The most cantankerous was the B-58 Hustler, America's first supersonic bomber. I flew as the back-seater in that fascinating craft for five years ( 1960-65 ). My job title was Defensive Systems Operator (DSO), but in reality I was more of a flight engineer. My primary job, de facto - if not de jur, dealt with aircraft performance, fuel management and serving as deputy assistant co-pilot without a stick, as it were. Oh yes, I also spent a lot of time on the radios. Our three-man crew consisted of a pilot, radar navigator and DSO.

The B-58's fuel system was fairly straight forward in design, with a complement of tanks used to both feed the engines and permit transfers for center of gravity (CG) control. There was a forward tank, an aft main tank, a reservoir tank and an aft balance tank. Supplemental fuel was carried in the bomb pod tanks, suspended beneath the fuselage and surrounding the weapon payload. A system of pumps kept transfer lines pressurized to feed the engines and another transfer manifold enabled tank-to-tank fuel movement for CG control. We additionally had an aerial refuelling receptacle with its associated plumbing.

Fuel gauges at the pilot's station and mine, in the rear in-line cockpit, enabled two of us to monitor the tanks. We also had a fuel flow gauge which reflected the total consumption of the four jet engines. And, there was a CG indicator which was supposed to tell us the position of our CG relative to the mean aerodynamic chord of the wing. It's readings depended upon both the fuel tank quantities and the weapon payloads. A somewhat-less-than-reliable computer was supposed calculate the CG for us, but we didn't trust it. So my job involved frequent monitoring of fuel quantities, locations and CG calculations.

All transfers of fuel from tank to tank were timed with a stopwatch, to verify the amount of fuel moved and to ensure that it went where it was supposed to go. The pilot had the only fuel controls, so we coordinated the transfers and double checked each other on valve openings and closings while I timed each operation.

With all this visibility, the control and timing of transfers, you might ask, "What could possibly go wrong?" In a word, everything.

Consider what could happen when one or more (and on rare occassions, all) fuel gauges failed.

How would you know where the fuel was and how much there was? The answers lay in the records and logs that the DSO kept. And the timing of transfers were recorded as well, to enable reconstruction of fuel movements. Further, there was a notation of what tanks were feeding the engines, and at what fuel consumption rates. So how could you lose track of your fuel and the CG?

Well, let's add the complication of fuel transfer valves that might or might to open or close on command. They did that with some regularity, just to keep us on our toes. And, oh yes, there was one more nuance to make life interesting. That was a valve between the aft main and the aft balance tank. It was normally kept closed, except when it wan't.

I recall one especially " interesting" flight when we had to return to home base because of smoke and fumes in the cockpit. Now that wasn't really unusual, for the vacuum-tube-based avionics systems located forward of the navigator and aft of the pilot was notorious for over-heating. Smoke and fumes in the cockpit was a way of life.

It's just that this time we had some serious fuel system problems too. Several of the fuel tank gauges and not a few transfer valves failed. And I seem to recall we lost one or more fuel pumps.

I was busy as a one-armed paperhanger trying to keep up with our fuel state and the CG position. In that delta-winged wonder maintaining proper CG was an important matter. We'd lost a few birds because the CG got out of hand. Serious business this fuel management and CG control.

The most troublesome malfunction turned out to be the little aft main-to-aft balance tank valve. It stuck in the open position. That would not have been a problem if we had actually landed on our first approach. But a vehicle or other obstacle on the runway forced us to go-around, meaning the pilot had to push the throttles forward, accelerate and climb out for another approach. So what, you ask?

Well, when you accelerate and climb, what happens to fuel in the aft main tank? With the aft to balance tanks valve stuck open, the fuel tends to move aft because of the acceleration. And when that happens the CG shifts aft too. But how much did it shift and how much fuel remained in each tank? A logical and important question. The answer? Who knows.

Only by estimating or working a problem backwards was I able to ascertain the CG and determine if it was safe for landing. The elevon ( combined elevator and ailerons ) position indicator gave us our clue. For any gross weight, altitude and airspeed there was a corresponding elevon position ( up or down ) which reflected CG and fuel distribution. A series of charts in the flight handbook allowed computation of the elevon position for specific conditions. If the CG was aft of normal, the elevon would reflect a lower-than-normal angle. Conversely, a forward of normal CG would result in a higher than normal elevon angle. Does that make sense? Sure it does.

By skill, cunning and a dose of luck, I was able to confirm that the CG would be within allowable limits for landing, provided, of course that we extended our downwind leg in the traffic pattern to allow the aircraft to cease accelerating, to stabilize and let the fuel that had shifted aft to flow forward into the aft main tank.

Simple enough, eh?