This dissertation centers around an empirical and mathematically oriented analysis of conflict events in 19 Latin American countries in the period 1946-1975. The events are three types of protest (riots, demonstrations, and political strikes) and one type of coercion (mass arrests). The study is divided in seven chapters, i.e., (1) Collective conflict as a concept (7 pp.); (2) Theories of collective conflict (28 pp.); (3) Models of collective conflict (15 pp.); (4) Methodology (13 pp.); (5) Preliminary data analyses (16 pp.); (6) Collective conflicts in Latin America (37 pp.); and (7) Summary and conclusions (12 pp.). Much of the evidence is presented in 19 tables and 9 figures. A FORTRAN program which was used for computations is listed in an appendix. In Chapter 1 regimes and oppositions are identified as parties in conflict. Mobilization processes that are going on within these parties are distinguished from the confrontation process between the parties. In Chapter 2 a set of theoretical distinctions (i.e., strain theory, control theory, cultural deviation theory, conflict theory, and social learning theory) is borrowed from delinquency theory to summarize the findings of prior research on collective conflict. In Chapter 3 three mathematical models are treated, which describe conflict events: (1) a linear model, closely related to Richardson's arms race model, (2) a perceptual model, based on Hamblin's arms race model, and (3) a nonlinear model, in which elements of (1) and (2) are combined. In Chapter 4 a discussion is presented of the problems and potentials of using cross-temporal as well as cross-national data to estimate the mathematical models. In Chapter 5 some preliminary issues are settled. One of these issues is that the Cuban Revolution did not cause a structural change in other Latin American countries. The difference between the slope parameters in the periods 1946-1959 and 1960-1973 is statistically not significant. In Chapter 6 it is shown mathematically that forms of dissident behavior and governmental repression have the tendency to return to the same level of equilibrium over and over again. In "direction fields" of the riot-arrest system, it is illustrated how these equilibrium levels are reached through time. In Chapter 7 the findings are evaluated and suggestions made for future research. This study of political instability can not escape agreement with the observation that, indeed, "Latin American history is a kind of Eternal Recurrence."