JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact support@jstor.org.. Ecological Society of America is collaborating with JSTOR to digitize, preserve and extend access to Ecology. Abstract. A model for the turnover of organic matter in soil, ROTHC-26.3, can be used to calculate how much organic C needs to enter a soil annually in order to maintain a specified stock of soil organic C. The annual return of organic C thus calculated, plus the amount of organic C removed annually from the site by harvesting, burning, etc., provides an estimate of the Net Primary Production (NPP) of that site, averaged over many years. The new method was used to calculate NPP for two adjacent savanna sites in the Nairobi National Park in Kenya, one grazed and one not, and for a dry Miombo woodland site in Zambia. Both the Kenyan and Zambian sites are taken to be at equilibrium, with soil organic C levels at steady state. Soils from the three sites were analyzed by layer for organic C, 814C, 813C, soil microbial biomass C, total N, pH, and clay content. Radiocarbon measurements were >100% modern in the surface layers (0-15 cm) of the Kenyan soils (both Vertisols) and in all three layers (0-15, 15-30 and 30-50 cm) of the Zambian soil (an Oxisol), presumably because of 14C coming from the testing of thermonuclear bombs. The 15-30 cm layer of the Kenyan soils dated at -500 yr and the 30-50 cm layer at -900 yr. The 14C data were consistent with the presence of a small inert fraction of organic C that accounted for an increasing proportion of total organic C with increasing soil depth. The 13C data indicated that the Kenyan soils had developed under C4 vegetation, whereas the Zambian soils had developed under vegetation dominated by C3 plants.From these results the annual input of C to soil from the ungrazed Kenyan site was calculated to be 388 g C.m-2-yr-l, to the grazed site 380 g C.m-2.yr-l, and to the Zambian soil 373 g C.m-2.yr-'. Taking the loss of C from the Kenyan sites by burning to be 40 g C.m-2.yr-l, the mean NPP for both Kenyan sites is 424 g C.m-2.yr-'. This value for NPP is compatible with earlier estimates of NPP by botanical methods from the same site in Kenya. Woodtaking is thought to be minimal in the protected Zambian woodland, so that here the annual input of C to the soil can be taken as the NPP without great error. This new method provides a long-term, integrated measure of NPP that should complement and enhance productivity measurements made by harvest methods over shorter periods.