S164 (2003). [Google Scholar]
  • 26.Marcellus, M., Goud, E. M., Swartz, N., Brown, E. & Soper, F. M. Evolutionary history and root trait coordination predict nutrient strategy in tropical legume trees. N. Phytol.243, 1711–1723 (2024). [DOI] [PubMed] [Google Scholar]
  • 27.Van Kleunen, M. & Fischer, M. Constraints on the evolution of adaptive phenotypic plasticity in plants. N. Phytol.166, 49–60 (2005). [DOI] [PubMed] [Google Scholar]
  • 28.Lei, Y. B. et al. Innate and evolutionarily increased advantages of invasive Eupatorium adenophorum over native E. japonicum under ambient and doubled atmospheric CO2 concentrations. Biol. Invasions13, 2703–2714 (2011). [Google Scholar]
  • 29.Zhong, C., Jian, S. F., Huang, J., Jin, Q. Y. & Cao, X. C. Trade-off of within-leaf nitrogen allocation between photosynthetic nitrogen-use efficiency and water deficit stress acclimation in rice (Oryza sativa L.). Plant Physiol. Bioch.135, 41–50 (2019). [DOI] [PubMed] [Google Scholar]
  • 30.Onoda, Y., Hikosaka, K. & Hirose, T. Allocation of nitrogen to cell walls decreases photosynthetic nitrogen-use efficiency. Funct. Ecol.18, 419–425 (2004). [Google Scholar]
  • 31.Averill, C., Bhatnagar, J. M., Dietze, M. C., Pearse, W. D. & Kivlin, S. N. Global imprint of mycorrhizal fungi on whole-plant nutrient economics. Proc. Natl. Acad. Sci. USA116, 23163–23168 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 32.Tedersoo, L., Bahram, M. & Zobel, M. How mycorrhizal associations drive plant population and community biology. Science367, eaba1223 (2020). [DOI] [PubMed] [Google Scholar]
  • 33.Kuyper, T. W. & Jansa, J. Arbuscular mycorrhiza: advances and retreats in our understanding of the ecological functioning of the mother of all root symbioses. Plant Soil489, 41–88 (2023). [Google Scholar]
  • 34.Lu, Y., Ma, Q., Chen, C., Xu, X. & Zhang, D. Effects of arbuscular mycorrhizal fungi on the nitrogen distribution in endangered Torreya jackii under nitrogen limitation. Planta254, 53 (2021). [DOI] [PubMed] [Google Scholar]
  • 35.Wu, F., Zhang, H., Fang, F., Liu, H. & Tang, M. Arbuscular mycorrhizal fungi alter nitrogen allocation in the leaves of Populus × canadensis ‘Neva’. Plant Soil421, 477–491 (2017). [Google Scholar]
  • 36.Zhong, Y. et al. Arbuscular mycorrhizal trees influence the latitudinal beta-diversity gradient of tree communities in forests worldwide. Nat. Commun.12, 3137 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 37.Hou, S. et al. A microbiota-root-shoot circuit favours Arabidopsis growth over defence under suboptimal light. Nat. Plants7, 1078–1092 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 38.Reich, P. et al. From tropics to tundra: global convergence in plant functioning. Proc. Natl. Acad. Sci.94, 13730–13734 (1997). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 39.Wright et al. The worldwide leaf economics spectrum. Nature428, 821–827 (2004). [DOI] [PubMed] [Google Scholar]
  • 40.Field, C. H. The photosynthesis-nitrogen relationship in wild plants. On the economy of plant form and function, 25–55 (Springer, 1986).
  • 41.Cui, E., Xia, J. & Luo, Y. Nitrogen use strategy drives interspecific differences in plant photosynthetic CO2 acclimation. Glob. Change Biol.29, 3667–3677 (2023). [DOI] [PubMed] [Google Scholar]
  • 42.Brown, R. H. A difference in N use efficiency in C3 and C4 plants and its implications in adaptation and evolution. Crop Sci.18, 93–98 (1978). [Google Scholar]
  • 43.Bossonia, G. E. B. et al. Assessing the cell wall nitrogen use efficiency–Can the differences between cell wall architectures contribute to the nitrogen economy of plants? Plant Physiol. Biochem.229, 110548 (2025). [DOI] [PubMed] [Google Scholar]
  • 44.Rogers, A. The use and misuse of Vc,max in Earth system models. Photosynth. Res.119, 15–29 (2014). [DOI] [PubMed] [Google Scholar]
  • 45.Thornton, E. Technical Description of version 4.0 of the Community Land Model (CLM). (NCAR, 2010).
  • 46.Anderegg, L. D. Why can’t we predict traits from the environment? N. Phytol.237, 1998–2004 (2023). [DOI] [PubMed] [Google Scholar]
  • 47.Harris, I. P. D. J. et al. Updated high-resolution grids of monthly climatic observations-the CRU TS3. 10 Dataset. Int. J. Climatol.34, 623–642 (2014). [Google Scholar]
  • 48.Hengl, T. et al. SoilGrids250m: global gridded soil information based on machine learning. PLoS One12, e0169748 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 49.Tian, H. et al. History of anthropogenic Nitrogen inputs (HaNi) to the terrestrial biosphere: A 5-arcmin resolution annual dataset from 1860 to 2019. Earth Syst. Sci. Data14, 4551–4568 (2022). [Google Scholar]
  • 50.Cayuela, L., Granzow-de la Cerda, Í, Albuquerque, F. S. & Golicher, D. J. Taxonstand: An R package for species names standardisation in vegetation databases. Methods Ecol. Evol.3, 1078–1083 (2012). [Google Scholar]
  • 51.Jin, Y. & Qian, H. V. PhyloMaker2: an updated and enlarged R package that can generate very large phylogenies for vascular plants. Plant Divers44, 335–339 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 52.Revell, L. J. phytools: an R package for phylogenetic comparative biology (and other things). Methods Ecol. Evol.3, 217–223 (2012). [Google Scholar]
  • 53.Pagel, M. The maximum likelihood approach to reconstructing ancestral character states of discrete characters on phylogenies. Syst. Biol.48, 612–622 (1999). [Google Scholar]
  • 54.Ramírez-Barahona, S., Sauquet, H. & Magallon, S. The delayed and geographically heterogeneous diversification of flowering plant families. Nat. Ecol. Evol.4, 1232–1238 (2020). [DOI] [PubMed] [Google Scholar]
  • 55.Soudzilovskaia, N. A. et al. FungalRoot: global online database of plant mycorrhizal associations. N. Phytol.227, 955–966 (2020). [DOI] [PubMed] [Google Scholar]
  • 56.Iversen, C. M. et al. Fine-Root Ecology Database (FRED): a global collection of root trait data with coincident site, vegetation, edaphic, and climatic data, Version 3. Oak Ridge National Laboratory, TES SFA, (U.S. Department of Energy, 2021).
  • 57.Breiman, L. Random forests. Mach. Learn.45, 5–32 (2001). [Google Scholar]
  • 58.Viechtbauer, W. Conducting meta-analyses in R with the metafor package. J. Stat. Soft.36, 1–48 (2010). [Google Scholar]
  • 59.E. C. Global database of leaf nitrogen fractions. Zenodo. 10.5281/zenodo.21624103 (2026). [DOI]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplementary Information (2MB, docx)
Reporting Summary (75.3KB, pdf)
Transparent Peer Review file (1.6MB, pdf)
Source Data (455.8KB, xlsx)

Data Availability Statement

The processed data underlying all figures and tables in this manuscript are provided as a Source Data file accompanying the paper. The raw datasets generated in this study, including leaf nitrogen fractions for 520 seed plant species spanning 98 families, have been deposited in Zenodo (https://doi.org/10.5281/zenodo.21624103)59. A complete list of paper source references is provided in the Supplementary Information. Source Data is provided as a Source data file. Source data are provided with this paper.

The codes used in this study are available are available via Zenodo at https://doi.org/10.5281/zenodo.2162410359.

Articles from Nature Communications are provided here courtesy of Nature Publishing Group

RESOURCES