ARTICLES DOI: 10.1038/s41559-017-0308-2 © 2017 Macmillan Publishers Limited, part of Springer Nature. All rights reserved. 1 Department of Aquatic Ecology & Center for Ecology, Evolution and Biogeochemistry, Eawag, 6047 Kastanienbaum, Switzerland. 2 Department of Fish Ecology and Evolution & Center for Ecology, Evolution and Biogeochemistry, Eawag, 6047 Kastanienbaum, Switzerland. Present Address: 3 School of Earth Sciences and Environmental Sustainability, Northern Arizona University, Flagstaff, AZ 86011 USA. 4 Institute of Ecology and Evolution, University of Bern, 3012 Bern, Switzerland. Present address: 5 Department of Biological Science, Florida State University, Tallahassee, FL 32306, USA. *e-mail: rebecca.best@nau.edu T here is mounting evidence that phenotypic differences can both shape and be shaped by dynamic interactions between organisms and their environment 1–5 . Previous studies suggest that recent evolutionary divergence can impact the same ecological conditions (for example, resource abundance and composition 6–9 ) that affect selection on heritable phenotypes 10–12 . Work across a broad range of systems shows resulting feedbacks between eco- logical and evolutionary processes when they occur on similar timescales 13,14 . However, moving from documenting the existence of such dynamics to understanding what controls their magnitude and direction across a variety of complex systems has only just begun and remains a major challenge at the intersection of ecology and evolution. Interest in quantifying the nature and impact of eco-evolutionary dynamics has fuelled the development of theory linking ecological processes to trait evolution 13,15 , as well as the exploration of increas- ing complexity in those dynamics observed in nature. Theoretical investigations of eco-evolutionary dynamics have demonstrated that they can either promote or constrain phenotypic divergence and that this outcome depends on the nature of resource use and supply 16 , the complexity of species interactions 13 , the temporal over- lap in rates of ecological and evolutionary processes 13 and the way that changing trait frequencies impact population density 15 . Despite this expanding body of theory, it remains challenging to formulate general predictions about the way eco-evolutionary dynamics alter evolutionary trajectories 17 . Empirical work in individual study systems mirrors the complex- ity of theoretical predictions, with eco-evolutionary feedbacks in nature depending on abiotic variation in time and space 18 , commu- nity complexity 19 , standing genetic variation 20 and density-depen- dent influence on resource availability 10,21 . The focus on population density in particular has served as an important bridge between theoretical and empirical work, with both viewing population density as the ecological link in a chain of evolutionary events 21,22 . This is the basis for a recent re-emphasis on soft selection 1 , whereby selection results from the interaction between population density and pheno- type frequencies, rather than from fixed links between phenotype and fitness. Work in the Trinidadian guppy system has demonstrated the importance of population density in eco-evolutionary dynamics, with predation intensity and density-mediated per-capita resource availability driving repeated life history divergence 1,10 . However, both theory and experiments have demonstrated that variation in predator density can impact not only total prey availability, but also prey composition 16,23 . While some increases in density may promote the simultaneous use of alternate resources and facilitate dietary differentiation among interacting competitors 24,25 , high population densities may also reduce the availability of alternative resources to a point where no dietary segregation can occur 23,26 . Thus, the eco- logical component of eco-evolutionary feedbacks is not simply the population ecology of the focal species (that is, its density), but also the community and ecosystem ecology of the whole system, which defines the composition and rate of supply of multiple available resources. One approach to translating the potential complexity of factors contributing to eco-evolutionary dynamics into general predictions is to use experiments that isolate the impacts of both phenotype and population density on the dynamics of multiple resources and test how this affects selection in subsequent generations. This approach can help pull apart mechanisms by which eco-evolutionary inter- actions might control the rate of evolutionary divergence, conver- gence or directional change. Here, we used a two-phase mesocosm experiment with adult and juvenile sticklebacks to investigate their ecosystem impacts and subsequent effects on selection dynamics. First, we independently manipulated the density and multivariate Transgenerational selection driven by divergent ecological impacts of hybridizing lineages Rebecca J. Best  1,2,3 *, Jaime M. Anaya-Rojas 1,2,4,5 , Miguel C. Leal 2 , Dominik W. Schmid  1 , Ole Seehausen 2,4 and Blake Matthews 1 Dynamic interactions between ecological conditions and the phenotypic composition of populations likely play an important role in evolution, but the direction and strength of these feedbacks remain difficult to characterize. We investigated these dynamics across two generations of threespine sticklebacks from two evolutionary lineages undergoing secondary contact and hybridization. Independently manipulating the density and lineage of adults in experimental mesocosms led to contrasting ecosystem conditions with strong effects on total survival in a subsequent generation of juveniles. Ecosystem modifications by adults also varied the strength of selection on competing hybrid and non-hybrid juveniles. This variation in selection indicated (1) a negative eco-evolutionary feedback driven by lineage-specific resource depletion and dependence and (2) a large perfor- mance advantage of hybrid juveniles in depleted environments. This work illustrates the importance of interactions between phenotype, population density and the environment in shaping selection and evolutionary trajectories, especially in the context of range expansion with secondary contact and hybridization. NATURE ECOLOGY & EVOLUTION | www.nature.com/natecolevol