Using an experimental pressure chamber the rates of descent and the maximum depths of buoyancy equilibrium of some cichlid fishes were tested. Tilapia mossambica juveniles could descend to depths greater than 30 metres very rapidly. At 22°C small sub adult male fish (length ± 160 mm) could descend to a maximum buoyancy depth of 20 metres, small sub adult females to 13 metres while large adults (length ± 200 mm) of both sexes could descend to 11 metres. The fish generally take 4 to 5 days to reach their maximum equilibration depth. At 30°C mature adult male and female T.mossambica increased their maximum equilibration depth to 20 metres and at 15°C showed a decrease to a maximum depth of 6.7 metres. Temperature affects the rates of descent in a similar manner. T.rendalli can descend to 7.5 metres at 22°C. T.sparmanii can descend to 15 metres at 22°C. Hemihaplochromis philander fry and young juveniles are not affected by pressure. Adult males can descend to 16 metres at 22°C and 20 metres at 30°C. Adult females can descend to 26 metres at 22°C and 27 metres at 30°C. In T.mossambica oxygen appears to be the only gas secreted into the swimbladder to compensate for buoyancy loss. The haemoglobin of T.mossambica showed a marked Bohr and Root effect. Increased rates of descent with increased temperature are believed to be due to increased O₂ uptake, heart output etc. rather than as a direct effect of blood chemistry. The swimbladder wall of T.mossambica is extremely thin (2.8 ∕∪ m) and consequently a large amount of gas is lost through passive diffusion. Diffusion is not affected by temperature, consequently the increased secretory rates at higher temperatures are not balanced by a proportionately higher diffusion loss, thus enabling the fish to maintain a deeper buoyancy equilibrium at higher temperatures. Summary, p. 62-63.

Allopatric populations of Pseudocrenilabrus philander were found to differ in male breeding coloration. These colours are described and attention is drawn to the possibility that male coloration might prevent population hybridisation. Behaviour of the allopatric populations was studied in the laboratory. The execution of behaviour patterns of the various populations was found to be indistinguishable, and very similar to that of Haplochromis. Behaviour of P. philander was recorded in three morphologically different natural water bodies. Diurnal activities were found to follow a set rhythm. Though P. philander was shown to have adapted to a variety of habitats they retained their behavioural Components in an unchanged form. Field and laboratory observations are used to propose a behavioural mechanism for the natural regulation of population density. Although aggression between conspecific territorial Sarot herodon mossambicus habituated so that neighbours could live close to one another, it was found that P.philander are intolerant of their neighbours and are consequently forced apart. Though the execution of behaviour patterns was apparently identical for all populations of P. philander, field observations indicated that the frequency of performance differed. An experimental procedure was established to quantitatively compare the behaviour of various populations. As a result of comparative and choice chamber investigations, it is suggested that one of the four populations of P. philander would be unlikely to hybridise with the others if they were to become sympatric. These popUlations would be kept apart because of colour and behavioural differences. Differences of egg-size, fry-size, developmental and interbrood periods were also found, providing further evidence of incipient speciation. Evolutionary divergence of the allopatric populations of P. philander is discussed against a background of geological and geographical evidence. Consideration is given to the role of male coloration, ethological barriers and the importance of ethological differences to taxonomy. Behavioural evidence suggests that P. philander has closer affinities with the haplochromids than was initially realised. It is concluded that the clinal populations of P. philander should not be taxonomically separated, and it is argued that it would be premature to give the Kuruman population an elevated taxonomic status.

A lek-breeding cichlid fish Pseudocrenilabrus philander was studied experimentally. Females in choice-chamber experiments showed no active choice for male size and colour, or for other male attributes, but preferred males which courted most actively. In a laboratory lek, the significant determinants of the strongly skewed mating success in males were territory size, side-shake and female chasing frequency. Together these three variables explained 79% of the variation in male mating success, with territory size alone accounting for 75% of the variation and the other two variables each accounting for 2% of the remaining variation. As there was no difference in territory quality in the laboratory lek, territory size became the principal measure of the effect of male-male competition since it was directly related to dominance. Both side-shake and female chasing could be identified as the basic factors influencing female choice, as they had an immediate effect on the display-response mating system of females. Thus, the relative importance of these three variables indicated that sexual selection in this particular lek mating species operated chiefly through the agency of intrasexual competition for dominance. However, both female behaviour and their requirement for a prolonged pre-spawning courtship had the effect of promoting male rivalry and favouring mating with dominant males. Although the intense male competition excluded subordinate males from practising normal courtship behaviour, competitively inferior males might "make the best of a bad situation" by facultatively adopting an alternative sneaking tactic to gain access to females. Spawning intrusions by females to steal freshly-laid eggs also occurred frequently. However, territorial males appeared to be relatively more tolerant of female intruders than male sneakers. Interference during spawning could lead to a longer pre-spawning courtship and even multiple-mating by females. The results of the present study and the behavioural evidence shown by males and females did not support the ʺrunaway selection modelʺ of the mating pattern in terms of sexual selection in leks, but conformed to the rival ʺwar propaganda modelʺ