Search

Reliable deterministic prediction of earthquake occurrence is not possible at present, and may never be. In the absence of a reliable deterministic model, we need alternate strategies to manage the seismic hazard or the risk. This involves making statements of the likelihood or earthquake occurrence in space and time, including a fair and accurate description of the uncertainty around statements used in operational decision-making. Probabilistic Seismic Hazard Analysis (PSHA) and Operational Earthquake Forecasting (OEF) have the role of providing probabilistic statements on the hazard associated with earthquakes on long-term (decades to centuries) and short-term (days to decades) time frames respectively. Both PSHA and OEF rely on a source model able to describe the occurrence of earthquakes. PSHA models are commonly modelled using a spatially-variable Poisson process to describe earthquake occurrence. Therefore, they are calibrated on declustered catalogues which retains only the largest earthquakes in a sequence. OEF models, on the other hand, are commonly time-dependent models which describes the occurrence of all the events above a certain magnitude threshold including dependent events such as aftershocks or swarms. They are calibrated on the full earthquake catalogue and provide accurate descriptions of the clustering process and the time-evolution of earthquake sequences. The Epidemic-Type Aftershock Sequence (ETAS) model is the most commonly used model as time-dependent seismicity model and belongs to the general class of Hawkes (or self-exciting) processes. Under the ETAS model, any earthquake in the sequence has the ability of inducing (or triggering) its own subsequence of earthquakes in a cascade of events, as commonly observed in nature. The earthquake catalogue is then the union of a set of events occurring independently from each other (background events) and a set of events which have been induced or triggered by another (aftershocks). The reliability of PSHA or OEF strategies depends upon the reliability of the source model used to describe earthquake occurrence. In order to improve the source model, we need the ability to (a) incorporate hypotheses on earthquake occurrence in a model, and (b) validate the model against observed data. Both tasks are problematic. Indeed, the complex mathematical form of the ETAS model requires ad-hoc methodologies to perform inference on the model parameters. These methodologies then need further modification if the classical ETAS model is adjusted to introduce new hypotheses. Comparing forecasts produced by models incorporating different hypotheses which are and calibrated with different methods is problematic because it is difficult (if not impossible) to determine where the differences in the forecasts are coming from. Therefore, a unique framework capable of supporting ETAS models incorporating different hypotheses would be beneficial. Similarly, the validation step has to be done on models calibrated on the same data and producing forecasts for the same spatio-temporal region. Moreover the validation must ultimately be done against future data, unknown in the moment in which the forecasts are produced, to ensure that no information about the data used to validate the models is incorporated in the models themselves. Hence, the Collaboratory for the Study of Earthquake Predictability (CSEP) has been founded with the role of gathering forecasting models and running fully-prospective forecasting experiments in an open environment. CSEP ensures that the models are validated fairly and using a set of community-agreed metrics which measure the agreement between forecasts and data on the outcomes. In this thesis, I present and apply a new Bayesian approximation technique for Hawkes process models (including ETAS). I also demonstrate the importance of one of the statistical properties that scores used to rank competing forecasts need to have in order to provide trustworthy results. The Bayesian framework allows an accurate description of the uncertainty around model parameters which can then be propagated to any quantity of interest. In the context of Bayesian statistics, the most commonly used techniques to perform inference are Markov Chain Monte Carlo (MCMC) techniques which are sampling-based methods. Instead, I use the Integrated Nested Laplace Approximation (INLA) to provide a deterministic approximation of the parameter posterior distribution instead of the random sampling. INLA is faster than MCMC for problems involving a large number of correlated parameters and offers an alternative way to implement complex statistical models which are infeasible (from a computational point of view) with MCMC. This provides researchers and practitioners with a statistical framework to formulate ETAS models incorporating different hypotheses, produce forecasts that accounts for uncertainty, and test them using CSEP procedures. I build on the work done to implement time-independent models for seismicity with INLA which provided a framework to study the effect of covariates such as depth, GPS displacement, heatflow, strain rate, and distance to the nearest fault but lacked the ability to describe the clustering process of earthquakes. I show that this work can be extended to include time-dependent Hawkes process models and run in a reasonable computational time using INLA. In this framework, the information from covariates can be incorporated both in modelling the rate of background events, and in modelling the number aftershocks. This resembles how information on covariates is incorporated in Generalized Linear Models (GLMs) which are widely used to study the effect of covariates on a range of phenomena. Indeed, this work offers a way to borrow ideas and techniques used with GLMs and apply them to seismicity analyses. To make the proposed technique widely accessible, I have developed a new R-package called ETAS.inlabru which offers user-friendly access to the proposed methodology. The ETAS.inlabru package is based on the inlabru R-package which offers access to the INLA methodology. In this thesis, I compared our approach with the MCMC technique implemented through the bayesianETAS package and shows that ETAS.inlabru provides similar results to bayesianETAS, but it is faster, scales more efficiently increasing the amount of data, and can support a wider range of ETAS models, specifically those involving multiple covariates. I believe that this work provides users with a reliable Bayesian framework for the ETAS model alleviating the burden of modifying/coding their own optimization routines and allowing more flexibility in the range of hypotheses that can be incorporated and validated. In this thesis, I have analysed the 2009 L'Aquila and 2016 Amatrice seismic sequences occurred in central Italy and found that the depth of the events have a negative effect on the aftershock productivity, and that models involving covariates show a better fit to the data than the classical ETAS model. On the statistical properties that scores needs to posses to provide trustworthy rankings of competing forecasts, I focus on the notion of proper scores. I show that the Parimutuel Gambling (PG) score, used to rank forecasts in previous CSEP experiments, has been used in situations in which is not proper. Indeed, I demonstrate that the PG score is proper only in a specific situation and improper in general. I compare its performances with two proper alternatives: the Brier and the Logarithmic (Log) scores. The simulation procedure employed for this part of the thesis can be easily adapted to study the properties of other validation procedures as the ones used in CSEP or to determine important quantities for the experimental design such as the amount of data with which the comparison should be performed. This contributes to the wider discussion on the statistical properties of CSEP tests, and is an additional step in determining sanity-checks that scoring rules have to pass before being used to validate earthquake forecasts in CSEP experiments.

Head and neck squamous cell carcinoma (HNSCC) is the eight most frequent cancer in the world, and approximately 2/3 of the patients are diagnosed at locally advanced stages (stage III or IV). Despite improvements in HNSCC management and the aggressiveness of first-line curative treatment, 65% of treated patients experience local recurrence or distant metastasis. Moreover, patients diagnosed with recurrence or distant metastasis have a poor prognosis (from 6 to 12 months) and 5-years survival less than 50%. From 2016 the clinical practice was revolutionized by the introduction of immuno-checkpoint inhibitors, approved for the treatment of recurrent/metastatic HNSCC patients. Nevertheless, only a small subset of patients respond to this therapy and currently predictive biomarkers are still under investigation. Herein, we investigated the tumor biology of R/M HNSCC patients, platinum-refractory, enrolled in the phase IIIb clinical trial Nivactor (EudraCT Number: 2017-000562-30), in which patients were treated with nivolumab. Across the study of single biomarkers and the extensive profiling through genomic and transcriptomic analyses, we aimed to characterize the tumor molecular peculiarity of patients that experienced response or those with the longer survival. While the prognostic/predictive role of Programmed Cell Death Ligand-1 (PD-L1; studied by IHC), Tumor mutational burden (TMB) and microsatellite instability (MSI) appeared to be relatively limited for R/M HNSCC patients, 8 expression signatures (retrieved from literature) showed up significant association with survival and contributed to highlight and extricate the extreme complexity of the tumor microenvironment of HNSCC, which appeared to be strongly immunosuppressive (suggesting and corroborating the activation of several mechanisms of immune evasion). Nevertheless, the testing of previously identified six HNSCC subtypes (De Cecco et al.) with specific biological and prognostic characteristics, indicated for two of them a strong prognostic role and a significant correlation with response. In conclusion, the current study demonstrated the strong relevance of gene expression signatures in HNSCC context (over the mere study of somatic mutations) to identify the biological features associated with benefits from immunotherapy. However, additional analyses for the validation of their significance are required.

This thesis provides the first extensive exploration of Italian crime fiction in the years that span between national unification in 1861 and the outbreak of the First World War in 1914. This is a transformative yet problematic period in Italian history, marked by political turmoil and social instability, turbulent processes of industrialisation and urbanisation, shifts in sexual and domestic organisation, controversial scientific discoveries, but also by a pervasive tension between the rational and supernatural, in which apparently incompatible fields such as science, technology, spiritualism, and mesmerism interacted and intertwined. Positivist criminology, which directly participated in the construction and normalisation of the new body politic by re-conceptualising the idea of delinquency while also turning to the scientific study of occult manifestations, perfectly encapsulates such ambivalence. Given its prominent role in the process of nation-building, its dialectical and ambiguous relationship with the occult as well as its inherently Gothic nature, positivist criminology constitutes the privileged angle from which to look at how different discourses on criminality engendered a multifarious response in the field of literary fiction. In this thesis, I adopt an interdisciplinary approach, framing readings of literary works with the study of contemporaneous developments in criminology, with a view to provide a fuller and more nuanced picture of late nineteenth and early twentieth-century Italian culture and literature. The aim of this thesis is twofold. First, it seeks to reassess the history and development of a literary form that has long been dismissed for being merely escapist, unproblematic, and conservative. I will bring to light an aesthetically and ideologically complex and even confrontational body of work, which taps into fears of gender, class fluidity, deviant selfhood and problems of insanity, and the unaccomplished disjunction of science and occultism. Second, it aims to provide new insights into the relationship between science and literature, by intervening in the ongoing debates concerning the ambiguity of positivist criminologists' work and legacy. Far from merely reflecting reality, crime literature intrudes upon science, manipulates its most controversial outcomes, and fuels socio-cultural fears and anxieties.

This thesis researches seventeenth- and eighteenth-century timber roof structures in Scotland,which had been largely forgotten up to now. Previous studies fail in characterizing the extent and nature of historic timber roofs in Scotland, especially post-1650, mainly because of the scarcity of measured surveys. This lack of knowledge, together with the lack of specific standards and training for the assessment of existing timber structures, hinders good conservation practice. Scottish professionals rely mainly on traditional methods and past experience, which makes it difficult for them to reach a confident structural assessment. The result is that many roofs are heavily altered or replaced. The thesis investigates the extent, nature and condition of seventeenth- and eighteenth-century timber roof structures in Scotland in order to raise awareness about their value and contribute in improving conservation practice. A database of 1500 Scottish buildings of the period has been created and a representative sample of 59 original timber roofs has been surveyed. Archival research has also been carried out. This has allowed tracing the historic development of seventeenth- and eighteenth-century timber roof structures in Scotland and demonstrating that they form a heritage of great extent and quality. For the first time overall structural arrangements, construction methods and details, the species and quality of the timber employed, the role of architects, wrights and patrons in the design process and foreign influences have been comprehensively described and classified. The data gathered during the surveys has also been used to identify typical pathologies and possible causes for each roof structural typology. Moreover, the assessment techniques used during the surveys have been evaluated in terms of effectiveness, reliability, costs and training needed, in order to inform specific training for conservation professionals.

We present a new model to simulate multispecies reactive transport on pore space images. We solve the Navier-Stokes equations and the advection-diffusion equation for concentration fields on an unstructured grid using the finite volume method. We couple it with the chemical model Reaktoro, which we use to calculate the chemical equilibrium in each grid cell, considered as a completely mixed batch reactor. We validate the model against analytical solutions and experimental data, and investigate, for a range of Péclet numbers, the interplay between transport and reaction for multispecies reactive transport in a 3D bead pack where two streams of reactants at different pH are injected in parallel. We analyse the distribution of species and the rates of formation and consumption and find that, despite the relative homogeneity of the bead pack, the concentration fields of the products can be asymmetric because of the interplay between transport and chemical equilibrium. We observe that lower Péclet numbers give rise to higher relative yields because of increased transverse mixing by diffusion. However, higher absolute yields are obtained at higher injection velocities because of larger amount of matter available for reaction. Reaction is more favoured in the faster-flowing regions. However, this effect is more marked for species for which advection is the dominant mechanism of transport to reactive sites, as opposed to diffusion-mediated reactions where the full velocity distribution is sampled before reaction occurs. Furthermore, we study multispecies mixing and reaction in a more heterogeneous carbonate sample. We use a micro-CT image of Portland limestone containing both macroporosity resolved at the image resolution, and sub-resolution microporosity which is quantified using difference imaging. We extend our solver to accommodate transport and reaction in microporous regions. We demonstrate how the highly variable flow field in carbonate allows reactants and products to disperse more rapidly compared to the more homogeneous bead pack, resulting in a highly non-uniform reaction rate and concentration distribution. This is due to a complex interplay between advection-dominated flow and reaction in the connected fast-flowing macroporous regions, and diffusion-mediated transport and reaction in microporosity. Multispecies mixing and reaction in natural rocks are much more complex than hitherto observed, which needs to be considered in reactive transport studies at different length-scales. Finally, we test the applicability of our code to study dissolution. We first validate our code on multispecies dissolution of a single sphere and a semi-infinite solid. Then, we study dissolution on a micro-CT image composed of dolomite and calcite in ratio 10:1. We find that fast channels are strongly associated with dissolved dolomite, high concentrations of reactants, and low concentrations of products.

This thesis is a study of similarity and diversity within a major US based hotel MNE encompassing transitional and peripheral countries. The study centres on the application of Human Resource (HR) policies and practices through evidence gathered from seven hotel properties representing the totality of fully owned subsidiaries within its European division, including developed economies (United Kingdom, Germany, and Switzerland) and the transitional periphery (Azerbaijan and Kyrgyzstan). As such, it includes the relatively unresearched transitional periphery countries in the Caucasus and Central Asia, while exploring the hotel industry, which has been rather neglected by mainstream studies in International Human Resource Management (IHRM). A comparative analysis highlights both variety and commonality across overseas national capitalist archetypes covering liberal market, coordinated market, and transitional periphery economies (LME, CME, and TPE, respectively). Information was collected from HR executives through quantitative questionnaire and in-depth interviews looking at similarity and diversity according to Whitley's (1999) defining national features of employment and work relations: employer-employee interdependence and employee delegation. Multiple research methods were employed in a process of triangulation, encompassing qualitative data analysis with small sample inferential statistics, document analysis, and participant observation. The key findings suggest considerable, but bounded, variety in the application of HR policies and practices reflecting conflicting pressures within subsidiaries, ranging from head office policies and directives, to those posed by differing institutional contexts and related enforcement capabilities. The latter compelled the MNE to devolve a certain degree of autonomy to overseas HR executives, which was especially evident in relation to compliance with local labour law. In the case of TPEs, this also reflected embedded clanism and clientelism. Despite the HR Manual centring on anticipatory visions and hopes of relentless advancement, this research indicates a persistent disconnect between corporate aspirations and subsidiary HR executive perceptions. Indeed, the firm's sophisticated HR system, in reality, only encompassed a minority of highly skilled core employees, and neglected the vast segment of blue collar workers. From a methodological perspective, the thesis also highlights how surveys and in-depth interviews can yield very different results; whilst the former may lack detail and nuance, they afford respondents a veil of anonymity to express potentially controversial views, especially as close-ended answers involve agreement or disagreement with statements, rather than the active articulation of viewpoints. The thesis concludes with theoretical implications, in three broad areas. Firstly, it locates the tensions between company aspirations and reality within the contemporary sociological literature on anticipation (Walsh, 2004); the latter suggests that expectations can serve as a powerful unifying narrative in the absence of corroborating material realities, even if promises of universal betterment remain elusive. Secondly, it highlights the centripetal forces within MNEs whilst rejecting the assumptions of universal country of origin and domicile effects; as such, it highlights the value of recent advances in business systems theory that see the MNE as a partially, and diversely, rooted actor that spans institutional domains in an uneven and contested fashion. Thirdly, the study sheds new light on institutional arrangements in the transitional periphery and deepens the emerging conceptualisation of the Transitional Periphery capitalist archetype.

Impulse responses or transfer functions are descriptions of acoustic and audio transmission channels, which completely characterise point-to-point propagation of sound in room acoustics or input-output relationships of linear systems in electroacoustics. Measurements of impulse responses or transfer functions are routinely carried out first to determine critical room acoustics parameters of enclosures such as concert halls, theatres and auditoria, or technical specification of electroacoustic transducers, i.e. loudspeakers and microphones. In room acoustics measurements, tradition techniques employ noisy testing signals as probe stimuli, which are unpleasant and intolerable to audiences. This hinders occupied measurements to be taken in many cases. Predicted in-use parameters from unoccupied measurements are known to be unreliable and problematic. It is also well appreciated in room acoustics research community that the use of musical or music-masked probe stimuli can mitigate problems of occupied measurements. It is therefore hypothesised as a starting point of this thesis that the use of musical tone like stimuli or musically masked testing signals can be used to determine impulse responses or transfer functions. Based on the above hypothesis, this thesis develops a new hybrid technique, in which narrow band linear chirps, called “presto-chirps” centred on musical notes are used to measure impulse responses in low to mid frequency bands, and music-masked maximum-length-sequences are deployed to obtain those in higher frequency bands. Broadband impulse responses are then obtained by combining the measured lower and higher frequency impulse responses. To test the hypothesis and identify the potential and limitations of the developed technique, mathematical formulation and analysis, computer simulations and real room measurements have been carried out and documented in this thesis. Investigation results show that purposely tailored and windowed narrow chirps that emulate musical tones can be used as probe stimuli to measure impulse responses or transfer functions with an uncompromised accuracy. It is found that Hanning windows are almost optimal for this application. This method covers frequency ranges commonly quoted in room acoustics investigations. Music-masked maximum-length sequences are found to be able to obtain in impulse responses or transfer functions in higher frequency. However, if completely masked stimuli are sought, the resulted signal to noise ratios in the measurements is limited, or the required averaging is going to be overly prolonged. Nevertheless, the masking music can still potentially be used as a distracter to make the audience more forgiving to the hissing noise from maximum length sequences, facilitating the occupied measurements.

The anned stmggle carried out by leftist groups in Italy during the 1970s and 1980s has been the subject ofmuch debate in tenns of both public opinion and the mass media from its very start to the present day. It represents a fundamental and dramatic phenomenon in recent Italian history, a phenomenon that involved the death of 128 victims (and of 68 militants), hundreds of serious woundings and robberies; it was a long-lasting and widespread movement, supported by - at least - 30,000-40,000 people and involving a number ofactive groups. This research concentrates on the personal image of some of the militants in the largest of such groups, the Brigate Rosse (Red Brigades). Two kinds of images are analysed: the ones created by the mass media and the ones the militants produced through their own autobiographical writings (more than twenty such texts have been published). This thesis analyses what was said or written about the militants in question during the Seventies and Eighties, not only in magazine and newspaper articles and historical essays, but also in movies, songs and biographies. The analysis aims to reach a better understanding ofhow Italian society views these individuals today, and how those views have been produced. Most Italians see the 'ex brigatisti' just as they did thirty years ago, and the juxtaposition of these public frozen identities with the personal ones held by the militants, as revealed in their autobiographies, is one of the two main goals ofthis study. A second, related goal of this research, is to carry out a detailed textual analysis of four selected texts (the autobiographies of Curcio, Peci, Franceschini and Balzerani), tracing the writers' stylistic and stmcturai choices in order to understand how they attempt to create their personal identity and present it to the reader. The Primary Sources of my research consist in the texts published by exmilitants, which include primarily autobiographies, but also essays, poetry, novels, and other material. Secondary Sources consist mostly of texts about the anned stmggle in Italy written by Italian and foreign experts. These include newspaper articles, biographies, interviews, and critical works on the creation ofpersonal and collective identity.

This study examined the second language (L2) development of adult learners of Spanish at three levels of proficiency during and after a semester of instruction. A fundamental goal was to identify cognitive and psychosocial individual differences (IDs) that can explain between-learner variation over time in order to expand our understanding of the role learner IDs play in L2 learning and retention. This study also advances a line of research aiming to improve the validity and reliability of tests designed to measure knowledge "of" (implicit) and "about" (explicit) L2 grammar. Eighty-seven learners in their 2nd, 4th, and 6th semester of university Spanish instruction completed four tests of cognitive ability measuring L2 aptitude, working memory capacity (WMC), phonological short-term memory (PSTM) and processing speed as well as two Likert-type questionnaires measuring different facets of L2 motivation. Implicit and explicit knowledge of ten linguistic structures in Spanish were measured through an elicited oral imitation task and untimed grammaticality judgment task used in previous research with several methodological improvements to strengthen their construct validity. A multi-faceted statistical approach was taken to answer research questions employing both traditional and innovative hierarchical linear modeling techniques. Results suggest that 2.5 months of classroom instruction significantly enhanced performance on both measures of L2 knowledge for all proficiency groups with continued gains one month later. Higher proficiency groups significantly outperformed lower proficiency groups at each point in time and there was evidence of faster and more efficient use of implicit linguistic knowledge with increasing proficiency. However, exploratory and confirmatory factor analyses revealed that the validity of a two-factor model varied according to learners' proficiency level, suggesting that implicit and explicit knowledge in instructed learners is best viewed as a continuum. For lower proficiency groups, several strong relationships were found between cognitive capacity, L2 motivation, and performance on measures of implicit and explicit knowledge; however, learner IDs did not explain performance variance for the advanced group. Relationships among ID constructs also highlight their likely interaction and dynamic influence on L2 development over time. Findings are discussed in terms of their implications for second language acquisition (SLA) theory, assessment, and pedagogy. [The dissertation citations contained here are published with the permission of ProQuest LLC. Further reproduction is prohibited without permission. Copies of dissertations may be obtained by Telephone (800) 1-800-521-0600. Web page: http://www.proquest.com/en-US/products/dissertations/individuals.shtml.]

The rise of quantum technology has put control at the centre of advancements in quantum mechanics. The union of quantum mechanics with mathematical control theory is a meeting that is leading to a much deeper insight into our interaction with the bizarre properties of quantum theory. Often, the study of discrete variable systems is the focus for making this union. Here, we look at how control theory may be applied to the continuous variable theory of Gaussian states. Special emphasis is given to control of the covariance matrix of these states, as it is here that we find the entanglement and entropic properties of the state. We begin by exploring some initial results for the geometry of Gaussian states, revealing different manifold structures dependent on symplectic eigenvalue degeneracy. In this geometrical setting a proposal for an extension of Williamson's theorem is put forward and partially completed. It is often interesting to look at restricted sets of Hamiltonians and ask what transformations can be performed with concatenations of their corresponding unitaries. Controllable systems are those for which the entire group of interest is possible to enact. We explore an uncontrollable system in a single mode and give a physical analysis as to why it behaves this way. This leads to ideas to move forwards for a necessary and sufficient condition for control on the symplectic group that has been conjectured since 1972. Later, we transfer to the question of open dynamics. We focus on a particular and ubiquitous channel known as 'lossy' or 'the attenuation channel'. An equation is derived describing the evolution for the symplectic invariants of a Gaussian state undergoing such dynamics. The equation of the former chapter is used to explore the evolution of entropy and entanglement. Optimal protocols are developed for the manipulation of these properties undergoing lossy dynamics.

A structural glycoprotein has been extracted from bovine ligamentum nuchae, using 5M guanidine hydrochloride containing a disulphide bond reducing agent, and purified by preparative gel electrophoresis. The isolated material appeared to be monodisperse with a molecular weight of ~ 34 OOO as shown by sodium dodecyl sulphate-polyacrylamide gel electrophoresis and by analytical ultracentrifugation. It contains 10% carbohydrate comprising mannose, N-acetyl-glucosamine, galactose and sialic acid in a 6:5:3:3 molar ratio. The glycoprotein has been assayed for peptidyl-lysine oxidase activity using 3H lysine-aortic elastin, prepared from 15- to 17-day-old chick embryos, as a substrate. In the absence of free lysine, the specific activity of the preparation over a 2h incubation was ~ 60 x 104 dpm/mg purified protein. Addition of 10mM lysine resulted in an ~ 50% decrease in the specific activity. Free lysine was shown to act as a substrate for the glycoprotein preparation as indicated by control experiments using 3H lysine in place of the aortic substrate. These results demonstrate that the glycoprotein exhibits a dual amine oxidase activity. In the presence of 0.27mM aminopropionitrile fumarate, a concentration which completely inhibits peptidyl-lysine oxidase activity in other lysyl oxidases, the glycoprotein preparation was inhibited by ~ 14%. In the absence of 5M guanidine hydrochloride and a reducing agent, the glycoprotein undergoes aggregation which in the presence of copper ions results in the formation of cylindrical tactoids, the diameter of which (11nm) corresponds closely to that of the fibrils which in the majority of connective tissue matrices constitute the microfibrillar component mainly associated with elastic fibres.

During embryogenesis, the inherited information on the spatial arrangement of a particular tissue results in the deposition of extracellular matrix components. The overall tissue architecture is the responsibility of the embryonic cells which use the physico-chemical properties of the synthesised matrix components to assure local stability and form. Trelstad and Birk (1984) have postulated that embryonic cells are inherently polarised in both structure and function and that the quality, quantity and orientation of the extracellular matrix is under cellular control. They have shown that the cellular polarity and matrix patterning in embryonic chick tendon and corneal matrix is very similar to that proposed in this thesis for the morphogenesis of the elastic fibre. During matrix secretion, epithelial cells in the cornea are observed to undergo a reversal of their polarity with up to 80% of the cells relocating their Golgi to the basal cell pole. Discrete packets of collagen are then transported to the cell surface where the vacuoles are brought into register with the orientation of the underlying matrix by the alignment of ordered filaments and/or microtubules in the basal cell cytoplasm. The fact that elastin appears in the extracellular matrix at a later stage in development than most other connective tissue macromolecules suggests that elastogenic cells require a highly structured or 'mature' extracellular matrix upon which to organise the complex elastic fibre and that only late in gestation is the matrix suitable for the fibre organisation. Accordingly, a component of the extracellular matrix such as the microfibrils may provide the information that signals the appropriate time for elastin synthesis and secretion. This view is supported by the cellular polarity and the presence of both secretory vesicles and aligned microtubules (Chapter Two) observed when elastin-producing cells are in close proximity to the microfibrillar component of developing elastic fibres. The belief that elastin-associated microfibrils play a crucial role in the morphogenesis of the elastic fibre has until recently been based primarily on electron microscopic studies of the development of elastic tissue. In Chapter Three, purified 35k-GP was shown by immunoblotting and immunochemical techniques to be a constituent of elastin-associated microfibrils in developing elastic tissues within the ear, skin, aorta and ligamentum nuchae of the foetal calf. No binding of anti-35k-GP was detected in adult bovine ligamentum nuchae, aorta or ear which suggests that this glycoprotein plays an important but transient role in the development of elastic tissue. It is proposed that the role of 35k-GP in the function of microfibrils is to induce the adhesion and morphogenetic movements of elastogenic cells during elastic fibre formation. The preliminary results described in Chapter Four support this view and we now plan to undertake a more detailed study with the following objectives in mind:- 1.(a)To examine by immunofluoresence the involvement of cytoskeletal structures such as microtubules and microfilaraents in the interaction of elastogenic cells with 35k-GP. (b)To determine the type and duration of the 35k-GP induced adhesion by time-lapse video interference reflection microscopy to discover whether a substrate specific focal contact is occuring resulting in cytoskeletal reorganisation and ultimately in the synthesis and extracellular deposition of elastin. 2. A study of the ability of 35k-GP fragments to induce cell adhesion and spreading. This will allow the elucidation of a possible specific cell-binding region of the molecule. 3. An investigation into the identity of a specific cell receptor, the presence of which appears essential to the primary recognition event in the binding of 35k-GP and the possible existence of additional cell surface molecules to the cell-substratum attachment mechanism. 4. In conjuction with the above, a gene library from human foetal aortic smooth muscle cells is being created with the intention of developing cDNA probes to 35k-GP specific mRNA from this tissue.

Section A of this thesis describes the non-degradative isolation procedure of elastin from the media of human thoracic aorta and the determination of its ponderal distribution as well as that of the other connective tissue matrix components in the aortic media of individuals of different ages. A cylindrical segment, free of complex atherosclerotic lesions, was resected at autopsy from each of fifty-nine descending human thoracic aortae by cutting just below the level of the first pair of intercostal arteries and 35 mm distal to this incision. In the isolated tunica media there was an age-related rise in the absolute concentration of the following components: (a) proteins and glycoproteins extractable in chaotropic agents, (b) collagen and (c) an insoluble polar protein(s) which was resistant to collagenase digestion but was solubilized by trypsin. In contrast, the absolute elastin concentration was constant at around 70 mg/cm in samples of all ages. Although after treatment with collagenase and trypsin, both the mechanical properties and amino acid composition of elastin did not vary with age, in samples tested prior to trypsin digestion the insoluble polar-protein contaminant affected the mechanical behaviour of the elastin network in a way akin to plasticisers in rubber. It is therefore suggested that the morphological changes .and stiffening observed in the ageing aortic wall are not due to degradation of its elastin network bu to variations in the supramolecular organisation of connective tissue components. Section B of this thesis describes the isolation of elastin and the determination of its distribution as well as that of collagen in the bulbus arteriosus and ventral aorta of trout (Salmo gairdneri) and salmon (Salmo salar). The isolation of elastin fibrils from these specimens did not require the use of proteolytic enzymzes as collagen and all the other components of the vessel wall were readily extracted in denaturing and reducing conditions. The arterial walls of trout and salmon are characterized by a high elastin content, with an elastin:collagen ratio exceeding 2.5:1. The elastin which display an amino acid composition highly at variance with that of the mammalian protein, is present as 250-A-diameter fibrils which have a strong affinity for electron-dense cations and which form in the intercellular matrix an isotropic reticulum. As in mammalian elastin, the fibrils can be resolved in negatively contrasted preparations into constituent primary filaments of about 25 A in diameter, aligned with an equatorial periodicity of about 55 A. The wide-angle X-ray diffraction pattern of the salmonid preparations showed broad reflections corresponding to spacings of 9.8, 4.5 and 2.2 A, similar to bovine elastin. The mechanical behaviour of the salmon preparation was characterized by a linear response to stress, with minimal hysteresis, a Young's modulus of 5.5 x 105 N m-2, and a breaking strain of 1.5.

Section A; Insoluble elastin was isolated from bovine aorta and ligaraentum nuchae by the use of guanidine and dithiothreitol in conjunction with collagenase, purified by affinity chromatography. The preparations were free from carbohydrate and exhibited amino acid compositions similar to that of alkali-purified elastin, with the exception of the concentration of some polar amino acid residues. N-terminal analyses indicated a very low level of polypeptide chain damage in the preparations. Upon mild alkaline treatment glycine was selectively liberated as amino-end group. Elastin from ligamentum nuchae was examined by electron microscopy, X-ray diffraction, optical polarisation analysis and dynamometry. It is suggested that elastin fibrils consist of a lateral array of primary filaments. Section B: The microfibrillar component of adult bovine ligamentum nuchae was isolated by solubilisation in guanidine-dithiothreitol followed by S-carboxymethylation and treatment with collagenase. The amino acid composition of the material was at variance to those previously reported for similar preparations. N-terminal analysis revealed glycine as the only end-group, at a concentration of 65.9 moles/106g of protein. This value, corresponding to a molecular weight of about 15,000 daltons, was in good agreement with the results of sedimentation equilibrium analyses.