3~o t hMonte Carlo-generated events for particular reactions a s well a s the real data, where it is possible to check this, indicate that there is little difference, generally less than 0.5 units, between 17 and the real longitudinal rapidity y = 4 ln[(E +p , , ) / ( E -prl)l.2"he different shapes of the data near v = 0 and near v = -6 result from the use of the pseudorapidity and the angle errors on the fast forward tracks.Measurements of ntp, K*p, pp, andpp elastic scattering a r e presented for incident momenta of 3, 3.65, 5, and 6 GeV/c and momentum transfers typically 0.03 to 1.8 Gev2. The q l e and momentum of the scattered particle were measured with the Argonne Effective Mass Spectrometer for 300 000 events, yielding 930 cross-section values with an uncertainty in absolute normalization of k4 %. Only the K + and proton data show any significant change in slope of the forward diffraction peak with incident momentum. The particle-antiparticle crossover positions are consistent with no energy dependence, average values being 0.14 10.03, 0.19010.006, and 0.162 +0.004 G~V ' for n's, K ' s , and protons, respectively; these e r r o r s reflect both statistics and the *1.5 % uncertainty in particle-antiparticle relative normalization. Differences between particle and antiparticle cross sections isolate interference terms between amplitudes of opposite C parity in the t channel; these differences indicate that the imaginary part of the odd-C nonflip-helicity amplitude has a ~~(r(-t)"') structure for -t ~0 . 8 Gev2, as predicted by strong absorption models. The cross-section differences for K * and proton-antiproton a r e in qualitative agreement with the predictions of w universality, the agreement improving with increasing energy. The corresponding quark-model predictions relating the n* and K* differences failed by more than a factor of 2. We have combined our r * cross sections with other data to better determine the nN amplitudes in a model-independent way; results of this analysis a r e presented.