The growth of P. aureofaciens PGS12 was followed in nutrient broth, on nutrient agar, and on plant roots by monitoring cell numbers, the production of the auto inducer N-acyl- L-hexanoyl homo serine lactone (HHL), and the antibiotic phenazine-l-carboxylic acid (PCA). In nutrient broth, as the growth rate declined, HHL synthesis increased rapidly. Up to 38% more HHL was produced compared to the increase in cell numbers indicating that transcription of the phenazine operon was auto induced. As the bacterial culture entered stationary phase, HHL concentration declined rapidly while PCA concentration continued to accumulate at a high rate. In stationary phase, HHL concentration continued to decline while PCA accumulated slowly. A promoterless mini-Tn5-luxAB transposon was used to generate isogenic strains of P. aureofaciens PGSI2. Strain BI03 was shown to have the luxAB reporter gene inserted in the phzB gene within the phenazine operon. Phenazine transcriptional activity (bioluminescence) was compared with the light output from a constitutive reporter, strain 117. Levels and pattern of bioluminescence from strain B 1 03 followed closely HHL production and indicated that gene expression was maximal in transition phase and silenced in stationary phase. PCA production continued in stationary phase suggesting that the protein products of the phenazine operon were maintained in the cell after down regulation. The induction of the phenazine operon started in nutrient broth when cell density was ca. 2x 108 cells mrl and HHL had accumulated to a threshold concentration of 0.63x 1 0- 6 ± O.3x 1 0- 6 ng celrl. ± On NA, cells were in a transition phase of growth for at least 9 h. The cell density was 55 to 75 times higher within a colony than in liquid culture. The maximal production of HHL and light output per calculated equivalent volumes were also between 50 and 65 times higher on NA than in broth. The maximal light output and maximal HHL accumulation per cell were similar on both media. Therefore, the increased levels on NA may be mainly due to a higher cell number in the colony. The production of the antibiotic PCA per cell was ca. 7 times higher in a colony than in NB, and the production in an equivalent volume was ca. 360 times higher in a colony. Therefore, the higher PCA concentration in the colony cannot be explained solely on the basis of an increase in cell density. The auto inducer concentration remained high within a colony for a prolonged period of time compared to the burst seen in NB. A high concentration of HHL per cell for a longer period of time may have sustained the greater production and accumulation of PCA in the colony. Similarly the transcriptional activity of the phenazine operon, as reflected by phzB::luxAB expression in strain BI03, remained maximal during this time. In contrast to laboratory culture studies, in all the experiments where P. aureofaciens PGS 12 was inoculated onto roots, neither HHL nor PCA was detected, although the bacteria colonised the bean roots and wheat seedlings efficiently. On wheat seedlings, the transcriptional activity from both reporter strains decreased during the experiment. On bean roots, bioluminescence per cell from strains 117 and B 1 03 increased 25-fold during the first 3 days and the ratio of bioluminescence from strains B 1 03 over 117 indicated an up to 5 times greater transcriptional activity from strain B 1 03 than from strain 117. The minimum levels for the detection of HHL and PCA were low. Therefore, either these compounds were produced in minute amounts, or HHL and phenazine were degraded or adsorbed onto the plant material lowering their levels below the detection limit.

The isolation procedure for the pMMO complex has been optimised to obtain a high specific activity enzyme from Methylococcus capsulatus (Bath). The enzyme is comprised of the pMMO hydroxylase (pMMOH) consisting of polypeptides 47,26 and 23kDa molecular mass. In addition to this, a putative pMMO reductase (pMMOR) was also found to be necessary to maintain propylene oxidising activity. This component was found to consist of two polypeptides of approximately 63 and 8kDa. Preliminary Nterminal sequence data of the large subunit ofpMMOR indicates that the sequence bears 70% similarity to the methanol dehydrogenase (MDH) from Methylococcus capsulatus (Bath). Therefore, we tentatively propose that the" MDH can act as a reductase component to the pMMOH. The significance of this result prompted investigations into the previous published proposals that electrons derived from the methanol oxidation reaction can be channelled back into the methane oxidation reaction by the methanol dehydrogenase, independent of NADH. Any effect of methanol to act as a reductant to pMMO in membrane preparations was lost upon isolation of the pMMO complex, indicating the necessity to maintain a fully functional methanol dehydrogenase (MDH) upon isolation. In addition to this, the in vitro electron donors of pMMO, NADH and duroquinol were found to act via distinct pathways of electron transfer (electron transport inhibitor studies). Electron paramagnetic resonance (EPR) spectroscopy data provided evidence that the copper in the active site of pMMO existed as a mononuclear copper (II) centre not a trinuclear copper centre suggested by Chan and coworkers (Chan et al., 1993; Nguyen et al., 1994, 1996a, 1996b, 1998). In addition to this preliminary data also indicates the presence of an iron centre which is only EPR visible after reduction of the complex suggesting the majority of iron in the complex is EPR invisible. The exact nature of this iron centre is still unclear. A structural study of the pMMO complex has also been undertaken using electron microscopy studies in conjunction with single particle analysis. This allowed low resolution projection maps of different views of the pMMO complex to be generated. The complex appears to exist in a polymeric state of at least a dimer, possibly a tetramer if the molecular weight analysis calculated by sedimentation equilibrium analysis is taken into account. This study has provided some insight into basic characteristics and the structure of a duroquinol-driven pMMO complex and its interaction with other electron transfer proteins.

Plankton are a signicant component of the biogeochemical cycles that impact on the global climate. Plankton ecosystems constitute around 40 % of the annual global primary productivity, and the sinking of plankton to the deep ocean (the so-called biological pump) is the largest permanent loss of carbon from the coupled atmosphere-surface ocean-land system. The biological pump need only increase by 25 % to cancel the anthropogenically-released ux of CO2 into the atmosphere. Mechanistic models of atmosphere-ocean dynamics have proved to have superior predictive capabilities on climate phenomena, such as the El Ni~no, than empirical models. Mechanistic models are based on fundamental laws describing the underlying processes controlling a particular system. Existing plankton population models are primarily empirical, raising doubts to their ability to forecast the behaviour of the plankton system, especially in an altered global climate. This thesis works towards a mechanistic model of plankton population dynamics based primarily on physical laws, and using laboratory-determined parameters. The processes modelled include: diusion and convection to the cell surface, light capture by photosynthetic pigments, sinking and encounter rates of predators and prey. The growth of phytoplankton cells is modelled by analogy to chemical kinetics. The equations describing each process are veried by comparison to existing laboratory experiments. Process-based model verication is proposed as a superior diagnostic tool for model validation than verication based on the changing state of the system over time. To increase our ability to undertake process-based verication, a model of stable isotope fractionation during phytoplankton growth is developed and tested. The developed model has been written to complement other process-based models of biogeochemical cycles. A suite of process-based, biogeochemical models, coupled to an atmosphere-ocean circulation model, will have superior predictive capabilities compared with present global climate models.

The purpose of this study was to investigate the regulatory protein (protein B) and component interactions of soluble methane monooxygenase (sMMO). sMMO is a multi-component enzyme which catalyses the oxidation of methane to methanol. It consists of three proteins, a hydroxylase, a reductase and protein B (Colby and Dalton, 1978). Protein B contains no metals, cofactors or prosthetic groups and has a molecular mass of 16 kDa. It has been shown that protein B is absolutely necessary for the hydroxylase activity of the sMMO complex and is a powerful regulator of the enzyme (Green and Dalton, 1985). It has also been found that 12 amino acids are cleaved from the N-terminus of protein B from Me. eapsulatus (Bath) to form an inactive truncate, known as protein B' and mutation of the Met12_Gly13 cleavage site to Met12 -GIn I3 to give the single mutant protein G13Q, improved the stability of the protein (Lloyd et al., 1997). Much of this work has concentrated on the study of the catalytic and regulatory significance of the 12 amino acids cleaved from protein B. Mc. eapsulatus (Bath) protein B appears to- cleave autocatalytically, generating the inactive protein B' truncate. The secondary structures of proteins B and B' were seen to be the same, although the overall structure was identified as differing slightly and protein B was shown to be capable of existing in a monomer-dimer equilibrium, whereas protein B' was identified as existing in a monomer form. An homologous protein B from Ms. trichosporium OB3b, identified as being more a-helical in character, has been shown to be more stable than Mc. eapsulatus (Bath) protein B but still undergoes the inactivating cleavage reaction to form truncates, although the cleavage sites differ between the two proteins. The construction, expression and purification of N-terminal truncates of Mc. capsulatus (Bath) protein B identified that the presence of the first 7 amino acids was essential for protein B activity within the sMMO system and a decrease in specific activity was observed as each amino acid from 1 to 7 was lost. Upon loss of the 7th amino acid, tyrosine, the truncate protein was observed to be totally inactive and also much more prone to cleavage, but unchanged in terms of secondary structure. Protein concentration was observed as having an effect on the stability of Mc. capsulatus (Bath) protein B and, the single mutant G13Q, with increased concentrations improving stability. This effect was not observed for the double mutant MI2A:G13Q, although it was shown to be more stable than the other proteins under more dilute conditions. The addition of a magnesium salt also improved the stability of protein B. Studies into the interactions of protein B with the other proteins within the sMMO complex have also been performed. Evidence that the hydroxylase undergoes a large conformational change upon the binding of the reductase and protein B has been obtained and modelled to suggest that one trimer of the hydroxylase dimer rotates by 1800 relative to the other upon complex formation. It also showed the sMMO complex to form in a stoichiometry of 1:2:2 hydroxylase:reductase:protein B. Other data suggest that -sMMO component binding occurs on only one trimer of the hydroxylase dimer under different conditions.

The capacity of reefs to recover after disturbance is fundamental to prediction of their stability. This is particularly relevant now, following the global decline of reefs during the last decades. A discrete, spatially explicit model (probabilistic cellular automaton) was developed to simulate a Caribbean coral community. Community complexity was generated from behaviour of fundamental units of corals, the polyps. Regarding background disturbance, area disturbed and patch size were investigated; both were equally important in driving coral community structure and diversity. A powerlaw model was developed to predict natural disturbances, and implemented in later testing of system dynamics. Corals were assigned differential susceptibilities to background disturbances. Results assessed against field data showed that most modelled species had realistic colony size frequency distributions (though 20% had insufficient comparison data). Following model development, recovery from single impacts (simulated warming events) was tested. Model responses indicate importance of local setting to community resilience. Individual susceptibility of species was mediated by life history strategy investment. Application of a warming sequence of predicted anomalies for this century was then introduced. Community composition changed betwee1 0-40 years from predominantly persistent, large, slow growing species to small, fecund, fast growing species. After 40 years a phase shift occurred in which algae dominated the community. It is concluded that the future may herald declines in the main Caribbean reef-building species, in ways that match several previous but largely untested speculations. This model indicates that there will be serious implications to reefs, including their numerous commercially important species. The model includes all known major life history attributes of the corals, based on real data. Structural properties of the model were tested for stability and computational efficiency. Disturbances of several types were investigated; natural background disturbance, and warming events, both as single and repeated incidents to assess recovery dynamics in the light of ongoing, intensifying climate-mediated global changes.

Variations in egg size and egg composition between females of the same species, as well as among eggs in the same clutch, have been studied in many avian species. The eggshell serves crucial functions in avian reproduction such as protection of the embryo from mechanical damage and from the invasion of micro-organisms, source of calcium to the embryo, control of gas exchange with the environment and conservation of water. But little attention has been paid to variation in eggshell, especially within-clutch variation. This thesis focuses on variations in eggshell characteristics in relation to laying order in a single species, the lesser black-backed gull (Larus fuscus). In order to evaluate a proper method for measurement of eggshell characteristics, this thesis used more than one technique to measure shell thickness, shell porosity and shell coloration. For the measurement of shell porosity, two techniques for counting pores were validated for the first time in this thesis. This study found within-clutch variations in shell porosity, mammillary layer contact area and shell coloration but not in shell thickness. The last-laid egg had a larger mammillary layer contact area and often had paler shell colour and streaks on the shell. This study found some relationships between shell structures and shell coloration. A calcium-supplementation experiment was used to investigate whether the shell formation is limited by calcium-availability. This thesis found effect of calcium-limitation on shell thickness, but no effect on shell background colour.

1-) The temperature of sea turtle nests in the Eastern Mediterranean was between 24 and 35 "C and rose by up to 10 "C during incubation. 2-) The mean incubation temperature can be used for estimating the incubation period but provides a poor prediction of sex ratio. 3-) The mean temperature during the middle third of the incubation period was closely correlated with the percent sex ratio. 4-) There was a female dominated sex ratio among the 22 nests and only one loggerhead turtle nest showed less than a 50 % female sex ratio. 5-) There was a consistent temperature difference within the nest with top eggs warmer, bottom eggs cooler and middle ones intermediate. Therefore the majority of hatchlings 11'0111 the top level in nests were females and those from the bottom level were predominantly males. 6-) Temperature differences within the nest also influenced the rate of development; the greater the difference in temperature between top and bottom the longer the time required to complete hatching of all embryos of the nest. The hatching intervals of green turtle nests were shorter than those at loggerhead turtle nests. Temperature variation between top and bottom of nests was low within green turtle nests. In general, a 1 "C temperature difference within the clutch caused a 4 day range in both hatching and emergence of hatchlings. 7-) Since the temperature within the nest and between the nests was so variable, sand air or sea water temperatures gave a poor prediction of the temperature of a nest and therefore the sex ratio. 8-) Although the predation pattern of sea turtle nests varied in relation to nest age, this predation can be reduced by screening the nest with mesh grids.9-) lnundation was one of the main abiotic factors lowering the hatching success on the beaches. Hatching success can be increased by relocating the nests to a safer area on the night of laying 10-) The mean grain sizes of sand ranged from 0.49 to 2.20 phi(<!»on 10 beaches but hatching success was not related to mean grain size of sand on the beaches. 11-) Simple embryonic staging of Mediterranean sea turtles was developed after measuring Cl set of selected morphological characteristics. The frequency of gross abnormalities among the samples was also calculated. Most common abnormalities were supernumerary and subnumerary scutes, albinos, head and jaw abnormalities and twinning l2-) The heavy metal concentrations III the tissues (yolk, liver and eggshell) of loggerhead turtle eggs and hatchlings were analysed. The concentrations of mercury, cadmium, lead, iron and copper were highest in the liver, while zinc concentrations were highest in the yolk. The concentrations of metals were similar on different beaches, except for lead concentrations in the eggshells, which varied between sites. 13 -) The genetic structure of loggerhead turtle samples from Cyprus exhibited haplotype B and green turtle samples haplotype XIII. No additional haplotypes were found. The presence of only single haplotypes suggests little variation in genetics within the Mediterranean and that these population were recently established by a small number of immigrants from the Atlantic.

The first phylogeny reconstructed solely for amblyceran genera is presented. This study, based on an extensive comparison of adult morphology and a rigorous cladistic analysis, considers generic exemplars from 4 families of amblyceran lice (Menoponidae, Boopiidae, Laemobothriidae and Ricinidae). The monophyly and evolutionary relationships of these families are strongly supported and there is good support for Menoponidae and Boopiidae as sister taxa. The relationships of the families are not concordant with the traditional hypothesis of a basal Menoponidae. The study identifies 4 supra-generic groups within the Menoponidae, which are discussed with reference to previous classifications and studies which have included amblyceran taxa. A preliminary assessment of host-parasite cospeciation is also provided. Whether a similar phylogeny would be produced from molecular data is investigated. The relationships of genera based on morphology are compared with phylogenies generated from the nuclear gene elongation factor 1a and the mitochondrial gene cytochrome oxidase I. Different methods of reconstruction used to assess their phylogeny and raw signal find that the data are largely incongruent, although there is little support for the topologies generated from the sequence data. The monophyly and relationships of families are compared between datasets and differences in rate heterogeneity between the data are also discussed. A first phylogeny for the genus Austromenopon (Amblycera: Menoponidae) and their close allies (based on the results of the morphological analysis) is reconstructed from molecular data using the mitochondrial genes COI and 12S rRNA. The molecular phylogenies obtained are generally incongruent, with most branch support located nearer the tips of the tree. No analysis recovered a monophyletic Austromenopon, although there is good support for a subset of the Austromenopon taxa, which repeatedly group together.

The aim of the thesis was the investigation of the intertidal macroinvertebrate distributions of the Dee Estuary and the factors controlling their distributions. The work consisted of a series of extensive systematic grid surveys over the Estuary in the Spring and Autumn covering a period from 1971-76, backed up with more intensive stratified random sampling in selected representative areas of the Estuary. The environmental factors investigated included salinity, temperature, particle size, percentage loss on ignition, tidal height but as the work progressed two other factors were found to be of great importance, these were the degree of protection from wave action and the effect of predation. Five consistent biological communities were defined for each of the surveys by objective numerical analysis techniques (Normal Association Analysis) and it was found possible to relate each of these communities to a unique set of physical/chemical factors. The data obtained in the initial stages of the study revealed the possibility of a balance in the relationship between the invertebrates and their predators which included other invertebrates, estuarine fish and birds. Analysis of the predation rate on Macoma balthica data revealed a mathematical relationship and an apparent minimum feeding density by the predators. The hypothesis was tested further by a variety of techniques including experiments with netted exclosures constructed on the Estuary and all the available evidence from the study substantiated the hypothesis.

For toxicity, B. thuringiensis must be taken into the larval midgut, where a community of other bacteria is already present. The culturable flora from the Aedes aegypti mosquito midgut was analysed and its role in larval growth and insect mortality was determined. In contrast to published reports concerning B. thuringiensis subsp. kurstaki, subsp. israelensis caused toxicity and larval death even in the absence of other bacteria. The pBtoxis plasmid of B. thuringiensis subsp. israelensis encodes all the mosquitocidal toxins and a number of other coding sequences. The potential effects of selected genes on host phenotype was assessed. No evidence was found for antibiotic production from putative antibiotic synthesis genes. The plasmid also carries potential germination genes organised in a single ger operon. Comparison of the germination responses of spores from strains with and without pBtoxis revealed that this plasmid could promote activation of the spores under alkaline conditions but not following heat treatment. Introduction of the ger operon on a recombinant plasmid to the plasmidless strain established this operon as the first with an identified role in alkaline activation. Mosquito midgets provide an alkaline environment and in which enhanced germination may occur. Co-feeding experiments showed that in competition to colonise intoxicated A. aegypti larvae, B. thuringiensis carrying pBtoxis, are able to outgrow the plasmid-cured strain. This indicates a selective advantage for the presence of pBtoxis. The strain carrying the recombinant ger genes also outgrew its plasmidless parent, indicating that the ger genes may be responsible for this effect, perhaps by allowing strains a head-start by germinating more rapidly in the insect gut.