I mentioned how SP was not a proponent of roller bearings, and here's one of the two steam locomotives that were so equipped: SP GS-5 #4458. The GS-5 class was composed of just two locomotives, the #4458 and the #4459. Identical to the GS-4s other than roller bearings on the tender trucks, lead and trailing trucks, drive axles, and side rods, these were considered the high water mark of SP 4-8-4s. The #4458 used Timken bearings, while the #4459 used SKF bearings, presumably to test both company's offerings, and when the pair of GS-5s were retired and scrapped in '58, the bearings were all examined and showed minimal wear.
So, if the roller bearings on the two engines were such a success, how come it didn't warrant widespread adoption? Part of it was timing; when the GS-5s were built in spring of '42. After that, getting roller bearings and lightweight materials became difficult and the War Production Board was cracking down on new designs or passenger-only locomotives, so when the GS-6s (and Western Pacific GS-64s) those were a rehash of the older GS-2/GS-3 designs with 73" drivers and solid bearings. After the war, and those materials became available again, SP was moving on to diesels, not purchasing any large, higher-speed mainline steam locomotives (The last of the 4-8-8-2 cab forwards were also built in '43).
The other issue was that Southern Pacific had developed their own system, the "Spring Pad Lubricator" which consisted of a wool carpet-like oil pad with wicks, resembling large unraveled carpet strands, dangling down into a bronze or aluminum oil sump. The lubricating pad is pressed against the bottom of the axle by an "X" frame and coil springs, and is also fed a stream of pressurized oil as well. Every conventional driver axle at the time used blocks of grease that had a concave surface that kept the axle journals lubricated over the full length of the journal bearing surface. However, the grease block had a steel distribution screen that allowed the full surface of the journal to get even grease application. As a driver journal began to warm-up , the screen 'metered' the grease flow, allowing the axle to return to its operating temp. However, grease blocks tended to wear-thin, at random rates. so after
every trip, when over an inspection pit, a loco could have one or more grease blocks that were in need of replacing. Replacing the thin grease blocks was a labor intensive job that could take 20 minutes per grease block to be replaced. SP was spending a lot of labor hours replacing grease blocks, since every 4-axle engine has 8 grease blocks, spread across a fleet of hundreds of locomotivess, with multiple axles and soon you're talking thousands of grease-blocks needing attention.
So, SP decided to solve the problem on its newer locos like cab-forwards and MT-1 Mountains and GS 4-8-4s. They switched to the oil-lubricated spring pad driver journal boxes. The driver journal boxes were fitted with round sight glasses so the inspectors could check the oil levels. Any that were low, were easily spotted and quickly re-filled, and if you've ever noticed an SP engine with white stars painted at the center of the drive wheels, that's to denote that they are equipped with spring pad journal boxes. However, not every SP loco was oil-cellar equipped, since there was no reason to be retrofitting old switch engines or 2-6-0s that were poking around on branch lines, and so manual labor and a substantial workforce was still required. Because of that large work force, adopting roller bearings for locos would not heavily reduce that labor force. No matter how many new locos they bought, if roller-equipped, they still had thousands of grease-blocks that needed inspecting. Also, the Southern Pacific didn't have the extended 100mph running of, say, UP or NYC, they were a 90mph railroad tops, and only for fairly short spurts, so the advantages of roller bearings weren't as pronounced.
Something that is interesting is that when Oakland requested a large steam locomotive for display, SP just grabbed the first GS off the yard track they were stored on, and that happened to be GS-4 #4449. Reportedly, right behind it was one of the two GS-5s, and someone said that Doyle McCormack has remarked that he wished they would have grabbed a GS-5 instead, since solid bearings have caused grief with hosting railroads. While still FRA legal, some Class Is like to refuse anything with solid bearings. Back in the day, solid bearings were fine, but as steam locomotives went away, a lot of the original manufacturers stopped making grease cakes, or environmental concerns have caused those that still make them to change the formula, and modern grease cakes tend to be troublesome (reminiscent of the flat-tappet camshaft issues of today). They have updated the #4449 with roller bearings on the lead and trailing trucks and the tender axles, since that's fairly easy to unobstrusive do, but haven't undertaken trying to retrofit the drive axles. And to install roller bearings on the rods requires a complete redesign of the rods and would be visually noticeable. Since the #4449 doesn't go out on Class I rails anymore (Or really any rails, since the Oregon Pacific Railroad is no longer allowing the #4449 to run on their tracks anymore), the friction bearing issue isn't an issue