The Knysna seahorse Hippocampus capensis is South Africa’s only endemic seahorse species, and is found in only three adjacent estuaries along the southern coast. The conservation of this endangered species is important on a national and international level. This study presents the first research on this species within the Knysna estuary since 2001 and specifically focuses on aspects of its ecology within a residential marina estate (Thesen Islands Marina). The physico-chemical and habitat features of the marina were described and the population ecology, habitat use, and behaviour of the Knysna seahorse were investigated. Physico-chemical conditions within the western section of the marina, characterised by high water current velocities, were similar to that of the adjacent estuary. The eastern section of the marina was characterised by lower water current velocities and higher turbidity. Four major habitat types were identified within the marina canals: (I) artificial Reno mattress (wire baskets filled with rocks); (II) Codium tenue beds; (III) mixed vegetation on sediment; and (IV) barren canal floor. Seahorse densities within the marina were significantly higher compared to densities found historically within the estuary. Highest seahorse densities were specifically found within the artificial Reno mattress structures and within the western section of the marina. Seahorse density varied spatially and temporally and the type of habitat was an important predictor for seahorse occurrence. An experimental investigation found that H. capensis chooses artificial Reno mattress habitat over Zostera capensis when given a choice. GoPro cameras were used successfully to investigate daytime seahorse behaviour within the Reno mattress habitat. Seahorses were more active during the morning, spent most of their time (> 80 %) feeding, and morning courting behaviour for this species were confirmed. However, during the summer holiday period (mid-December to mid-January) few seahorses were observed on camera, which suggests that the increase in motor boat activity and the related increase in noise had a negative effect on H. capensis feeding and courting behaviour. The marina development, and in particular the Reno mattresses, created a new habitat for this endangered species within the Knysna estuary. In addition to the protection and restoration of natural habitats in which H. capensis is found, the conservation potential of artificial structures such as Reno mattresses should be realised.

The purpose of this thesis was to develop a method by which economic value can be imputed from an economic activity to a non-market ecological function or service which contributes to that economic activity. The Knysna River in South Africa was chosen as the ecological function which supported three economic activities from which value was to be imputed; these were the Knysna Municipal Water Supply, Fish Production in the Knysna Estuary and Production of Indigenous Forest within the Knysna Catchment. Three underlying assumptions and two functional operations were required in order to implement the suggested method. The underlying assumptions were: - The ecological and economic activities considered are within a single catchment. - The allocation of value imputed for a specific economic activity to the ecological function or service under consideration (in this case the Knysna River) is proportional to the total contribution of ecological functions or services contributed to the economic activity. - The valuation of the economic activity for the purposes of obtaining a price-quantity point on a demand function is to be full cost pricing with no producer surplus. The two functional requirements were: - Diagram or map the linkages between an economic activity and the supporting ecological functions. - Determine the consumer surplus related to an incremental change in quantity under a demand function where the original price and quantity are known. A value from each of the economic activities was imputed to the Knysna River. However, the method was not tested. Nonetheless applying the equations and collecting the required data allowed several methodological needs to be clearly pointed out. The most acute deficiency was difficulty in obtaining secondary data from governmental agencies, commercial representatives and existing published academic research to ensure a robust price. Also, scientific information was not sufficiently available for allocating ecological contributions to the economic activities. Even with the shortage of credible data the method appears to allow non-market ecological functions to be valued in context of an existing economic system.