Sunday, May 30, 2021

Freight car handling and distribution: a response

 In recent posts about car fleet proportions, I described my understanding of the ways in which freight cars usually moved about the country, and gave some specific examples. Part of this description was a summary of the Gilbert-Nelson idea (of which, more below). You can find the newest post here: .

A very interesting comment on that post was provided by Dan Smith and is appended to that post. Because I wanted to comment in more detail than seemed appropriate in the “comments” segment, I am doing so here. To begin, here is Dan’s comment (I have taken the liberty of dividing it into paragraphs):

“On thinking about the proportion of foreign and small road cars, it occurs to me that the assumption that foreign road cars were free-running, especially in the 1950s, is probably misleading. In my experience railroads typically divided other railroads into three categories: connections (e.g. UP, to SP), competitors (e.g. ATSF, to SP), and Neutrals (e.g. PRR or NYC to SP). 

“Railroads would generally see many inbound cars from connections, and would be somewhat likely to spot those cars for return loading. Railroads would see few inbound cars from competitors, and would return them empty rather than willingly spotting them for return loading. You might see ATSF reefers in peak ag season when SP was short, but SP would actively avoid providing ATSF cars to on-line customers. (It is true that some customers would specifically make troublesome requests as bargaining leverage.) Pooled cars would be an exception, but it would also be rare for a railroad to enter into pooling with a competitor. 

“Neutrals would show up depending on where on-line customers bought their goods, and would be returned empty. If Jupiter Pumps had a supplier on the DSSA in Duluth, you might see DSSA boxcars regularly. These practices changed in later years as multi-carrier pooling for major customers, such as GM and Ford, became more common. But free-running cars have always been the exception. That exception was why the incentive per-diem boxcar program led to colorful short-line boxcars all over the country in the 1980s.That's also why TTX stenciled "Next load any road" on Railbox cars - it was the exception.
“So bring that DSSA boxcar onto the layout once a month with a load of castings for Jupiter, but be sure to send it back empty!”

Let  me begin by saying that Dan omits one very, very important aspect of freight car supply and handling: the state of the economy. When the economy is slack, and empty car supply accordingly plentiful, every road has surplus empties and, as Dan says, is happy to send them homeward. But when the economy is booming, cars soon are in short supply, and now the opposite case occurs: roads are scrambling for enough empties to serve their shippers, and will use anything available, even in violation of the Car Service Rules, to get that done.

Beyond that, Dan is quite right that railroads were fully aware of their close competitors (like Santa Fe versus SP, as Dan mentions), and would strive never to spot one of the competitor’s cars for loading. The same relationship applied between PFE and Santa Fe’s SFRD. 

There were also connecting roads that might be “friendly” connections, such as Northern Pacific in Portland, for the SP. These might well get a little better treatment in interchange. Dan seems to think that most arriving loads would be in cars owned by these connections, and I have no idea what that is based on. I have never seen any data that point in that direction, and perhaps Dan can direct me to examples. 

My only thought is that Dan is thinking of later years when Special Car Order 90 (SCO 90) began to dominate movement of empty cars, and more and more railroads added themselves to SCO 90 to obtain empties directly homeward. But that’s a later era than I was describing.

But now let’s look at what Dan calls “neutrals,” essentially all other railroads in the country. Dan thinks that the cars of neutrals were not free runners. I know of no basis for this in the 1950s, which was the period I emphasized. In fact, several authorities, including Eugene W. Coughlin (a manager in AAR’s Car Service Division), in his book, Freight Car Distribution and Car Handling in the United States (AAR, 1956), specifically identify the entire national fleet as free-running. This of course does not apply to specially-equipped or assigned-service cars, but certainly to most box cars, gondolas and flat cars. 

Moreover, Dan states that the DSS&A would use its own box car to ship to California, and, he adds, would get it back empty. This violates the Car Service Rules in both directions, and is exactly the reason those rules were set up. Not every car movement was in accord with those Rules, to be sure, but in the early 1950s, over two-thirds of all car movements were in accord. 

In fact, a quite likely car in which that shipper on the DSS&A would load for California would be an SP box car, which is what the Car Service Rules would direct.

Lastly, let’s again remember Gilbert-Nelson. This interesting idea states that at least on main lines of most railroads, the “free-running” freight cars, particularly box cars and gondolas, would move around the country somewhat randomly, as needed by shippers. That in turn suggests that the frequency of observation of any particular railroad’s freight cars, anywhere in the nation, would be in proportion to the size of that railroad’s total freight car fleet, relative to the national fleet.

A number of pieces of actual train consist data from the 1950s supported Gilbert-Nelson, so many of us interested in car distribution accept it as at least broadly true for that era. But if we accept Dan’s description of car handling, it could not be true. “Neutral” empties would always be getting sent home, instead of being loaded, and the further away a particular neutral might be located, the less frequently its cars would show up.

Dan is perhaps thinking primarily about railroading after the 1950s, when, as he says, more and more specially-equipped cars were put in service, and the “general purpose” box car or gondola became a smaller and smaller proportion of the national fleet. Certainly he is right that the IPD box cars, and the Railbox fleet, were aimed at countering those trends; but now we are talking about an era twenty or more years later than what I was discussing.

I will repeat what I said in response to Dan’s comment on my previous post: it is an interesting and though-provoking set of remarks, and I have enjoyed thinking through the topics that Dan raised. Moreover, as stated above, I agree with some of it. But there are other parts that I think are simply wrong, in the 1950s era that I discussed, and I’ve explained why, above. But I still have to thank Dan, for the stimulus and for taking the time to think and comment.

Tony Thompson


  1. Tony,

    Dan does present some thought provoking assertions and I agree with his categorization of railroads by business relationship and distance. But I absolutely disagree with his view of the traffic impact of those relationships. I've sent you an email from Keith Jordan with copies of several Pacific Coast Demmurage Bureau car request forms from the seventies for the ATSF at Stockton. On them, note the preferred use of home road cars for on line hauls but also note the use of an SP car to an SP termination that were spotted for loading by the Santa Fe.

    1. Thanks, John, these are interesting. Would like to find comparable documents a couple of decades earlier -- as I'm sure you would also.
      Tony Thompson

  2. Tony - Once again, a post that makes me think! This subset of car distribution is important to my future ops (and one I hope I can get my guys to follow) as the one modeled end of my layout has 9 interchange connections radiating from it. I'd like to try to "mind the rules" (the layout is set in 1951, so this is a timely discussion), especially since it is a terminal operation with about two dozen industries to service. Hopefully, my crews wont be so overwhelmed by activity (I'd say at least 3/4 of the cars operating thru this part of the layout will be overheads), and they'll be able to play the role of the car service department as well as engineer/conductor.


    1. Ralph, that does sound like an intriguing challenge for your crews, and really a perfect opportunity to implement some of these distribution ideas. I hope you'll pass on how it works.
      Tony Thompson

  3. Tony, the one point that confuses me here is where you state "the frequency of observation of any particular railroad’s freight cars, anywhere in the nation, would be in proportion to the size of that railroad’s total freight car fleet, relative to the national fleet."

    How does the AAR's "Home Districts" routing come into play here?

    I had taken away from previous posts of yours that the Home District routing would tend to favour a higher proportion of cars from nearby districts - or at least that the farther flung the district, the less likely to see a car from that region.

    So that is to say, the odds of a PRR car making it all the way to district 1 or 2 (west coast) without being routed homewards would lower the likelihood or frequency of PRR cars showing up.

    1. Good question, Brian. But I don't really see how the Home District rule affects Gilbert-Nelson. A PRR car, say, might end up anywhere in the country, for example on an outbound load from PRR rails, and then be reloaded to move on anywhere at all (not always homeward). In fact, the underlying hypothesis in Gilbert-Nelson is exactly that.

      Whether cars of the most distant roads from a particular place would be less prevalent might be hard to study, but the guideline we have, Gilbert-Nelson and its supporting data, says no.
      Tony Thompson