FIGURE 3 Oxygen isotope records of planktonic and benthic foraminifera. Reprinted from Shackleton and Opdyke (1973), with permission from Academic Press, Inc.

provided at least one absolute calibration point for estimating average rates of sediment accumulation.

Nick Shackleton, a British marine geologist, was the foremost figure in promoting another proxy for climate change, stable isotopes. Working in England, he used a high-resolution mass spectrometer to analyze the down-core oscillations in the ratio of the heavy oxygen isotope, ¹⁸O, to the light oxygen isotope, ¹⁶O. Based on the correlation with the biostratigraphy, these variations were clearly correlated with changing climate, but it was unclear whether the isotopic variations were caused by changes in ocean temperature or in terrestrial ice volume. With the encouragement of NSF, Shakleton became the first international corresponding member of NSF's CLIMAP program, which sought to decipher Earth's paleoclimate during the last glacial maximum. U.S. researchers were intrigued by Shakleton's stable isotope work, and Shakleton badly needed better samples on which to work. He had been using samples collected 100 years earlier by the HMS Challenger! Under CLIMAP sponsorship, Shakleton came to the United States and worked on core V28-238, a high-resolution core in the Lamont data bank collected by the Vema from the Ontong Java Plateau (Figure 3). This core contained well-preserved benthic and planktonic foraminifera, which showed the same oxygen isotopic signal. The argument was that whereas surface waters are very prone to temperature changes, the deep sea is roughly isothermal. Therefore, the fact that the signal was the same in the surface waters as the deep sea argued that the ultimate cause was climate-related changes in ice volume, not temperature directly.

The impact of the development of the stable isotope proxy on paleoceanography was substantial. On the assumption that sedimentation rates were constant throughout the entire Bruhnes epoch, the oscillations in the stable isotopes became the paleoclimate equivalent of the magnetic reversals for plate tectonics. The pattern could be used for global correlation. But unlike the magnetic reversal signal, which defies prediction and is likely an excellent example of chaos, there was a pattern to the variations in the oxygen isotopes. In 1976, Hays at Lamont, working with Imbrie at Brown and Shackleton, applied spectral techniques to the signals from cores that were thought to be fairly well dated such that the isotopic signal as a function of depth could be accurately converted to a time series. The result was the identification of spectral peaks that matched the predictions of the Milankovitch hypothesis (Figure 4). According to this theory, variations in Earth' s orbital parameters (eccentricity, tilt, and precession of the equinoxes) caused variations in solar insolation that resulted in changes in climate.

Although there was some cause to question how well core depth had been converted to time, the strength of the spectral peaks and the repeatability of the pattern won many converts—so much so that now cores with poor age control are assigned dates by assuming that the isotopic peaks and troughs should correspond in time to what is predicted by the Milankovitch hypothesis ("orbital tuning"). Not all is completely understood, however. For example, northern and southern hemispheres would be predicted tc be out of phase for the precession period, but they are not. Overall, phase relationships demonstrate that regional insolation is not important. The net effect on the whole globe ,with its unequal distribution of continents and oceans must be taken into account. In addition, the strength of the spectral peaks is not consistent with the hypothesis that it is variations in solar