Deep ocean circulation is dominated by two "rivers" of sinking seawater, Broecker explained. One deep-water source originates in the North Atlantic, the other off the coast of Antarctica. In this way all the water in the oceans is ventilated at the surface and mixed at depth about once every 800 years.

The mystery lies in the fact that different chemical tracer methods for measuring the rate of deep-water formation have yielded different results -- a faster rate over the long term and, more recently, a slower rate for the Antarctic source -- while other evidence suggests that deep ocean circulation runs at a steady rate, with roughly equal inputs from north and south over the 800-year mixing cycle.

Broecker's team turned to data on the pollutant chlorofluorocarbon-11, better known as freon, to double-check the recent rates of deep-water formation. Freon makes a perfect tracer, because it disperses quickly in the atmosphere, is highly soluble in water, can be measured very accurately, and was released only recently. Thus, the levels of freon in deep-water samples provide a measure of how much surface water has subsided to the depths over the past few decades.

The freon data was analyzed in a study led oceanographer Alejandro Orsi of Texas A&M University.

According to Broecker's reading of that study, the current rate of Antarctic deep-water formation is only one-third that of the North Atlantic, where the rate matches the expectations of a steady-state scenario.

The freon data also suggests that no other significant, as yet undiscovered deep-water source areas exist. Accommodating all the data suggests to Broecker that the rate of Antarctic formation must have changed dramatically over the last 800 years, from fast to slow.

"If I'm right, deep ocean circulation is not steady state process. That concept is mind-boggling, and certainly open to debate," Broecker said. "But it's reasonable to wonder whether deep ocean fluctuations could be related to fluctuations in global climate. The potential consequences of that are enormous, and provide a lot of fodder for scrutiny and further research."

Broecker believes that such a relationship does exist. After all, scientists have demonstrated that El Nino cycles are strongly tied to fluctuations in the patterns of shallow ocean currents.

Why wouldn't changes in deep ocean currents also express themselves climatically? He finds it tantalizing to look at the evidence for the Antarctic slow-down and consider the possibilities.

Of particular interest is the relationship between the new data and three well-established climatic phenomena.

• First, that about 12,500 years ago deep-water formation in the north virtually ceased during a pronounced cold event, while Antarctic deep-water formation accelerated in conjunction with a relative warming there.

• Second, that pronounced ice-rafting events related to cyclical warming trends occur in the North Atlantic every 1,500 years.

• And third, that a cooling event called the "Little Ice Age" occurred between 1350 and 1880, since which the planet has warmed in two roughly equal steps -- from 1880 to 1945 and from 1975 to present.

If true, regardless of the mechanism by which deep ocean currents affect climate, natural global temperature changes can be expected to occur on a millennial time scale. The question then becomes -- are we now in the midst of just such a natural temperature shift and, if so, what would today's climate be like in the absence of greenhouse warming?

"There's really no way we can tell at this point. It is just as likely that natural forces currently retard the effects of greenhouse warming as enhance it," Broecker said. "We don't know enough about these natural climate shifts or how quickly they occur to be able to predict."

Broecker does believe that the key to understanding humanity's role in the current global warming trend lies in unraveling the demise of the Little Ice Age -- what part of the two-step warming since 1880 is due to greenhouse gases versus natural forces. And he hopes that his paper will stimulate more interest and research in that direction.

That further research might disprove his notion of a deep ocean cycle tied to climate is fine with Broecker. However, to those of his colleagues who would attribute the current global warming trend entirely to such natural causes, he says nonsense.

"It is ridiculous to argue that greenhouse gases are having no effect," he said. "The preponderance of evidence that they are warming the planet is overwhelming, regardless of any underlying natural trend."