Mending plates, bent to fit and screwed to the frame's back, but not to the stretcher. Still a method recommended by CCI.
I think in all of the places I have worked, (general retail to conservation), and all of the museum
back rooms and workshops I have been graced to work in.... I have seen brass mending plates
and screws used about four times. Nails through the stretchers into the frame..... all of the time.
[remember - commercially viable screws aren't common prier to the 1850s. But square nails
pre-date art in a frame by a few centuries.]
I have been consolidating my files and couldn't find the side image of the large Titan painting
in the North Upper floor of the Louvre, where the square nail is clearly securing the stretcher bar,
through the canvas into the frame. (Which wasn't why I was taking the photo... it was for the
sliding dove-tails that secured the miter that after a few centuries still wasn't cracking open.)
More commonly, I have seen this treatment, where the wall of the frame was built-up.
One would think they would also build up around the rabbet.... and nail sideways like a framing
point... but, sadly, I have seen the build-up to block view of the working behind the frame,
and then nails driven through stretchers.
The dun colored build-up is to make a "prettier" side view. Please also note that the painting/frame
is utilizing no bumpers or wire to let air behind.....
This Bearstat is in room 241 of the Louvre.
And if I remember right, the frame on it's left - - is hung only on a single hook/nail.
This has always been the war between the Ivory Tower, and Reality.
In my other place that lost my respect; Italy; if the Uffizi can't care enough about the treasures
to close the windows against the rotting putride methane gasses rising from the sewage in the
river, why would they worry about driving a nail through the canvas?