The mapping u↦u∗ of L(E;F) into L(F;E) defines by passing to the subspaces an isometric linear mapping of L1(E;F) into L1(F;E).
Let u∈L1(E;F). Since u is a Hilbert–Schmidt mapping, the same is true of u∗ (EVT, V, p. 54). The mapping v↦v∗ is a bijection of the set of v∈L2(E;F) such that ∥v∥⩽1 onto the set of w∈L2(F;E) such that ∥w∥⩽1; for every v∈L2(E;F) with ∥v∥⩽1, one has ⟨v∣u⟩=⟨u∗∣v∗⟩ (EVT, V, p. 54, formula (42)), whence the result by Prop. 14.
Bourbaki sets the proof directly after the statement, and the corpus does not separate them, so what is above is the statement and its proof together.