Neogene dolomites are common in island settings, and they have been used to understand the complex processes of dolomitization. The island of Bonaire was one of the first locations where the concept of reflux dolomitization was applied. Given this historical significance, here we re-evaluate the importance and nature of reflux using a previously unstudied set of outcrops in Bonaire. Mio-Pliocene units in Seru Grandi in northwestern Bonaire show well-defined bodies of dolomite that extend along clinoform surfaces beneath a subhorizontal erosional unconformity. The dolomite distribution suggests early dolomitization as a result of fluids moving downwards through a succession of subtidal facies, as expected from the reflux model. However, non-stoichiometric compositions, weak 18O enrichment, low trace element concentrations and the sedimentological context of the dolomites suggest that refluxing fluid was more likely to be mesohaline in composition rather than hypersaline as previously proposed. In addition, our findings suggest that the fluids did not dolomitize the succession uniformly, but rather elongate bodies of dolomite were developed within selected clinoforms. This variability may be due to temporal changes in the flux and chemistry of reactive fluids and textural changes of original sediment, whereas downdip trends may reflect evolution of magnesium exchange efficiency with distance from the brine source and reactivity of the rock.