Radiocarbon dating of East african lake levels.

The fluctuations of the key East African lakes discussed are summarized in Fig. 4 which also includes the available evidence from Lake Rukwa (42) and Lake Chad (43) Exceot for Lake Victoria, all of these now lack surface outlets and are situated in much drier climates than the major lakes of the Western Rift Valley, which remain filled to their overflow levels. The apparent differendes among the fluctuations of the lakes are partly due to differendes in the nature of the evidence or the intensity of research or both, although there must also have been important local differences in the histories of the lakes Yet the consistencies are far more striking, most notably the coincidence of early Holocene high stands. Between 10,000 and 8,000 years ago, it seems that lakes in many parts of tropical Africa were greatly enlarged. Where evidence for the previous span of time is well resolved, it appears that transgressions leading to this high stand began about 12,000 years ago, and evidende from three basins (Victoria, Nakuru, and Chad) indicates a pause or minor recession just at or before 10,000 years ago. Wherever information is available for the period preceding 12,000 years ago, it can consistentlybe shown that lakes were much small-er . Several basins (Rudolf, Nakuru, and Chad) also show traces of much earlier phases of lake expansion. which are not yet well dated but which all occurred more then 20,000 years ago. The Holocene record subsequent to the maximum of 10,00 to 8,000 years ago is more complex. Three basins (Rudolf, Nakuru, and Chad) show an apparently concordant, positive oscillation at some point between 6000 and 4000 years ago, but it is uncertain how widely this episode is represented. Although many of these lakes that are now closed filled to overflowing at least once during the late Quaternary, it is evident from Fig. 4 that the periods of expansion were short-lived compared with phases of contraction to levels near those of today. This pattern may be in accord with fragmentary evidence from lower and middle Pleistocene formations, such as those of Olduvai(44)and Paninj (45), within which some relatively short-term lake expansions can be documented, but which lack evidence for any marked long-term departure from a balance of evaporation and precipitation similar to the present one Further, this pattern of brief moist pulsations, with a duration of perhaps 2000 to 5000 years, is also suggested by other late Pleistocene and Holocene sequences (based primarily on geomorphological and palynological evidence) from the Saharan area, Angola, and South Africa (46). In default of radiometric dating, such complex successions of relatively brief moist intervals provide few stratigraphic markers of broad applicability. This, together with the fact that vegetation, weathering processes, montane glaciers, lake size, lake salinity, and so forth are all likely to reflect the diverse aspects of Climatic change differently, underscores the strictures of Cooke (2) and Flint (3) against the use of pluvials and intrlvasas a basis for subdividing Quaternary time in Africa. Positive correlations between high-latitude glacial advances or maxima and intervals of high lake levels have been demonstrated or suggested for many areas of mid-latitude North America and Eurasia (47), and similar patterns have often been regarded as probable for tropical Africa as well. However, the evidence summarized above shows a notable lack of such correlations for the tropical lakes considered here. If glaciation and tropical lake levels were connected at all, then a far more complex-delayed, multiplefactor, or inverse-relationship must be sought for the late Quaternary (48). This renders the introduction of new climato-stratigraphic terms such as hypothermal and interstadial (49) of questionable value in East Africa. Further, whereas the so-called pluvial lakes of higher latitudes were probably due primarily to reduced evaporation (50), our computations for the early Holocene lakes Nakuru and Naivasha, as well as for the oscillations of Lake Rudolf and Lake Victoria in recent decades, suggest that many or most of the high tropical lake levels where associated with a modest but significant increase in precipitation.