Rivets on canvas doesn't make sense to me. Stitching sounds like the logical one for me.
I looked in my P-40 books.
When on the ground the elevator looks very smooth, so you might think it is metal.
But in flight I have E anf F pictures that clearly show the ribbing effect at the leading part of the elevator. Probably by airflow.
So, at least E anf F versions had canvas elevators.
And another interesting observation is that in flight probably the leading edges of canvas surfaces show 'bigger' ribbing then trailing edges.
Something we can think about to simulate while skinning
