Jeff Aley notes:

"It was the same WestVaCo, eventually sold to Food Machinery Corp (FMC)."

There were, however, two [ at least ] Westvaco companies. As Jeff says, FMC [ a Missouri Corp ] acquired Westvaco Chemical Corp. (chlorine and caustic soda used to produce organic insecticides and pesticides) in 1946. This Westvaco apparently leased the SHPX covered hoppers. To confuse the situation somewhat, another Westvaco (originally the Piedmont Pulp and Paper Company and then The West Virginia Pulp and Paper Company), and then Westvaco operated an extensive paper [ boxes, etc. ] oriented business and still does as MeadWestvaco. As far as I am aware, this Westvaco leased no covered hoppers from SHPX. One has to wonder how many lawyers each Westvaco kept on staff.

"With regard to the overall fleet composition, one must be careful when analyzing data for the UP (and, I presume, other RR’s as well). Not all trains looked the same. In fact, I believe many professional railroaders could often identify a train by its consist, because the car types were similar on a day-to-day basis."

In fact, even some professional model railroaders can identify a train by its consist which is useful since Fraley [ at least ] did not identify a single train in his conductor's book...although he didn't hesitate to name the engineer and brakeman but not the fireman [ odd ]. Fortunately, he did list the locomotive.

"The UP obviously operated “Fruit Trains” that were dominated by reefers (mostly, but not exclusively PFE)."

Yes, but I prefer the term "reefer train" because, as Fraley shows, by far the majority of reefers were carrying spuds. Of course, commodities like fruit and spuds were seasonal.

"In addition, we can learn from Mark Amfahr’s excellent articles in The Streamliner that they operated merchandise trains (almost entirely loaded box cars for the freight houses), lumber extras (many double-door box cars), and drag freights (mostly empties, plus low-value loads such as soda ash, sand, gravel, and coal).
Given that the trains were different, and given that we usually only have data for a SAMPLE of the trains, a statistician can see that our data might be highly biased."

For sure. Using the criteria that if about a half of a train carried a specific item [ like spuds or coal or lumber ], here's how 34 trains might be labeled in March/April '49:

1. Lumber-9
2. Spuds/apples-4
3. Cattle/sheep- 3
4. MT [ PFE ]-3
5.MT coal-3
6.Gas-1
7.Ore-1
8.MT-tank-2

That is 26 of the 34 trains in the book. The other 8 might be described as carrying "stuff".

"Oh, and it gets worse. The UP would also re-arrange the trains based on tonnage. So they might combine all of a fruit train with the first quarter of a merchandise train to fill out the tonnage, and then send it out. This has the effect of convoluting the data even further, but doesn’t mix the cars up “enough” to make each train “random”.

As Jack Nicklaus said, "Nobody said it had to be fair".

"This is not to say that we should ignore data completely and just do whatever we like. I’m am simply saying (as others have before) that we must be cautious when we draw conclusions from the data."

Particularly since we have so little of it.

Mike Brock