During long-term studies of dolphins, the number of individuals in the population being studied are constantly monitored using the technique of photo-identification. This constant monitoring makes use of different researchers over time. Therefore, measurement of photographic quality and individual distinctiveness for photo-identification analyses was incorporated in this dissertation to provide an additional data set for analysis. Researchers with differing levels of experiences did not obtain the same information from the same photograph and were unable to reliably quantify variables of photo quality and individual distinctiveness, but experienced researchers were found to be more adept than inexperienced researchers in counting notches on the dorsal fin of bottlenose dolphins. These results highlight the necessity for researchers to be trained in photo-identification techniques prior to carrying out their study. This study theodolite tracked dolphins off Durban from June 2004 to Feb 2005 to assess habitat utilization of bottlenose dolphins (Tursiops aduncus) in the Durban bay region. Dolphins were seen in all months surveyed and throughout the survey area. Temporal distribution of dolphins was skewed with 91% of dolphins seen before midday and 98% of sightings observed in water depths less than 30m. Six behavioural categories were recorded, including: social, fast travel, slow travel, feeding, resting and milling. The most dominant behaviour exhibited by dolphins was slow travel (46%) followed by feeding (27%). Resting was not observed at all. Of the feeding behaviour 88% occurred in the southern end of the bay whereas other behaviours occurred randomly throughout the survey area. Additional theodolite tracks were conducted during experimental boat approaches (before, during and after boat approaches) to determine potential short-term reactions of dolphins to dolphin watching boats. Two speeds of approach (slow ~ < 5 km/hr and fast ~ > 40 km/hr) and two distances of approach (20m and 80 m) were tested. The bottlenose dolphin groups did not change their behaviour in response to boat approaches during any of the periods of experimentation. Short-term changes in group speed, group size and spread were not statistically significant. Dolphin groups continued with their ‘normal’ behaviour and spent the same amount of time in the bay when compared to their distribution and behaviour in the absence of the experimental boat. These findings indicate that the experimental boat did not affect the behaviour of dolphins at either a slow or fast approach and even at a close distance. This is interpreted as being as a result of habituation of the dolphins due to their residency in a busy port. This work is crucial in developing guidelines for the development of a sustainable dolphin watching industry off Durban. / Thesis (M.Sc.) - University of KwaZulu-Natal, Durban, 2008.

Coastal dolphins on the South African east coast are threatened by degradation and loss of habitat as a result of increasing coastal development, industrial effluent and agricultural runoff. In addition, dolphins off the coast of KwaZulu-Natal have, for more than four decades, been heavily exploited through unchecked incidental capture in shark nets set at 45 beaches. In light of the high rate of mortality and apparent depletion of both species, the persistence of bottlenose (Tursiops truncatus) and humpback (Sousa chinensis) dolphins in that region has been questioned. Genetic variation in south east African dolphin populations was determined as a means of assessing the fitness of the populations and their resilience to demographic disturbances. Furthermore, in order to determine the effects of continued mortality on the KwaZulu-Natal subpopulations, it was necessary to determine whether they are open or closed to immigration from the adjacent East Cape region, which represents a relatively unstressed region, characterised by a lack of shark nets and less intensive coastal activities. Genetic variation and differentiation in the maternal genome was assessed by determining the sequence of the first 400 bases of the mtDNA control region in bottlenose and humpback dolphins from KwaZulu-Natal and the East Cape. Nuclear variation and differentiation was estimated at six microsatellite loci and compared with earlier estimates determined from allozyme electrophoresis. Random amplified polymorphic DNA (RAPD) was assessed as a means of identifying population subdivisions and diagnostic population markers. Both bottlenose and humpback dolphins on the South African east coast are characterised by low nuclear and organellar genetic variation, consistent with a possible genetic bottleneck, the inferred date of which coincides with the onset of the last glacial period. Genetic variation in South African bottlenose dolphins was lower than that reported elsewhere for the species, while an intraspecific comparison supported lower genetic variation in South African humpback dolphins than in humpback dolphins sampled off Hong Kong. An analysis of molecular variance (AMOVA), performed on mtDNA haplotype frequency data indicated, for both species, significant genetic subdivision, concordant with geographic location. The data suggested female bottlenose dolphins demonstrate regional philopatry, displaying limited movement between KwaZulu-Natal and the East Cape. Female humpback dolphins tend towards strict local philopatry, with significant maternal differentiation evident both within and between regional subdivisions. Differentiation in microsatellite allele frequencies was also demonstrated between KwaZulu-Natal and the East Cape for both species, suggesting that the movement of male bottlenose and humpback dolphins may also be restricted. Nonetheless, considerably higher nuclear gene flow estimates suggested that males of both species represent the principal vectors of gene dispersal. The implications of historically low genetic variability and population subdivision in South African dolphins are important in view of the current rate of mortality in KwaZulu-Natal. The persistence of coastal dolphin populations relies on their ability to recover following a bottleneck event. Continued removal of demographically important age-sex classes such as occurs in shark nets, may not only further reduce the genetic variation, but would ultimately deplete dolphin populations in KwaZulu-Natal beyond a sustainable number, resulting in eventual local extinction. The differentiation of the two regions implies that, in the event of local extinction occurring, dolphins, particularly females, from adjacent regions will not readily re-colonise the area. This would result in fragmentation of the south east African populations and ensure reproductive isolation from neighbouring populations on the east African coast.