We investigated the source mechanism of the 1975 Kalapana, Hawaii, earthquake (MS=7.2) by modeling the tsunamis observed at three tide-gauge stations, Hilo, Kahului, and Honolulu. We computed synthetic tsunamis for various fault models. The arrival times and the amplitudes of the synthetic tsunamis computed for Ando's fault model (fault length=40 km, fault width=20 km, strike=N70¿E, dip=20¿SE, rake=-90¿, fault depth=10 km, and slip=5.6 m) are ~10 min earlier and 5 times smaller than those of the observed, respectively. We tested fault models with different dip angles and depths. Models with a northwest dip direction yield larger tsunami amplitudes than those with a southeast dip direction. Models with shallower fault depths produce later first arrivals than deeper models. We also considered the effects of the Hilina fault system, but its contribution to tsunami excitation is insignificant. This suggests that another mechanism is required to explain the tsunamis. One plausible model is a propagating slump model with a 1 m subsidence along the coast and a 1 m uplift offshore. This model can explain the arrival times and the amplitudes of the observed tsunamis satisfactorily. An alternative model is a wider fault model that dips 10 ¿NW, with its fault plane extending 25 km offshore, well beyond the aftershock area of the Kalapana earthquake. These two models produce a similar uplift pattern offshore and, kinematically, are indistinguishable as far as tsunami excitation is concerned. The total volume of displaced water is estimated to be ~2.5 km3. From the comparison of slump model and the single-force model suggested earlier from seismological data we prefer a combination of faulting and large-scale slumping on the south flank of Kilauea volcano as the most appropriate model for the 1975 Kalapana earthquake. Two basic mechanisms have been presented for explaining the deformation of the south flank of Kilauea: (1) pressure and density variation along the rift zone caused by magma injection and (2) gravitational instability due to the steep topography of the south flank of Kilauea. In either mechanism, large displacements on the south flank are involved that are responsible for the observed large tsunamis. ¿ 1999 American Geophysical Union