c ­ • { -1.6 a(K+p) > 0, a(K~p) < 0 c - -1.6 is ruled out because it gives a(K-p) - -0.65. Hence we conclude that a(K+p) < 0 and a(K-p) J. V. Allabye data. pOints, for a{K+p) were plotted under this assumption in Fig. 6. The errors are clearl¥ very large, and allow us to safel¥ ignore the subthreshold singularities. The general features of 1EP dispersion relations appear also in Kp. The logarithmic behavior is magnified because 2a{"')... 4: mb. However, at present energies the bulk of the real part seems to come from the sub­ traction term, and not from the logarithmic one. In fact, these appear to have opposite signs. Thus we expect Ia I to actuall¥ fall until very high energies, when a changes signs and Ia I begins to grow again. As in 1EP, the real part does not dominate untU extremel¥ high energies. The difference 'between the pion and the kaon amplitudes lies in the energy range below the cutoff point. The usual Regge picture -- which assumes the Pomeranchuk theorem to hold...... is compatible with experiment for the pions, but appears not to be so for the kaons. In the latter case, the existence of an additional real term seems to be implied by the data. ---...•............. ~- ~,. -12­ 1. lHEP-CERN Collaboration D. Horn~ al., Physics Letters roB, 500 (1969). 2 G. HShIer, G. Ebe1, and J. Giesecke Caltech report CALT-68-228 (to be published). 3

4. s Phys. Rev. 159, Bl6:38

5 A. Citron et al., University of California at Riverside report UCR-MP107­ 105 (1969). 6 Hly's. Rev. 144, llOl

Phys. Rev. !etters !b 505 (1965) Group, L. R. }bysics Letters 28B, 615 (1969). We use the fit of the ParticleDa.ta K. J. FoleyPrice ~ al., UCRL 20000 K+N (September 1969). ll

et a1., Phys. Rev. Letters!§., 45 (1965) M cds.JOURNAL>Phys.Lett.</cds.JOURNAL> B24 454