As I found out the hard way, the Winchester 1894 rifle in .38/55 was chambered for the shorter 2.082-inch case. The longer 2.125-inch is only applicable to early single-shot rifles chambered for the .38-55 cartridge, being the original case length as designed by Marlin for its Ballard rifles about 1884. I bought some of the Starline 2.125-inch cases on the assumption that my 1905-vintage Model 1894 rifle would be chambered for the longer case in view of its age. Cartridges loaded in the longer case will not work though the Model 1894 rifle's feed mechanism. After I trimmed the longer cases to 2.082-inch, everything worked fine. Although I didn't do so, if one single-loaded and fired a cartridge having the longer case, it very well might raise chamber pressures to a dangerous level as "Colt Fanning" has suggested, even with a cast bullet.
I got good accuracy with a load of 45 grains of Swiss 1-1/2 powder behind a Buffalo Arms 380261 bullet cast of 1:20 alloy and a CCI BR2 primer. Like virtually all older .38/55 rifles, my Model 1894 has a groove diameter of .379-inch, so cast bullets of .380-inch were in order. I suspect, though, that 47 grains of either Olde Eynsford or Swiss FFFG (3F) powder in concert with either the Accurate Moulds' 38-245D or Dick Dastardly's .38/55 255-grain "Big Lube" bullet, as suggested by "Blackpowder Burn," would work better.
I'll just mention that 19 grains of AA 5744 powder behind either the Buffalo Arms bullet or Desperado Cowboy Bullets' 260-grain bullet worked very well and was not position sensitive. I did chronograph smokeless powder loads, and while I don't recall offhand what the exact velocities were, those of the preceding AA 5744 load were very consistent. Loads employing the often earlier recommended IMR4198 and IMR3031 powders were very position sensitive and gave wildly varying velocities depending upon how the powder charge was positioned in the case. But we're really not that interested in that mercurially behaving nitrocellulose stuff anyway, are we?