A proposal has been made by Broecker et al. (1990) that rapid changes on a time scale of a thousand years or so, seen over much of the last major glacial in the Greenland ice core record, represent significant climate changes and are caused by a salt oscillator in the glacial Atlantic. This proposal is examined in terms of a rudimentary quantitative model. Scale analysis asserts that heat transported to the high-latitude atmosphere when the thermohaline circulation is turned on, is large enough to produce the melting rates found by Fairbanks (1989) for the time interval around that of the Younger Dryas event and that these melting rates are of the same order of magnitude as the mass flux associated with water vapor flux to the Pacific Ocean estimated by Broecker (1989). Scale analysis also suggests that the salinity fluxes associated with 1) the water vapor flux mechanism, 2) the rapid melting episodes of Fairbanks, 3) possibly ice sheet growth events, 4) net transport by the thermohaline circulation and 5) net transport by turbulent eddy mixing are roughly of the same order of magnitude and therefore may be important mechanisms for producing salinity oscillations on a time scale of a few thousands of years, (see Broecker, 1990). By integration of simple salt conservation equations, it is found that model oscillations with a period of a few thousand years occur over a significant range of salinity fluxes; estimated fluxes fall well within the range for which oscillations exist. The model also suggests that there may exist close coupling between the European-Scandinavian ice sheets and the bimodal response of the oscillator; that is, significant increases or decreases in continental ice volume may accompany each complete cycle of the oscillator. In addition, it appears that continental ice may be required for the salt oscillator to function. A crucial element, which cannot adequately be investigated with the present model, concerns the local effect of salinity source/sinks associated with melt water production. The proximity of these source regions on the neighboring ice sheets to the local regions where production of deep water occurs may play a critical role in the functioning of the proposed salt oscillator. In addition, further treatment of thermodynamics is needed to investigate the feasibility of a salinity driven oscillator.