Introduction Rugby Union (hereinafter referred to as ‘rugby’) is a contact sport with players being exposed to repetitive collisions throughout a match. As the risk of injury is relatively high, incidence surveillance studies within rugby has become popular. However most of the studies have focussed on senior players. The data on injuries among youth rugby players are limited. This makes it difficult to develop the game to make it safer for youth of all ages. Objectives The first objective of this study was to establish if any injury trends exist across different ages of youth rugby players (13 to 18 years). The second objective was to determine the patterns of injuries changed over four years (2011 to 2014). Methods The South African Rugby Union (SA Rugby) hosts four local youth tournaments annually to for local rugby talent: Craven Week under-13, Grant Khomo under-16, Academy Week under-18 and Craven Week under-18. Injury data were collected from the four SARU Youth Week Tournaments between 2011 and 2014. These data were compiled into one central SARU injury surveillance database. Injury categories were used to group data: ‘Type’, ‘Location’, ‘Event’ and ‘Severity’ of injury were assessed. Injuries were defined as either ‘Time-loss’ (those injuries that prevented a player from match participation for one or more days), or ‘Medical attention’ (injuries that required the player to seek medical attention at the time of or after injury but were not required to miss a match). Injury rates were represented by injury incidence densities (IIDs) (corresponding 95% confidence intervals (95% CIs) for IID were calculated for the number of injuries regardless of whether one person was injured more than once) per 1000 hours of match play. Incidence densities were considered to be significantly different from each other if their 95% CIs did not overlap and using Poisson regression analysis. Results The ‘overall’ combined IID across all four years was 54.6 injuries per 1000 hours of match play (95%CI: 51.0-58.2). The combined ‘time-loss’ IID was 18.9 injuries per 1000 hours of match play (95%CI: 16.8-21.0). ‘Time-loss’ injuries were greatest in 2011 (23.2 per 1000 match hours (95% CI: 18.5-28.0)). However, ‘time-loss’ injuries rates were significantly reduced in 2013, when compared to these injury rates in 2011 (13.3 (9.7-17.0). Craven Week under-13 presented significantly greater ‘overall’ injury incidence densities when compared to the older age groups (71.9 per 1000 match hours (95% CI: 62.4-81.4)). Overall, joint/ligament/tendon injuries were most common ‘overall’ and ‘time-loss’ injury sustained by players between 2011 and 2014 (30% and 33% respectively). This was followed closely by concussion injuries, which accounted for 29% of ‘time-loss’ and 12% of ‘overall’ injuries. A large proportion of both ‘overall’ (57%) and ‘time-loss’ (55%) injuries occurred during the tackle event, with the tackler being injured more often than the ball-carrier (37% and 18% respectively). However, there were no statistically significant differences when comparing ‘overall’ and ‘time-loss’ IID between the different tournaments from 2011 until 2014. Discussion Significant differences were found when comparing ‘overall’ and ‘time-loss’ IID between the different tournaments from 2011 until 2014. Craven Week under-13 presented significantly greater ‘overall’ injury incidence densities. This finding contradicts previous literature within youth rugby research. The tackle (combination of tackler and ball-carrier) still accounts for the highest proportion both ‘time-loss’ and ‘overall’ injury events (57% and 55% respectively). This is in accordance with previous studies. However, a point of concern was that concussion accounted for 29% of all ‘time-loss’ injuries and 12% of all ‘overall’ injuries. This finding suggests a gradual increase in the number of concussions suffered during the SARU Youth Week Tournaments between 2011 and 2014. Further research is required to determine the reason for this pattern. Conclusion Further research within youth rugby cohorts is required to determine the risk associated with involvement at various level of participation. Injury prevention programs should place focus on reducing the prevalence of concussion at youth level by educating players and coaches about safe tackle techniques. Future studies should focus on local youth cohorts for seasonal

Includes bibliographical references / The studies described in this thesis were the first to investigate in - depth body composition, exercise performance and neuromuscular characteristics of elite Paralympic athletes with cerebral palsy (CP). In the first study, in - depth whole body and site specific body composition was investigated in six athletes with CP using dual - energy x - ray absorptiometry. There were no differences between non - affected and affected sides with respect to bone mineral density and fat mass. Fat free soft tissue mass was lower on the affected side in bot h upper and lower limbs of the athletes. The novel findings of this study provided the first insight into anthropometric and bone physiology of elite Paralympic athletes with CP, and the possible residual effect of CP in these individuals. In the second study, five athletes with CP and 16 able - bodied (AB) age and performance matched controls performed a 30 second Wingate sprint cycle test. Power output was significantly higher in the AB group, although fatigue indices were statistically similar between groups. Muscle activity changed similarly in all muscle groups tested, in both affected and non - affected sides, in both CP and AB groups. However, certain neuromuscular irregularities were identified in the CP group. The similarity in fatigue profile was a novel finding. It was proposed that this similarity in fatigue was the result of long term high level athletic training required for Paralympic competition. Study three tested the similarity in fatigue between CP and AB athletes (that was described in the second study), using an externally paced fatiguing running trial. Six athletes with CP and 12 AB athletes performed one 40 m sprint test and vertical jump tests off both legs, the affected leg individually and the non - affected leg individually, before and after an adapted multistage shuttle run test to exhaustion. The 40 m sprint test, vertical jump off both legs and vertical jump off the affected leg were significantly compromised in the CP group, while vertical jump off the non - affected leg was similar between groups. Both groups fatigued similarly with regard to performance and muscle activity. The third study's finding s generally supported those of the second study. However, it was shown that although athletes with CP may represent a group of individuals who have achieved maximal physiological adaptation toward AB levels, the activity generated by both legs was performed towards the capacity of the affected leg. Study four attempted to elucidate explanations for the novel findings in studies 2 and 3 through investigation of pacing strategies employed by these athletes. Six athletes with CP and 13 AB athletes performed two trials of eight sets of ten shuttles (totalling 1600 m). One trial was distance deceived and the other was non distance deceived. The CP group ran slower than the AB group in both trials, and differences in pacing were observed in the deceived trial in the CP group. This novel study provided evidence for a possible pacing strategy underlying the exercise performance and fatigue profiles observed in the athletes with CP documented in the previous studies. The work described in this thesis lends novel insights and understanding to the physiology and physiological adaptations of highly functioning ambulant athletes with CP. The findings might have important implications with respect to the understanding of rehabilitation, coaching and clinical management of individuals with CP.

Static stretching is commonly performed by athletes and clinicians on the assumption that it increases joint range of motion (ROM). However, observations from our laboratory indicate that there is an apparent inter-individual variance in the change in ROM in response to static stretching. Furthermore, prolonged repetitive loading has also been shown to affect ROM of a joint or series of joints. In particular, runners have a significantly decreased hamstring ROM. The aim of this study was to investigate the factors which are associated with a change in ROM in response to 1) a static stretch session and 2) prolonged repetitive loading. The findings showed that there is a variable response in the change in ROM following both a SSS intervention and in response to participation in a 42.2 or 56 km road race. More specifically, about 10% of the participants had a reduction in ROM after the SSS while the majority of participants had a reduced ROM after completing the marathon or ultra marathon.

The overarching purpose of this study was to investigate the multifactorial aetiology of Achilles tendinopathy. Variables such as training and injury history, lower limb alignment, running biomechanics and movement variability were investigated. This study also aims to understand how different sensory inputs, such as shoe or pain may affect biomechanics. Thirty four uninjured runners (UN) and twenty one runners with Achilles tendinopathy (AT) composed the population sample for this study. Questionnaire and lower limb measurements were used to investigate the multifactorial aetiology of the injury. Selected kinetic, kinematic and muscle activity parameters were employed to analyse the biomechanics aetiology of the injury. Runners performed 10 running trials at a self-selected speed in two running shoes with different hardness. Additionally, the UN runners ran for 10 min on the treadmill while the AT runners ran on the treadmill until they developed pain in the Achilles tendon. The results of the study showed that AT runners have a higher incidence of previous injury, run for more years, have reduced hamstring flexibility, and are heavier and taller than uninjured runners. The study also showed a reduced integrated electromyography activity (IEMG) of tibialis anterior and rectus femoris in the AT group during the running cycle. Stride to stride variability was similar between UN and AT runners but the biomechanics variability between participants were lower for the AT runners. No specific biomechanical adaptations were found between the two different shoe conditions (soft vs. hard). Similarly, biomechanics parameters were not altered at the onset of pain, but the reduced IEMG activity of tibialis anterior and rectus femoris were presented before and during the pain condition. It can, therefore, be concluded that runners with Achilles tendinopathy present reduced muscle activity during running which may be either a novel aetiological factor, or an adaptive response to the injury. The lower variability between runners with AT may indicate that these runners are less able to adjust their biomechanics according to their different functional behaviour or external input signals but this may require further investigation. Finally, it can also be concluded that the sensory inputs such as shoes and pain do not change this muscle activity pattern.

The central governor model (CGM) proposes that muscle recruitment is regulated by the brain through subconscious homeostatic control of afferent feedback. It has been suggested that the dopaminergic system plays a key role in the CGM, with dopaminergic activation leading to lower fatigue thresholds. Key neural circuits, including the central autonomic network (CAN), may also play a role in altering thresholds by reducing conscious bodily awareness, known as interoception. However, few studies have directly examined the dopaminergic neurotransmitter system and CAN connectivity, during exercise-induced fatigue. Although there is reason to suspect that subjects with high activity (HA) and low activity (LA) levels have different fatigue thresholds, potential differences in dopaminergic function and related circuity have not been compared across these groups. In this thesis, I was therefore interested in examining the impact of a dopaminergic agent, methylphenidate (MPH) on: a) brain connectivity at rest before exercise and during exercise, b) on heart rate (HR) at rest before exercise and during exercise. Furthermore, c) I was interested in whether the effect of MPH administration (MA) on HR and functional connectivity (FC), between CAN regions, pre-exercise impacted HR and FC during exercise and whether this differed between HA and LA subjects. In order to measure the effect of MA on brain connectivity at rest and during exercise I used a functional magnetic resonance imaging (fMRI) scanner. To measure the effect of MA on fatigue thresholds, I used a MRI compatible handgrip device which recorded grip force output during a fatiguing handgrip task. To measure HR before and during handgrip exercise I used a MRI compatible electrocardiogram (ECG) machine. My first study aimed to determine the effect of MA on resting state (R-fMRI) FC between CAN regions and its correlation with HR before exercise. My second study aimed to see the effect of MA on task HR, force output and, FC and EC between CAN regions, during a fatiguing handgrip task. My third study aimed to see the effect of MA on the relationship between HR and FC pre-exercise and HR and FC during exercise and whether this differed between LA and HA subjects. I found that: 1.) MA decreased R-fMRI FC between CAN regions, and that this was associated with an increase in HR during resting state 2.) MA decreased FC between CAN regions with an associated increase in task HR and force output during a fatiguing handgrip task 3.) After MA, R-fMRI FC determines task FC independent of HR 4.) MA significantly increased task HR in LA but not HA subjects 5.) After MA, R-fMRI FC determines task FC independent of HR in LA but not HA subjects. Taken together, the functional uncoupling of CAN regions and increased HR after MA suggests that MA increases the fatigue threshold by diverting attention away from interoceptive cues, such as HR. However, this finding may be more relevant in LA subjects, where MA is also associated with increased HR during exercise.

Includes bibliographical references. / Previous findings have suggested that functional variants within collagen encoding genes are associated with several musculoskeletal soft tissue injuries and other exercise-related phenotypes. Specifically variants within the functional COL5A1 3’- untranslated region (UTR) have previously been associated with (1) chronic Achilles tendinopathy, (2) Anterior Cruciate Ligament (ACL) ruptures in females, (3) endurance running performance and (4) range of motion (ROM). Since this gene encodes for an important structural component of the collagen fibril it has been hypothesised that variants within other collagen fibril encoding genes, such as COL3A1, COL6A1 and COL12A1, will also be associated with these and/or other musculoskeletal soft tissue injuries and exercise-related phenotypes. The COL5A1 rs12722 and COL12A1 rs970547 gene variants have been previously associated with risk of ACL ruptures in females [153;154] and/or chronic Achilles tendinopathy [131;181]. The first aim of this thesis was therefore to investigate the COL3A1 rs1800255 and COL6A1 rs35796750 gene variants as risk factors for these musculoskeletal soft tissue injuries.

Introduction: In recent years, there has been a move to understand the environment and context in which athletes develop. South Africa’s unique context provides an opportunity to understand how environmental factors could influence talent development in cricket. Since democracy, there has been limited representation of Black African cricketers at the elite levels in South Africa. Therefore, the aim of this thesis was to determine the role that socio-ecological factors may play in the development of cricket talent in a diverse society. Methods: Qualitative research methods were used to explore the experiences and perceptions of South Africa’s male cricketers as they progressed through the talent pathway from exposure to the game to the elite level. The perceived effectiveness of the introduction of an ethnic target policy was also explored. Seventy-one semi-structured interviews were conducted with a purposive sample of players from all ethnic groups (n=43), and with knowledgeable and experienced key informants (n=16). A thematic analysis of the data resulted in the identification of themes which are presented using a multi-level socio-ecological framework. Results: All players progressed to the elite level; however, their access points to and routes through the pathway varied. This progress was influenced by the inter-relationship of distal and proximal socio-ecological factors that they experienced during their cricketing careers. These influences can be summarised into five talent development components that acted either as barriers or enablers to progress: (1) access to opportunities and competition, (2) holistic player development, (3) effective support networks, (4) inclusive team environments, and (5) adaptive mind-sets. In addition, various intrapersonal characteristics were identified that further affect a player’s ability to achieve elite cricketing success. Finally, it was determined that an ethnic target policy alone is not an effective intervention for developing cricket talent in a diverse society undergoing transition. Conclusion: A socio-ecological framework to talent development lends additional support to the idiosyncratic, multifactorial, dynamic and complex way in which cricket expertise is achieved, particularly in diverse societies. It provides stakeholders involved in the talent development process with evidence to inform policy and practice, as well as design effective interventions.

Includes bibliographical references. / Barefoot running is a subject of significant interest, both in scientific publications and in the lay media as a result of its alleged benefits for runners. These benefits include the potential to reduce injury risk, more economical running and broadly speaking, a better understanding of running biomechanics. Although there are numerous scientific publications describing differences between barefoot and shod running, there is a dearth in understanding whether all runners are able to adapt to the proposed benefits and how this may affect long-term injury risk. Thus, we sought to investigate the biomechanical, neuromuscular and metabolic changes associated with habitually shod runners during the transition to pure barefoot running over an 8- week progressive training programme. This thesis begins with a critical review of the literature, which evaluates the theories and evidence for barefoot running, as well as describing the necessary future research to confirm or refute the barefoot running hypotheses. Our first study aimed to describe acute changes occurring in habitually shod runners when first exposed to barefoot running. We were particularly interested in the variability in response, and whether we could identify factors that predicted potentially favourable changes in kinematic and kinetic outcomes. Fifty-one runners were recruited and assessed using a 3-D motion capture system and integrated force platforms using conventional methods. We found that loading rate was significantly greater in the barefoot condition, but that high individual variability existed, particularly in the barefoot trials. We found that an increase in ankle dorsiflexion is associated with an increase in initial loading rate when in the barefoot condition, supporting previous findings in this regard. We then performed a supervised, pure barefoot running training programme, over 8 weeks, to determine whether the biomechanics of barefoot running would adapt gradually to habituation. Twenty-three runners were recruited for participation, and performed comprehensive biomechanical and neuromuscular assessments before and after the 8-week programme. The first finding was runners do not adapt similarly to barefoot training, and that biomechanics do not change significantly over the 8-week period. High variability in ankle kinematics and loading rate were found, with three sub-groups identified, namely positive responders (reduced loading rate after training), non-responders (no change in loading rate) and negative responders (increase in loading rate after training). We found significant associations between initial loading rate the changes in ankle flexion angle at initial ground contact, presumably as a result of its influence on footstrike. This finding suggests that conscious instruction might be necessary in order to achieve reductions in collision forces during barefoot running. With respect to neuromuscular variables, a persistently higher gastrocnemii muscle preactivation was found in the barefoot condition before and after the training intervention. Increased gastrocnemius pre-activation was associated with lower initial loading rate. An increase in gluteus medius and peroeus longus and a decrease in tibialis anterior pre-activation were also associated with a reduction in initial loading rate after barefoot training. This finding suggests a refined neuromuscular activation strategy prior to ground contact in the barefoot condition to stabilize the hip and centre of mass. Lastly, oxygen cost of transport was found to improve as a result of the barefoot training programme in the male runners and this improvement was found to be associated with a decrease in ground contact time and increase in stride frequency, but no a change in ankle flexion angle at initial ground contact. The outcomes from this thesis elucidate the highly variable response of individuals to barefoot running. This advises individuals choosing to transition to barefoot running to do so with caution. With this in mind, we suggest certain characteristics that may be used as screening mechanisms to indicate individual suitability to barefoot running based on the “collision force theory”. Further, benefits associated with barefoot running other than varied responses in initial loading rate such as improvements in oxygen cost of running are pre-dominantly hypothesised to be a result of musculo-tendinous adaptations, neuromuscular strength and motor control.

Includes bibliographical references. / A collection of published studies establishing the central governor model of exercise physiology, and the aetiology of the hyponatraemia of exercise.

Submaximal step tests are used to predict maximal oxygen consumption and work capacity. However, if the external workload is not controlled the interpretation of the test results may be inaccurate. The purpose of the research was to develop a submaximal test of cardiorespiratory fitness using a novel step test designed specifically to overcome the weaknesses of the previously published step tests. A series of studies contributed to the theoretical development of the submaximal step protocol, piloting the protocol, reliability studies, validation of the protocol and finally a cross-validation of the protocol. The first study tested the hypothesis that stepping tests configured for the same external workload, but varying in stepping frequency, elicit the same physiological stress. Participants (n = 31) performed three step tests at 16, 20 and 24 steps per minutes in random order. External workload was standardised at 45 kJ. Energy expenditure, heart rate recovery, rating of perceived exertion, maximum heart rate and total heart beats were significantly different between tests (p < 0.05) with the biggest differences occurring between 16 and 24 steps per minute. Maximum heart rate as a percentage of age predicted heart rate increased from 70% at 16 steps per minute to 81% at 24 steps per minute. The study concluded that standardisation of external workload with different exercise intensities does not result in the same physiological responses. The second study tested the reliability of the step test. Participants (n = 34) performed a step test three times in a week at a cadence of their choice (16, 20 or 24 steps per minute). The study showed that the step test is repeatable for most variables measured and therefore is a reliable test of fitness. The third study used the outcome variables measured during the step test to develop equations which predicted VO₂max measured directly in a maximal test on a treadmill. A diverse sample of participants (n = 273), differing in sex, level of habitual physical activity and age were recruited for the study. Several models for predicting VO₂max were determined. The most parsimonious equation was: VO₂max (ml.kg⁻¹.min⁻¹) = -0.10911 (age) - 0.06178 (body mass) - 0.75481 (body fat %) +0.00208 (METS) + 0.11636 (HRR) - 0.019551 (MHR) + 0.07955 (Av HR) + 83.34846 (R² = 0.75, standard error of estimate = 5.51 ml.kg.min⁻¹) where METS is metabolic equivalent, HRR is heart rate recovery, MHR is maximum heart rate and Av HR is average heart rate. Cross validation was done (n = 50) to test the accuracy of the prediction equation. The relationship between the predicted VO₂max and the measured VO₂max was r = 0.87. In conclusion the standardised step test can predict VO₂max in a heterogeneous population of males and females, varied ages (20 to 60 years), physical activity levels and fitness levels.