Thesis (MSc)--Stellenbosch University, 2012. / ENGLISH ABSTRACT: The decline of Haliotis midae (perlemoen) populations together with the ensuing collapse of commercial abalone fisheries in South Africa have shifted the responsibility to abalone farms to meet the demand for perlemoen. Attention has recently turned to the genetic enhancement of cultured abalone in order for the farms to remain competitive in the international aquaculture market. To develop a successful breeding programme it is imperative to draw on a good foundation of high levels of genetic diversity and to successfully maintain these levels in order to create an enhanced strain of cultured abalone. A Performance Recording Scheme (PRS) was established as the first breeding programme for Haliotis midae to utilise molecular tools. This programme was aimed at enhancing the growth rate of abalone in order to shorten the production times on farms. The current study made use of 12 species-specific microsatellite markers to assign parentage to a group of faster-growing PRS animals, as selected by the abalone farms, in order to select a diverse on-farm generation of broodstock. Additionally, the influence of standard selection practises on the genetic diversity of a population compared to genotypic selection was investigated. This data was also used to study the differentiation and levels of genetic diversities within and between cultured and wild populations. Selection based on genotypic traits successfully retained genetic diversity while some diversity was lost in phenotypically selected populations. These phenotypic populations differed significantly from each other and wild populations, while the genotypic populations were similar in genetic composition to each other and wild populations of the West coast. The broodstock populations used in the PRS spawning event were representative of the wild populations from where they were sourced, with no significant differentiation between the broodstock and West coast population. When these broodstock populations were compared to their corresponding offspring populations, only two populations displayed a significant loss in diversity; although all of the offspring populations showed significant differentiation with their corresponding broodstock populations. This was attributed to the differential contribution of broodstock and the effect of artificial selection. It was established that the cultured populations of the participating abalone farms should be used with caution in ranching and reseeding programmes. These populations differed significantly from both the East and West coast wild populations. This study concluded that it is possible to retain genetic diversity by selecting breeding animals based on genotypic traits. The loss of diversity in some cultured populations and significant differentiation from the wild populations indicate that animals are exposed to different selection pressures in the cultured environment. The results found in this study highlight the need for the effective management of hatchery practices and the genetic monitoring of the breeding animals. / AFRIKAANSE OPSOMMING: Die afname in Haliotis midae (perlemoen) populasies en die daaropvolgende ineenstorting van die kommersiële perlemoen bedryf in Suid-Afrika het die verantwoordelikheid om in die aanvraag na perlemoen te voorsien, na perlemoen plase verskuif. Die genetiese verbetering van verboude perlemoen geniet tans aandag in ‘n poging om kompeterend te bly in die internasionale mark. Dit is noodsaaklik vir die sukses van ‘n broeiprogram om gebruik te maak van ‘n goeie genetiese basis met hoë vlakke van genetiese diversiteit en die suksesvolle behoud van die vlakke om so ‘n verbeterde lyn te skep. ‘n Groeiprestasie aanteken stelsel [Performance Recording Scheme (PRS)] is gestig as die eerste broeiprogram vir Haliotis midae wat gebruik maak van molekulêre tegnieke. Die doel van hierdie program was om die groeitempo van verboude perlemoen te verbeter om produksie tye te verkort. Die huidige studie het gebruik gemaak van 12 spesie-spesifieke mikrosatelliet merkers om ouerskap toe te ken aan ‘n groep vinnig-groeiende PRS-diere, soos geselekteer deur die perlemoen plase, om ‘n diverse generasie gekultiveerde diere te selekteer wat as broeidiere kan dien. Die invloed van standaard seleksie metodes op die genetiese diversitiet van ‘n populasie in vergelyking met genotipiese seleksie is ook ondersoek. Die ouerskap data is ook gebruik om differensiasie en vlakke van genetiese diversiteit tussen verboude perlemoene en wilde populasies vas te stel. Seleksie gebasseer op genetiese eienskappe het daarin geslaag om genetiese diversiteit te behou, terwyl diversiteit verlore gegaan het in die fenotipies geselekteerde populasies. Hierdie fenotipiese populasies het ook beduidend met mekaar sowel as met die wilde populasies verskil, terwyl genotipiese populasies soortgelyk was in hul genetiese samestelling en nie van die wilde populasies van die Weskus verskil het nie. Die broeidiere wat in die PRS broeiprogram gebruik is, was verteenwoordigend van die wilde populasies vanwaar hulle oorspronlik gekom het, met geen beduidende differensiasie tussen die broeidiere en die Wes kus populasies nie. Met die vergelyking van die broeidiere en hul ooreenstemmende nageslag, het dit geblyk dat slegs twee populasies ‘n beduidende verlies aan genetiese diversiteit getoon het, alhoewel al die nageslag beduidende populasie differensiasie met hul ouers getoon het. Hierdie bevindinge is toegeskryf aan oneweredige bydraes van die broeidiere tydens gameetvrystelling en die invloed van kunsmatige seleksie. Hierdie studie het ook vasgestel dat die verboude perlemoen populasies met sorg gebruik moet word om wilde populasies te herstel, aangesien hierdie populasies beduidend verskil het van wilde populasies van beide die Oos en Wes-kus. Hierdie studie het gevind dat dit moontlik is om genetiese diversiteit te behou deur diere te selekteer op grond van genotipiese eienskappe. Die verlies van diversiteit in sommige van die verboude perlemoen populasies en die beduidende verskil met die wilde populasies dui daarop dat diere in die gekultiveerde omgewing blootgestel word aan verskillende tipes seleksiedruk. Hierdie bevindinge beklemtoon die belang vir effektiewe bestuur van broeiery praktyke en genetiese monitering van broeidiere.

Thesis (MSc)--Stellenbosch University, 2012. / ENGLISH ABSTRACT: Haliotis midae, or more commonly known as Perlemoen, is an abalone species found along the coast of South Africa. It is the only cultured abalone species in South Africa and has a high demand abroad. Due to its popularity as a seafood delicacy, illegal harvesting has taken its toll on Perlemoen numbers. This increases the need for sustainable farming efforts and efficient implementation of law enforcement practices against poachers. Abalone farms make use of a limited number of broodstock for breeding, so it is necessary to ensure that genetic effects such as inbreeding and bottlenecks do not interfere with the viability of the offspring. Research that focuses on the genetics of Perlemoen will greatly aid the farms to continue sustainable production of this species as well as enhance their breeding efficiency. This study focuses on the construction of a linkage map for H. midae that will allow the future identification of markers associated with genes important to production, such as growth and disease resistance. Identification of these genes will allow breeders to select genetically superior abalone that will be used for breeding programmes in which the phenotype of the offspring will be enhanced. For the construction of a linkage map it is necessary to have enough informative markers for mapping. In this study, gene-linked microsatellite markers were developed by screening a contig assembly of H. midae’s transcriptome. Ninety-eight primer pairs could be developed from the contigs and 60 loci produced amplification products. Twenty-six microsatellites were found to be polymorphic (27% success rate). In addition to these markers, 239 previously developed microsatellites and 48 gene-linked SNPs were used to develop sex-average and sex-specific linkage maps in four full-sib families consisting of approximately 100 offspring each. Of these markers 99 were informative in family DS1 (31% success rate), 81 in family DS2 (26%), 77 in family DS5 (24%) and 71 in family DS6 (23%). These markers were used for linkage analysis (LOD>3). The average number of linkage groups for the sex-average maps ranged from 17-19. The average genome length for these maps ranged from 700cM to 1100cM with an average marker spacing of 8cM. The sex-specific maps’ linkage groups ranged from 13-17 with an average genome length of 600cM to 1500cM. The average marker spacing was approximately 16cM. The integrated map was constructed by merging the sex-average maps. This map contained 25 linkage groups with an average genome length calculation of 1700cM and an average marker spacing of 9.3cM. The linkage maps created in this study are the first to utilize SNPs in H. midae. Further incorporation of SNPs into linkage maps will enhance the density. The maps created in this study are of medium-density (65%) and provide a link to the development of high-density linkage maps to facilitate associations of phenotypic traits to certain markers, to so that QTL mapping can be performed. This information can be used for marker-assisted selection to produce genetically superior abalone. / AFRIKAANSE OPSOMMING: Haliotis midae, of meer algemeen bekend as Perlemoen, is 'n klipkous spesie wat langs die kus van Suid-Afrika voorkom. Dit is die enigste gekweekte klipkous spesie in Suid-Afrika en het 'n hoë aanvraag in die buiteland. As gevolg van sy gewildheid as 'n seekos lekkerny, het onwettige stropery sy tol geneem op Perlemoen getalle. Hierdie verhoog die behoefte vir volhoubare boerdery pogings en doeltreffende implementering van wetstoepassing teen stropers. Perlemoenplase maak gebruik van 'n beperkte aantal broeidiere vir teling, dus is dit nodig om te verseker dat genetiese effekte soos inteling en genetiese bottelnekke nie inmeng met die lewensvatbaarheid van die nageslag nie. Navorsing wat fokus op die genetika van Perlemoen sal grootliks die plase steun om die volhoubare produksie van hierdie spesie voort te sit, sowel as hul teling doeltreffendheid te verbeter. Hierdie studie fokus op die ontwikkeling van 'n genetiese koppelingskaart vir H. midae, wat die toekomstige identifisering van die merkers wat verband hou met die gene wat belangrik is vir die produksie, soos groei en weerstand teen siektes sal verbeter. Identifisering van hierdie gene sal toelaat dat telers genetiese voortreflike Perlemoen kan kies vir teelprogramme waartydens die fenotipe van die nageslag sal verbeter word. Vir die ontwikkeling van 'n genetiese koppelingskaart is dit nodig om genoeg informatiewe merkers vir die kartering te hê. In hierdie studie, is geen-gekoppelde mikrosatelliet-merkers ontwikkel deur ‘contig’ data van H. midae se transkriptoom te ondersoek. Agt en negentig inleier pare kon ontwikkel word uit die ‘contigs’ en 60 loki kon ‘n amplifiseringsproduk lewer. Ses-en-twintig mikrosatelliete was polimorfies (27% suksessyfer). Bykomend tot hierdie ontwikkelde merkers is 239 voorheen ontwikkelde mikrosatelliete en 48 geen-gekoppelde SNPs gebruik om geslagsgemiddelde en geslagspesifieke koppelingskaarte in vier volsib families, wat uit ongeveer 100 nageslag elk bestaan, te ontwikkel. Van hierdie merkers was 99 informatief in familie DS1 (31%), 81 in die familie DS2 (26%), 77 in die familie DS5 (24%) en 71 in die familie DS6 (23%). Hierdie merkers is gebruik vir 'n koppelingsanalise (LOD>3). Die gemiddelde aantal koppelingsgroepe vir die geslagsgemiddelde kaarte het gewissel van 17-19. Die gemiddelde genoom lengte vir hierdie kaarte het gewissel van 700cM tot 1100cM met 'n gemiddelde merker spasiëring van 8cm. Die koppelingsgroepe van die geslagspesifieke kaarte het gewissel van 13-17 met 'n gemiddelde genoom lengte van 600cM tot 1500cM. Die gemiddelde merker spasiëring was ongeveer 16cm. Die geïntegreerde kaart is saamgestel deur die samesmelting van die geslagsgemiddelde kaarte. Die kaart toon 25 koppelingsgroepe met 'n gemiddelde berekende genoom lengte van 1700cM en' n gemiddelde merker spasiëring van 9.3cM. Die genetiese koppelingskaarte wat in hierdie studie ontwikkel is, is die eerste om SNPs te gebruik in H. midae. Verdere insluiting van SNPs in koppelingskaarte sal die digtheid verhoog. Die kaarte wat in hierdie studie ontwikkel is, is van medium digtheid (65%) en bied 'n stap nader aan die ontwikkeling van hoë digtheid koppelingskaarte om fenotipiese eienskappe met sekere merkers te assosieer, vir kwantitatiewe kenmerk lokus kartering. Hierdie inligting kan gebruik word vir merker bemiddelde seleksie om geneties verbeterde Perlemoen te produseer.

Prestonella bowkeri and Prestonella nuptialis are montane specialists endemic to the southern Great Escarpment of South Africa. Phylogeographic analyses of these species based on mitochondrial markers CO1 and 16S reveal extremely high levels of divergence between populations indicating a lack of gene flow between populations. This is not surprising, because P. nuptialis and P. bowkeri have limited dispersal capacity, low vagility, a highly fragmented distribution and are habitat specialists that are restricted to isolated mesic refugia associated with waterfalls and montane seepages. A relaxed Bayesian clock estimate suggests that populations diverged from one another during the mid-late Miocene (12.5-7 MYA) which coincides with the modern trends of seasonal aridity which began during the Miocene. This result should be viewed with caution because the rates used are at best imprecise estimates of mutation rates in snails. There is no clear dichotomy between the two species and P. bowkeri is paraphyletic with respect to P. nuptialis, as a consequence the taxonomy is unclear. Due to the high levels of sequence divergence between populations they may be considered as evolutionary significant units (ESU’s). An assessment of haplotype diversity (h) and nucleotide diversity (π) reveals that populations in the western part of the Great Escarpment are more genetically depauperate than populations in the east. Correlations between genetic diversity and climatic variables show that genetically depauperate populations are found in areas that have lower annual rainfall, less reliable rainfall and higher potential evaporation, all factors associated with a drier, less mesic environment that increases the chances of a population bottleneck. This indicates that a shift towards a more arid environment may be a driver of genetic erosion. Historical climate change may thus have affected the amount and distribution of genetic diversity across the Great Escarpment since the Miocene. This has serious future implications for the survival of Prestonella. With predicted increase in global temperatures, climate change in South Africa is likely to result in range contraction and an eastward range shift for many species in the drier central and western areas (Erasmus et al. 2002) and regions along the Great Escarpment are likely to become more arid. Prestonella populations found living on inselbergs along the Great Escarpment are already restricted to site specific watercourses and seepages. An increase in the periods between stream flow, and increasing rainfall variability and mean annual potential evaporation are likely to have an adverse affect on species living in these habitats, resulting in further bottlenecks and possibly local extinction. An IUCN assessment of P. nuptialis and P. bowkeri suggests that these two species are probably endangered. The issue surrounding the conservation of Prestonella species is that they are threatened by global climate change, which cannot be simply restricted or prevented, which makes dealing with the threat of climate change difficult. Assisted migration (MA) may be considered as a method to prevent possible future extinctions of Prestonella populations, but will only be considered as a last resort. The thermal tolerance (Arrhenius breaking temperature and flat-line temperature) of individual snails from three Prestonella populations (one forest population and two thicket populations) were assessed using infrared sensors that detected changes in heart rate with increasing temperature. The forest population had a significantly lower Arrhenius breaking temperature (ABT) and flat-line temperature (FLT) than the two thicket population (p<0.05). Our results do not show a correlation between upper thermal limits and maximum habitat temperatures or other climatic variables in Prestonella populations. Although no correlation is found between ABT and maximum habitat temperature, it is likely that the differences seen between these populations are due to local micro-climate adaptation. The climatic variables used in this experiment are coarse estimates from GIS data and do not reflect actual microhabitat conditions. Forest environments are less heat stressed than thicket environments due to the forest canopy which may explain the lower ABT and FLT of the forest population.

MSc (Microbiology) / Department of Microbiology / See the attached abstract below

Poultry provide affordable animal protein products compared to other animal products in agricultural industry. The demand of organic food by world health organisation and call for discard of conventional laying cage production method led to this research study. The aim of the study was to determine how (4 x 4) full diallel crossbreeding of the Potchefstroom Koekoek (PK), Naked neck (NN), Lohmann Brown (LB) and White Leghorn (WL) had an effect on production performance, egg parameters, genetic and phenotypic characteristics of F1 crossbreed offspring. The study was conducted at the Agricultural Research Council (ARC), Livestock Production Improvement at the Irene Campus, which is situated about 25 km south of Pretoria. The (4 x 4) full diallel crossbreeding design used on four chicken breeds to produce four pure breeds, six crossbreeds and six reciprocal crosses. The total number of 352 chickens with16 treatments (2 cocks and 20 hens) used in phase 1 and 384 chickens 16 F1-treatments (3 cocks + 21 hens) used in phase 2. Data was analysed by full factorial analysis of variance (ANOVA), General Linear Model procedures and Scheffe post-hoc for multiple comparison of the means of different variable data. The outcome had shown that crossbreeding had an effect on the production performance, genetic and phenotypic characteristics. The performed F1 crossbreeds emerge from crossbreeding between the local dual-purpose PK and commercial LB chicken breeds. PKLB dominated on growth and production performance traits compared to other crossbreeds. All set null hypothesis differ significantly at (p < 0.05), the outcome of all five hypothesis of this study were rejected. In conclusion PKLB was the best performing F1 crossbreed, based on its best performance on growth, FCR, cost of rearing, productive, high quality safe eggshell, economic efficiency and consumer preference (brown eggshell and yolk colour). / Dikgogo di neelana ka dikumo tsa poroteine ya diphologolo go tshwantshanngwa le dikumo tsa diphologolo tse dingwe mo intasetering ya temo. Tlhokego ya dijo tse di bolang mo mekgatlhong ya boitekanelo ya lefatshe le pitso ya go latlha mekgwa ya kumo ya dikgetshe tsa go beela tsa tlwaelo di ne tsa isa kwa thutong ya patlisiso eno. Maikaelelo a thuto eno ke go tlhomamisa gore tsadiso ya kgabaganyo ya dilo tse pedi kgotsa go feta go tshwantshanya kgolagano ya mofuta wa dijene le tikologo tse di tletseng tsa (4 x 4) tsa Potchefstroom Koekoek (PK), Naked Neck (NN), Lohmann Brown (LB) le White Leghorn (WL) di na le ponalo mo tiragatsong ya kumo, diparametera tsa mae, le dijene le diponagalo tsa kgolagano ya mofuta wa dijene le tikologo tsa ditsadiso tsa kgabaganyo tsa ngwana wa F1. Thuto e ne ya diragadiwa kwa Agricultural Research Council (ARC) le Tokafatso ya Kumo ya Diruiwa kwa khempaseng ya Irene, e e agilweng bokana ka 25 km jwa borwa jwa Pretoria. Ditsadiso tsa kgabaganyo tsa dilo tse pedi kgotsa go feta go tshwantshanya kgolagano ya mofuta wa dijene le tikologo tse di tletseng tsa (4 x 4) di ne tsa dirisiwa mo mefuteng ya ditsadiso tsa dikgogo go ntsha mefuta ya ditsadiso e e tletseng e mene, ditsadiso tsa kgabaganyo tse thataro le dikgabaganyo tse di tshwanang tse thataro. Palo e e tletseng ya dikgogo tse di 352 ka ditiragatso di le 16 (mekoko e le 2 le dithole di le 20) di ne tsa dirisiwa mo letlhakoreng la 1 le dikgogo di le 384 ka ditiragatso tsa F1 di le 16 (mekoko e le 3 + dithole di le 21) di ne tsa dirisiwa mo letlhakoreng la 2. Data e ne ya tshetshereganngwa ka tshetshereganyo ya dintlha tse di tletseng tsa pharologantsho (ANOVA), dikgato tsa General Linear Model le tshwantshanyo ya bontsintsi ya morago (ANOVA), dikgato tsa General Linear Model le tshwantshanyo ya bontsintsi ya morago ga tiragalo ya Scheffe ka mekgwa ya data ya pharologantsho e e farologaneng. Ditlamorago di ne tsa bontsha gore ditsadiso tsa kgabaganyo di na le ponalo mo tiragatsong ya kumo, ga mmogo le diponagalo tsa dijene le setlhopha sa kgolagano ya mofuta wa dijene le tikologo. Go ne ga diriswa mefuta ya ditsadiso tsa kgabaganyo ya F1 tse di tlhagelelang go tswa mo ditsadisong tsa kgabaganyo magareng ga mefuta ya ditsadiso tsa dikgogo tsa PK tsa lebaka la gabedi la selegae le LB ya kgwebo. PKLB e ne ya fekeetsa metlhala ya tiragatso ya kgolo le kumo go tshwantshanngwa le mefuta ya ditsadiso tsa kgabaganyo tse dingwe. Setlhopha sotlhe sa dikakanyo tsa lefela se x farologana mo go bonagalang ka (p < 0.05) le ditlamorago tsa dikakanyo tse tlhano tse tsotlhe tsa thuto eno di ne tsa kganediwa. Kwa bokhutlong, PKLB e ne ya nna mofuta wa ditsadiso tsa F1 o o diragatsang go gaisa, go ikaegilwe ka tiragatso mabapi le kgolo, FCR, tshenyegelo ya go tsadisa, kumo, boleng jo bo kwa godimo jwa dikgapetla tsa mae tse di babalesegileng, bokgoni jwa ikonomi le boikgethelo jwa modirisi (dikgapetla tsa mae tse di tshetlha le mmala wa tlhae). / Agriculture and  Animal Health / Ph. D. (Agriculture)

MSc (Microbiology) / See the attached abstract below

DNA is identified as a powerful tool in the solving of rape cases, but it is often destroyed either by members of the public or the police officials who attend to the scene. The aim of the study was to evaluate the management of DNA evidence in rape cases in the Bishop Lavis Policing Area. To address the research topic under investigation, research questions, a legal framework and policies were used. The outcome of the study indicated poor performance in securing the crime scene and ensuring that physical evidence is preserved and not tampered with. On this basis, it was recommended that developmental workshops and intensive training on the management of DNA evidence be conducted to all members of the South African Police Service attend to rape crime scenes. This should be done to equip them with knowledge and an understanding of the management of DNA evidence. / Police Practice / M. Tech. (Forensic Investigation)

The Russian wheat aphid (RWA) (Diuraphis noxia Kurdjumov) is one of the important insect pests of wheat (Triticum aestivum L.), barley (Hordeum vulgare L.) and other grasses. To date, there are four RWA biotypes identified in South Africa. The virulent biotypes emerged, partly due to climate change and new genetic variations within populations of RWA; hence there is a need to improve host-plant resistance, as an effective control measure. Bread wheat (Triticum aestivum L.) accession Cereal Introduction (CItr) 2401 is known to be resistant to all RWA biotypes worldwide. The goal of this study was to use a backcrossed near-isogenic line (NIL) BC5F5 mapping population, developed from a cross between CItr 2401 and susceptible Kavkaz, to identify and validate single nucleotide polymorphism (SNP) markers linked to the resistance phenotype in CItr 2401. This was achieved by (i) conducting a preliminary study that evaluated the suitability of simple sequence repeat (SSR) markers previously reported in literature for discriminating stacked RWA resistance genes and, (ii) employing SNP markers for the first time in a RWA resistance study as a future alternative to the widely used SSR markers. None of the tested SSR markers showed potential use in marker-assisted selection (MAS). The mapping population was phenotypically evaluated for RWA resistance using the four South African biotypes, viz. RWASA1, RWASA2, RWASA3 and RWASA4. Analysis of variance (ANOVA) showed significant (P<0.001) differences of genotypes after confirming the normality of residuals and homogeneity of variance. The Illumina iSelect 9,000 wheat SNP platform was used to genotype the two crossing parents and a selection of 24 NIL genotypes from the mapping population. Eight SNP markers found to be linked to the phenotype were converted to breeder-friendly and high-throughput Kompetitive allele-specific polymerase chain reaction (KASP) markers. The designed KASP markers were validated on the two crossing parents, the 24 NIL sent for SNP genotyping, on the mapping population and on the preliminary study genotypes for their effectiveness. The KASP assays developed in this study will be useful for stacking the RWA resistance from CItr 2401 with other Dn genes effective against the RWA. / Life and Consumer Sciences / M. Sc. (Life Sciences)

Organisms inhabit environments that have many dimensions, each of which can vary temporally and spatially. The spatial-temporal variations of environmental stressors and disturbances may have major but different effects on indigenous and invasive species, favouring either of them at different times and places. The invasive mussel Mytilus galloprovincialis invaded the South African coast 30 years ago and, on the south coast of South Africa, it now competes and co-exists with the indigenous Perna perna in the lower eulittoral zone (referred to here as the mussel zone) The invasive and indigenous species dominate the upper and the lower mussel zones respectively, while the two co-exist in the mid-zone. My results show that intertidal mussels experience, and respond to, spatial and temporal fluctuations of several biotic and abiotic stressors. The invasive and the indigenous species adopt different strategies when reacting to environmental factors and their physiological and behavioural responses vary in time and in different habitats as different pressures become of overriding importance. Attachment strength of both species decreased in summer and increased in winter, and was higher on the open coast than in bays for both species, showing a strong positive correlation with wave force in time and space. P. perna had significantly higher attachment strength than M. galloprovincialis but, contrary to previous studies, the difference in gonad index between the two species varied according to the habitat. In bay habitats, M. galloprovincialis had a higher maximum reproductive effort than P. perna, however, on the open coast, there was no significant difference between the two species, suggesting that for the invasive species wave action is a limiting factor not only in terms of the attachment strength but also of energy availability for reproductive tissue development. Major spawning events occurred during periods of low wave action while minor spawning coincided with periods of intense hydrodynamic stress. On the open coast, gonad index was negatively correlated with attachment strength for both species while, in bays, there was no correlation between these two factors for either. The two species also showed different behaviour. In the field, M. galloprovincialis moved significantly more than P. perna over a period of six months. The higher mobility of the invasive species was also confirmed in the laboratory where, in general, M. galloprovincialis formed clumps more readily than P. perna. Taken collectively, these results suggest that channelling more energy into attachment strength limits reproductive tissue development and that, while the indigenous species invests more in byssal production, the invasive species adopts a more dynamic strategy looking for aggregation or a safer arrangement. Higher endolithic infestation and a greater expression of heat shock proteins (Hsps) in mussel populations on the open coast than in bays indicate that this habitat is a more stressful environment not only in terms of wave action. Endolith damaged mussels had significantly lower attachment strengths and condition indices than clean mussels, probably due to the need to channel energy into shell repair. The constant shell repair and expression of Hsps typical of open coast populations are energetically demanding processes. These observations suggest that on the open coast, mussels are subjected to more severe energetic constraints than in bay habitats. Wave and sand stress fluctuated seasonally with the former having a greater effect on mussel mortality on the open coast and the latter a higher impact on bay populations. Overall, mussel mortality rates were higher on the open coast than in bays. My results show that populations on the open coast had fewer private haplotypes and less genetic endemism than those inside bays. Gene flow analysis showed the relatively stable bay habitats act as source populations with greater genetic migration rates out of bays than into them. These differences in genetic structure on scales of las of kilometers show that coastal configuration strongly affects selection, larval dispersal and haplotype diversity. Environmental gradients that are key factors in species distribution over large geographical scales can also be responsible for micro-scale distributions. My results show that M. galloprovincialis colonizes the upper mussel zone where temperature is high, but is less tolerant to this stressor and has to maintain a high expression of Hsps. This suggests that temperature is probably a limiting factor in its invasion towards the sub-tropical east coast. There are inter- and intra-specific differences in responses to the environment which highlight the efforts of M. galloprovincialis and P. perna to optimize resource utilization for survival and reproduction. Determining these differences is crucial to understanding patterns of co-existence between competing indigenous and invasive species.

The dispersal of freshwater fishes in the Cape Floristic Region of South Africa has been attributed to river capture events and confluence of rivers during sea level regression. The role of low drainage divides and inter-basin water transfers have received less attention. A unique lineage of Galaxias zebratus (hereafter the Joubertina galaxias) occurs in two currently isolated river systems, the Gamtoos and Krom. The present study mapped the distribution of the Joubertina galaxias and used mitochondrial and nuclear DNA sequences to assess the processes that could have influenced its current distribution pattern. Analyses of both mitochondrial cytochrome ♭ and nuclear (S7) sequences revealed that observed genetic differentiation cannot be explained by isolation between the Gamtoos and Krom River systems. No genetic differentiation was found between the Krom River System and the Twee River (a tributary of the Gamtoos River System). Shallow genetic differentiation (0.4% for cytochrome ♭ and 0.3% for S7) was found between the Krom and the remaining populations in the Gamtoos River System. High levels of genetic structuring were observed within the Gamtoos River System with most tributaries having one or more unique alleles. Inter-basin dispersal during pluvial periods or recent human mediated translocation seems to be the most plausible explanations for presence of the Joubertina galaxias in the Krom River System. The present study also assessed the threats and habitat preferences of the Joubertina galaxias, and an evaluation of the conservation status of this lineage was done. The Joubertina galaxias is threatened mainly by alien fish invasion, habitat loss and loss of genetic diversity due to fragmentation of its populations. Due to its very restricted geographic range, small known population sizes and the intensity of threats to this lineage’s survival, this lineage has been assessed as Endangered. The lineage has a restricted Area of Occupancy (AOO) and Extent of Occurrence (EOO). The extent of occurrence has declined in all tributaries and is continuing to decline in all except two tributaries that are secure. The lineage may have had natural fragmentation among its populations, but recent threats have completely isolated the populations. The threats affect the lineage’s survival potential in the four tributaries which have small populations that are not presently viable. The densities observed for the Joubertina galaxias ranged from 0.16 - 1.3 fish /m² and the number of mature individuals for the whole population seems to be less than 2500 mature individuals. There is therefore a need for specific conservation actions to ensure the long-term survival of this unique lineage.