"I got it a half hour ago on the lateral array," Jones said. He referred to a subsystem of the BQQ-5 multifunction submarine sonar. Its main component was an eighteen-foot-diameter dome located in the bow. The dome was used for both active and passive operations. A new part of the system was a gang of passive sensors which extended two hundred feet down both sides of the hull. This was a mechanical analog to the sensory organs on the body of a shark. "Lost it, got it back, lost it, got it back," Jones went on. "It's not screw sounds, not whales or fish. More like water going through a pipe, except for that funny rumble that comes and goes. Anyway, the bearing is about two-five-zero. That puts it between us and Iceland , so it can't be too far away."

"Let's see what it looks like. Maybe that'll tell us something."

Jones took a double-plugged wire from a hook. One plug went into a socket on his sonar panel, the other into the jack on a nearby oscilloscope. The two men spent several minutes working with the sonar controls to isolate the signal. They ended up with an irregular sine wave which they were only able to hold a few seconds at a time.

"Irregular," Thompson said.

"Yeah, it's funny. It sounds regular, but it doesn't look regular. Know what I mean, Mr. Thompson?"

"No, you've got better ears."

"That's cause I listen to better music, sir. That rock stuff'll kill your ears."

Thompson knew he was right, but an Annapolis graduate doesn't need to hear that from an enlisted man. His vintage Janis Joplin tapes were his own business. "Next step."

"Yessir." Jones took the plug from the oscilloscope and moved it into a panel to the left of the sonar board, next to a computer terminal.

During her last overhaul, the Dallas had received a very special toy to go along with her BQQ-5 sonar system. Called the BC-10, it was the most powerful computer yet installed aboard a submarine. Though only about the size of a business desk, it cost over five million dollars and ran at eighty million operations per second. It used newly developed sixty-four-bit chips and made use of the latest processing architecture. Its bubble memory could easily accommodate the computing needs of a whole squadron of submarines. In five years every attack sub in the fleet would have one. Its purpose, much like that of the far larger SOSUS system, was to process and analyze sonar signals; the BC-10 stripped away ambient noise and other naturally produced sea sounds to classify and identify man-made noise. It could identify ships by name from their individual acoustical signatures, much as one could identify the finger or voice prints of a human.

As important as the computer was its programming software. Four years before, a PhD candidate in geophysics who was working at Cal Tech's geophysical laboratory had completed a program of six hundred thousand steps designed to predict earthquakes. The problem the program addressed was one of signal versus noise. It overcame the difficulty seismologists had discriminating between random noise that is constantly monitored on seismographs and genuinely unusual signals that foretell a seismic event.