Although the breeding behaviour of Cichlid fishes in aquaria has been studied by several authors, very little is known about their behaviour in natural waters. This paper presents data from field studies on the breeding behaviour of T. karomo Poll and T. variabilis Boulenger. Tilapia species can be divided into three groups according to differences in their breeding behaviour; 'guarders', 'male mouth-brooders', and `female mouth-brooders'. Both the species considered here are female mouth-brooders. Breeding males of T. karomo congregate on the spawning grounds and each male establishes a territory wherein he prepares a nest. The males have a brightly coloured breeding dress, including a long genital tassel; they are slightly larger than the females, which are less brightly coloured. The nests are plaques of clean sand often raised on mounds; nest cleaning is done by 'mouthing', 'nosing' and `fanning' the nest; the other activities of the male before and between spawnings are described. Females cruise over the spawning grounds singly or in small shoals; the male swims out to meet them as they enter his territory, turns and leads back to the nest. The sequence of movements of male and female in the nest during spawning is described. Often there is very little pre-spawning display and spawning may be complete in two to five minutes. After laying, the female collects the eggs in her mouth and leaves the spawning ground; the male remains in his territory and immediately starts courting other females. T. karomo females stay among the waterplants when brooding. They probably have three or four broods in succession; it is not known for how long the young are brooded. The breeding behaviour of T. variabilis in Lake Salisbury (Uganda) is described and was found to be very similar to that of T. karomo. The similarities and differences in behaviour between T. karomo and T. variabilis are discussed. T. variabilis has a distinct piebald-and-orange colour form; nearly all these piebald fish are females. 'Normal' males must, it seems, breed with these. The areas which are suitable for spawning may be determined by the nature of the bottom and by the availability of dissolved oxygen. Young T. karomo live in shoals; they 'skittered' at the surface near the spawning ground, which may indicate that the oxygen content of the water was low at times. Nests cleaning activities in T. karomo suggest incipient 'behaviour forms'. The long genital tassel, present in T. karomo and T. variabilis, is not glandular, it may attract the attention of the female; it emphasizes movements and increases the apparent size of the male. Problems raised by aquarium studies are examined in the light of field observations. The long pre-spawning display in T. mossambica in aquaria as opposed to the rapid spawning of T. karomo in the field may be due to specific differences and not to aquarium conditions. 'Surface territories' present in aquaria were never seen in natural waters. In 'guarder' species of Tilapia the territory is used both for spawning and for the protection of the young. In mouth-brooders the spawning ground provides a well-advertised meeting place for ripe fish. The continuous succession of broods produced in aquaria is probably due to the confined conditions which make it impossible for the females to segregate from the males after spawning. Reference is made to the behaviour of T. macrocephala (= T. heudeloti) a 'male mouthbrooder'.