f∘g(x) = f(g(x)), so the domain of f∘g can only contain elements which are elements of the domain of g, which is { 2,4 }. g(2) = 5, so f∘g(2) = f(g(2)) = f(5) = 4, so (2,4) ∈ f∘g g(4) = 3, so f∘g(4) = f(g(4)) = f(3) = 4, so (4,4) ∈ f∘g Therefore f∘g = { (2,4), (4,4) } Edwin