I think by "Steam impact," the RJR is referring to dynamic augment. I have never seen this force taken into account for bridges any more than any other track, but I am not sufficiently well versed to judge, so I leave that to those who know more.
I do know that, during WW II, a Frisco spot-series 2-10-2 (#41, IIRC) kicked the rails out from under herself on a curve on a trestle bridge approach near Ft. Leonard Wood in MO. It took the track gangs quite a while and a LOT of trouble to jack the engine up and put the rails back while hanging off the side of the trestle many feet above the Big Piney River. The spot-series was limited to 35 mph because they were old drag-ear engines with insufficient balancing and had a bad habit of damaging track. Apparently, the engineer got in too much of a hurry on the bridge/curve.
