a consequence of Cenozoic cooling? (Brain, 1981b). This was an 

 early attempt to correlate some of the important episodes of human 

 evolution with climatic events, which formed part of a long-term 

 cooling trend over many millions of years. My attention had been 

 drawn to this particular topic in the course of our Swartkrans cave 

 investigation where we found that repeated cycles of erosion were 

 interspersed with ones of deposition. As it seemed likely that these 

 cycles were climatically mediated, I started to familiarize myself 

 with the published record of temperature changes and to speculate 

 on how they might have affected African habitats, as well as the 

 fauna and flora within them. It seemed as if some interesting cor- 

 relations were possible, not only for human evolution, but for that 

 of other mammals as well. At that time, Elisabeth Vrba was working 

 at the Transvaal Museum and she took up this topic with her usual 

 energy and enthusiasm; with various colleagues, she organised sev- 

 eral international conferences that resulted in many new publications 

 that clarified the issues involved. 



Early in the 20th century, geologists realized that a remarkable 

 record of the Earth's history had been preserved in deep-sea sedi- 

 ments, steadily accumulating in a protected environment under water 

 up to 3 km in depth. Samples of the sediments retrieved by deep- 

 sea drilling frequently contain the calcareous shells, or tests, of fo- 

 raminiferans, which had been studied taxonomically throughout the 

 century, and of which water temperature preferences were generally 

 known. It is also possible to say which species were planktonic, 

 living in surface waters, and which were benthic, living on the ocean 

 floors. So the various species present in a sediment core gave some 

 indication of the sea temperature at the time they lived. But a far 

 more precise method of estimating past sea temperatures could be 

 gained from determining the proportions of oxygen isotopes in the 

 shells of foraminiferans, a method first used by Cesare Emiliani 

 (1955). The oxygen in the calcium carbonate of foraminiferal tests 

 consists of two isotopes: oxygen 16, the more common one, and 

 oxygen 18, about 500 times less common. Thermodynamic theory 

 predicts that the ratio of these two oxygen isotopes in carbonate, 

 such as that of foraminiferal shells, will vary according to the tem- 

 perature at which the mineral was deposited. For this reason, mea- 



