Jules Verne thought you could get to the center of the Earth through the chimney of an extinct Icelandic volcano. At the bottom, he envisioned a vast inland sea -- really inland, as in 4,000 miles down -- and a bunch of dinosaurs.

This, it turned out, was science fiction.

These days we know more, and this week Paramount Pictures will release "The Core," in which a team of intrepid "terranauts" try to reverse the collapse of the Earth's magnetic field by traveling to the metal center of the planet and setting off a nuclear bomb.

This is somewhat closer to reality, since it reflects the prevailing view that the Earth's core is made of partially crystallized iron and nickel, and makes use of the idea that heat from the cooling core contributes to the magnetic field that repels solar radiation and keeps it from frying the Earth to a crisp.

Still, no one knows exactly what's down there and, despite "The Core," there are no terranauts getting ready to find out.

The question is still open, and maverick geophysicist J. Marvin Herndon has a theory in many ways as radical as Verne's was 150 years ago: The center of the Earth, he believes, is a nuclear fission reactor.

And in the Proceedings of the National Academies of Science earlier this month, he argued that the mix of helium isotopes rising in lavas to the Earth's surface suggested that the "demise of the georeactor" is approaching. The reaction could cease anytime from 100 years to 1 billion years from now, collapsing the Earth's magnetic field with monumental consequences.

Nevertheless, "I'm not a bit worried," Herndon said in an interview. "I don't know how long it will take when it starts -- maybe hundreds of years. In this paper, we say that we have perhaps the first warning."

Herndon and Paramount have forged a publicity partnership of sorts. Although Herndon said he has not received any money from the film company, he is unabashedly lauding "The Core" in hopes of raising the "curtain of silence" that has stymied debate on his theory since he propounded it 12 years ago.

"I like the way he does his science, but I can't say the same about the people who review his work," said Hatten S. Yoder Jr., former director of the geophysical lab at the Carnegie Institution of Washington and one of Herndon's most influential boosters. "The geophysical community has been doing this to him for years. He's had a terrible time, and I hope this latest paper will generate some civilized discussion."

It does not appear to have done so. "It's a very controversial topic," said Don L. Anderson, a geophysicist at the California Institute of Technology. "I'm a little more open-minded than many of my colleagues," he added, noting that georeactors may help explain some anomalies in the Earth's crust and mantle. But it is "ridiculous" to suggest that the Earth's geomagnetic field is threatened, he said.

Most scientists agree that the Earth was formed about 41/2 billion years ago from an amalgam of fiery material that contained all of the elements found in nature. Disagreement begins over what happened next.

The view held by most Earth scientists is that iron and nickel migrated downward, taking with them all the trace elements that readily combine with these metals. The rest of the trace elements, including uranium, combined with oxygen to form oxides that remained in the Earth's mantle and crust.

These scientists use common meteorites as their model, but Herndon used a rare meteorite with a small amount of oxygen as his example, arguing that the uranium would remain metallic and, as the heaviest element in nature, would migrate to the Earth's core, forming a sphere about five miles in diameter -- a natural nuclear fission reactor.

"It's a self-sustaining critical reaction," said nuclear engineer Daniel F. Hollenbach of Oak Ridge National Laboratory, a longtime collaborator of Herndon's until the two parted ways last year. "Depending on how much it fissions, that's the power."

Hollenbach explained that the core would be composed primarily of two uranium isotopes. Atoms of the isotope U235 would split, giving up neutrons, which would be absorbed by the isotope U238, transforming it into an isotope of plutonium -- Pu239. The numbers signify the number of protons and neutrons in the atomic nucleus, known as the atomic weight.

This reaction, the same produced in some nuclear power plants, eventually creates radioactive waste isotopes much lighter than uranium. These migrate upward and outward from the core, "like fizz from a soft drink," Hollenbach said. The heat from the reaction is what drives the Earth's magnetic field.

Hollenbach said it is important to regard the core not as one large, controlled nuclear reaction, but as billions of smaller reactions that "shut down and resume" as waste products are created and expelled. Hollenbach and Herndon theorize that the aggregate fluctuations in the reaction's intensity are what cause the Earth's magnetic field to weaken every 200,000 years and change direction when it starts up again.

Further evidence of the georeactor comes from Hawaii and Iceland, Hollenbach said, where young lava basalts have been recovered that contain the helium isotopes He3 and He4. While He4 is a byproduct of the decay of natural uranium, He3 can only be produced deep within the Earth in a nuclear reaction.

Herndon's latest paper went even further, suggesting that the ratios of He3 to He4 indicated that the georeactor is reaching the end of its life -- albeit in perhaps a billion years. Hollenbach said he disagrees with that conclusion.

"Marvin likes to make sweeping grand statements without the data to back them up," Hollenbach said.

Herndon said the two parted ways because Discover magazine didn't pay enough attention to Hollenbach in an article last year: "There is a still a lot of bad blood," he acknowledged.

Anderson dismissed Herndon's latest conclusion, but he noted that French scientists showed in 1972 that a georeactor did exist for millions of years in a uranium deposit in Gabon. Most geophysicists believe that the He3 found in lava is "primordial" -- already there when the Earth congealed.

Also, Anderson noted, scientists have managed to flip the Earth's magnetic field in lab simulations that envision a cooling core. Finally, he said, Herndon's experiments use mathematical models based on data gathered from controlled nuclear reactions. "But there's no reason to put the uranium in the core," he said. "It does interesting things even if it's in the crust or upper mantle."