The processes controlling total carbon (C) storage and release from the terrestrial biosphere are still poorly quantified. We conclude from analysis of paleodata and climate biome model output that terrestrial C exchanges since the last glacial maximum (LGM) were dominated by slow processes of C sequestration in soils, possibly modified by C starvation and reduced water use efficiency of trees during the LGM. Human intrusion into the C cycle was immeasurably small. These processes produced an averaged C sink in the terrestrial biosphere on the order of 0.05 Pg yr-1 during the past 10,000 years. In contrast, future C cycling will be dominated by human activities, not only from increasing C release with burning of fossil fuels, and but also from indirect effects which increase C storage in the terrestrial biosphere (CO2 fertilization; management of C by technology and afforestation; synchronous early forest succession from widespread cropland abandonment) and decrease C storage in the biosphere (synchronous forest dieback from climatic stress; warming-induced oxidation of soil C; slowed forest succession; unfinished tree life cycles; delayed immigration of trees; increasing agricultural land use). Comparison of the positive and negative C flux processes involved suggests that if the C sequestration processes are important, they likely will be so during the next few decades, gradually being counteracted by the C release processes. Based only on tabulating known or predicted C flux effects of these processes, we could not determine if the earth will act as a significant C source from dominance by natural C cycle processes, or as a C sink made possible only by excellent earth stewardship in the next 50 to 100 yrs. Our subsequent analysis concentrated on recent estimates of C release from forest replacement by increased agriculture. Those results suggest that future agriculture may produce an additional 0.6 to 1.2 Pg yr-1 loss during the 50 to 100 years to CO2 doubling if the current ratio of farmed to potentially-farmed land is maintained; or a greater loss, up to a maximum of 1.4 to 2.8 Pg yr-1 if all potential agricultural land is farmed.