Orca (Orcinus orca), also known as killer whales, are more widely recognised than other marine mammals. Although they have been reported from all oceans of the world, including the seas around New Zealand, information above anecdotal notes exists for only a few places. Orca are an apex marine predator that exhibits cultural differences in diet, vocalisations, and behaviour, between and within populations. This study was established to determine baseline information on New Zealand orca and to provide recommendations for future management and conservation. The conservation status of orca worldwide is poorly known, although two populations of the Pacific North West Coast of North America have recently been classified as ‘Threatened’ and ‘Vulnerable’. Photo identification was used to determine the population size, distribution around New Zealand waters, as well as range use and association among individuals. The total New Zealand orca population is small (range 65-167 animals, with 115 calculated alive in 1997). Resighting rates were high, with 75 % (n = 88) of the animals seen on more than two occasions. The mean number of sightings for the 117 photo-identified animals was 5.4, the mode was one sighting, and the median 9 – 10 sightings. One orca was photographed over a 20 year period. Population structure, frequency of association with others, and other social behaviours were used to determine population demographics. The New Zealand orca population appears to be made up of at least three sub-populations based on geographic distribution (North-Island-only, South- Island-only and North+South-Island sub-populations). Preliminary mtDNA analysis supports the hypothesis that some New Zealand orca do not mix. The mean Association Indices within the North-Island-only and South-Island-only sub-populations are significantly greater than within the North+South-Island sub-population. Those animals sharing food had higher Association Indices than those who did not share food. Sex ratios appear similar within each sub-population and calves were present in each, suggesting all sub-populations are breeding. Feeding behaviour was observed to assess habitat use and differences between foraging strategies and prey preferences. Twenty four different species of prey have been recorded in the New Zealand orca diet. Of these, ten have not been recorded elsewhere. The prey consists of four types; rays (the most common food type), sharks, fin-fish and cetaceans (pinnipeds have not been identified as a prey source). Foraging strategies were different for each prey type, with benthic foraging for rays in shallow waters the most diverse strategy used in New Zealand. Food sharing was observed for all prey types. One of the three proposed New Zealand subpopulations appears to be generalist or opportunistic foragers, feeding on all four prey types, another sub-population slightly less so, feeding on three prey types, and the third sub-population appears to be a more specialist forager, only recorded taking one prey type (cetaceans). Potential threats to orca, in addition to small population size, such as bioaccumulation of toxic chemicals, oil spills, boat strikes and shootings are considered and recommendations for conservation and future management are offered. Whether the three sub-divisions within the New Zealand orca population are reproductively isolated and hence require separate management, and whether there is further sub-division within the proposed North+South-Island sub-population, requires further study including genetic analysis. Some level of ongoing monitoring is recommended to ensure that the population of New Zealand orca does not decline. In addition, records of stranding locations and details of strandings are appended. Twenty-four live strandings occurred, involving 63 killer whales, of which 17 animals were successfully refloated and two of these resighted. One was seen after three years (nine resightings) and the other after four months (10 resightings). Refloating stranded orca is recommended.

The biochemical pathways involved in nitrogen (N) utilisation by marine phytoplankton have received considerable attention over the last forty years, but our understanding of these processes, and how they are affected by environmental change is still far from complete. This study investigates N metabolism in marine phytoplankton in both a controlled laboratory environment (using the coastal marine diatom Phaeodactylum tricomutum), and in the open ocean (e.g. Jellicoe Channel and the Subtropical Convergence Zone, New Zealand). Although the characteristics of ammonium uptake have been extensively studied in marine phytoplankton, comparatively little information exists on rates of assimilation. In this study, a robust method for measuring the rate of ammonium assimilation after a transient addition of ammonium is described. The method relies on the measured ability of the protonophore carbonyl cyanide m-chlorophenylhydrazone (CCCP) to release unassimilated ammonium from the cell and prevent further assimilation. There was little or no correspondence between the rate of ammonium assimilation and rates of ammonium uptake or maximum glutamine synthetase (GS) activity in Phaeodactylum tricomutum. Moreover, in N-limited cells maximum GS activity was a poor measure of N incorporation under steady-state conditions. However, GS activity did provide reliable information on N status (e.g. increased GS activity with increased N-limitation). Comparisons of the effects of varying N-source suggest that nitrate-grown cells are not disadvantaged under conditions of N-limitation due to the extra costs associated with nitrate reductase (NR) and nitrite reductase (NiR) activity. The metabolic costs of growth on nitrate may be significantly increased under iron (Fe)-limitation, as both NR and NiR require Fe. Fe-limited chemostat cultures excreted nitrite and ammonium when grown on nitrate. This release is probably a response to insufficient photoreductant under Fe-limited conditions. However, under Fe-limitation cellular N and C was similar to that of Fe-replete cells, suggesting that the N-source used for growth (nitrate or ammonium) did not influence N-assimilation (i.e. that nitrate-grown cells were able to secure the extra reductant required to support growth) under Fe-limited, light saturating conditions. The Gln:Glu ratio (an index of the cellular N-status) was significantly reduced under N-limitation, but not under Fe-limitation. Measurement of several biochemical indicators of the physiological state of phytoplankton cells (e.g. Gln:Glu ratio, GS activity, and Fv/Fm ratio) permitted the nutrient status of phytoplankton populations to be investigated during the NIWA Ocean Fronts programme over the Subtropical Convergence Zone, New Zealand. Low Gln:Glu ratios suggested that phytoplankton in both Subtropical and Subantarctic waters were N-limited, with a marked increase in this ratio when Fe was added to Subantarctic phytoplankton. The temporal utilisation of N by neritic phytoplankton was also investigated in Jellicoe Channel, northeastern New Zealand. Again, several biochemical indicators (e.g. Gln:Glu ratio, GS activity, and Fv/Fm ratio) were used to identify the N-status of this neritic phytoplankton assemblage both during bloom and non-bloom periods.

The impact of no-take marine reserves on spiny lobster, Jasus edwardsii, populations and the adjacent lobster fishery was assessed in a study that determined: the response of lobster populations to protection, if lobsters within marine reserves moved out into the surrounding fishery, examined the behavioural characteristic of lobsters that contributed to their recovery and spillover, and contrasted catch characteristics around a north-east New Zealand marine reserve with 2 unprotected sites. Surveys of 4 no-take marine reserves of different ages and 4 unprotected control sites confirmed that J. edwardsii increase in mean size and abundance within protected areas. Total lobster density was estimated to increase by 6.1% per year of protection. The relationship between age of reserve and the density of lobsters above the legal size limit (>100 mm carapace length) was complicated by depth related interactions between reserves, due to seasonal changes in the depth distribution of lobsters and differences in the timing of surveys among locations. However, density increases in legal sized lobsters were greater than those of the overall population. Mean size of protected lobsters increased in a non-linear fashion, reflecting the asymptotic growth characteristics of the species. Mean carapace length of lobsters increased by 11.24 mm after 3 years of protection and only by 16.25 mm after 2l years of protection. Corresponding increases in biomass (kg.500 m-2) and egg production (eggs.500 m-2) were estimated to be 7.4% and 6.6% respectively per year of protection. A tag recapture program indicated that the limited scale of movements and relatively high site fidelity of mature J. edwardsii aided the recovery of lobster populations in protected areas. Of 737 lobsters tagged in and around the Leigh Marine Reserve,212 were subsequently resighted. Seventy percent of resighted lobsters were relocated less than 250 m along the shore from their initial site after a mean period of liberty of 146 (se = 13.42) days. However, seasonal offshore movements took lobsters beyond the seaward boundary of the reserve where they were susceptible to capture. Of the 521 lobsters tagged on inshore reefs within the Leigh Marine Reserve. l.9% were caught over the seaward boundary by commercial fishers, and 14.4% of the 181 lobsters tagged over the seaward boundary were resighted on inshore reefs within the reserve. Movements into or out of the reserve across the longshore boundaries were also recorded for 7 out of the 27 lobsters moving further than 250 m along the shore. Acoustic tracking confirmed that J. edwardsii have a high site fidelity, with tracked lobsters spending a median of 84% of their time at their home site. Twenty one percent of the 26 lobsters fitted with acoustic tags never left their tagging site and, of those that did leave 56% eventually returned. Lobsters moved up to 3.1km away from their home sites but managed to find their way back after periods of 1 to 103 days away. Five lobsters had strong associations with two separate sites and made repeated movements between these alternate home sites. The straight line distance between alternate home sites ranged from 200 m to 1.3 km, and the period between visits to their primary or secondary home sites ranged from 1 to 93 days. Movement activity varied throughout the year, and seasonal patterns differed between males and females. Females displayed one peak in movement activity per year, around the time of larval release in September-October. Males displayed a summer peak centred on January, and a winter peak centred on July. The onset of high movement activity was often accompanied by a shift to deeper water and l0 of the 14 lobsters tracked for over 6 months spent periods away from the inshore reef system. Exposed groups of lobsters were observed aggregating in offshore areas of sand and low lying patch reef during the day. Offshore aggregations were observed over 7 months of the year and lobsters within the aggregations displayed behaviour consistent with the use of mutual defence. The sexual composition of the aggregations reflected seasonal patterns in male and female movement rates. Male J. edwardsii dominated the aggregations in winter and summer; times when peak feeding rates of males held in captivity were also recorded. Females dominated the offshore aggregations in September-October when larvae are released. Seven of the 26 lobsters fitted with acoustic tags moved over the boundaries of the Leigh Marine Reserve and Tawharanui Marine Park during offshore movements. Despite the fact that there is no fishing access to inshore reefs within the Leigh Marine Reserve, no significant difference was detected between the catch per unit effort (kg.trap haul-1) of lobsters caught around the reserve, nearby coastal Leigh or Little Barrier Island. However, catches around the marine reserve contained fewer (P=0.0009) but larger lobsters than at Little Barrier Island. The catch characteristics of lobsters from coastal Leigh were intermediate between the other two sites. As access to inshore reefs around the Leigh Marine Reserve was limited to a small area at either end of the reserve, traps tended to be set in offshore locations and the sexual composition of the total catch around the Leigh Marine Reserve reflected the seasonal movements of males and females into, and out, offshore areas. At coastal Leigh and Little Barrier Island the location of traps was not restricted and greater use was made of coastal fringing reefs. As a result the sexual composition of the catch at these locations did not reflect the movement of various components of the lobster population into and out of specific habitats or locations. Overall, the results of this study suggest that relatively small (~ 5 km2) no-take marine reserves increase lobster biomass and egg production, without adversely affecting catch rates in the surrounding fishery. Marine reserves may therefore represent a viable management tool, which if used in conjunction with other management regimes, could provide a more precautionary approach to the management of the J. edwardsii fishery at minimal cost to the industry.

This study focuses on an assessment of the health status of the yellowbelly flounder (Rhombosolea leporina) from two estuarine locations (site 1 – the mouth of the Henderson Creek; site 2 – the mouth of the Whau Creek) in the Waitemata Harbour. This harbour borders the highly urbanised and industrialised Auckland City metropolitan area. Whangaparaoa Peninsula, located approximately 30 km north of the other two collecting sites (away from the main urban area), was chosen as a reference site for comparative purposes. Physico-chemical analyses revealed differences in water quality at the sampling sites. A lower pH, oxygen deficiency and higher temperature were recorded in both the Waitemata Harbour locations in comparison with the reference site. Histopathological analyses revealed significantly higher prevalences and severity of pathological changes in the gills, blood, liver, kidney and gonads of the yellowbelly flounder from both harbour locations in comparison with fish from the reference site. ln addition, some types of lesions (eg. neoplasms) were observed in fish from the two harbour locations only. Abnormalities in the gill structure of fish from both harbour sites included: epithelial swelling (hyperplasia and hypertrophy), necrosis, and lifting with oedema; the fusion of secondary lamellae; aneurysms; filamental deformities; mucous cell proliferation, and infestation by Trichodina. The abnormalities found in the blood of these fish were manifested as: polycythaemia; erythrocytosis; erythroblastosis; leucocytosis (increased neutrophils); poikilocytosis; anisocytosis and an increased prevalence of erythrocytes undergoing necrosis. Vacuolar degeneration of the hepatocellular parenchyma due to lipid or glycogen accumulation was the most prominent liver change observed in fish from all sampling sites. The other liver abnormalities observed in flounder from the two harbour sites included: foci of cellular alterations (clear, basophilic and necrotic), congestion of the sinusoids, infestation by nematodes, and anaplastic growths (cholangiocellular carcinoma and teratoma). Pathological changes found in the kidneys of flounder collected at both harbour sites were classified as glomerular abnormalities (atrophy and dilatation of the glomerular tuft; enlargement of Bowman's space) and tubular vacuolar degeneration and necrosis. The presence of myxosporean parasites was also a common finding in the kidneys of harbour fish. Follicular atresia was the most prevalent change observed in the ovaries of flounder from the two harbour sites. Biochemical analysis of plasma proteins and electrolytes of flounder from the three sampling sites revealed hyperbilirubinaemia, hypoalbuminaemia and uraemia in those inhabiting both harbour sites. In addition, the concentration of total liver microsomal proteins was significantly depressed in flounder from harbour sites 1 and 2 when compared to that of fish from the reference site. Concentrations of heavy metals in the livers of flounder from different sampling localities were found to decrease in the order: site 2 > reference site > site 1, and did not correlate with the prevalences of liver abnormalities. Significantly more prominent pathological changes were thus observed in fish from both estuarine harbour sites in comparison with those from the reference open water site. The pathological changes noted are believed to occur in response to environmental changes. Contamination by different xenobiotics in the Whau and Henderson Creeks, which have been recorded in previous studies, suggest the possibility of direct toxic effects of the water contaminants on flounder from the estuarine parts of these creeks. In addition, the relatively high temperature, low pH and low oxygen levels recorded at the two sites in the Waitemata Harbour are believed to have induced oxygen deficiency-related tissue hypoxia which could then have led to the expression of a variety of diseases of which some have been detected in this study. However, the possibility that some unknown and unmeasured causal factors may have produced the observed pattern of flounder diseases cannot be eliminated.

The population biology and ecology of scallops in Greater Omaha Bay (a semi-oceanic bay) and Kawau Bay (a estuarine bay) was quantified by observation and experiment. Information was collected to extend the knowledge base on northern New Zealand scallop population dynamics, and for application to potential scallop enhancement in the region. Contagious scallop population organisation was found at all spatial scales examined, ranging from bay wide through to individual bed patchiness, down to the scale of inter-animal distances. Such clumping has strong implications for a range of population processes, including fishing susceptibility and fertilisation success. Monitoring of adults found two main spawning events to occur; in late October and in mid January. Changes in the gonado-somatic index (GSI) were well synchronised between individuals within populations. Subsequent monitoring of spat-fall in artificial collectors documented two main recruitment events, probably the outcomes of the two local spawning events. These spat-fall events occurred on collectors separated by 1Os of km. However, substantial density variations occurred between sites, indicating that local hydrodynamics may have played a significant role in modifying local spat-fall intensities. Smaller spat-fall events were also present between the two major events. The number of spat collected at a number of combinations of site and time were sufficient to support commercial spat catching operations, although problems were encountered with spat detaching at sizes too small to be retained by the collectors. In the 1993/94 summer a large algal bloom event completely eliminated scallop recruitment to collectors for the first three months of that season. Mass mortality events were a major contributor to overall benthic scallop population mortality. Probable causes included intensive scallop harvesting (commercial and recreational), a major storm episode, and a large algal bloom. These effectively eliminated scallop populations from Greater Omaha Bay. The adjacent Kawau Bay was not affected by any of these particular events, but populations there did not survive long after reaching adult sizes. Estimates of M (natural mortality) were higher for all scallop populations than have been previously documented in New Zealand studies. Growth trajectories were reasonably consistent in waters shallower than 19 m, but a progressive decline occurred in both maximal size reached and average growth rates with increasing depth after this point. Food limitation may have been the mechanism involved, which is likely to vary significantly for other locations depending on local environmental conditions. Average time to recruitment to the fishery (100 mm shell width) was three years for the shallower populations. A slight reduction in average size of adults at higher densities was found for some populations, indicating a possible density-dependent effect. Examination of a high density scallop bed found animals to display distinctive substratum preferences over small spatial scales, with higher abundances occurring on coarser materials such as shell gravel, marl and grit. Mud was not favoured as a habitat type. Movements of tagged animals at this location were spatially limited to within the particular habitat patch in which an individual was tagged and released, i.e. at a scale of 1s to 1Os of metres. No animals moved between adjacent patches of similar habitat (100 m scale). A B.A.C.I type experiment was undertaken to assess incidental mortality effects of commercial scallop dredging on undersize scallops, at the spatial scale of beds. Significant negative effects were quantified, with the number of undersize animals killed per legal animal harvested estimated at 1.7 and 2.8 : 1, depending on the size frequency structure of the fished bed. Modelling of likely improvements in the number of animals surviving at the end of fishing, given a reduction in the minimum legal size from 100 to 90 mm, indicated improvements of 20 to 41% of the original population remaining after fishing, depending on animal size and assumed dredge efficiencies, A 90 mm MLS has subsequently been adopted by the Coromandel Scallop Fishery. The results from this work provide detailed population based estimates of parameters required for successful management and optimal harvesting strategies of Hauraki Gulf scallop populations. The large variability in parameters such as mortality, and strong abundance correlations with habitat type, has strong implications for such activities. This work also provides essential information for the undertaking of locally based enhancement operations, such as the spatial and temporal magnitude and variability of spat-fall events, and growth rates with respect to habitat features (i.e. depth).

The research presented in this thesis was undertaken in order to develop an understanding of the biology of Perna canaliculus sufficient to allow for commercial hatchery based production of Greenshell™ mussel spat. Hatchery production is an alternative to unreliable and inconsistent wild spat collection. In a Perna canaliculus population followed for one year spawning occurred in early spring and late summer. Three quantitative histological measures of gonad maturity utilising image analysis technology and a qualitative classification system were compared. Measuring the relative surface area comprised of gametes on histological sections was found to be the most reliable method. A practical gonad visual index to determine the reproductive condition of adults for the selection of broodstock was developed and found to be highly effective as a means of predicting induced spawning success. Serotonin was not effective for inducing spawning of Perna canaliculus. Temperature shock and the use of stripped gametes was however found to be a reliable spawning induction method. Relative gamete concentration, gamete age, temperature, sperm half life and gamete contact times were all found to have effects on fertilisation success for Perna canaliculus. Sperm concentration and the conditions of sperm aging were particularly important. Fertilisation kinetics of Perna canaliculus gametes modelled using the Vogel-Czihak-Chang-Wolf method suggested that 5% of sperm-egg contacts lead to successful fertilisation. Broodstock management protocols that could be used to condition the adult of Perna canaliculus were investigated in order to enhance and prolong the natural reproductive season. Research suggested that for successful broodstock conditioning animals should already have begun gametogenesis at the time conditioning is commenced. Successful conditioning of Perna canaliculus was achieved at temperatures between l0 and 16°C over a period of about 50 days. A diet ration above 2-3% of the dry meat mass per day is suggested. A trial examining non-algal diet supplements suggest a mixture of yeast and lipid emulsion may have some potential value. Photoperiod manipulation did not effect the reproductive condition of Perna canaliculus. The yield of veliger larvae was significantly enhanced if embryo culture water was treated with 1.0 mg/l EDTA. Veliger yield was not significantly affected at densities below 50 embryos/ml. Perna canaliculus larvae grew most rapidly and survived well at the salinity of 35 ppt. Larvae grew most rapidly when cultured at low densities. Experiments suggest that early larvae can be cultured at 5-10/ml, however late stage larvae grew most rapidly when cultured at l/ml. Perna canaliculus larvae displayed best growth and good survival if fed a mixed flagellate-diatom diet comprising Isochrysis galbana (T-Iso) and Chaetoceros calcitrans. The optimal diet ration, as a function of larval size, increased from about 20 cells/μl Isochrysis galbana (T-Iso) to around 150 cells/μl through the larval development period. Thyroxine between the concentrations of l0-5 and l0-8 M did not have an observable effect on larval developmental rate or eye spot development. Down welling settlement systems were found to be generally successful for Perna canaliculus lanrae. L-DOPA was also demonstrated to enhance the settlement and metamorphosis of Perna canaliculus pediveligers.

The marine pulmonate limpet Siphonaria Zelandica is commonly found in the mid-eulittoral zone in shallow pans or rock pools with dense algal growth. Though it resembles a true limpet externally, it is active when emersed during the ebbing of the tide browsing on small succulent algae, rather than grazing the surface film. The structure and function of the digestive and respiratory systems of Siphonaria were studied, particularly in relation to its adaptations to the intertidal mode of life. The digestive system is relatively simple consisting of a large buccal mass with paired salivary glands, a spacious esophageal crop, a slightly muscular stomach with two digestive diverticula and a simple intestine-rectum. The arrangement of the 28 muscles of the buccal mass and the odontophore is typical of patelliform pulmonates. The radula is a broad sheet with 144 ± 31 rows of teeth, each row having a tooth formula of (18±4) + (18±3) + (1) + (18±3) + (18±4). The radular teeth are continuously produced at a rate of 4.9 rows per day similar to those of active feeding herbivorous gastropods. The characteristics of the teeth and the movement of the mouth and the feeding traces showed that the feeding processes do not involve rasping of a hard substratum but browsing and scooping of soft lush algal growth. The whole alimentary tract is lined by ciliated columnar cells of a basic form but with slight structural variations in different region. Those in the oesophagus and stomach contain apical vesicles and lysosome-like bodies indicating involvement in intracellular digestion, while those in the intestine contain various amounts of lipid and glycogen, functioning in active absorption and storage. The ciliated cells of the post-intestine, having a much folded basal plasma membrane are apparently involved in osmoregulation. The proteinaceous secretory cells in the crop and mid-intestine and the glycoprotein cells in the pro-intestine are either producing enzymes for extracellular digestion or secreting the faecal-binding sheath. Mucous cells interspersed among the ciliated cells of the oesophagus intestine and rectum produce mucoid material for lubrication and transport of food particles. The cuticular strip in the anterior chamber of the stomach, probably a vestigial gastric shield consists of tall non-ciliated cells with thick microvilli embedded in a dense fibrous matrix. The salivary gland contains six cell types: grain cells, duct mucocytes, mucocytes I and mucocytes II, duct ciliated cells and undifferentiated cells. The first four are secretory, the grain cell being serous secretory and the other three muciparous, with varying amounts of mucopolysaccharides. The tubules of the digestive gland are made up of five cell types, acinous digestive cells, neck digestive cells, crypt cells, vacuolated cells and undifferentiated cells. Both types of digestive cells participate in absorption and digestion. The high lipofuscin content in the neck digestive cell suggests its possible role in lipid digestion. The crypt cell with elaborate GER whorls and proteinaceous globules is apparently secretory, producing enzymes for extracellular digestion. Apocrinal secretion of the globules from the crypt cells was observed for the first time. The vacuolated cells appear to be degenerated crypt cells with an excretory function. The digestive gland tubule was found to undergo cyclic changes, correlated with the feeding activity which is related with the tides. Enzyme assays showed the presence of various carbohydrases, proteolytic and lipolytic enzymes in the different region of the digestive system. The digestive gland is the main site of enzyme production. The pH optima of amylase, laminarinase and sucrase from various regions of the digestive tract were found to be within the pH range of the alimentary tract. The activities of various enzymes in the salivary gland and the oesophagus but not digestive gland were synchronous and correlated with feeding. When emersed, Siphonaria respires atmospherically with the pneumostome wide open, and the mantle cavity functioning as a lung without any active ventilatory movements. When submerged, the mantle cavity is filled with water, a strong water current through the gill being created by the ciliated dorsal and ventral raphe behind the gill. The available respiratory surfaces include the sides of the foot, the anterior mantle roof and the single plicate gill. All these are lined by a thin epidermis with large blood spaces beneath it the non-ciliated epidermal cells in different regions carry different lengths of microvilli, apparently related to the degree of their protective role. The respiratory physiology of Siphonaria was also studied. No tidal or diurnal rhythm in oxygen consumption was revealed. In unagitated conditions the aerial respiratory rate was much higher than the aquatic rate but with agitation, the aquatic rate increased to near the aerial rate. Cutaneous respiration constituted about 25% of the total. The respiratory rate varied with body weight, the coefficient b being temperature-independent (0.791 ± 0.122). Both aerial and aquatic respiratory rates increased with temperature. The highest temperature-sensitive range was 10-25°C in winter and 15-30°C in summer which coincided with the normal environmental temperature. No seasonal temperature acclimation was found within the normal thermal range. Respiratory pigments haemocyanin and myoglobin were detected in the blood and buccal mass respectively. Their characteristics were studied and their role in oxygen transfer system was postulated. The reverse Bohr shift of the haemocyanin may facilitate oxygen uptake in the lung during burses of activity at low tide. A high oxygen-combining capacity of the buccal mass myoglobin (21.2 vol%) indicated a role of oxygen storage during bursts of feeding activity. The distribution of carbonic anhydrase in various tissues was consistent with a transfer system facilitating the release of metabolic CO2 from the buccal mass. Finally the phylogenetic relationships of Siphonariidae with other groups of marine pulmonates were discussed. The unspecialised digestive system, the advanced form of nervous system and the possession of a secondary gill suggest that Siphonariidae, as a family, may not be an advanced derivative of any of the more primitive living basommatophoran groups, but could represent, along with the Gadiniidae, a specialized group which has originated directly from a primitive pulmonate stock and are adapted to life on exposed rock shores.

Two subtidal reef communities characterised by sponges were examined to assess two distinct aspects of marine benthic community ecology which have not previously received much attention. The first study focused on the role of settlement events in the origin and maintenance of community structure and involved investigation of factors affecting dispersion, settlement and recruitment of sponge propagules. The second examined how natural gradients in ambient physical conditions, in the absence of disturbance, affected size structure and species interactions within the community. Settlement events were examined on a reef flat which was covered by a shallow sediment overlayer. The distribution and abundance of established sponges was found to be highly correlated with sediment regimes. The sponge community was shown to be stable over time and able to resist disturbance by storms because of the presence of turfing algae. Sponge settlement was inhibited by the sediment overlayer. Established sponges were found clumped into oblong groups orientated east-west and there were many instances where sponges were found closely associated. Adjacent areas were characterised by deeper unstable sediments and fewer sponges. Storms contoured sediments into ripples and scoured oblong east-west orientated patches of basal rock. Sediments were also sorted during storms such that particles of large grain size were left around the edges of ripples and over the scoured rock patches. The basal reef of both the stable and unstable areas was covered in a mosaic of encrusting algal and sponge species which tolerated the sediment overlayer. This assemblage was also exposed during storms. Sponges reproduced in a number of ways, but asexual propagules were the most successful. Sponges from several orders produced buds which developed in a complex manner. Buds adhered to rock and shell fragments of large size which acted as anchors, or attached to other buds irrespective of genotype. The change in weight and shape increased the likelihood that buds would gravitate into the scoured patches of basal rock. Bud loss off the reef was minimal. Buds were able to settle onto the basal rock patches and did so almost exclusively onto bare rock. A high degree of 'selectivity' of settlement site by buds was imparted by altering developmental rate during the settlement process in response to the substrate type. Polymastia granulosa buds exhibited an additional developmental sequence not previously reported for any colonial organism. Buds either settled directly in suitable conditions, or developed further along a different pathway. Buds which did not come into contact with basal rock elongated and appeared to be capable of digging into sediments. The elongated bud then broke up along its length into smaller particles, termed 'beads', which were capable of movement and invariably settled onto gravel fragments. Sponges were shown to settle onto scoured rock patches during storms and were eventually covered by sediments. Subsequent survivorship and recruitment depended species specifically on sediment depth and quality. Settlement patterns matched the observed dispersion patterns of established sponges and suggested a mechanism whereby the community had originated and how structure was maintained. These hypotheses were tested experimentally in situ and in the laboratory. The importance of asexual modes of reproduction, and settlement mediated by physical disturbance appears to be a common formula for recruitment by colonial marine organisms. Recruitment events were found to explain the distributional patterns of the established population. The structure of populations of thin encrusting species on vertical reef walls was examined in a novel manner whereby the mean size and density of individuals was quantified and related to gradients in the ambient physical environment. This was achieved by examining the benthic communities on a range of subtidal reef walls of uniform slope and aspect, but which were subject to natural gradients in physical conditions. The species composition of each of the six reefs studied was similar but the size structure of communties changed within each reef system in response to gradients in light intensity and water movement. Fish and urchin disturbance was shown to be unimportant and there was no indication that storms disrupted communities. The implications of changes in size structure of populations on interactions between component species in different ambient conditions were examined. The eastern wall of two cave, archway and open reef systems was surveyed according to a stratified sampling design where each reef was divided by depth and distance along the wall. Different taxonomic/morphological groups were found at particular positions on reef walls. Thin encrusting algal species were found to be one of the most abundant groups. It was assumed that the individual size of encrusting species at each wall position reflected long term ambient environmental conditions as well as the ability of each species to maintain itself against overgrowth by neighbours. By relating the population size structure of different taxa with indices for light intensity and water movement recorded at each wall position within each reef, it was shown that community structure varied markedly on the same reef wall as a function of ambient physical conditions. Changes in proportional species composition, notably from thin encrusting algal dominated assemblages to sponge characterised assemblages, with decreasing light intensity were accompanied by changes in the individual size structure of populations. At each wall position all species exhibited similar individual size. Individual size decreased and density of individuals of all species increased with decreasing light intensity. In conditions where light levels were not sufficient to support algae, individual size of encrusting fauna increased with increasing water movement. The implications of these observations on competitive interactions between encrusting species was examined. The outcome of competitive interactions between the same species was found to vary depending on wall position.

‘Infaunal haloes’ of either decreasing or increasing abundances of individual soft-sediment species with distance from reefs have been suggested to be caused by reef-associated predators. A large-scale mensurative experiment was used to investigate the distribution of two size classes of macrofauna with distance from the reef edge across three locations in northeastern New Zealand. The role of reef-associated predators, the snapper (Pagrus auratus Sparidae) and rock lobster (Jasus edwardsii Palinuridae), was investigated using established marine reserves at each location. Consistent patterns were found in a few large-bodied fauna. The hermit crab Pagurus novizelandiae occurred more frequently near the reef edge, whilst the heart urchin Echinocardium cordatum and bivalve Dosinia subrosea were more abundant further away from the reef. Dosinia subrosea and another bivalve, Myadora striata, exhibited lower biomass at sites with higher densities of snapper and rock lobster. In contrast, small-bodied macrofauna showed no consistent patterns with distance from the reef or among sites with different predator populations. It was hypothesised that predation was driving the distribution of large bivalves. An experiment was done to investigate this model using D. subrosea. Equal densities of this bivalve were established in plots either with or without cages at sites either inside or outside of reserves. Significant predation was detected, but only inside reserves. Much of this mortality could be specifically attributed to predation by large rock lobsters, given the distinctive marks on the valves of dead D. subrosea. Inside reserves, predators are not only more abundant but also larger. It was hypothesised that different size classes of predators would result in different levels of predation. Laboratory feeding experiments were used to investigate this model. Lobsters of all sizes chose D. subrosea over the heavier shelled D. anus. Small lobsters chose to prey on small D. subrosea and large lobsters more frequently chose larger prey. The distributions of these two bivalve species at protected (large predators) and fished sites (small predators) reflected the feeding choices observed in the laboratory. Results suggested that rock lobster populations are capable, where their size structure is not truncated by fishing pressure, of controlling population-level dynamics of bivalve communities adjacent to reefs.

This thesis presents a descriptive and a functional analysis of the ecology of an island host-parasite system consisting of the Polynesian rat, Rattus exulans (Peale) and its gut helminths. The results, which include an historical perspective, are presented in the form of 7 papers or sections. Each of these examines a particular aspect of this host-parasite relationship. A review of the origin and an update of the theorised dispersal of this rat from Southeast Asia to New Zealand is given in the first section of chapter one. Previous theories have derived the New Zealand populations from a line which passed through Micronesia. In accordance with new information from the Lapita cultural assemblage, this rat is now theorised to have accompanied these "Lapita" peoples through the Bismarck Archipelago and Solomon Islands, arriving in the Tonga- Samoa region about 3600-3000 Before Present (B.P.). From here, the Proto- Polynesians then dispersed further east, taking with them the commensal R. exulans, pig, dog, and chicken. This rat is thus thought to have arrived in New Zealand, the most southern and last-settled landmass in Polynesia, in the canoes of the Maori about 1000 years ago. Information on the ecto- and endoparasites of the Polynesian rat from throughout its geographical range is collated and presented in section two. This includes the results of the two surveys (one being part of this thesis) done on the parasites of this rat in New Zealand. All populations of R. exulans sampled in these two surveys came from offshore islands, to which this rat is almost totally confined, and where, on many, it is the only rodent species present. In contrast, most of the populations sampled beyond New Zealand are now sympatric with other rodent species. For the New Zealand populations only, it was also possible to identify those parasites only accidentally associated with this host; these are listed as "transients". In section three, an attempt is made to determine the probable biogeographical origins of parasites recorded from populations of this rat on "exulans only" offshore islands of New Zealand. Such a study was possible only because of the archaeologically documented commensal relationship between rat and Polynesian man. This information, detailed in section one, together with the parasitological data base assembled in section two, provided the material for this analysis. Several "heirloom" species are identified, theorised to have been inherited by this rat during speciation somewhere in Island Southeast Asia. Parasites acquired during dispersal are divided into "old" and "new souvenirs"; the former are thought to have been acquired from sympatric rodent species in Near Oceania sometime prior to 3000 B.P., and the latter from R. rattus, R. norvegicus or Mus musculus introduced in the last 200 years during European settlement in New Zealand. The conclusions further suggest that some at least of the "new souvenir" species have been acquired by R. exulans on "exulans only" offshore islands of New Zealand by cross-transfer from other rodent species which have temporarily gained access to these islands. This theory is examined in more detail in the fourth section, and reports of such accidental colonizations of offshore islands are presented as supporting evidence. In Chapter two, the influence of habitat on the population demography of the host is investigated. Nearly 1000 rats were trapped and necropsied over a 17 month period in three different habitats on Tiritiri Matangi, an "exulans only" island at the entrance to the Auckland harbour. Rank grassland which covers most of this island formed one habitat; a second consisted of forest remnants confined to gullies, and the third consisted of the small, inhabitated, lighthouse station and farmed area at one end of the island. Between-habitat differences were observed in diet, adult longevity, mean weight of immatures, the time of onset of sexual maturation, and annual reproductive output. These results suggested several modifications to existing models of this host's demography in New Zealand. Shelter in particular appears to play an important role in the demography of this species in temperate latitudes. The effects of parasitism on potential fitness parameters e.g. reproduction, and adult mortality/survival, are examined in chapter three. Based on the results obtained in chapter two, a number of hypotheses were developed, and the predictions arising from these were tested. Few significant results were obtained; these revealed habitat and some sexual interactions with the level of infection, at certain times of the year. However, no causal relationship could be established between these effects and host reproduction or mortality. It is concluded that the helminth parasites of this rat on this island have little or no effect on these host parameters, and support the suggestion that these species constitute a depauperate and well-adapted rodent parasite fauna. The last chapter presents the results of an analysis of the effects of habitat, season, host age, and sex on the distribution and abundance of the helminths of this rat on this island. Together, the graphs and the statistical analyses demonstrate that habitat has the most important influence, significantly affecting all 7 species; this effect is of greater magnitude than the other 3 variables on 5/7 of these species. Season and age also have important effects, while sex had no apparent influence. Explanations for the observed patterns are sought in known aspects of the biology of the host in the three habitats described in chapter two, and in the life cycles of the parasites. In total, this thesis provides a comprehensive account of the ecology of the Polynesian rat and its helminth parasites on Tiritiri Matangi Island. It also identifies gaps in the existing data base, formulates certain hypotheses, and makes a number of predictions all of which will hopefully stimulate further interest in this rat and its parasites.