Oikos

Investigating the factors that stabilize biological communities is a central topic in ecology. Response diversity, defined as variation in species responses to environmental change, has been proposed as a key mechanism underlying the biodiversity-ecosystem functional stability (BEFS) relationship, whereby greater species diversity enhances ecological stability. Previous studies have shown that response diversity promotes ecological stability by generating asynchronous population fluctuations and the resulting compensatory dynamics. Although several metrics have been proposed to quantify response diversity, they do not explicitly consider the presence of insensitive species whose performance is unaffected by current environmental conditions. To examine how insensitive species influence response diversity, species persistence, and ecological stability, we conducted numerical simulations of a generalized Lotka-Volterra model under environmental forcing. We first confirmed that increasing variation among sensitive species increased the response diversity index and stabilized community dynamics. We then examined a scenario in which response diversity depended solely on the proportion of sensitive and insensitive species, assuming that all sensitive species responded identically to environmental change. Under this assumption, the response diversity index was maximized when sensitive and insensitive species occurred in equal proportions, whereas increasing the number of sensitive species monotonically destabilized community dynamics. Consequently, the relationship between response diversity and community stability depended on how response diversity was generated, such that higher response diversity could even be associated with lower community stability. These findings demonstrate that overlooking environmentally insensitive species can obscure the mechanisms linking response diversity and ecological stability. More broadly, our results reveal that response diversity comprises at least two distinct biological components--species sensitivity and response variation among sensitive species--that can have contrasting consequences for community stability. We therefore highlight the need to quantify sensitive species empirically and to develop response diversity metrics that distinguish these components. Author SummaryUnderstanding why some communities remain stable despite environmental change is a longstanding goal in ecology. Response diversity, which refers to differences in how species respond to environmental change, has been proposed as a key mechanism explaining why greater biodiversity (species richness) can promote ecological stability. Because species respond differently to changing environments, declines in some species can be compensated by increases in others, helping to stabilize community dynamics. However, previous studies have rarely considered species that are insensitive to current environmental changes. Using a mathematical model, we show that response diversity can arise from two distinct biological components--the number of sensitive species and variation in their responses--and that these components can have contrasting effects on ecological stability. When response diversity reflects variation among sensitive species, greater response diversity stabilizes community dynamics, as expected. In contrast, when response diversity changes only because of the proportions of sensitive and insensitive species, higher response diversity can be associated with lower community stability. Our findings highlight the importance of quantifying the number of sensitive species and developing response diversity metrics that distinguish species sensitivity from variation in responses among sensitive species.

As ecosystems face a growing number of threats, coextinctions (resultant extinctions following a primary extinction) are expected to proliferate. However, less is known about the conditions under which coextinctions could outpace primary extinctions. Because coextinctions often occur through lost species interactions, we posit that aspects of food web structure and complexity can help predict differences in vulnerability to coextinction across ecosystems. To test this, we leverage Bayesian network models to assess the extent to which variation in ecosystem vulnerability to coextinction varies with food web structure. We find that food webs with high maximum trophic level are most vulnerable to coextinction, and that maximum trophic level is a better predictor than other aspects of food web structure, such as species richness or trophic connectance. Extending this approach, we also find that maximum trophic level uncovers the relative vulnerability of ecosystem services to species coextinction across 12 empirical food webs.

Nutrient availability in many temperate ecosystems is shaped by soil properties and historical land use. Especially in otherwise nutrient-poor landscapes human-induced, local fertilisation can generate fine-scale mosaics of nutrient hotspots. Whether and how large herbivores respond to such heterogeneity remains poorly understood. We tested whether spatial variation in soil-derived nutrient availability structures habitat selection by large herbivores, using full-year GPS tracking data from 7 red deer (Cervus elaphus) in the Veluwe, the Netherlands. We used soil types as a proxy for nutrient availability and assigned nutrient scores based on soil pH, cation exchange capacity and soil structure. We then evaluated habitat selection across multiple components of space use: (i) home range size; (ii) use of relatively nutrient-rich parts within home ranges; (iii) selection among soil types; and (iv) selection of locally enriched former agricultural patches. Red deer used relatively nutrient-rich within their home range more than expected based on availability, including local patches enriched by former agricultural use. However, site selection did not consistently follow nutrient scores among soil types. These results show that nutrient availability does shape habitat selection, but primarily through fine-scale, localised nutrient enrichment rather than broad-scale variation in soil properties. Our findings demonstrate that nutrient-related foraging contributes to habitat selection in a large wild herbivore, while also revealing that this process is scale- and context-dependent. By repeatedly concentrating their foraging in nutrient-rich patches, large herbivores may contribute to nutrient redistribution across the landscape, with the potential to reinforce or modify existing spatial heterogeneity in resource availability and ecosystem functioning.

1. Species loss can erode food-web functioning not only through secondary extinctions, but also through biomass redistribution, weakened energy pathways, and threshold-like functional collapse. Common topology-, connectivity-, and extinction-based robustness metrics provide valuable summaries of structural disassembly and cascade risk, but they are not designed to quantify continuous biomass retention, collapse-associated species sets, and non-additive group-level effects within a single dynamic framework. 2. We develop a dynamic biomass-based framework for assessing food-web robustness under progressive species removal. The framework introduces Dynamic Area-based Robustness (DAR), which quantifies the weighted area between slow- and fast-collapse reference trajectories of total ecosystem biomass retention. Building on these trajectories, we operationally define the Minimal Vital Species Set (MVSS) as the smallest fast-collapse-prefix species set whose removal first drives biomass below a predefined functional-collapse threshold. We further propose Cluster Influence (CI), which compares the biomass effect of simultaneous group removal with the mean effect of removing the same species individually. 3. We evaluated the framework using 120 niche-model virtual food webs spanning controlled gradients of species richness and connectance, and further demonstrated its applicability on 16 empirical stream food webs. We compared DAR with AUC- and secondary-extinction-based robustness metrics and assessed the sensitivity of DAR, MVSS, and CI to key bioenergetic parameters and parameter uncertainty. 4. DAR captured biomass-based robustness patterns that were only partly aligned with structural and extinction-based metrics, indicating that dynamic functional degradation provides complementary information. In virtual food webs, MVSS subsets were strongly enriched in basal species or basal resource nodes, and smaller MVSS proportions were associated with stronger positive CI under fast-collapse trajectories. Together, DAR, MVSS, and CI provide a reproducible framework for linking food-web structure, biomass dynamics, collapse thresholds, and non-additive species-set effects, offering a practical tool for dynamic robustness assessment in theoretical and empirical food webs.

Although cities are increasingly recognized as ecological islands, a unified framework explaining their susceptibility to alien plant invasion remains lacking. Using the most recent and comprehensive global dataset of urban alien plants, we modelled alien richness, mimicking island biogeography theory (IBT). Across all models, neither city size nor geographic isolation independently explained alien richness. Instead, richness was consistently associated with their interaction, supporting the central IBT prediction. However, the strength of this interaction depends on how city size was quantified, with socio-economic dimensions exhibiting stronger positive interactions with geographic isolation than physical measures of city size. Introduction-hub identity further modified these relationships. North America was the only hub for which the interaction between city size and isolation was consistently weakened, indicating that donor regions of alien plants are not ecologically equivalent. Simulations of simultaneous increases in city size and isolation showed that larger, more connected cities generally accumulated more alien plants despite increasing geographic distance, but the magnitude and direction of these responses are hub dependent. Our findings inspire an extension of classical IBT to a mechanistic explanation for global variation in urban alien plant richness in this increasingly urbanized and globally connected world.

Climate change is disrupting life's seasonal rhythms, altering the timing of key phenophases and reshaping how communities assemble. Beyond shifting flowering times, climate change can modify the extent of floral overlap and the sequence in which species bloom, generating novel assemblages with uncertain consequences for plant-pollinator interactions. Here, we ask whether flowering order generates priority effects in plant-pollinator communities, much like germination order does in plant communities. We tested how flowering order influences bee foraging behaviour and plant reproductive success in two co-flowering species, Hypochaeris radicata and Campanula rotundifolia, using a greenhouse experiment in which we manipulated the sequence of floral availability while allowing bees to forage repeatedly. We quantified changes in visit frequency, interspecific switches, handling time, and seed production. Our findings reveal priority effects in bee foraging that were strong enough to affect plant fitness: both species received more visits when flowering earlier than their co-occurring counterpart, and seed production declined when species flowered later. Overall, our results show that flowering order is an underappreciated driver of plant-pollinator interactions, suggesting that climate-driven phenological shifts could alter priority-effect dynamics with broader implications for community assembly. Key questions remain: How will climate-driven phenological shifts rearrange flowering sequences, and how will these priority effects emerge in more diverse communities in the wild? Our controlled experiment reveals strong flowering-order effects, underscoring the need to evaluate how widespread and impactful such dynamics are under accelerating climate change.

Land-use change, such as the transformation of woody ecosystems into open pastures, acts as a strong ecological filter, favouring some species while excluding others according to differences in ecological niche breadth. Understanding how differences in niche breadth influence species responses under anthropogenic filters is crucial to anticipate their persistence or displacement. In this study, we quantified realized niche breadth in two sympatric ecosystem engineers, the Neotropical leaf-cutter ants Atta cephalotes and Atta laevigata, to test whether breadth differences are consistent with specialist and generalist ecological strategies. We characterized realized niche breadth across fine-scale environmental gradients by integrating hemispherical photography, microclimatic data, mound architecture, and edaphic profiles from 114 colonies across a regional transect in the Colombian Andes, alongside macroclimatic data from Copernicus. Principal Component Analysis (PCA) and PERMANOVA identified canopy openness and bushes- and tree-type vegetation density as the principal axes of interspecific niche partitioning. The observed differences in realized niche breadth were consistent with specialist and generalist ecological strategies. A. laevigata was predominantly associated with open-canopy areas, warmer micro- and macroclimatic conditions, and narrower edaphic dispersion. In contrast, A. cephalotes occupied a wider range of microhabitat conditions. This broader realized niche breadth is compatible with previous reports of A. cephalotes occurring in urban areas. Together, these findings suggest that niche breadth may influence how Neotropical leaf-cutter ants respond to habitat transformation, helping to understand the ecological consequences of land-use change.

Plants competing belowground may produce extra roots, fewer roots, or no detectable change compared to plants growing alone. This inconsistency is often attributed to plants altering their root allocation in response to diverse cues, including neighbor detection and resource depletion by neighbors, but isolating these cues experimentally is challenging. Here, we hypothesize that water depletion alone can generate the range of root allocation strategies reported in the literature. We present this hypothesis as a water-explicit optimization model of root allocation that predicts a non-monotonic response. The model identified a critical depletion rate at which allocation shifted from increasing to decreasing with depletion. We tested this prediction using artificially rooted pots that imposed controlled water depletion while excluding living neighbors and their cues. A continuous artificial depletion gradient revealed the predicted hump-shaped pattern. These results reframe root overproliferation and underproliferation as positions along a single depletion-response curve.

Frugivory is essential for the maintenance of tropical forests, influencing seed movement, spatial structure, and ecosystem functioning. Although frugivory interactions are well documented at local scales, we still know relatively little about how network structure varies across biogeographic regions, particularly in equatorial lowland forests. In this study, we compare frugivory networks in lowland forests in the Guiana Shield (South America), Malaysia (Southeast Asia), and Central Africa (Gabon and Cameroon). Because these sites occur at similar latitudes, they provide a useful setting to examine how evolutionary history and biogeographic context, rather than climate alone, might shape the structure of the network. We analyze patterns across regions, focusing on implications for frugivory interactions and, ultimately, seed dispersal. Guided by the Integrative Hypothesis of Specialization (IHS), we focus on the role of megafauna and other large-bodied frugivores, which are important consumers of large fruits and dispersers of large seeds, and have experienced different historical trajectories in each continent. We also considered whether smaller frugivores, such as rodents, might partially compensate for the loss of Megafauna in the Americas through seed handling and caching behavior. The pantropical comparison revealed structural and functional differences between the studied frugivory networks. In the Guianas, lacking megafauna, the frugivory network showed strong modularity with nestedness within modules, forming a marked compound topology. Alternatively, in Central Africa, with elephants and great apes, the network showed intermediate modularity and nestedness, reflecting a balance between local compartmentalization and regional integration. Finally, in Malaysia, the network was predominantly nested, with a few highly connected species, especially figs and flying foxes, that linked most of the partners. Across continents, frugivory networks tend toward a compound topology, but the relative influence of modular and nested components shifts with biogeography, evolution, and ecology.

Squid abundance has been reported to increase globally between 1970 and 2010. This increase has been hypothesized to result from two primary factors: the loss of top predators due to overfishing and rising ocean temperatures. The decline in apex predators may lead to the expansion of squid populations either through reduced predation pressure or diminished competition with juvenile predators. Concurrently, increased temperatures could enhance the somatic growth rates of squid, thereby accelerating their population growth. However, empirically disentangling the impacts of predator loss and temperature on squid biomass remains challenging, especially in a food-web context. In this study, we used a size- and trait-based model of upper trophic levels that resolves the ecosystem structure -- biomass and trophic interactions of fish and squid -- for varying depth, temperature, and secondary production, to investigate two hypotheses of the historical expansion of squid, i.e., the effects of predator depletion from fishing and rising temperatures on squid biomass. Our model reveals that intensified fishing of squid predators -- specifically large demersal fish in shelf systems and large pelagic fish in open oceans -- leads to a slight increase in squid biomass. Conversely, elevated temperatures are associated with a decline in squid biomass. This temperature-driven reduction in biomass is attributed to an increased metabolism of squids beyond the available food supply. If historic overfishing on large marine predators continues to be curtailed, we expect a corresponding reduction in global squid biomass and fisheries potential, which could be further exacerbated by rising temperatures.

Environmental disturbances are ubiquitous in microbial ecosystems, yet their effects on community composition and interactions remain poorly understood. Here, we combine analytical derivations and large-scale generalized Lotka-Volterra simulations to examine how recurrent disturbances, varying in intensity (C), frequency (1/T), regularity, and evenness, affect microbial communities. We find that context-dependent interactions (ICD)--representing the net effect of the entire community on each species--are robust to changes in disturbance intensity and frequency. Community composition is influenced by disturbance, but this effect is primarily captured by a single composite parameter T/log(C), reducing the two-dimensional disturbance space onto a one-dimensional axis. Species abundances and ICD remain largely independent of specific values of T and C except near critical transition regions (survival thresholds and no-interaction zones). Sensitivity analyses show that disturbance-sensitive cases occur in the vicinity of these boundaries. This robustness extends to when the disturbance is irregular (even up to 50% stochastic variation in T or C), with uneven disturbances, when the disturbance is species-specific, and holds at any time point throughout the inter-disturbance growth phase. The patterns are independent of species richness, growth rates, carrying capacities, and interaction strengths. Our findings highlight that microbial community organization is largely conserved under recurrent environmental stress, with extinction acting as the dominant route by which disturbances reshape communities. These findings indicate that disturbance is not a strong driver of community structure and that minor variations in experimental or environmental conditions or sampling time are not expected to substantially impact microbiome composition.

Long-lived plants facing anthropogenic disturbance often exhibit recruitment failure despite persistent reproductive adults, generating extinction debt masked by longevity. However, whether adult presence reliably reflects environmental suitability for recruitment remains unclear. Here, we examine ontogenetic niche differentiation and its consequences for recruitment in Jubaea chilensis, an endangered long-lived Mediterranean palm with an aging population. We assigned individuals within the largest known population to four ontogenetic stages and characterized their environmental niches using climatic, edaphic, topographic, and vegetation variables. We then applied spatial and multivariate analyses, including Random Forest models to evaluate environmental segregation and identify predictors of seedling establishment. Age classes occupied significantly different environmental niches, with the greatest differentiation between seedlings and reproductive adults. Saplings and adult differentiation reflected mainly topographic variables at landscape scale, whereas seedling establishment was primarily predicted by microhabitat conditions (vegetation cover heterogeneity, east-facing slope orientation, and soil texture). This pattern is consistent with niche reconfiguring throughout the life cycle, suggesting that adult occurrence and recruitment suitability respond to distinct environmental conditions. Over one-fifth of sampled individuals occupied high-suitability sites without recruitment, suggesting that ontogenetic niche shifts are associated with a spatial decoupling between adult persistence and recruitment, consistent with demographic collapse independent of habitat degradation. This failure is likely mediated by insufficient effective seed dispersal, as the sole disperser (Octodon degus) preys on most seeds before dispersal. Conservation strategies based solely on adult distribution may therefore overestimate effective habitat and underestimate extinction risk in long-lived species.

Abundance correlations cannot reveal ecological interactions without an assumption about the covariance of environmental noise. A natural biological expectation is that similar species respond similarly to environmental fluctuations, generating positive correlations. Yet the same species also tend to overlap more strongly in resource use and therefore compete more intensely, generating negative correlations. The simplest plausible benchmark is thus to take environmental-response correlations proportional to niche overlap. We show that, under this assumption and across a broad class of stochastic community models, the two effects cancel exactly: equal-time abundance correlations vanish, independently of interaction strength, heterogeneity, and system size. Away from this matched point, the observed correlations measure primarily the mismatch between shared environmental response and competition, rather than the interaction matrix itself. Correlations can recover information about niche overlap when competitive feedback is delayed relative to environmental forcing, but the inference then depends on a resource-response timescale that is generally not determined by the abundance time series alone. When stochasticity enters through the mechanism that generates similarity itself--for example, through fluctuating shared resources--nonzero correlations may persist, but they reflect yield-depletion mismatch rather than niche overlap. Abundance correlations therefore report how environmental variability reaches the community at least as much as they report who competes with whom.

Symbioses are widespread and highly successful but vulnerable to the stress sensitivity of either symbiont. In some symbioses, turnover of symbionts has been shown to confer resilience to stressors. While similar mechanisms have been proposed for lichen symbioses, direct evidence for rapid adaptive symbiont turnover has not been shown. We tested the photobiont community composition and physiological responses of the foliose lichen symbiosis Flavoparmelia caperata-Trebouxia to urbanization-induced stress in a transplant experiment. We found evidence for significant compositional change in the photobiont community along an urbanization gradient (measured as vegetation cover), reflecting a turnover in dominance of Trebouxia OTUs from A46 to I05 in more urbanized transplant sites. This change in symbiont composition is associated with a greater physiological tolerance for urbanization, consistent with the hypothesized adaptive role of photobiont turnover. These findings support rapid photobiont turnover as a potential adaptive response to environmental change in lichen symbioses.

Biodiversity frequently peaks in fluctuating micro-oxic environments such as marine oxygen minimum zone interfaces, rhizospheric aggregates, sediments, microbial mats, and gut mucus layers. Yet, classical ecological theories do not adequately explain why intermediate oxygen tensions repeatedly favor coexistence and diversification. Herein, we propose a murburn ecological formalism wherein oxygen acts not merely as a metabolic substrate but as a generator of dynamic redox heterogeneity through partial reduction and diffusible reactive species (DRS) and redox-intermediates formation. Integrating empirical observations from marine, gut, soil, and aquatic-interface ecosystems with a reaction-diffusion framework, we show that intermediate oxygen tensions naturally maximize radical-field heterogeneity and produce dynamically shifting fitness landscapes. Numerical simulations demonstrate spontaneous coexistence, biodiversity maxima within micro-oxic zones, localized diversification, and coexistence stabilization without externally imposed niche partitioning. Additional simulations suggest that aquatic macrofauna indirectly enhance biodiversity by restructuring oxygen gradients and generating ecosystem-scale diffusional redox architectures (ESDRA). The framework proposes that fluctuating redox interfaces function as potential ecological zones of elevated adaptive turnover across biological scales.

Elevational gradients provide a framework for understanding how environmental filtering reorganises communities and interactions, but plant-pollinator interactions along temperate forest elevational gradients remain overlooked. We studied early-spring understorey communities at four forest sites spanning the foothills towards the timberline (450-1,000 m a.s.l.) in the Krkono[s]e Mountains, Czechia. Across six transects per elevation, we quantified flowering plant species richness, floral resources and traits, and video-recorded flowers, yielding 4,003 pollinator visits. We analysed elevational patterns in species richness, community composition, floral traits, and quantitative network characteristics. Visitation frequency and flowering plant and pollinator species richness peaked at intermediate elevations. The contribution of dipteran relative to hymenopteran pollinators increased towards higher elevations, principally because of non-hoverfly flies, whereas individual bee groups showed no uniform response. Floral resources and traits showed no uniform elevational responses, although total nectar sugar availability peaked at the highest site because of the dominant Vaccinium myrtillus. Most notably, both network-level specialisation and mean species-level specialisation were generally greater at the two higher elevations, whereas nestedness was lower and other network characteristics showed no consistent patterns. These findings suggest that shifts in pollinator composition and dominant floral resources potentially shaped interactions along the gradient. The increasing specialisation with elevation contrasts with the generalisation often expected under reduced partner availability and indicates that forest networks may follow elevational patterns not predicted from open habitats. Despite limited site-level replication, this study provides, to our knowledge, the first community-wide characterisation of plant-pollinator interactions along a temperate forest elevational gradient and identifies patterns requiring evaluation across replicated gradients.

Dispersal differences between hosts and their symbionts can generate mismatched population structure, potentially destabilizing beneficial interactions across space. We tested this possibility in the carnivorous pitcher plant Darlingtonia californica and its obligate arthropod associates, the midge Metriocnemus edwardsi and the mite Sarraceniopus darlingtoniae, sampled across sites spanning the hosts patchy range in Oregon and northern California, USA. Comparing nuclear and chloroplast genomic data from D. californica with mitochondrial COI data from both arthropods, we tested how range position, landscape connectivity, and dispersal mode influence population genetic structure across this mutualistic metacommunity. Host plant populations supported the central-marginal hypothesis: nuclear diversity declined toward the range margins, and marginal populations showed greater nuclear genetic differentiation. Chloroplast variation was more weakly structured, most clearly separating the northern Oregon Coast populations and revealing cytonuclear discordance consistent with historical seed-mediated movement or chloroplast capture near the boundary between neighboring regions. Landscape connectivity estimated from an ecological niche model was also associated with genetic exchange. Circuit-theoretic current flow was positively related to effective migration inferred independently from plant genotypes. Further, landscape resistance explained variation in plant and mite differentiation beyond geographic distance alone. Both arthropods showed significant spatial congruence with the host plant but not with one another, a pattern inconsistent with co-dispersal and suggesting that each associate tracks the shared landscape according to its own dispersal biology. These results show that regional genetic concordance among obligate ecological partners can coexist with substantial differences in the processes governing their movement and local connectivity.

The shape of density-dependence governs species persistence, and ecosystem stability. Yet, whether per-capita growth declines sublinearily, or superlinearily with density remains hotly debated. Growth rates across the tree of life have been shown to decline sublinearly with density, whereas theory founded on resource competition predicts the opposite. Here, we resolve this discrepancy and show that sublinearity can readily emerge from geometric constraints on consumer interactions. By linking inter individual spacing, movement and interference rates, we derive two limiting-interference regimes, one of which the well-mixed limit recovers the form of classic Beddington DeAngelis interference response. We then developed an individual-based model from first principles which reproduces the derived sublinearity response, and further use empirical data from published consumer-resource experiments that also bears the signature of sublinear density-dependence. Further, embedding the interference mechanisms underlying the emergence of sublinear density-dependence in coexistence theory opens a new regime for species coexistence where classical theory fails to predict. Our framework indicates that non-consumptive interactions are not merely a correction to resource competition but might be a distinct axis along which diverse communities may potentially coexist.

Adaptive phenotypic plasticity is expected to evolve when environmental conditions change predictably over time. This has led to the hypothesis that ectotherms in environments with low temperature seasonality, such as the tropics, should evolve lower thermophysiological plasticity than those from more seasonal environments (the Climate Variability Plasticity Hypothesis, or CVHP). Yet, empirical support for the CVHP is incredibly low, creating a need to identify other factors that can help explain how thermal plasticity evolves. Here, we use numerical models to show that the evolution of constitutive thermal tolerance breadth greatly affects the evolutionary benefits of thermal plasticity. In particular, tolerance breadth interacts with within- and between-season temperature variation in ways that can confound expectations of the CVHP, including conditions in which organisms from less seasonal environments benefit 30 most from expressing plasticity. Our findings indicate that a more holistic view of the relationship between thermophysiology and environmental temperature is needed to explain the evolution of thermal plasticity across climatic gradients.

Cross-feeding mutualisms, in which partner species exchange essential metabolites, are ubiquitous in microbial communities. In spatially structured environments, motility can improve access to partner-produced resources but also impose metabolic costs and displace cells from nutrient-rich regions, so its net benefit depends on the spatial dynamics of the interaction. Here, we combine competition experiments in a cross-feeding mutualism between Escherichia coli and Salmonella enterica with a spatially explicit consumer-resource model to determine what drives selection on motility. Spatial structure imposes asymmetric selection between partners i.e. S. enterica benefits from motility regardless of partner motility, whereas selection on E. coli switches from favourable to unfavourable depending on whether its partner can move. Competition in well-mixed culture suggests that this reversal reflects the loss of a spatial benefit rather than an increased cost. Our model attributes the asymmetry to three interacting factors: the ratio of metabolite production to consumption which sets whether the cross-fed resource is scarce or abundant; the number of growth-limiting resources which determines whether an alternative gradient can rescue the benefit of motility; and partner motility and growth rate, which shape where metabolites are produced. When a metabolite is scarce, motile cells gain by dispersing into regions it has reached but not yet been depleted from. When it accumulates, this gradient is eroded, and the motility costs offset any benefit it provides. Selection on motility therefore depends on the metabolic structure of the interaction and the spatial behaviour of partners.