Geophysicist Louise Kellogg studies the Earth's mantle, a vast internal realm so hot that, even though it is solid rock, it circulates like a pot of impenetrable soup. "That motion of the mantle is driving the surface motion of the continental plates, it's driving mountain building, it's driving earthquakes and hot spots, like volcanoes-all that interesting geology at the surface," Kellogg says. "All the energy that goes into them is provided by the motion of the mantle. If you want to understand the rise of the highest point on the planet, you have to look into the deep Earth." That's not easy, because the mantle is almost wholly inaccessible-"as remote as the far reaches of outer space," says Kellogg. It begins about 22 miles below our feet and continues for 1,800 miles, but no one has drilled more than about 10 miles into the Earth. So mantle analysts must rely on indirect observations, and the resulting division of opinions has become the hottest debate in geophysics. At the geophysicists' annual meeting, "People will get up with this kind of picture, or they'll get up with that kind of picture, and they'll alternate, and they'll yell at each other and scream and get all heated up. It's amazing," Kellogg says. "After years of watching this, I thought, well, we need a new approach." Heat convection and fluid dynamics So Kellogg attacked the problem with her expertise in heat convection and computer modeling of fluid dynamics. Now, at age 39, just 10 years out of Cornell University and seven years after receiving one of the first $500,000 Presidential Faculty Fellowships, Kellogg is helping to forge the first understanding of the forces at the middle of the Earth. Her most important research report to date, which merges the existing conflicting opinions about mantle movement into one new explanation, appears today in the prestigious journal Science. To get a sense of the Earth's interior, think of the planet as a basketball. Its surface is the crust-the thin, rubbery sphere bearing the continents and ocean basins. Inside is another ball, filling slightly more than half the basketball. This is the molten outer core, where temperatures start at 5,000 degrees Fahrenheit. And within that ball is the tennis-ball-sized inner core, blazing away at 12,000 degrees but solid because it is under so much pressure. Between the crust and the outer core lies the mantle. It is solid rock, and yet heat (from the core and from its own internal radioactive decay) sets it to slowly flowing like a liquid. Its flow rate is unknown, but it is probably similar to that of the crustal plates-1/3 inch to about 3 1/3 inches a year. The big question is: What is the circulation pattern of that ever-moving rock mass? On one hand are seismologists-geophysicists specializing in the study of earthquakes. The seismologists say quake records suggest that when crustal plates collide and send slabs of surface rock diving down into the mantle, these "cold sinkers" travel very deep, perhaps all the way to the outer core boundary, before they recirculate throughout the whole mantle. History and snapshots merged On the other hand are geochemists-geologists specializing in the study of the chemistry of rocks. The geochemists say rock samples from "hot spots"--mid-ocean ridges and volcanic islands--show that the cold sinkers encounter an impassable geological barrier at about 420 miles down. They rebound and, driven by convection, circulate in the upper mantle. After hundreds of millions of years, the plate material re-emerges at the hot spots. "This is where I come in. I try to merge the time history of the geochemist and the snapshot of the seismologist," says Kellogg, who is specifically a geodynamicist-a geophysicist who studies how physical properties, such as convection and fluid dynamics, influence the Earth's interior. In her Science paper, Kellogg and two other geophysicists from the Massachusetts Institute of Technology suggest that there is indeed a geological barrier in the mantle. It starts at about 1,140 miles down, deeper than the seismologists thought, but it can produce the seismic records they have. When the sinking plate reaches this barrier, it is deflected and begins mixing throughout the mantle. Eventually, it will re-emerge at mid-ocean ridges. And, at superhot spots on the surface of that lower mantle layer, some slab material and a bit of deep-mantle material rise together in a plume, forming a volcano. That would satisfy the geochemists' findings. Jeff Mount, chair of the UC Davis geology department, says the cross-disciplinary thinking that produced the Science paper is typical of Kellogg's work, and one of the keys to her success. "I think the big breakthroughs in science now are taking place at the bridges between disciplines. The big leaps come when you integrate," Mount says. "Louise has taken that extra step of learning the fundamentals of specialties outside her own so that she can build that bridge between them. That takes a tremendous amount of creativity and a whole lot of talent." Almost an artist There was a time when those qualities almost led Kellogg into a different field of endeavor. In high school, she planned to be an artist or historian, until a calculus class made her think it would be more fun to be a scientist. Even when she entered Cornell's master's program in applied physics, she was still drawn to the performing arts. "I was very interested in modern dance-Merce Cunningham, Mark Morris, Alvin Ailey and so forth," Kellogg recalls. So while she studied fluid dynamics, she also explored dance theory. "Unfortunately, I don't have the talent or physical build required to dance seriously. The requirements are like those for a professional athlete." She eventually discovered that her knowledge of dance informed her research. "The thought processes are the same, although the specific skills are very different," Kellogg explains. "Dance involves visualization of complex patterns in space and time to create a unified whole. Geophysics requires extracting patterns from complex and sometimes seemingly unrelated data. "All good science involves this, but for a complex, messy system like the Earth, it's especially tricky." Those patterns exist in another of Kellogg's current research projects-a study of creeping crustal deformation around the White Wolf Fault and its neighbors in the Southern California desert. As in her mantle studies, Kellogg is looking for relationships. Geological forces are the dancers and Kellogg is the reverse choreographer, recording each movement to reveal the secret workings of the planet's inner spaces.