"Maybe if you used your hand," Burne suggested.

Richman tried swiping at the handle with his palm. In a few moments, the side of his hand was red, and the lever remained firmly up and locked.

"Okay," he said, sitting back in the seat. "I get the point."

"It can't be done," Bume said. "It simply can't be done. An uncommanded slats deploy is impossible on this aircraft. Period."

From outside the cockpit, Doherty said, "Are you guys finished screwing around? Because I want to pull the recorders and go home."

As they came out of the cockpit, Burne touched Casey on the shoulder and said, "See you a minute?"

"Sure," she said.

He led her back in the plane, out of earshot of the others. He leaned close to her and said, "What do you know about that kid?"

Casey shrugged. "He's a Norton relative."

"What else?"

"Marder assigned him to me."

"You check him out?"

"No," Casey said. "If Marder sent him, I assume he's fine."

"Well, I talked to my friends in Marketing," Burne said. "They say he's a weasel. They say, don't turn your back on him."

"Kenny…"

"I'm telling you, something's wrong with that kid, Casey. Check him out."

With a metallic whir from the power screwdrivers, the floor panels came away, revealing a maze of cables and boxes under the cockpit.

"Jesus," Richman said, staring.

Ron Smith was directing the operation, running his hand over his bald head nervously. "That's fine," he said. "Now get the panel to the left."

"How many boxes we got on this bird, Ron?" Doherty said.

"A hundred and fifty-two," Smith said. Anybody else, Casey knew, would have to thumb through a thick sheaf of schematics before he answered. But Smith knew the electrical system by heart.

"What're we pulling?" Doherty said.

"Pull the CVR, the DFDR, and the QAR if they got one," Smith said.

"You don't know if there's a QAR?" Doherty said, teasing him.

"Optional," Smith said. "It's a customer install. I don't think they put one in. Usually on the N-22 it's in the tail, but I looked, and didn't find one."

Richman turned to Casey; he was looking puzzled again. "I thought they were getting the black boxes."

"We are," Smith said.

"There's a hundred and fifty-two black boxes?"

"Oh hell," Smith said, "they're all over the aircraft. But we're only after the main ones now-the ten or twelve NVMs that count."

"NVMs," Richman repeated.

"You got it," Smith said, and he turned away, bending over the panels.

It was left to Casey to explain. The public perception of an aircraft was that it was a big mechanical device, with pulleys and levers that moved control surfaces up and down. In the midst of this machinery were two magic black boxes, recording events in the flight. These were the black boxes that were always talked about on news programs. The CVR, the cockpit voice recorder, was essentially a very sturdy tape deck; it recorded the last half hour of cockpit conversation on a continuous loop of magnetic tape. Then there was the DFDR, the digital flight data recorder, which stored details of the behavior of the airplane, so that investigators could discover what had happened after an accident

But this image of an aircraft, Casey explained, was inaccurate for a large commercial transport. Commercial jets had very few pulleys and levers-indeed, few mechanical systems of any sort. Nearly everything was hydraulic and electrical. The pilot in the cockpit didn't move the ailerons or flaps by force of muscle. Instead, the arrangement was like power steering on an automobile: when the pilot moved the control stick and pedals, he sent electrical impulses to actuate hydraulic systems, which in turn moved the control surfaces.

The truth was that a commercial airliner was controlled by a network of extraordinarily sophisticated electronics-dozens of computer systems, linked together by hundreds of miles of wiring. There were computers for flight management, for navigation, for communication. Computers regulated the engines, the control surfaces, the cabin environment.

Each major computer system controlled a whole array of sub-systems. Thus the navigation system ran the ILS for instrument landing; the DME for distance measuring; the ATC for air traffic control; the TCAS for collision avoidance; the GPWS for ground proximity warning.

In this complex electronic environment, it was relatively easy to install a digital flight data recorder. Since all the commands were already electronic, they were simply routed through the DFDR and stored on magnetic media. "A modern DFDR records eighty separate flight parameters every second of the flight."

"Every second? How big is this thing?" Richman said.

"It's right there," Casey said, pointing. Ron was pulling an orange-and-black striped box from the radio rack. It was the size of a large shoe box. He set it on the floor, and replaced it with a new box, for the ferry flight back to Burbank.

Richman bent over, and lifted the DFDR by one stainless-steel handle. "Heavy."

That's the crash-resistant housing," Ron said. "The actual doohickey weighs maybe six ounces."

"And the other boxes? What about them?"

The other boxes existed, Casey said, to facilitate maintenance. Because the electronic systems of the aircraft were so complicated, it was necessary to monitor' the behavior of each system in case of errors, or faults, during flight. Each system tracked its own performance, in what was called Non Volatile Memory. "That's NVM "

They would download eight NVM systems today: the Flight Management Computer, which stored data on the flight plan and the pilot-entered waypoints; the Digital Engine Controller, which managed fuel bum and powerplant; the Digital Air Data Computer, which recorded airspeed, altitude, and overspeed warnings…

"Okay," Richman said. "I think I get the point."

"None of this would be necessary," Ron Smith said, "if we had the QAR." 

"QAR?"  

"It's another maintenance item," Casey said. "Maintenance crews need to come on board after the plane lands, and get a fast readout of anything that went wrong on the last leg."

"Don't they ask the pilots?"

"Pilots will report problems, but with a complex aircraft, there may be faults that never come to their attention, particularly since these aircraft are built with redundant systems. For any important system like hydraulics, there's always a backup-and usually a third as well. A fault in the second or third backup may not show in the cockpit So the maintenance crews come on board, and go to the Quick Access Recorder, which spits out data from the previous flight They get a fast profile, and do the repairs on the spot"

"But there's no Quick Access Recorder on this plane?"

"Apparently not" she said. "It's not required. FAA regulations require a CVR and a DFDR. The Quick Access Recorder is optional. Looks like the carrier didn't put one on this plane."

"At least I can't find it" Ron said. "But it could be anywhere."

He was down on his hands and knees, bent over a laptopcomputer plugged into the electrical panels. Data scrolled down the screen.

???? A/S PWR TEST 00000010000

???? AIL SERVO COMP 00001001000

???? ADA INV 10200010001

???? CFDS SENS FAIL 00000010000

???? CRZ CMD MON INV 10000020100

???? EL SERVO COMP 00000000010

???? EPR/N1 TRA-1 00000010000

???? FMS SPEED INV 00000040000

???? PRESS ALT INV 00000030000

???? G/S SPEED ANG 00000010000

???? SLAT XSIT T/0 00000000000

???? G/S DEV INV 00100050001

???? GND SPD INV 00000021000

???? TAS INV 00001010000

"This looks like data from the flight control computer," Casey said. "Most of the faults occurred on one leg, when the incident occurred."

"But how do you interpret this?" Richman said.