Three-neutrino mixing schemes suggested by Cardall & Fuller and Acker & Pakvasa are compared and contrasted. Both of these schemes seek to solve the solar and atmospheric neutrino problems and to account for the possible neutrino oscillation signal in the LSND experiment. These neutrino oscillation schemes have different atmospheric and solar neutrino signatures that will be discriminated by Super-Kamiokande and SNO. They will also have different signatures in proposed long-baseline accelerator and reactor experiments. In particular, both of these schemes would give dramatic (and dramatically different) signals in an "intermediate baseline" experiment, such as the proposed ICARUS detector in the Jura mountains 17 km from CERN. © 1997 Elsevier Science B.V.