Yes, Frank is correct. However, many journals are asking for forest plots (I suspect at the request of peer reviewers) even when there are fewer than 10 studies. It's not great science, but I think it has come to be viewed as a sort of "quality metric" of a meta-analysis (for better or worse). I completely agree with Frank, I just want to share my experience of being asked by a journal to provide funnel plots in less ideal situations.