Theoretical and experimental results for the fine-structure separation of the lowest 3 P state of the helium atom disagree significantly. The experiment is well checked and the disagreement is most likely due to deficiency in the theoretical evaluation, which is based upon power expansion of the Bethe-Salpeter equation [6], this yields the separation n 01 = 29 616 946.4 (2) kHz, which deviates significantly from the experimental results. Adjusting the fine-structure constant so that the experimental and theoretical separations coincide, yields the value α −1 = 137.035 987 2(18), which deviates >7 SD from the CODATA value. The latter value is largely based upon the anomalous magnetic moment of the electron, g − 2, measured for a single electron in an ion trap by the Seattle group and the corresponding QED calculations by Kinoshita [7]. This result is supported by measurement of h/m, using photon impact by Chu [6]. Very recently the Chu result has been confirmed by the Paris group (F. Nez, private communication, 2005). The calculations of Kinoshita are also well checked, and the only remaining explanation of the discrepancy seems to be that the theory of the helium fine structure is in error or at least incomplete.The accurate treatment of the helium fine structure is a complicated theoretical problem, involving (i) strong electron correlation, (ii) higher-order QED effects, as well as (iii) quasi-degeneracy. No numerical technique is presently available that can handle all these problems in a systematic fashion. The standard many-body perturbation technique (MBPT) can handle strong correlation, particularly