L’exploitation de données radiographiques, en particulier pour la reconstruction 3D du rachis de patients scoliotiques, est un prérequis à la modélisation personnalisée. Les méthodes actuelles, bien qu’assez robustes pour la routine clinique, reposent sur des ajustements manuels fastidieux. Dans ce contexte, ce travail de thèse vise à la détection automatisée de points anatomiques spécifiques des vertèbres, permettant ainsi des ajustements automatisés. Nous avons développé premièrement une méthode originale de localisation de coins de vertèbres cervicales et lombaires sur les radiographies sagittales. L’évaluation rigoureuse de cette méthode suggère sa robustesse et sa précision. Nous avons ensuite développé un algorithme pour le problème pertinent cliniquement de localisation des pédicules sur les radiographies coronales. Cet algorithme se compare favorablement aux méthodes similaires dans la littérature, qui nécessitent une saisie manuelle. Enfin, nous avons soulevé les problèmes, relativement peu étudiés, de détection, identification et segmentation des apophyses épineuses du rachis cervical dans les radiographies sagittales. Toutes les tâches mentionnées ont été réalisées grâce à une combinaison originale de descripteurs visuels et une classification multi-classe par Random Forest, menant à une nouvelle et puissante approche de localisation et de segmentation. Les méthodes proposées dans cette thèse suggèrent un grand potentiel pour être intégré à la reconstruction 3D du rachis, utilisée quotidiennement en routine clinique. / Exploitation of spine radiographs, in particular for 3D spine shape reconstruction of scoliotic patients, is a prerequisite for personalized modelling. Current methods, even though robust enough to be used in clinical routine, still rely on tedious manual adjustments. In this context, this PhD thesis aims toward automated detection of specific vertebrae landmarks in spine radiographs, enabling automated adjustments. In the first part, we developed an original Random Forest based framework for vertebrae corner localization that was applied on sagittal radiographs of both cervical and lumbar spine regions. A rigorous evaluation of the method confirms robustness and high accuracy of the proposed method. In the second part, we developed an algorithm for the clinically-important task of pedicle localization in the thoracolumbar region on frontal radiographs. The proposed algorithm compares favourably to similar methods from the literature while relying on less manual supervision. The last part of this PhD tackled the scarcely-studied task of joint detection, identification and segmentation of spinous processes of cervical vertebrae in sagittal radiographs, with again high precision performance. All three algorithmic solutions were designed around a generic framework exploiting dedicated visual feature descriptors and multi-class Random Forest classifiers, proposing a novel solution with computational and manual supervision burdens aiming for translation into clinical use. Overall, the presented frameworks suggest a great potential of being integrated in current spine 3D reconstruction frameworks that are used in daily clinical routine.

South Africa is more than ever, in need of a public transport system integrated within its surroundings. A successful transportation network needs an environment where users can change transport systems with ease and peace of mind. The ultimate goal is to achieve a pedestrian orientated city. This dissertation will attempt to address these problems by redesigning Loftus Metrorail Station located in Hatfield, Pretoria, South Africa. Within the transport node, Loftus station, the challenge is to combine the different movement systems intersecting at this point, reinforcing connectivity and accessibility. The transition from public to private needs to be clear and legible - this can be achieved through the appropriate design of the building interface, and clearly defining spatial boundaries. The sense of place is lacking in this area, already containing various landmarks associated with Hatfield. These landmarks assist with the legibility of the project area. This orientation device assists with ease of movement and circulation within the area. The defragmentation of facilities in the Loftus precinct area is due to the large scale of the city block. Loftus station becomes a barrier between the University of Pretoria and Loftus Versfeld Stadium. It is an isolated site – dislocated from its surroundings. The lack of facilities has lead to the station becoming underutilized and dilapidated. By dividing the study area into human scaled city blocks, the area becomes manageable. This division reinforces already existing pedestrian movement patterns. The activation of edges and transparency of the structure ensures a secure environment. The different movement systems at this node influenced the design and choice of materials used in the project. The user requirements had to be satisfied. The careful integration of commerce within the station and surroundings ensure not only a transition space, but also a controlled and activated destination place. / Dissertation (MArch(Prof))--University of Pretoria, 2009. / Architecture / unrestricted

Zielstellung: Die Verwendung von Computertomographie und Magnetresonanzto-mographie zur Diagnostik von pathologischen Weichteilveränderungen, wie Läsionen der Speicheldrüsen, wird empfohlen. Diese begründet sich in Ihrer hohen Bildauflö-sung und dem guten Bildkontrast. Nach unserem Wissen gibt es aktuell keine publi-zierte Literatur, welche sich mit der CT- oder MRT-Anatomie der Speicheldrüsen der Katze beschäftigt. Ziel der vorliegenden Arbeit war es daher, Schnittbildmerkmale der Speicheldrüsen zu definieren und Landmarken für eine sichere Identifikation zu be-schreiben. Material und Methoden: Es wurden zwei Gruppen von Katzen untersucht. Die pros-pektiv untersuchten Tiere bestanden aus 16 Katzen, welche aus Gründen euthana-siert wurden, die nicht in Zusammenhang mit der Studie standen. Diese erfuhren ein CT (1 mm Schichtdicke) und 0,5 Tesla MRT (T2W, T2W SPIR, PDW, T1W) des Kop-fes. Im retrospektiven Anteil der Studie wurden CT-Untersuchungen (0,8-1 mm Schichtdicke) von 25 Katzen und MRT-Untersuchungen (T2W, T1W vor und nach Kontrastmittelgabe) von 21 Katzen verwendet. Diese Untersuchungen wurden zwi-schen 2005 und 2013 in unserer Klinik angefertigt. Dabei konnten keine Pathologien in Zusammenhang mit den Speicheldrüsen festgestellt werden. Alle Speicheldrüsen wurden bezüglich Identifizierbarkeit, Abgrenzbarkeit, Größe und Signalverhalten un-tersucht. Wenn vorhanden, wurden anatomische Landmarken die Speicheldrüsen umgebend notiert. Zusammenfassung 80 Ergebnisse: 94,3% der Glandulae parotidae, 90,7% der Glandulae mandibulares und 96,8% der Glandulae zygomaticae konnten mit dem CT identifiziert werden. Beim MRT waren es 94,1% der Glandulae parotidae, 99,2 % der Glandulae mandibulares, 88,9% der Glandulae zygomaticae und 51,4 % der Glandulae buccalis ventralis. In beiden Gruppen konnten keine der restlichen kleinen Speicheldrüsen identifiziert werden, dies galt für beide Schnittbildverfahren. Als hilfreiche anatomi-sche Landmarken konnten der äußere Gehörgang, der M. masseter, der M. pterygo-ideus medialis und der Bulbus oculi herangezogen werden. Die medianen Dichtewer-te lagen für die Glandula parotis, die Glandula mandibularis und die Glandula zygo-matica bei 65 HU, 62 HU und 57 HU. Die Parotis war überwiegend dezent hyperin-tens zur Muskulatur und hypointens zum Fett. Die Mandibularis, Zygomatica, und Glandula buccalis ventralis stellten sich hauptsächlich hyperintens zur Muskulatur und hypointens zu Fettgewebe dar. Dieses Signalverhalten war für T1W, T2W und T2W SPIR (fettunterdrückt) gewichtete Sequenzen gleich. Die Größe war für beide Gruppen und Modalitäten vergleichbar und lag bei circa 17 x 6 mm (gemessen von lateral nach medial und von dorsal nach ventral) bei der Glandula parotis, 8 x 17 mm bei der Glandula mandibularis, 10 x 7 mm bei der Glandula zygomatica und bei der Glandula buccalis ventralis 3 x 10 mm. Schlussfolgerungen: Landmarken und Aussagen über die Signalintensität bezie-hungsweise Dichte und die Größe der Glandula parotis, Glandula mandibularis, Glandula zygomatica und der Glandula buccalis ventralis konnten erhoben werden. Die mangelhafte Identifizierbarkeit der restlichen kleinen Speicheldrüsen, im Ver-gleich zu den oben genannten, kann vor allem mit dem unzureichenden Weichteil-kontrast zum umliegenden Gewebe im CT und der niedrigen Auflösung des Nieder-feld-MRTs erklärt werden. Neuere MRT-Techniken können detailliertere Bilder er-zeugen und so eine Beschreibung der kleinen Speicheldrüsen ermöglichen. / Objective: Due to their high resolution and useful contrast resolution, CT and MRI are recommended methods for diagnosing pathological changes of soft tissues such as lesions of the salivary glands. To our knowledge, no studies about the CT or MR anatomy of the feline salivary glands have been published yet. The aim of the study was to define anatomical characteristics of the salivary glands in cross-sectional im-ages and to describe landmarks for a reliable identification. Materials and methods: Two groups of cats where examined. In a prospective trial, 16 cats, euthanized for reasons unrelated to the study, underwent a head CT (1 mm slice thickness) and 0.5 Tesla MRI scans (T2W, T2W SPIR, PDW, T1W). In a retro-spective trial, the CT scans (0.8-1 mm slice thickness) of 25 cats and the MR scans (T2W, T1W, pre- and post-contrast) of 21 cats were used. These scans were per-formed at our clinic between 2005 and 2013. No signs of salivary-gland-related pa-thologies were visible. All images were reviewed with respect to the following criteria: the presence, delineation and size of the salivary glands as well as the signal intensi-ty. Anatomical landmarks surrounding the glands were also recorded. Results: 94.3% of the parotid glands, 90.7% of the mandibular glands and 96.8% of the zygomatic glands could be delineated using CT. With MRI, 94.1% of the parotid Summary 82 glands, 99.2 % of the mandibular glands, 88.9% of the zygomatic glands and 51,4 % of the ventral buccal glands could be identified. In both groups the other small salivary glands could not be visualized neither in CT nor in MRI. Anatomical landmarks such as the external ear canal, the M. masseter, the M. pterygoideus medialis and the Bulbus oculi facilitate the identification. The mean density of the parotid gland, the mandibular gland and the zygomatic gland was 65 HU, 62 HU and 57 HU, respectively. The parotid gland was mainly modestly hyperintense to muscle and hypointense to fat. The mandibular, zygomatic and ven-tral buccal glands were hyperintense to muscle and hypointense to fat. Signal intensi-ties were similar for T1W, T2W and T2W SPIR images. Sizes were comparable for both groups and modalities and measured about 17x6 mm (lateral to medial and dor-sal to ventral) for the parotid gland, 8 x 17 mm for the mandibular gland, 10x7 mm for the zygomatic gland and 3x10 mm for the ventral buccal gland. Conclusions: Landmarks were recorded and data regarding the density/signal in-tensity and the size of the parotid gland, the mandibular gland, the zygomatic gland and the ventral buccal gland could be gathered. The discrepancy between depicting the small glands and the above mentioned glands can be explained mainly by the lack of contrast to the surrounding tissue in CT and the low spatial resolution in low-field MRI. Advanced MRI techniques are expected to produce more detailed images and therefore those techniques could be used to also describe the small salivary glands.

Cancer is one of the leading causes of death in the world, but advances in research and development of treatment methods is constantly ongoing to reduce the number of deaths and the amount of suffering. One of many approaches is radiation therapy, which uses high doses of radiation to kill tumors. Radiation therapy requires advanced software in image analysis to create careful treatment plans, evaluate treatment responses and to perform dose accumulation, among other things. One important tool for this is deformable image registration (DIR) which is used to find a correspondence between the images. The aim with this master thesis is to improve the DIR method ANACONDA by automatically provide additional information to the algorithm before the registration is performed.This work focuses on the registration of internal tubular structures in lung and liver images (bronchial and vascular tree, respectively). Two challenges in registering lung images are the sliding motion of lung surfaces and large motion of small internal structures. Several DIR methods have been proposed for solving the challenging internal structures, however most of them do not take into account the alignment of surrounding tissues. DIR methods applied to the liver are published less frequently, but accurate registration of the liver is of high interest since, for example, knowledge of the anatomy of the vascular tree is essential when removing tumors through liver surgeries. In this work, corresponding (anatomical) points are automatically found in two images and added to the DIR algorithm. The points are found by extracting and comparing the tubular structures between the images, and with use of different distance requirements, nearby points are paired.The new method manages to achieve good registration of both internal structures and surrounding tissue. Mean target registration errors for the internal structures of lungs was 1.11 ± 0.75 and for liver 1.67 ± 1.15 mm.

<p>We investigated 1) how objects come to serve as landmarks in spatial memory and more specifically how they form part of an allocentric cognitive map and 2) how humans encode multiple connected spatial environments. In both sets of experiments, participants performing a virtual driving task incidentally learned the layout of a town and locations of objects or stores in that town. Their spatial memory and recognition memory for the objects or stores were subsequently tested. To assess whether the objects were encoded allocentrically, we developed a new measurement, pointing consistency. We found that when participants had more limited experience of the environment spatial memory for objects at navigationally relevant locations was more consistent across tested viewpoints than for objects at navigationally less relevant locations. When participants’ attention was focused on the appearance of objects, the navigational relevance effect was eliminated, whereas when their attention was focused on the objects’ locations, this effect was enhanced, supporting the hypothesis that when objects are processed in the service of navigation, rather than merely being viewed as objects, they engage qualitatively distinct attentional systems and are incorporated into an allocentric spatial representation. The results were consistent with evidence from the neuroimaging literature that when objects are relevant to navigation, they not only engage the ventral “object processing stream”, but also the dorsal stream and medial temporal lobe memory system classically associated with allocentric spatial memory. Moreover, in the connected environments, our data were more consistent with the formation of local maps, regardless of whether the neighborhoods were learned together or separately. Only when all visible distinctions between neighborhoods were removed did people behave as if they formed one integrated map. These data are broadly consistent with evidence from rodent hippocampal place cell recordings in connected boxes, and with hierarchical models of spatial coding.</p> / Doctor of Philosophy (PhD)

Cette thèse étudie l'intégration de repères phonétiques dans la reconnaissance automatique de la parole (RAP) continue à grand vocabulaire. Les repères sont des événements à temps discret indiquant la présence d’événements phonétiques dans le signal de parole. Le but est de développer des détecteurs de repères qui sont motivés par la connaissance phonétique afin de modéliser quelques événements phonétiques plus précisément. La thèse présente deux approches de détection de repères, qui utilisent l'information extraite par segments et étudie deux méthodes différentes pour intégrer les repères dans le décodage, qui sont un élagage basé sur les repères et une approche reposant sur les combinaisons pondérées. Alors que les deux approches de détection de repères présentées améliorent les performance de reconnaissance de la parole comparées à l'approche de référence, elles ne surpassent pas les prédictions phonétiques standards par trame. Ces résultats indiquant que la RAP guidée par des repères nécessite de l'information phonétique très hétérogène pour être efficace, la thèse présente une troisième méthode d'intégration conçue pour intégrer un nombre arbitraire de flux de repères hétérogènes et asynchrones dans la RAP. Les résultats indiquent que cette méthode est en effet en mesure d'améliorer le système de référence, pourvu que les repères fournissent de l'information complémentaire aux modèles acoustiques standards. / This thesis studies the integration of phonetic landmarks into standard statistical large vocabulary continuous speech recognition (LVCSR). Landmarks are discrete time instances that indicate the presence of phonetic events in the speech signal. The goal is to develop landmark detectors that are motivated by phonetic knowledge in order to model selected phonetic classes more precisely than it is possible with standard acoustic models. The thesis presents two landmark detection approaches, which make use of segment-based information and studies two different methods to integrate landmarks into the decoding, which are landmark-based pruning and a weighted combination approach. While both approaches improve speech recognition performance compared to the baseline using weighted combination of landmarks and acoustic scores during decoding, they do not outperform standard frame-based phonetic predictions. Since these results indicate that landmark-driven LVCSR requires the integration of very heterogeneous information, the thesis presents a third integration framework that is designed to integrate an arbitrary number of heterogeneous and asynchronous landmark streams into LVCSR. The results indicate that this framework is indeed ale to improve the baseline system, as soon as landmarks provide complementary information to the regular acoustic models.

O propósito deste trabalho foi mensurar as dimensões dos arcos dentários de crianças com fissuras labiopalatinas, de 3 a 9 meses, antes das cirurgias primárias, e comparar com crianças sem fissura labiopalatina. A amostra foi composta de 223 crianças, divididas em grupos: sem deformidades craniofaciais (G1), fissura pré-forame incompleta (G2), fissura pré-forame completa (G3), fissura transforame unilateral (G4) e fissura transforame bilateral (5). Para avaliação, os arcos dentários superiores das crianças foram moldados com silicona de condensação. Os modelos passaram por um processo de digitalização, por meio de escâner 3D e as medidas utilizadas para a correlação entre os grupos foram realizadas diretamente nas imagens escaneadas. Nas imagens, foram determinados pontos de referência, necessários para a realização das medidas de avaliação. Os resultados mostraram diferença estatisticamente significativa na distância intercaninos (DIC) entre G1-G4 (p= 0,000), G1-G5 (p= 0,007), G2-G4 (p= 0,000), G2-G5 (p= 0,000), e G3-G4 (p= 0,008). Para distância intertuberosidades (DIT) foram encontradas diferenças estatisticamente significativas entre: G1-G4 (p= 0,000), G1-G5 (p= 0,000), G2-G4 (p= 0,000), G2-G5 (p= 0,000), G3-G4 (p= 0,009) e G3-G5 (p= 0,000). Para amplitude anterior da fissura unilateral (AAFuni), o G3 apresentou diferença estatisticamente significativa com relação ao G4 (p= 0,000) e para amplitude anterior da fissura esquerda (AAFe) e direita (AAFd) com relação ao G5 (p= 0,011 e p= 0,030). Para amplitude posterior da fissura (APF), houve diferença estatisticamente significativa entre G4 e G5 (p= 0,013). Com base nos resultados obtidos, foi possível constatar que as crianças com fissura labiopalatina apresentaram dimensões transversais dos arcos dentários maiores que as crianças sem fissura labiopalatina, de 3 a 9 meses, antes das cirurgias primárias. A DIC foi maior nas crianças com fissura transforame unilateral e a DIT foi maior nas crianças com fissura transforame bilateral em relação aos grupos estudados. Houve aumento significante da AAF e APF nos diferentes tipos de fissura com relação à severidade da fissura de lábio e/ou palato. / The purpose of this study was to measure the dimensions of the dental arches of 3-9 month children with cleft lip and palate, before the primary surgery, and compared them with children without cleft lip and palate. The sample consisted of 223 children, divided into groups: without craniofacial deformities (G1), unilateral cleft lip (G2), unilateral cleft lip and alveolus (G3), unilateral cleft lip and palate (G4) and bilateral cleft lip and palate (5). For evaluation, impressions of the upper dental arches of the children were performed with condensation silicone. The casts underwent a process of scanning through 3D scanner and the measurements used for the correlation among groups were executed directly on the scanned images. On the images, the landmarks required to perform the evaluation measurements were determined. The results showed statistically significant differences in intercanine distance (ICD) between G1 - G4 (p = 0.000), G1 - G5 (p = 0.007), G2 - G4 (p = 0.000), G2 - G5 (p = 0.000), and G3 - G4 (p = 0.008). For the posterior arch width (PAW), the following statistically significant differences were found: between G1 - G4 (p = 0.000), G1 - G5 (p = 0.000), G2 - G4 (p = 0.000), G2 - G5 (p = 0.000), G3 - G4 (p = 0.009), and G3 - G5 (p = 0.000). The unilateral anterior cleft width (UACW) exhibited statistically significant differences between G3 and G4 (p = 0.000); the left (LACW) and right anterior cleft width (RACW) showed differences between G3 and G5 (p = 0.011 and p = 0.030). For posterior cleft width (PCW), statistically significant differences between G4 and G5 (p = 0.013) were found. Based on these results, it was found that 3-9 month cleft lip and palate children had transverse dimensions of the dental arches higher than those without cleft lip and palate, before the primary surgeries. The ICD was higher in children with unilateral cleft lip and palate and PAW was higher in children with bilateral cleft lip and palate among the groups studied. A significant increase in the ACW and PCW of the different cleft types with respect to the severity of cleft lip and / or palate was found.

<p>In this thesis, the shape of the distal humerus trochlea is analysed using landmark-based morphometrics and multivariate methods, with the aim of exploring locomotor evolution in carnivorans. Elbow joint morphology is used together with body size and craniodental morphology to characterize past and present carnivorans. Evolutionary implications are studied at the ordinal, familial, and species levels, testing specific hypotheses about scaling, morphological constraints, evolutionary trajectories, and potential for social pack-hunting behaviour. The circumference of the distal humerus trochlea is found to be highly correlated with body mass, and appears to scale similarly throughout the order Carnivora. A general predictive model for carnivoran bodymass is presented (a=0.601; b= 2.552; r2=0.952, SEE=0.136, p<0001, n=92), which removes the need for the investigator to actively choose between the diverging estimates that different predictors and their equations often produce. At the elbow joint, manual manipulation and locomotion appear to be conflicting functions, thus suggesting mutually exclusive lifestyles involving either forelimb grappling or pursuit. At large body sizes, carnivorans are distributed over a strongly dichotomised pattern (grappling or locomotion), a pattern coinciding with the postulated threshold in predator-prey size ratio at 21.5-25 kg. This pattern is compared to that of two carnivoran faunas from the Tertiary. In the Oligocene (33.7-23.8 Myr BP), the overall pattern is remarkably similar to that observed for extant Carnivora. In the Miocene (23.8-11.2 Myr BP) carnivores show a similarly dichotomised pattern as the Oligocene and Recent, although the whole pattern is shifted towards larger body sizes. This difference is suggested to be a reflection of the extraordinary species richness of browsing ungulates in the early Miocene of North America. Such an increase in prey spectrum would create a unique situation, in which large carnivores need not commit to a cursorial habitus in order to fill their nutritional requirements. Finally, the elbow joints and craniodental morphology (14 measurements) of fossil canids were examined with the aim of assessing the potential for pack-hunting in fossil canids. It is clear that small and large members of the Recent Caninae share similar craniodental morphologies. However, this pattern is not present in Borophaginae and Hesperocyoninae. In the latter, large representatives are characterized by being short-faced, with reduced anterior premolars and enlarged posterior premolars, thus approaching a “pantherine-like” craniodental configuration. These traits are interpreted as an adaptation for killing prey with canine bites. It is similarly determined that, unlike recent Caninae, all analyzed species of borophagines and hesperocyonines have retained the ability to supinate their forearms. It is therefore likely that manual manipulation was part of their hunting behaviour, thus removing an essential part of the argument for social pack-hunting in these forms, as the benefits of such a strategy become less obvious.</p>

Taking the establishment of Ice Age theory as its point of departure, the present dissertation examines aspects of geological, plant geographical and archaeological research on shoreline displacement conducted in Sweden during the period 1860-1930, and the significance of this research for the perception of "the Swedish landscape" and its post-glacial history. The research is analyzed on three levels under the rubrics "The Highest Shoreline and the Ancylus Lake", "The Question of Land Elevation", and "Charting Swedens Past", respectively. Taken together, these levels capture the varying perceptions and exchanges of opinion of the nature of shoreline displacement and the contexts in which they were applied. The present study is conducted via a theoretical and methodological approach where both the ideas and the practices of science are studied: activities in the field and at the various institutions (primarily the Swedish Geological Survey, its museum and the Geological Society of Stockholm); arguments and hypotheses presented in artides and handbooks, including visual images, diagrams and maps; social networks, career paths and controversies. Particular attention is paid to the relationship between science and nationalism, and comparisons are made with research conducted in an imperialist context. Ice Age theory helped initiate research whose purpose was to discover how the Swedish naturai and cultural landscape carne into being. The foremost task of geology became studying geographic evolution during the Quaternary epoch, how the land had "risen from ice and water". Plant geography studied how and whence plant life had migrated and how vegetation had evolved under the influence of biological, geological and climatological factors. In a similar manner, archaeologists studied the migration and dissemination of mankind during the StoneAge. When natural scientists primarily used "natural landmarks" as its source material, archaeologists relied on "archaeological finds", which were invested with scientific value but also became symbols of national collective memory. Through this survey, national identity was unifìed with the territory itself and its evolutionary history. Knowledge about shoreline displacement became significant for geology, plant geography and archaeology, which in turn encouraged interdisciplinary collaboration, but also locked the researchers into a similar way of thinking about the nature of shoreline displacement. According to this "thought style", the phenomenon was first and foremost the result of the vertical movement of the land rather than movement in the ocean surface. Up until the 1870s, the Ice Age was thought to have been followed by one single subsidence and elevation; during the 1880s, two such land oscillations; three during the 1910s and by the 1920s, five. Only toward the end of the 1920s did Swedish researchers begin to accept a multi-factor explanation, which succeeded in finally subverting the reigning thought style. According to this explanation, shoreline displacement was not solely the result of changes in the land or the sea, but of both. / digitalisering@umu

In this thesis, the shape of the distal humerus trochlea is analysed using landmark-based morphometrics and multivariate methods, with the aim of exploring locomotor evolution in carnivorans. Elbow joint morphology is used together with body size and craniodental morphology to characterize past and present carnivorans. Evolutionary implications are studied at the ordinal, familial, and species levels, testing specific hypotheses about scaling, morphological constraints, evolutionary trajectories, and potential for social pack-hunting behaviour. The circumference of the distal humerus trochlea is found to be highly correlated with body mass, and appears to scale similarly throughout the order Carnivora. A general predictive model for carnivoran bodymass is presented (a=0.601; b= 2.552; r2=0.952, SEE=0.136, p&lt;0001, n=92), which removes the need for the investigator to actively choose between the diverging estimates that different predictors and their equations often produce. At the elbow joint, manual manipulation and locomotion appear to be conflicting functions, thus suggesting mutually exclusive lifestyles involving either forelimb grappling or pursuit. At large body sizes, carnivorans are distributed over a strongly dichotomised pattern (grappling or locomotion), a pattern coinciding with the postulated threshold in predator-prey size ratio at 21.5-25 kg. This pattern is compared to that of two carnivoran faunas from the Tertiary. In the Oligocene (33.7-23.8 Myr BP), the overall pattern is remarkably similar to that observed for extant Carnivora. In the Miocene (23.8-11.2 Myr BP) carnivores show a similarly dichotomised pattern as the Oligocene and Recent, although the whole pattern is shifted towards larger body sizes. This difference is suggested to be a reflection of the extraordinary species richness of browsing ungulates in the early Miocene of North America. Such an increase in prey spectrum would create a unique situation, in which large carnivores need not commit to a cursorial habitus in order to fill their nutritional requirements. Finally, the elbow joints and craniodental morphology (14 measurements) of fossil canids were examined with the aim of assessing the potential for pack-hunting in fossil canids. It is clear that small and large members of the Recent Caninae share similar craniodental morphologies. However, this pattern is not present in Borophaginae and Hesperocyoninae. In the latter, large representatives are characterized by being short-faced, with reduced anterior premolars and enlarged posterior premolars, thus approaching a “pantherine-like” craniodental configuration. These traits are interpreted as an adaptation for killing prey with canine bites. It is similarly determined that, unlike recent Caninae, all analyzed species of borophagines and hesperocyonines have retained the ability to supinate their forearms. It is therefore likely that manual manipulation was part of their hunting behaviour, thus removing an essential part of the argument for social pack-hunting in these forms, as the benefits of such a strategy become less obvious.