Let ( α ) (\alpha ) be a property of separated locally convex spaces. Call a locally convex space E [ J ] E[\mathcal {J}] an ( α ¯ ) (\bar \alpha ) -space if J \mathcal {J} is the final topology defined by { u i : E i [ J i ] → E } i ∈ I {\{ {u_i}:{E_i}[{\mathcal {J}_i}] \to E\} _{i \in I}} , where each E i [ J i ] {E_i}[{\mathcal {J}_i}] is an ( α ) (\alpha ) -space. Then, for each locally convex space E [ J ] E[\mathcal {J}] , there is a weakest ( α ¯ ) (\bar \alpha ) -topology on E stronger that J \mathcal {J} , denoted J α ¯ {\mathcal {J}^{\bar \alpha }} . Kōmura’s closed-graph theorem states that the following statements about a locally convex space E [ J ] E[\mathcal {J}] are equivalent: (1) For every ( α ) (\alpha ) -space F and every closed linear map u : F → E [ J ] u: F \to E[\mathcal {J}] , u is continuous. (2) For every separated locally convex topology J 0 {\mathcal {J}_0} on E, weaker than J \mathcal {J} , we have J ⊂ J 0 α ¯ \mathcal {J} \subset \mathcal {J}_0^{\bar \alpha } . Much of this paper is devoted to amplifying Kōmura’s theorem in special cases, some well-known, others not. An entire class of special cases, generalizing Adasch’s theory of infra-(s) spaces, is established by considering a certain class of functors, defined on the category of locally convex spaces, each functor yielding various notions of “completeness” in the dual space.