We investigated the accuracy of the basic fusion data for the T(d, n)4He, 3He(d, p)4He, T(t, 2n)4He, D(d, n)3He, and D(d, p)T reactions in the 10- to 100-keV bombarding energy region, and assessed the effects of inaccuracies on the design of fusion reactors. The data base for these reactions /particularly the most critical T(d, n)4He reaction/ rests on 25-yr-old experiments whose accuracy (often assumed to be ±5%) has rarely been questioned: Yet, in all except the D + D reactions, there are significant differences among data sets. The errors in the basic data sets may be considerably larger than previously expected, and the effect on design calculations should be significant. Much of the trouble apparently lies in the accuracy of the energy measurements, which are difficult at low energies. Systematic errors of up to 50% are possible in the reactivity values of the present T(d, n)4He data base. The errors in the reactivity will propagate proportionally into the errors in fusion probabilities in reactor calculations. The 3He(d, p)4He reaction cross sections could be in error by as much as 50% in the low-energy region. The D(d, n)3He and D(d, p)T cross sections appear to be well known and consistent. The T(t, 2n)4He cross section is poorly known and may be subject to large systematic errors. Improved absolute measurements for all the reactions in the low bombarding energy region (10 to 100 keV) are needed, but until they are done, the data sets should be left as they are [except for T(t, 2n)4He data, which could be lowered by ∼50%]. The apparent uncertainties of these data sets should be kept in mind.