Although sexual selection often explains the evolution and maintenance of a single male ornament, it is unclear how the multicomponent nature of most sexual displays evolves. Theoretical models suggest that handicap signalling should converge on a single most informative quality indicator, whereas additional signals are more likely to be arbitrary Fisherian traits, amplifiers, or exploitations of receiver psychology. Despite the predictions that multiple handicap signals are unlikely to be stable, the male nuptial plumage of the highly polygynous (ca. 3, but up to 9 actively nesting females) red-collared widowbird Euplectes ardens comprises two classic quality-indicating avian ornaments (handicaps); a long graduated tail (22 cm) and a red carotenoid throat patch (collar). To investigate the evolution and maintenance of these handicaps in the red-collared widowbird, a population was studied in the Hilton district, KwaZulu-Natal, South Africa, where these small (males ca. 20 g) African weaverbirds (Ploeceidae) inhabit a grassy valley during the breeding season. Multivariate selection analyses, used to investigate net, direct and indirect female selection, demonstrated an unusually strong fitness effect of natural tail length (47%) on male reproductive success (the order and total number of nests acquired). There were no other effects of morphology, ectoparasite load, display rate, territory quality, tail asymmetry or collar measures (hue, brightness and size) on reproductive success, suggesting that females base their mate choice exclusively on only one handicap signal, extreme tail elongation. Although ignored in mate choice, there is strong evidence that the variable carotenoid collar (ranging from yellow to red) functions as a status or agonistic signal in male contest competition for territories. Compared to 'floating' males (that did not establish a territory in the area), resident males had a 60% larger and 23 nm 'redder' (longwave hue) collar. Model presentations also corroborated the status signalling function of the collar as territorial males were less aggressive towards conspecific models with intense collar displays, and males with greater carotenoid investment responded more aggressively to the models. In captive experiments, non-breeding brown males painted with red 'collars' dominated orange painted, control brown, novel blue collared and testosterone-implanted males in dyadic contests over food resources. In addition, experiments in the field demonstrated that males manipulated with larger and redder collars established and maintained territories in the area, whereas most males with small, orange or blackened collars failed to establish or retain territories. Thus the size and particularly redness of the costly carotenoid collar reliably signals male status and fighting ability in male contests. The unique negative phenotypic relationship between the expression of tail length and carotenoid pigmentation, suggests strong overlapping developmental costs (and allocation conflicts) between the two handicap ornaments. This tradeoff is predicted to be strongest between signals with the same or similar costs. Although current theory predicts that multiple handicaps should be evolutionary unstable, the coexistence of multiple costly ornaments in the red-collared widowbird is stable because of selection by different receivers, females and males (i.e., multiple receivers). / Thesis (M.Sc.)-University of Natal, Pietermaritzburg, 2001.

Sexual selection is usually invoked to explain the evolution of elaborate epigamic characters in animals. However, the mechanism by which female choice operates is poorly understood, and it is not clear whether female choice is purely aesthetic or related to male genotypic quality. It has been suggested that Moller's fluctuating asymmetry (FA) hypothesis may resolve the 'arbitrary trait'-'good gene' debate. However, tests of this controversial hypothesis have yielded equivocal results. I examined the allometric patterns of FA in the redcollared widow in order to test the FA hypothesis. In addition, I documented intrapopulation variation in trait size to determine whether females could distinguish between males on the basis of ornament size and/or symmetry. Male tail length was found to be more variable in size than other traits, suggesting that sufficient variation exists in ornament size for females to distinguish between males on this basis. In addition, the prediction of the FA hypothesis that ornaments would display higher degrees of asymmetry than non-ornamental traits was supported. However, no significant linear or second-order polynomial relationship was found between trait size and asymmetry for tail, wing and tarsus length. Furthermore, neither tail length or symmetry was correlated with indices of body condition. Although this might suggest that the FA hypothesis is invalid, I argue that the assumptions of the hypothesis are too simplistic for it to hold true for all species without taking the basic biology of the species into account. / Thesis (M.Sc.)-University of Natal, Pietermaritzburg, 1997.