Correct. The R-4360 engine used in the B-36 had other issues, too, mainly overheating and carburetor icing problems, made worse by the pusher configuration. Engine fires were a very real possibility on the B-36.
They did dual duty actually. My grandfather was a flight engineer (and top turret gunner) on a B-17 Flying Fortress for 25 missions over Germany. He had begun his USAAF career as a volunteer who wanted to be a mechanic, but due to the needs of the service, got shifted to gunnery instead. But since he had that mechanic background, he was tapped to be a flight engineer.
To hear him tell it, the flight engineer job was more of a in-flight fix-it sort of guy and/or person who accomplished tasks that were mechanical in nature, but that didn’t involve piloting. For example, he was the guy who’d crank down the landing gear if the hydraulics got shot up. Or he’d be the guy who went back into the bomb bay and took out the pins to arm them once they were near their targets. He also transferred fuel between tanks if they were leaking, etc… I’m sure there were more things, but those are the ones I specifically remember him saying he’d had to do on combat missions.
Of course, he was first and foremost a gunner- the top turret gunner in fact, probably because the top turret was located immediately behind the cockpit and he was right there if he was needed for flight engineer stuff.
In general, as described in the posts I linked to earlier, the FE has a very big role in preparing the airplane for launch up through getting airborne. And likewise from post-landing through walking away from it. Once in cruise there was little to do on B-17 level tech, more on B-36 tech, and a bit less on 707-era tech.
As such FE’s could have other cruise duties such as gunner, battle damage repairer, bomb armer, etc.
Setting aside battle damage, the need for inflight mechanic was large on the unreliable pre-WWII designs and dwindled to near-zero at the very end of the FE in early 1970s designs, while the need for inflight systems operator peaked at about the B-36 / DC-7 / Lockheed Constellation era and again dwindled to near-zero at the very end of the FE in early 1970s designs.
When I was a 727 FE it was normal to tighten loose panel screws in the cockpit or interior, replace lamps in idiot lights, bang on stuck galley carts, and, based on systems knowledge and personal experience, pull and reset circuit breakers to try to reset various systems that were acting funky. All that is verboten now.
Well, he didn’t have to worry about landing gear or brakes.
I read somewhere that the Concorde was extremely complicated, maybe even moreso than the DC-7 or Connie. It did require a Flight Engineer. I know that part of his job was to pump fuel to different tanks to change the CG for different speeds.
Ref my comment a couple posts ago about B-36 FEs, I have since located an operating manual for a B-36H.
That airplane carried 2 FEs and they sat side by side both working a monster panel stuffed full of gauges, switches, and levers. The earlier B-36D had one FE and a different panel arrangement.
And navigational errors could be lethal, not always for obvious reasons: Wander too far off course and a paranoid superpower might shoot you down. (In fact, that incident was a major impetus behind the Reagan Administration’s decision to make GPS available for civilian use.)
Here are some examples of engine failure and prop runaway:
Crash of a Douglas C-54-DO Skymaster in Seattle: 28 killed | Bureau of Aircraft Accidents Archives.
This will give some flavour of how FEs earned their money.
A slight O/T trip with a full cockpit on an interesting DHL charter from Kandahar to Bagram in Afghanistan, 2002.
DHL hired a Romanian (could have been Bulgarian) Antonov AN-12 for package delivery for NATO forces. It had a set rate for a certain minimum number of kilograms and flew from Bahrain to Kandahar-Bagram-Lahore (for fuel) and back daily. The pilot was in standard uniform; cap, white shirt with epaulets and gold stripes. The co-pilot had just stepped off a tennis court with sneakers, tighty-whities, and a polo shirt. The navigator slept the entire time 
dressed in shorts and a t-shirt. The engineer sat behind the Captain on a swing-out seat facing right initially dressed in sweatpants and shirt. . The AN-12 is similar to a C-130 (4 engine turboprop) with an unpressurized cargo hold, a pressure bulkhead up front. I think two rows of seating in theater style, and the flight deck “right there.” We (a few passengers and crew) climbed up the ladder by the front landing gear to get in/out of the plane. After the takeoff checklist and an engine start; the engineer pulled a pin and his whole seat swung out and relocked between/slightly behind the pilots in front of a massive set of gages for all the engines/generators/fuel etc… After takeoff, he was also responsible for manually pumping the handle to retract the nose wheel the last few inches. He and the pilot remained attentive; the co-pilot read a book until landing, and the navigator slept. We never did land at Bagram, massive spring dust storm (120 days of wind!); diverted to Kabul much to the surprise of the DHL representative flying with me and trucked the cargo the rest of the way the following day. We (the pilot) had to pay a landing fee in Kabul - cash money in dollars. The DHL had empty trucks brought on (free) but we “diverted” through the Turkish compound (Turks were security at the airport at that time) to avoid paying more to get loaded trucks out the airport proper gate. I slept at the DHL compound in the shadows (literally) of the US Embassy. We had steaks, fried potatoes, salad, and genuine Czech Budweiser beer. An interesting trip.
That’s my experience with an aircraft with flight engineer. Back to the topic!
That great story reminds me of an old joke.
For reference, here’s a generic B-727 F/E panel. Note the shallow yellow desk/table at the lower edge. That’s covered with a bunch of tables and graphs under plexiglass. Most engineer panels on other aircraft types have a similar table to hold books, clipboards, log forms you’re keeping as the flight goes along, etc. It also made a real nice surface to set your meal tray on while eating. And when the non-work got too boring you could also lay your head on it.
Anyhow, with that background, one of the standard jokes of the day was:
Q: What’s the difference between a Flight Engineer and a dog?
A: The dog sleeps under the table after dinner.
I was looking something else up, and found this:
“It was the first land-based aircraft to include a flight engineer as a crew member”
Real life has kept me busy and away from participating in this conversation. Thank you all for your interesting contributions.
Unlike an ocean-going ship, the engines of an aircraft are pretty much out of reach during the voyage. If something goes wrong, the F/E’s options have got to be pretty limited. In a military aircraft, might he be crawling around in the wings trying to effect repairs? In a commercial aircraft, I can imagine that such a thing would be rather upsetting to the passengers. 
Are there modern accidents, such as the recent problems with the 737Max, that may have been recoverable with a F/E working on the problem? Are we less safe without the third set of eyes in the cockpit?
United Airlines Flight 811 out of Honolula was 1989, which is 30 years ago now. In the cockpit recording, the Flight Engineer can be heard asking if he should dump fuel, then making the decision himself, because the Pilot was suffering a near-panic narrowing of focus. The aircraft landed safely, but overspeed: the fuel load would have made the aircraft more difficult to fly, and more difficult to land.
It’s not common now to have a section of the side of an airplane blow out, but if it does happen, they won’t have a FE to go back and report on the state of the hull, nor do the fuel dump* while the pilot is asking “where is the closest airport?”
* And I though that being in charge of the fuel balance sounded like a boring job…
I don’t know, but I recently watched an interview on YouTube with a British Airways Concorde pilot in which he claimed that a contributing factor (among many) to the fatal Concorde crash was the FE shutting down one of the engines (due to the fire alarm) even though it was still producing power, and not announcing this clearly. Even if that engine was not producing full power, what power it was providing before being shut down might have given the aircraft enough altitude to attempt an emergency landing at the nearest airport, as opposed to crashing in a village.
That’s not to say that having the FE made that flight less safe, of course (and the above was not mentioned in the official accident report). In general you make a good point that having another person in the cockpit could help the pilot out. Though that only works if the culture in the cockpit/airline allows - sadly, a common factor in aviation accidents is pilots ignoring inputs from their colleagues and/or said colleagues not being forceful enough due to seniority. Hopefully less so nowadays, as I believe Crew Resource Management training has become standard across the commercial aviation industry. Actual experts will be along to comment shortly, I hope.
That’s true today. But was much less true in the past.
Way back in the large biplane days some airplanes had a small open cockpit right behind each engine out there in the pod hanging between the wings. Some poor schlub sat out there in the prop blast the whole time. What exactly did he do? I can’t say for sure; perhaps he was fire watch & had fuel shutoff controls, watched for signs of oil leaks, etc. Heck, maybe he had the only throttle & operated it in concert with hand signals from the pilots, a la the engine room telegraph common on steamships of the day.
Fast forward to the pre-WWII flying boats and the post-WWII B-36. The F/E could get out to the engines to replenish the oil supply. The B-36 engines had a 55 gallon tank per engine, but that wasn’t enough for a long mission, so they’d be periodically refilled from drums carried in the fuselage.
As airplanes became more complicated, but engines more reliable, the engineer duties morphed from airborne engine repairman to airborne engine operator / monitor to finally in the jet era, engine starter and “everything else” operator and monitor.
Back in my own 727 experience we had a procedure to confirm the main gear was down if the cockpit idiot lights failed. It involved going back into the cabin, peeling up the aisle carpet amidships, locating a 8" hatch in the floor, opening it, then peering through a periscope the supposedly gave a view of the main gear’s mechanical over-center locking arrangement. If we saw the mechanism was properly over-center, the gear would hold on landing. If not, not. Most likely we’d see nothing through the accumulated grease, grunge, and rubber residue stuck to the outside of the lens; gear wells are not clean rooms.
I never had to do that for real, but it sure would have excited the audience no end. I did go back a time or two to investigate a slight vibration felt by the pilots in the flight controls. Yup, one spoiler panel was riding slightly high, or one flap segment was drooping slightly low. About as significant as your car pulling slightly left or right on the freeway. In other words, of interest to the mechanics, but not a safety concern.
But boy did that alarm the folks when I was peering intently out the windows at the wing. “Oh, It’s nothing important” did nothing to assuage them. So it was time to explain it all, show them what I’d seen and how minimal it was, yet we’d noticed it. That’s how careful we are about the integrity of our machines for your safety. That made them a lot happier.
In a surprising coincidence …
On my most recent trip as we’re g’buy-ing at the pax during deplaning one guy pulled aside, showed me a diagram of the airplane on his phone and told me he’d seen something loose or vibrating out on the wing. It’s probably been 5 years since the last time someone did that with me.
Between his weak English, the fact his pic was a 747, not a 737, and how hard it is to point precisely on a phone screen I wasn’t sure what he’d seen where. I had flown that leg and felt nothing odd. Anyhow I thanked him and after everyone had deplaned I looked out the cabin windows and also went outside on the ground to see what I could see. Which was exactly nothing.
I’m betting it was a bit of gap seal which is like super-duty weather stripping where the flap segments tuck up under the wing. They’re wear components that get scrubbed a bit each time the flaps move and eventually they crack and develop peeling bits that flail in the breeze until they tear off. Maintenance checks them for integrity every so often and once they’re ratty enough they get replaced.
That’s very perceptive. A third person in the cockpit is a valuable asset regardless of their role. They can often see the “big picture” before the working pilots can; it’s kinda hard to synthesize that from any one seat up close to the panels & lights and all. And for darn sure when the flying gets difficult two people can both become task saturated. Having a third person to delegate tasks to is very valuable in those bad-to-worst case scenarios.
Or just to monitor the two up front and remind them when they drop a ball unnoticed. Two people can also amp one another up emotionally when the adrenaline is running; somehow three tends to leave one odd man out in a calmer frame of mind who can sooth the situation. The goal is to be bored {yawn} while your pants are on fire. Easier to say than to do.
Given the unlikelihood of serious in-flight emergencies, a 3rd crewman, be that pilot or FE is not economically viable. And would suffer from severe skill atrophy / boredom given how rarely they’d do anything besides eat.
The original design for the 767 included an engineer which was removed before the final prototype version ever flew. They had had to invent a bunch of make-work for the job, given how automated they’d made the systems compared to the previous 707 / 727 machines. The airlines told Boeing in no uncertain terms that they wouldn’t buy the airplane with that “union featherbedding” built in.
So since then medium-haul flights up to 9 hours ~= 4,500 miles carry just 2 pilots, not 2 pilots + F/E. On the longer 9+ hour flights we carry a 3rd or even 4th pilot so somebody can nap while two folks work. It’s actually better / safer that way, since back in the F/E days we didn’t carry any extra anybody so all 3 were on-duty the whole time. (This paragraph is a simplified sketch but matches the big picture)
There’s still a vestigial engineer panel there on the 757/767 sidewall behind the FO, but all the gizmos on it are for maintenance use only on the ground. It’s so unimportant I had a hard time finding a pic of it to share. See the “Auxiliary Panel” item on the third slide in this slideshow; that pic is rotated sideways; the top is shown on the left. Other than a comm panel and an oxygen mask for the observer seat nearby, it’s just a few switches & Cannon plugs for the maintenance folks. And in another quaint touch there’s a circular pull out ashtray near the bottom. Those have long since been removed from near the pilot seats.
The original design for the 737 also expected to carry an engineer. But there would be no sidewall panel; all his (in those days “his” was fully appropriate) stuff was integrated onto the overhead panel and he’d sit facing forward right behind the center console. That too didn’t last more than a few months after introduction into service in 1968. United Airlines almost had a meltdown as the pilots protested the loss of the 5 jobs represented by each 737 that would (eventually) replace a 727. (it takes ballpark 5 crews to keep one airplane busy year round). But the economics were not to be denied. By that era, the F/E was an extravagance, not a necessity.
And yet the vestiges remain. This is a pic of the overhead panel on a shiny new 737 MAX. You’ll notice that most of that panel background is a dark gray color. But there are 4 sections that are a light gray/silver color. That (and only that) was the flight engineer’s stuff; pilots were not supposed to touch it back in 1966 when that was designed. Here in almost-2021 the color remains as a mute reminder of a bygone era.
As has been discussed upthread and in other threads, the Lion Air MAX airplane that crashed had flown a 90 minute flight the previous day with the same AOA malfunction and consequent MCAS runaway in place. That crew was surprised, but handled it OK as Boeing had expected and flew uneventfully to their destination.
There was a third pilot riding in the jumpseat on that flight. But not on the accident flight. Was he/she the difference? For whatever reasons fair or foul, the official Indonesian NTSB-equivalent report is silent on all this. But that extra set of eyes, hands, and brain may may well have been invaluable.
Finally …
As @Melbourne and @Dead_Cat suggest, more people is also more opportunity for stuff to become uncoordinated. We can all practice the relatively scripted failures in the checklists. But when stuff is really going berserk, you’re probably better off not having uncoordinated activity even if that leaves some stuff undone, or done later than it might have been. Folks hastily shutting down an engine producing useful thrust or worse yet the wrong engine are common ways of converting a problem into an accident.
Engineers’ relatively myopic focus on mechanical health can lead to mistakes in priority. Having proper oil pressure is nice, but it’s altitude above the rocks and speeds above stall that will keep us alive another minute; oil pressure be damned. This accident
was arguably a similar situation, where a too-quick reaction to a developing hydraulic problem led to the engineer rendering the aircraft uncontrollable before the pilots were ready to deal with the consequences.
Sidetrack into fuel dumping:
Fuel dumping is another vestige. All the old 3- & 4-engine long range airplanes had it. Even the pitifully underpowered 727 had it. Most modern twins don’t. Because it’s really not needed. It’s purpose is not, and never was, about reducing fire hazard on landing. It was about lightening the airplane enough that it could perform a go-around from an engine-out landing. And back in the far more under-powered recip days, it was about lightening the airplane enough that it could maintain an altitude above reasonably high terrain while engine-out.
For reasons of fuel supply safety if a malfunction triggers a rogue dump, the actual rate of fuel jettison is not real impressive compared to the size of the tanks. If you’re in deep shit and need to land ASAP, the amount of weight difference between dumping or not is trivial bordering on negligible. Conversely, if you’ve got a couple hours to waste, and only a skimpy airport nearby, then dumping is useful for the corner case of problems right after take-off with nearly full tanks. But if you’ve got a couple hours to waste, you should not hastily be starting that dump. And you might be able to find a less skimpy airport within the thousand or so miles you’ll cover anyhow droning in circles over those two hours.