The Flowers of Edo: The Firebombing of Tokyo, Franklin and the R-3350

The seventieth anniversary of the firebombing of Tokyo comes with the first darkness of March 10th, and the burning of Franklin not for another nine days. So this is not a perfect anniversary post. It is certainly not a useful 70th anniversary of the Wright Duplex Cyclone, the 55 litre engine which, at this point in the war, had a habit of catching fire in mid-air. Close enough, I'll say, and leave it at that.

The  connection between the three is death by fire, but I will leave the burning men and cities below the fold. Here's a more hopeful intimation of a new world order to come. Like a phoenix from the ashes, etc, etc.

I. The morning of 10 March 1945, somewhere in Tokyo.

Source

II. Sometime afternoon, 19 March, 1945, between the waters off Tokyo and New York Navy Yard:

There exist pictures from the Franklin, or more likely Wasp or Bunker Hill, which make the photograph from the streets of Tokyo humane by comparison. An author on the subject took it upon himself to publish BuShips pictures of the carpet of bodies outside the air plot, basically an Essex-class's air group, overcome by smoke inhalation as they tried to escape. Google didn't turn them up, The Internet can be very restrained. 

Double standards. The firebombing on the night of 10/11 March 1945 cost an "estimated" 83,793 Japanese lives. Those are very definitely on show. I'm using scare quotes, by the way, because while we know that 267,171 buildings were destroyed, and that 20th Air Force aerial reconnaissance assessed it at "fifteen square miles" burned --dehoused, if will, the human losses are a bit more uncertain. They have been reconstructed over the last seventy years. It's not that the number is inaccurate, in the way that some of the estimates for the losses at Dresden are inaccurate. It's that we really need to stop and appreciate what the effort to put that number together implies. The archictectural statistics reveal a brutal body blow to the city of Tokyo. The human losses reveal a society that was picked apart that terrible night, and which is still putting itself together. Technically, the bombardment of a closed city is not a war crime, but this one attack killed a beautiful, vibrant city, which then put itself back together in a new and much more fireproof social order.

Speaking of moral dilemmas, the damage done to Franklin was not inflicted by kamikazes, but the Special Attack Corps would make its appearance in the afternoon of 19 March in its most horrific form. As with firebombing, the kamikazes tie us up on knots. Nominal volunteers, ordered to attack. Do we honour them as heroes, regret them as victims, or reject Japanese militarism in all of its forms? If the latter, I note that Captain Okamura, Admiral Onishi, Admiral Ugaki and others died to atone for their actions; Lieutenant General Pedro del Valle lived on an unrepentant --as far as we can tell-- twenty five years after he began his quixotic effort to organise an American storm trooper/death squad movement as the Defenders of the American Constitution. Kevin Coogan, I think tongue-in-cheek, wonders "how much of this existed in reality as opposed to Walter Mitty-like fantasy?" But that's not the point. The point, other than the bizarre life of General dal Valle, the weirdest thing I've learned about the 1940s and 50s since discovering that the Andrews Sisters were the Madonna of their day, is that if Japanese militarism produced particularly repellent tactics, it also produced more serious and consequential apologies.

There is death by fire, technically-not-war-crimes, and, again, death, so many deaths. And there are causes and consequences, things to take on board to better understand the history of technology in this world of ours. I think, not being smart enough to write meaningfully about the ethical issues in play here, I will move on to that.  

The context of the firebombing of Tokyo is well explained here.

Some time before dawn on 10 March, 1945, 279 of 345 B-29s despatched attacked Tokyo with loads of 6 tons each of M-68 napalm bombs attacked Tokyo at altitudes of between 4900 and 9200ft. Fourteen B-29s were lost, and the damage was beyond horrific. The results were also far from inevitable. The even  heavier attack on Nagoya the next night was a damp squib. "2.05 square miles" were reported to be burned out, and that sounds like it might be a wishful thinking. Nagoya was a new city in the Tohoku, the colder and less agriculturally productive coastal areas of northern Honshu that were virtually a Japanese frontier of settlement in the Nineteenth Century, but even the older city of Osaka, attacked on 13 March, got away with the burning of 8 square miles and 12,451 deaths. (Useful summary here.) 

While all of this was going on, Task Force 58 was dragging its tail off Japan's eastern waters, supporting ongoing operations on Iwo Jima, or the air offensive, or preparing for the Okinawa landings --in general, doing stuff because it existed, and there wasn't a better use for it while the Japanese retained significant land-based air striking forces, and its remaining fleet contented itself with force-in-being tactics. (One might think that the carrier strike force was such a powerful force that it might be able to impose its will on the Japanese air forces. One would be wrong. In the wake of his educational experience, upcoming, Admiral Mitscher would respond to major kamikaze attack waves against the Okinawa landings by pulling his carriers out of range, striking the aircraft below, and degassing them. See Morison 14:209) 

That is getting ahead of things. To quote myself:

On 18–31 March the fast carriers are sent in on an air superiority campaign preparatory to Okinawa. On 18 March the TF is rumbled, TG 58.4 attacked shortly after dawn by no more than 50 a/c. Enterprise is hit by a dud. Intrepid takes 2 dead, 43 wounded to fragments that start fire on hangar deck. 3 dive bombers attack Yorktown, 1 500lb bomb passes through 1 deck and “explodes near ship’s side,” holing hull and 5/26. Listed as minor damage.

On 19 March TG 58.2 attacked shortly after dawn. Wasp bombed by weapon that penetrates hangar deck, bursting in no 3 deck galley. Fires on 5 decks controlled in 15min. Another weapon detonates alongside. 101/269. Listed as moderate damage. Franklin is hit by 2 bombs dropped by an intruder not even detected before it bombs. One explodes on hangar deck, one above it. Internal explosions and enormous fire result. Ship not underway until 0300 20 March. 724/265. In the course of 20 March TF 58 withdraws slowly from its launching area, providing air cover for the slowly proceeding Wasp and Franklin. Heavy air attacks appear several hours after dawn and have relatively little effect. A weapon misses Hancock, which suffers minor damage. Fires due to AA splinters occurs on Enterprise flight deck. The most important aspect of this operation is a squadron-strength attack by 27 Mitsubishi G4M“Betties,” so-called heavy bombers actually lighter than Mosquitoes. They are carrying the first operational Yokosuka MXY7 Okha manned rocket glide bombs. The weapon's operational debut was not a success, as the G4M was very cumbersome with an Okha external load (it is not clear that any other combination would have been more successful), and the result was the pointless loss of both bombers and rocket bombs. On the other hand, finishing off the victims of the 18/19 attacks would have been quite a fillip for the Japanese. As it is, the operation forced  Intrepid, Wasp and Franklin to retire from the theatre, and Franklin would have been written off were it not for the propaganda consequences. (Though it is not clear to me who besides the Air Force might have been the target of the propaganda campaign.) 

Drilling down, Franklin was attacked at 0708 by a twin-engined "leaker" that penetrated TF 58.4's air defence screen without being detected by either radar or by Franklin's lookouts, although Hancock had raised a general warning via TBS at 0705. In defence of the Franklin, it was even more heavily overmanned than Hancock as a flagship, and was in the midst of flying off a full strike, with 53 aircraft in the air and 45 aboard, 31 armed and gassed up on the flight deck. (Source.) 

The attacker struck the Essex-class carrier with two weapons, one probably a 550lb semi-armour-piercing bomb, the other a 550lb general purpose one, then made a successful escape into the I'm-guessing-hazy-morning-visibility and complete anonymity. It was a very neatly executed attack, and it is unfortunate that we do not know who made it. The bombs, striking at low altitude had relatively little velocity. It was enough, however, to pass through the lightly built flight deck, although the SAP might have ricocheted. 

The GP bomb detonated in mid-air above the hangar deck. The SAP functioned as both bomb and ship designer intended. The fuze seems to have been triggered by the initial impact, so that the bomb did not penetrate intact to the armoured floor of the hangar deck.  designed, Nevertheless, as intended in the design, the explosion was fully effective in holing that the armour and penetrating the citadel of the ship. BuShips is quite pleased that the original concept of the Essex-class protective scheme is vindicated in this case, in that the retarding, fuzing effect of the flight deck penetration was counted on in protecting the deck armour from a penetrating weapon impact. Though not even BuShips thinks that this justifies the unarmoured flight deck concept --that bit of sophistry was reserved for modern contrarians. (This isn't to say that armoured flight decks are any better justified. It's a question of carriers that get destroyed before they can be used effectively versus carriers that don't get destroyed before they can't be used effectively due to their minute air grouips. What it is saying is that at a displacement of less than 30,000 tons, you really cannot have a useful fleet carrier.) 

Any self-congratulation, BuShips admits, would be beside the point. With all the munitions and gasoline on flight and hangar deck, the fires and secondary explosions that must inevitably follow will, at the minimum, wreck the ship. This was made clear the previous October, when Franklin was set on fire by a single Mitsubhish A6M Zero kamikaze armed with a single 250lb GP, killing 56 men and forcing the ship to retire to Puget Sound for repairs.  I have already said as much as needs to be said about the consequences of this more serious fire. Pictures, thousand words, etc. 

Franklin arriving in New York.

I don't think that the consequences of setting fire to the many, many flammable things onboard a fleet carrier really need to be dwelled upon. It is a little hair-raising to realise that of the 765 crew deaths, many occurred on the hangar deck, which only 2 men escaped alive, and that includes the entire staff of the Air Plot and CIC. I could dwell on the circumstances that led TF 58 to be "rumbled," or just point you to someone who has dwelled on them in the past, but here, again, too much can be made of it. The experts understandably want to make sure that this will not happen again. The frequency with which it did happen seems to me to sufficiently underline the difficulties of gaining full situational awareness of the air-sea battle from the command spaces of a flagship underway. Clearly, we need some way of not only gathering more information by extracting aircraft plots from the raw data produced by radar equipments, some way of turning jagged lines into soft curves, "smoothing" the "signal" to produce useful "data," if you will allow some novel turns of phrase, and then turning that data into a coherent picture of the situation with some kind of machine-aid to the figurative leader-helmsman --a "cybernetic system" if you will.

If you are a regular reader, you might have encountered this "the computer revolution started in WWII!!1!" self-indulgence of mine before, and I do not want to rehash old ground. I will leave Franklin to its complex failure of an old machine in the face of unparalleled stress and move on another fire from the air.

It was intended that the B-29 force attack Japan in daylight, with precision high altitude bombing in closed formations. It did not work out like that.

Toyama alight. It's a small city by Japanese standards, so this fire engulfs less than 2 square miles, but it sure looks impressive!

Supposedly, the failure was due to the unexpected atmospheric conditions over Japan. Who could have anticipated that the well-known phenomena of the polar jet stream might exist without doing something unthinkable and researching it? As presented to the public, the high stratospheric wind speeds encountered over Japanese targets affected bombing by reducing the amount of time available to take bomb sightings, by complicating aerial navigation, and by increasing flight times.

In reality, all of these difficulties were encountered over Germany and overcome. It certainly helped that Germany isn't in the polar jetstream, but stratospheric wind speeds can get pretty high over Germany, too. B-17s and B-24s did not fly as high as the pressurised B-29s, but the whole point of going to the higher altitudes was to gain range and immunity from interception and AA, all of which would gain operational advantages. 

The issue here is that not that the operational advantages did not exist, but that they did not pay off adequately. And the reason for that is that the B-29s flying over Japan did not achieve the necessary range. Pushing the range of the B-29 up to the target meant operating its engines lean, which produces more heat. And when wartime R-3350s were overheated, they caught fire. All of these tradeoffs, excepting the unsatisfactory engines, had already been encountered over Germany.

To hear more about the reasons that led General LeMay launched a firebombing attack on Tokyo on 10 March, I can only recommend again that you watch the documentary. There's a grim satisfaction in seeing all the arguments that Bomber Command laid out in a vain attempt to convince 8th Air Force to adopt the night area bombing campaign in Europe come back and, this time, be effective. Mike Sherry, if I recall correctly, as it is a long time since I read his book, makes a racial argument. Americans didn't want to be seen terror bombing Germans, even if they eventually ended up doing that in all but name, because Germans are White, but they were fine with doing it to the Japanese, because they weren't. I tend to think that you can get more complicated than that, that we have really not done all the work we need to do in understanding how early-Twentieth Century Americans came to be isolationist with respect to Europe, internationalist in the Far East. (If that rings oddly, and it may by the time I'm done, it is because I am writing this and last week's posts in reverse order.)

I also think that whether or not we take Claire Chennault as a typical American thinker is rather pointless. Eighth Air Force did not shift to night bombing because there was no way that it could have afforded to re-equip its planes for night flying, or retrain its aircrew. Twentieth Air Force flew much more modern aircraft that were suitable for night bombing, and the force was smaller and better trained, making night attacks entirely plausible. Moreover, the B-29 could carry a large enough bomb load to make area bombing seem feasible. I am going technologically determinist here. 

It's pretty simple, really. Here's the copy-pasted B-29 data box from Wikipedia:

General characteristics

• Crew: 11 (Pilot, Co-pilot, Bombardier, Flight Engineer, Navigator, Radio Operator, Radar Observer, Right Gunner, Left Gunner, Central Fire Control, Tail Gunner)
• Length: 99 ft 0 in (30.18 m)
• Wingspan: 141 ft 3 in (43.06 m)
• Height: 27 ft 9 in (8.45 m)
• Wing area: 1,736 sq ft (161.3 m²)
• Aspect ratio: 11.50:1
• Empty weight: 74,500 lb (33,800 kg)
• Loaded weight: 120,000 lb (54,000 kg)
• Max. takeoff weight: 133,500 lb (60,560 kg) ; 135,000 lb plus combat load
• Powerplant: 4 × Wright R-3350-23 and 23A Duplex Cyclone turbosupercharged radial engines, 2,200 hp (1,640 kW) each
• Zero-lift drag coefficient: 0.0241
• Drag area: 41.16 ft² (3.82 m²)

Performance

• Maximum speed: 357 mph (310 knots, 574 km/h)
• Cruise speed: 220 mph (190 knots, 350 km/h)
• Stall speed: 105 mph (91 knots, 170 km/h)
• Range: 3,250 mi (2,820 nmi, 5,230 km)
• Ferry range: 5,600 mi (4,900 nmi, 9,000 km, ^[63])
• Service ceiling: 31850 ft ^[21] (9,710 m)
• Rate of climb: 900 ft/min (4.6 m/s)
• Wing loading: 69.12 lb/sqft (337 kg/m²)
• Power/mass: 0.073 hp/lb (121 W/kg)
• Lift-to-drag ratio: 16.8

Armament

• Guns:
  • 10× .50 in (12.7 mm) Browning M2/ANs in remote-controlled turrets.^[64]
  • 2× .50 BMG and 1× 20 mm M2 cannon in tail position (the cannon was later removed)^{[N 11]}
• Bombs: 20,000 lb (9,000 kg) standard loadout.^[65]

74,500lb of engineered material, taking off with a typical 46,000lb combat load under almost 9000hp when running with catalytically-reformed and extensively doped 115/120 octane gasoline. Forget the low rate of climb and ridiculous 3-1 speed ratio. Four-engined bomber project after project had failed around the world over the previous decade, floundering out of the factory too late to compete with state-of-the-art performance, and too overweight to be useful in its own terms. In that light, the fact that the B-29 pilot has to "fight for airspeed" instead of "height" in the takeoff is a pretty trivial complaint, especially given that it was hurried into production two years ahead of a reasonable schedule. 

In fact, I'm not even going to talk about the plane today --that's mainly because I am still rather dilatorily casting around for technical material bearing on its fire-control system. Today I am going to talk about the engine. It is, above all, the failure of the Double Cyclone to deliver adequate long range performance under cruise control that is driving the shift to night combat. The B-29 fleet is being kept serviceable by the replacement of the five upper cylinders every 25 flight hours, and a full overhaul every 75. The relevant Wiki articles (1,2) tell us what is going on, but not why, which is that the carburettor is delivering a stratified charge, which is too rich in tetraethyl lead in some cylinders, and too low in it in others --the air-fuel mixture may vary for the same reasons, but my source focusses on TEL ratios--. The starved cylinders burn out quickly, while the enriched ones suffer lead attack to the valves. 

So that's a minor engine overhaul every sortie, and a full one for every three flights! It's also very high numbers. Even at its worst the Merlin did not push below a major rebuilding every hunded hours, and the Wright 3350s still in use, mainly by the US Forest Service, get 2000 hours. One way of looking at that is that only a very rich country can afford a plane so profligate of human technical skills. Another is that if that's what it takes to win WWII in 1945 and without an invasion of Japan, then it would be worth the cost, even if the "cost" were not, in effect, the training-by-doing of a country's worth of mechanics on the tropical isles of the Mariannas. When I put it that way, an engine that kept catching on fire in the air was practically a national-industrial strategic asset!

A postwar order is being born, and it's all thanks to the crap engineers at Curtiss-Wright! thanks to the imperial club for these great images, by the way.

Okay, totally not fair. Well, it is fair in the sense that Pratt & Whitney struggled through the same difficulties and managed to produce better and more reliable engines, and I am extremely suspicious about the relationship between Curtiss-Wright and Wright Field, but the distinction between the two is small, and it was Chrysler, not Curtiss-Wright, that actually made most of the Cyclones that fought the Battle of Japan.

And. . . wow. Here's my source again. I mean, I knew that these things were incredible industrial efforts, but did you know that the supercharger impellers rotate at 10,000rpm, and each of the twenty gears of their teeth have to take 100hp? That to deliver this performance, the machining has to be to such precision that the one-millionth of an inch expansion of a piece of steel with a change in temperature of one degree Fahrenheit is unacceptable, and that this was why 22 of the 84 acres of the Chicago-Dodge plant had to be air-conditioned, and that all of it would have been air-conditioned were it not for the limit on equipment? I like this bit of detail because I am currently on a kick of seeing refrigeration as one of the Big Things about the postwar era.   

Nor is it just metallurgy. In order to achieve adequate cooling, designers increasingly came to use the internal circulation of lubrication oil, which must pass through an oil cooler, as a means of cooling the engine overall. That the Duplex Cyclone was going to have an overheating problem was realised early. To meet this need, the internal oil circulation was pushed up to 160lbs per hour. With little irrigation pumps literally playing cooling flows of oil onto cooling fins machined into the underside of the cylinders, there was a demand for a system of internal pumps and sprayers with minimal oil loss. 

Dodge employees, who would be under the Air Inspectorate Directorate in Britain rather than being counted as engineering department employees, explaining much of the discrepancies of the Fedden report, for those interested in my ongoing critique of the Barnett thesis, are not necessarily checking the tolerances of oil pumps, but they are demonstrating the nature and range of the learning-by-doing associated with this great project.*

At the end of this commemorative post, I do not reallly come back around to a stronger unifying theme behind the near-loss of the Franklin, the destruction of Tokyo, and the B-29 production battle than the birth of the Information Age in the destructively transformative years of WWII. Tokyo would never have burned like it did were it not an old city, ill-designed against the fires that had long bloomed in its night, the so-called "blooms of Edo." Wright Aeronautical would not have turned out the R-3350 as a sort of perverse treasure box out of which Chrysler would create a new kind of America at its Chicago Dodge plant were it not for prewar American insouciance which assumed that American society would be able to create a world-beating military technology without paying for it. The unwillingness of some (strawman alert!) to learn from the fact that unfunded research and development leads to preventable disasters like the near loss of the Franklin.

You get what you pay for.  "Yankee ingenuity," or Yamato Spirit, or whatever exceptionalism floats your boat is not going to change that. And in many circumstances, it is better to pay in money in advance than to wait for your city, your ship, or your  plane catches fire. But, hey, if you don't want to pay for something, be my guest: don't pay. Just don't imagine that the calipers you don't apply to the pumps delivered by the subcontractors would show that they're up to snuff.

*By the way, I notice that the total number of "graduate engineers" at Dodge-Chicago was 120.