We reviewed papers that compared intraspecific variation in territoriality vs. alternative forms of spatial or behavioral organization with three goals: (1) to discover which ecological variables act as determinants of territorial behavior and how they might act; (2) to extract and evaluate predictions and evidence for determinants of territoriality and (3) to suggest ways for future studies to build upon what the review revealed. Twenty ecological variables have been predicted, correlated with or experimentally demonstrated to relate to territoriality within vertebrate species. These variables include several characteristics of food: quantity, predictability, distribution, quality, renewal rate, type, density and assessibility. Other variables include nonfood resources, population density, habitat features, mates, space, refuges/spawning/home sites, predation pressure, host nests (for brood parasites) and energy availability. We suggest several reasons why food resources are cited most often, including their biological significance, ease of study and publishability of negative results. Certain groups of animals lend themselves to certain methods of study and, therefore, constrain the variables measured. Many variables are the subjects of apparently contradictory reports, i.e., some papers report that an increase in a given variable increases territoriality and others report that a decrease in the variable increases territoriality. After summarizing these reports we hypothesized U-shaped relationships between the ecological variables and behavior that could accommodate all these findings. However, these hypotheses cannot be tested rigorously by most current studies because of methodological limitations. We recommend a shift to quantification of intraspecifically varying spacing systems combined with simultaneous quantification of several ecological variables. Relative importance of different determinants of particular spacing systems can be revealed via multiple regression analysis. Hypothesized causal pathways, in which one ecological variable determines another variable that, in turn, determines territoriality, can be tested by path analysis.