This paper measures and compares total factor productivity growth in African agriculture under contemporaneous and sequential technology frontiers over the period 1970–2004. The paper further investigates the sources of agricultural productivity growth using a fixed-effects regression model and a second-degree polynomial distributed lag structure for agricultural research. The conventional estimates show an average productivity growth rate of only 0.3% per year over the period 1970–2004. In contrast to conventional measures, however, the improved measures under sequential technology show that agricultural productivity grew at a higher rate of 1.8% per year. Technical progress, rather than efficiency change, was the principal source of productivity growth. Agricultural research has turned out to have positive and significant impacts on productivity. The estimated productivity elasticity with respect to agricultural research is 0.04 and suggests that doubling research investments would lead to a 4% increase in total factor productivity. Consistent with the induced intensification hypothesis, population pressure has a positive and significant effect on agricultural productivity. We find a negative and significant relationship between rainfall variability, confirming that drought is a major constraint to agricultural production in Africa.