Hutchinson (1961) first posed the paradox of the plankton: Why do so many phytoplankton species coexist while competing for a limited number of resources? High biodiversity has been explained in terms of the phytoplankton system not reaching an equilibrium state. Spatial and temporal variability can be achieved through externally imposed physical variability or internally-induced behaviour including periodic oscillations or irregular, chaotic behaviour. The research presented in this thesis investigates whether the non-equilibrium, chaotic response of the phytoplankton community is a likely outcome within the aquatic ecosystems. The thesis addresses the extent that chaotic behaviour remains a robust response with externally-imposed environmental variability. The sparsity of long-term time-series data and infrequent sampling inhibits the ability to verify whether marine ecosystems exhibit complex behaviour. The analysis of the time-series records of phytoplankton taxa in the English Channel suggests that chaos might occur within diatom and dinoflagellates abundance time series. However, simulations using a chemostat model for phytoplankton and nutrients suggests that time series sampled every 1-2 days for more than 5 years are required to confidently distinguish deterministic chaos from noise. The model simulations suggest that the community response depends on the phytoplankton requirement for nutrients and attributed physiological traits allowing each species to be a stronger competitor for a different resource. A wider inter-species specialization increases the likelihood of oscillatory and chaotic responses, with competitive exclusion decreasing from 50% to 20% of the cases. Higher departures from the Redfield ratio in the elemental composition of species favour complex community behaviour and act to increase biodiversity. Whether chaotic response can be sustained is sensitive to the strength of the diffusive feedback between nutrient supply and ambient nutrient concentration that acts to sustain steady-state nutrient concentrations. Including seasonal and stochastic variability in the nutrient supply reveals that the frequency of chaotic dynamics increases by 20% and 45% respectively. In addition, seasonal forcing leads to temporal variability in the strength of the chaotic response, with chaos becoming more prevalent in the summer. In contrast to a well-mixed, homogeneous environment, physical dispersal can stir different phytoplankton communities together, which might act to inhibit chaos, but at the same time enhance phytoplankton diversity. Idealised model simulations are conducted to mimic the small and large scale transport processes by including 2 or 3 well-mixed boxes. Locally generated chaotic response is sustained if: 1) there is a low rate of exchange with a strong nutrient competitor that maintains the contrasts in the community structure; 2) a strong competitor is inhibited by a high mortality rate. In addition, if the local community is outcompeted, chaos can be exported through the advection of stronger competitors that exhibit chaotic fluctuations. This study highlights the importance of understanding the interactions between ambient nutrients and phytoplankton community. The variability in the nutrient supply and connectivity between ecosystems shape the community response to inter-species competition. Complex behaviour arising from inter-species competition is suggested to have a significant contribution in driving biodiversity. Future research on assessing the extent of chaos requires extending and analysing the available time-series data obtained from stable or isolated marine provinces.

Intramammary infections (IMI) in sheep have a major economic impact through reduced milk production, premature culling and even death of ewes. This study hypothesizes that the sheep mammary gland could host a microbiome with certain members affecting SCC. Previous studies have been cross-sectional, using only one sample per subject and not conducted in sheep. This limits understanding causality; that is, how infection develops and what triggers development of disease. A longitudinal study of 30 sheep, each with two mammary gland halves, collected over 8 weeks, provided 379 milk samples and data on ewe parity and milk SCC. DNA was extracted from milk samples and processed using a bacterial 16S rRNA gene targeted PCR. Bacterial community diversity was visualised using denaturing gradient gel electrophoresis (DGGE). DGGE fingerprints were analysed in a mixed effects regression model to identify associations between individual DGGE bands and changes in SCC. Those bands associated with SCC were sequenced. Corynebacterium efficiens, Psychrobacter maritimus, Streptococcus uberis, Burkholderia cepacia, Fusobacterium necrophorum, Trueperella pyogenes, Pseudomonas chlororaphis and Psychrobacter faecalis were significantly associated with a higher SCC. Achromobacter xylosoxidans, Nocardia globerula or Rhodococcus qingshengii, Atopostipes suicloacalis, Mannheimia haemolytica, Jeotgalicoccus psychrophilus and Sharpea azabuensis were significantly associated with a lower SCC. A protocol to analyse all study samples using Illumina MiSeq sequencing was developed to elucidate the complex interactions between the sheep mammary gland microbiome and SCC. The DGGE and MiSeq results show a persistent community has been detected over time, with similarities and differences by mammary gland half, lactation and age. Associations between individual bacterial species and SCC were identified through mixed effect modelling. The DGGE results were comparable to the MiSeq results from 5 sheep. Analysis of all 379 samples by MiSeq sequencing and mixed effects models will be used to directly test the study hypotheses.

The Amaryllidaceae have characteristic isoquinoline alkaloids including galanthamine that is approved for treatment of Alzheimer’s disease. The daffodil (Narcissus pseudonarcissus) is an industrial source of this alkaloid. This project undertook analysis of the daffodil transcriptome as an approach to understanding this alkaloid biosynthetic pathway. Material from the basal plate of var. Carlton was analysed using the Roche 454 GS FLX Titanium and Illumina HiSeq platforms to assemble reference transcripts (45324 transcripts from 454, 165065 from Illumina). Annotation was via a bespoke BLAST pipeline utilizing UniProt, TAIR, Rfam and RefSeq. Further functional annotation and enrichment studies were carried out using the DAVID platform encompassing KEGG, GO and EC annotations. Illumina HiSeq sequencing of a second variety, Andrew’s Choice, was used alongside the reference transcripts to identify SNPs and transcript level differences. A bioinformatics method to determine ploidy indicated both varieties were triploid, in agreement with microscopy results. The level of selected transcripts was also assessed using qPCR. Several transcripts putatively involved in alkaloid biosynthesis were identified. Comp75950_c0_s1 showed homology to a C4H gene from peppers and could be involve in protocatechuic acid biosynthesis in daffodils. Two transcripts, Daff106212 and Contig1404, were predicted to catalyse the synthesis of norbelladine from protocatechuic acid and tyramine, and its subsequence conversion to 4’-O-methylnorbelladine. Finally, transcripts HDA57HA0AK3FX and Daff88927 were suggested for the final step in galanthamine biosynthesis, an intermolecular phenol coupling. This is the first transcriptomic comparison of two daffodil varieties and is an important resource for further investigation into genes involved in Amaryllidaceae alkaloid biosynthesis.

There is an overwhelming experimental and theoretical evidence indicating SF6 arc burning in a supersonic nozzle (known as the switching arc) is turbulent and in local thermodynamic equilibrium (LTE). Such an arcing arrangement is commonly used as the interrupter in gas blast circuit breakers. In order to reduce the development cost of gas blast circuit breakers, it is highly desirable to predict the arc behaviour under the operational conditions encountered in a power system. The major difficulty in achieving full computer aided predictive design of gas blast circuit breakers is the satisfactory prediction of the thermal interruption capability of an arc under turbulent conditions. Mathematical modelling of turbulent SF6 switching arcs, thus, forms the subject matter of this thesis. The approach for the modelling of turbulent switching arcs is similar to that for turbulent shear flows due to a direct resemblance between a nozzle arc and a round free jet both of which are dominated by shear flow. The conservation equations for switching arcs are, therefore, derived using Reynolds’s approach. The closure of these equations is based on the adoption of Boussinesq assumption to relate Reynolds stress to the time averaged velocity gradients through eddy viscosity. The turbulent heat flux is assumed to be related to Reynolds stress through a constant turbulent Prandtl number. Additional equations are introduced to determine the turbulence length scale and velocity scale required by eddy viscosity, which are provided by turbulence models. There are numerous turbulence models but none of them are specifically devised for switching arcs. The objective of the present investigation is, therefore, to choose relevant turbulence models to model turbulent SF6 switching arcs. Our choice of turbulence models is restricted to those which have been applied with success to similar flow conditions as those of a switching arc as well as their suitability for engineering application. We therefore choose the standard k-epsilon model and its two variants (the Chen-Kim model and the RNG model) for the modelling of SF6 turbulent switching arc. Since the application of the Prandtl mixing length model to SF6 switching arcs has met considerable success, this turbulence model is included in our investigation for comparison. In order to demonstrate the effects of turbulence, results based on laminar flow model are presented. Therefore, altogether five flow models have been used to study the nozzle arcs. Computational results are obtained by the five flow models under a wide range of discharge conditions in terms of different nozzle geometries, the rate of change of current (di/dt) before current zero and the stagnation pressure (P0). A detailed analysis of the physical mechanisms encompassed in each flow model is given to show the adequacy of a particular model in describing the rapidly varying arc during current zero period. The computed values of the critical rate of rise of recovery voltage (RRRV) are compared with corresponding measurements. It is found that RRRV predicted by laminar flow model is a few orders of magnitude lower than that measured, which indicates that turbulence plays a decisive role in the determination of thermal interruption capability of a nozzle arc. Of the four turbulence models, the Prandtl mixing length model gives the best prediction of RRRV when compared with experimental results. The drawback is that the value of the turbulence parameter of the Prandtl mixing length model needs to be derived from one test result for a given geometry. With our current understanding of the physics of turbulent arcs, the Prandtl mixing length model is the only turbulence model which can be used to predict the thermal interruption capability of a nozzle arc arrangement.

Entamoeba histolytica is a parasitic protozoan that infects the human digestive tract. Infection results from ingestion of cyst-contaminated food or water. To date, E. histolytica infection remains a major worldwide public health problem in worldwide endemic areas. The spectrum of clinical manifestations ranges from asymptomatic carrier to mucous and bloody diarrhea or even extraintestinal amoebiasis, usually amoebic liver abscess. Several molecular studies have been carried out to reveal novel aspects of E. histolytica infection. However, the studies focused on genomic-wide analysis comparing between E. histolytica strains are still limited. Thus, the aims of this project are to comprehensively study the comparative analysis of the whole and small RNA transcriptomes amongst nonvirulent and virulent strains of laboratory cultured E. histolytica trophozoites as well as to integrate such transcriptomic findings with the genomic data for advanced understanding of the molecular pathogenesis and virulence in amoebiasis. In this study, genome-wide transcriptome analysis using Illumina RNA-Seq technology can illustrate significant expression differences between nonvirulent and virulent E. histolytica strains. Differential gene expression analysis between nonvirulent Rahman strain and other three virulent strains (i.e. PVBM08B, HM-1:IMSS and IULA:1092:1) reveals that transcripts involved in host cell killing and mucosal invasion, nucleic acid interaction and response to oxidative stress are notably upregulated in the virulent trophozoites. InterProScan results show the upregulation of genes encoding proteolysis-related domains and the co-upregulation of cytoskeleton and actin-modulating domains in the virulent strains. Also, process ontologies related to protein degradation, cellular biosynthesis, DNA metabolism, repair and recombination, mitotic cell division, actin dynamics and response to stress are highly enriched as a core metabolism in the virulent strains, indicating the rapid growing and active metabolic state are the main drivers of virulence. However, the striking underrepresentation of ontologies involved in signaling and regulatory processes was observed in the virulent parasites. It could be inferred that reduced regulation of sensing and correctly responding to the environmental stimuli potentially enable the parasites to become virulent and subsequently cause the invasive infection. Also, NanoString validation reveals the spectrum of virulence-associated gene expression among these four strains, reflecting their different degrees of virulence. Gene copy number variation (CNV) is widespread among the genomes of the E. histolytica strains, reflecting genomic plasticity and variability in gene family content. Herein, this present data show that patterns of CNV contribute to differential expression profiles, therefore it can be extrapolated that differences in gene copy number between genomes could contribute to the variation in phenotypic attributes, including virulence, among E. histolytica strains. Also, genome plasticity can also be seen in Trypanosomes and Leishmania, suggesting that CNV is a potentially important mechanism in generating genetic diversity and regulating gene expression levels in almost exclusively asexual parasite group. For small RNA transcriptomics, the size-fractionated sRNA sequencing data demonstrate the inverse relationship between antisense sRNA abundance and target gene expression levels, strongly suggesting the sRNA-mediated regulation. Differential sRNA regulation in virulence-associated gene expression was found among strains, indicating that sRNA-mediated post-transcriptional regulation may be important in shaping the parasite virulence. In addition, this study identified the novel putative miRNA from the sRNA sequencing data using the biogenesis-based bioinformatic analysis and qPCR validation, implying that miRNA potentially play a regulatory role in E. histolytica. In summary, it can be inferred that genomic plasticity and sRNA-mediated regulation are important mechanisms of virulence modulation in E. histolytica.

Reversible protein phosphorylation is a key regulatory mechanism for controlling various cellular processes that determine cell fate. Phosphorylation status is controlled by kinases and phosphatases, which phosphorylate and dephosphorylate specific residues in target proteins respectively. Protein Phosphatase 1 (PP1) is a major serine/threonine phosphatase that is highly conserved in eukaryotes. Its catalytic subunit (PP1c) associates with numerous subunits that target and regulate its activity to specific subcellular localisations and substrates to define its role in various processes. PP1 Nuclear Targeting Subunit (PNUTS) is one of the most abundant regulatory subunits of PP1 in the nucleus. The aim of this work was to characterise a Drosophila orthologue of PNUTS and identify interacting proteins to further understand the role of PNUTS-PP1 in nuclear processes using various genetic and biochemical approaches. Mutational analysis revealed dPNUTS is essential for cell growth and survival as mutant animals die as 1st instar larvae and mutant cells are eliminated from developing tissues. PNUTS-PP1 co-localise on Drosophila polytene chromosomes, with many sites also marked by RNA Polymerase II (RNAPII). Biochemical analysis revealed Serine 5 of the C-terminal domain of RNAPII (CTD-Ser5) is a possible substrate of PP1-dPNUTS. Various RNAPII-dependent genes are misregulated in dPNUTS mutant animals, including genes involved in metabolic processes, most likely as a consequence of deregulated CTD-Ser5 phosphorylation. Another possible mechanism could be through regulation of histone modifications that determine gene expression patterns since clonal analysis revealed various histone marks are upregulated in dPNUTS and PP1 mutant cells. Methodologies to screen for regulatory enzymes affecting histone phosphorylation were also assessed. To further understand the role of dPNUTS in cell growth and development, a yeast two-hybrid approach together with biochemical analysis was used to identify dPNUTS interacting proteins. The top hit was dTOX4, the Drosophila homologue of LCP1/TOX4, which binds to PNUTS in humans. Mutational analysis revealed dTOX4 is essential for fertility in female and male adult flies. Homozygous mutant females displayed defects in nurse cell chromosome dispersal and abnormal dorsal-ventral patterning in the oocyte. Homozygous mutant males failed to complete spermatogenesis and exhibited a range of abnormal phenotypes in the testes, including defective cytokinesis, small nuclei, decondensed chromatin and Stellate crystal formation. Novel dPNUTS-interactors were also identified, which could offer further mechanistic insight into dPNUTS-PP1 function.

Toxoplasma gondii and Neospora caninum are intracellular protozoan parasites from the phylum Apicomplexa. Both parasites share many morphological and genetic features, but have diverse host preferences. While T. gondii can infect any warm-blooded animal including humans, N. caninum does not. The basis of host preference in these two parasites is unknown, but could be due to differences in tropism to specific host cells. The discovery of differences between host cell responses could lead to a better understanding to why T. gondii can lead to disease in humans and N. caninum cannot, specifically related to infection of the CNS. To investigate the differences in host cell response, the differential expression patterns of host-cell transcripts and proteins as well as protein secretions and dopamine level during the invasion of human astrocyte (HA) cells by both Toxoplasma and Neospora were investigated. Proteomic investigations were achieved by global proteomic profiling of human astrocytes during infection using two dimensional difference gel electrophoresis (2D-DIGE) and a quantitative label-free approach. Transcriptomic investigation of infected human astrocytes was carried out using an RNA-Seq approach. A number of differences in host cell response genes/proteins were noted between Toxoplasma and Neospora infection. Transcriptome analysis revealed differential expression of host cells transcripts associated with cell adhesion and cytokine receptor interaction pathways. Differences were observed in cell adhesion molecules (such as VCAM1 and NCAM2) and chemokines (such as CCL2), which were up-regulated in T. gondii infected cells compared to N. caninum infection. Whereas up-regulation of neutrophil activation was identified in N. caninum infected cells compared to T. gondii infection. Host cell secretome analysis during infection identified significant changes in many secreted proteins such as cytokines chemokines and cell adhesion molecules (i.e. CXCL1, CXCL2, IL8 and ICAM1). This data suggested that the two parasites stimulate different host cell responses. Dopamine acts as a neurotransmitter in the brain and the induction of dopamine has been found to provoke behavioural changes in the host during infection by T. gondii. Dopamine levels were measured in dopaminergic PC-12 cells during infection by T. gondii and N. caninum through tachyzoite conversion to the bradyzoite stage. Increased levels in dopamine production was observed in dopaminergic cells infected with T. gondii, compared to no changes in cells infected with N. caninum. This result suggested that T. gondii modifies the host’s behaviour (rodents), whereas the successful vertical transmission of N. caninum in cattle has led the parasite to not need to alter changes in the host’ brain. Overall, differences were mostly observed in molecules associated with cell adhesion and cytokine receptor signalling pathway, which may suggest why T. gondii is more successful in host cell infection of the CNS compared to N. caninum. Dopamine measurements of host cells during infection has also confirmed that T. gondii induces significant production of dopamine in infected host cells compared to N. caninum infection, suggesting different modulation of the host cell between the parasites in the CNS.

Iron in coastal and sea water is readily complexed with organic matter (ligands) causing the formation of different species of iron, which affects its biogeochemistry. For this reason it is of interest to determine iron speciation in addition to its concentration. The chemical speciation of iron in coastal and sea water is typically determined by cathodic stripping voltammetry (CSV), making use of ligand competition between an electroactive ligand added to obtain the CSV signal and the natural ligand to determine the complex stability of the natural species. This thesis aims to identify the main ligand responsible for binding iron. Thereto, I first set out to improve the existing method of iron speciation in seawater using salicylaldoxime (SA) (as an added competing ligand) and subsequently used this method to determine the complex stability and concentration of the unknown Fe-binding ligands. Using another CSV method the concentration of humic substances (HS) was determined in samples from estuarine, coastal, shelf and ocean origin. Particular attention was given in this dissertation to the comparison between humic substances and natural Fe-binding ligands in all samples. The improved SA method has a much better sensitivity due to two effects: 1) the use of dissolved oxygen (DO) as an oxidant was found to give a catalytic effect; this improved the sensitivity greatly, and also stabilized the pH. 2) the analytical conditions of using SA were re-optimised and best sensitivity is now obtained at 5 µM SA (~5 x less than the pre-existing method). Interpretation of the experiments showed that the complex responsible for adsorption on the mercury drop electrode is FeSA, whereas the FeSA2 species does not adsorb. The complex stability for complexes of Fe with SA (FeSA and FeSA2), in pH 8 seawater, were calibrated over a range of SA concentrations and salinities. The similarity of the stability constants (log K’Fe’SA and log B’Fe’SA2) determined via calibration against EDTA to those calibrated without EDTA shows that the speciation of Fe with SA and EDTA is well understood. The improved method is applied to a mixed depth Celtic Sea sample, and two GEOTRACES samples from the Atlantic, at a SA concentration of 5 µM. Ligand concentrations were 1.47 and 1.49 nM in the GEOTRACES water (log K’Fe’L values of 11.1 and 11.9) and 2.53 nM in the Celtic Sea water (log K’Fe’L =11.5). Application of the method to ligands added to seawater gave log K’Fe’L values of 11.6  0.1 for humic acid (Suwannee River) and 12.2 0.3 for a siderophore (desferrioxamine B). The improved SA method was also applied to samples from the Mersey estuary and Liverpool Bay (salinity between 19 and 32). The concentration of unknown Fe-binding ligands was virtually the same as that of Fe-binding humics and both are greater than [Fe] in all samples, indicating that there was an excess of the ligand concentration and that almost the entire ligand concentration (95%) could be ascribed to humic substances. The average complex stability (log K’Fe’L) was 11.2  0.1, the same as for the iron-humic species (log K’Fe’-SRHA). Copper additions demonstrated competition between Cu and Fe for the natural ligands. Multiple analytical windows (MAWs) were used to determine the iron speciation in the Mersey estuary by varying the concentration of salicylaldoxime (SA). Data fitting of the individual titrations was compared to fitting of all data simultaneously (MAW fitting) giving good agreement. Individual titration fitting as well as the MAW fitting demonstrated the presence of only one ligand class in all samples (which was identified as humic substance in the previous work). Measurement of the composition of dissolved organic carbon (DOC) indicates the presence of terrestrial as well as microbial sources of organic matter in the estuary. The fraction of HS in the DOC amounted to between 47 and 25 % between salinities of 19 and 31. Application of the improved method to samples from the shelf and the Atlantic ocean (vertical profile) shows that the Fe-binding ligands concentration is greater than Fe concentration in all shelf and ocean samples. The complex stability of the Fe species in the Atlantic and over the shelf is identical within a narrow range of log K’Fe’L values of 11.2  0.1 (Atlantic profile) and 11.2  0.2 for the shelf samples. Comparison of the iron-binding humics and complexing ligands shows that the humic substances account for 95 to 98 % of the total ligand for iron in the shelf and ocean waters. This revolutionary finding challenges the paradigm that iron-binding ligands largely originate from in-situ production by microorganisms and much facilitates the modeling of the background ligand concentration that controls the global oceanic distribution of Fe. Data fitting of the complexometric titrations in all samples from the estuarine, coastal, shelf and ocean origin demonstrated the presence of only one ligand class.

The research reported here examines the usability of Computer Assisted Assessment(CAA) and the development of domain specific heuristics. CAA is being adopted within educational institutions and the pedagogical implications are widely investigated, but little research has been conducted into the usability of CAA applications. The thesis is: severe usability problems exist in GAA applications causing unacceptable consequences, and that using an evidence based design approach GAA heuristics can be devised The thesis reports a series of evaluations that show severe usability problems do occur in three CAA applications. The process of creating domain specific heuristics is analysed, critiqued and a novel evidence based design approach for the design of domain specific heuristics is proposed. Gathering evidence from evaluations and the literature, a set of heuristics for CAA are presented. There are four main contributions to knowledge in the thesis: the heuristics; the corpus of usability problems; the Damage Index for prioritising usability problems from multiple evaluations and the evidence based design approach to synthesise heuristics. The focus of the research evolves with the first objective being to determine If severe usability problems exist that can cause users d?ffIculties and dissatisfaction with unacceptable consequences whitct using existing commercial CAA software applications? Using a survey methodology, students' report a level of satisfaction but due to low inter-group consistency surveys are judged to be ineffective at eliciting usability problems. Alternative methods are analysed and the heuristic evaluation method is judged to be suitable. A study is designed to evaluate Nielsen's heuristic set within the CAA domain and they are deemed to be ineffective based on the formula proposed by Hanson et al. (2003). Domain specific heuristics are therefore necessary and further studies are designed to build a corpus of usability problems to facilitate the evidence based design approach to synthesise a set of heuristics, in order to aggregate the corpus and prioritise the severity of the problems a Damage Index formula is devised. The work concludes with a discussion of the heuristic design methodology and potential for future work; this includes the application of the CAA heuristics and applying the heuristic design methodology to other specific domains.

This thesis presents a surgical needle designed for bone biopsy, based on an ultrasonic-sonic drilling mechanism. Bone biopsy is an invasive diagnostic procedure where a bone sample is extracted for clinical analysis. For conventional bone biopsy methods, closed biopsy is normally adopted and uses a core needle. An intact and viable biopsy sample is required for clinical analysis. However, a particular limitation of closed biopsy is that the microarchitecture of the biopsy sample can be easily damaged due to the large force which is applied through the core needle to penetrate bone. In some cases, the bone biopsy samples are fractured or crushed during the biopsy process. Power ultrasonic surgical devices have improved many aspects of bone cutting procedures, such as lower cutting force, higher accuracy, and better preservation of the tissue around the cutting site. In this study, an ultrasonic-sonic needle (US needle) system is designed and used to extract an intact biopsy sample and the penetration performance is evaluated by the effective impulse delivered to the target. The ultrasonic-sonic drilling mechanism was originally invented for rock drilling in low environments. In the US needle system, a free mass oscillates between an ultrasonic transducer-horn and a surgical needle, converting the ultrasonic frequency vibration of the horn to sonic frequency vibration of the needle. Compared to other ultrasonic surgical devices that directly transfer the ultrasonic vibrations from the cutting tip to the tissue, the US needle allows sufficient time between impacts with the free mass for the tip vibration amplitude to be re-established in the horn. This can maintain penetration progress of the needle into bone, where the rate of progress has been shown to be proportional to the effective impulse delivered by the needle to the bone. To maximise the effective impulse, a numerical model is developed to simulate the dynamic behaviour of the needle system and optimise the US needle. To build the US needle system, the design and optimisation of the ultrasonic transducer-horn were investigated with the finite element method and experimental modal analysis, ensuring that the transducer-horn operates at the tuned frequency (50kHz) with a pure longitudinal mode. The configuration of the ultrasonic horn determines the momentum transferred to the free mass and hence also affects the effective impulse delivered to the target. The shape and dimensions of the ultrasonic horn were determined through the finite element model of the ultrasonic horn impacting the free mass, which focused on maximising the post-collision velocity of the free-mass. The dynamic components of the US needle were also investigated. A numerical model representing the dynamic behaviour of the needle system was developed, allowing the optimisation of each dynamic component, maximising the effective impulse delivered to the target. Each dynamic component of the US needle was modelled as a mass-spring-damper (MSD) system, which constituted the whole system dynamic model. The numerical model was validated by experiments using a prototype needle. The free-mass velocity, needle velocity and impact force predicted by the numerical model were compared with the results measured from experiments using 3D laser Doppler vibrometry, an ultra-high speed camera and a load cell, respectively. The numerical model results exhibit good agreement with the experimental results, indicating the numerical model can be used as a predictive tool to evaluate the performance of the US needle when different configurations are implemented. The configuration of the US needle is studied to maximise the effective impulse by the numerical model, through optimisation of the mass of the free mass, spring rate and spring pre-load. An optimised configuration of the US needle was determined by the numerical model and validated by experiments. The resulting prototype of the needle device was tested in ovine femur in vitro and was demonstrated to retrieve a cortical bone biopsy sample with a more cylindrical geometry, smoother surface and more intact sample than a cortical biopsy sample retrieved using a conventional trephine needle. Moreover, the penetration performance of the US needle was also compared with an ultrasonic resonant needle where the ultrasonic transducer and surgical needle resonate at the same frequency.