Drew McCormack wrote: > Is this one legal? > > real, pointer :: a(:), b(:) > nullify( a ) > b => a > print *, 'Is b associated?', associated(b) Yes, it is (section 7.4.2.1 says that the pointer association status of b becomes that of a). Otherwise making linked-lists would be a big pain. There have been ideas to extend this also to cases where a is something like a nonpresent optional dummy argument... Aleks