In his recent 'State of the Art' survey 'Unnatural Attitudes: Realist and Instrumentalist Attitudes to Science' (Mind, I986, pp. I49-79), Arthur Fine claims to establish some metatheorems about the tenability of realist and antirealist attitudes to science. In this note, I shall dispute his first metatheorem, and its purported proof. Notoriously, some recent realists have presented an 'explanationist' defence of realism about scientific theories and the entities referred to therein. The explanandum is the instrumental success of science; by applying these theories, we are able to perform laser surgery, improve crop yields, and so forth. How could this be, unless the theories are, by and large, true, and unless the entities they refer to actually exist? Fine's first metatheorem attacks this defence: