Syntactic methods of picture generation have become established
  during the last decade or two.  We use random context picture grammars 
(rcpg s) to
  generate pictures.  The productions of such a grammar are
  context-free, but their application is regulated---permitted or
  forbidden---by context randomly distributed in the developing
  picture.

  In previous work we investigated the most obvious subclasses of
  rcpgs, namely context free picture grammars (cfpgs), random
  permitting context picture grammars (rPcpgs) and random forbidding
  context picture grammars (rFcpgs) and developed characterisation
  theorems for them.

  We now introduce the notion of table-driven context-free picture grammars, 
and prove
  two characterisation lemmas for the corresponding picture sets.  We
  show that these grammars are more powerful than context-free picture grammars, but
  strictly weaker than random forbidding context picture grammars.

  Moreover, we give results which show how these lemmas are limited in
  their application to galleries generated using permitting context
  only.