Duality gives a satisfying connexion between two different areas of strong interaction physics, Regge poles at high energy and resonances at low energy. This interlocking gives powerful bootstrap conditions, and together with the assumption that certain channels do not contain resonances it gives strong restrictions on the hadron spectrum. Since there is some confusion about the term duality, we shall explain what is meant by the various forms of duality (f. e. s. r. (finite energy sum rules) duality, local duality), and what is meant by ‘building up’, and we shall show in what way antidual models (such as the generalized interference model) come into conflict with basic empirical facts. Duality expresses the relation between two descriptions of the hadronic scattering amplitude. At low energy (l. e.) the description by direct-channel resonances is simple and useful (see figure 1). At low energy the data show prominent peaks as a function of energy, and one may try the approximation of resonance saturation, i. e. of neglecting the non-resonating background. The second description is the exchange of Regge poles, and it is useful at high energies (h.e.), where typical features are forward peaks, energy dependence s
α
, and structure at fixed
t
(see figure 2). The two descriptions are very different; resonance formation corresponds to poles in the
s
channel, Regge exchange to poles in the
t
channel. Duality says that there are direct relations between these two descriptions, that they are equivalent in a certain sense. In complete contrast, the interference models postulate that one must add the two descriptions. (If lowest order perturbation theory was relevant to strong interactions, one would be led to adding the diagrams.)
What is duality?
