(over) stimulated by the discussion on another thread about Flyer curved track, I rummaged around in my notes and miraculously found some measurements I had made a few years ago of a circle of Gilbert flyer track (the only track I have is the original flyer track). I then supplemented these with additional measurements shown below. I figured if I posted the results here I'd at least be able to find them again via the search function...!
My original measurements were done in millimeters, so I'll give all these in mm first and then convert.
The diameter across the circle, from the inside of the inside rail was 950 mm - that's probably good to +/- 4 mm. I'm sure I "squared" the circle, so I got similar measurements across several directions, but my notes only report one value. I should have made and recorded several measurements, then averaged, but it appears I didn't. Today I measured D, the distance across the rails from outside to outside, 27.1 mm, and T, the length of the tie, 47.4 mm. With all that, here's what you get.
R = 475 mm = 18.7 in. ~ 18 11/16 in (to the nearest 1/16 in.). So the diameter is 950 mm = 37.4 in ~37 3/8 in.
D above is 27.1 mm, so the radius to the track centerline is 475 + 13.55 = 488.6 mm = 19.2 in ~ 19 1/4 in and the radius to the outside of the outer rail is 475 + 27.1 = 502.1 mm = 19.8 in ~ 19 3/4 in
T above is 47.4 mm, so the radius to the outer edge of the tie (not the roadbed) is 488.6 + 23.7 = 512.3 mm = 20.2 in ~ 20 3/16 in
Just for giggles, the radius to the inside edge of the tie is 488.6 - 47.4 = 441.2 = 17.4 in (so that's not how Lionel gets S36!)
BTW, for those of you who remember high school geometry (or can google 'chord to radius'), I used that method as another way to estimate the track radius (to the inner edge of the inside rail) and got ~477 mm...
In any event, there doesn't seem to be any dimension that is 'exactly' 20 in (or R20, as Lionel sez) and certainly nothing even close to S36 (or R18). Still, Tom (AMFLYER) on the other thread got it right, that the values for Gilbert's original 20 in radius, 40 in diameter curve is approximately the dimension of the table top that would (just) hold a circle of track.