Stickleback males, which are held solitary under experimental conditions described above, give a more or less constant aggressive response toward a test-male (aggression test) when tested once or twice daily during a 4-day period following nest-construction. A distinct waning of this response is found in such males that have a continuous view of another male in a neighbouring territory behind a glass partition ('rival situation') : the frequency of biting decreases and becomes nil or fluctuates about a low level. Rival males and controls do not differ in the response toward a test-female (sex test). Therefore, rival males are said to have attained to a 'hypo-aggressive' condition. The stimulus situation that produces the waning of aggressiveness appears to have some specific properties, e.g. a goldfish neighbour does not lead to the same results. Waning of aggressiveness is also found toward the neighbour if it is used as a test-male. Waning is qualitatively independent of the presence of the glass partition, it occurs also in situations where neighbours can interact freely. Neighbours need not have a continuous view of one another to become hypo-aggressive, a (frequent) intermittent view leads to the same results. We conclude from our experiments that a recovery of aggressiveness, if present, takes place slowly (as a matter of days). A further analysis of the rival responses leads to the separation of three types: I, an unmodified response (similar to controls) ; II, a hypo-aggressive response without an increase of sexual activities toward the test-male; III, a hypo-aggressive response with an increase of sexual activities in addition. A large part of the type III males do not seem to discriminate a test-male and a test-female, the numbers of bites and zigzags in both test situations being nearly equal. Nest-building, or perhaps the presence of a complete nest, is a relevant factor for the occurrence of hypo-aggressiveness and the increase of sexual activities toward a test-male. The interaction of the aggressive and sexual tendencies is found to be entirely consistent with SEVENSTER'S (1961) results. In controls, biting and zigzagging result from the same aggressive and sexual tendencies irrespective of whether they are shown in aggression or in sex tests, and both tendencies are mutually inhibitive. The same conclusion is drawn for the interaction of aggressive and sexual tendencies in rivals. Though the interactions take place on quantitatively different levels, the relations of the tendencies remain qualitatively similar.