Climatic and environmental changes in Africa during the last 2 centuries have been examined, using both systematic rainfall records and proxy information concerning lakes and rivers and the occurrence of famine and drought. The rainfall records provide excellent detail for the 20th century. The proxy data have been used to produce a semi-quantitative data set spanning most of the continent and having an annual time resolution. These provide an overview of conditions during the 19th century. Various issues related to the causes of these 2 centuries of variability are also considered: atmospheric and oceanic processes, desertification, surface albedo, mineral dust and hydrological feedbacks. The most significant climatic change that has occurred has been a long-term reduction in rainfall in the semi-arid regions of West Africa. This was on the order of 20 to 40% in parts of the Sahel. There have been 3 decades of protracted aridity. Nearly all of Africa has been affected by increased aridity, particularly since the 1980s. Few changes in temperature have been demonstrated. These have occurred on a much smaller scale and are of considerably lower magni- tude than those over the continents. The rainfall conditions over Africa during the last 2 to 3 decades are not unprecedented. A similar dry episode prevailed during most of the first half of the 19th cen- tury. By mid-century, conditions more typical of the 'normal' for the current century again prevailed. Thus, the 3 decades of dry conditions evidenced in the Sahel are not in themselves evidence of irre- versible global change. On the other hand, the processes controlling rainfall over most of the conti- nent are now reasonably well understood. One of the most important factors, particularly in the Sahel, is sea-surface temperatures. It has been hypothesized that anthropogenic changes in the land surface, particularly land use change and desertification, have contributed significantly to the decline in rainfall. Current evidence suggests that if changes in the land surface (e.g., vegetation cover, sur- face albedo, soil moisture) signficantly impact climate, they are much more strongly controlled by natural climate variations, such as the recent decline in rainfall, than by human-induced land-use change or degradation. Unfortunately, we still do not have any accurate large-scale assessments of the extent, nature and degree of such changes. The dreaded 'desertification' process appears to be confined to relatively small scales. However, there is still cause for concern because the net effect of the various feedback processes involved in land degradation appears to be destabilization of ecosys- tems. Thus, a priority must be large-scale monitoring of the land surface and estimates of the degree of change.