The topic of the paper is the "realism-Instrumentalism" debate concerning the status of scientific theories. Popper's contributions to this debate are critically examined. In the first part his arguments against instrumentalism are considered; it is claimed that none strikes home against better versions of the doctrine (specifically those developed by duhem and poincare). In the second part, Various arguments against realism propounded by duhem and/or poincare (and much discussed by more recent philosophers) are evaluated. These are the arguments from the (...) use of idealisations in science, From the "underdetermination" of scientific theories, And (especially) from the existence of radical scientific revolutions. A maximally strong version of realism-After due allowance has been made to these arguments-Is stated and defended. This position is close to popper's own "conjectural realism" but involves dropping entirely the idea that science has developed "via" theories possessing increasing verisimilitude. (shrink)