Reverse Osmosis (RO) uses semi-permeable membranes to remove contaminants from water, producing highly purified water from saline or polluted water. This technology is gaining increased attention for treating wastewater for potable reuse, an application that requires reliable removal of contaminants, such as carcinogenic N-nitrosamines, considered dangerous to human health, . N-nitrosamines, including the disinfection by-product N-nitrosodimethylamine (NDMA), can pass through RO membranes. Biofouling of RO membranes is a challenging problem in the application of RO technology for seawater desalination and water reclamation. Biofouling occurs when microorganisms that feed on nutrients in waters form a layer on the RO membrane surface, reducing water passage. Importantly, this process generally reduces rejection of contaminants by the membrane. Currently it is not known to what extent biofouling affects the passage of small, neutral organic compounds such as NDMA and by what mechanisms biofouling diminishes membrane performance. This thesis examined the effect of biofouling on the passage of five nitrosamines and evaluated the relative contributions of four proposed fouling mechanisms using physical models. The four models evaluated were: enhanced concentration polarization (ECP), hydraulic resistance (HR), surface masking (SM), and charge neutralization (CN). Fouling by surface masking is a newly proposed mechanism by which a certain fraction of the membrane area is completely blocked to water passage. Results show the changes in rejection of small neutral organic compounds caused by biofouling, quantify the contribution of concentration polarization to biofouling, and introduce surface masking as a biofouling mechanism to attribute the portion of biofouling not explained by ECP. In laboratory experiments, the transport of water, sodium chloride and nitrosamines through clean NF90 membranes were evaluated with a combined solution-diffusion concentration polarization (SD-CP) model. The sensitivity of the mass transfer coefficient, k, on modeling solute rejection was evaluated by estimating k in three ways: Sherwood estimation, Sutzkover estimation and Murthy Gupta fitting of k to the SD-CP model. The SD-CP model best modeled the rejection of small neutral organic solutes by the NF90 membrane by using k as a fitting parameter. Values of k obtained in this way were 1.5 times higher than those obtained by the traditional Sherwood approach. A technique was developed to build biolayers on RO membranes. Biolayer development occurred over four distinct phases: (1) lag-phase bacterial growth, (2) rapid bacterial growth, (3) slowed bacterial growth, and (4) stagnated bacterial growth due to depletion of the electron donor. Establishment of a biofilm of 24.3 [unknown symbol]m resulted in 75% flux decline, 2.5-fold increase in salt passage, and 13-30% increase in the passage of five different nitrosamines. These increases occurred primarily over bacterial growth phases (2) and (3). To evaluate concentration polarization caused by a biolayer, its porosity needs to be estimated. For realistic porosities (> 0.2), enhanced concentration polarization only marginally contributed to biofouling. Charge neutralization was not considered significant because changes in rejection of charged compounds (salts) and uncharged compounds (nitrosamines) were comparable. Although hydraulic resistance of exopolymeric substances (EPS) was not ruled out, the majority of the observed flux decline was attributed to surface masking.