N&W 611 (and other big 4-8-4)

First, to record an interesting publication:

"Trains" (that is, Kalmbach Publishing Co) has put out a "special" in honour of the return to service of Norfolk & Western 611: title is "^11 in Steam," also identified as "Special No. 2, 2015," and "Trains Magazine Special Edition No. 15-2015." Magazine format (i.e. glossy but floppy paper as covers, some ads), but apparently they want to market it as a book after it comes off the news-stands: it has been assigned an ISBN,
978-1-62700-305-6. $6.99 in the U.S., $8.99 in Canada.

Contains, among other things, some old articles by long-time "Trains" editor David P. Morgan, reprinted from old issues of "Trains": one from 1954, describing N&W's continued operation of modern steam locomotives into what elsewhere in the U.S. was the diesel era. Also lots of pictures.

Second, to maybe start a conversation. "Trains" editor Jim Wrinn's preface (p.3) contains the sentence
"My friend and retired Smithsonian Curator of Transportation Bill Withuhn says the Class J locomotives reached and often exceeded 5,100 hp, topping their nearest rivals on the New York Central at 5,000 hp and Santa Fe at 4,600."

(i) Note that these are drawbar horsepower.
(ii) The comparison to the "runner up," New York Central's S-1 "Niagara," could do with a bit of context. A N&W J tipped the scales at 494,000 pounds (engine alone), a New York Central S-1a or S-1b at 471,000 pounds. A J had a boiler maximum diameter of 102 inches, and S-1 of 100 inches. A J had a grate area of 108 sq. ft., an S-1 or 100 sq. ft. So… The J weighs a bit less than 5% more, has a boiler cross-sectional area 4% more, a grate area 8% more… and power output 2% more than the Niagara. New York Central's chief mechanical engineer Paul Kiefer and his collaborators at Alco didn't really have anything to apologize about for having only the second most powerful 4-8-4!
(iii) The much lower rating of the Santa Fe's famous 4-8-4 is an eye-opener. These were locomotives of roughy the same size as the Norfolk and Western's, with 102 inch boilers an 108 sq. ft. grates. Why didn't they match the J in performance? An obvious thought is that Baldwin's designers (with Santa Fe looking over their shoulders) just weren't quite as skillful as Norfolk & Western's or the New York Central - Alco team, but I wonder if there is more to be said.
--- (a) The Santa Fe's is the earliest of the three designs, so maybe the later designers had learned something new?
--- (b) The Santa Fe's 4-8-4 boilers had much shorter combustion chambers than the other two. I believe Baldwin recommended short combustion chambers for oil-burning locomotives, but (heat transfer through firebox walls -- the combustion chamber being an extension of the firebox -- being more efficient than through tubes) maybe it just made for a less efficient boiler.

Allen Hazen wrote:[someone says] "Bill Withuhn says the Class J locomotives reached and often exceeded 5,100 hp"

Wonder what he actually said. How could he know they often did that?

Hmmm… I don't know how Withuhn would have known. I suppose that as part of his job at the Smithsonian he would have pored over N&W records (which wouldn't have said much about the horsepower achieved in service, since most trains didn't have dynamometer cars!) and perhaps interviewed old engineers?
---
I just discovered something useful. In the late 1940s (title page says June 1947, but the reverse title page gives a copyright date of 1948) there was a pamphlet (60-odd pages) titled "A Practical Evaluation of Railroad Motive Power": publisher "The Steam Locomotive Research Institute" (of New York -- anybody know off hand what builders and or railroads and or coal mining companies it was affiliated with?), author Paul W. Kiefer (chief mechanical officer of the New York Central system, in charge of designing the Hudsons, later Mohawks, and Niagaras), preface by Joseph B. Ennis (long-time Vice President of Engineering, involved in both steam and diesel design: he retired from Alco as Senior Vice President (i.e. second in command) in 1947).

This compares steam and diesel performance: it is the primary source for the oft-cited comparison of the S-1 Niagara with 2- and 3-unit E-7 diesels. (And also discusses electrics: there is a lovely artist's rendition of a proposed New York Central mainline electric -- think GG-1 running gear with carbody of a double-ended Alco-GE PA.)

I hadn't had a chance to look at it since the 1970s, when I had access to a library that had a copy. (But I was an impecunious graduate student in those days and didn't photocopy it when I had a chance!)

Last night, while thinking about the Norfolk & Western J and the NYCRR S-1 I tried Googling and (Oh frabjous day! Calloo, cal lay!)
it turns out that the whole booklet is available, as a downloadable pdf, on the WWWeb at
https://milwaukeeroadarchives.com/Steam ... er1948.pdf" onclick="window.open(this.href);return false;

Enjoy!
----
Unfortunately the booklet was published before test result were in on the poppet-valved S-2 Niagara, but road tests had been done on the S-1, and there is a graph showing power vs. speed. Eyeballing it, it looks as if the S-1 had a drawbar horsepower of over 5000at speeds between about 54mph and 67mph, and the text says that peak d.b.h.p. was 5050 at 63 mph. This with 275 pounds boiler pressure (the J operated at 300 psi): when tested at 290 psi, the Niagara was capable of well over 5100 hp.

(Item: for comparison, figures are given for the Pennsylvania Railroad's T-1: a hair (about 1%) lower than the Niagara in cylinder hp, but drawbar hp marked with a question mark. Which suggests to me that the PRR had tested their unit on the stationary test plant, but hadn't run a full series of road tests with a dynamometer car. Somewhere in the "Trains" special announced in the first post there is a quote from a Norfolk & Western man, expressing disdain for the PRR's reliance on the stationary plant rather than real-world road tests.)

Allen, thanks very much for posting that link. I was aware of the Kiefer book, but had not previously seen it. From an initial “speed read”, it is most interesting.

For the most part, Kiefer has quoted steam locomotive power outputs to two significant digits. The 5050 hp case may be seen as 2½ digit precision. This is, I think, in line with the likely precision of the test methods that were available at the time, and may even have been on the optimistic side. I suspect that an ASTM-type precision statement for these methods, covering repeatability and reproducibility, would not be all that flattering. So power output claims and counterclaims need to be looked at after sprinkling with “precision salt”, as it were.

I think too that one needs to bear in mind that the Stephensonian steam locomotive was amenable to tuning and tweaking, as demonstrated by Chapelon, Porta and Wardale amongst others. Thus by attention to detail in the front end and elsewhere, quite significant peak power increases could be obtained from the same underlying dimensions, although not always with accompanying significant increases in utility. I suppose it was something like tuning and tweaking passenger car engines; taken too far it could result in a vehicle that although more powerful, was much less useful for the routine daily tasks. Wardale’s “Red Devil” project in South Africa might have been a case like this. With appropriate attention to detail, one might find that most of the “big” American 4-8-4s, those with 100 ft² grates and 100 inch OD boilers, would end up in pretty much the same place, power-wise, unless they had egregious basic design faults. They may have started from different places, though. As I recall, Wardale described C&O 614 as having a “couldn’t be worse” front end.

An American example of the power output increase obtainable without significant change of major dimensions is available in the progression of the N&W Y-class 2-8-8-2, as shown in the attachment. Note though that at the low-speed end, the change was minimal.
Kiefer notes (on page 12) the importance of a high factor of adhesion, a spring equalization system that included series coil springs at the attachment points, the use of lateral motion devices, and lower initial truck resistance, all conspiring to increase the flexibility of the driving machinery and allow it to adjust to cumulative wear. This was very much the Alco philosophy, in particular developed for the UP 4-8-4s, and evidently carried over to the Niagara. British engineer and writer E.S. Cox, who described himself as a lifelong student of American locomotive practice, had made the connection between the UP FEF and NYC Niagara in one of his books. In fact the UP FEF-2/3 strikes me as having been a well-rounded machine that, like diesel locomotives, would have worked on a wide variety of roads, and so was not road-specific. It wasn’t excessively heavy, did not have oversized cylinders, did not rely on an optimistic factor of adhesion, and had running gear that was quite flexible both laterally and vertically. That Alco philosophy was also found in its WWII standard 4-8-4, particularly the D&H K-62 variant. In fact Farrell and Pearsall (in “The Northerns”) even suggested that the K-62 was an antecedent of the Niagara.

Some information on the NYC S-2 power output is provided in Brian Reed’s Loco Profile #20, excerpt attached.
The Santa Fe 4-8-4 does look to have been something of a “brute force” approach by Baldwin, but no doubt it did the intended job to ATSF’s satisfaction, even if its measured power output was lower than one might have expected.

On locomotive testing, a useful reference is chapter 6 of “The Concise Encyclopedia of World Railway Locomotives”. That chapter was written by Sam Ell of BR, and covers both stationary plant and road-testing, each of which has its useful place.

Cheers,

Pneudyne--
So, it's not just DIESEL history where I have much to learn from you! (Grin!)
Re: testing accuracy. I ***think*** road tests involved multiple runs, with instruments (pressure gauges?) installed on the front end of the locomotive to measure cylinder horsepower, and drawbar horsepower measured by the instrumented drawbar of the dynamometer car. The New York Central had the reputation of being more thorough in its testing than most railroads (it was the second largest U.S. railroad in economic terms, so could afford to put more into r&d than most). So: warning noted, but I'm inclined to think that the test figures are reliable enough for discussion. (Particularly since tests of other railroads' engines are even more dubious!)

Re: Locomotive profile #20. That particular fascicule (on North American 4-8-4) is one I own and have re-read repeatedly (but thanks for posting the Niagara pages here, since not everybody started haunting second-hand bookstores in the 1970s!). Note that the last sentence on the lefthand page says that NYC 6000's 290 psi tests were on the Selkirk(*) stationary test plant whereas the others were road tests. That explains why, in Kiefer's pamphlet, the drawbar horsepower for 290 hp is marked as estimated, whereas other db hp and cylinder hp are given without comment.
((*) No, no connection to the Canadian Pacific! New York Central's test plant was located at their Selkirk Yard, just outside Albany (so, conveniently, not too far from Alco's Schenectady plant: I'd guess that Alco's engineers were frequent visitors).)

Re your paragraph "I think too that one needs to bear in mind… …"couldn't be worse" front end." Very definitely. The Niagaras had a front end design that had been worked out by the New York Central through tests in the ?1930s?, called, I think, the "Selkirk front end." So probably about as good as could easily be achieved with a conventional, single blast pipe, configuration. Whether it could have been improved by using "Kylchap" exhaust nozzles or a Giesl ejector or… I don't know: I have never studied front end design seriously (but I have the impression that these devices gave their greatest benefit on locomotives operated at much lower firing rates than late mainline steam in North America. (I think I have read-- perhaps in the "Locomotive Profiles" booklet on the New York Central's Hudsons-- that the NYC did experiment with a European-designed multiple-nozzle blastpipe, and decided to stick with their simpler (cruder?) design: the multiple nozzle device cost too much to replace, and wore out from steam erosion too quickly under the conditions of mainline usage.) … The Union Pacific used double blast pipes on at least some of its 800-class
4-8-4: motive power departments at different U.S. railways seem to have had different views about what was best!

As for other detail improvements… I am pretty sure I have read (but don't remember where!) that Kiefer, in designing the Niagara, WAS influenced by Chapelon's work: I think Chapelon's emphasis on "internal streamlining" of steam passages and on adequate cross section for gas passage. So the Niagara's boiler, at least, might well have been more efficient (and more "European inspired") than those of other large U.S. 4-8-4. … As I said in my first post, the basic design of the Santa Fe's 80-inch drivered
4-8-4 (3765,3776 and 2900 classes) was at least a bit earlier than the N&W J, and several years earlier than the Niagara. So its boiler may not have been as refined.

Re: Delaware & Hudson K-62. Visually, this is perhaps the type that most resembles the Niagara, but it is significantly smaller, and internally very similar to Alco 4-8-4 built for the CRI&P and (I think) CMstP&P. How close they were in boiler internal anatomy and other mechanical designs I don't know. (The New York Central seems to have had strong views about counterbalancing: their later steam designs had less cross balancing, and so came closer to "perfect" balance on each side, than most. Or so I gather from reading a variety of sources and imagining that I can see something between the lines!)

Bibliography. "Trains" had an article on the Niagara design some years back: "The strange story of S class." I will try to find my copy and post author and date shortly. (One curious fact it reports: the combustion chamber on the production Niagaras was shorter than that on the prototype, S-1a 6000, and the boiler tubes correspondingly longer. So they were somewhat less powerful: the prototype was closer to right on the optimum balance between combustion chamber and tube length.) There was also, apparently, a series of three articles by Tom Gerbracht in "Central Headlight" (the magazine of the New York Central Historical and Technical Society) which I have not seen, but which are cited in Steinbrenner's book on Alco.

Promised reference:
Robert A. Le Massena, "The strange story of Mr. Kiefer's superlative S class 4-8-4's," in "Trains," March 1984, pp. 46-51.

As usual, Le Massena wondered why NY Central was so stupid. Everyone knows big combustion chambers are better-- why didn't the RR know it?

In the 4-8-4 Profile, Reed mentions the NY Central 4-8-4's acceleration tests, which are the most interesting test results of any American 4-8-4. They probably didn't measure the speed quite right, but the distance-vs-time results are impressive enough.

I should have re-read the RME article on the NYC Niagara, as reprinted in Train Shed Cyclopedia (TSC) #56, before the above post. This gives some clues as to the origins of the Niagara design.

I think that the TSC series has had wide circulation, but it is worth quoting from that article.

“The development of the Class S-la locomotive, known locally as the Niagara type, began with the intention of bringing out a new class of L type locomotives with some expansion of boiler capacity. As the development progressed, however, it became evident that to attain the full possibilities of increased boiler capacity a 4-8-4 type wheel arrangement was required. The Class S 4-8-4 locomotive is expected to develop not less than 6.000 i. hp. This compares with a maximum indicated horse-power of nearly 4,800 for the Class J-3 4-6-4 locomotive and 5,400 for the Class L-4 4-8-2 type.”

So the dual-service L-4 was the starting point, and not surprisingly, NYC found that it could not be made much bigger without any extra carrying axle. Once it was clear that a 4-8-4 was needed, one may wonder if NYC went straight to the 6000 ihp idea, essentially going for a 4-8-4 of the maximum size that the road could handle, or whether it considered any intermediate ideas. In that event, the D&H K-62, with NTYC refinements, could well have been considered or at least put on the table by Alco.

More from the article:

“The boiler has been increased both in diameter and length from that of the Class L-4 locomotive. The outside diameter of the third course is 100 in. and the top of the boiler is practically at the clearance limit, which precludes the employment of a steam dome. Grate area has been increased from 75.3 sq. ft. to 100 sq. ft. and the depth of the combustion chamber from 63 to 921/2 in. The length of the tubes has been reduced from 20 ft. 6 in. to 19 ft. The boiler has been designed for a maximum working pressure of 290 lb. which is required with the 79-in. driving wheels.”

With 75-inch driving wheels, the working boiler pressure was 275 lbf/in².

The driving wheel diameter choices warrant consideration. Basis its trial work with a pair of rebuilt L-2 class 4-8-2 locomotives, NYC had decided that 69 inch drivers were satisfactory for dual-service locomotives that were required to operate in passenger service at up to 80 mile/h. Thus the L-3a class 4-8-2 fleet was built that way. But one suspects that in practice NYC found that larger diameter drivers would be preferable, and so the L-4a fleet had 72-inch drivers. I don’t know this to be the case, but perhaps the 72-inch size was the largest that the basic design could be stretched to accommodate without major rework. I think that the L-4a order predate the advent of WPB controls, so WPB restraints on new designs would not seem to have applied.

Generally, driver sizes of 73, 74 and 75 inches seem to have been those most popular for dual-service 4-8-2s and 4-8-4s. So, if the 72 inches of the L-4a was better than the 69 inches of the L-3a, but still not seen as optimum, possibly NYC had chosen 75 inches as the best option for its new dual-service locomotive, which as said, was initially conceived as an enlarged L-4. Once the scope allowed by the 4-8-4 wheel arrangement was realized, NYC might have seen that it could also be more-or-less the 4-8-4 homologue of its J-3 4-6-4, with the same speed capability and with 33% greater haulage capacity, and that would have led to the option of using 79-inch drivers, as on the J-3.

From the Kiefer book, one may infer that he saw lateral flexibility of the running gear as a key desideratum, and so would not have been at all happy with the heroic measures that N&W used on its J class, namely very stiff lateral controls, to enable the use of 70-inch drivers.

Whatever the original intent, NYC did end up with what was, in overall terms but not in the detail, a 4-8-4 homologue to its J-3, and also a locomotive that in its key dimensions, was very close to the UP FEF-2/3. At around 6600 ihp, it was 10% above the nominal target of 6000 ihp, and just above the 6400 ihp that would have made it directly proportional to the 4800 ihp J-3.

The RME article closed with the following comments:

“Before locomotive No. 6000 was completed an order had already been placed for 25 additional S-1 locomotives. These locomotives, Class S-1 b, are now being delivered. Because of the inability to test the S-la locomotive before the others were ordered, the length of the combustion chamber in the boilers of the latter have been reduced from 92½ in. to 81¼. in., and the tubes and flues correspondingly lengthened. These locomotives have 79-in. driving wheels and the cylinder diameter has been increased from 25 to 25½ in. to accommodate a working pressure of 275 lb.

“A further change in the S-lb locomotive is the trailer which has two pairs of 41-in. wheels. This change from 36-in. front and 44-in, rear wheels has permitted a further increase in ash-pan volume to 98 cu. ft. and has permitted the design of a pan with better slopes.

“A twenty-seventh locomotive will be built to the same proportions as the 25 S-1b class. It will differ from them in that the Franklin poppet-valve system of steam distribution will be installed. This locomotive will be designated as Class S-2a and is to be subjected to comprehensive acceleration, capacity and performance tests for direct comparison with the S-1 class, which has large piston valves and Baker valve gear.”

As regards the S-1b, it looks as if NYC was taking a conservative line, avoiding any developments that, whilst likely to increase efficiency, might lead to higher maintenance costs in the future. So it settled for a slightly shorter combustion chamber, the more familiar (to it) 275 lbf/in² boiler pressure, and 79 inch drivers with slightly larger cylinders. The use of 79-inch drivers would have removed any doubts about its ability to match Hudson speeds on a sustained basis. NYC was not alone; UP had moved from 77-inch drivers on its FEF-1 (which diameter appears to trace back to the Harriman Pacifics) to the 80-inch size on its FEF-2, one assumes to remove any doubts about meeting the intended speed parameters.

Talking of the UP FEF-2/3, it might be noted that it had a 90-inch combustion chamber, so there was some pertinent Alco experience available at this length. Possibly fuel quality was a factor in the selection of combustion chamber length; as far as I know the UP used soft coal of around 11 800 BTU/lb; whereas I suspect that the NYC used rather better bituminous coal.

Steinbrenner, in his landmark Alco book, related a broadly similar story. He lightly hinted (page 234) at the possibility that when the S-1 was initially mapped out, the WPB restrictions on new designs still applied obtained, so that it might have been conceived, or I think perhaps more likely cloaked, as a development of Alco’s WWII standard 4-8-4 design.) The specification table for the NYC S-variants on page 235 included the D&H K-62 for comparison purposes. And the reason for the shorter combustion chamber in the S-1b is stated as unknown.

What does not escape notice is that the NYC L-4a had what looks to be quite a high power output, 5400 ihp, considering that it was a moderately-sized 4-8-2, with 75 ft² grate area and 94-inch BMOD. One wonders how this would have compared with other 4-8-2s. The biggest I think had 88 ft² grates, but with boilers that derived from USRA 2-10-2 practice, so might not have been as powerful as their dimensions suggested. Of the later 4-8-2 designs, perhaps the B&M R-1 would have been comparable to the L-4a. Aside from lowish boiler pressure and largish and odd-sized cylinders, the R-1 strikes me as being about where one might end up if attempting to design a 4-8-2 homologue to the Alco WWII standard 4-8-4, same adhesive weight but smaller boiler. An interesting case would have been the 1927 MoPac MT-73. It is not clear whether this was somewhat backward looking or whether it was as modern as say the NYC L-2. MoPac had previously developed a 73-inch driver derivative of the USRA light 4-8-2, and to some its extent its 1927 effort looks like a similar development of the USRA heavy 4-8-2, but with a larger boiler (98-inch BMOD) and with high enough boiler pressure (250 lbf/in²) to allow the use of the “light” sized cylinders. It could be though that the L-4a was better than the norm, and showed just would the NYC could achieve with appropriate attention to detail over the whole thermal cycle.

Cheers,

Thanks for that thoughtful essay. I'll think some more, consult my "library" (i'm VERY glad I was able to find a copy of the Locoshed Cyclopedia volume with the reprint of the Railway Mechanial Engineer article on the Niagara: as I recall, I last consulted it to confirm that the boilers of the Pittsburgh & Lake Erie Berkshires were in fact, as they look, shortened versions of the Niagara boiler, with the same gas cross section), and see if I have anything to add!

Re: "Because of the inability to test the S-la locomotive before the others were ordered, the length of the combustion chamber in the boilers of the latter have been reduced from 92½ in. to 81¼. in., and the tubes and flues correspondingly lengthened."
--My head shakes with sadness and dismay. But your guess that NYCentral management decided to play it safe even if theory suggested the longer combustion chamber would filed a more powerful and efficient locomotive sound plausible. (Theory. I have never read Chapelon's book on "La Locomotive à Vapeur," but I suspect he had, by the time the Niagara was designed, worked out equations from which he could have predicted that the longer chamber would be better.)

I have the Chapelon book, the 2000 English edition published by Camden. It’s a good read, perhaps a little rose-tinted, and I would temper it with other sources, such as Cox’ “World Steam in the Twentieth Century”.

Surprisingly, Chapelon says very little about combustion chambers, and they do not appear to have been an essential part of his lexicon. They were used to some extent in French practice, such as in the 232R/S/U and 242A1, and of course the 141R.

When discussing the Altoona test plant results for the Pennsy Q-2 (page 450), Chapelon made the following comments:

"If these results are related to those already obtained with the M1a and I1s classes it may be stated that the efficiencies of the three boilers, as a function of the number of heat units introduced into the firebox per square metre of grate area per hour, are essentially the same. The straight graph lines showing these efficiencies as a function of firing rate virtually coinciding

“The very large combustion chamber volume of the Q-2 locomotive in relation to grate area (V/G = 2.31, V = firebox volume, G = grate area) and that of the M1a (V/G = 2.07) have not resulted in any improvement in efficiency compared with the relatively small ratio (V/G = 1.58) for the I1s 2-10-0.

“Thus, the determining factor in the combustion efficiency for all these locomotives was the firing rate/m² grate/hour, and this was independent, within certain limits, of their firebox volume."

So perhaps there was some uncertainty about the benefits of extra-long combustion chambers on the one hand, but reasonably certainty that they would adversely affect maintenance costs on the other.

Chapelon was complimentary in respect of the NYC Niagara (page 336). He reported upon, but made no comment about the reduction in combustion chamber length. The power output figures he quoted seemed to differ from others reported. For the S1a, the numbers were 5900 ihp between 120 and 145 km/h, and drawbar power 4550 hp between 88 and 113 km/h. For the S1b production version, they were 6300 ihp between 128 and 161 km/h, and 4850 drawbar hp between 96 and 121 km/h.

He also seemed to like the UP FEF, particularly its running gear, and marvelled at the remarkably small tyre diameter differences he found on an example that he examined that had done 185 000 km since its last tyre turning. He quoted 5040 ihp at 130 km/h for the FEF-1, and noted that it burned fuel of only 6560 cal/kg. The book includes the best picture I have seen of the Alco lateral motion device, and a picture of the four UP FEF axles in-line, showing three of them with that device.

Then I guess the NYC folk were also looking at the bigger picture. The diesel locomotive was past adolescence and into young adulthood, old enough to vote, so as to speak, but still on quite a steep improvement curve. The steam locomotive was in its “grey hair” years, and pretty much on a plateau as far as non-incremental improvements were concerned, and very soon would be there just to make up the numbers whilst the diesel fleet was built up to the numbers required to do the whole job. If the production Niagara was short a couple of hundred or so ihp as compared to what might have been, it was not of great consequence overall in respect of running the railroad.

Cheers,

Re:
"The diesel locomotive was past adolescence and into young adulthood, old enough to vote, so as to speak"
Well, what is called "the first commercially successful" diesel locomotive in the U.S. (a 300 hp boxcar switcher, Central Railroad of New jersey 1000) was built in 1925, it would have reached voting age(*) about the time the first Niagaras we're built! (Grin!)
(*) For the young: voting age was 21, not 18, until about 1970.

Pneudyne--
Thank you for the quotations from Chapelon's book!
---
Since the comparison to and relationship with the New York Central's 4-8-2 "Mohawk"(*) designs has come up, I looked up a bit of their history…
The New York Central got its first 4-8-2 (class L-1) in 1916. Slightly smaller than a USRA heavy 4-8-2, the L-1 was apparently unremarkable, and some were scrapped before the last later 4-8-2 were built.

In 1925 (recall that the J-1 Hudson was introduced in 1927), they got a much improved version, the L-2. Ultimately they got 300 L-2 locomotives, which were their main fast freight engine. Given the timing,it makes sense that the Hudson would be similar in technology, but it was slightly smaller: about 5% lighter in engine weight. (Since weights varied a bit over the production of both classes, I won't try to be more precise than that-- typical L-2 were about 365,000 pounds, typical J-1 about 350,000.) The J-1 had a larger firebox than the L-2 (81 sq. ft. grate area versus 75.3 sq. ft.), but n other respects its boiler was a bit smaller (87 inch maximum diameter against 94). They had the same (20 feet 6 inches) tube length, but since the Hudson's boiler barrel was shorter, it eliminated the L-2's combustion chamber. (The J-3 Hudson of the late 1930s shortened the tubes and restored a combustion chamber.)

In the late 1930s the New York Central felt a need for more heavy passenger power, and investigated the possibility of using Mohawks on passenger trains. Two L-2 were modified (roller bearings, light-weight rods, cylinders lined to reduce bore) for 80 mph service (as built the L-2 was limited to 60 mph), and were deemed successful.

Hence the L-3. Not too much bigger than the L-2 (engine weight up to 388,000 pounds for the first, up to 399,000 for later production), and built like the L-2 with 69 inch drivers… but spaced (76 inches instead of 72 between driver axles) to allow 72 inch drivers: two were so re-equipped. The L-4 (built during the war) was basically an L-3 repeat, but built with 72 inch drivers.

The L-4 were described as "dual service," but they (and the passenger equipped L-3) did much more passenger than freight work during WW II. (At a guess, troop trains-- heavy through passenger consists, but not schedule quite as fast as the commercial a med trains like the 20th Century Limited-- would have had L-3 and L-4 power on a regular basis.) There were 65 L-3 and 50 L-4, so by the time the Niagara was built the New York Central had had extensive experience with big 4-8-2 in high speed service.

(*) Most American railroads referred to the 4-8-2 type as a "Mountain" (the first ones, I believe, were ordered specifically for use on passenger trains in mountainous territory), but the New York Central -- which advertised itself as "The Water Level Route" -- called theirs "Mohawks," after a river the NYC main line parallels across part of New York State.

(Oh. forgot to cite source. The information on New York Central "Mohawk" locomotives in my last post is basically from Alvin Stauffer's "Steam Power of the New York Central System." Stauffer's books aren't the most scholarly of sources, but the story here is consistent with what I have read elsewhere.)

As for the power of the L-4 Mohawk in comparison to the Niagara… 5400 is an astonishingly high figure for such a comparatively small locomotive, but I guess believable: the Niagara is some 20% heavier (roughly 480,000 pounds to roughly 400,000), so the power/weight ratios are comparable. Since the Mohawk's maximum boiler diameter (94 inches) is about 6% less than the Niagara's, the outputs of the two types are roughly in proportion to their boiler cross sectional areas.

George Ellwood's "Fallen Flags" rail image site has available a New York Central locomotive diagram book, which might provide some other dimensions to compare. I'll look when I get a chance and report back if I find anything fascinating!

One place where the Mohawk seems undersized is the firebox: it has only about three quarters the grate are of the Niagara. So my guess would be that the Mohawk, in order to sustain its maximum power, would have had to be fired at a higher rate than the Niagara. The New York Central used, I think, the best coal West Virginia could provide, and didn't like wasting it. (Its mainline steam locomotives were equip with "valve pilots," which as I understand it were gauges showing the locomotive driver, at each speed, what the most efficient power setting were for that speed.) It has been said several times that the Niagaras were more powerful than the New York Central needed: J-3 Hudsons could handle most of the express passenger schedules, and freight train weights were such that the Mohawks could do time freights. The move to a bigger locomotive, then, was probably motivated as much by a desire for better fuel (and water) efficiency as by a desire for higher power.

Thanks for that Allen. As you said, economy of operation might have been one rationale for the Niagara design.

As well, once the NYC designers stepped into 4-8-4 territory, it was likely tempting for them to build a notable example, one that would be to the 4-8-4 community what its J-3 was to the 4-6-4 community.

The common rationale for the 79/80-inch driver 4-8-4 was to have a passenger locomotive that would match a 73/74-inch 4-8-2 or 4-8-4 on all but the steepest mountain grades, and yet be able to run as fast and as freely as a 79/80-inch 4-6-2 or 4-6-4 on the flat. It would inevitably be overpowered for typical trailing loads on the flat, but that trade-off was acceptable if it allowed much longer locomotive runs and so better utilization.

Against that, one might say that for the NYC, with its “water-level” route, a big 4-8-4 inherently was going to be overpowered. On the other hand, there was probably a legitimate case for some increase of output as compared with the L-4 4-8-2, and rail stress considerations would have dictated that this larger locomotive be a 4-8-4. Axle loading aside, the 4-8-2, typically with a fully-loading tailing truck well back of the rearmost drivers, tended to suffer from what might be called reverse bending moment relief. And a smallish boiler – say just a bit bigger than that of the L-4 – on a high-drivered 4-8-4 chassis – would have looked a bit silly.

Steam locomotive efficiency curves (coal consumption vs. drawbar power) tend to be inverted (vertex down) parabolas, with the vertex (best efficiency) somewhat below maximum power output. Quite likely the Niagara was more efficient, with similar loads and schedules, than the L-4 in situations where the latter was required to operate at outputs above its maximum efficiency point. That could have been an argument in favour of the Niagara, or a defence of it if anyone suggested that maybe it was too big.

Grate area is a vexed question, and the empirical evidence suggests that it would be difficult to derive a general formula that would lead one to choose 100 ft² rather than say 90 ft² for a given type of locomotive.

During the 1920s, 4-8-2 grate areas ranged up to around 84 ft² (MoPac 1927 and Wabash), with 88 ft² appearing later with the 2-10-2 rebuilds. But for the B&M R-1, a “big modern” 4-8-2, presumably with higher power output than most that had gone before, 79 ft² was chosen. Presumably its designers saw that as large enough, or perhaps that was as much as the trailing truck axle loading would allow.

At the other end of the scale, Lima chose 100 ft² for its A-1 2-8-4, but that may have been an overshoot; beyond what was optimal for locomotive operation but a better fit for the accompanying “hype”. There was no increase for the homologous 2-10-4, and Lima came back to 88 ft² (a number associated with the USRA heavy 2-10-2) for the MoPac 2-8-4s (those that were later rebuilt as 4-8-4s). 88 ft² was also chosen for some early 4-8-4s, in particularly the Rock Island R-67B. The Van Sweringen 2-8-4s had 90 ft² grates. It might also be said the A-1 and its 63-inch 2-8-4 clones did not seem to leave a lasting impression on the roads that operated them. They either made the logical upward step to the 4-8-4, or took a sideways step to the 4-8-2. The implied argument for the latter seemed to be that a heavy 4-8-2 was better at realizing the speed capability (at reasonable track stress levels) that its boiler would allow than was the case for a 2-8-4 of similar axle loadings, albeit with the tradeoff of a higher firing rate per ft² of grate area.

Pennsy chose a 92 ft² grate area for its T-1, which had a 100-inch BMOD. One imagines it was a considered number, as the design would seem to have allowed a 100 ft² grate had it been desired. Whilst the duplex concept can be seen as a case of “what were they thinking”, one would expect the Pennsy to have chosen appropriate boiler proportions nonetheless. Given that the same could be said [in respect of boiler proportions] for the NYC Niagara, and that both roads might have used similar quality coal, the conclusion is that there was quite a large “plus/minus” factor when it came to optimum grate areas.

As a sidebar, something that comes out of this kind of analysis is that there is not a sharp division between locomotives with two-wheel trailing trucks and those with the four-wheel kind; one might perhaps challenge the standard notion that a four-wheel trailer is sine qua non for the “superpower” moniker.

Cheers,

Has anyone come across an explanation of why the Wabash had both 4-8-2s and 4-8-4s that were nearly identical in all respects, except for the extra trailing truck axle on the 4-8-4s?