Most emission estimates of the greenhouse gas nitrous oxide (N2O) from African agriculture at a continental scale are based on emission factors, such as those developed by the IPCC Guidelines. Here we present estimates from Africa from the EDGAR database, which is derived from the IPCC emission factors. Resulting estimates indicate that N2O emissions from agriculture represented 42% of total emissions from Africa (though that rises to 71% if all savannah and grassland burning is included), or roughly 6% of global anthropogenic N2O emissions (or 11% including burning). Emissions from African agriculture are dominated by grazing livestock; 74% of agricultural N2O excluding biomass burning was from paddocks, ranges, and pasture. Direct soil emissions represent 15% of agricultural emissions; substantial changes in direct emissions from North Africa helped drive a 47% continental increase in direct soil emissions from 1970 to 2005. Future trends based on the Millennium Ecosystem Assessment scenarios indicate that agricultural N2O emissions may double in Africa by 2050 from 2000 levels. Any regional or continental estimates for Africa are, however, necessarily limited by a paucity of direct measurements of emissions in sub-Saharan agro-ecosystems, and the heavy reliance on emission factors and other default assumptions about nitrogen cycling in African agriculture. In particular, a better understanding of livestock-related N inputs and N2O emissions will help improve regional and continental estimates. As fertilizer use increases in sub-Saharan Africa, emission estimates should consider several unusual elements of African agriculture: farmer practices that differ fundamentally from that of large scale farms, the long history of N depletion from agricultural soils, seasonal emission pulses, and emission factors that vary with the amount of N added.