As biologists, we seek to understand the sources of biodiversity, and the consequences of that diversity for ecological and evolutionary dynamics. But what do we mean by biodiversity? An evolutionary biologist might consider the traits of a population, while an ecologist may invoke differences among species. A key complexity of biological systems is that these facets of biodiversity occur simultaneously-all biological systems span levels of biological organization. In the first half of my thesis, I use gall-makers and dragonfly larvae to address a key facet of natural selection; species live, reproduce, and die in complex webs of interacting species, all of which can directly and indirectly shape evolutionary change. I show that subtle differences in the traits of 'keystone' individuals (Chapter 1), populations (Chapter 2), and indirectly interacting species (Chapter 3) all modify interactions and natural selection. These links can be further modified by environmental conditions (Chapter 4), and spatial processes that have random (Chapter 5) and deterministic (Chapter 6) effects on species composition and traits; traits, interactions, and selection are intimately linked. In the second half of my thesis, I investigate some of the ways in which a process occurring at one level of organization (e.g. evolution) can influence processes and their consequences occurring at other levels (e.g. community assembly and trophic cascades). Using dragonfly larvae, I show that ontogenetic differences (Chapter 7), individual differences (Chapter 8), and community-level differences generated by macroevolution (Chapter 9) all shape predation. In a unifying experiment, I determine that intraspecific variation may be particularly consequential for trophic dynamics and ecosystem functioning (Chapter 10). To understand classic ecological patterns (e.g. biodiversity-ecosystem functioning) requires an understanding of traits and their functions across biological scales of organization. The broad conclusions of this thesis are two-fold. First, trait variation across biological levels can reshape natural selection. Second, it will be difficult to understand patterns of diversity in ecological communities without considering trait variation at other biological levels. To understand biological systems, we require a synthesis of the causes and consequences of diversity across levels-we require a rapprochement of evolutionary biology and community ecology.