This behaviour-genetic study concentrates on intra-specific aggressiveness in the three-spined stickleback (Gasterosteus aculeatus L., forma leiura). Aggressiveness was studied under standardized conditions in five different test situations, referred to as juvenile aggressiveness, female aggressiveness, territorial aggressiveness, courtship aggressiveness, or dominance ability. The aim of the study is two-fold: 1. To assess the extent to which variation of aggressiveness in each of the different test situations is attributable to genetic causes. 2. To assess the extent to which variation in these various manifestations of aggressiveness is influenced by common genetic factors. The paper starts with an analysis of the variation of aggressiveness in the base population, composed of individually isolated progeny of wild-caught parents (chapter 2). In each of the different test situations aggressiveness is highly variable across individuals. Repeated measurements with the same individuals as well as similarity of the levels of aggression between full sibs shows that phenotypic variation is to a considerable extent attributable to genotypic variation in each of the investigated situations. The genetic influence on variation of aggressiveness is further analysed with the aid of selection experiments (chapter 3). Bidirectional selection is exerted upon juvenile aggressiveness in juveniles of both sexes (juvenile aggression lines), upon territorial aggressiveness in males and female aggressiveness in adult females (territorial aggression lines), and lastly, upon dominance ability in males (dominance lines). Besides these six selection lines an unselected control line was maintained. Selection is highly successful in downward direction in each of the different contexts. However, enhancement of the level of aggression is less successful in most lines, with the exception of female aggressiveness. Possible causes for these asymmetries of responses are discussed in detail. It is argued that the lack of response in males to selection for territorial aggressiveness and dominance ability in upward directions is probably due to an upper selection limit for territorial aggressiveness and dominance ability, that is already present in the base population. The combined two-way responses yield heritability-estimates of at least 0.30 for aggressiveness in the different test situations. In the single selection lines the realized heritabilities range from 0 to at least 0.60. Across individuals of the base population the levels of aggression in the different test situations are correlated with each other to various degrees. In females there is a distinctly positive phenotypic correlation between juvenile and female aggressiveness, but in males correlations between juvenile, territorial and courtship aggressiveness are weaker. Dominance ability is uncorrelated with aggressiveness in the other test situations. Since fish of the various selection lines are not only scored for their aggressiveness in the particular situation in which they are selected, but also for their aggressiveness in the other test situations, these phenotypic correlations can be translated into genetic correlations (chapter 4). The genetic correlations between the levels of aggression in the different test situations are comparable to the corresponding phenotypic correlations. The genetic basis of juvenile aggressiveness is most likely identical for both juvenile males and juvenile females. In adult females variation of aggressiveness remains most likely governed by the same genetic factors. The genetic factors that contribute to variation of territorial aggressiveness are only partly identical to those that contribute to variation of juvenile aggressiveness. Changes of courtship aggressiveness parallel those of territorial aggressiveness in the various selection lines, suggesting that common genetic influences underlie variation in both manifestations of intra-specific aggressiveness in reproductive males. Lastly, variation of dominance ability is probably governed by genetic factors different from those governing variation of juvenile or territorial aggressiveness. Indirect determination of the level of androgens in reproductive males of the various selection lines, by means of kidney-size measurements, reveals that selection for territorial aggressiveness most likely acts on the level of androgens. In view of changes accompanying selection for juvenile aggressiveness (viz. a difference between high and low line fish in age at sexual maturity, in age at onset of juvenile aggression, and in incidence of ripeness) selection for juvenile aggressiveness likely acts on the level of gonadotropic hormones. Finally, selection for dominance ability acts on factors (probably androgen-independent) that influence the brightness of colouration. Throughout this paper the influence of a number of variables on aggressiveness in the various test situations, such as age, degree of ripeness, location of the nest, age of sexual maturity, experience, length, isolation and responsiveness, are analysed. Furthermore, attention is paid to threatening displays, to a peculiar behaviour called snapping at air-bubbles, to aggressiveness of juveniles in groups of fixed size, and to measures of aggressiveness other than the criterion applied during selection in a particular situation. Finally, the literature is reviewed with respect to selection studies on aggressiveness and to selective forces acting on aggressiveness in the three-spined stickleback.