The dynamics of the California Current system are studied using two ocean models, one with a shelf and one without. Both models are viscid and linearized about a background density field Pb(Z), Solutions are forced by steady and annually periodic winds with and without curl, and by an idealization of the observed wind field off California. Solutions forced by a steady, equatorward, curl-free wind τy all have an equatorward surface coastal jet and a poleward undercurrent. Due to the β-effect and horizontal mixing, the circulation is not necessarily confined within a Rossby radius of the coast. The strength and structure of the currents vary considerably with parameters, the currents being stronger and broader when the forcing includes remote winds to the south and when ρb has a near-surface pycnocline. If τy oscillates at the annual cycle the response is qualitatively quasi-steady, but it also involves a poleward, offshore and vertical propagation of waves, and the maximum coastal current leads τy by several weeks.Solutions forced by a steady, positive wind curl τyx develop a deep, broad, poleward surface current near the coast, consistent with Sverdrup theory. Interestingly, there is also an equatorward surface flow located farther offshore, which exists because of the vertical mixing in the model. Solutions are not very dependent on model parameters, because they are primarily interior currents directly in balance with the wind curl and do not require the coast for their existence. If τyx oscillates at the annual cycle, the response is not at all quasi-steady, and the maximum coastal current lags τyx by 1–2 months. Solutions forced by an idealization of the observed wind field off California compare favorably with observations, but only if Pb has a realistic pycnocline and the forcing includes remote winds off Baja California. Forcing by positive τyx accounts for both the poleward Davidson Current during the winter and the equatorward flow located more than 100 km offshore throughout the year. The coastal jet forced by τy provides summertime equatorward flow within 100 km of the coast that is strong enough to reverse the poleward flow driven by τyx.