<p>Sexual selection arises through variation in reproductive success. This thesis investigates different aspects important in sexual selection, namely nest building, sperm competition, paternity and paternal care, and their mutual interrelationships. In the studied species, the sand goby (<i>Pomatoschistus minutus</i>) and the common goby (<i>Pomatoschistus microps</i>), sperm competition did arise when small males, so called sneakers, sneaked into other males nests and released sperm. They seemed to use female behaviour as their prime cue for a sneaking opportunity. However, also nest-holders, both with and without eggs, were found to fertilize eggs in the nests of other males.</p><p>Clearly, nest-holding males tried to prevent other males from spreading their sperm in their nests, since they showed aggression towards such males. A nest building experiment indicated that the small nest-openings found in the sneaker male treatment were sexually selected through protection against sneaking or by female choice. Yet, no behavioural or genetical support for the hypothesis that the nest functions as a physical or visual defence, or that sneaker males prefer to sneak upon nests with wide nest-openings, were found in the other studies. Still, individual nest-holding males showed a higher mucus preparation effort inside the nest in the presence of a sneaker male than when alone. In close relatives, such mucus contains sperm, suggesting an importance in sperm competition. However, the mucus may also have pheromone and anti-bacterial functions and may constitute a mating effort, as found in other gobies. Both a behavioural and a mate choice experiment suggested that the males were not less eager to spawn in the presence of a sneaker male.</p><p>Sneak intrusion did not affect nest defence, fanning or filial cannibalism, nor had paternity an effect on filial cannibalism. This and various life history aspects, together with the fact that the parasitic male only fertilized a fraction of the clutches, would predict females to ignore sneaker males. This was also the case, as the presence of sneaker males was found not to affect female spawning decision. Still, several females spawned in two nests, which coincided with parasitic spawnings, suggesting a cost of disturbance for the females and thus a substantial cost to the nest-holding males in terms of lost mating success. However, females paid attention to other traits in their choice of mate since spawning was associated with sand volume of the nest, but not with nest-opening width. Also, female (but not male) courtship was correlated with partial clutch filial cannibalism, indicating that females are able to anticipate future male cannibalism.</p><p>In a partial correlation of nest opening, sand volume, male courtship display, displacement fanning and male size, a large number of traits were correlated both positively and negatively with regard to how we may expect them to be appreciated by females. For instance, males which fan well also build large nests or display intensely (but not both). Together with all the other results of this thesis, this shows the entangled selection pressures working on breeding animals, as well as the different male and female tactics employed to maximize their reproduction.</p>

Sexual selection arises through variation in reproductive success. This thesis investigates different aspects important in sexual selection, namely nest building, sperm competition, paternity and paternal care, and their mutual interrelationships. In the studied species, the sand goby (Pomatoschistus minutus) and the common goby (Pomatoschistus microps), sperm competition did arise when small males, so called sneakers, sneaked into other males nests and released sperm. They seemed to use female behaviour as their prime cue for a sneaking opportunity. However, also nest-holders, both with and without eggs, were found to fertilize eggs in the nests of other males. Clearly, nest-holding males tried to prevent other males from spreading their sperm in their nests, since they showed aggression towards such males. A nest building experiment indicated that the small nest-openings found in the sneaker male treatment were sexually selected through protection against sneaking or by female choice. Yet, no behavioural or genetical support for the hypothesis that the nest functions as a physical or visual defence, or that sneaker males prefer to sneak upon nests with wide nest-openings, were found in the other studies. Still, individual nest-holding males showed a higher mucus preparation effort inside the nest in the presence of a sneaker male than when alone. In close relatives, such mucus contains sperm, suggesting an importance in sperm competition. However, the mucus may also have pheromone and anti-bacterial functions and may constitute a mating effort, as found in other gobies. Both a behavioural and a mate choice experiment suggested that the males were not less eager to spawn in the presence of a sneaker male. Sneak intrusion did not affect nest defence, fanning or filial cannibalism, nor had paternity an effect on filial cannibalism. This and various life history aspects, together with the fact that the parasitic male only fertilized a fraction of the clutches, would predict females to ignore sneaker males. This was also the case, as the presence of sneaker males was found not to affect female spawning decision. Still, several females spawned in two nests, which coincided with parasitic spawnings, suggesting a cost of disturbance for the females and thus a substantial cost to the nest-holding males in terms of lost mating success. However, females paid attention to other traits in their choice of mate since spawning was associated with sand volume of the nest, but not with nest-opening width. Also, female (but not male) courtship was correlated with partial clutch filial cannibalism, indicating that females are able to anticipate future male cannibalism. In a partial correlation of nest opening, sand volume, male courtship display, displacement fanning and male size, a large number of traits were correlated both positively and negatively with regard to how we may expect them to be appreciated by females. For instance, males which fan well also build large nests or display intensely (but not both). Together with all the other results of this thesis, this shows the entangled selection pressures working on breeding animals, as well as the different male and female tactics employed to maximize their reproduction.

This study assessed the importance of overwash and breaching events on the ichthyofaunal community structure in the medium-sized temporarily open/closed Mpekweni Estuary located on the southeast coastline of southern Africa. The fish in the littoral zone of the estuary were sampled using a 5m seine net while the channel region was sampled using two nets, a smaller meshed 30m seine net to target the estuarine spawning species and the juvenile estuarine-dependant marine spawners, and a larger meshed 50m seine net to target the larger marine and freshwater spawning species. Intensive monthly sampling over two years provided data on selected physico-chemical and biological parameters. During the two year sampling period from November 2005 to October 2007 the estuary breached in late July 2006 and remained open till the sandbar re-formed across the mouth in late December 2006. Thus, sampling encompassed three open/closed phases 1) initial closed period, 2) open period and 3) re-closed period after the berm was re-formed. The open period was divided into two phases 1) the out-flow phase and 2) the tidal phase. A total of 36 fish species representing 19 families were sampled using the various seine nets employed during the investigation. In the littoral zone, the estuarine spawners (Estuarine Utilisation Category, [EUC] I), mainly the Gobiidae, Glossogobius callidus, dominated the samples numerically and by biomass. The smaller estuarine spawning species sampled in the channel were numerically dominated by Gilchristella aestuaria in conjunction with two other EUC I species, Atherina breviceps and G. callidus. The estuarine-dependant marine spawners (EUC II), however, dominated the ichthyofaunal biomass of the channel. The abundance and biomass of the larger species targeted were dominated by estuarine-dependant marine spawning species (EUC II), principally Rhabdosargus holubi. During the closed periods of the estuary, total fish abundance and biomass displayed weak seasonal patterns. The breaching event and subsequent open period was associated with a decrease in the total abundances of fish in the littoral zone and channel of the estuary, reflecting the out-flow of estuarine biomass-rich water into the marine environment. The breaching event coincided with a shift in the community composition of the ichthyofauna, reflecting the recruitment of marine spawning species into the estuary. Numerical analysis identified two distinct spatial fish communities within the estuary, a community associated with the mouth region and one comprising the rest of the estuary. The absence of any further spatial patterns in the ichthyofaunal community structure within the Mpekweni Estuary appear to be ascribed to the virtual absence of horizontal patterns in physico-chemical and biological parameters recorded in the system. Cohort analyses were employed to determine possible recruitment events for selected estuarine and marine spawning species. The estuarine spawning species displayed continuous recruitment patterns throughout the study, which appeared to be unaffected by the breaching event. Conversely, the larger marine spawning species displayed multiple cohorts, indicating non-continuous recruitment. Subsequent retrospective analysis of cohorts for the different species identified summer recruitment peaks that coincided with the breaching event and open period. Minor recruitment of marine spawning species also occurred during overwash events. The recruitment of ichthyofauna into the estuary was quantified during three distinct hydrological events: overwash, out-flow phase immediately after breaching and tidal phase during the period when the mouth was open. Estimates of fish recruitment were highest during the outflow phase immediately after the estuary breached and declined as the estuary became tidally inundated with marine water. Although not as high as the outflow and tidal phases, recruitment was evident during overwash events. Results of the current study highlight the importance of both breaching and overwashing events in structuring the ichthyofaunal community composition in a medium-sized southern African temporarily open/closed estuary. These results are broadly in agreement with similar studies conducted both locally and in other regions of the world.

Les Sicydiinae amphidromes constituent une part importante des peuplements piscicoles insulaires de la région Indo-Pacifique et sont vulnérables à de nombreuses pressions anthropiques (i.e. pêcherie, dégradation et fragmentation des habitats). La reproduction a lieu en eau douce, puis les jeunes dévalent en mer où ils débutent leur croissance pendant quelques mois avant de coloniser les rivières. L’objectif est d’acquérir des connaissances sur les traits de vie de S. lagocephalus (cosmopolite) et de C. acutipinnis (endémique). Les travaux de terrain démontrent que le choix de l’habitat quotidien est fortement lié aux interactions sociales, alors que l’habitat de fraie montre une forte sélection pour des conditions morphodynamiques favorisant l’oxygénation des oeufs. Les mâles sélectionnent des sites de ponte et gardent les oeufs. L’étude expérimentale de la survie en eau douce confirme que tous les embryons libres possèdent la capacité à rejoindre la mer. L’analyse des traits de vie marins à partir des otolithes des post-larves révèle des variations saisonnières d’âge et de taille au recrutement, en lien avec la croissance et la température marine. L’examen histologique des ovaires permet de décrire les variations spatiales et temporelles de l’activité de reproduction chez les femelles en rivière et de déterminer la fécondité et la taille de première reproduction. Les variations inter et intra-spécifiques des traits de vie sontcomparées en lien avec la répartition géographique des deux espèces. Ces travaux permettent d’identifier des perspectives concrètes dans le domaine de la gestion et de la conservation sur la base des caractéristiques de la stratégie amphidrome. / Amphidromous Sicydiinae are particularly widespread among fish freshwater assemblages of the Indo-Pacific region and are vulnerable to anthropogenic pressures (i.e. fishery, degradation and fragmentation of habitats). Spawning and hatching happen in freshwater, then offspring drift downstream to the sea where they begin their growth for a few months before recruiting in the rivers. This study aims at acquire knowledge about life traits of S. lagocephalus (cosmopolite) and C. acutipinnis (endemic). Field sampling shows that the usual habitat choice is related to social interactions, whereasspawning habitat shows a strong selection for morphodynamic conditions which favor eggs oxygenation. Males select spawning sites and care for the eggs. The experimental study of freshwater survival confirms that free embryos survival and swimming capacities are enough to drift to the ocean. Marine life history analysis, based on postlarvae otoliths examination, reveals a seasonal variation of both age and size-atrecruitment related to the larval growth rate and the sea temperature. Histological examination of ovaries is used to describe the spatial and temporal variation of reproductive activity of mature females in the rivers, and to determine fecundity and size at first reproduction. Inter and intra-specific variations of life history traits are compared in relation to the respective geographical distribution of both species. This study identifies concrete prospects for management and conservation based on the features of amphidromous life history strategy.