One of the most concerning trends associated with the ongoing problem of civil wars is that conflicts often flare-up a short period after they appeared to have ended. While significant progress has been made in the study of post-civil war peace building and the causes of civil wars, the tendency for civil wars to recur is one factor which has been largely overlooked. This thesis addresses this shortcoming by analysing the causes of recurring civil war using statistical methods. Relevant civil war research was consulted and hypotheses pertaining to the variables which might influence civil war recurrence were formulated. These factors are organised in a contingency framework which suggests that conflict recurrence is dependent on both pre- and post-conflict environments as well as factors associated with how the original conflict was fought. The Uppsala Conflict Data Program/Centre for the Study of Civil War Armed Conflict Dataset was used to produce a dataset of 238 civil wars which were fought between 1946 and 2004. Additional data pertaining to specific hypotheses was collected from a range of other sources. Statistical analysis was conducted to determine the strength and direction of relationships between different variables and civil war recurrence. Several factors were found to have a significant relationship with civil war recurrence: ethnic diversity, conflicts which were fought over territorial issues and conflicts which were not ended by military victory, particularly those which ended as a result of low or no fatalities. These findings are discussed with reference to improving civil war management and policy recommendations are presented.