For some time, Knolls Atomic Power Laboratory has been working on a $65 million nuclear propulsion system for NASA to power rockets and space stations, which could eventually take humans to Jupiter's moons or even the stars.

The secretive facility hiding in plain view recently dribbled a few rare details about the space project.

The program has changed a bit, even over the past few weeks, as NASA has rearranged priorities under a new chief. One thing remains constant - the $65 million NASA is funneling into the lab.

That money also has brought new hires from all over the country, many of them engineers with advanced degrees. This year alone, the lab hired 350 employees, 150 for the space program, with plans to add more in 2006.

For decades, space jockeys have turned their gaze toward exploring the solar system's other planets. But catapulting spacecraft far beyond the Earth's orbit calls for a propulsion system previously drawn only by Hollywood. NASA needs an energy source more compact and faster than current rocketry, so the agency has contracted the Knolls Atomic Power Lab to design the engine.

Nuclear-powered rockets have been on NASA's radar since the late 1950s, although some of the earlier ideas might seem like a bad joke if they weren't true - igniting nuclear bombs beneath rockets to propel them spaceward, for instance. That idea never flew.

The Air Force did launch a small reactor in 1965, according to NASA documents. The outcome of that experiment isn't known.

The projects have had their detractors. Articles posted over the years on the space industry's online resource, space. com, have offered warnings that nuclear rockets could start the militarization of space.

The Bulletin of the Atomic Scientists, a nuclear science industry bible, has reported warnings that problematic launches or re-entry into Earth's atmosphere could spread radiation throughout the globe. For example, imagine what might have happened if the Columbia shuttle had a nuclear engine when the spacecraft disintegrated while attempting to land. The debris stretched across four states.

NASA's case for nuclear rocketry also suffered when it was made public that the program was revived in the late 1980s as part of President Reagan's "Star Wars" plan. The nuclear systems would propel the heavy missile defense system into space, according to the Bulletin.

Less militarized uses for nuclear propulsion in space were broached in the early 1990s.

The lure of nuclear power lies in efficiency. The basic science is the same as nuclear power plants: When uranium atoms split, they release energy. On Earth, that energy can level cities or light lamps. In rocketry, that energy can heat hydrogen gas, which shoots out of a rocket to propel a spaceship - or power robots and heating systems.

Knolls already designs and tests nuclear power systems for naval submarines, so the Niskayuna facility was a natural choice when NASA joined with the National Nuclear Security Administration to study nuclear power, according to Knolls spokesman Anne LaRoche.

The Jet Propulsion Lab at the California Institute of Technology is charged with building the actual rocket.

The Navy owns Knolls, but Lockheed-Martin runs it. The lab has a $450 million annual budget.

It sits behind a bluff overlooking the river, with chain-link fences wrapping the property and security maintained by armed guards.

A few years ago, NASA pitched sending a rocket to Europa, one of Jupiter's moons, a program dubbed JIMO, or Jupiter Icy Moons Orbiter.

In mid-May, NASA leaders told Congress they were shelving JIMO indefinitely in favor of focusing on a nuclear-powered base for the moon by 2015, which could then launch astronauts to Mars. "I will say categorically we cannot effectively explore space without nuclear power and, in the longer run, nuclear propulsion," Michael Griffin, NASA's administrator, told a subcommittee of the Senate Appropriations Committee on May 12. "The first thing we will need is surface nuclear power for our astronauts when they return to the moon in a decade or so."

That won't affect Knolls, LaRoche said, because the nuclear technology remains essentially the same. "They are deliberately general on the terminology," LaRoche said. The engines would likely power not only the thrusters, but also the general craft, robots and experiments from a case the size of a dorm room refrigerator. "This reactor will be the source of electrical power for the spacecraft when it's a zillion miles from here," LaRoche said.

The rocketry work marks one of the only Knolls research projects that have public components, and even then the answers to basic questions, such as the number of employees, must be vetted by several layers of federal officials.