Thanks Robert, I noticed the problems with the arguments of F only later - I was using a specific form before and just used F for illustration on this post. Interestingly, if you use F(k,u,v) then in addition to u->mu [which is strange, i agree! the known=F should be the default option] it swaps the first two arguments, which is definitely not right! Anyway, your explanation makes sense - I don't know whether that would be a bug or not...

Hi David Thanks for the advice, but it's still not working - say I need sin(20/x) from 0..1 - I would need at least 1000 points if it were to be presented in print, but it currently looks awful. I assume there's no way to hack the ps interpreter? At the moment I am having to use maple 10 in windows running in parallels! Best Chris