We present constraints on the number of relativistic species from a joint analysis of cosmic microwave background (CMB) fluctuations and light element abundances (helium and deuterium) compared to big bang nucleosynthesis (BBN) predictions. Our BBN calculations include updates of nuclear rates in light of recent experimental and theoretical information, with the most significant change occuring for the d(p,gamma)^3He cross section. We calculate a likelihood function for BBN theory and observations that accounts for both observational errors and nuclear rate uncertainties and can be easily embedded in cosmological parameter fitting. We then demonstrate that CMB and BBN are in good agreement, suggesting that the number of relativistic species did not change between the time of BBN and the time of recombination. The level of agreement between BBN and CMB, as well as the agreement with the standard model of particle physics, depends somewhat on systematic differences among determinations of the primordial helium abundance. We demonstrate that interesting constraints can be derived combining only CMB and D/H observations with BBN theory, suggesting that an improved D/H constraint would be an extremely valuable probe of cosmology.