So, from this discussion and one I found in the archive, I'mFrank, per Dominion railroad statictics only about 10% of all Canadian
carloadings were destined to points in the US. Some of those would have
been using mty US marked cars being sent home.
As the CN and CP would have owned enough cars to take car of Canada's
traffic in a given year, it's a reasonable probability then that over time
there would be an average of not more than 10% of their cars beeing seen
south of the border at any given time. Some days would have more, some
less, but over time it would average out to around 10%.
If you take 10% of Canada's boxcar fleet and compute those cars as a
percentage of the total US boxcar fleet it works out to around 1.4% (early
1950's data). Now, looking at wheel reports, the numbers of Canadian cars
is unusually low relative to the total CP / CN fleet -- see, if there were
no issues at the border whatsoever you'd see about 15% of all boxcars were
marked CP or CN as that's what percentage their total fleet would be in the
US... but the analyized wheel reports show the numbers close to the expected
1.4%. Now you'd probably find distortions regionally... such as near
Buffalo or Seattle for instance as compared to, say, some tiny burg south of
Phoenix Arizona or in central Georgia. But I am of the opinion that once
they cleared the border choke points, the cars probably dispersed fairly
So again, you need to first think of national averages -- 10% of ladings
going to the US means roughly 10% of boxcars seen in the US would be
Canadian... and then make whatever adjustments you deem correct to account
for seasonality, location, and possibly lading (somewhat lessor odds of
Canadian lumber being delivered to, say, a lumber center like Burns,
In the final analysis, I think you can justify that 1.4% w/ no difficulty.
But making it 15-20% will be either take a real stretch of imagination of a
fortuitous choice of locations.