The principal models in vogue today for dolomitization are the mixing zone and the sabkha models. Despite the wide acceptance of these models, there has been little critical assessment of their validity. Such an assessment is the objective of the present paper. A close look at the "Dorag" mixing-zone model (Badiozamani 1973) reveals several serious weaknesses: 1) the range of freshwater-seawater mixtures that meet the Dorag requirement for dolomitization shrinks to a small "window" if the geologically more realistic disordered dolomite is used in the calculations instead of the ordered dolomite on which the model is based; 2) in none of the known modern coastal mixing zones in limestone or lime sediments has replacement by dolomite been observed; 3) in a number of dolomites interpreted to be of mixing-zone origin, dolomite has precipitated without dissolution of the calcite substrate, evidence that negates the fundamental premise of the Dorag model. In addition to these weaknesses, isotope and trace-element data used to identify mixed-water dolomite are inadmissible because 1) isotopic fractionation factors for dolomites remain unresolved; 2) isotope values (uncorrected for temperature) for a host of dolomites, interpreted to be of different origins, overlap; 3) nonisomorphous trace elements, such as Na, in dolomite cannot, on theoretical grounds, be relied on to identify dolomitizing fluids. Similar overall objections can be brought against Folk and Land's (1975) "schizohaline" version of mixing-zone dolomitization. In summary, mixing-zone models have such weak underpinnings that they should be questioned as viable explanations for massive dolomitization. Contemporaneous dolomite formation in modern sabkhas is well documented, but the important question of whether the mechanism of dolomite formation is replacement or direct precipitation remains to be resolved. In the well-studied sabkha at Abu Dhabi, brine chemistry changes have been used as evidence of a replacement origin for the dolomite, but it is shown here that this evidence is far from unequivocal. An alternative origin of direct precipitation of dolomite is offered, an origin in keeping with dolomite precipitation known from other modern sabkhas and saline lakes, and in line with our laboratory experience with dolomite synthesis. Moreover, it is suggested here that, in general, contemporaneous dolomite will form at low temperatures only by direct precipitation, a mechanism that requires special conditions of highly super-saturated waters of high Mg/Ca ratio and elevated CO 3 -HCO 3 concentrations. This may explain why modern dolomite is mainly restricted to evaporitic environments, a bias not shared by ancient dolomites. In contrast, replacement dolomite appears to require, at low temperatures, long reaction times > or = 10 4 yr?), a requirement that is mainly to be met in large, regional groundwater-flow systems, both marine and nonmarine. A third dolomitization model considered here is that of Baker and Kastner (1981), based on the experimental finding that sulfate ions inhibit or retard dolomitization. A number of modern sedimentary dolomites are not in accord with this model in that they are forming from brines with large sulfate concentrations, 2 to 70 times that of seawater. Thus, the Baker-Kastner model should be held in abeyance until these serious contradictions are resolved. The current emphasis on mixing-zone and sabkha dolomitization has diverted attention from other promising avenues of approach to the dolomite problem. Four of these avenues, each of which deemphasizes the "special water" approach, are briefly addressed and are as follows: 1) influence of temperature and time; 2) mass transfer processes; 3) burial diagenesis of epigenetic dolomites; 4) fluid-inclusion studies. At elevated temperatures the dolomite problem essentially disappears (ordered dolomite can be made in the laboratory in days at 100 degrees C). What is more, at temperatures above 60 degrees C, Ca-rich waters become dolomitizing fluids, which makes most natural subsurface waters capable of dolomitization. At low temperatures, time may be the key element, so that seawater will become a major dolomitizing fluid only where stable circulation systems, such as Kohout convention, can drive seawater through carbonate platforms for many thousands to millions of years. This paper has tried to show that currently favored models of dolomitization carry serious uncertainties, enough to warn us to look more critically at the validity of these models, and also, it is hoped, enough to spur new efforts to find new solutions to the problems of dolomitization.