Prior to the widespread use of natural gas, manufactured gas plants (MGPs) produced the fuel for cooking and lighting by gasifying coal. The processes used resulted in the production of wastes that are now known to pose a threat to human health and the environment. Costs to actively remediate soil and groundwater contaminated with MGP wastes can range into the millions. Natural processes exist that decrease the mass, toxicity and/or mobility of the contaminants. Methods of assessing the potential for these processes to achieve remedial goals have previously been applied to petroleum hydrocarbon release sites. However, little work has been done to assess the potential for natural attenuation (NA) at MGP sites. The two main objectives of this research were to evaluate methods for assessment of NA at former MGP sites and assess the natural attenuation processes occurring at two study sites. Six characteristic compounds were selected for the assessment. Several methods were selected to assess the occurrence of NA processes at the two study sites. Statistical and graphical methods were used to evaluate historical data for direct evidence of contaminant attenuation. Indirect evidence was evaluated by designing a protocol to collect groundwater samples to analyze for geochemical indicators of natural attenuation processes. Indicators of natural attenuation include concentration of dissolved oxygen, nitrate, sulfate, iron, alkalinity, oxidation-reduction potential and pH. Computer modeling with both analytical and numerical models was explored. The results of this study indicate that dissolved contaminant plumes for six target compounds were approximately stable at the two sites. Qualitative evidence of biodegradation exists in the form of observed changes in geochemical indicators of natural attenuation corresponding to the observed hydrocarbon plume. Computer modeling indicates that attenuation (including biodegradation) is necessary to explain the observed plume size. Qualitative evidence indicated that natural attenuation was likely occurring. However, limitations imposed by the monitoring well network design and analytical error prohibits reliable quantitative estimates of attenuation rates.