This book defends the 'disambiguating project' (DP) with respect to the units of selection, the thesis that the expression 'units of selection' is polysemous, in that it 'refers to at least three non-co-extensional functional concepts: interactor, replicator/ reproducer/reconstitutor, and manifestor of adaptation/type-1 agent' (1). Lewontin (1970) inaugurated the 'recipe' approach to the units of selection issue, the idea that we can identify a small number of conditions, such as variation, differential fitness, and heritability, individually necessary and jointly sufficient for natural selection to take place, and for units of selection to be identified. Later, Hull and Dawkins argued that 'units of selection' could refer to either 'replicators'—entities that pass on their structure in replication—or 'interactors/vehicles'—entities whose interaction with the environment causes replication to be differential. Subsequently a third entity was added: 'manifestors of adaptation/type-1 agents', entities that are cohesive and exhibit accumulated engineering adaptations.