THE JOINT PROBABILITY DISTRIBUTIONson between the theoretical curve and the experimental results obtained by (13) is shown in Fig. 2. The agreement between theory and practice is satisfactory. In particular, in spite of the presence of a heavy atom (C1), no serious deviation from the theory can be noted in structure (b). Furthermore, it is interesting to note that, owing to the presence of the factor 1/2N in the (EaEaIE1,E2) expression, a high percentage of quartets present A values crowded round the zeroThis behaviour is, therefore, enhanced in structure (c) in comparison with the structures (a) and (b).
ConclusionsIn this paper, as in the Harker-Kasper inequalities (3) and (4) and in the Woolfson relation (15), no use is made of the signs S(E2u)and S(E2k). E2h and E2k, nevertheless, are strongly correlated, as is well known, with the sign of the product Eh-kEh+k. Consequently, from a general point of view, the probability density P(Eh, Ek, Eh-k, Eh+k) worked out in this paper can be considered as a useful marginal probability function of the more exhaustive density function P (Eh, Ek,., Eh + k, E2,, E2k). This aspect of the problem will be considered in a further paper.