“…(One should note that this result, essential to our approach, was found also by Dikranjan, Giordano Bruno, and Milan [23] (4.11).) Suppose not only that some N ∈ Λ(G) has a proper, dense pseudocompact subgroup D, but also that D may be chosen so that r 0 (N/D) c. Then, there is a selection set X for the coset space G/N , constructed recursively with some care, such that the pseudocompact subgroup H := D ∪ X of G is proper, indeed it is proper for the good reason that r 0 (G/H ) c. This fulfills two purposes at once: It obviously gives a proper, dense, pseudocompact subgroup of G, thus responding to Question I.…”