Interactions among large brown macroalgae, sea urchins, and fishes were investigated in northeastern New Zealand during the period 1988 - 1993. The Cape Rodney to Okakari Point Marine Reserve was the site of many of these investigations. The patterns of abundance of large brown macroalgae and urchins down depth gradients over a wide geographic range were compared with those reported from earlier studies, and 3 major trends were identified. First, the fucoid alga Carpophyllum flexuosum now occurs at many sites which are exposed to wave action, in contrast to earlier studies. This alga occurred most abundantly on urchin-grazed coralline flat areas. Second, at four sites in the Marine Reserve, the densities of the echinometrid urchin Evechinus chloroticus decreased with increasing depth, rather than reaching maximal densities at mid-depths, as had previously been described. Finally, at sites of decreased exposure to wave action, the coralline flats habitat did not occur at all, and dense stands of. C. flexuosum occurred, in conjunction with the ubiquitous laminarian alga, Ecklonia radiata. Following the discovery of this new algal component to exposed rocky reefs, a more detailed study of its population characteristics was initiated. The population size structure of C. flexuosum on coralline flat areas was markedly different from that of the same species in sites sheltered from wave action. These differences occurred at both offshore islands and sites near Leigh, suggesting that it was a general pattern. C. flexuosum plants on coralline flats were smaller than those from sheltered sites, and had a greater number of smaller laminae, heavier stipes, and a greater degree of branching. There was some evidence of temporal change in the morphology of C. flexuosum from coralline flats. Comparisons between a site with C. flexuosum and a site with coralline flats suggested that the activity of fish was 75% lower in the vegetated than in the unvegetated site, and the feeding rate in the vegetated site was less than 50% that in the unvegetated site. I speculate that future effects on fish activity of the invasion of C. flexuosum into a habitat which previously lacked macroalgal vegetation may depend on changes in the morphology of plants. An investigation of many aspects of the biology of E. chloroticus in different habitats was undertaken. Analysis of the body dimensions of E. chloroticus suggested that this species was relatively tall compared to other echinometrids (average ratio of test height: test diameter = 0.54), Comparisons among habitats with differing amounts of vegetation revealed only small differences in the relationship between test diameter and test height. Small E. chloroticus (<40 mm test diameter) lived in crevices, while larger individuals grazed freely over the substratum. In vegetated habitats, the crevice-dwelling habit was maintained at test diameters about l0 mm greater than in unvegetated habitats. Very small (<20 mm test diameter) E. chloroticus frequently covered themselves with shell. Population size structures of E. chloroticus within the Cape Rodney to Okakari Point Marine Reserve were bimodal; other localities had unimodal populations. Modal sizes varied among localities, with smallest modes (50-60 mm TD) being found at Inner Hauraki Gulf sites, and largest modes at the offshore Mokohinau Islands (70-80 mm TD). Habitat did not predictably affect population size structure. A bimodal population structure was maintained at Waterfall Reef rock flats throughout the 5-year study period. Gonad size showed seasonal fluctuations at several sites, being greatest in summer. There were few consistent differences in gonad size between biological habitats. Gonad colour varied among sites and habitats, with orange gonads generally being more prevalent in vegetated habitats, and black gonads being represented more in unvegetated habitats. Smaller urchins had greater proportions of orange gonads, while larger urchins had greater proportions of brown and black gonads. Although highly variable among individual urchins, movement of E. chloroticus was greater at unvegetated sites (0.7 m per 5 days) than at vegetated sites (0.4 m per 5 days), in the Marine Reserve. Feeding of E. chloroticus was studied at a number of sites in the Marine Reserve. Urchins frequently consumed drift algae, particularly E. radiata. C. flexuosum was consumed at less than half the rate of other macroalgae in several laboratory feeding experiments, and was chosen least frequently in a field assay of feeding preferences among 8 species of macroalgae. Boosting densities of E. chloroticus in stands of E. radiata to 60 m-2 led to destructive grazing of plants over a 2 month period - at lower densities, the urchins dispersed. Densities of C. flexuosum were effectively unchanged when urchin densities were increased to these elevated levels. As a result of these observations I speculate that feeding preferences of E. chloroticus may have a role in allowing C. flexuosum to survive on coralline flats. In a laboratory experiment, urchins from a feeding aggregation did not graze algae at higher rates than individuals from outside aggregations. Diets of both E. radiata and C. flexuosum consistently produced similar gonad volumes in urchins held in the laboratory, although gonad volumes produced were low. A preliminary experiment suggested that C. flexuosum from exposed sites was consumed at lower rates than C. flexuosum from sites which were sheltered from wave action. These differences in palatability are mirrored in the formation of stable borders between coralline flats and C. flexuosum of the sheltered morphology, and the ability of C. flexuosum of the exposed morphology to survive in the coralline flats habitat. The fish fauna of the Cape Rodney to Okakari Point Marine Reserve was shown to be different from that of a nearby area. A number of species were more abundant within the Marine Reserve. Subsequent surveys showed that there were differences in abundances of 3 large carnivorous fishes among sites within the Marine Reserve, and that population size structure and the distance within which divers could approach one species, (Pagrus auratus), clearly varied between areas within the Marine Reserve. Mean standard length of P. auratus in the central marine reserve was 40% larger than that of P. auratus outside the central marine reserve, and the average minimum approach distance was 70% less in the central marine reserve. Feeding of fish by humans in the central part of the Marine Reserve was suggested to be the main cause of the differences in responses to divers. Population size structure of, and crevice occupancy by, E. chloroticus, clearly differed between the Marine Reserve and an adjacent area, with bimodal population size structures and a 10 mm greater size of crevice occupancy occurring in the Marine Reserve. The implications of these findings for extrapolating from experiments done in one area to other areas are discussed. The major biological components of rocky reef habitats identified in this study were broadly similar to those identified in previous studies in northeastern New Zealand, and have parallels in overseas studies. Long term changes to the flora of rocky reefs in northeastern New Zealand have occurred, and appear to persist by a mechanism which had previously been discounted. Similar processes to those observed in overseas studies appear to maintain habitats (consistent recruitment of algae or urchins to habitats which they dominate), or cause them to change from one habitat state to another (e.g. grazing outbreaks by urchins). However, the predictability of the persistence of these habitats at a particular site appears to be low. Further, the precise mechanisms whereby habitats may change from one to another may also be unpredictable. I argue that there is little scope for general statements concerning the spatial and temporal occurrence, or mode, of habitat transitions on temperate subtidal reefs. This study emphasises the value of repeated descriptions of patterns of abundance, and highlights problems of extrapolation and generalisation in marine ecology. Insufficient information exists at present to comment adequately on the persistence of subtidal habitat types. This may in part stem from the types of information which have been collected in the past. Methodological problems with the use of quadrats to sample densities of organisms in areas of differing topography are therefore addressed. In conclusion, it is suggested that sampling protocols which incorporate a variety of information, gathered over as wide an area, and as intensively as possible, should be used in future research of this type.

The New Zealand green-lipped mussel, Perna canaliculus, is farmed by an aquaculture industry (> NZ$ 150 million/year) that is dependent on mussel juveniles (spat) collected from unpredictable and unreliable wild sources for more than 80% of its mussel seed requirements. Most wild-caught spat is collected from the surf zone at Ninety Mile Beach, northern New Zealand, where unique environmental conditions cause the accumulation and transport of spat attached to drift algae, which arrive to the shore in great quantities (up to 100 tonnes at once). This study is the first to investigate the ecology of mussels at Ninety Mile Beach throughout their life history, including reproductive behavior, micro-scale settlement patterns on filamentous macroalgae, accumulation and transport of mussel spat to the shore, colonization of the rocky intertidal, and adult population dynamics. Histologic investigation of 4 intertidal and 2 subtidal populations revealed that females and males were well-synchronized throughout their reproductive cycles, with a prolonged spawning season from June to December, when temperatures were lowest and rising. Comparisons of gonad indices and maximum shell length indicated high productivity in certain populations, which likely contribute to the high larval availability of the area. Experiments showed that mussel spat preferentially settled on fine-branching natural and artificial substrata, with correlation evident between mussel shell size and degree of branching. Furthermore, greater numbers of mussels settled on node versus internode areas within natural and artificial substrata. Chemical cues for mussel settlement also were studied using phytogel plates spiked with algal extracts, which were preferred over control plates by mussel larvae/post-larvae in the field, and by hatchery-reared larvae in the laboratory. Three intertidal populations were investigated at different temporal and spatial scales. Mussel concentrations in seawater were higher after spawning for both small mussels (< 0.25 mm) in August, and for larger mussels (> 0.5 mm) in March. Settlement patterns within quadrats cleared of mussels in 2 habitats (adult mussel bed and adjacent areas covered with algae) were studied from July 1999-March 2001. Primary settlement (mussels < 0.5 mm) was found to dominate the algal habitats at the beginning of the spawning season in August, while secondary settlement (mussels > 2.0 mm) was higher in the adult mussel bed late in the spawning season (November-March). Monthly surveys of undisturbed quadrats indicated that a peak in new recruitment coincided with a peak in adult mortality in August. At Scott Point, massive mortality for 2 years in August was followed by a dramatic re-colonization of the empty spaces by juveniles. Settlement patterns of mussels on suspended ropes in the water column were investigated at 3 water depths inside and outside Ahipara Bay during 2 spawning seasons (1999-2000). Mussel settlement was higher for small mussels (< 0.49 mm) in shallower water (2 m water depth) in August, and higher for larger mussels (> 1.0 mm) at greater depths (18 m water depth) in September-December. Mussels found on shallow-water ropes may have settled directly from the plankton, whereas mussels on ropes near the bottom (18 m) may have transferred from macroalgae tumbling on the seafloor. Environmental conditions associated with mussel spat arrival to the beach (spatfall events) were studied by statistical analysis (1990-1998; daily, monthly, inter-annually) of wind speed/direction, tidal range, water temperature, swell height/direction, and records of spatfall events/amounts. Spatfall events/amounts were more abundant during days of strong offshore winds. Swell height in the onshore direction was significantly correlated with spatfall events/amounts. Storm events (wind speeds > 20 m/s) were most frequent between May-August; spatfall events/amounts were more numerous 4 months later (September-October). Years with more storm events (El Niño/La Niña episodes) were associated with significantly higher spatfall events/amounts.

This thesis examines sublethal stress responses in the common New Zealand estuarine bivalve, Austrovenus stutchburyi (cockle). The approach used throughout this thesis is a refinement of the biological indicator approach of Adams (1990), in which biomarker data are collected primarily at the individual and sub-organismal levels, but with additional measurements at the population and community levels. In this research several biochemical and physiological biomarkers are used to assess effects of contaminants on energetics, fecundity and growth of cockles. Adenylate energy charge (AEC) and total adenylate nucleotide pool (TANP) are biomarkers that have previously not been used with A. stutchburyi, and provide new information on energy cycling for this species. RNA concentration has also not previously been used with cockles, and this biomarker is correlated with growth. Glycogen concentration and condition indices are used to indicate reproductive potential and overall health. In order to fully investigate the utility of this comprehensive suite of biomarkers, their performance in laboratory and field situations was tested. Temporal variability in AEC, TANP and glycogen was analysed over two years for two populations of cockles in a northern harbour. The results indicated seasonal patterns for all biomarkers, with AEC reaching a maximum value in spring, TANP being lowest in winter, and glycogen concentration being high over spring and summer. Variability in biomarker response between cockles sampled at low tide and high tide was analysed in a laboratory experiment simulating exposure to air during low tide. Adenylate energy charge was found to be significantly lower in cockles after exposure to air for 4 hours, highlighting the importance of standardised collection protocols. These initial patterns, as well as spatial variability in biomarker response, were further investigated at several sites within a second harbour south of Auckland. There was little among site difference in glycogen concentration, and the data suggested that differences in site characteristics, such as sediment quality and type, were driving the mixed response patterns of AEC, TANP and condition. Because of inherent variability in field conditions, a series of laboratory experiments was then undertaken to assess responses under carefully controlled conditions. Cockles were challenged with PAHs, chlordane and tributyltin at two different doses, and biomarker response (AEC, TANP, glycogen and RNA) measured. One trial examined the effects of a one-off pulse of contaminant, in which sediment containing cockles was dosed at the outset of a 14-day experiment with a high concentration of contaminant. A daily dose of this same concentration of contaminant was supplied to the sediment for 14 days in a second trial to assess effects of a more continuous discharge. There were virtually no significant differences in biomarker response between treatments and controls for either trial, suggesting that the experimental setup may have been causing sublethal stress in some way. Finally, in order to maximise experimental control whilst maintaining environmental realism, a manipulative field experiment was undertaken in which cockles were transplanted from an uncontaminated site to a series of uncontaminated and contaminated sites in one of Auckland's major harbours. Biomarker response (AEC, TANP, glycogen and RNA) was measured in the transplanted and the autochthonous populations two weeks and eight weeks after transplantation. An ability to regulate adenylate nucleotides was detected, as cockles transplanted to contaminated sites had reduced TANP but maintained AEC levels. There were significant differences in glycogen and RNA among sites, with highest levels detected in cockles transplanted to an uncontaminated site. This suite of biomarkers showed great utility for use in environmental quality assessment. It is recommended that an approach such as that used in the Mussel Watch programme in the United States be employed for cockles, in order to monitor estuarine ecosystem health in New Zealand. The transplantation of cockles to estuaries where there are concerns about environmental quality, and the use of TANP, glycogen and RNA concentration to assess sublethal stress, has potential as a sensitive and cost-effective environmental monitoring technique for estuaries in New Zealand.

The eggs and larvae of the New Zealand snapper Pagrus auratus are pelagic with early buoyancy provided by dilute body fluids. The swimbladder begins to develop on the third day after hatch from a dorsal evagination of the gut tube. Communication w1h the gut is lost on about the tenth day following pneumatic inflation at around day eight. At this age the gas gland system appears fully functional and capable of secreting gas. By the age of settlement at around 30 days the swimbladder is a fully functional replica of the adult form except for the lack of a resorbent capillary system which does not develop until later in juvenile life. The swimbladder of the adult is of the euphysoclist form with a dorsally located resorbent oval area and sits high in the pleural cavity. The ventral tunica externa is firmly attached to the connective tissue lining the pleural space. The adult swimbladder displaces 5.6% of the volume of the body and its volume is regulated to provide near neutral buoyancy. The connective tissue integument provides almost no restriction to volume changes brought about by vertical movements of the fish and the swimbladder obeys Boyle's Law for physiological pressure changes. The ability of the connective tissue of the tunica externa to accommodate large tissue strains is due to massive regular crimping of otherwise straight collagen fibrils allowing reversible extensions up to 130%. In all other respects however the tissue structure of the tunica externa is consistent with a tissue providing an active mechanical role. The fibrillar morphology and physicochemical properties of swimbladder collagen is consistent with the vertebrate type I form however there are interesting variations in collagen form distributed throughout the swimbladder. Fibrillar morphology of the highly extensible tunica interna is significantly different to that of the tunica externa and appears to play very little mechanical rote. The extensibilty of the tunica externa appears to be regulated by physiological stress and related to the past history of tissue strain.

The characterisation of the universal features of intertidal zonation on rocky coasts (Stephenson and Stephenson 1949; 1952; 1954; 1961) has stimulated many attempts t o explain the factors responsible for the observed patterns of distribution. Initial investigations which were essentially confined to the effects of a single factor upon different species suggested that ecological position was often correlated with physiological response. Recent studies have indicated that the subtidal distributions of algae show characteristic patterns (Kitching 1941 ; Bergquist 1960; Kain 1960; 1971; Morgans 1961 ; McLean 1962; Schwenke 1966; Neushul 1967; Clark and Neushul 1967; Larkum et al.1967; John 1971). However, there have been relatively few attempts to examine in detail and subsequently integrate the physiological response to environmental factors of major belt-demarcating algae of the sublittoral . The four New Zealand species of Carpophyllum are inhabitants of the uppermost sublittoral and the sublittoral fringe (Bergquist 1960; Dromgoole 1965). This particular study was undertaken in an attempt to elucidate the subtle differences in response which determine the distribution limits characteristic of each species. There is little information on the biology of Carpophyllum with the exception of some anatomical and embryological work (Delf 1939; Dawson 1940; Naylor 1954) and an earlier investigation of C. maschalocarpum (Dromgoole 1965). Thus to provide a basis for physiological experiments preliminary studies were directed to the following aspects: ( i ) a brief review of the morphology, anatomy and taxonomy of the genus, ( ii ) growth of sporelings in laboratory culture using various seawater media, ( iii ) definition of the geographical and vertical limits of the four species and a general study of their ecology with emphasis on population distributions, periodicitiy of reproduction and plant composition in relation to environment. The physiological response of algae to environmental factors has been examined by previous workers at the protoplasmic level using the criterion of resistance (e.g. Biebl 1956, 1962; Kanwisher 1957); at the metabolic level by measurement of photosynthesis and respiration (e.g. Chapman 1961n) ; and at the level of the whole plant by growth rates in laboratory cultures (e.g. Kain 1965) or of selected individuals in the field (e.g. Klugh and Martin, 1927). Attempts to culture Carpophyllum were not entirely successful. An earlier investigation indicated that the growth rates of Carpophyllum in situ are difficult to assess (Dromgoole 1965) and hence the metabolic approach, which does allow some integration of response to several factors by means of carbon balance calculations (e.g. Brown and Johnson 1964), has been used extensively in this study. The metabolic activity of large brown algae varies considerably with age and part of plant (Clendenning and Sargent 1957a, b; Chapman 1961a). To eliminate sampling problems in Carpophyllum a cuvette allowing continuous long-term monitoring of oxygen exchange of a single piece of material was developed. This laboratory apparatus has shown that the photosynthesis and respiration of the various organs is sensitive to environmental changes in pH, C02 supply, oxygen tension, light , water flow, dehydration and temperature. The nature and possible causes of the response t o each factor were examined in detail as this was considered essential to any interpretation of their ecological significance. Finally, the information obtained from these gas exchange measurements has been integrated by means of metabolic balance calculations and a descriptive synthesis of field and laboratory results has provided a general picture of the biology of these plants.

The scope of the present work is the systematic revision of the Demospongiae of the N.Z. region, here defined as including North and South Is. and all outlying islands as far north as Three Kings and South to Auckland and Cambell Is. The Kermadec Is. are not considered as part of the N.Z. region, but the sponges recorded from these islands are revised in the sytematic account. At present, 243 species of Demospongiae are recorded for this region, these belonging to 124 genera and 35 families. Of this total 49 species have been described as new and forty-six species are recorded as new to the fauna in this work.

The humpback whale was almost driven to extinction by commercial whaling in the Southern Ocean. Little is known about the degree of interchange among the remaining Southern Hemisphere populations. This thesis aimed to assess the connectivity among breeding grounds, feeding areas and migratory corridors of humpback whales using mitochondrial and nuclear DNA markers. The population structure of humpback whale populations in breeding grounds across the South Pacific and eastern Indian oceans was investigated, with an interest in the origins of whales in eastern Polynesia, using an extensive collection of mitochondrial DNA (mtDNA) sequences (n = 1,112; 470 base pairs in length). Samples were obtained from living whales at six breeding grounds: New Caledonia, Tonga, Cook Islands, French Polynesia (Society Islands), Colombia and western Australia. We found significant differentiation, at both the haplotype and nucleotide level, among the six breeding grounds (FST = 0.033; ΦST = 0.022) and for most pair-wise comparisons. The differentiation of the eastern Polynesia humpback whales is consistent with the hypothesis of a relic subpopulation, rather than vagrancy or colonization from known neighbouring breeding grounds. Regardless of their origin, it seems probable that islands of eastern Polynesia are now the primary breeding grounds for humpback whales feeding in the management Area VI (170°W – 120°W) of the Antarctic, as defined by the International Whaling Commission. A population of humpback whales migrates along the western South American coast, with breeding grounds mainly off Colombia and Ecuador and feeding areas off the western coast of the Antarctic Peninsula and in the channels and fjords ofsouthern Chile. We analysed the genetic relationship between humpback whales from these two summer feeding areas of the eastern South Pacific population to assess the potential heterogeneity in the migratory pattern of the population. We compared mitochondrial DNA control region sequences from 132 whales from the Antarctic Peninsula and 52 whales from the Magellan Strait areas. An AMOVA showed significant differences between the two feeding areas (FST = 0.180; ΦST = 0.169). A phylogenetic analysis showed both areas are represented in the AE clade that is only found in the Southern Hemisphere in the Colombian breeding ground. Genetic and previous demographic data (based on photo-identification) strongly suggest that both feeding areas are related to the same breeding ground but that heterogeneity exists among the feeding areas of this population, similar to that observed in the North Pacific and North Atlantic populations of humpback whales.Humpback whales migrating through eastern Australia and New Zealand have been linked with those breeding off northeastern Australia, New Caledonia, Fiji and Tonga, forming a single stock (Breeding Stock E). We investigated the relationship between the New Caledonian and Tongan breeding grounds, based on interchange of individual whales (genetically identified) and population genetic differentiation (mitochondrial DNA control region sequences and nuclear DNA microsatellites). We found significantly higher recapture probabilities within each breeding ground compared to the recapture probability between them using genetic identification, and significant differences at the population level in the FST and ΦST for mitochondrial and nuclear markers. These analyses suggest differentiation among the Breeding Stock E, supporting a proposed sub-stock division for New Caledonia (E2) and Tonga (E3). Historically, humpback whales off New Zealand coasts were caught during their migratory journeys between Antarctic feeding areas and tropical breeding grounds in the South Pacific. Here we investigated the genetic diversity of New Zealand humpback whales, comparing mitochondrial DNA control region sequence data with that from breeding grounds across the South Pacific (New Caledonia, Tonga, Cook Islands, French Polynesia and Colombia) and eastern Indian (western Australia) Oceans. We analyzed 30 samples collected around New Zealand, revealing 20 haplotypes. All haplotypes were found in New Caledonia and some were also found in other breeding grounds. New Zealand humpback whale haplotype diversity and nucleotide diversity were similar to those from the compared breeding grounds, but were significantly different at haplotype level from the Cook Islands, French Polynesia and Colombia breeding grounds. We found significant differences at haplotype level with the same three locations when a pair-wise AMOVA was performed. Three breeding grounds (western Australia, New Caledonia and Tonga) did not show significant differences at either nucleotide or haplotype levels. This genetic evidence and the available demographic data suggest a closer relationship of the New Zealand stock with New Caledonia and to a lesser extent with Tongan whales, supporting New Zealand as a primary migratory corridor for the humpback whales breeding in these western Pacific Islands.

This thesis examined the relative importance of natural and human influences to the population and assemblage dynamics of reef fishes in northeastern New Zealand. In particular, how different reef fishes responded to the implementation of no-take and partial marine reserve protection, and physical factors responsible for spatial differences in fish abundance. Included were data from prior to the establishment of a no-take marine reserve, multiple fished reference locations, biannual sampling and the use of two independent methods to provide quantitative estimates of fish abundance and size. This combination of factors is rare in studies of marine reserves was an important strategy leading to an improved understanding of the mechanisms structuring fish communities. Responses of the reef fish assemblage to changes in fishing mortality were examined at the Poor Knights Islands Marine Reserve. Full no-take marine reserve protection was implemented on the 1st Oct 1998 but for the prior 17 years, the Poor Knights Reserve comprised only two small no-take zones and allowed recreational fishing in the rest of the reserve. Following implementation of no-take marine reserve status the reef fish community changed rapidly; there were no obvious changes at either reference location. Species targeted by fishers, such as Pagrus auratus (snapper) and Caprodon longimanus (pink maomao), responded most strongly to protection. An increase in the density of some non-targeted species can probably be attributed to climatic effects, rather than a reduction in fishing pressure. A decline in the abundance of several species at the Poor Knights may have been a result of natural mortality, or competitive or predatory interactions with snapper. Along with human influences (fishing), physical variables are important in determining the distribution and abundance of reef fish. Four locations (two mainland and two island) were surveyed in northeastern New Zealand to determine spatial patterns for seven labrid species, one of the most abundant and widespread taxa of reef fish in New Zealand. The underlying mechanisms were then explored through an examination of the relationship between swimming ability (as examined through pectoral fin morphology), exposure and depth. Each of the four locations consistently displayed distinct labrid assemblages, likely due to the influence of the East Auckland Current. Regardless of location, there was a consistent depthrelated trend for most labrids and a trend for some species to be associated with certain levels of wave exposure. By analogy with tropical labrid assemblages, it was expected that there would be a clear relationship between pectoral fin aspect ratio and depth and/or exposure. However, this relationship was not strongly evident suggesting that wave exposure may not be as important for labrids on northeastern New Zealand reefs as hypothesised for tropical coral reef systems. The response of snapper, an important recreational and commercial finfish, was investigated following the cessation of all fishing at the Poor Knights. The rate of response of snapper to protection was rapid, in areas that had previously been partially protected as well as in those that had been fully protected from fishing, with the overall density of legal sized fish increasing by 7.4 times over 4 years, likely a result of recolonisation rather than recruitment. The 818% increase in snapper biomass has the potential to enhance areas outside or within the reserve through the export of biomass (eggs and/or larvae and adults) - the daily batch fecundity was 11 to 18 times higher at the Poor Knights compared to the reference locations. The effects of partial protection on reef fish were further examined at the Mimiwhangata Marine Park, an area where recreational fishing is permitted but all commercial fishing has been prohibited for 18 years. Snapper showed no difference in abundance or size between the Mimiwhangata Marine Park and adjacent control areas, with the density of snapper most similar to fished reference locations. The lack of any recovery by snapper within the Marine Park, despite the exclusion of commercial fishers and restrictions on recreational fishing, and results from the Poor Knights, indicates that partial fishing regulations are ineffective as conservation tools for protecting targeted species or for fish communities in general (i.e. through reduction in by-catch). Results from this study provide evidence that recreational fishing has significant impacts on reef fishes.

Bottlenose dolphins (Tursiops truncatus) are frequently brought into contact with humans, due to their coastal habitat use. Over the last decade there has been considerable growth in the commercial dolphin-based tourism industry. This growth exposes coastal populations to boats and swimmers. New Zealand has allowed commercial operations offering swimming with wild dolphins since 1989. Little is known about the effects of boats and swimmers on dolphin behaviour. The present study examines a population of bottlenose dolphins that utilise the Bay of Islands as part of their range along the northeastern coast of New Zealand, and their behavioural response to commercial swim-with-dolphin operations. Data were collected from an independent research boat and the commercial dolphinwatching boats from December 1996 to February 2000. Photo-identification showed a closed population of 446 (95% C.I. 418-487) dolphins, with an 81% resight rate of the 378 catalogued individuals. Group size ranged from 2 – 50 dolphins, with 80% of groups (n = 160) containing 2 - 20 dolphins. The number of groups with calves present increased from 32% (n = 17) in 1997, to 63% (n = 47) in 1999. This could be explained by the increased number of neonates born; from four in 1997, to nine in 1999. Photo-identification in areas outside the Bay of Islands showed that individuals ranged minimum distances of 82 km north, and 388 km south of the Bay of Islands. All photographs taken outside the Bay of Islands were matched to known animals, suggesting that the majority of the northeastern coastal population is seen in the Bay. With the next nearest population of bottlenose dolphins approximately 1,000 km south, these dolphins probably form an isolated population ranging along the east coast of the northeastern North Island. Habitat use by the dolphins was examined using Geographic Positioning System (GPS) points simultaneously linked to behavioural states. In total, 31 groups were followed, and 847 GPS points were recorded. Position of initial encounter was recorded for 138 groups of dolphins. There was a significant correlation between water depth and sea-surface temperature, with dolphins found in deeper waters in summer when water temperature was highest, and in shallow waters in winter when the water temperature was lowest. Only one behavioural state, slow travel, was significantly different by season. Although all behaviours were observed in all zones, dolphins had preferential habitat use throughout the four zones within the Bay. To determine the level of impact on dolphin behaviour due to boat traffic, focal-group follows using a scan sample methodology were conducted from an independent research boat. A CATMOD analysis was used to determine dolphin behaviour by season, number of boats, and group size. Dolphin behaviour differed by boat number; in particular, resting behaviour decreased as boat number increased and milling behaviour increased as boat number increased. Behaviour differed when the research boat was the only boat present. Closer examination of boat type, i.e., permitted dolphin watching boats, non-permitted tour boats, and recreational boats, found a difference in dolphin behaviour in the presence of the permitted boats when compared to the other types. Dolphins rested less and engaged in more milling behaviour when the permitted boats were present. Responses to swim attempts from the commercial boats were compared from 1994 - 1995 to 1997 - 1998. In total, 255 groups were encountered of which 36% (n = 93) were exposed to at least one swim attempt. Successful swims (i.e., where at least one dolphin remained within 5 m of at least one swimmer) decreased from 48% in 1994 - 1995 to 34% in 1997 - 1998 and avoidance responses increased from 22% to 31%. Dolphin response varied according to swimmer placement from the boats. The ‘in path’ placement had the highest level of avoidance. Only ‘line abreast’ placement resulted in a decrease in avoidance. Based on the photo-identification sighting records, it was estimated that an average dolphin was exposed to 31 swim attempts per year. This suggests that individual dolphins have, with cumulative experience, become sensitised to swim attempts. If a swim attempt was successful, on average, it involved 19% of the group. Juveniles were more likely to interact with swimmers than adults. The present study suggests that the wide-ranging population of bottlenose dolphins that frequents the Bay of Islands is exposed to levels of tourism which can affect the dolphins behaviour. In particular, the permitted boats’ presence results in decreased resting behaviour by the dolphins. The apparent sensitisation of this population to swim attempts, and the relatively small number of individual dolphins that interact with swimmers, are grounds for concern about the long-term impact of interactions. Ongoing monitoring of this populations’ response to swimmers and boats, combined with a precautionary approach to management of boats and swimmers, are recommended.

This thesis uses molecular genetics as a tool to uncover information about the population structure and genetic variation in Hector’s dolphin (Cephalorhynchus hectori), to track population declines and to assess the evolutionary origins and taxonomic status of this species. A high-resolution genetic analysis of population structure was considered important for the determination of population boundaries and delimitation of conservation management units due to potentially unsustainable fisheries-related mortality. Population structure and dispersal rates were assessed using 281 samples collected from individual Hector’s dolphins of ten population groups representing the known geographic range of this species. Variation among mitochondrial DNA sequences (ΦST = 0.545) and microsatellite allele frequencies at six loci (RST = 0.252) indicated the presence of four genetically isolated regional populations, North Island (n = 29), East Coast South Island (n = 110), West Coast South Island (n = 122) and South Coast South Island (n = 19). Significant levels of genetic differentiation were not detected within local sub-populations of the East Coast and West Coast regional populations. However, the estimated geneflow between these sub-populations fitted a one-dimensional stepping-stone model (r2 = 0.6225) suggesting a vulnerability of local populations to fragmentation. A measure of expected mtDNA diversity (Tajima’s D statistic) suggested decline in eight of the ten populations. Microsatellite heterozygosity was also lower than expected in the East Coast and North Island regions, suggesting either further regional sub-structuring (Wahlund effect), loss of diversity due to population decline or the presence of null alleles. Examination of all Hector’s dolphin museum specimens of known origin (n = 55) enabled comparison of historic (1870 - 1987) genetic diversity to contemporary (1988 – 1999) diversity in two regional populations to assess the possibility that these populations have undergone recent declines. Over the last 20 years the North Island population has been reduced from at least three lineages (h = 0.41) to a single lineage (h = 0, p < 0.05). The diversity of the East Coast, South Island population has declined significantly from h = 0.65 to h = 0.35 (p < 0.05). These results suggest that the low abundance currently observed is due to recent population declines and that the North Island population is threatened with extinction in the near future. Based on a trend analysis of the mtDNA, it can be predicted that the East Coast South Island population may lose all mtDNA diversity within the next 20 years. Alternatively, detection of a one dimensional dispersal pattern may indicate that some populations are at risk of extirpation while others may not be in decline. If this is the case then the East Coast regional population is at risk of fragmentation. On a wider evolutionary scale, Hector’s dolphin is one of four species of the genus Cephalorhynchus, all of which suffer fisheries–related mortality. To describe the origin and radiation of these species, 485 bp of the mitochondrial DNA control region was sequenced from 320 individuals (including previously sequenced 200 Hector’s dolphins) representing nine of the ten species in the sub-family Lissodelphininae. The hypotheses that either Cephalorhynchus is a monophyletic genus or that the four species have arisen separately from pelagic Lissodelphine species and have converged morphologically were tested. The mtDNA phylogeny supported the monophyly of the genus and suggested that the genus Cephalorhynchus originated in the waters of South Africa and, following the West Wind Drift, colonised New Zealand and then South America. Secondary radiations resulting in two genetically isolated populations were found for the Kerguelen Island Commerson’s dolphin and the North Island Hector’s dolphin. A comparison of the genetic differentiation between the Commerson’s dolphins of the Kerguelen Islands (n = 11) and the coast of South America (n = 35), and between the North Island (n = 14) and South Island (n = 185) Hector’s dolphins, was conducted in order to assess the conservation and taxonomic status of these populations. A single fixed substitution in the mtDNA control region was diagnostic for the Kerguelen Island compared to South America (FST = 0.306, ΦST = 0.602) and the North Island compared to the South Island (FST = 0.442, ΦST = 0.495). Population differentiation of four microsatellite alleles (including unique alleles in each of the four populations) between the Kerguelen Island and South American Commerson’s dolphin (FST = 0.036, RST = 0.0493) and between the North and South Island Hector’s dolphins (FST = 0.391, RST = 0.3197) indicated restricted nuclear as well as maternal geneflow. These data, combined with additional evidence of morphological and geographic isolation, indicated that the Kerguelen Island Commerson’s dolphin and the North Island Hector’s dolphin are likely to be reproductively isolated from their alternate con-specific populations. Examination of various species concepts and definitions of conservation units leads to the conclusion that these four populations should each be considered unique at the subspecies level for the purposes of management, protection and evolutionary potential. These results lead to the conclusion that the Hector’s dolphin consists of highly subdivided populations. As a result of this and a low reproductive potential, Hector’s dolphin populations are vulnerable to extirpation through even low levels of human induced mortality. To manage such populations, it is appropriate to consider each of the two islands as separate sub-species. Within the South Island, the populations may be further subdivided into three demographically independent Management Units – the East, West and South Coasts. The South Coast management unit is vulnerable due to its low abundance and isolation and requires further investigation. Population modelling will need to reflect the fact that the local populations within the East and West coast regions share only limited dispersal with immediately adjacent populations and are thus susceptible to fragmentation. These results also show that the population declines of the East Coast South Island and the North Island populations are of recent origin thus implicating fisheries-related mortality as the principal threat to Hector’s dolphin. To prevent further decline or fragmentation of South Island populations more stringent control of inshore gillnet fisheries is required. By contrast, current decline of the North Island population may be a result of inbreeding depression. Given the low abundance and rapid decline of the North Island population, it is imperative to evaluate the potential for inbreeding depression while continuing to mitigate all human-related threats.