The "story" in railfan literature about bridge loading limitations outruling a 4-8-4 is one of those that might have a single source. It may be worth checking the Railway Mechanical Engineer (1935 October) and Railway Age (1935 August 17) articles, if they can be found.
A “light” 4-8-4, such as the NC&StL design, would probably have fitted within the total weight limitation that applied, but might have been short on adhesive weight against what B&M saw as desirable for the envisage duties. It is reasonable to assume that B&M wanted driving axle loadings at the maximum allowable, say 67 500 lb given that the R-1 had an adhesive weight of roundly 270 000 lb. The NC&StL 4-8-4 (derived from the USRA 4-8-2A) had an adhesive weight of around 220 000 lb.
If total weight was the limiting factor, then to a first approximation, the R-1 is probably where you’d arrive if you started at say the Rock Island R-67B 4-8-4 (which could accommodate 73/74 inch drivers even though built with 69 inch) and downsized the boiler enough to get it on a 4-8-2 chassis at the same adhesive weight. The Rock Island R-67B had a grate area of 88 ft² and a BMOD of 98 inches, whereas the B&M R-1 had 79 ft² and 96 inches.
Although that said, one may note that the MoPac 1927 4-8-2, which was probably a precursor to the R-67B, and whose lineage can be traced back to the USRA 4-8-2B, had 84 ft² and 98 inch BMOD. Probably though a modernized version of this would have been above the weight limit, whatever that was.
Perhaps the B&M could have considered a 69/70 inch drivered 2-8-4, such as the Van Sweringen (NKP) design, which as initially built was similar in weight to the R-1. It had 90 ft² grate area and 98 inch BMOD. Adhesive weight was probably less than the B&M desired, but later iterations of that 2-8-4 design had much higher adhesive weights, so building it to 270 000 lb on drivers should not have been a problem. But this kind of locomotive might have been of lower utility to the B&M. At a given level of track stress, and with comfortable riding, the R-1 could probably run quite a bit faster, this due to its higher diameter drivers and its four-wheel leading truck. It was evidently fast enough for some passenger services. Whilst the 2-8-4 might have been able to go as fast, its top allowable speed in regular service was probably determined by tracking and riding considerations deriving from its 2-wheel leading truck. And those limitations made it not worth going above 69/70 inch drivers. One may wonder if the Van Sweringen 2-8-4 had more power than it could fully use in typical service.
With a large 4-8-2 such as the B&M R-1, the wheel arrangement may have dictated a lesser firebox size than would be considered ideal, in turn resulting in specific combustion rates at the high end of the acceptable range. Evidently that was an acceptable engineering tradeoff considering the constraints and the intended duties.
The B&M, as a user of the much-hyped original Lima 2-8-4 design, was not alone in not further developing this wheel arrangement.
Of some of the others, the IC also turned to the 4-8-2. One could argue that this was in part an exercise in convenience, as it rebuilt some of its 2-10-2 fleet as 4-8-2s (with 88 ft² grate area). But it then perpetuated the 4-8-2 with new designs. In its 1930s rebuilding programme, whereas the 2-8-2s were central to it, its 2-8-4s presented a conundrum, with one unsuccessful rebuild to a 4-6-4 before updating in kind was settled upon.
B&A’s use of the original 2-8-4 appeared to have had no influence on parent NYC, who instead elected to progressively develop the 4-8-2 for fast freight.
MoPac moved from the 2-8-4 to the 4-8-4 for fast freight, and rebuilt its 2-8-4s as 4-8-4s. Interestingly, its 2-8-4s had had an 88 ft² grate area rather than the original 100 ft², suggesting that Lima had in the first place gone “over the top” with this parameter. (The 90 ft² grate area of the Van Sweringen 2-8-4 also points to this.)
Perhaps it could be said that the original 63-inch 2-8-4 probably had more power than it knew what to do with.
At a given adhesive weight, as intermediate between say a 63-inch driver heavy 2-8-2 and a 73/74-inch 4-8-4 one could have either a 2-8-4 or a 4-8-2. Both solutions were chosen, and it would be difficult to say that one or the other predominated. Rather individual railroads chose that which best fitted their specific requirements. The simplistic conventional wisdom might have been that the 2-8-4 must have been better because it had a four-wheel trailing truck, which in turn allowed a much bigger grate area and was also one of the key elements of the “super-power” type, to which august group the 4-8-2 was denied entry because it had only a two-wheel trailing truck. But reality was not quite like that. There was not always such a marked distinction. For example, consider that the original UP 4-6-6-4, generally considered to be in the superpower class, was in some ways an articulated derivative of the 4-12-2, and retained the same 108 ft² grate area. The extra trailing axle was probably required mostly to accommodate the extra weight. The DM&IR 2-8-8-4 was also superpower, but had the same firebox dimensions as the non-superpower WP 2-8-8-2 from which it was derived (and actually a smaller grate area, 125 ft² as a compared with a notional 145 ft²). The extra trailing axle was there to accommodate the weight growth, not a larger firebox. The NKP 4-6-4, as I understand it, was essentially a USRA 4-6-2A with a 4-wheel trailing truck to spread the weight, but with the same grate area,
An interesting feature of the B&M R-1 was that its trailing axle was 12 ft behind the hind driving axle, a bigger distance than was typical for heavy 4-8-2 locomotives. This may have been to get the trailing axle somewhat out of the zone where bending moment augmentation occurred. The high railhead stresses caused by heavily loaded trailing axles of x-y-2 locomotives had been noted in the 1920 AREA proceedings. Trailer axle loading for the R-1 was probably about the same as that for the driving axles. With four-wheel trailing trucks, the typical axle spacings were probably more likely to be in the range that produced bending moment relief. A four-wheel trailing truck could probably have been accommodated under the R-1 firebox, but that would have increased the weight noticeably, and might have required, for example, a slightly smaller boiler in order to get back down to the target maximum weight. Evidently the chosen trade-off point was a two-wheel trailer placed as far back as reasonably possible.