Show description

72) coming from keeping the next term from the sinh and cosh in Eq. (65) are of order (qW)3. They can easily be included to all orders in a numerical treatment but give much too small an effect to account for the observed structure in P and Q. The only remaining assumption in Eq. (72) is that of equal geometries; hence, the assumption must be at fault. This surprising result-at high energies, equal spin-orbit and central geometries lead to structureless spin observables (or too far and too little structure)-has been obtained by a number of authors, including Auger and Lombard, (23) Osland and Glauber, (24) and Bertsch and Schaeffer, (25) but without leading them all to the cure in terms of geometry differences.

In this approximation of no distortion, an interesting energy sum rule (Ko 72) for the correlation cross section follows from a generally valid sum rule for a spectral density [see, for example, (FW 71)]. As we will see after including the initial- and final-state interactions, the form of Eq. (3) may be retained, but S Es(k3) is modified. This modified function is occasionally called the distorted-spectrum junction, although the connection with the hole propagator ceases to exist and the energy sum rule also, breaks down.

In Eq. (46), the firstphase factor arises from the asymptotic form of the Hankel functions, which, as we will see, leads to the Blair phase rule (recall, L + M is even). The factor of bl-1iq comes from the Tassie form. Had we used a less singular transition density, it would have been the value of that density evaluated at the stationary-phase point. However, because it is less singular, it would not have the factor of q. Thus, if both a Tassie-like and non-Tassie piece are present, the Tassie-like piece dominates at large q, accounting for the phenomenological dominance of the Tassie part.