<p> The separation of distilled water and coconut fatty acid (CNFA) was investigated in a 4 in. x 16 in. vertical gravity decanter to determine the most significant design variables. The design variables studied were tangential versus perpendicular feed flow, height/diameter ratio, the total feed flow rate, the temperature and the intertace position with respect to feed inlet. The criteria used were the calculated immiscible concentration of water in CNFA overflow (Cw) and that of CNFA in water underflow (Co), The test conditions were arranged according to a Central Composite Design based on Box Method of Experimentation. Two levels of inlet geometry and five levels of each of the other four variables were studied. The range of variables were 1. Inlet geometry Tangential and Perpendicular Feed Flow 2. Height/Diameter Ratio 3/1 to 4/1 3· Total Feed Flow rate 160 cc/min. to 320 cc/min. 4. Temperature 40° to 80°c. 5· Interrace Position 2" below feed inlet to 2" above feed inlet The Overall residence time in the decanter ranged from 7.5 to 12.5 minutes. The drop size distribution of the incoming dispersion was measured by taking photomicrographs for all the runs. </p> <p> It was found that the tangential feed flow was slightly better but not significantly better for the separation of two immiscible liquids. Statistical analysis revealed that 95% confidence level height/diameter ratio, flow rate and temperature, respectively, are the most important design variables for the separation of CNFA from water underflow. The dependence of Cw on height/diameter ratio of Co on the interface position was interpreted to mean that the theoretical, plug flow residence time within each phase is the most important design variable. Interaction between variables and second order constants were not found significant at 95% confidence level. Sedimentation rather than coalescence appeared to be the controlling factor in the present study. </p> / Thesis / Master of Engineering (MEngr)