Previous analyses of civil war trends tend to be informal and consider only post 1945 data. We examine data on civil wars over the period 1816–2005, using new methods for evolutionary growth processes. We find a number of new patterns and trends in civil war that have received little attention in previous research, including a structural break in frequency of conflict with decolonialization, as well as evidence of periodicity in civil conflict. We develop new measures of civil war intensity and impact, and find that conflicts have been generally more severe in the 20th than in the 19th century. We also find that the frequency-severity distribution of civil war does not appear to follow a power-law distribution, unlike data on many other types of conflict. Although structural trends suggest an increase in future civil wars, we discuss possible limiting factors that might prevent this in light of the recent observed decline in civil wars after the Cold War.