This thesis is concerned with the interaction between the nervous system and muscles during the metamorphosis of the waxmoth Galleria mellonella. The mesothoracic dorsolongitudinal muscles of the waxmoth arise from dorsal muscles in the larval mesothorax. Several of these dorsal larval muscles begin their transformation shortly before the larva-pupa ecdysis when: they break down to be replaced by presumptive myoblasts. These appear to arise from the nuclei of the larval muscle. In the 60-72 hours pupa all three mesothoracic dorsolongitudinal. (2dl) muscles are made up of the characteristic number of fibres and the first dorsolongtudinal muscle (dl1) has its characteristic adult shape. From the 82 hour stage onwards the muscle fibres increase in diameter until the 168 hour stage when the muscles have achieved the adult condition. Cautery of the first thoracic ganglion of the larva or section of the nerve cord between the first two thoracic ganglia leads to an adult in which most of the 2dl muscles are absent. Only the dorsal part of 2dl1 is present as normal. Up until the 24 hours pupa cautery of the first thoracic ganglion has the same results as cautery of the larva. Cautery of the 25-30 hours pupa results in an adult in which the ventral fasciculae of 2dl1 are represented by a few tiny muscle fibres. At this stage there is no sign o f 2dl2 or 2dl3 except for a few small cells which are about the same size as presumptive myoblasts. Cautery of the 40 hours pupa leads to an adult in which all of the 2dl1 fasciculae are present as smaller versions o f the normal case. Up until the 108 hours pupa cautery of the first thoracic ganglion leads to an adult in which the 2dl1 muscles are smaller than normal. An ultrastructural investigation of the small cells found in the denervated adult in the place of the 2dl muscles showed these to be similar to presumptive myoblasts. It therefore appears that the nervous system affects the ability of the presumptive myoblasts both to differentiate and then to grow as normal. Denervation of the fifth and sixth instar larvae by section of one of the connectives between the first two thoracic ganglia has the same effect on the operated side as denervation of the last larval instar. Cobalt staining of the nerve supplying the dorsal muscles of the larva shows that the motor neuron complement in both the fifth and seventh in star larvae is very similar. Cobalt staining of the nerve to the 2dl muscles of the adult reveals cell bodies which are similar in position and size to those found in the larva. The interaction between nerve and muscle during development in insects and vertebrates is discussed. From this work it is hypothesized that the nervous system affects muscle development by preventing mitosis in the presumptive myoblasts. It is also hypothesised that the adult muscles are served by the same motor neurons which were present in the later instars of the larva.

The aim of this study is to test for photoperiodic control of larval and pupal development in the New Zealand moth Planotortrix octo Dugdale. The photoperiodic response curves for larval and pupal development and especially for instar number at 17°C and 21°C indicate that a photoperiodic mechanism is involved. Superimposed on this response is the suggestion that daylength affects development rate directly, with larvae and pupae developing faster under longer photophases. This effect is especially strong for pupal development (Chapter 3). The effects of thermophotoperiods (Chapter 4), night interruption and resonance experiments (Chapter 6) provide further evidence for photoperiodic involvement. The response to resonance experiments suggests the involvement of an hourglass rather than a circadian mechanism. Larvae reared under short days accumulate significantly more lipids in the 5th and 6th instars than larvae reared under long days (Chapter 4). This finding, combined with the suppressed development rate and higher instar number under short days, suggests that a weak form of diapause may be present in this insect. This is significant in being the first recorded incidence of a photoperiodically induced diapause in a phyllophagous New Zealand insect for which a year round food supply is available (Chapter 1). By transferring insects from long to short days I found that long days have more influence than short days on larval development (Chapter 7). An attempt was made to measure juvenile hormone titres under long and short days using a Galleria bioassay. The test used was not sensitive enough however to measure any significant amounts of juvenile hormone (Chapter 8). Simulations of the experimental results were performed using a damped circadian oscillator model (Chapter 9). This model was considered the most appropriate to use, based on the experimental results and on a review of the literature (Chapter 2). Simulations showed good similarities with experimental results in most cases, but could not account for resonance responses.

The vegetable brassicas of cabbage, broccoli, and cauliflower are grown in Pukekohe for the Auckland fresh-produce markets. These brassicas are attacked by three major lepidopterous pests: diamondback moth (DBM) (Plutella xylostella (L)) (Yponomeutidae), white butterfly (WB) (Pieris rapae (L.)) (Pieridae), and soybean looper (SBL) (Thysanoplusia orichalcea (F.)) (Noctuidae). Current grower strategy to combat these pests is calendar-scheduled insecticide applications. The goal of this thesis is to develop pest management alternatives. The seasonality of these three pests is discussed. DBM and WB are each under biological control by a larval and a pupal parasitoid, but this natural control is not sufficient to allow economic harvests in cabbage and was not synchronized. No parasitoids of SBL were found. The importation of additional natural enemies is discussed. A scouting system of the percent of cabbage plants infested coupled with an action threshold of. 15%-20% infested plants, resulted in good yields in field trials and allowed up to a 50% reduction in insecticide applications over the growth period when compared to a 14-day calendar schedule. Implementation of the 15% infested threshold in commercial cabbage fields resulted in up to an 83% reduction in insecticide applications with no yield decrease in quality or quantity. Application of this 15% infested plant threshold to broccoli and cauliflower decreased insecticide applications by 40% and 17%, respectively. Study of larval biology indicated that all of the lepidopterans preferentially fed on leaves; timing of the first insecticide application in broccoli and cauliflower to coincide with floret initiation decreased insecticide applications by 80% and 67%, respectively. Laboratory and field trials comparing DBM oviposition preference, larval survivability, and parasitism rates between cabbage, broccoli, and cauliflower are discussed. Knowledge of lepidopterous pest seasonality and biology, linked to careful timing of insecticide applications to coincide with threshold levels of pests, can take full advantage of natural enemies and reduce insecticide input in the vegetable brassicas of cabbage, broccoli, and cauliflower with no decrease in crop quality.

Examination of the terminal abdominal ganglia of the New Zealand maritime earwig Anisolabis littorea (White) revealed a paraldehyde fuchsin stainable material in animals which had undergone some degree of dehydration. A Malpighian tubule bioassay preparation enabled the physiological activity of extracts of the terminal abdominal ganglia and pharmacologically actives substances to be monitored. Results obtained suggest that substances are present in the terminal abdominal ganglia which alter the Malpighian tubule secretion rate. These results are discussed in view of current theories of Malpighian tubule secretion in insects.

Biological disturbance was found to be one of the most important mechanisms control1ing community organisation in the temperate sub-tidal region. The different types of biological disturbance structuring three encrusting communities were investigated on the east coast of Northland, New Zealand. The operation of each type was determined and the rate of disturbance measured. Experimental exclusion treatments were set up to demonstrate the effects of the major disturbance agents on community structure. The urchin Evechinus chloroticus was the most abundant agent of biological disturbance and affected the widest spectrum of encrusting organisms. The abundant balistid fish Navodon scaber was another major agent of biological disturbance in this region. Disturbance of algal populations also resulted from a guild of abundant herbivorous gastropods. Two episodes of fungal/bacterial infection degraded numbers of the large sponges Ancorina alata and Polymastia fusca. The operation of the different disturbance agents was found to be generally unpredictable in both time and space. There was no escape from biological disturbance for encrusting organisms in either small or large size. Re-occupation processes on patches of free primary space were investigated both experimentally and by using settlement plates (artificial free space patches). Recruitment was found to be irregular in space and time, especially for long-lived sessile organisms. Settlement processes as well as growth and mortality of newly settled organisms were investigated with a view to understanding community development. Only one verifiable example of substrate preparation or biological succession was found to operate in the communities studied. It is postulated that community organisation is flexible and not rigidly directed along a single successional pathway. Multiple developmental pathways and multiple stable configurations are possible in the same locality, resulting from the operation of different disturbance regimes.

Invasive ant species represent a major threat to biodiversity through their detrimental impacts on native ecosystems. An important step to managing invasive ant species is determining the factors responsible for their current and potential distribution. This thesis examines the spatial distribution of invasive ant species at a range of scales. The focus is on the Pacific region, specifically Fiji and New Zealand, which have a relatively high number of invasive ant species. Taxonomic checklists of the ant fauna of Fiji and New Zealand are presented. Climate variables were used to model the distribution of, and predict suitable areas for, particular invasive ant species. In general, models had high evaluation success, especially for models that were built and tested within a region. However, the utility of models was far less when transferred and tested in new regions. A better understanding of how climate variables directly and indirectly affect a species is needed to improve the utility of species distribution modeling. On a finer spatial scale, habitat partitioning was evident at Colo-i-Suva Park in Fiji, where distinct ant communities were sampled from litter and canopy habitats. Results suggested that body size, habitat and resource utilisation interact to create opportunities for invasion, and influence the susceptibility of different habitats. Habitat also shaped the assembly of ant communities and mediated inter-specific competition in the Yasawa Islands, Fiji. Similarly, habitat was important in partitioning ant communities in New Zealand; endemic ant species were in forest, and invasive species in scrub and urban habitats. However, there was no evidence that inter-specific competition influenced community composition or the distribution of invasive ant species at regional or local scales in New Zealand. The extent and diversity of exotic ant species arriving at the New Zealand border was also investigated. A high proportion (>64%) of intercepted ants originated from the Pacific. The effectiveness of detecting exotic ant species at the New Zealand border ranged from 48-78% for different pathways, indicating a number of species remain undetected. Future work on invasive ants should focus on species-specific tolerances, and how the physical environment and small scale abiotic conditions influence distribution.

In insects, olfaction is mediated by a large family of integral membrane proteins, called olfactory receptors (ORs), that mediate the transduction of odorant binding into a neuronal signal. A functional assay for insect ORs was developed utilising calcium imaging in Sf9 cells. The Drosophila melanogaster OR, Or22a, was expressed using transient transfection, and its activity measured by monitoring increased intracellular calcium levels using a calcium–sensitive dye. The interaction of the odorants ethyl butyrate, pentyl acetate and ethyl acetate with Or22a were both dose–dependent and sensitive, with EC50 values of 1.53 x 10−11 M, 5.61 x 10−10 M and 3.72 x 10−9 M, respectively. Furthermore, Or22a expressed in Sf9 cells has a similar response profile to a range of odorants previously tested in vivo. This assay system will provide a useful tool for the investigation of insect olfactory receptor structure and function. A consensus of eleven transmembrane (TM) domain prediction algorithms suggested a model for Or22a that contains seven TM domains, reminiscent of GPCRs. To test this model empirically, the membrane topology of Or22a was determined using epitope–tagging of predicted loops followed by immunochemistry. These experiments revealed that Or22a has seven TM domains but that its orientation in the membrane is opposite to that of GPCRs, having a cytoplasmic N–terminus. This orientation was also observed for Epiphyas postvittana Or1, which suggests that this inverted topology may be common to all insect ORs. To test whether Or22a forms higher order structures, fluorescence resonance energy transfer (FRET) between cyan and yellow fluorescent proteins inserted into the intracellular loops of Or22a was employed. The third intracellular loop interacts strongly with itself in homo–multimers, with interactions between the first and first loops and first and third loops also observed. These experiments show that ligand binding ORs can form multimeric structures in heterologous cells. The co–transfection of Or83b into S2 cells had no impact on these interactions, however Or83b is likely expressed in this cell line. Finally, models of how a ligand binding OR interacts physically with the ion channel Or83b are presented, and approaches that could be used to distinguish between these models are discussed.

The Australian sheep blow fly Lucilia cuprina is an economically important dipteran pest whose circadian behavioural rhythms have been the subject of considerable scrutiny. The underlying biochemical nature of these rhythms however, has remained a mystery. The primary objective of this thesis was therefore to investigate the molecular control of circadian rhythms in L. cuprina using an integrative approach. To these ends, a dynamic molecular simulation model for L. cuprina was formulated using existing biochemical data on insect circadian clocks. The validity of this simulation model was subsequently tested at both molecular and behavioural levels. The basic molecular assumptions of the simulation model were tested by cloning a full length L. cuprina per cDNA and analysing its mRNA and protein expression levels. Isolation of the 4 Kb L. cuprina per cDNA revealed the conservation of three functional domains known to be important for circadian clock function; namely the PAS dimerisation motif (with 92% identity to D. melanogaster at the amino acid level), and the cytoplasmic and nuclear localisation domains (with 85% and 80% identity respectively). A fourth domain, the threonine-glycine (TG) repeat region, was also found to be conserved, but severely truncated in L. cuprina. No length variation was found in the TG repeat of flies collected from several different latitudinal zones, and no correlation was detected between sequences flanking the repeat and latitude of collection of flies. Thus, the contention that the TG repeat region plays a role in temperature compensation of the circadian clock is cast in doubt. Expression analyses (using quantitative RT-PCR) showed per mRNA levels to undergo diel oscillations with a period (24 h) and peak phase (Zt 12) consistent with the Drosophila data. PER-immunoreactive protein oscillations were also demonstrated, with peak immunoreactivity lagging approximately 3 h behind peak mRNA levels. The behavioural predictions of the model were tested by recording adult locomotor activity under different light regimes. The simulation model successfully predicted free-run, entrainment, the effect of short light pulses, and the effects of constant lighting on behavioural rhythms. Disparities between the simulated and real phase response cnrves for L. cuprina are hypothesised to be indicative of an ealier nuclear entry time of the PER-TIM dimer in L. cuprina compared with D. melanogaster. The three different approaches of simulation modellingo molecular analysis and behavioural investigation are integrated in the discussion in order to help provide a comprehensive explanation of circadian function in L. cuprina. The benefits of an integrated approach to the analysis of circadian function are discussed, as is the relevance of the present findings to the development of a clock-based control strategy for this economically important pest species.

The species status and host range of Heliothis armigera was investigated. DNA profiling, mitochondrial DNA sequencing, RAPD's, along with traditional morphological and crossing methods were used to investigate micro, macro and mega-population structuring. Thirty-six new host records were added, of which a number are important common weeds and crops. Mortality due to parasitoids and fungal infections were quantified on many host species. Genetic analysis of the COII and the AT-rich regions of the mitochondrial DNA showed very high levels of variation, as did the DNA profiling using the probes (CA)n and 33.15. The morphological analyses also showed variation, but to a lesser degree and without statistical significance. The variation found at all levels and in all aspects is discussed with respect to caterpillar host-plants and geographical location. All host-plant populations of caterpillars showed very high levels of genetic variability. However, the population of caterpillars found on Sulla (Hedysarium coronarium) was significantly more variable than those found on Lotus and Lucerne when compared using DNA profiling. The sequences obtained from the two mtDNA regions also showed considerable variation, a great percentage of which was uninformative. This variation did not allude to any structuring of caterpillar populations with respect to host-plant or geographical location. H. armigera is genetically a very variable species which does not equate with any population structuring present in the host-plant or geographical populations investigated here. This study has greatly increased the general understanding of this insect, and has elucidated a portion of the genetic makeup, but not helped in the development any new control methods.

Although Argentine ant(Linepithema humile), a highly invasive ant species, has been in New Zealand for at least l4 years, little is known about their ecology and potential for invasion. Increasing spread and establishment of populations throughout New Zealand is disturbing because of the devastating impacts documented on native invertebrate biodiversity overseas. The primary aim of this study was to determine the impacts of Argentine ants on arthropods in native forest habitats in west Auckland. Pitfall traps at invaded and uninvaded sites were used to quantify ant and non-ant arthropod faunas. Argentine ants did not adversely affect native host ant communities. Moreover, two ant species appeared to be resistant to invasion. Argentine ant invasion reduced the abundance of a few orders of invertebrates while several taxa were more abundant in the presence of Argentine ants. Distribution and foraging activity of Argentine ant populations were monitored in this study from 2000-2003. Also, rate of spread was investigated to evaluate whether native forest habitats would be at risk from invasion. Measurements of foraging ant trails on monitored tree trunks revealed seasonal distribution patterns involving high activity in summer/autumn and low activity in winter/early spring. Argentine ants were found to be established primarily along the edge of the forest and did not invade into the interior of the forest during the study period. An Argentine ant poisoning operation on Tiritiri Matangi Island in January 2001 provided the opportunity to document the results of the eradication trial. In addition, pitfall traps placed at two treated sites and one untreated site were used to compare pre-poison and post-poison effects on ant and non-ant invertebrate communities. Fipronil baiting at 0.01% effectively reduced Argentine ants at the study sites and very few ants were observed in both tree count and pitfall trap recordings two months after poisoning. The invasion of Argentine ants on Tiritiri Matangi Island decimated native host ants and no recovery was detected throughout the study. Several groups of invertebrates appeared to benefit from the removal of Argentine ants while a few showed no detectable changes. Conservation implications resulting from the findings of this study are discussed Also, potential future research involving Argentine ants are outlined.