Anatomical, Biomechanical and Physiological Loading During Human Endurance Performance at Selected Limb Cadences Via Triathlon

Anatomical, biomechanical and physiological loading during human endurance performance at selected limb cadences via triathlon

GRANT JUSTIN LANDERS BSc (Hons 1A)

Department of Human Movement and Exercise Science The University of Western Australia

This thesis is presented in fulfilment of the requirements for the degree of Doctor of Philosophy at The University of Western Australia

September 2002

Supervisors Professor Brian A. Blanksby Associate Professor Timothy R. Ackland

Dedication

I wish to dedicate this thesis to three generations of love and support.

Rebecca Elizabeth Landers Who passed away October 2000. She has always given me inspiration to achieve; to keep my eye on the ball.

Mum and Dad Yet again for their continual support in all aspects of life, making every opportunity a possibility.

Holly Bruse The love of my life; the one who has kept me on track yet create enough distractions, such as a marriage, to ensure my enjoyment.

Triathlon has had a rapid evolution from its origins 25 years ago in the United States to becoming a full summer Olympic sport in Sydney 2000. It is a sport that combines the three disciplines of swimming, cycling and running linked together with two transitions. It is this combination of events that gives triathlon its uniqueness in the area of exercise science. As a very young sport, the body of knowledge is somewhat limited, but is steadily growing. The following document aims to shed some new light on a range of aspects within the sport of triathlon and highlight triathlon as a sport in its own right with very specific demands when compared with each of the three individual sub-disciplines.

iii

Acknowledgements

With the help of those below, the completion of my PhD has been made not only easy but also a pleasure. I thank all of you for your time, effort and understanding throughout the last three years.

Study One Anthropometry of 1997 triathlon world championships To the competitors who gave up their time and participated so willingly. To Bob Eklund for helping me understand factor analysis so that I could decipher the large amounts of data. Again thank you to those who helped with data collection.

Study Two Swim significance on race outcome Ron Monson for Mathmatica statistical analysis and commentary.

Study Three Cadence during a race (Age group selection race) Renae Landers for her camera skills Holly Bruse for the loan of her camera Mark Batten (Triathlon Western Australia) for allowing the event to be filmed

Study Four Cadence during a race (ITU 2000 triathlon world championships) Uniswim for shirts EventsCorp for permitting filming Holly Bruse for her videography skills on camera one from the spa Brett Landers & Scott Beatty for being my helpers on the day and running around like little troopers. Veronica Wilkinson from EventsCorp/Tourism WA for providing the results in excel Stuart Fuller for help in deciphering the female run course

Study Five Body size and triathlon running Those from study one and study four who helped again with the collection of data

Study Six Unloaded cadence Rob Duffield and Dr Brian Dawson for help with the physiological aspects. Nataphoon Benjanuvatra for his assistance with EMG To the subjects who came and participated without question. iv

The Other Special People

Technicians; for their untiring support. For always asking plenty of questions and helping me get the most out of each testing session.

Darren Smith; who gave me an opportunity to experience elite triathletes first hand and who acted as both an educated scientist and triathlon expert when discussing ideas and possible investigations. Your time has been greatly appreciated.

Nat, a fellow swim coach, student and friend who has been available throughout for all sorts of discussions, whether it be pure or social science.

Tim Ackland; for his unbiased observation of the thesis and its 'contribution to new knowledge'.

To Professor Blanksby. The countless hours that you have devoted to me and this research program over the last four years and the support offered to me during my undergraduate studies has not gone unnoticed. I thank you for your commitment to me and your belief in what I could achieve. Those lengthy discussions where we spread those antlers both within and outside university have opened my eyes to the greater world. The man of so much wisdom, a friend for life. v

Preface

This thesis contains five studies that sought to systematically use sport science to determine important physical, physiological and biomechanical aspects of triathlon, which could improve classic distance triathlon performance.

Study One The initial research continued on from an earlier study in which the author examined the influence of human morphology on triathlon race outcome. As with other endurance and weight bearing sports, low levels of adiposity and measures of proportionality were found to be predictors of triathlon performance. Those with proportionately longer limbs recorded faster total times but no common variance was found between measures of anthropometry and cycle performance during triathlon competition.

Study Two Having examined the morphology of triathletes, this study aimed to investigate the relative importance of each of the three sub-disciplines in triathlon. To date, the cycling section has been credited with the greatest impact as it contributes approximately 50% of the total race time. With the legalisation of drafting in classic distance events for elite competitors, the level of significance of each of the three disciplines has altered. It was found that 80% of winners came from the first pack of swimmers to exit the water and that the run discipline exerted the greatest influence on overall finishing position.

Study Three No previous research had focused on limb cadences used during a triathlon. Study two had suggested that the run section of a triathlon had the greatest bearing on final race outcome. Therefore, it was considered important to investigate how the run might be improved by altering the cycle mechanics. Thus, study three examined what leg speed was actually used by male and female age-group triathletes during both the ride and run portions of a race, and whether there was any relationship between the two disciplines. Cycle cadence was found to be greater at the end of the cycling leg when compared with vi the start of the cycling section and the run stride rate. The initial cycling cadence was similar to the running stride rate selected by the triathletes during the event. Faster stride rates were found in the first 200 m of the 10 km run, for both male and female competitors.

Studies Four & Five This study continued the examination of selected cadences during the cycle and run sections while in competition. The sample in this study included senior elite male and female triathletes competing at the 2000 Triathlon World Championships. Cadence remained consistent through most of the cycling stage but was significantly increased during the final few kilometres for both males and females. The females also pedaled faster during the first 3 km. During the run, stride rate again remained relatively constant and the variation in stride length was related to run and triathlon performance. Male triathletes used a faster cadence and stride rate, and longer strides than the female competitors. During the cycling and running stages, leg frequencies were comparable for both males and females.

Study Six This study attempted to integrate the knowledge surrounding body size and shape of triathletes with the stride rate and stride length selection during competition. Previous research had highlighted the importance of lever lengths and low levels of adiposity on performance, and that those who maintained longer strides and faster rates ran faster and performed better. Correlations were conducted between measures of body size (height and mass) with stride rate and stride length. The results from 58 senior elite triathletes (30 female and 28 male) indicated that taller and heavier triathletes used longer strides.

Studies Seven & Eight The final study sought to investigate various cycling cadences of triathletes in an attempt to improve the transition from cycle to run phases, and thereby, enhance final triathlon performance. The cycle cadences were presented to triathletes in random order and in an unloaded fashion. Following the laboratory based cadence trials, the vii transitioning run mechanics were then analysed using electromyography and videography. The results indicated a linear increase in the oxygen cost of pedaling with an increase of cadence from 50 to 110 rpm before a significant increase at 130 rpm. Initially, a higher oxygen cost of running was noted in most cases, and there was a sustained increase in the oxygen cost of running following cycling at 130 rpm. Some biomechanical variations were noted in the initial stages of the run but these returned to the normal pattern within 2 minutes of running.

viii

Publications

Scientific Publications

Landers, G.J., Blanksby, B.A., Ackland, T.R. and Smith, D. (2000). Morphology and performance of world championship triathletes. Annals of Human Biology, 27(4), 387-400.

Ackland, T.R., Blanksby, B.A., Landers, G. and Smith, D. (1998). Anthropometric profiles of elite triathletes. Journal of Science and Medicine in Sport, 1(1), 53-56.

Ackland, T., Blanksby, B., Landers, G. and Smith, D. (1998). Anthropometric correlates with performance among world championship triathletes In; Norton K., Olds, T. and Dollman, J. (eds.) Kinanthropometry VI, Proceedings of the Sixth Scientific Conference of the International Society for the Advancement of Kinanthropometry, 1998, 92-104.

Landers, G.J. (1998). Kinanthropometry and performance of elite triathletes. Unpublished Honours Thesis, The University of Western Australia.

Conference proceedings

Landers, G.J., Blanksby, B.A., Ackland, T.R. and Monson, R. (2001). Swim Position and its effect on triathlon outcome. International Triathlon Coaching Symposium, July 23-24 2001, Edmonton, Canada, page 11.

Landers, G.J., Blanksby, B.A. and Ackland, T.R. (2001). Cadence selection and performance. International Triathlon Coaching Symposium, July 23-24 2001, Edmonton, Canada, page 13.

Landers, G.J., Blanksby, B.A., Ackland, T.R. and Smith, D. (1999). Kinanthropometry and performance of elite triathletes. The Book of Abstracts: Fifth IOC World Congress, 31st October -5th November, 1999, Sydney, Australia, 129.

Landers, G.J., Blanksby, B.A., Ackland, T.R. and Smith, D. (1999). Kinanthropometric differences between world championship senior and junior elite triathletes. Maximising Olympic Distance Triathlon Performance: A multidisciplinary perspective, Proceedings from the Gatorade International Triathlon Science II Conference, Noosa, Australia, November 7-8, 1999, 74-87.

Ackland, T.R., Blanksby, B.A., Landers, G. and Smith, D. (1998). Anthropometric correlates with performance among world championship triathletes - abstract. In; Australian Conference of Science and Medicine in Sport, Adelaide Convention Centre, Adelaide, 13-16 October 1998 : Abstracts, Sports Medicine Australia, 1998, Canberra, p.58. http://www.ausport.gov.au/fulltext/1998/acsm/smabs058.htm

Table of Contents i Dedication iii Acknowledgements v Preface viii Publication ix Table of contents xiv List of figures & tables

Chapter One 1 Introduction 2 Aims 2 Research hypotheses 3 Significance of the study 4 Limitations 4 Delimitations 5 Operational definitions

Chapter Two 8 Literature review 8 Morphology and performance 10 Well trained/elite triathletes (classic distance) 12 Recreational triathletes 13 Female triathletes 13 Overview 14 General principles - male and female differences 15 Cadence 16 Self selected cadence 20 Biomechanical interaction of cycling and running 25 Anthropometry relationship with cadence 26 Relationship between cadence and blood lactate 28 Relationship between cadence and electromyography 31 Cadence variation prior to change in disciplines 32 Muscle flexibility 33 Cadence and ground reaction forces

Chapter Three 35 Study one: Morphology and performance of world championship triathletes 37 Introduction 39 Methodology 39 Sample 40 Testing protocol 41 Data analysis 42 Results 42 Performance correlations - all competitors 43 Elapsed times for elite vs junior competitors x

44 Factor analysis 45 Stepwise linear regression 48 Discussion 48 Performance correlation for the entire field 48 Junior vs senior competitors 49 Factor analysis 50 Stepwise linear regression 53 Summary

Chapter Four 55 Study 2: Race analysis 56 Introduction

Chapter 4A 58 Study 2A: Swim positioning and its effect on triathlon outcome 58 Methodology 58 Sample 58 Testing protocol 58 Data analysis 59 Results 63 Discussion 65 Summary

Chapter 4B 67 Study 2B: Relative contribution of each discipline and the effect on triathlon outcome 67 Methodology 67 Subjects 67 Data analysis 68 Results 68 Descriptive data & correlation 69 Winners position 71 Groupings 73 Discussion 73 Descriptive data & correlation 74 Winners position 75 Groupings 76 Summary

Chapter Five 77 Studies 3, 4 & 5: Cadence selection and performance 79 Introduction 80 Cadence determination, selection and race performance

82 Study 3: Cadence selection and performance of age-group triathletes 82 Methodology 82 Sample xi

82 Testing protocol 83 Data analysis 84 Results 84 Performance times 84 Cadence 86 Relationship with performance 88 Discussion 89 Relationship with performance 90 Summary

92 Study 4: Cadence selection of senior elite male and female triathletes at the 2000 Triathlon World Championships 92 Introduction 92 Methodology 92 Subjects 92 Testing protocol 93 Data analysis 95 Results 95 Female cycle results 96 Male cycle results 97 Run data 98 Female run results 98 Descriptive statistics for all female triathletes 99 Correlations between performance measures 101 Male run results 101 Descriptive statistics for all male triathletes 102 Correlations between performance variables 104 Discussion 104 Female cycle results 105 Male cycle results 105 Female run results 105 Descriptive statistics for all female triathletes 107 Correlations between performance measures 108 Male run results 108 Descriptive statistics for all male triathletes 109 Correlations between performance variables 110 Summary

111 Study 5: Relationship between cadence selection and performance of senior elite male and female triathletes at the 2000 Triathlon World Championships 111 Methodology 111 Subjects 111 Data analysis 111 Results 111 Female cycle 114 Male cycle 117 Female run 118 Male run xii

120 Discussion 120 Female cycling 121 Male cycling 123 Female run 124 Male run 126 Summary

Chapter Six 127 Study 6: Running stride rate, stride length and body size relationship for senior elite triathletes at the 2000 Triathlon World Championships 128 Introduction 131 Methodology 131 Sample 132 Testing protocol 132 Data analysis 133 Results 134 Discussion 134 Anthropometry 135 Female 135 Male 137 Summary

Chapter Seven 138 Study 7 & 8: The effect of cycle cadence on subsequent running 140 Introduction

144 Study 7: Physiological effect of cadence alteration on cycling and subsequent running 144 Methodology 144 Sample 144 Testing protocol 145 Data collection 146 Data analysis 146 Results 147 Physiological response to unloaded cycling 149 Physiological response to running following cycling 151 Discussion 151 Physiological response to unloaded cycling 152 Physiological response to running following cycling 153 Summary

Chapter Eight 154 Study 8 Biomechanical effect of cadence alteration during cycling on subsequent running 154 Methodology 154 Sample 154 Data analysis xiii

155 Results 155 Stride rate and stride length 158 EMG during cycling 161 EMG during running 164 Discussion 164 Stride rate and stride length 165 EMG

166 SR, EMG and O2 167 Summary

Chapter Nine 168 Summary, conclusions and future study recommendations 168 Summary 168 Study 1 169 Study 2A 169 Study 2B 169 Study 3 170 Study 4 171 Study 5 171 Study 6 171 Study 7 172 Study 8 172 Conclusions 173 Recommendations for further study

References 175 References

Appendices Appendix A. Publications Appendix B. Abbreviations Appendix C. Summary of Triathlete Characteristics Appendix D. Study 1: Informed Consent Form Appendix E. Study 1: Data Collection Card Appendix F. Landmarks and Techniques Appendix G. Formulae Appendix H. Study 2: Raw Data Appendix I. Study 2: Group Sizes Appendix J. Study 3: Raw Data Appendix K. Study 4: Race Results Appendix L. Study 5: Consent Form Appendix M. Study 5: Data Collection Card xiv

List of Figures

Chapter 1 7 Figure 1.1 Different hand positions while cycling; upright, drops and aero.

Chapter 4A 60 Figure 4.1 Number of male triathletes out of the water in the first pack and, subsequently, those who finished in the top 10 of each race during the 1999 Triathlon World Cup Season. 61 Figure 4.2 Number of female triathletes out of the water in the first pack and, subsequently, those who finished in the top 10 of each race during the 1999 Triathlon World Cup Season. 63 Figure 4.3 Number of triathletes in the first pack of swimmers from the total number of starters for all male and female events. 63 Figure 4.4 Total number of top 10 finishers from those triathletes who exited the water in the first pack for both males and females.

Chapter 4B 70 Figure 4.5 Average male placing after swim cycle & run for the winner, top 5 & top 10. 70 Figure 4.6 Average female placing after swim cycle & run for the winner, top 5 & top 10. 71 Figure 4.7 Male average pack number and size at end of swim, cycle and run. 72 Figure 4.8 Female average pack number and size at end of swim, cycle and run.

Chapter 5 85 Figure 5.1 Cadence employed by males, females, and combined males and females during a triathlon competition.

93 Figure 5.2 2000 Triathlon World Championships camera locations on the cycle and run course. 95 Figure 5.3 Male and female cycle cadence over time. 97 Figure 5.4 Male and female stride rates over time. 97 Figure 5.5 Male and female Stride lengths over time.

114 Figure 5.6 Female cycle cadence by pack. 115 Figure 5.7 Male cycle cadence by pack.

Chapter 7 148 Figure 7.1 Oxygen cost of unloaded cycle at different cadences. 149 Figure 7.2 Heart rate response to different unloaded cycle cadences. 150 Figure 7.3 Oxygen cost of running after unloaded cycling at various cadences. 151 Figure 7.4 Heart Rate values at each minute after different unloaded cycling cadences.

Chapter 8 156 Figure 8.1 Running stride rate following unloaded cycling at various cadences. 157 Figure 8.2 Average knee angle during the first minute of running following unloaded cycling at various cadences. 158 Figure 8.3 Average hip angle during the first minute of running following unloaded cycling at various cadences. 159 Figure 8.4 Tibialis Anterior activity during unloaded cycling at various cadences. 160 Figure 8.5 Gastrocenemious activity during unloaded cycling at various cadences. 160 Figure 8.6 Vastus Medialis activity during unloaded cycling at various cadences. 161 Figure 8.7 Biceps Femoris activity during unloaded cycling at various cadences. 162 Figure 8.8 Tibialis Anterior activity during running after unloaded cycling at various cadences. 163 Figure 8.9 Gastrocenemious activity during running after unloaded cycling at various cadences. 163 Figure 8.10 Vastus medialis activity during running after unloaded cycling at various cadences. 164 Figure 8.11 Biceps femoris activity during running after unloaded cycling at various cadences. xvi

List of Tables

Chapter 2 27 Table 2.1 Results from Ito et al. (1983) for each subject at OBLA.

Chapter 3 43 Table 3.1 Rank order correlations of swim, cycle, run and end bike position rank with final rank. 44 Table 3.2 Descriptive statistics & ANOVA summary of comparative times for elite versus junior competitors. 46 Table 3.3 Complete principal component matrix coefficients by which variables were multiplied to obtain factor scores and total variance explained. 47 Table 3.4 Factor B weights and regression coefficients in the final regression model for the entire sample. 47 Table 3.5 Factor B weights and regression coefficients in the final regression model for the male sample. 47 Table 3.6 Factor B weights and regression coefficients in the final regression model for the female sample.

Chapter 4A 60 Table 4.1 Number of male triathletes out of the water in the first pack and, subsequently, those who finished in the top 10 of the race during the 1999 Triathlon World Cup Season. 61 Table 4.2 Number of female triathletes out of the water in the first pack and, subsequently, those who finished in top the 10 of the race during the 1999 Triathlon World Cup Season. 62 Table 4.3 Totals and percentages of the first pack of triathletes to alight from the water and the relationship with final finishing position.

Chapter 4B 68 Table 4.4 Means, standard deviations and percentages for male and female triathletes across 10 events. 69 Table 4.5 Rank order correlations of the final finishing position with the time for each of the swim, cycle and run disciplines. 69 Table 4.6 The average, mean position of triathletes at the end of the swim and cycle for the winners, top 5 and top 10 finishers.

Chapter 5 84 Table 5.1 Average split times (h:m:s) and total times (h:m:s) for all male and female triathletes. 85 Table 5.2. Cadence (rpm) used throughout the cycle and run portions of a triathlon for all competitors combined, and male and female competitors separately. 86 Table 5.3 Level of skewness and kurtosis of split and total times, for both male and female triathletes combined. xvii

87 Table 5.4 Correlation matrix: relationships between cadence and times for all male and female triathletes.

94 Table 5.5 Cycle cadence data collection points. 94 Table 5.6 Data collection points during the run. 96 Table 5.7 Descriptive cycle cadence statistics for all female triathletes. 96 Table 5.8 Descriptive cycle cadence statistics for all male triathletes. 98 Table 5.9 Descriptive run stride rate and stride length statistics for all female triathletes. 100 Table 5.10 Pearson correlation of run and triathlon performance with stride rate and stride length for female triathletes. 101 Table 5.11 Descriptive run stride rate and stride length statistics for all male triathletes. 103 Table 5.12 Pearson correlation of run and triathlon performance with stride rate and stride length for male triathletes.

113 Table 5.13 Descriptive cycle cadence statistics for female triathletes split by pack. 116 Table 5.14 Descriptive cycle cadence statistics for male triathletes split by pack. 117 Table 5.15 Descriptive stride rate and stride length run statistics for female triathletes split by run time. 119 Table 5.16 Descriptive stride rate and stride length run statistics for male triathletes split by run time.

Chapter 6 133 Table 6.1 Anthropometric data of body mass & height for male and female senior elite triathletes. 134 Table 6.2 Stride rate, stride length data & correlations with mass and height for male and female senior elite triathletes.

Chapter 7 147 Table 7.1 Descriptive anthropometric data for male, female and combined subjects. 147 Table 7.2 Physiological characteristics of male, female and all triathletes combined. 147 Table 7.3 Weekly training. 148 Table 7.4 Mean oxygen cost and hear rate values for cycling at various cadences. 149 Table 7.5 Mean oxygen cost at the start and end of run after different cycle cadences. 150 Table 7.6 Mean heart rate values during running after unloaded cycling at various cadences.

Chapter 8 155 Table 8.1 Selected running stride rate of triathletes following unloaded cycling at various cadences. xviii

156 Table 8.2 Correlations of triathlete stride rate and stride length at the end of the run with anthropometric variables. 157 Table 8.3 Trunk, thigh and leg angles during the first minute of running following unloaded cycling at various cadences. 159 Table 8.4 Muscle activity during unloaded cycling at various cadences. 162 Table 8.5 Muscle activation during running of biceps femoris, tibials anterior, gastrocenemious and vastus medialis after unloaded cycling at various cadences.

Chapter 1

Introduction

Humans are animals who can locomote on land and in water under their own power. However, humans locomote bipedally whereas most other animals tend to use quadrupedal ambulation. Terrestrial, bipedal locomotion can be either walking or running. Napier (1967) suggested that human walking is primarily an adaptation for covering long distances economically. The decision as to which of these two gaits is to be used is determined by the speed of travel and the anatomical structure, specifically lower limb length. As humans usually aim to walk or run with the greatest economy of motion, gait is adjusted to reduce energy cost (Alexander, 1984). Within each gait, Taylor (1985) showed that an optimal frequency of movement exists which minimises metabolic energy cost and maximises mechanical energy. This frequency, or stride rate, is related directly to whole body mass and could relate also to the resonant frequency of the whole body (Taylor, 1985).

Triathlon is an endurance sport that includes the three separate disciplines of swimming, cycling and running, performed in a continuous manner. Thus, efficiency of motion is a necessity throughout. Triathlon has a history of less than 30 years, following its origin in the United States of America in the early 1970's. Now, it is one of the world’s fastest growing sports and caters for all ages through a variety of distances. The sport took its place in the summer Olympic program for the first time in Sydney 2000.

There is limited research on triathlon and triathletes. In one of the few studies of elite triathletes Landers (1998) suggested that triathlon was similar to other endurance events where high performers had lower levels of adiposity than those who were less successful. From this study (Landers, 1998), normative morphological data were established for male and female, elite and junior triathletes (Ackland et al., 1998a). Further research was required to clarify performance and morphological issues in triathlon. 2

Aims The aims of this research program were to develop further morphology and kinanthropometry findings, and their relationships with performance. Hence, the following were investigated: • Determine which groups of variables are correlated to successful triathlon performance. • Analyse the effect of prior movement patterns on subsequent activities by examining how the cadence of limb oscillation is related to triathlon performance. • Determine cadences used by triathletes throughout competition and their relationship with performance. • Determine the effect of prior cycling on running mechanics and economy. • Determine the effect of an intervention program, based on a theoretical model, which alters triathlete cadences for improved biomechanical efficiency and metabolic energy cost of performance.

Research Hypotheses It was hypothesised that: (i) Triathletes will possess similar morphological characteristics to those in other endurance sports. (ii) Triathletes choose a similar leg cadence for all three disciplines (within the range 80-90 rpm) in order to optimise economy of motion. (iii) The swim portion has a greater importance in draft legal events than cycling. (iv) The best runners in the first group of swimmers have the greatest chance of winning. (v) There is very little change in athlete position between the end of the swim and the commencement of the run, that is during the cycle discipline. (vi) Cadence in prior disciplines (especially cycling) will affect the cadence of the following discipline (especially cycling to running). (vii) Posture during the cycling discipline will affect both mechanics and economy of motion during the run portion of a triathlon and (viii) There is an underlying relationship between anthropometry and cadence. 3

Significance of the Study Currently, there is limited research directed at the relatively new sport of triathlon. Originally, triathlon was seen as a combination of three separate sports, swimming, cycling and running. However, as the earlier disciplines of swim and cycle can have an impact on the disciplines that follow (cycle and run), for success in competition, triathlon must be viewed as a single sport when training and, for research purposes also.

A review of literature has highlighted that minimal research has been conducted on multidisciplinary physical activities. Research on multidisciplinary athletes has rarely used continuous multidisciplinary tasks. Therefore, it was considered important to investigate the effects of multidisciplinary activities such as triathlon in terms of morphological, biomechanical and physiological demands. Only then can the extent of the interaction of prior activities on those that follow be determined.

Preliminary studies by the author have provided anthropometric characteristics of senior and junior elite, classic distance triathletes (Ackland et al., 1998a; Landers, 1998). Hence, normative data are now available for use by athletes or coaches. However, there has been no analysis of how performance is related to these characteristics. The inclusion of triathlon in the Olympic Games has highlighted the void of information to be filled from triathletes who compete in the Olympic distance event. Thus, it was considered important to identify any physical characteristics that might be advantageous for competitors over this classic distance.

Studies have been carried out in other sports to assist coaches in talent identification and performance (Bloomfield et al., 1994). Humans demonstrate anatomical differences, some of which are advantageous for one sport, but not for others. For example, weight lifters and gymnasts have relatively long torsos and short limbs. High jumpers are tall with relatively long lower limbs in comparison with the trunk, and a high crural index (Bloomfield et al., 1994). Therefore, it is useful for the coach or sport scientist to have a set of morphological preferences that identify athletes who are anatomically most suited to a particular sport. 4

The relationship of limb cadence with economy of movement is another area requiring investigation. While considerable research has been conducted in the individual sports of swimming, cycling and running, only three previous studies have examined triathlon and cadence (Hausswirth et al., 1997; Quigley & Richards, 1996; Vercruyssen et al., 2002). It is assumed that the most efficient stroke rate/stroke length, cadence/power output, stride rate/stride length relationships are adopted by successful athletes to optimise economy of motion and, consequently, performance (Alexander, 1984; Craig & Pendergast, 1981; Hausswirth et al., 1999; Marais et al., 1999; Moritani et al., 1993; Napier, 1967; Takaishi et al., 1996; Taylor, 1985; van der Woude et al., 1989). However, with the three disciplines combined in triathlon, the optimal rates may vary from those used in the individual sports. Different cadences could be necessary to reduce the effect of prior movement patterns on those that follow.

Therefore, triathletes' average "cadences" during cycling and running need to be determined. The relationship with performance or trends in spontaneous cadence selection could add significant insight as to how to optimise training and competition strategies. It could also resolve the issue of whether the selection of cadence in the prior discipline has an immediate or prolonged effect on the cadence and/or economy of the following discipline. This information forms the basis of further investigations in the areas outlined below in the related literature.

Limitations • Symmetry is assumed between left and right leg movements. • Most data collection took place in the field in order to obtain competition condition results and could be influenced by a number of uncontrollable variables such as wind, rain, temperature and the effect of crowd. Some laboratory testing was conducted.

Delimitations • Data collection of triathletes in study one was delimited to that which could be obtained from the 1997 Triathlon World Championships (TWC) via athletes who were present and made themselves available for testing. 5

• Data collection for study three was restricted to those triathletes competing in the age group selection race for the 2000 TWC held in Perth, Western Australia. • Studies four, five and six data were delimited to male and female senior elite triathletes competing in the 2000 TWC. • The results only pertain to “Classic” distance triathlons and are not intended to show differences across the different sprint or ironman distances.

Operational Definitions 1. Event distances Classic distance triathlon refers to an event which comprises a 1.5 km swim, 40 km cycle and a 10 km run. It has also been known as the Triathlon Distance, Standard Distance or Olympic Distance.

Sprint Distance events are typically 750 m swim, 20 km cycle and 5 km run; whereas Ironman distance triathlons consist of a 3.8 km swim, 180 km cycle and 42.2 km run.

2. Drafting “Drafting” is currently a controversial rule in triathlon. Drafting is where one cyclist is permitted to position him/herself as close behind another rider as possible in order to decrease the wind resistance. It has been shown that drafting can decrease energy expenditure by up to 30% (Faria, 1992), thereby leaving a larger reserve to use at other times in the race such as during the run (Lehénaff et al., 1998), or to increase speed on the bike. Most of the international classic distance races, including the world championships, are now “draft legal”, but for the elite category only. The age-group competitors do not normally have this strategy at their disposal. The 1997 TWC was the first race where drafting was legal for the junior elite competitors.

3. Body size and composition The Phantom Stratagem is a technique used to analyse proportionality by removing size and gender differences by scaling (Ross & Ward, 1986).

Five-way fractionation subdivides the body mass into five compartments (Kerr & Ross, 1994; Ross & Kerr, 1991). These compartments are skin mass, bone mass, muscle mass, adipose mass and residual mass. Residual mass relates to what mass remains, it is not the result of subtracting the other four compartments from total mass, but has its own value and is related to internal organ masses.

4. Cadence Cadence is the term used throughout this thesis to describe leg oscillation rates in all three disciplines of triathlon. It is calculated as the number of repetitions per minute of a complete cycle of one of the legs. For example, in swimming, each time the right leg completes a cycle from top to bottom and back to the top; in cycling, one complete revolution of the pedals through 360 degrees is one cycle; and, during running, a revolution is deemed as successive foot strikes of the same foot.

5. Transition Triathlon combines the three separate disciplines of swimming, cycling and running. Changing from one activity to the next is termed the transition. The sport of triathlon should be viewed as a whole, including transitions (Millet & Vleck, 2000).

6. Cycling Postures There are three cycling postures typically used by cyclists and triathletes depending on terrain, environmental conditions, race format and rules. The "upright" position relates to the cyclists placing their hands on top of the brake hoods on the handle bars. The "drops" refers to the lower curve of the handle bars and the "aero" position requires the use of bars which tend to put the cyclists in a more streamlined position (Figure 1.1). 7

(a) Drops (b) Hoods (c) Aero

Figure 1.1 Different hand positions while cycling; upright (hoods), drops and aero.

7. Spinning Spinning relates to triathletes increasing their cadence such that there is usually a decrease in the force required for each pedal stroke. That is their legs are ‘spinning’ faster. This is usually used prior to the second transition in preparation for running.

Chapter 2

Literature Review

This literature review has been separated into two sections for clarity. The first area of review is the effect body morphology plays on triathlon performance. Morphological data relating to elite, recreational and female triathletes are presented here. The second section introduces the role cycling cadence influences subsequent running performance during triathlon.

Morphology and Performance Triathletes have been found to be long limbed and have low levels of body fat. (Ackland et al., 1998a). It has been suggested that triathletes will have a morphology which resembles the average of the three disciplines (Travill et al., 1994). Leake and Carter (1991) compared female triathletes with athletes from the individual sports which comprise triathlon and found they tended to have a similar shape to that of cyclists.

However, with the changing nature of the sport of triathlon and the inclusion of a draft legal cycle leg, it could be hypothesised that the triathletes will have a physical structure more similar to runners and swimmers than cyclists. This is due to a decrease in the importance of this discipline to the overall outcome of a race. That is, during the cycling section, triathletes are permitted to ride as close together as they wish to create a slip-stream and conserve energy. This has increased the importance of the run and swim where drafting plays less of a role.

A study of world championship swimmers (Carter & Ackland, 1994) revealed a relationship between body morphology and performance. Other studies also have shown that body composition is related to endurance performance (Tittle & Wutscherk, 1988; Woolford et al., 1993) including both running (Wilmore & Brown, 1974) and cycling (Olds et al., 1995). Tittle and Wutscherk (1988) described the critical anatomical features of runners as being thigh and leg length, and Bloomfield and 9

Sigerseth (1965) noted the importance of longer levers (arms) for greater swimming speed. This has been supported by Ackland et al. (1998b) where limb lengths correlated positively with triathlon performance.

Anthropometric data from small samples of triathletes were collected at the Hawaiian Ironman, an ultra endurance event lasting in excess of eight hours (Holly et al., 1986; O’Toole et al., 1987; O’Toole et al., 1989), or other ultra endurance events (Dengel et al., 1989; Khort et al., 1987; Khort et al., 1989). These studies were also conducted in the early 1980’s during the formative years of the sport. The athlete undertaking such an event over 8 h was possibly different from one with a primary focus on classic events lasting around two hours.

Other studies have focused on female triathletes (Leake & Carter, 1991; Sleivert & Wenger, 1993) but elite females were not included. Laurenson et al. (1993) compared elite with club level females over the classic triathlon distance. Others have studied well-trained classic distance triathletes (De Vito et al., 1995; Hue et al., 1998; Rowbottom et al., 1997; Schneider et al., 1990; Sleivert & Rowlands, 1996; Zinkgraf et al., 1986), or recreational athletes only (Sleivert & Wenger, 1993). These studies have investigated the basic measurements of height and mass to characterise their subjects, and will enable only limited comparisons with the more comprehensive kinanthropometry of the present study.

The data obtained when triathlon was in its infant stages may be redundant, as it is possible that a “standard” body type may have evolved since then. Such changes were noted by Carter and Heath (1990) when they compared swimmers and runners via a photoscopic procedure to determine somatotypes over a 60-year period. Stepnicka (1986) also commented that the somatotypes of athletes today are dissimilar to those of 8-15 years ago. Triathlon participants in half ironman (1.9:90.3:21.1) distance showed differences in height, mass and age from ultra endurance and classic distance triathletes (Khort et al., 1989). De Vito et al. (1995) noted that variation in the length of a triathlon places different physiological demands on the body and that the energy sources (anaerobic v aerobic) vary according to the length of the event. This highlighted the need to characterise triathletes by the distance over which they train and compete. There is a morphological difference between 100 m sprinters and marathon runners 10

(Bloomfield et al., 1994) and it would be expected that differences exist between triathletes who compete for two hours versus eight hours.

Travill et al. (1994) measured 18 of the 70 highest ranked South African triathletes. The measurements included a restricted profile of 20 anthropometric measurements to enable the determination of somatotype. The triathletes were males of mean age 23.1 y who were younger than present senior elite triathletes (27.5 y) (Ackland et al., 1998a). A comparison was made with seven professional triathletes from the United States who were measured by Dolan (1987). However, at this time, the sport only had existed for about 10 years, drafting was not legal and the aero cycling position (Figure 1.1) was only just introduced. It might be imprecise to compare triathletes of 1985 with 1993 and, even more so with World Championship triathletes of 1997. However, both the amount of run training and the sum of six skinfolds were found to be related to triathlon performance (Travill et al., 1994).

Well trained/elite triathletes (Classic Distance) A study of physical and physiological factors associated with success in triathlon found triathletes to be tall, average to light mass and have low levels of body fat relative to the wider community (Sleivert & Rowlands, 1996). The authors considered that this benefited a triathlete by having large leverage and optimal power to body surface area or body mass ratios. Triathletes were closer in height and mass to cyclists (distance of training or racing was not stated), taller and heavier than 5 km-marathon runners, and shorter than swimmers that competed over 800/1500 m (Sleivert & Rowlands, 1996). The values were similar to those stated in previous studies. Body composition was not determined in this study.

De Vito et al. (1995) measured and reported age, mass and height of six well-trained, male triathletes (Appendix C) to determine a decrease in endurance performance during the classic distance triathlon. They found that race times over a reported classic distance triathlon were only slightly slower than those achieved by elite athletes for a single event over the same distance. However, this might not have been a classic distance triathlon as the cycle leg was only 32 km (normally 40 km). The 1500 m mean swim time of 29 min 30 s was also slow when compared with that achieved by elite 11 triathletes in open water. The authors concluded that performance of a 10 km run decreases when it follows a swim and a cycle, as opposed to a straight 10 km run.

Hue et al. (1998) and Guezennec et al. (1996) showed an increase in the oxygen cost of running following a swim and a cycle, and hypothesised that it was due to physiological and/or biomechanical causes. The triathletes in these studies were highly trained, with seven subjects from Hue et al. (1998) having been in the French national champion team for the previous four years. Again, age, height and mass characteristics were reported for each triathlete. Ten triathletes were measured by Guezennec et al. (1996) with mean age, height and mass the only anthropometric measurements in a study which focused on the energy cost of running. The anthropometric variables are presented in Appendix C.

The eight male triathletes studied by Rowbottom et al. (1997) recorded a mean personal best time for the classic distance triathlon of 118 min, which is comparable with elite triathletes. These subjects competed at state and national level. This study investigated biological changes in triathletes but only minimal anthropometric data reported. Mean age (29.6 y), height (180.0 cm), mass (73.9 kg) and body fat (7.9%) were recorded.

Schneider et al. (1990) and Zinkgraf et al. (1986) collected basic anthropometric data of height, mass and age on highly trained and elite triathletes (Appendix C). Schneider et al. (1990) examined ventilatory threshold and O2max during cycling and running. They concluded that O2max of triathletes was equal to that of single sport athletes in cycling and running, and that triathletes were comparable with the single sport athletes physiologically.

A study by Lehénaff et al. (1998) reported age and height of eight elite triathletes (Appendix C). This study examined the effects of drafting during the cycle leg of a triathlon on the subsequent run. It was concluded that drafting improved run performance.

The somatotypes of 21 short course triathletes were investigated by Formby (1989) to determine the relationship with performance. Twelve finished in the first 26% (classed 12 as elite) and the other nine from the last 26% of finishers (non-elite). No data were provided other than summaries of significant findings. It was reported that the mean age, height, mass and sum of skinfolds did not differ significantly. However, 67% of the “elite” competitors were classified as balanced mesomorphs, whereas only 33% of “non-elite” were in this classification.

Recreational triathletes Sleivert and Wenger’s study (1993) measured mean age, mass and height (Appendix C) for recreational triathletes. They sought to determine the physiological predictors of short-course triathlon success. The distance covered was a 1 km swim, 30 km cycle and 8 km run. The subjects were 18 male and seven female beginner triathletes. They concluded that the O2max values were less than those reported for triathletes competing in ultra endurance triathlons, but similar to the values reported for triathletes participating in events of comparable length. Again, differences were noted between the ultra endurance and classic distance triathletes. But it could lend support to the assumption that, if consistent physiological variables are reported for classic distance events, there might also be a preferential phenotype.

The mean O2max values were less than the means reported for athletes competing in the individual disciplines of swimming, cycling and running. This is in agreement with other studies reviewed by Sleivert and Rowlands (1996).

Deitrick (1991) examined the physiological effect that an increase in body mass would have on the cycling and running performances of 14 male, recreational triathletes. Seven were considered to be ‘typical’ triathletes in terms of morphology (height of 176.6 cm and mass of 66.6 kg), while the other seven were deemed ‘heavy’ for triathletes, being greater than 90 kg. Results in this short course triathlon showed that the lighter triathletes performed better, recorded a significantly lower sum of skinfolds and were significantly shorter in stature.

Female triathletes Leake and Carter (1991) compared the body types of female triathletes (n=6) with Olympic swimmers and runners, and used the classic distance event (1.5 km swim, 40 km cycle & 10 km run) as the race for which the triathletes were training. These triathletes were competitive but not elite. The study compared their somatotypes and skinfolds with swimmers from the Montreal (1976) Olympics, and runners from the Munich (1972) Olympics. The swimmers had competed in a variety of events and the runners had competed in the 800 m and 1500 m events (The longest Olympic running event for women prior to 1984 was 1500 m (Drinkwater, 1986)). It should also be noted that, as yet, women do not compete in swimming events greater than 800 m in Olympic competition. Hence, the comparisons made were between different events from those of swimming and running subjects. Also, the swimmers and runners had competed 17 to 22 years before the triathlon measurements were made. However, the mean somatotype of the female triathletes was 3.1-4.3-2.6, or balanced mesomorphy. Triathletes were less ectomorphic, more mesomorphic, and recorded greater skinfolds at all of the six sites than did the middle distance runners. Because the triathletes were also found to be heavier than the runners but lighter than the swimmers, the authors suggested this could be due to a cross-training effect. No significant differences were found between the somatotype and skinfold patterning of the swimmers and triathletes. The triathletes were shorter and lighter than the swimmers, and taller and heavier than the runners.

A comparison of the top ten finishers (n=5) with the rest (n=10) revealed no significant differences in somatotype with 2.5-4.2-2.8 for the best and 3.4-4.2-2.6 for the rest (Leake & Carter, 1991). Again, it should be noted that the subjects were not elite. When elite and club level female triathletes were compared by Laurenson et al. (1993), the sum of four skinfolds showed a significant difference, with the higher performers having lower total scores. Comparisons of age, mass and height of female triathletes (Laurenson et al., 1993; Leake & Carter, 1991; Sleivert & Wenger, 1993) showed that there are similarities (Appendix C).

Overview Generally, triathletes have more closely resembled cyclists, while being taller and heavier than runners, and shorter and lighter than swimmers (O’Toole et al., 1987 (long 14 distance & ironman); Sleivert & Rowlands, 1996 (classic)). Possibly, cyclists are an average of runners and swimmers in terms of height and mass or, as previous research suggests, the cycle discipline’s importance is greater during non-drafting events. Triathletes also have low levels of body fat as measured by hydrostatic weighing or skinfold calipers. Somatotype should be taken into consideration when determining morphological similarities of triathletes to athletes in each of the single discipline sports. Only Leake and Carter (1991) have studied somatotypes, and these were of non- elite, female triathletes.

Other indices such as limb lengths and proportions of elite triathletes have not been determined or compared with athletes in the individual disciplines of swimming, cycling and running.

General principles - Male and female differences Hip width is approximately the same in both males and females. However, the shoulders and thoracic cavity grow more rapidly in men. Therefore, men tend to have proportionately wider shoulder-to-hip measures than women in a normal population (Carter, 1982; Sloane, 1980). Females have relatively longer trunks when compared with males, and this is largely determined by relative sitting height (RSH) (Norton et al., 1996).

In the general population, greater muscle mass in males and higher percentage of body fat in females is attributed to the differing levels of hormones in each gender. Higher levels of testosterone are found in males and greater oestrogen levels in females.

The difference in somatotype between males and females can be shown graphically with the lines connecting male and female somatoplots generally being at right angles to the ectomorph axis (Carter & Heath, 1990). Complementing this, Bloomfield et al. (1994) summarised data from a large number of high level athletes in various sports. Female athletes were generally more endomorphic and less mesomorphic, but had similar ectomorphic ratings as their male counterparts.

Cadence The cycle section of a triathlon has been termed by some coaches as "just a means of getting from the swim to the run" and others have referred to triathlon as a "wet run". If this is so, the importance of being in the first pack out of the water and, therefore, in the first pack during the cycle; and be a good runner, has increased. This importance will be examined further through the course of the present studies.

Numerous studies (De Vito et al., 1995; Guezennec et al., 1996; Hausswirth et al., 1997; Hue et al., 1998; Kreider et al., 1988 and Lehénaff et al., 1998) have shown that, after cycling, there is a decrease in efficiency; or greater oxygen cost of running following cycling when compared with just running. Hue et al. (1998) and Lehénaff et al. (1998) also have shown that, if drafting is permitted, there is less decrement in run performance. Drafting can reduce the energy cost of cycling by up to 30% (Faria, 1992). Therefore, energy is conserved and this enables a faster run, an increase in cycle speed, or a combination of both. This is in contrast with a 6.5% decrease in energy cost due to drafting during running at 4.5 m/s (Pugh, 1971). In swimming, a 3.2% increase in performance was noted over 400 m while drafting (Chatard et al., 1998).

It is generally more difficult to change running biomechanics to improve performance as it is a skill that has been developed from childhood. Cavanagh and Williams (1982) have shown that, when running, the average difference between optimal and freely chosen stride length is 4.5% of lower limb length, or within 4.2 cm of the SL to minimise energy cost. This is in agreement with other research (Holt et al., 1991). Stroke rate - velocity curves of swimmers who had not competed for a number of years are the same as those produced by competitive swimmers. Hence, skill, as opposed to general fitness, is a factor in the selection of stroke rates at certain velocities (Craig & Pendergast, 1981). Cyclists tend to modify gear selection throughout training and racing, depending on the environmental conditions or terrain, in order to maintain a consistent cadence. Thus, because it is a machine, it is easier to modify the bike set-up or mechanics without a large decrement in performance, as opposed to refining running technique.

Hence, it would be valuable to try and develop a cycling technique, position or training modification which might reduce cycling efficiency, but will reduce pre-running fatigue and allow greater ability to run "off the bike" and, hence, improve overall performance. For example, changing the leg oscillation speed, muscle force generation and the interaction with SL of the triathlete perhaps could be altered in order to improve run efficiency. The following are a number of areas related to cadence which were to be examined in the literature and the present studies: • Self-selected cadence in sport • Biomechanical interaction of cycling and running • Anthropometry and cadence • Cadence and blood lactate (HLa) levels • Cadence and electromyography (EMG) • Variation of cadence prior to changing disciplines • Cadence and hamstring flexibility • Cadence and Ground Reaction Force (GRF)

These are discussed below.

Self-selected cadence Self-selected cadence occurs by chance and could be a naturally successful or unsuccessful choice. Also, it could be related to a natural frequency of the body, muscle firing pattern, or muscle contraction efficiency. It appears that elite athletes choose similar leg frequencies in a variety of individual sports, irrespective of limb lengths, fibre type, O2 or anaerobic threshold.

It has been shown previously (Gotshall et al., 1996; Hagan et al., 1992; Marsh & Martin, 1993 & 1997) that cycling efficiency is greatest, in terms of lowest oxygen cost, at cadences of between 50 and 60 rpm. These cadences are lower than those freely selected. This also has been shown not to change with a variety of workloads from 75 – 200 W. However, these test sessions were relatively short (5-15 min).

Gotshall et al. (1996) focused on the hemodynamic changes that occur as a result of different cadences used while cycling. Seven healthy men cycled at 200 W using a 17 cadence of 70, 90 or 110 rpm. Oxygen cost increased with cadence, as did heart rate (HR), stroke volume (SV), cardiac output (Q) and blood pressure (BP). However, vascular resistance decreased and it was found that the skeletal muscle pump was more effective at the higher cadences.

Hagan et al. (1992) analysed cardiorespiratory variables of five fit and healthy young men (27 y) when cycling at each of two work loads (127 & 166 W), and at two cadences (60 & 90 rpm) for 45 minutes. Cycling at 90 rpm at both workloads produced higher HR, ventilation and oxygen cost when compared with the same workload at 60 rpm.

Marsh and Martin (1993) studied the preferred cadence of eight cyclists and eight non- cyclists who were of equal fitness, and compared the difference between the self- selected cadence and the most oxygen efficient cadence. Again, the workload was set at 200 W and subjects cycled at 50, 65, 80, 95 and 110 rpm for 4-6 min. The preferred cadences for both the cyclists (85.2 rpm) and non-cyclists (91.6 rpm) were significantly greater than those which utilised the least oxygen (56.1 & 62.9 rpm, respectively).

A follow-up study compared cyclists (n=12), runners (n=10) and untrained cyclists (n=10) in terms of their preferred and most economical cadence (50-110 rpm) at different work loads (75-250 W) (Marsh & Martin, 1997). Again, the preferred cadence was found to be significantly greater than the most economical cadence for all groups and at all work loads.

Elite cyclists choose a higher cadence than non-cyclists (90 rpm vs 60 rpm) (Takaishi et al., 1996 & 1998) and it is generally considered to be a result of cycling skill or experience. However, during a similar study Marsh & Martin (1993), used high class runners as non-cyclists in order to maintain similar O2max values between groups (67.8 ml·kg-1·min-1 vs 66.2 ml·kg-1·min-1). They found that the runners chose a similar cadence to that of the cyclists during cycling at 200 W. This suggested that cycling skill might not be the most important factor and that the fitness level also could be a contributor. Fitness level also was offered as an important factor by Marsh and Martin (1997) when comparing cyclists, runners and less trained non-cyclists in a follow-up 18 study. They suggested that the history of running training was similar to that of cycling (high repetitions, relatively low forces and relatively fast joint angle velocities), and resulted in similar force-velocity properties of the lower limb muscles.

Faulkner et al. (1971) also showed that, when testing cardiovascular responses to cycling and running, the cycling tests were performed with a cadence of 80 rpm to approximate the 87 strides/minute taken with each leg when running at 8 mph (12.8 km/h).

Takaishi et al. (1998) selected non-cyclists (n=7) from team sports to try to nullify the possible lower limb training effect and compared cycling cadence between equally fit groups. Subjects pedalled at 45, 60, 75, 90 and 105 rpm for 10 min. He found that cyclists (n=7) had greater pedalling skill and chose higher cadences (75 or 90 rpm) than non-cyclists (60 or 75 rpm).

Most data presented on SR and SL of running have shown that they change with speed. First, there is an increase in SL, and this is followed by an increase in SR (Williams, 1985). The point at which each increases is not understood clearly, but it appears to be related to optimal efficiency. Endurance athletes are looking for an optimal SR/SL relationship that allows greatest running speeds for least effort. Hay (1993), has shown that SL is related to lower limb length (LLL) or stature.

Generally, swimming data are presented in terms of stroke rate and stroke length (Craig & Pendergast, 1981; Hay, 1993). However, analysis of two middle distance swimmers suggests that swimmers also choose a leg cadence speed (for kicking) similar to that of cyclists (90 rpm). e.g. 44 strokes/50 m/30 s, with a 4 beat kick ~ 88 kicks per minute 50 strokes/50 m/50 s, with a 6 beat kick ~ 90 kicks per minute van der Woude et al. (1989) found that wheel chair propulsion appeared to have an optimum arm cycle frequency which was closely related to the freely chosen frequency (no significant difference) at any given speed. Subjects were split into two groups of six based on their wheelchair experience. Frequency was found to increase (0.67 - 1.03 19

Hz) with increasing velocity (0.55 - 1.39 m/s) and the results were similar for both groups. These results are also similar to swimming speeds and arm stroke rates.

Marais et al. (1999) investigated freely chosen cadences in upper body exercise in eight well trained competitive kayakers. Subjects performed arm cranking exercise at various workloads (50 - 120% max power) and at three cadences; freely chosen and +/- 10% of freely chosen. Marais et al. (1999) also found that spontaneously chosen rates were most efficient and that increased power demands (50 - 80% max power) resulted in an increased cadence (58 - 67 rpm).

Millet et al. (1997) found significant correlations between velocity and cadence in 15 cross-country skiers when tested on roller skis over two kilometres at a range of 37 rpm at 2.2 m/s, to 50 rpm at 3.6 m/s. They suggested that this might be the result of increased resistive force due to the uphill surface and a different surface friction than snow.

Sprint swimmers tend to have greater distance per stroke and proportionately lower SR when compared with distance swimmers. The slower SR requires increased muscle force per stroke which leads to undue muscle fatigue (Craig & Pendergast, 1981). Hay (1993) reported that, as swimming distance increased, stroke frequency decreased and SL increased, except in butterfly. He also showed that the male SL was longer than that of the females but there was a similar frequency. Because male swimmers have longer limbs than females, the difference in velocity between the two genders could be attributed to limb length. This related directly to SL, and perhaps is not influenced much by SR (Carter & Ackland, 1994). However, it is questionable as to whether this is a natural frequency. If muscle strength is lacking, swimming technique can be altered by flexing the forearm to produce a more effective lever system. This shortens the upper limb and highlights the need for strong musculature to use an increased oscillation rate with increased lever lengths (Bloomfield et al., 1994).

It could be concluded from the above studies that athletes in a variety of sports spontaneously choose the most efficient rate at which to perform a certain task.

Biomechanical interaction of cycling and running Research generally has shown that, with increasing fatigue during running, there is a decrease in SL and, if velocity is to be maintained, an increase in SR is required (Ackland, 1997; Elliott et al., 1981; Elliott and Ackland, 1980; Elliott and Roberts, 1980; Hausswirth et al., 1997, Williams et al., 1991). This also has been found with running mechanics following cycling (Hausswirth et al., 1996 & 1997, Vercruyssen et al., 2002). These results disagree with the findings of Quigley and Richards (1996) and Hue et al. (1998) who found no change in running biomechanics with prior cycling. Also, Witt (1993) found a decrease in running SR immediately following cycling, but it then stabilised after 1000 m. Elliott et al. (1981) showed that, with either continuous or intermittent running training, SL/SR can be maintained.

The effect of cycle posture also has been investigated to determine the metabolic efficiency while cycling but only one study has investigated subsequent run performance (Garside & Doran, 2000). Eight male triathletes performed a 40 km cycle o o at 70% O2peak on a bike set-up with either a 73 or 81 seat tube angle (STA). This was followed by a timed 10 km run on a treadmill. Results indicated that the triathletes ran faster after cycling on the steeper STA and this difference was mostly accounted for during the initial stages of the run.

Heil et al. (1995) compared the cardiorespiratory responses of 25 competitive triathletes while cycling only, at various STAs (69o, 76o, 83o, 90o). Subjects cycled for 10 minutes at 73% O2peak at 90 rpm and were in the aero position. Lower HR and O2 values were recorded during the steeper STA trials (83o & 90o) when compared with 69o STA. Greater hip extension and ankle plantar flexion were reported when using a steeper STA.

Price and Donne (1997) continued this investigation of STA but also included the variable of seat height. Fourteen competitive cyclists cycled at 200 W for 4 min at 85 rpm for each of the nine trials. The trials included three seat heights; 96%, 100% and o o o 104% of trochanteric height, and three STAs; 68 , 74 and 80 . The O2 was found to be lowest, and power efficiency highest, at a STA of 80o at all heights. Measured HR was lower at a seat height equal to 100% trochanteric height. 21

Faria et al. (1978) showed an increase in maximum oxygen consumption during maximal cycling exercise tests when using the drop position as opposed to the traditional upright posture. Nine subjects were used (8 male & 1 female), cadence was maintained at 60 rpm and workload was increased by 180 kpm/min every 2 min. Faria et al. (1978) suggested that the increase was due to the greater amount of muscle mass required during exercise to hold the upper body in position. Conversely, Origenes IV et al. (1993) found no differences in ventilatory parameters or maximal oxygen consumption when conducting two O2max tests in either the aero or upright position. Here, the investigators increased the work load 50 W every 3 min and analysed metabolic gas exchange in 10 moderately trained males.

Two studies which investigated oxygen cost during submaximal exercise (Berry et al., 1994; Ryschon & Stray-Gundersen, 1991) found no differences for any of the respiratory variables, HR or energy cost while cycling in either the aero, drop or upright posture. Berry et al. (1994) compared the aero and upright positions using 11 cyclists, six of whom had previously used aero bars. Subjects cycled at 80% O2max for 60 min before increasing 5% O2 every 15 min until exhaustion. It should be noted that four subjects did not complete the first 60 min and times ranged from 15 to 95 min. Also, although not reported, it appears that a significant difference could exist between the two conditions for those cyclists without aerobar familiarity. Ryschon and Stray- Gundersen (1991) compared four body positions for 10 subjects (8 males & 2 females); hands on hoods (upright) at 80 and 60 rpm; hands on drops at 80 rpm; and standing at 60 rpm. The workload was set via a treadmill at 19.3 km/h on a 4% grade which equated to approximately 50% O2max. No differences were found between the energy expenditure while cycling in either the upright position or with hands on the drops. However, a greater energy cost was noted while standing when compared with sitting.

In contrast, Gnehm et al. (1997) showed an increased HR, O2 and respiratory exchange ratio when cycling in the aero position during submaximal cycling. Fourteen elite cyclists cycled at 70% O2max in each of the three different positions (aero, hoods, drops) for five minutes in a continuous, but random order. A cadence of 90 rpm was used. 22

Finally, Sheel et al. (1996) conducted an investigation on a velodrome at a constant speed, where air resistance and drag affects efficiency. Results showed a significantly lower O2, HR and ventilatory parameters when using the aero position when compared with either the upright or drops posture (Sheel et al., 1996).

Previously, improved performance via aerobars was sought via reduced frontal area by bringing the arms in closer to the body and lowering the torso by flattening the back to reduce drag. Since 1995 for senior elite, and 1997 for junior elite, the cycle portion of a triathlon has been draft legal. This has led to bicycle specification changes in order to ensure safety while cycling (International Triathlon Union Competition Rules, rule E.3 Cycling Conduct, 1999). The aerobars are shortened which reduce the amount of forward lean of the triathlete when in the aero position. Drafting also has meant that triathletes tend to spend more time in the upright position than previously, as drag is reduced by riding in a group and handling skills are improved in this posture.

Also, during this time there has been a decrease in 10 km run time "off the bike". This is considered mostly to be due to the reduced effort created by drafting (De Vito et al., 1995; Guezennec et al., 1996; Hausswirth et al., 1999b, c; Hue et al., 1998; Kreider et al., 1988 and Lehénaff et al., 1998). However, it is possible that the cause might be related to other indirect changes (Berry et al., 1994; Hausswirth et al., 1997).

De Vito et al. (1995) set out to quantify the endurance impairment of running after cycling in six male Olympic distance triathletes. Ventilatory threshold and O2peak during running were measured via an incremental treadmill test either fresh or after a 1.5 km swim and 32 km cycle. Both measures were significantly reduced after the swim and cycle.

Guezennec et al. (1996) determined the increase in energy cost associated with running at the end of a triathlon. The study included testing 11 male triathletes over a classic distance triathlon and then a 10km run at the same pace used during the triathlon, but with no prior exercise. The results revealed higher O2, e and HR during the triathlon run when compared with the control run. Body mass and plasma volume decreased by 23 a greater amount during the triathlon run. No differences in blood lactate were reported at the end of the runs. It was concluded that there was a decreased running efficiency at the end of a classic distance triathlon.

Hausswirth et al. (1999b) compared the 5 km running performance of 8 male triathletes after swimming 750 m and cycling for 20 km, either alone or in a full drafting situation. Cycle speed was constant in both conditions. There were significantly lower values for

e, O2, HR and blood lactate (HLa) when cycling in a drafting position. Running speed also was significantly faster after drafting during the cycle.

Hausswirth et al. (1999c) further investigated the drafting concept on running performance utilising 10 male triathletes, all of whom completed a classic distance event in less than 1 h 55 min. The subjects completed two sprint distance triathlons on different days. In the first session, the triathlete rode for 500 m in front of another cyclist and then 500 m behind the same cyclist in a drafting situation throughout a 20 km ride. In the second session, the triathletes cycled at the same speed but in a drafting position for the entire 20 km. Significantly faster runs were recorded after drafting for the entire ride. Oxygen cost was higher during the cycle section when alternating the lead position, thereby indicating a greater energy demand during this format.

Seven competitive male triathletes took part in four testing sessions to analyse the effects of prior cycling (40 km) on the biomechanical and cardiorespiratory response during running (10 km) (Hue et al., 1998). The study required four testing sessions; a maximal cycling test, a maximal running test, a 10 km run following a 40 km cycle, and a 10 km control run. Results highlighted the increased oxygen cost of running following cycling. No change in SL or SR measures were reported between the two run conditions.

Nine male triathletes performed a simulated 75 min (40 km) control bike test and a 40 min (10 km) control run at 70% of maximal oxygen uptake (Kreider et al., 1988). These control data were compared with data derived from a simulated triathlon. Results demonstrated that prior swimming significantly decreased triathlon cycling work output 24

and, as running was completed at the same workload, significant increases in O2, e,

HR, a- O2diff and core temperature were noted.

Lehénaff et al. (1998) examined eight elite male triathletes in three outdoor testing sessions. The first session was a maximal 5 km run. The second and third sessions involved completing a sprint distance triathlon where drafting was permitted and then where drafting was illegal. Results showed that cycling in a sheltered position during the entire bike leg demanded significantly less energy expenditure which, in turn, was available for a significantly faster run. The control run was still significantly faster than either of the other two runs performed after cycling.

Hausswirth et al. (1997) showed that, after cycling, a significant increase in trunk flexion angle was noted at foot strike (3.5o - 2.1o) when compared with just running, and running in a fatigued state at the end of a marathon (0o). Average trunk flexion during the first five minutes of running also was different from that of just running and running at the end of a marathon (6.6o vs 10.3o). Berry et al. (1994) showed a significant difference in mean hip (91.3o vs 126.3o) and shoulder (100.6o vs 118.0o) angles when cycling in the aero and upright postures, respectively. Garside and Doran (2000), as reported earlier, showed an increase in running speed after cycling with a steeper seat tube angle (81o) compared with the norm (73o). It may be possible that, with the changing nature of the cycle leg during triathlon, cycling posture changes could lead to improved run performances.

Vercruyssen et al. (2002) reported an increased O2, e and SR in eight male triathletes during running following cycling at various cadences when compared to an isolated run. The results also indicated greater oxygen cost while pedalling at higher cadences.

The above studies summerised the current level of knowledge of the interaction between cycling and running in triathlon. Cycling efficiency can be highly influenced by altering cycling posture or bike setup (eg STA). Fatigue has also been shown to alter run mechanics and that fatigue elicited by prior cycling can reproduce these mechanical changes. If drafting is permitted during the cycle, oxygen cost is reduced and less influence on running performance is noted. 25

Anthropometry relationship with cadence A strong relationship was found between lever lengths and swim performance during a triathlon, and anthropometry accounted for over 50% of the variance of total triathlon performance (Ackland et al., 1999; Landers, 1998; Landers et al., 1999). Others have shown a relationship between lower limb length (LLL) and SL during running, but usually at higher speeds (Hay, 1993). The thigh length has a negative correlation with limb frequency during running (Tittle & Wutscherk, 1988). That is, there is a high crural index (CI), which is the ratio between leg length and thigh length (Appendix E). A high CI is related to achieving higher stride frequencies with a similar SL, thus greater speed for the same effort. Studies relating cycling cadence to anthropometry measures have shown equivocal results. Gonzalez and Hull (1989) suggested that optimal cadence decreases with cyclist size. There also appear to be economy relationships between LLL and seat height (Price & Donne, 1997) which require greater clarification.

Increased running speed results from an increase in both SL and SR. Initially, the most rapid increase is in SL, which then plateaus and then SR increases once SL is maximised (Elliott & Blanksby, 1976; Williams, 1985). Therefore, at lower speeds, the relationship between SL and LLL may not be as strong. Cavanagh and Williams (1982) showed no relationship between SL and LLL during moderate running of 3.83 m/s (13.8 km/h).

Knuttgen (1961) investigated O2 and HR with determined and undetermined SL at different speeds. It was found that the two male subjects could not maintain speeds higher than 11.6 km/h with a SL of 77 cm because of the inability of the body parts to move fast enough. At this speed, a cadence of 126 strides per minute would be needed. This is greater than 2 strides per second, nearly 4.5 steps per second and approaching sprinter rates.

Chapman and Caldwell (1983) examined the kinetic limitations of maximal sprinting and found that SR was limited, and therefore sprinting speed, by the ability of the knee to absorb the energy of the leg toward the end of the swing phase.

However, size might not have as great an influence if shape is similar (Dern et al., 1947; Wilkie, 1950). Wilkie (1950) found that, with a singular contraction of biceps brachi, although maximal force varied (12.5 - 20 megadynes), the maximal velocity was relatively constant across all subjects. He suggested that it followed similar lines of geometrical similarity in animals and noted the different sizes, but similar shapes, of greyhounds and race-horses which generate similar running speeds. In contrast, Taylor (1985) noted the strong relationship between stride frequency and whole body mass in bipedal hopping, quadruped galloping and human hopping. This relationship suggests that the entire body mass determines SR, and that there is a frequency where muscle- tendon spring efficiency is optimised via minimum metabolic cost and maximal mechanical advantage through the cyclic action. Taylor (1985) also noted that animals (bipeds and quadrupeds) maintain the same frequency across a two to three fold increase in speed.

The research is equivocal as to the relationship between economy of motion and SR/SL. It is still not clear if taller/long limbed athletes choose a lower cadence when at race pace, or choose the same cadence and longer strides, and therefore go faster (or shorter strides/strokes and go the same speed).

Relationship Between Cadence and Blood Lactate

Blood lactate levels have been shown to be related to O2 levels (Cox et al., 1994; Ito et al., 1983). As exercise intensity increases, O2 increases as well as the level of HLa produced. During a race, triathletes are competing in an endurance event where they try to maintain blood lactate levels below threshold. Thus, it is hypothesised that there will be a relationship between cadence and Lactate Threshold (LT), and possibly Individual Anaerobic Threshold (IAT).

Cox et al. (1994) evaluated the effect of pedal frequency on the 4 mM LT of elite cyclists. Seven male cyclists completed three ramped protocols to exhaustion using three pedal cadences (50, 70 & 90 rpm). The results showed a significant decrease in work rate at which 4 mM LT was reached. However, there were no significant differences in the peak watts obtained at exhaustion. This suggested that, although 4 mM LT was reached approximately one minute earlier at 90 rpm (vs 50 rpm), it was 27 still possible to perform at the same maximum workload and for the same total time. It could be that 4 mM is not actually threshold, as these cyclists are highly trained.

Gotshall et al. (1996) investigated the skeletal muscle pump being affected by varying cycle cadences (70, 90 & 100 rpm) at 200 W. Seven cyclists pedalled for five minutes at each of the cadences. The HR, SV, and BP were increased, and vascular resistance decreased with increasing cadence. This suggested an improved pump action by skeletal muscles. Gotshall et al. (1996) suggested that the results obtained by Cox et al. (1994) could be the result of the enhanced muscle blood flow which is associated with higher cadence. This would result in “washing out” the lactate faster from the muscle into the blood stream and cause higher readings.

Prior to these studies, Hagberg et al. (1981) had proposed a similar rationale for muscle blood flow and blood lactate measures. Hagan et al. (1992) found no significant differences between blood lactate levels at pedal frequencies of 60 and 90 rpm at either 127 or 166 W. In their study, five trained males cycled continuously for 45 min to complete the above workloads and frequencies in a random order.

Ito et al. (1983) examined the role of elastic potentiation on mechanical efficiency in three male runners at speeds between 7 and 22 km/h. Kinematic and mechanical energy were analysed via film recordings, and energy expenditure was determined from expired air. No relationship was noted between height and SL, and no data were presented on LLL. By revisiting the results (Table 2.1), it was possible to determine that onset of blood lactate (OBLA) occurred at a SR of approximately 90 strides per minute. The variation in speed obtainable in the three subjects at OBLA was affected by SL and not SR. Hence, a run cadence of approximately 90 strides per minute could be the most effective for constant speed, endurance running.

Table 2.1 Results from Ito et al. (1983) for each subject at OBLA. Subject Speed (km/h) SL (m) SR (/min) Height (cm) A 15.0 1.375 91 182.9 B 15.5 1.443 90 166.9 C 18.5 1.644 94 167.0

It could be that a triathlete chooses a cadence that allows greater speeds when pushing towards IAT, above LT, which may not be as efficient in terms of O2 consumption. Perhaps they may choose a higher cadence as it increases blood flow and removes lactate from the muscle faster, improving contraction efficiency. This could be preferable to working at the same workload and lower cadence where lactate accumulates to a greater extent in the muscle and leads to decreased muscle efficiency and resultant decrease in speed.

Results from Moritani et al. (1993) revealed a significant correlation (r=0.92) between oxygen consumption at neuromuscular fatigue and at anaerobic threshold during cycling. Neuromuscular fatigue was defined as the change in integrated EMG (iEMG) as a function of time, or the gradient of the iEMG. However, the mean oxygen consumption at neuromuscular fatigue (1.84 L/min) was significantly greater (p<0.01) than at anaerobic threshold (1.72 L/min). During this study, cadence was freely chosen and the workload was varied.

Relationship Between Cadence and Electromyography (EMG) Studies on cyclists (Moritani et al., 1993; Takaishi et al., 1996) found that there might be a point where neuromuscular fatigue is minimised (~ 90 rpm), yet does not optimise

O2 efficiency (~ 60 rpm). This point of neuromuscular fatigue minimisation has been noted to coincide approximately with the preferred cadence of cyclists (~90 rpm) (Moritani et al., 1993; Takaishi et al., 1996). It also has held true for triathletes during cycling (Hausswirth et al., 1999) but it is not so clear-cut in the other two disciplines.

Marsh and Martin (1995) hypothesised that the average and peak activation of lower limb muscles would display a minimum activation at the preferred cadence of both cyclists and non-cyclists. Again, non-cyclists with similar O2max results (>60 ml·kg- 1·min-1) were chosen as subjects. It should be noted that they were distance runners and their similar training experiences could have affected the results as no differences were found between the two groups (Marsh & Martin, 1997). A positive and significant linear trend was noted between gastrocnemius peak and average EMG with cadence; vastus lateralis average EMG and cadence; and a quadratic relationship between EMG and rectus femoris with a minimum at 95 rpm. Data were recorded 1 min after subjects 29 had reached and maintained the required cadence. Previous literature has indicated that cyclists and non-cyclists maintain cadence (+ 5 rpm) over a 30 min period and a 2 min period for determination of preferred cadence is valid (Marsh & Martin, 1993). However, it could be possible to maintain these cadences accurately but it is not known what happens to the firing rate, or efficiency, after 10 min or 60 min. The cadence might not change but EMG could vary significantly and show that efficiency changes during exercise.

Hagan et al. (1992) investigated cardiorespiratory response during continuous exercise at two pedalling rates and found that O2 was significantly higher during a cadence of 90 rpm when compared with 60 rpm, and at higher power outputs. However, the respiratory exchange ratio (RER) was found to be significantly lower at 90 rpm than at 60 rpm and was not affected by power output. A decreased RER, in conjunction with an increase in O2, suggests that greater amounts of CO2 are either produced and/or removed. During this study ventilation was significantly greater at the higher cadence and there were no significant differences between blood lactates (Hagan et al., 1992).

Fuel utilisation could change at different pedal frequencies. Ahlquist et al. (1992) found that glycogen depletion of type II muscle fibres was greater when at 50 rpm compared with 100 rpm, while cycling at 85% O2max. It is well known that O2 increases linearly with time when exercising at a constant workload. Takaishi et al. (1996) found a similar trend with EMG data. As exercise duration increased, there was a linear increase in muscle firing amplitude. However, the gradient varied with changes in cadence, even when at equal workloads. Takaishi et al. (1996) suggested that, if the relationship between increasing O2 and EMG is close and direct, the cyclist might choose a cadence which minimises the increase in O2 over time.

The optimal cadence to minimise oxygen cost significantly increased from 70 rpm at 4 min to 86 rpm at 29 min during prolonged cycling (Brisswalter et al., 1999). The preferred cadence (80 rpm) during the early stages of this experiment was significantly greater than the economically optimum (70 rpm), which is in agreement with most research (Gotshall et al., 1996; Hagan et al., 1992; Marsh & Martin, 1993 & 1997). A second data collection showed that minimal energy cost was achieved with a cadence 30 similar to the freely chosen cadence at this time (83 rpm). In a follow-up study by

Hausswirth et al. (1999), examination of both EMG and O2 at different cadences over 90 min of cycling found that, after 90 min of freely chosen cadence, the cadence did not differ from the cadence which minimised both muscle activation and oxygen cost.

Edwards et al. (1977) found 80-90% maximum voluntary contraction (MVC) is obtained with a stimulation of 30 Hz and MVC is obtained at approximately 50 Hz. Relaxation time of the quadriceps muscle to 50% peak force was 104.7 + 9.4 m/s following a muscle stimulated at 30 Hz for 1 s to produce tetanus. No difference was obtained between quadriceps and adductor pollicis, which suggested that indices of muscle function are independent of muscle mass and unaffected by the presence of inactive muscles in parallel. Moulds et al. (1977) also found no difference in relaxation times between different muscles throughout the body, no significant differences between in vivo and in vitro testing and found a negative and significant correlation between type II muscle fibres and relaxation time. Miler-Brown and Stein (1975) agreed that surface EMG and muscle force was linearly related, and that, as force increases, it is achieved first by an increase in motor unit recruitment followed by an increase in firing rate.

Faria (1992) found that quadriceps muscle twitch response time to be 250 ms, or half the crank revolution when pedalling at 120 rpm during cycling. With a relaxation time of 105 ms (Edwards et al., 1977; Moulds et al., 1977) accounting for another 72 degrees, Faria (1992) then suggested that the inability of the muscle to contract and relax any faster could explain why the force continues into the crank cycle further than desirable. Wilkie (1950) suggested that there appeared to be a fundamental firing frequency of about 50 cycles/s as a result of synchronous discharge of the anterior horn cells. There is currently no evidence to suggest that a decreased neuromuscular fatigue during cycling will improve performance at the end of a triathlon, that is during the run; or that the increased oxygen cost related to using a higher cadence will lead to a diminished run performance due to a lack of fuel available.

Cadence variation prior to change in disciplines It has been suggested by coaches, that 'spinning' the legs prior to dismounting the bike and commencing the run, relaxes the legs and allows easier transition from cycle to run. One hypothesis is that it increases blood flow to the muscles of the lower limbs which can flush-out waste products and/or move blood into the muscles which will be working next. It also reduces the force of contraction of the muscles and enhances relaxation (Takaishi et al., 1996 & 1998). However, this would only occur if gears were changed to allow a similar velocity to be maintained with reduced force per revolution and not impede or occlude blood flow.

This would also hold true in the swim to bike transition, where swimming is predominantly an upper body activity and, when wetsuits are permitted, the legs are not required to work as hard because buoyancy is increased and drag reduced (Chatard et al., 1995; de Lucus et al., 2000). Thus, changing from a four beat to a six beat kick coming into transition may have similar benefits. Triathlon coaches have been known to use "kick sets" following hard running or cycling sessions to increase blood flow to the legs and remove waste products and decrease recovery time.

However, a change from a four to six beat kick could also decrease efficiency and cause

O2 debt and/or increase in blood lactate levels. This could negatively affect cycle performance in the initial stages of this discipline. The increased blood flow to the legs may also lead to difficulty in the transition, as there is less blood flow to the head, and with a change in posture from prone to standing, there is a tendency to feel light-headed or disorientated upon exiting the water.

Also, in the bike-run transition, ‘spinning’ generally requires a drop in speed as triathletes tend to gear down to increase cadence. The decrease in speed might outweigh the benefits of spinning. If 20 s was lost in the last 1.5 km of the ride, the triathlete may not be able to run 20 s faster over the 10 km run following ‘spinning’. On the other hand, this increased spinning should result in a decreased muscle force and an increased blood flow which might permit greater recovery prior to the commencement of the run.

Gotshall et al. (1996) investigated seven male cyclists and found that increasing cycle cadence (70, 90 & 110 rpm) at a constant workload, leads to a decrease in systemic vascular resistance (SVR), a decrease in a- O2 difference and an increased cardiac output ( ). This suggests that is excess at 110 rpm (cf. 70 & 90), and supports the hypothesis that the skeletal muscle pump increases in effectiveness with increasing cadence. This, in turn, leads to increased muscle blood flow and venous return and may result in selection of higher cadences even though there is a reduced economy as measured by O2.

Takaishi et al. (1996) measured both EMG of vastus lateralis and O2 consumption, while cycling at a constant workload and varying cadences (50-100 rpm). They found a linear increase in EMG throughout the duration of exercise at all cadences. However, the gradient was dependent on the selected cadence and they suggested this could be linked to O2 consumed. Therefore, it might be possible to vary cadence throughout a steady workload to minimise both EMG and oxygen cost with respect to time. This raises the question as to whether varying cadence throughout steady state exercise could reduce fatigue.

People normally change from walking to running at the appropriate speed to ensure they use a more efficient gait for the chosen speed (Alexander, 1984). Hoyt and Taylor (1981) showed that horses avoided awkward, inefficient speeds at the borderline between gaits, and adopted preferred speeds for walk, trot and gallop. At velocities on the borderline, animals tend to speed up, slow down or change between the two gaits. Altering cadence during cycling could fall into the same category as altering gait because it has been shown that cycling biomechanics differs with changes in cadence (Neptune et al., 1997). However, these changes on a machine may not constitute a change in gait.

Muscle Flexibility After cycling, running feels difficult, and it takes some time to "warm up" or adjust. This could be due to shortening, or tightening, of the hamstring muscle group as a result of cycling. If so, the tighter hamstrings, or even hip flexors, could alter running 33 biomechanics by reducing ROM and shortening SL. Therefore, to maintain running speed, an increase in cadence may be necessary.

Hence, it is necessary to biomechanically assess running in terms of SL, cadence, and running with and without cycling preceding the run. The second section would require measurement of hamstring flexibility prior to, during and after, both runs in both tests.

Cadence and Ground Reaction Forces (GRF) Cavanagh and Lafortune (1980) showed that the GRF produced by a runner is approximately three body weights at a velocity of 4.5 m/s, which is slower than velocities of elite male and female triathletes (~5 m/s). After 6000-8000 steps (3000- 4000 strides), during a 10 km run, 9000-12000 body weights have impacted on each leg.

Reduced velocity has decreased GRF and high GRFs have been linked to many lower limb (LL) injuries in running. Triathletes do not want to decrease velocity and it could be possible to decrease GRF by another means to perhaps decrease LL injuries or enable an increased training volume or intensity with less risk of injury. This is in agreement with Martin and Marsh (1992), who reported increased vertical GRF when SR is decreased (thus increased SL) during walking.

Mero and Komi (1986) suggested that men have a longer SL than women due to a greater GRF. However, perhaps the higher GRF is a result of the longer SL, attained by men who have longer legs than females or the result of larger mass of male athletes. If this is the case, an altered SL-SR relationship which maintains velocity, without a decrement in economy, could reduce the GRF of each stride. It is well known that GRF decreases when running up hill and increases when running down hill. The SL decreases when running uphill and prevents the LL from accelerating completely before making contact with the ground. Running downhill has the opposite effect.

Takaishi et al. (1996 & 1998) reported that an increase in cadence while cycling leads to a significant decrease in peak pedal forces and force required to turn the cranks, in both cyclists and non-cyclists, even with the same power outputs.

It is possible that, although there may be a reduced GRF with each contact, the total GRFs over a 10 km run might not change as the runner will take more strides due to increased SR. Thus, the aim is to investigate the relationship between GRF at various SL-SR combinations while maintaining a constant velocity.

The above literature review highlighted the lack of documented knowledge of the sport of triathlon and the numerous potential areas of study. The focus of the following studies aims to shed new light on the areas of triathlete morphology and the influence cadence plays on triathlon performance. 35

Chapter 3

Study 1 Morphology and performance of world championship triathletes Based on Landers, G.J., Blanksby, B.A., Ackland, T.R. & Smith, D. "Morphology and performance of world championship triathletes." Annals of Human Biology, 27(4), 387- 400, (2000).

Abstract Performance is related to body morphology in many sports. With triathlon making its debut into the Olympic programme in 2000, it was deemed important to determine which physical characteristics of elite level triathletes were significantly related to performance. Seventy-one elite and junior elite triathletes, from 11 nations, who competed at the 1997 World Triathlon Championships were measured on a battery of 28 anthropometric dimensions (Landers et al., 2000).

A factor analysis was conducted which reduced the number of variables to four and these were used in a stepwise linear regression to determine which morphological factors contributed most to performance.

Elite triathletes were significantly (p<0.05) faster over the complete triathlon than their junior counterparts (males 1:52:26 vs 2:03:23 and females 2:07:01 vs 2:14:05) and showed less variation in performance times. Run time variation was the largest of the component disciplines and tended to show the importance of this discipline to the final outcome.

Following a factor analysis, the four distinguishable morphological factors that emerged were; robustness, adiposity, segmental lengths and skeletal mass. Relating these factors

36 to the total time obtained by the triathletes in this study yielded a regression equation that correlated significantly with all triathletes, accounting for 47% of the variance in total triathlon duration.

The regression equations illustrated the importance of low levels of adiposity for elite triathletes for total time and most of the sub-disciplines. The other important factor showed that proportionally longer segmental lengths contributed to successful swimming outcome. Raw data have previously been reported in Ackland et al. (1998a).

Introduction Triathlon originated only 30 years ago and many changes have taken place in that short amount of time. It is a rapidly growing sport, which took its inaugural position in the Olympic programme in Sydney 2000. The Olympic triathlon, also known as the classic or traditional distance, includes the continuous performance of the three separate disciplines of swimming (1.5 km), cycling (40 km) and running (10 km), respectively; and with quick transitions from one to another.

There is a void of information on elite triathletes who compete over the Olympic distance event and, with the inclusion of triathlon into the Olympic programme, further research was deemed necessary. Some research has been conducted on physiological responses during triathlon performance and the physiological characteristics of triathletes (Schneider et al., 1990). However, if only physical characteristics determined athletic success, it would be easy to predict winners. A combination of many aspects are involved with perhaps body shape or body structure a contributing factor to one’s ultimate athletic accomplishment (Atwater, 1990).

Longer lever lengths have been found to be advantageous in swimming (Bloomfield and Sigerseth, 1965). Various authors (Tittle & Wutscherk, 1988; Sleivert & Rowlands, 1996) stated that long limbs allow for greater economy of effort because an increase in stride or stroke length is more efficient than an increase in frequency. Body fat correlates negatively with running success because it is essentially a dead mass that must be carried throughout the event (Atwater, 1990). Therefore, body measurements that are directly related to the motor process, such as stride frequency and length, could provide the most important values for performance prediction.

Anthropometric data have been collected from small numbers of triathletes competing at the Hawaiian Ironman, an ultra endurance event lasting in excess of eight hours (Holly et al., 1986; O’Toole et al., 1987; O’Toole et al., 1989), or other ultra endurance events (Dengel et al., 1989; Kohrt et al., 1987; Kohrt et al., 1989) (Appendix C). While these studies were conducted in the early 1980s, during the formative years of the sport, it is also reasonable to suggest that the physical and physiological requirements for success in such a long endurance event could well be different from an event lasting

38 approximately two hours (Sleivert & Wenger, 1993). No kinanthropometric data were found on elite level triathletes competing over the World Championships classic distance of a 1.5 km swim, 40 km cycle and a 10 km run.

Other studies have focused on female triathletes (Laurenson et al., 1993; Leake and Carter, 1991; Sleivert & Wenger, 1993) with only one of these comparing elite with club level females who compete over the classic distance. Others have studied groups of well-trained classic distance triathletes (De Vito et al., 1995; Hue et al., 1998; Rowbottom et al., 1997; Schneider et al., 1990; Sleivert & Rowlands, 1996; Zinkgraf et al., 1986), or recreational athletes only (Sleivert & Wenger, 1993). All these preceding studies have focussed on physiological aspects, describing only the height and body mass of their subjects, and therefore have not analysed the morphology relationships with performance.

The identification of any attribute that predicts good performance is important. It can be used to design training programmes that attempt to modify the plastic phenotype characteristics of the athletes towards the optimum morphology. It could also be used in talent identification programmes to direct a young athlete having the desired genotype and phenotype into the sport of triathlon.

Thus, it was considered valuable to try and identify any physical characteristics that might be advantageous to the sport over this distance. This project profiled male and female, senior and junior elite, world championship triathletes in Western Australia during the 1997 Triathlon World Championships (TWC). Body type and composition, along with phantom proportionality values were also recorded and correlated with performance. Prediction equations were developed on the basis of factors calculated from a factor analysis, and differences between elite and junior competitors were assessed.

At present it is not known which body types, proportions or shapes best suit each of the three components of a triathlon event. It could be that a champion all rounder will out- perform one who is an exceptional athlete in only one or two disciplines. Initial analysis of the raw data by Ackland et al. (1998b) revealed that successful male

39 triathletes possessed an ectomorphic shape, large chest breadth and a high proportional thigh length. Poorer performances were associated with high levels of adiposity and measures of body bulk such as hip and thigh girths. Successful female triathletes showed similar characteristics to the males, being advantaged by factors such as a linear physique. Those athletes with greater adiposity, or greater thigh and hip girths, did not perform as well. Therefore, this current study sought to continue to describe the kinanthropometric features of elite traditional distance triathletes, and to make performance, age and gender-related comparisons.

Methodology Sample Male and female triathletes, from the senior and junior elite categories participating in the 1997 Triathlon World Championships (TWC) were invited to take part in the study. Eighty-seven triathletes from 11 nations were measured from the 1997 TWC. Of this sample, only 71 actually completed the race. The male sample consisted of 20 elite male triathletes (25.5 y, 179.8 cm, 72.3 kg), eight of whom finished in the top 20; and 29 junior male triathletes (19.1 y, 175.7 cm, 67.0 kg), three of whom finished in the top 20. The female sample comprised 18 elite female triathletes (29.3 y, 168.3 cm, 59.5 kg), six of whom were placed in the top 20; and 20 junior female triathletes (19.0 y, 164.9 cm, 56.71 kg), of whom eight placed in the top 10.

In order to compete in the elite division at the TWC, triathletes must be ranked by the International Triathlon Union (ITU). Each country can select a maximum of six competitors in each category, with the elite field limited to 80 triathletes. If more than six triathletes from any one country are ranked in the top 80 ITU rankings, a decision is made by the country as to which six triathletes will compete. Only volunteers participated in the research and a wide cross-section of countries were approached. All testing was carried out one week prior to competition to minimise disruption to the final championship preparation.

Testing protocol Approval for the study was obtained from the Human Rights Committee of The University of Western Australia, and all participants signed a consent form after being

40 informed of the study requirements (Appendix D). Data were collected at the Department of Human Movement and Exercise Science, or at another suitable accommodation site using a mobile testing unit. All equipment other than electronic scales and stadiometer could be transported. For these occasions, a mobile stadiometer and beam balance were used and calibrate before use.

Personal details and demographic data were obtained at the first test station and anatomical landmarking of the subject took place before moving to other stations where various data were obtained. One station was set up to record seven skinfold thicknesses, another for height, mass and five segmental lengths; two stations were available for four girth measurements each and another for five segment breadths. At each measurement station there was a scribe to record values to enable the tester to measure continuously and reduce testing time (Appendix E). The data were then entered into a computer on an SPSS spreadsheet (SPSS Inc.) for subsequent analysis.

Measurements were taken according to standards set by the International Society for the Advancement of Kinanthropometry (ISAK) (Norton & Olds, 1996) and followed the procedures set out by Carter and Ackland (1994) in the Kinanthropometry in Aquatic Sports Project (KASP) study of world championship swimmers (Appendix F).

The same qualified anthropometrist (level 2 ISAK) was posted at each station for the entire testing programme. This strategy was adopted to limit inter-observer variability and maximise the reliability of results. The technical error of measurement (TEM) was calculated on a sample of 20 subjects randomly chosen. The TEM for all variables was within the standards set by ISAK for level two anthropometrists (Norton & Olds, 1996).

After the first measurement of each variable, the sequence was repeated. If a difference in scores exceeded 0.2 mm in skinfold, or 0.5 cm for other variables, a third measure was taken. The final value used in further analysis was the mean of the two scores or the median of three if taken.

Data analysis Data were entered onto a spreadsheet from which means and standard deviations for

41 each characteristic were calculated. Standard formulae were employed to derive somatotype (Carter & Heath, 1990) (Appendix G). Proportionality was calculated via the phantom stratagem for all measurements (Ross & Marfell-Jones, 1991) (Appendix G). The Phantom Stratagem is a technique used to analyse proportionality by removing size and gender differences (Ross & Marfell-Jones, 1991). This method is generally preferred as, in contrast to standard indices, it allows the analysis of results in parametric statistical techniques (Ross & Ward, 1986).

Body composition was determined using the five compartment fractionation model (Kerr & Ross, 1994; Ross & Kerr, 1991) (Appendix G). Five-way fractionation subdivides the body’s mass into the following five compartments: skin mass, bone mass, muscle mass, adipose mass and residual mass. Residual mass relates not to the mass remaining after subtracting the other four compartments from total mass, but has its own value describing internal organ masses. Ackland (1994) stated that this technique was more accurate than that of Drinkwater and Ross (1980) in predicting adipose mass and was no worse than other regression equations for predicting compartmental masses. The Kerr and Ross (1994) five-way fractionation method does not rely on assumptions of constant density of fat mass or lean body mass. The derivation is independent of body mass and, as all scaling and massing of items is done using the phantom stratagem, it is not population specific.

Tables containing all of the measures and derived variables were constructed for males and females separately, as well as elite and junior groups (Ackland et al., 1998a). Split times for each discipline (excluding transition time) and total time (including transition time) were recorded and used as performance variables.

A factor analysis was undertaken in order to reduce the number of independent variables prior to their entry into a step-wise linear regression analysis. The aim was to identify underlying factors that explain the correlations among variables. The extraction process, via the ‘principal component’ method, was used to determine four factors. Variables were sorted by size and cases excluded listwise. The initial solution was varimax rotated to obtain the final solution. This was conducted using all z-scores and somatotype scores and, thus, the complete battery of measures was used in

42 determining the factors.

Body characteristics were then regressed with each of swim, cycle, run and total time via a linear, stepwise regression. This enabled the determination of favourable variables from each sub-discipline and overall triathlon performance. Equations for all the triathletes were determined as well as separate equations for all male and all female triathletes. Actual times were then correlated with times predicted from these equations.

In order to determine the relative importance of each discipline and enable comments on morphology with respect to performance, the rank order of all triathletes who completed the course was determined for final position. These data were then correlated, using a Spearman Rank Order correlation technique, with position at the end of the swim and cycle legs, as well as with the ranked position for the discrete run and cycle legs.

A 2-way repeated measures ANOVA was used to determine any differences between elite and junior competition times in all three disciplines, and total time. A Tukey’s HSD post hoc test was used. The sample was split by gender as male and female triathletes are anatomically different (Ackland et al., 1998a), and are assumed to be biologically different.

Results Performance correlations - all competitors A total of 58 elite male, 58 elite female, 34 junior female and 84 junior male triathletes completed the course at the 1997 Triathlon World Championships. These triathletes remained in their respective categories and were ranked on the basis of their time to finish each leg (swim, cycle and run), as well as end of the cycle leg and the complete triathlon. The rank of triathletes for the swim leg is equivalent to the rank of position at the end of the swim.

Correlations of each rank position with total position are included below (Table 3.1). The final finishing position correlated significantly with the ranking in the swim, cycle

43 and run. Only swim ranking for the junior females was not correlated significantly with final finishing position.

Table 3.1 Rank order correlations of swim, cycle, run and end bike position rank with final rank.

Elite male Elite female Junior male Junior female Final position Final position Final position Final position Swim rank 0.603* 0.533* 0.568* 0.377 Cycle rank 0.750* 0.875* 0.855* 0.855* End bike pos. 0.832* 0.904* 0.892* 0.875* Run rank 0.871* 0.835* 0.856* 0.533* Note: * signifies significance at p<0.05

Elapsed times for elite vs junior competitors The elite male triathletes were significantly faster than the junior males in all three disciplines, as well as the total event time (Table 3.2). They also recorded a smaller time variation than the juniors, as indicated via a smaller standard deviation. The greatest variation in times for all male triathletes was obtained in the running section of the triathlon.

The elite female triathletes were also significantly faster than the junior females in cycling, running and the overall time to complete the event (Table 3.2), but not in the swim leg. In contrast to the males, the female triathletes revealed the greatest time variation in the cycle leg of the triathlon.

Total elapsed time for both male and female triathletes were significantly correlated (p<0.05) with swim (r = 0.586, 0.522 respectively), cycle (r = 0.950, 0.929 respectively) and run (r = 0.932, 0.875 respectively) times. A high inter-correlation was noted between cycle and run times (r = 0.801, 0.684 respectively).

Table 3.2 Descriptive statistics & ANOVA summary of times for elite versus junior competitor comparisons Combined group Elite Junior F ratio p Males mean SD Mean SD mean SD TOTAL s 7114 414 6746 145 7403 313 67.334 0.0001 hms 1:58:34 1:52:26 2:03:23 SWIM s 1190 78 1158 26 1211 94 6.073 0.0001 hms 0:19:50 0:19:18 0:20:11 CYCLE s 3717 183 3555 59 3844 141 65.96 0.0001

hms 1:01:57 0:59:15 1:04:04 RUN s 2104 210 1922 96 2247 158 58.182 0.0001 hms 0:35:04 0:32:02 0:37:27

Combined group Elite Junior F ratio P Females mean SD Mean SD mean SD TOTAL s 7819 403 7621 333 8045 363 11.097 0.002 hms 2:10:19 2:07:01 2:14:05 SWIM s 1287 99 1274 56 1298 126 ns ns hms 0:21:27 0:21:14 0:21:38 CYCLE S 4105 211 4022 195 4199 195 6.143 0.0001 hms 1:08:25 1:07:02 1:06:59 RUN S 2313 187 2197 130 2446 151 23.388 0.0001 hms 0:38:33 0:36:37 0:40:46 Note: ns = no significant difference s = seconds hms = hours:minutes:seconds

Factor analysis Table 3.3 illustrates the results of a principal components factor analysis of kinanthropometric data based on a 0.4 suppression factor. The following four factors were identified by the predominance of values exceeding 0.4 (seen in bold in Table 3.3):

• Factor 1 = "robustness" • Factor 2 = "adiposity" • Factor 3 = "segmental length" • Factor 4 = "skeletal mass"

Whilst most of the variables showed a single dominance, there were three that showed dominance in two categories. These variables were proportional biacromial breadth with robustness and skeletal mass; and proportional humerus and femur breadths with skeletal mass and robustness. Proportional anterior-posterior breadth had a correlation coefficient of 0.317 with robustness (Factor 1). Thus, it is included in this factor, even though it is less than 0.4. The level of suppression was set arbitrarily to assist in revealing factors. Table 3.3 shows all the coefficients by which variables are multiplied in order to obtain the four factor scores.

Stepwise linear regression Attempts were made to develop multiple, linear prediction equations for total, swim, cycle and run times for the complete sample, as well as for male and female triathletes separately (Tables 3.4-3.6). Performance predictions were made using the four factors (robustness, adiposity, segmental lengths and skeletal mass) to judge their relative importance to this process, and correlation between actual race times and times predicted using the developed regression equations are included.

Table 3.3 Complete principal component matrix coefficients by which variables were multiplied to obtain factor scores and total variance explained. Factor Factor Factor Factor 1 2 3 4 Proportional mass 0.871 0.362 -0.113 -0.179 Ectomorphy -0.87 -0.352 0.109 0.189 Proportional arm girth 0.869 0.24 -0.019 0.011 Proportional chest girth 0.867 0.043 0.245 0.043 Proportional arm flexed girth 0.866 0.122 0.129 0.122 Mesomorphy 0.855 0.02 -0.168 0.341 Proportional forearm girth 0.814 0.038 -0.049 0.28 Muscle mass percent of total mass 0.806 -0.306 -0.004 0.051 Proportional waist girth 0.701 -0.047 0.365 0.203 Proportional chest breadth 0.662 -0.104 0.118 0.135 Proportional calf girth 0.642 -0.06 -0.448 -0.124 Proportional biacromial breadth 0.575 -0.065 0.198 0.468 Proportional anterior-posterior 0.317 -0.009 0.262 0.137 breadth Sum of 8 skinfold Z-score -0.02 0.977 -0.11 -0.099 Endomorphy 0.113 0.959 0.086 -0.085 Adipose mass percent of total mass -0.399 0.877 -0.143 -0.095 Proportional abdominal skinfold 0.108 0.839 0.006 0.101 Proportional tricep skinfold -0.068 0.833 -0.225 -0.294 Proportional iliac skinfold 0.104 0.829 0.159 0.186 Proportional calf skinfold -0.175 0.806 -0.209 -0.255 Proportional subscapular skinfold 0.325 0.785 0.044 0.122 Proportional bicep skinfold 0.045 0.758 -0.333 -0.016 Proportional front thigh skinfold -0.113 0.711 -0.385 -0.349 Proportional hip girth 0.358 0.645 -0.357 -0.259 Proportional sitting height 0.233 0.232 -0.699 0.126 Proportional tibiale-laterale length -0.069 -0.146 0.686 -0.025 Proportional arm span 0.346 -0.199 0.669 0.144 Proportional acromiale-radiale length 0.132 -0.173 0.62 -0.217 Proportional radiale-stylion length 0.169 -0.039 0.411 0.125 Bone mass percent of total mass 0.053 -0.169 0.045 0.856 Proportional humerus breadth 0.471 -0.277 0.144 0.591 Proportional femur breadth 0.422 0.249 -0.212 0.585 Proportional trochanterion-tibiale length 0.016 0.228 0.379 -0.447

Initial Eigenvalues 9.25 9.12 2.66 2.05 % of variance (cumulative %) 28.0 (28.0) 27.6 (55.7) 8.0 (63.7) 6.2 (69.9)

Final Loading 8.69 8.3 3.33 2.74 % of variance (cumulative %) 26.3 (26.3) 25.1 (51.5) 10.1 (61.6) 8.3 (69.9) Note: Variables highlighted in bold exceed the arbitrarily set suppression value of 0.4.

Table 3.4 Factor B mass and regression coefficients in final regression model for entire sample Dependant B0 B1 B2 B3 B4 Regression F Variable coefficient Total time (s) 7435.0 ... 324.8 -180.2 ... 0.688 29.195** Swim time (s) 1232.6 ... 31.9 -44.3 ... 0.549 17.096** Cycle time (s) 3892.6 ... 149.5 -99.9 ... 0.659 24.877** Run time (s) 2201.0 ... 146.2 ...... 0.618 40.844** Where: Y = dependant variable X1 = robustness X2 = adiposity X3 = segmental lengths X4 = skeletal mass

Table 3.5 Factor B mass and regression coefficients in final regression model for male sample Dependant B0 B1 B2 B3 B4 Regression F Variable coefficient Total time (s) 7204.6 ... 195.6 ...... 0.368 5.794* Swim time (s) 1222.9 ... 48.9 -32.8 ... 0.539 9.206** Cycle time (s) ...... ns Run time (s) 2152.9 ... 110.6 ...... 0.410 7.655** Where: y = dependant variable x1 = robustness x2 = adiposity x3 = segmental lengths x4 = skeletal mass

Table 3.6 Factor B mass and regression coefficients in final regression model for female sample Dependant B0 B1 B2 B3 B4 Regression F Variable coefficient Total time (s) 7704.8 ... 233.1 ...... 0.533 10.309** Swim time (s) 1268.3 ...... -42.8 ... 0.410 6.880* Cycle time (s) 4057.8 ... 96.5 ...... 0.421 5.615* Run time (s) 2252.0 ... 131.7 ...... 0.643 18.280** Where: y = dependant variable x1 = robustness x2 = adiposity x3 = segmental lengths x4 = skeletal mass Note: for tables 3.4, 3.5 & 3.6 general format for multiple regression equation is: y = B0 + B1x1 + B2x2 ... + B4x4 * = significant p < 0.05 ** = significant p < 0.01

Discussion Performance correlations for the entire field From the results (Table 3.1) of all the elite and junior male and female triathletes competing at the 1997 TWC, it would appear that the most important determining factor of finishing place is the position of the athlete at the end of the cycle leg. At this stage, the race time is approximately 67% complete (Table 3.2). The elite male triathletes showed continued predictability throughout the race with increasing correlation coefficients, to the largest correlation with the triathlete’s ranking in the run component. The elite male triathletes showed high correlations between all disciplines and final position. Discarding the position at the end of the cycle leg, there is a noticeable trend of increasing correlation between finishing position and each successive leg, showing a greater significance of running ability to the overall position.

Results of the elite female and junior male triathletes were similar in that the race outcome was equally strong with cycle rank (r=0.875 & 0.855) and run rank (r=0.835 & 0.856), with position at the end of the cycle slightly greater (r=0.904 & 0.892) (Table 3.1). The elite females and junior males posted similar results for each discipline and the complete triathlon which could be assumed to be as a result of similar average times recorded during the race (Table 3.2).

The junior female triathletes differed in that the final outcome positions (r=0.533) had less to do with run rank than the other three groups and relied heavily on a good cycle performance (r=0.855). With the exception of junior female competitors, all categories showed positive, significant correlations with swim rank (or position out of the water).

Junior vs senior competitors The elite male triathletes were significantly faster than the junior male triathletes in each of the three disciplines and during the entire event. The elite males were more homogeneous in their performances in each discipline than the juniors. A smaller variation in discipline times for the senior males (closer racing) may mean that a weakness in any stage of the event can be clearly exposed. The greatest variation in times for the male triathletes occurred in the final discipline of the triathlon, the 10 km run leg. This large variation in times in the run leg could increase the probability that

49 the winner will be a triathlete with strong running ability and that the variation found in the run may be due to fatigue. After completing approximately 20 minutes of swimming and 60 minutes of cycling, energy systems are partially depleted and running efficiency declines (De Vito et al., 1995; Guenzennec et al., 1996; Hue et al., 1998; Lehénaff, 1998). This also indicates that the junior triathletes have room for improvement in all disciplines, and are not yet at the same level of the elite triathletes. This improvement might be as a result of increased training or could be related to differences in body morphology between the two levels of triathletes.

The elite female triathletes also were significantly faster than the female juniors in cycling, running and overall time. While there were no significant differences in swim times, the elite females tended to be faster than the junior females. The female triathletes recorded the greatest absolute time variation in completing the cycling leg of the triathlon. However, when expressed as a proportion of elapsed time for that component, there was more variation in swim and run times for the female triathletes. Once again, this demonstrated the importance of running ability for successful performance.

Factor analysis After the completion of the factor analysis, four discrete factors emerged containing similar variables (Table 3.3). The first comprised those revealing a robust body type and included variables of mass, girths and breadths of the trunk region, as well as variables relating to mesomorphy and muscle mass. The negative relationship between this factor and ectomorphy (-0.870) demonstrates a lack of linearity in this factor. Hence, it was named the “robustness” factor.

The second factor was linked with seven proportional skinfold measures, a proportional sum of eight skinfolds, percentage adipose mass and proportional hip girth. These measures had correlation coefficients greater than 0.7, thereby indicating a strong relationship with this factor. As it was also assumed that proportional hip girth would be influenced by fat storage, especially in females, this factor was termed “adiposity”.

Factor three included proportional segment lengths of leg, upper arm and forearm,

50 sitting height and arm span. Therefore, this factor related to the biomechanics of movement and included most of the propulsive levers of the body. The factor was named “segmental lengths” and is important for performance prediction as these body measurements are directly related to the motor processes of stride or stroke length and frequency, where long limbs allow for greater economy of effort (Sleivert & Rowlands, 1996; Tittle & Wutscherk, 1988).

The final factor of bone breadths and bone mass related to the skeletal mass of the triathlete. Proportional humerus and femur breadths tended to affiliate with the robustness variable; as did the biacromial breadth (in the robustness factor) with the "skeletal mass" factor.

As this analysis reduced the large number of variables to four succinct factors, meaningful regression equations could then be developed. The factors were concise in their meaning and separated the variables into different categories.

Stepwise linear regression All competitors Prediction equations for total time and separate discipline times were developed for all triathletes using the four factors obtained from the factor analysis. All four equations involved the adiposity factor for time prediction, thereby demonstrating the importance of low levels of body fat for greater endurance performances (Tittle & Wutscherk, 1988; Wilmore & Brown, 1974). That is, as the magnitude of adiposity increased, there was an increase in non-functional body mass and the level of performance declined. This is in agreement with initial findings of Ackland et al. (1998b) which suggested that successful triathletes possessed a more ectomorphic shape and those less successful had higher levels of adiposity.

Swim and cycle performances also rely on the segmental length factor for time prediction. That is, those with proportionally longer limb lengths tended to perform better in the swim and cycle legs. Long limbs in swimming have been shown to provide a biomechanical advantage if sufficient strength is available (Bloomfield & Sigerseth, 1965). Foley et al. (1989) found that time trial cyclists had relatively longer

51 lower limb lengths (RLLL) when compared with other cyclists (road and track). The prediction equation derived for cycle time in this study suggested that larger RLLL, with respect to stature, benefited cycling performance. This factor, segmental lengths was also highly related with phantom sitting height (-0.699), which gives similar results to RLLL; and was positively related to proportional tibiale-laterale height (0.686). It was originally hypothesised that this might not be such an important factor now that drafting is permitted in the cycle leg. Drafting refers to one cyclist riding in the slip- stream of another, reducing the energy cost to overcome drag forces by up to 30% (Faria, 1992). However, as it was the first draft-legal event for the junior competitors and most of the elite triathletes have had a background in non-drafting events, it follows that these triathletes have a trait similar to time trial cyclists (Foley et al., 1989).

The robustness of the athlete or skeletal mass added little to performance prediction presumably because the level of adiposity was such an important factor. It could also follow that, because skeletal mass shows some relationship with robustness, phantom humerus breadth, phantom femur breadth and phantom biacromial breadth, as they are in both factors, the relationship of power-to-mass ratio might already be optimised.

When male and female triathletes were combined, body morphology accounted for 47% of the variance in total elapsed time. The remaining portion presumably is due to a combination of physiological (Schneider et al., 1990; Zinkgraf et al., 1986) or psychological characteristics of the triathlete, or environmental factors. Nevertheless, body morphology appears to be an important ingredient in the fields of talent identification and development for elite triathletes.

Male triathletes Despite significant associations between one or more of the morphological factors and total, swim and run times for the male triathletes (Table 3.5), none of the factors explained a significant portion of the cycle leg performance. This finding is most likely explained by the formation of large packs (generally larger than the female packs) during the cycle leg where the usual race strategy is to maintain the average speed of the pack while also conserving energy through drafting (Faria, 1992). Generally, there are few triathletes who are strong enough on the bike to bridge the gaps up to preceding

52 packs individually.

Total time and run time again were found to be slower with increasing levels of adiposity. Therefore, the results of this study confirmed that triathlon is another endurance sport in which adipose mass needs to be minimised (Atwater, 1990; Wilmore & Brown, 1974). As running involves the greatest mass bearing of the three disciplines within triathlon (Flynn et al., 1990; Sleivert & Rowlands, 1996; Tittle & Wutscherk, 1988), it was not surprising that the run time was highly influenced by the adiposity of the triathlete.

Swim times for the male triathletes were found to be positively related to the adiposity, but negatively with proportional segmental lengths. The positive relationship of adiposity with swim performance could indicate that triathletes have sufficient muscle power to propel themselves through the water and do not rely on body fat for buoyancy. The swimmers who possess greater levels of adiposity but similar muscle mass may show greater girths. Thus, in theory, having a greater frontal area exposed to the water which leads to an increased drag and decreased performance (Hay, 1993). It must also be noted that the swim in the current study was contested in a salt water river. Salt water has a greater density than fresh water and thus increases the buoyancy of the swimmer.

Female triathletes Significant prediction equations for female triathlete times were derived in all three triathlon disciplines as well as for total time. Adiposity had a strong positive relationship with total, run and cycle times; whereas the factor of proportional segment lengths significantly predicted swim time. This also illustrates the importance of low levels of body fat and proportionally long levers for superior endurance performance.

Adiposity hampers run time because it increases the energy demands of an athlete attempting to keep pace with a runner of equal body mass but reduced fat mass (Atwater, 1990). The correlation between predicted times determined from prediction equations (Table 3.6) and actual times (Table 3.2) was greatest for female triathletes when predicting run time, suggesting a close relationship between low adiposity and

53 improved run performance.

Relating adiposity to cycle performance has been discussed above. Cycling is a partial mass-bearing discipline and, therefore, excess mass will hamper performance. This is especially so when a course includes many hills and corners. As the mass of the cyclist increases, there is an increase in the energy required to ride up hills and the kinetic energy needed to accelerate (Olds et al., 1995). This course consisted of 60 corners and 10 hill climbs (365 m vertical displacement) resulting in reduction of velocity and a need to accelerate after completing the turn or the climb. Increased adiposity also leads to increased girths and increased frontal cross sectional area, thereby creating greater drag forces (Olds et al., 1995; Sleivert & Rowlands, 1996).

Segmental lengths are important in swimming performance as longer levers create a biomechanical advantage when adequate strength is present to move the levers through the water. The equation for predicting swim time in female triathletes demonstrated a negative relationship with segmental lengths, highlighting the importance of longer limbs in swimming for higher performance.

Summary Elite triathletes were faster than their junior counterparts and showed less variation in performance times. Run time variation was the largest of the component disciplines and tended to show the importance of this discipline to the final outcome.

Following a factor analysis the four distinguishable morphological factors that emerged were; robustness, adiposity, segmental lengths and skeletal mass. Relating these factors to the total time obtained by the triathletes in this study yielded a regression equation that correlated significantly with all triathletes, accounting for 47% of the variance in total triathlon duration.

The regression equations illustrated the importance of low levels of adiposity for elite triathletes for total time and most of the subdiscipline times. The other factor that showed importance was that proportionally longer segmental lengths contributed to successful swimming outcome.

Chapter 4

Study 2 Race Analysis Based on “Swim Position and its effect on Triathlon Outcome” International Triathlon Coaching Symposium, July 23-24 2001, Edmonton, Canada, page 11.

Abstract Understanding the important aspects of an event allows greater understanding of the physical and physiological characteristics of triathletes by which to assist coaches in devising training practices. This study sought to identify the association of being in the first pack of swimmers out of the water and winning the event (Study 2A), and the relative contributions of each discipline to the final order of triathlon finishing positions for all competitors (Study 2B). Data were gathered from 10 male and 10 female ITU world cup events during the 1999 triathlon season. Mathmatica was used to analyse more closely the significance of pack formation on race outcome. Study 2A found that an average of 80% of the eventual winners exited the water with the first pack of swimmers. This occurred during 90% of elite male races and 70% of elite female races. Study 2B indicated that the run carried the greatest weight in determining final finishing position. This was followed by the swim with the cycle leg the least important of the three disciplines.

Introduction Previous investigations have suggested that the swim portion of the triathlon might be the least important leg because of the lower correlations between swim times and overall times, when compared with those of cycle or run times with total times (Dengel et al., 1989; Landers, 1998; Sleivert & Wenger, 1993). One explanation could revolve around the different durations of each individual discipline as a proportion of the total event time. For example, the 1500 m swim takes approximately 20 min to complete, the 40 km cycle around an hour, and the 10 km run approximately 35 min. This represents a total time of just under 2 hours for males and just over 2 hours for females. Hence, approximately 18% of the total time is accounted for by the swim, 52% by the cycle and 30% by the run. Other authors have reported roughly equal contributions of swim, cycle and run with final time or finishing position (Landers et al., 2000; Zinkgraph et al., 1986).

However, the more common, subjective view expressed by coaches and athletes, is that the swim plays a greater role than suggested by the above mentioned equivocal research results. They consider that the cycle might be the least important because it is really just a means of transiting from the swim to the run. Most likely, this is caused by the recent legalisation of drafting during the cycle leg. Hence, triathletes can position themselves behind other competitors during cycling and be 'shielded' from air resistance. Drafting can reduce energy cost by as much as 30% (Faria, 1992), thereby conserving energy for a faster run, or permitting an increase in cycle speed, or a combination of both. It is difficult for triathletes to change position during the cycle because a group of cyclists can travel at a greater speed than an individual rider. Several studies (De Vito et al., 1995; Guezennec et al., 1996; Hausswirth et al., 1997; Hue et al., 1998; Kreider et al., 1988 and Lehénaff et al., 1998) have shown that, after cycling, there is a decrease in efficiency, or greater oxygen cost of running when compared with the oxygen cost of just running. Hausswirth et al. (1999b & 1999c), Hue et al. (1998) and Lehénaff et al. (1998) also showed that, when drafting is permitted, there is less decrement in run performance.

Body morphology is another aspect that could lend support to the notion that the cycle discipline does not carry as much importance in elite level, draft legal events. Landers 57 et al. (2000) and Landers (1998) reported no significant correlations of body dimensions with cycle time, and the regression equations could not be developed to predict male cycle performance during a triathlon (Table 3.5). In an attempt to clarify this further, the first part of this study sought to determine the importance of being placed in the first pack of swimmers out of the water in relation to the overall finishing position of a triathlon. The second part of the project set out to characterise the importance of packs during the swim and cycle disciplines and the relative importance of each discipline on final race outcome. The results sought to characterise both male and female senior elite triathletes during the 1999 ITU world cup. 58

Chapter 4A

Study 2A Swim Positioning and its Effect on Triathlon Outcome

Methodology Sample Results from 11 International Triathlon Union (ITU) World Cup Races held during the 1999 season were collected via the ITU website (www.triathlon.worldsport.com) for analysis (Appendix H). The results included both males and females separately for 10 events each, which gave a total of 20 sample groups to investigate. The 11 events were conducted in eight countries over the classic distance of a 1.5 km swim, 40 km cycle and a 10 km run, and were draft legal.

Testing protocol After data collection, triathletes were categorised in terms of final finishing positions, and the pack of swimmers in which they alighted from the water and commenced the cycle leg. Visual data were not sufficient to clearly identify the number of swimmers in each pack for all events. Hence, it was decided that, if a swimmer was more than 5 s in front of the following swimmer, that was deemed to be the termination point for that pack. The subsequent swimmer was considered to be the leader of the next pack.

Data analysis In order to determine the importance of the swim leg in triathlon competition, the number of winners that came from the first pack of swimmers (pack1) were tallied. Secondly, the number of triathletes in pack1 that finished the triathlon in the top 10 were tallied. This enabled determination of the percentage of pack1 swimmers to finish in the top 10 and the percentage of top 10 places filled by pack1 swimmers. Knowing the percentage of top 10 places filled by pack1 swimmers can highlight the importance of being out of the water in the first pack. Conversely, knowing the percentage of pack1 swimmers to finish in the top 10 can reveal the probability of placing in the top 10 at the end of the race if an athlete is out of the water with the first pack, especially if the first pack is greater than 10 in number. 59

Results Raw data of the number of triathletes in the first pack of swimmers (pack1) and, subsequently, those who finished in the top 10 for each triathlon are in Table 4.1 for males (Figure 4.1) and Table 4.2 for females (Figure 4.2). The male races held in Ishigaki (Japan), Gamagori (Japan), Kona (USA) and Montreal (Canada) revealed no discernible differences between triathletes at the end of the swim leg. Based on a 5 s separation criteria, video analysis and race reports indicated that, only one pack of swimmers emerged from the water and that same pack cycled together in these events. Revisiting the results also showed no differences between athletes at the end of the cycle, with all athletes entering the transition from bike to run together.

In two other men’s races (Belgium & Hungary), the number of pack1 swimmers was greater than 10. Hence, it was not possible for all of the pack1 swimmers to finish in the top 10. Thus, a secondary analysis was undertaken which limited the sample to those races where pack1 included 10 or less triathletes (Table 4.3).

Table 4.1 Number of male triathletes out of the water in the first pack and, subsequently, those who finished in the top 10 of races during the 1999 Triathlon World Cup Season. Location Date Finishers/ Triathletes Top 10, Starters in pack1 pack1 triathletes Ishigaki, April 11 72/80 72(all) 10* Japan Gamagori, April 18 66/82 66(all) 10* Japan Kapelle-op-den Bos, June 13 50/60 33 10* Belgium Monte Carlo, June 20 41/47 3 0 Monaco Kona, June 26 24/46 30 10* USA Tiszaujvaros, Aug 8 51/53 20 7* Hungary Corner Brook, Aug 15 32/41 2 2* Canada Lausanne, Aug 29 52/70 4 1* Switzerland Montreal, Sept 12 67/76 67(all) 10* Canada Noosa, Nov 7 40/50 8 5* Australia Note: all = entire field exited the water in one pack. * = event winner came from pack one

Figure 4.1 Number of male triathletes out of the water in the first pack and, subsequently, those who finished in top 10 of each race during the 1999 Triathlon World Cup Season.

80 70 60 50 Starters 40 Pack 1

athetes (n) 30 Top 10 o.

N 20 10 0 12345678910 Event 61

Table 4.2 Number of female triathletes out of the water in the first pack and subsequently those who finished in the top 10 of races during the 1999 Triathlon World Cup Season. Location Date Finishers/ Triathletes Top 10, Starters in pack1 Pack1 triathletes Ishigaki, April 11 68/71 4 3* Japan Gamagori, April 18 63/69 2 2* Japan Sydney, May 2 40/52 3 3* Australia Monte Carlo, June 20 37/42 10 4* Monaco Kona, June 26 20/27 4 4 USA Tiszaujvaros, Aug 8 39/46 1 1* Hungary Corner Brook, Aug 15 27/31 2 1 Canada Lausanne, Aug 29 39/49 2 1* Switzerland Montreal, Sept 12 52/62 6 3* Canada Noosa, Nov 7 25/30 5 2 Australia Note: all = entire field exited the water in one pack. * = event winner came from pack one

Figure 4.2 Number of female triathletes out of the water in the first pack and, subsequently, those who finished in top 10 of each race during the 1999 Triathlon World Cup Season.

80 70 60 50 Starters 40 Pack 1

athletes (n) 30 Top 10 o.

N 20 10 0 12345678910 Event 62

A total of 605 males and 479 females started in the 11 events under examination. In the complete male sample, 305 (50%) competitors exited the water in pack1 (Table 4.3). In the reduced male sample of four races, this included 208 starters. Seventeen (8%) of these male triathletes exited the water in the first pack of swimmers. Only 39 (8%) of the 479 female starters exited the water in the first group of swimmers over the 10 races (Figure 4.3 & 4.4).

Table 4.3 shows that, across four male races, 20% (8/40) of top 10 places were filled by triathletes exiting the water in the first pack. Of the male swimmers that exited the water in the first pack, 47% of these placed in the top 10. By comparison, if all 10 races were analysed, 65% (65/100) of the top 10 places were filled by pack1 swimmers. Of these swimmers that exited in the first pack, 21% (65/305) placed in the top 10.

Of the ten female races studied, 24% (24/100) of top 10 places were filled by triathletes exiting the water in the first pack of swimmers. Thirty-nine female triathletes in 10 races had been in the first pack out of the water and, of these, 24 (61%) finished the race in a top 10 position.

Across the 10 male races, 90% of winners came from pack1 swimmers. In the female events, 70% of winners had exited the water with the first pack of swimmers.

Table 4.3 Totals and percentages of the first pack of triathletes to alight from the water and the relationship to final finishing position. Sample n races Total n Total n % triathletes Total n top 10 % top 10 % pack1 in starters triathletes in in pack1 places filled by places filled top 10 pack1 pack1 by pack1 Male 10 605 305 50 65 65 21 Male A 4 208 17 8 8 20 47 Female 10 479 39 8 24 24 61

Note: % = percent n = number Male A = male sample including only those which have pack1 swimmers < 10

Figure 4.3 Number of triathletes in the first pack of swimmers from the total number of starters for all male and female events.

700 600 500

400 Starters 300 Pack 1 atheltes (n) o.

N 200 100 0 Male Female

Figure 4.4 Total number of top 10 finishers from those triathletes who exited the water in the first pack for both male and female.

350 300 250

200 Pack 1 150 Top 10 athletes (n) o.

N 100 50 0 Male Female

Discussion In events where all competitors exited the water together, there are no differences between swim performances. Hence, there can be no relationship with overall triathlon performances. It highlights the point that, if a break can not be made during the swim, then it comes down to the final run to determine the race outcome. This is because little positional change takes place during the cycle leg possibly due to drafting. As a group, cyclists who are drafting can maintain higher velocities for a similar workload due to an 64 air resistance reduction of up to 30% (Faria, 1992). However, it could be possible that other factors, such as position in the cycle or swim pack, play a role in the final outcome. Those who just 'sit in' the pack and do not contribute to the pace of the group by taking a 'turn' at the front will conserve more energy during the cycle, which then can be utilised during the run for a faster time (Hue et al., 1998; Lehénaff et al., 1998). Those who dismount earlier might create an advantage with a clear transition and the ability to start the run first. In a large pack of 50 triathletes, there may be 30 s between those who dismount at the front of the pack and those who are on the back of the same pack. This time difference could be increased during the transition as those dismounting last must negotiate their paths around the other triathletes throughout the transition.

When analysing the 10 race male sample, it was possible that 100% of the top 10 places could be filled by pack1 swimmers (305/100). When removing the events where there were greater than 10 pack1 swimmers, it is obvious that there will be fewer competitors to fill the top 10 places. Therefore, this will contribute a lower percentage (20% vs 65%). In this case, a maximum of 42.5% (17/40) of top 10 positions could be filled by triathletes exiting the water in the first pack. However, there is a marked increase in the percentage of the pack1 swimmers placing in the top 10 (47% vs 21%) when the initial pack of swimmers is less than 10 competitors. Hence, if pack1 has less than 10 triathletes, there is a two to three times greater chance that a member of that group will finish in the top 10.

The female events had smaller numbers of competitors and smaller groups of swimmers exiting the water in the first pack (8%) when compared with the total field of male triathletes (22%). Again, it is logical that the results show a smaller percentage of top 10 places filled by female, pack1 swimmers (24%) and to be less than the male competitors (65% for the 10 race sample). However, the female sample (61%) had a greater percentage of pack1 swimmers finishing in the top 10 when compared with the 21% of males. When this is considered in conjunction with the male sample that contains smaller leading packs (47% pack1 swimmers in top 10), it is similar to the female results (61%). The smaller the first pack to exit the water, the greater the chance of those triathletes finishing in the top 10 for both males and females. Thus, it could be said that if the first pack of swimmers is less than 10 competitors, they have 50% 65 chance of finishing in the top 10. This is also strengthened by the finding that 70% of female winners and 90% of male winners exited the water with the first group of swimmers.

These results did not agree with previous research which suggested that the swim section was not related to finishing time (Dengel et al., 1989; Landers, 1998; Sleivert & Wenger, 1993); or that all three disciplines carry the same weight (Landers et al., 2000; Zinkgraph et al., 1986). However, these prior investigations usually have only examined non-drafting events where the formation of packs during the cycle leg was not permitted and would not have influenced the final outcome. It could be inferred that the lack of correlation between swim time and total time is due to the relatively small amount of time spent swimming (17%) compared with cycling (52%) and running (30%) (Landers, 1998). However, considering the race position of an athlete at the end of each section rather than the total time can remove some of the bias of the amount of time spent in each discipline. Landers et al. (2000) used rankings for this purpose, but did not take into account the formation of packs. The run was found to play a greater role in final outcome than previously reported. Also, these previous studies have only considered single events. They have not allowed for other factors which could play a role in race outcomes over a number of events such as course design and environmental conditions, athlete preparation and training experience.

The results do not unequivocally demonstrate that, by exiting the water in pack1, there is a greater chance of winning or placing in the top 10. It could be that the better overall triathletes who can win, will be among the first out of the water to form pack1. However, whilst not conclusive, the results do suggest that superior swimming ability, as indicated via finishing position, is an important factor in determining the final outcome of a race.

Summary These results tend to confirm the subjective views of triathlon coaches, who believe that the swim leg of the triathlon is important. This is especially so since the introduction of draft legal events. In 90% of elite male races and 70% of elite female races, the 66 eventual winner exited the water in the first pack. Results also showed that, if a small group of less than 10 swimmers do break away during the swim leg, the possibility of these triathletes finishing in the top 10 is even greater (47% male & 61% female). This is compared to the top 10 finishing prospects of triathletes who exit the water in larger packs (21% male). It was found that 50% of male and 8% of female starters exited the water with the first pack of swimmers.

These results have training implications for both swimming and running. A triathlete must develop swimming ability to be able to maintain a pace fast enough to swim with the front pack without expending excessive energy which reduces subsequent cycling and running performance. There is also a need to develop early speed and lactate tolerance to start the swim hard and be positioned with the lead pack from the outset. Both aspects can be worked on relatively easily through anaerobic threshold/lactate tolerance traditional swimming training. Finally, the ability to run well 'off the bike' in order to improve a top 10 position into winning, is an area that is not well understood and invites further investigation. 67

Chapter 4B

Study 2B Relative contribution of each discipline and the effect on Triathlon Outcome

Methodology Subjects The subjects recruited in this study competed in 10 ITU events held during the 1999 season (Appendix H). Summation of split times during the Mexican triathlon for both male and female events, did not equate with the supplied total time and were excluded from this second phase of the study. Athletes who did not record full results or did not finish the event were removed from subsequent analysis.

Data analysis The results were analysed in a number of ways using Mathmatica (2000) to perform the large number of calculations. Firstly, standard deviations from the mean of each discipline were computed to determine where the greatest amount of variation existed within the classic distance triathlon. This was followed by calculating what percentage of the total discipline time for which the standard deviation accounted. This was done to ascertain whether a greater standard deviation would exist during the cycle discipline due to the greater time triathletes spent cycling compared with the other two legs.

Correlations were established between finishing order for the whole triathlon and the rank order that only included either the swim time, cycle time or run time. That is, it did not correlate between final position and the position at the end of the cycle leg. The position of the winner was ascertained after the swim for both male and female events. This procedure was also carried out for the top five and top 10 finishers in each triathlon, and the mean positions were reported. 68

In order to confirm findings of the previous study (2A), numerous calculations were conducted via Mathmatica to analyse groupings more closely. It had already been established (2A) that, in order to improve ones chance of winning an event, it was an advantage to be in the first pack out of the water. The following procedures set out to determine the importance of being in the first pack out of the water on final finishing position. A criterion measure of 5 s separation time had been used in the first instance because 5 s during the cycle equates to approximately 50 m of distance where no draft advantage can be gained. This was deemed a significant gap between athletes to cause them to cycle in different packs. In this study, the variations created by changing the separation time from 2 s to 6 s (ie 2, 3, 4, 5 & 6 s) on pack groupings were analysed. The size and number of the groups were calculated for each event and averaged for each separation condition.

Results Descriptive data & correlations The average of the standard deviations for all 10 events were calculated for both men and women in all three disciplines (Table 4.4). The greatest raw variation was noted for the run discipline followed by the cycle and swim. The percentage of the total time that the standard deviation represented was then calculated. This equated to a higher percentage for both males and females in the swim and run, than in the cycle (Table 4.4).

Table 4.4 Means, standard deviation and standard deviation percentage for male and female triathletes across 10 events (time in seconds). Swim Cycle Run mean sd sd % mean Sd sd % mean sd sd % Male 1082.0 16.06 1.48 3565.4 34.76 0.97 1929.3 56.05 2.9 Female 1136.7 42.43 3.73 3956.8 49.70 1.26 2216.7 79.88 3.6

A rank order correlation was undertaken between finishing order for the whole triathlon and a finishing order that only included either the swim time, cycle time or run time in order to relate these draft legal results to previous non-drafting data. Both the male and 69 female triathletes showed a small common variance between swim time and final time, and a strong correlation between the final time and run time. Cycling was in between these variables (Table 4.5).

Table 4.5 Rank order correlation of final finishing position with the time of each of the swim, cycle and run disciplines. n Swim Cycle Run male 605 0.49 0.67 0.86 female 479 0.39 0.67 0.85

Winner position The position of the winner was calculated after the swim and cycle for both male and female events. On average, the male winner placed 10th after the swim, and 6th after the cycle. Males who finished in the top five were placed 13th and 7th on average, after the swim and cycle, respectively. A top 10 male finisher averaged 15th out of the water and 10th after the cycle (Table 4.6).

On average, the female winner was placed 9th after the swim, and 5th after the cycle. A top five female finisher placed 11th and 7th, after the swim and cycle, respectively. A top 10 finisher averaged 13th out of the water and 9th after the cycle (Table 4.6, Figure 4.5 & 4.6).

Table 4.6 The average, mean position of triathletes at the end of the swim and cycle for the winners, top 5 and top 10 finishers. Male Female end swim end cycle end run end swim end cycle end run Winner 10 6 1 9 5 1 top 5 13 7 3 11 7 3 top 10 15 10 5 13 9 5

Figure 4.5 Average male placing after swim cycle & run for the winner, top 5 & top 10.

16 14 12 10 Wi nner 8 Top 5 6 Top 10 4 2 0 Swim Cycle Run

Figure 4.6 Average female placing after swim cycle & run for the winner, top 5 & top 10.

14 12 10 8 Winner Top 5 6 placing Top 10 4 2 0 Swim Cycle Run

Groupings Varying the separation for selection into packs of values between 2 s and 6 s also did not result in significant changes in pack size or composition. A 4 s separation time was chosen for subsequent analysis as it is the median of 2 s to 6 s. It is also a reasonable time difference to allow separation of triathletes. Using a separation of 4 s, the average number of groups after the swim and cycle, were 8.0 and 8.3, respectively. This increased to 35.3 groups following the run in the male events with 4 s (Figure 4.7). A similar pattern emerged from the female triathletes where the average number of groups after the swim and cycle, were 10.2 and 8.6, respectively. Again, this increased to 33.8 groups following the run (Figure 4.8). Complete pack data can be viewed in Appendix I.

Figure 4.7 Male average pack number and size at end swim, cycle and run.

20 18 16 14 12 swim cycle run 10 8 frequency (n) 6 4 2 0 1 2 3 4 5 6 7 8 9 101112131415161718192021222324252627282930313233343536 pack no.

Figure 4.8 Female average pack number and size at end of swim, cycle and run.

16 14 swim cycle run 12 10 8

frequency (n) 6 4 2 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 pack no.

A separation distance of 4 s to define groupings at the end of each leg was used to see if membership of a particular "smallish" group out of the swim was critical in winning the event or at least finishing in the leading places. This could demonstrate the importance of the swim, since, ordering established during this phase would have been shown to persist throughout the remainder of the triathlon. This did occur in the Hungarian triathlon, however, for all others this effect was not observed. Even in the Hungarian event, the effect became less apparent when viewing the top 10 finishers. For example, the triathlons in Belgium, Canada (Montreal) and the USA all had the top 10 finishers within one group. However, this was not significant given that the three groups were not "smallish" and consisted of 94 %, 65 % and 96 % of the total number of competitors, respectively. In the triathlons where smaller groups emerged after the swim, the effect was not noticeable as the top 10 finishers invariably were spread over many groups. For example, the top 10 finishers in the Canadian (Corner Brook) and Australian triathlons were spread out over six different groups out of the water. In the Hungarian event, when one considers the top 10 finishers instead of the top five, the athletes were spread over three groups 73

Discussion Descriptive data & correlations Considering measures of dispersion (ie standard deviation) highlights that the greatest variation during a triathlon occurs during the run. It is not known if this is due to greater levels of fatigue at this stage in the race, but there is a tendancy for an increasing SD throughout the event for both men and women. This was also noted by Landers (1998) following an investigation of just one event, the 1997 TWC. However, despite the SD spread revealing a greater spread of scores during the cycle than the swim, when the SD is expressed as a percentage of the time spent in each discipline, the situation changes. The swim appears to play a greater role in determining the eventual winner in female events than in male events. However, both show higher variation during the run and low difference in times during the cycle. That is, there is a larger percentage variation in run and swim times than cycle times.

There is a greater percentage variability of the women's times compared with those of the men and also greater raw variability in the run times when compared with the other legs for both genders. Fatigue is possibly a factor, but it could be that one can make up more places in the run and emphasises its importance upon which to concentrate one's training. This suggestion is further confirmed by the significant correlation existing between run time and final time. This suggests that the relationship between competitors’ running ability, measured by run time, is a better predictor of triathlon performance than swimming or cycling ability. These results differ from previous findings of non-drafting events (Dengel et al., 1989; Landers, 1998; Sleivert & Wenger, 1993). In addition, cycling time is a better predictor of final event time than is swimming. This does not take into account the effects of groupings, but only reflects that those who run faster usually have a faster total time. It could be that the better triathletes are those who can be at the front of the pack at the end of the cycle and have a faster transition. Thus, a faster cycle time may be the result of better positioning at the end of the cycle rather than cycling speed. There was a small standard deviation during the ride of 34 s and 50 s, for males and females, respectively. One standard deviation accounts for 34% of the competitors. These triathletes then are leading 74 at the commencement of the run and, psychologically, could have the advantage over the chasing athletes and run faster. The factor of fatigue on final outcome requires further clarification. Another explanation could be that the better triathletes tend to show their superiority more towards the end of an endurance event simply because of better staying power. That is, the faster triathletes are those who still have more fuel in the tank for the run. This could be a result of drafting during the cycle section having been utilised better by some than others. It could be that all triathletes have the ability to run the same time for 10 km, but it is those who are fresher at the start of the run, or who get into a rhythm first, who will have the faster run after a swim and ride.

Hausswirth et al. (1999b & c) showed that the physiology and biomechanics of running at the end of a triathlon and marathon are similar. Hence, it could be concluded that this correlation of run time with final time is more related to the overall endurance effects rather than skilled running. This again highlights the importance of the run phase of triathlon and the need to be able to run effectively after prior activity (swim and cycle).

Winners position Although the winners position is worthy of note (Table 4.6) it does not account for packs developed during the swim and cycle. As was noted in the previous study (2A), some of the groups included the entire field at the end of the swim. However, it does reveal two things. It is advantageous to be in the first pack exiting the water and, even if the initial pack is large, it is still important to attain a good position out of the water. This is possibly related to the tactical consideration of having a clear transition (Millet & Vleck, 2000) and ensuring a good position during the cycle. Also, it highlights that one does not have to be the best swimmer on the day, but placed in the first exiting group. Again, it is hard to determine if this position is due to the lesser swimming ability, or if it could be a successful tactic were the eventual winner to possibly draft and conserve energy during the swim. Drafting has been shown to decrease 400 m swim time by 3.2% (Chatard et. al, 1998).

Groupings It had already been established previously that, in order to have a chance of winning an event, it is advantageous to be in the first pack out of the water (Study 2A). This may have been influenced by the size of some of these first packs. While not highlighting the importance of packs, it does indicate the influence of draft legal events where large packs are formed during the swim and cycle. Therefore, only the final run leg, where drafting has minimal impact on performance, remains to rank athletes into finishing positions.

The most striking aspect of analysing the groups that formed, were the amount of grouping in the swim, and especially the cycle leg, in comparison with the run leg. This confirms the expectation of the effect of drafting on the event. For example, using a separation of 4 s, the average number of groups after the swim and cycle, were 8.0 and 8.3, respectively. However, this increased to 35.3 groups following the run. The small number of groups after the cycle leg and the large size of one or two of the most dominant suggests that it is important to expend less energy than one’s competitors in the first two legs than gaining a significant time advantage. In turn, this tends to suggest that the final finishing order is determined more by the run than the swim or the cycle. Hence, data support the previous calculations in study 2A concerning spread and correlation calculations associated with each of the three stages. Although perhaps less important than the run, the first two stages remain important as noted by calculations showing the top finishers moving sequentially through the field from phase to phase (Table 4.6).

Table 4.6 presented the average position of the winner, top five and top 10 place getters after the swim and cycle. However, these were averages and it remains possible for a superior runner to move through the field into the top places. For example, in the male events, one winner was 33rd after the swim and another was 22nd after the cycle. However, the winner was still within the first pack after the swim. A top five finisher had been 33rd after the cycle whilst there had been a top 10 finisher who was 43rd after the swim and 39th after the cycle. This also occurred in the female events where one winner had been 25th out of the swim and another 17th after the cycle. A top five finisher had been 52nd after the swim and another 27th after the cycle. 76

For athletes wanting to improve their finishing positions it is probably advisable to swim well enough to be within striking distance of joining onto a big cycling pack and cycle well enough to stay in the main pack to maximise drafting and optimise energy efficiency. The need to improve one's running for the final effort after two other disciplines, irrespective of skill, is paramount. The minimum swimming and cycling performances required to do this could be more reliably ascertained by studying more triathlons and more closely considering the conditions under which they were staged. In terms of talent identification, an athlete who is "just" a good swimmer and cyclist but an outstanding runner will probably be a better prospect than someone who is a good all rounder. A similar story emerged in the women’s triathlons where there were also no significant correlations between positions and the leading placegetters were spread evenly amongst the packs. Using a separation distance of 4 s, the average number of groups after the swim and cycle, were 10.2 and 8.6, respectively. This increased to 33.8 groups following the run.

In turn, this tends to suggest that the final finishing order is determined more by the run than either the swim or the cycle and supports earlier calculations concerning spread and correlation calculations associated with each of the stages. Although the first two stages perhaps are less important than the run, they are still not insignificant. This is indicated by calculations showing the top finishers moving evenly through the field from stage to stage.

Summary The greatest amount of variation in performance is found in running times when examining both the raw SD and the percentage SD. The swim leg appears to be the second most important discipline based on findings from study 2A and the large variation percentage SD in swim times. There are few changes in position during the cycle leg, hence, the importance of being in the first pack to exit the water and being able run quickly off the bike is increased. 77

Chapter 5

Studies 3, 4 & 5 Cadence selection and performance

Abstract Cycling has been shown to alter running physiology during a triathlon but biomechanical changes that might occur have not been quantified. This chapter examined cycling cadence, running stride rate (SR) and stride lengths (SL) used in competition by male and female triathlon competitors. These factors were then related to triathlon performance.

The chapter is sub-divided into three sections for clarity but with a common introduction and methods section. Study 3 reports cycling cadence and SR data of age group triathletes. The sample consisted of 20 male and 10 female finishers of an Australian selection race for the 2000 Triathlon World Championships (TWC). Similar data for senior elite male (n=51) and female (n=46) triathletes are presented in study 4. Finally, study 5 reports on the elite data in order to relate cycling cadence, SR and SL to performance.

Results indicated that significantly higher cycle cadences occurred at the end of the cycle when compared with the start, and higher run SRs occurred at the beginning of the run for both male and female age group competitors when compared with the mean. The initial SR was similar to the initial cycling cadence for both male and females, and the faster triathletes recorded higher cadences and SRs. Senior elite male triathletes used higher cadences, SRs and longer SLs than the senior elite female triathletes. Again, the selected cadence during the cycle was comparable with that used during the run for both males and females. Female triathletes used a significantly faster cadence at the commencement and completion of the cycle than during the middle stages.

Variations in running SLs appear also to be associated with running ability and triathlon performance. That is, the better performers maintained a longer and more consistent SL. 79

Introduction The classic distance triathlon sequentially combines a 1500 m swim, 40 km cycle and 10 km run linked together by two transitions. Most of the available triathlon data are of the physiological aspects of triathletes. Many authors have found that an endurance oriented physiological profile plays an important role in performance (De Vito et al., 1995; Dengel et al., 1989; Khort et al., 1987; Laurenson et al., 1993; Miura et al., 1997; O’Toole et al., 1987, 1989a, 1989b; Schneider et al., 1990; Sleivert & Rowlands, 1996; Sleivert & Wenger, 1993; Zhou et al., 1997). However, it should be noted that, in the shorter triathlon events, up to and including the classic distance triathlon, there is a tendancy for the physiological variables to correlate with the swim and run only (Sleivert & Wenger, 1993).

Landers et al. (2000) and Ackland et al. (2000) highlighted the morphological characteristics of elite triathletes and their roles in performance. Although absolute body size did not appear to be a determinant of triathlon performance, the limb length proportions did contribute (Landers et al., 2000). Also, no common variance was found between any of the anthropometric variables and cycle performance (Landers, 1998).

Landers et al. (2001) reported that triathletes who alighted from the water in the first pack of the swim enhanced their probability of finishing in the top 10, and that the winner will come from this pack 80% of the time. This study demonstrated the importance of swimming and being able to run well following a cycle.

A comparison of the anthropometric data (Landers et al., 2000) and pack formation results from the swim with race outcomes (Landers et al., 2001b) revealed the decreased importance of the cycle discipline for elite competitors where drafting is permitted. Landers et al. (2000) showed no significant correlation between cycle time and body size or shape, whereas the swim and run were related to proportional lever lengths and sum of skinfolds. Landers et al. (2001b) determined that triathletes exiting the swim in the first pack had a greater chance of winning, and that the swimming and running abilities of triathletes were highly related to final finishing position.

Cadence determination, selection and race performance Prior exercise in the form of cycling and/or swimming causes a decrease in endurance capabilities which, in turn, increases the energy cost of running at the end of a triathlon (De Vito et al., 1995; Guezennec et al., 1996; Hausswirth et al., 1996 & 1997; Hue et al., 1998; Kreider et al., 1988b).

A decreased SL is generally coupled with an increase in SR when observed during the initial stage of the run during a triathlon and could be the result of prior cycling (Hausswirth et al., 1996 & 1997). This might be due to the cycling action causing some fatigue of the biceps femoris muscles which are important in both cycling and running (Hausswirth et al., 1997). This would support the work of Lehénaff et al. (1998) who found less decrement in run performance following a drafting cycle. Drafting decreases energy cost by up to 30% and should decrease fatigue (Faria, 1992). Hue et al. (1998), Marino and Goegan (1993) and Quigley and Richards (1996) found no change in run SR at the commencement of the run. On the other hand, Witt (1993) found a decrease in SR immediately following cycling. Thus, the cadence used during cycling might influence the adoption of a similar cadence at the transition of the cycle to the run.

Generally, increased fatigue during running results in a decreased SL and maintenance of SR. Hence to maintain velocity, a corresponding increase in SR is required (Ackland, 1997; Elliott et al., 1981; Elliott & Ackland, 1980; Elliott & Roberts, 1980; Hausswirth et al., 1997, Williams et al., 1991). This also has been shown in running mechanics following cycling (Hausswirth et al., 1996 & 1997). However, Quigley and Richards (1996) and Hue et al. (1998) reported no change in running biomechanics with prior cycling. Gohlitz and Witt (1993) found a decrease in running SR immediately following cycling but this stabilised after 1000 m at the subject’s normal rate. However, Quigley and Richards (1996) analysed running mechanics by analysing 10 successful 40 m runs at race pace through a testing area. Because it was intermittent running and provided a recovery period, this study might not give a true indication of the alterations with continuous running. Cavanagh and Williams (1982) found that runners typically chose within 4.2 cm of the most efficient SL by altering the SR/SL combinations for each change in velocity. Elliott et al. (1981) showed that SL/SR could be maintained with either continuous or intermittent running training.

The first part of this study sought to determine self selected cycling and running cadences of triathletes during competition. Only then can triathlon performance be related to the cadences used throughout the race and an understanding of how the cadences of each discipline relate to each other. 82

Study 3 Cadence determination, selection and performance of age group triathletes

Methodology Sample Subjects included both male and female high performance age group triathletes from within Australia. These athletes competed in the first national series race, 1999/2000 in Perth, Western Australia. The race was used for selection into the national team for the 2000 TWC. Competitors in the junior (16-19 y), 20-24 y and 25-29 y age groups were used as subjects as this corresponds to the average age of junior and senior, elite triathletes (Ackland et al., 1998). Total subject numbers included the top 20 male and top 10 female triathlete finishers, with complete data for 15 males and nine females. In some instances, triathletes were running side by side and screened other competitors from the camera field of view. The male sample was larger due to the greater numbers of high level male triathletes compared with the female competitors.

The sample of age-group triathletes were chosen initially as a wider range of performances would be expected which perhaps would provide a greater insight into the relationship between cycling cadence, running SR and performance, and preparation for further research with elite triathletes.

Testing protocol Prior to the race, each subject was marked with a number and letter code to aid identification. From videotapes of the competition, individual self-selected cadences in each of swimming, cycling and running under race conditions were ascertained. Unfortunately, it was not possible to get close enough to the swimmers to determine the leg kick cadence during this discipline. Two cameras, recording at 25 frames per second, were placed in relatively flat sections of the course during the cycle and run phases in order to minimise the effect of terrain on the choice of cadence.

The event constituted the classic distance triathlon and included a one-lap 1500 m swim in a fresh water lake with an exit and re-entry at approximately 1200m. This was 83 followed by a one lap, 40 km cycle over undulating roads, including six 90o left hand turns, one 90o right hand turn and one 180o right hand U-turn. Drafting was not permitted during the cycle leg. Finally, the triathlon finished with a 10 km run, conducted over a three lap, undulating graded dirt course.

During the cycle, a camera was placed 1000 m from the transition area (c1). This allowed determination of cadence near the start (1 km) (c1) and at the end (39 km) (c2) of the cycle. On the run course, a camera was placed 130 m from the cycle to run transition (r1), to enable determination of cadence following the end of the cycle leg. A second camera was located at the 3.7 km point into the 10 km run (r2). Then, the first camera was relocated to 50 m from the finish (r3). A 30 m section was marked on the run course using cones to designate the data collection area. As each runner uses a SR of approximately 90 rpm, approximately 15 foot strikes should be available for examination inside the 30 m of marked ground.

Data analysis After the event, the video was replayed and the cadence for each triathlete was calculated at each of the points previously described during the cycle and run. Cadence was determined as "revolutions per minute" (rpm). During the bike leg, this was the number of times one leg completed a full revolution in one minute. When running, the number of strides (or half the number of steps) per minute were counted from the number of times the left foot contacted the ground.

Data were entered onto an SPSS spreadsheet where the means and standard deviations for each collection point were determined for males, females and combined groupings. These were graphed against time. A two-way repeated measures ANOVA was conducted to determine whether any significant differences existed between the selected cadences in each discipline, or throughout each discipline. A Tukey’s HSD post hoc test was used to isolate significant differences. Due to the small sample size and multiple comparisons, the level of significance was set at p < 0.01.

Split times for each discipline and total time were recorded for use as performance variables. Cadences for each discipline then were correlated with the corresponding discipline split time using Pearson product moment correlation.

Results Performance times All discipline times and total times for males, females and all triathletes combined are in Table 5.1. The male triathletes were significantly faster than their female counterparts in all disciplines and total times. Analysis of the raw data (Appendix J) highlights this gender difference as only one woman was a faster swimmer than three males, and a second woman swam faster than one male.

Table 5.1 Average split times (h:m:s) and total times (h:m:s) for male female and combined triathletes. All (n = 24) Male (n = 15) Female ( n = 9) Gender difference mean sd mean sd mean sd p sig Swim 0:20:50 0:01:48 0:19:49 0:01:17 0:22:32 0:01:08 0.0001 ** Cycle 1:07:59 0:03:23 1:05:39 0:01:35 1:11:52 0:01:16 0.0001 ** Run 0:40:55 0:03:21 0:38:41 0:01:13 0:44:38 0:02:16 0.0001 ** Total 2:09:42 0:07:46 2:04:10 0:02:29 2:18:56 0:03:10 0.0001 ** Note: ** = significant difference between male and female competitors p < 0.01 sd = standard deviation h:m:s = hours, minutes, seconds

Cadence The mean (+ sd) cadence used at each collection point during the triathlon for males, females, and combined males and females is presented in Figure 5.1 and Table 5.2.

Male and female triathletes combined Male and female triathletes used significantly different cadences at the end of the cycle discipline (c2) and at the 3.7 km point into the run (r2) leg. Males recorded a significantly higher c2 cadence and significantly smaller r2 cadence than the females.

Overall, the male competitors used a significantly higher average cadence during the cycle ((c1+c2)/2 = 109.2 Vs 98.8, p < 0.05) and lower average SR than the females. However, males and females were not significantly different during the run portion of the triathlon ((r1+r2+r3)/3 = 90.1 Vs 93.6). 85

Figure 5.1 Cadence employed by males, females, and combined males and females during a triathlon competition.

Cadence variation during a triathlon

130

120

110

100 male ce

en female

cad 90 combined

60 c1 c2 r1 r2 r3 time

Table 5.2 Cadence (rpm) used throughout the cycle and run portions of a triathlon for all competitors, and male and female competitors separately. All (n = 24) Male (n =15) Female (n = 9) Gender difference Cadence sd Cadence sd Cadence sd p sig c1 94.1 6.1 95.8 6.3 91.2 4.6 0.071 c2 116.4 12.0 122.5a 5.2 106.3a 13.4 0.0001 ** c average 109.2 14.7 98.8 12.4 r1 93.3 4.1 92.1b 3.8 95.2b 4.1 0.082 r2 91.9 6.1 88.4abc 4.3 97.8a 3.4 0.0001 ** r3 89.0 3.4 89.9ab 4.4 87.9bcd 2.4 0.189 r average 90.1 4.2 93.6 5.4 Note: c1 = cadence at 1 km into cycle ** = significant difference between male and c2 = cadence at 39 km into cycle female competitors at p < 0.01 r1 = cadence at 130 m into run a = significantly different from b1 (p < 0.01) r2 = cadence at 3.7 km into run b = significantly different from b2 (p < 0.01) r3 = cadence at 9.95 km into run c = significantly different from r1 (p < 0.01) c average = (b1 + b2)/2 d = significantly different from r2 (p < 0.01) r average = (r1 + r2 + r3)/3 sd = standard deviation rpm = revolutions (strides) per minute

Male triathletes The cadence at the beginning (c1) and end (c2) of the cycle was significantly greater than each of the final two running cadences measured (r2 & r3). The cycle cadence at the end of this discipline (c2) was also significantly higher than that used during the initial stages of the cycle (c1) and early run phase (r1) during the triathlon. However, c1 and r1 were not significantly different. The SR at the commencement of the run (r1) 86 was significantly faster than that used 3.6 km later (r2) and at the end of the race (r3). The SR for r2 and r3 were the same.

Female triathletes No difference was noted between c1 and either of r1 or r3 for the female triathletes. The second cadence measured during the run (r2) was significantly greater than that used at the start of the cycle discipline (c1). Pedalling frequency used at the end of the cycle leg (c2) was significantly greater than the initial (c1) pedalling rate and the initial (r1) and final (r3) SRs during the run. No significant differences were found between c2 and r2, or between r1 and r2. The two measures of initial run cadence (r1 and r2) were significantly greater than the cadence used at the end of the run (r3).

Relationship with performance Combining all triathletes allows a typical bell shaped curve to be developed by having an even spread of scores across the means of each measure of time (Table 5.3). Therefore, one can investigate the relationship between these two variables.

Table 5.3 Level of skewness and kurtosis of split and total times for both male and female triathletes combined. Skewness Kurtosis Swim -0.007 -0.981 Cycle 0.323 -1.134 Run 0.733 0.674 Total 0.498 -1.383

Typically, the faster swimmers had higher cadences at the end of the cycle leg (c2) and lower running SRs during the mid point of the run (r2) (Table 5.4). It was found that triathletes with a faster cycle split tended to have a higher cadence at c1 and c2, and had a lower initial running SR (r1 & r2).

Superior runners tended to have lower running cadences at the beginning of the run (r1 & r2) and higher pedalling frequencies at the end of the cycle (c2). A significant and positive correlation was found between r1 and r2, and a significant but negative correlation found between c2 and r2. Faster finishers had a higher c2, lower r1 and lower r2 cadences. The male competitors showed a positive correlation between swim time and initial cycle cadence (i.e., faster swimmers had a lower initial cycle cadence) 87 and a significant positive correlation between r2 and r3. The female triathletes recorded significant positive correlations between cycle time and r2. However, significant, negative correlations were noted between run time and r3, and between total time and r3.

Table 5.4 Correlation matrix: relationships between cadence and times for all, male and female triathletes. c1 c2 r1 r2 r3 swim time cycle time run time c2 All 0.225 Male 0.055 Female -0.139 r1 All 0.035 -0.191 Male 0.230 -0.199 Female 0.139 0.257 r2 All -0.165 -0.570** 0.490 Male 0.212 -0.218 0.420 Female 0.170 -0.086 0.233 r3 All 0.045 0.089 0.148 -0.010 Male -0.014 -0.162 0.470 0.535 Female -0.245 -0.151 -0.183 -0.443 swim time All 0.113 -0.633** 0.253 0.680** 0.002 Male 0.660** 0.033 0.126 0.327 0.309 Female 0.565 -0.580 -0.304 0.107 0.329 cycle time All -0.438 -0.552 0.418* 0.755** -0.332 0.595** Male -0.273 0.254 0.127 0.036 -0.093 -0.193 Female -0.181 0.138 0.429 0.789 -0.644 -0.446 run time All -0.353 -0.588** 0.453 0.743** -0.380 0.593** 0.875** Male -0.114 0.154 0.036 0.277 -0.079 0.018 0.285 Female 0.003 -0.075 0.545 0.257 -0.747 -0.392 0.644 total time All -0.313 -0.646** 0.438 0.810** -0.314 0.746** 0.951** 0.953** Male 0.114 0.258 0.167 0.333 0.063 0.410 0.687** 0.689** Female 0.155 -0.254 0.476 0.611 -0.739 -0.085 0.764 0.894**

Note: c1 = cadence at 1 km into cycle c2 = cadence at 39 km into cycle r1 = cadence at 130 m into run r2 = cadence at 3.7 km into run r3 = cadence at 9.95 km into run ** = significant correlation between variables at p < 0.01

Discussion Both male and female triathletes showed a similar r1 cadence to c1 cadence and a significantly higher end cycle (c2) cadence than initial cycle (c1) or run (r1) cadence. More measurements of pedalling frequency are needed to determine more accurately the average cadence used throughout the cycle discipline. It could be that the initial cycle cadence is maintained for up to 90% of the course. Hence, this self-selected rate is the typical pedalling frequency used during training and has a carry-over training effect on the runners’ competition SR. The higher cadence at the end of the cycle leg (c2) might be the result of several factors. It has been proposed that gearing down and "spinning" (i.e. increasing cadence) at the end of the cycle leg helps to increase blood flow to the muscles (Gotshall et al., 1996). This could improve removal of waste products and reduce the force demands on the musculature (Takaishi et al., 1996 & 1998), thereby enabling an easier transition into the run. A second possibility could be that, when approaching the final transition, triathletes ‘jockey’ for position and increase their cadences in an attempt to increase velocity and reach the transition area at the front of the pack.

For the male triathletes, r1 was found to be significantly greater than r2 and r3, but there were no differences between r2 and r3 SRs. The female competitors revealed two initial SRs which were not significantly different from each other but were significantly higher than the SR used at r3. The elevated SR in the early stages of the run could be the result of fatigue causing a shorter SL (Hausswirth et al., 1997). Thus, to compensate for the resultant speed loss, an increased cadence is adopted. It also could be that the initial run SR is influenced by the cycling pedalling frequency. Perhaps this is why similar cadences were recorded at the start of the cycle and the start of the run. The higher cadences at the end of the cycle leg possibly influenced the choice of higher run cadence at the commencement of the third discipline.

Previous studies have shown that SR decreases with fatigue (Ackland, 1997; Elliott et al., 1981; Elliott and Ackland, 1980; Elliott and Roberts, 1980; Hausswirth et al., 1997, Williams et al., 1991). In the case of the female triathletes, this could have occurred because there was a significant decline in cadence from 97.8 (r2) to 87.9 (r3) strides per minute (~ 10% decrease) at the end of the run. Decreased SRs occurred for males 89 within the first third of the run and, although significant, were not as pronounced. The SR decreased from 92.1 per minute (r1), to 88.4 per minute (r2) (~ 4% decrease). This also could suggest that the cycle cadence plays a role in the initial elevation of run cadence. However, in the faster male competitors, a "normal" running cadence is resumed earlier than for the females.

Relationship with performance Typically, the faster swimmers had higher cycling cadences at the end of the cycle leg (c2) and lower running SRs during the mid point of the run (r2).

Triathletes with a faster cycle split time tended to have a higher cadence at c1 and c2, and lower initial running cadences (r1 & r2) (p<0.05). This suggested that the faster triathletes were less affected by the cycle portion of the triathlon than those who were slower. It might mean that less fatigue occurred in the biceps femoris muscles during the cycle or that the process of "spinning" the legs at the end of the cycle had a positive effect on the triathlete leading into the transition for the run.

Faster runners tended to have lower running cadences at the beginning of the run (r1 & r2, p<0.05) and higher pedalling frequencies at the end of the cycle (c2, p<0.01). A significant and positive correlation was found between r1 and r2, and a significant negative correlation existed between c2 and r2. There were no significant differences between the cadence used at r1 and r2. Thus, the positive correlation between r1 and r2 strengthens this finding, and shows that those with lower cadences at the start of the run maintain a lower cadence 4 km later (r2). The negative relationship between c2 and r2, and the finding that faster runners were using lower cadences at the start of the run, also suggested that there could be a positive effect of "spinning" the legs at the end of the cycle. Slower cyclists also employed a similar cadence selection at the end of the cycle (c1) and during the run (r2).

Faster finishers had a higher c2, lower r1 and lower r2 cadences which summarises all of the above. The faster cyclists who used higher cadences at the end of the bike leg also tended to be the faster runners, and used lower cadences at the start of the run leg than the slower competitors. Whether the higher cycling cadences influenced the lower running SR cannot be ascertained because several other factors could be involved. For 90 example, the training effect could mean that the faster competitors were adapted better to the sport of triathlon. Hence, they could be affected less by the prior disciplines through either greater training volume and/or specificity have resulted in a learning effect or reduced fatigue following the cycle. It is possible that a higher O2max could delay the onset of fatigue or the superior performance might be related to an increase in skill level which enables superior economy.

The male competitors exhibited a positive correlation between swim time and initial cycle cadence (i.e. faster swimmers had a lower initial cycle cadence) and a significant positive correlation between r2 and r3 (i.e. faster swimmers used a higher run SR).

The female triathletes recorded significant, positive correlations between cycle time and r2. That is, the faster female competitors had a lower r2 SR. This is in agreement with other findings of this study. For example the females reported significantly greater r2 SRs than the male competitors and recorded slower times than the males. Significant, negative correlations were noted between run time and r3, and between total time and r3. That is, as time for the run increased, SR decreased (i.e. less distance per stride because of fatigue). These latter two correlations showed the importance of maintaining a cadence throughout the run discipline as r3 was significantly lower than both r1 and r2 for the female competitors. This decline in cadence was not coupled with an increased SL and the resultant decrease in velocity was probably due to fatigue. Further study is required to clarify the role that the cycle leg plays in influencing running mechanics and if the same pattern is utilised by the elite level triathletes.

Summary Previous research has shown that prior cycling has a physiological effect on a subsequent run caused by fatigue. Similarly, there could be a biomechanical effect such as body posture or stride length. This study sought to determine whether running SR was affected by prior cycling during a triathlon competition and the relationship cadence exerted on triathlon performance. The sample consisted of 20 male and 10 female finishers of an Australian selection race for the 2000 Triathlon World Championships. 91

Study 4 Cadence selection of senior elite triathletes at the 2000 Triathlon World Championships Based on “Cadence Selection and Performance” Proceedings of International Triathlon Coaching Symposium, July 23-24 2001, Edmonton, Canada, page 13.

Introduction Study four sought to determine self-selected cycling and running cadences of senior elite male and female triathletes during competition.

Methodology

Sample Subjects included 51 males and 46 female senior elite triathletes who competed and finished the ITU, Triathlon World Championships (TWC) in Perth, Australia, 2000.

Testing protocol Each competitor was marked with a number code to aid identification and was then video taped during competition to determine individually selected cadence in cycling and running, and SL during running under race conditions. One camera was placed in a relatively flat section during the cycle stage and another to record the run phases in order to minimise the effect of terrain on the choice of cadence (Figure 5.2).

The event was the classic distance triathlon commencing with the 1500 m, one lap swim in a salt water river after a dive start from a pontoon. Competitors were permitted wetsuits during the swim as water temperature was 18.8oC. This was followed by a six lap, 40 km cycle. Each cycle lap consisted of four 90o right hand turns, two 90o left hand turns and one 180o right hand U-turn. In addition, one large ascent and descent covering 50 vertical metres, and one smaller climb and descent of 10 vertical metres were included. Drafting was permitted during the cycle leg and the triathlon finished with a 10 km run conducted over a flat three and a half lap course.

Figure 5.2 2000 Triathlon World Championships camera locations on the cycle and run course.

Transition Camera 1

Camera 2

During the cycle, a camera, recording at 25 frames per second, was placed 1500 m from the transition area. Data were collected at eight points during the male 40 km cycle, namely 1.5 km (post-T1), 5.3 km (lap 1), 8.1 km (lap 2), 14.7 km (lap 3), 21.3 km (lap 4), 27.9 km (lap 5), 34.5 km (lap 6) and 38.3 km (pre-T2). During the women's cycle, only seven data points were collected for analysis because no recording occurred during lap 3 (14.7 km) due to equipment malfunction.

On the run course, a camera (25 Hz) was placed 1200 m from the cycle-to-run transition. This enabled determination of cadence at 1.2 km (lap 1), 3.0 km (lap 1A), 3.64 km (lap 2), 5.44 km (lap 2A), 6.08 km (lap 3), 7.88 km (lap 3A) and 8.52 km (lap 4). Due to an error on the part of the race organisers, the women triathletes only completed 7.8 km (2.5 laps) rather than the standard 10 km. This limited the data collection for the women to five points and could have skewed the results by the females misjudging their energy expenditure due to thinking they had further to run before completing 10 km and finished prematurely rather than the usual way.

Data analysis After the event, the video was replayed and cadences for each triathlete were calculated at each of the points described above from the cycle and run. Cadence was determined as "revolutions per minute" (rpm). During the cycle leg, this was the number of times one leg completed a full revolution in one minute. It was not possible to determine the cycle cadence for every triathlete at every point as, at times, they were obscured by other competitors due to the draft legal nature of the event. When running, the number 94 of strides (or half the number of steps), were counted each minute. The length of stride and instantaneous velocity was also calculated. The data collection used digitising via the computer program, Video Expert II Coach.

Table 5.5 Cycle cadence data collection points. Distance (km) Lap 1.50 post-T1 5.30 1 8.10 2 14.70 3* 21.30 4 27.90 5 34.50 6 38.30 pre-T2 * No female data collected at this point T = transition

Each data point during the run was coded either SR for stride rate or SL for stride length. The subsequent 1, 1A, 2, 2A, 3 or 3A explains during which lap the data were measured. SR1A refers to SR data collected at the end of the first lap of the run, whereas SL3 was the stride length measured at the commencement of lap 3. For full details, see Table 5.6.

Table 5.6 Data collection points during the run. Distance (km) Lap Stride rate Stride length 1.20 1 SR1 SL1 3.00 1A SR1A SL1A 3.64 2 SR2 SL2 5.44 2A SR2A SL2A 6.08 3 SR3 SL3 7.88 3A SR3A SL3A 8.52 4 SR4 SL4

Data were entered onto an SPSS spreadsheet and descriptive statistics for each collection point were determined for male and female subjects, and graphed against time. As it was deemed important, two-tailed t-tests (p < 0.01) were conducted to determine whether any significant differences existed between the initial two cadence measures (post-T1, lap1) and the final cadence measure (post-T1) for both male and female triathletes. This was also undertaken for SR and SL data. To improve the understanding of the results, competitors also were split into groups depending on their running and cycling ability. Cyclists were categorised on the basis of the pack in which they rode whereas, during the run component, competitors were grouped according to 95 their run time. Consequently, two-way repeated measure ANOVAs were undertaken to compare between groups. Tukey’s HSD post hoc test were employed to compare between time points, with significance set at p < 0.01. Male and female data were treated separately.

Split times, rankings for each discipline, and total times were recorded and used as performance variables. Cadences for each discipline were then correlated (Pearson Product Moment) with the corresponding discipline split time and total time. The SL was also correlated with run performance.

Results Figure 5.3 Male and female cycle cadences over time

101 99 97 Male 95 Female 93 cadence (rpm) 91 89 post- 123456pre- T1 T2 lap

NB: no female data collected during the 3rd lap due to equipment malfunction.

Female cycle results Twelve cycle data collection points were possible (Table 5.5). At the commencement of the cycle, there were three distinct packs. As the cycle continued, the slower swimmers, who were initially riding individually, merged, and a fourth group of cyclists developed. During the third lap, pack 3 caught pack 2 to form a larger chase pack and reduced the number of packs to three. Significant differences were found between the cadences used at different points of the cycle (F = 0.9241, p = 0.0001, Table 5.7). The t-tests of important time points revealed the initial cadence (post-T1 = 98.2 rpm) and 96 final cadence (pre-T2 = 99.7 rpm) measures were significantly greater than were used at any other stage of the ride (90.4 - 94.4 rpm).

Male cycle results All data describing the freely chosen cadences of male triathletes during the cycle leg are presented in Table 5.8. When all male triathletes were analysed as a complete sample, significant cadence differences were found at different points of the cycle (F = 2.259, p = 0.029). The t-test analyses of important data points revealed that the cadence used during lap1 (98.2 rpm) was significantly greater (p < 0.01) than that used at the start of the final lap (lap 6 = 94.3 rpm).

Table 5.7 Descriptive cycle cadence statistics for all female triathletes.

Lap n Min Max Mean sd sig Post- 60 85.71 107.14 98.02 5.66 c, e, f, g T1 1 39 78.94 107.14 94.26 7.11 h 2 49 73.75 107.14 94.39 7.14 4 45 83.33 107.14 93.44 5.75 5 34 83.33 103.45 93.42 6.11 6 39 78.94 103.45 90.38 5.97 Pre-T2 29 83.33 107.35 99.74 7.18 b,c,e,f,g

Table 5.8 Descriptive cycle cadence statistics for all male triathletes.

Lap n Min Max Mean sd sig Post- 70 75.00 107.14 96.75 7.37 ns T1 1 69 83.33 107.14 98.17 5.36 g 2 67 75.00 107.14 95.78 6.60 3 66 85.71 107.14 96.46 4.94 4 57 88.24 107.14 97.47 4.98 5 56 75.00 107.14 96.08 6.15 6 57 75.00 107.14 94.28 6.68 Pre-T2 49 75.00 107.14 97.47 7.61 ns Note for tables 5.7 & 5.8: sd = standard deviation sig = significant difference between variables (p<0.01) a = significantly different than at post-T1 b = significantly different than at lap1 c = significantly different than at lap2 d = significantly different than at lap3 e = significantly different than at lap4 f = significantly different than at lap5 g = significantly different than at lap6 h = significantly different than at pre-T2 ns = not significant

Run data Figure 5.4 Male and female stride rates over time. Note: female triathletes ran 8 km only

95 94 93 Male 92 Female

stride rate 91 90 89 11a22a33a4 lap

Figure 5.5 Male and female stride lengths over time.

Note: female triathletes ran only 8 km

3.4 3.3 3.2 3.1 Male 3 Female 2.9 2.8 stride length (m) 2.7 2.6 11a22a33a4 lap

Female run results Descriptive statistics for all female triathletes Means and standard deviations of SRs and SLs for female triathletes are presented in Table 5.9. No significant differences were noted between the mean SRs used at any point during the run. This is noteworthy as this first collection point might have been too far from the actual transition in order to record a significant change in SR in the initial stages of the run.

Significant differences were noted between four of the mean SLs. However, no differences were apparent between SL1 and SL1A, SL1 and SL2A, or between SL2 and SL3. The initial SL was significantly greater than that at half way (SL2) and at the end (SL3). The SL at point 1A (3000 m) was significantly greater than that used at any other point further in the race. The SL at 3640 m (SL2) was significantly less than that used prior to SL1 & SL1A and after SL2A, but not the final SL (SL3). The length at SL2A was significantly greater than that adopted at the final collection point (SL3).

Table 5.9 Descriptive run stride rate and stride length statistics for all female triathletes.

Variable Minimum Maximum Mean sd sig n=46 SRL1 85.71 107.14 94.24 4.31 ns SRL1A 83.33 107.14 92.50 4.93 ns SRL2 80.36 107.14 93.24 4.87 SRL2A 83.33 100.00 93.61 4.76 SRL3 86.54 103.45 92.80 4.57 ns

SLL1 2.39 3.32 2.89 0.21 c, e SLL1A 2.50 3.50 2.94 0.20 c, d, e SLL2 2.35 3.43 2.70 0.24 SLL2A 2.52 3.22 2.85 0.17 SLL3 1.40 4.34 2.68 0.37 a, b, d Note: sd = standard deviation sig = significant difference between variables (p<0.01) a = significantly different than at SR1/SL1 b = significantly different than at SR1A/SRLA c = significantly different than at SR2/SL2 d = significantly different than at SR2/SL2A e = significantly different than at SR3/SL3 ns = not significant

Correlations between performance measures Correlations between all measured run variables for all female triathletes are presented in Table 5.10. Significant and positive correlations (p < 0.0001) were recorded between running time and final triathlon times.

No significance correlations were recorded between the SR and the performance variables. Significant and positive correlations between SRs throughout the run were recorded between the SRs at SR1A, and three of the five subsequent data collection points (SR2, SR2A & SR3). Positive correlations also were found between SR2 and SR2A & SR3, and finally between SR2A and SR3 (p < 0.0001).

Both run time and run rank showed negative and significant association with the choice of SLs employed by the female triathletes (p < 0.01). The strongest correlation was noted at the approximate midpoint of the run (3640 m, r = -0.627).

A significant and positive correlation was found between SLs at most collection points. The strongest associations occurred during the middle stages of the run with significant correlations (p < 0.001) between SL1A and SL2 (r = 0.600), SL1A and SL2A (r = 0.752), and SL2 and SL2A (r = 0.649). Weaker correlations were noted at the start and finish of the run. 100

Table 5.10 Pearson correlation of run and triathlon performance with stride rate and stride length for female triathletes.

RNK_RUN TOTAL RNK_TOT SR1 SR1A SR2 SR2A SR3 SL1 SL1A SL2 SL2A SL3 RUN 0.982** 0.761** 0.871** 0.011 -0.310* -0.132 -0.116 -0.251 -0.448** -0.387** -0.627** ,-0.413** -0.535** RNK_RUN 0.737** 0.865** -0.001 -0.295* -0.116 -0.088 -0.222 -0.418** -0.410** -0.636** -0.429** -0.542** TOTAL 0.877** -0.045 -0.272 -0.067 -0.019 -0.178 -0.629** -0.280 -0.393** -0.181 -0.580** RNK_TOT 0.016 -0.209 -0.082 0.025 -0.188 -0.597** -0.347* -0.603** -0.333* -0.560** SR1 0.112 0.160 0.146 0.061 SR1A 0.699** 0.573** 0.679** SR2 0.698** 0.754** SR2A 0.672** SL1 0.318* 0.480** 0.183 0.382** SL1A 0.600** 0.752** 0.426** SL2 0.649** 0.452** SL2A 0.338* Note: ** = Correlation is significant at the 0.01 level * = Correlation is significant at the 0.05 level

RUN = run time RNK_RUN = run time in rank order TOTAL= total triathlon time RNK_TOT = final finishing position

101

Male run results Descriptive Statistics for all male triathletes Descriptive data for male SRs and SLs can be seen in Table 5.11. The average SRs during the last 4 km (SR3, SR3A & SR4) were significantly less than those used during the first 1200 m (SR1) of the run. Also, the SR at the second last collection point (SR3A = 7880 m) was significantly less than that measured at SR2 and SR2A.

Table 5.11 Descriptive run stride rate and stride length statistics for all male triathletes.

Minimum Maximum Mean sd sig n = 51 SR1 83.33 97.83 92.02 2.85 e, f, g SR1A 83.33 100.00 91.39 3.48 ns SR2 83.33 97.83 91.22 3.20 SR2A 83.33 100.00 91.30 3.81 SR3 83.33 97.83 90.63 3.76 SR3A 80.36 97.83 89.92 3.82 a,c,d SR4 83.33 97.83 90.44 3.27 a

SL1 2.41 3.47 3.01 0.21 b, d, e, f, g SL1A 2.82 3.69 3.37 0.19 a SL2 2.58 3.91 2.97 0.27 SL2A 2.66 3.61 3.24 0.20 SL3 2.61 3.57 3.21 0.19 SL3A 2.39 3.60 3.21 0.21 a, b, c SL4 2.40 3.57 3.12 0.22 a, b, c, d, e, f Note: sd = standard deviation sig = significant difference between variables (p<0.01) a = significantly different than at SR1/SL1 b = significantly different than at SR1A/SRLA c = significantly different than at SR2/SL2 d = significantly different than at SR2/SL2A e = significantly different than at SR3/SL3 f = significantly different than at SR3A/SL3A g = significantly different than at SR4/SL4 ns = not significant

The initial SL at 1200 m (SL1) and SL2 was significantly shorter than that used during the remainder of the run. The SL was significantly less at the final data collection point (SL4) than at any other point during the run other than during SL1 and SL2. No significant differences were noted during the middle stages of the run. That is, the average SL used by the triathletes did not change from 5440 m to 7880 m (SL2A - SL3A).

102

Correlations between performance variables All the male performance variables of total time, final rank, run time and run rank, were significantly and positively interrelated (Table 5.12).

Total time was significantly (p < 0.01) and positively correlated with the SR used at the second measurement point (3000 m into the run). Typically, SRs at various stages of the race were significantly and positively correlated with each other throughout the run. The only exception was the second data collection point (SR1A), where no significant correlation was found. The SR used at the start of the second lap (SR2) was significantly (p < 0.01) and positively correlated with all subsequent SRs.

Stride length was negatively related to run time at all points except SL2 where there was no significant correlation. Run rank, total time and final finishing position also revealed significant, negative correlations with all SL points except the first (SL1), third (SL2) and fifth (SL3). The results suggest that there is a strong and positive correlation between SL measures throughout the run, with the exception of SL2 and SL3. 103

Table 5.12 Pearson correlation of run and triathlon performance with stride rate and stride length for male triathletes.

RNK_TOT TOTAL RNK_TOT SR1 SR1A SR2 SR2A SR3 SR3A SR4 SL1 SL1A SL2 SL2A SL3 SL3A SL4 RUN 0.917** 0.857** 0.835** 0.045 0.137* -0.134 -0.144 -0.145 -0.093 -0.236 -0.296* 0.084** -0.120 -0.604** -0.321** -0.765** -0.841** RUN_RNK 0.831** 0.893** -0.065 0.258 -0.243 -0.252 -0.235 -0.146 -0.326* -0.223 -0.747** -0.175 -0.491** -0.322* -0.664** -0.794** TOTAL 0.945** 0.001 0.385** -0.255 -0.275 -0.343* -0.132 -0.278* -0.023 -0.819** -0.049 -0.435** -0.196 -0.644** -0.729** RNK_TOT -0.054 0.327* -0.294* -0.297 -0.357 -0.138 -0.297 -0.123 -0.755** -0.191 -0.468** -0.262 -0.605** -0.693** SR1 0.216 0.336* 0.242 0.175 0.497** 0.395** SR1A 0.001 0.164 -0.196 0.302 0.204 SR2 0.515** 0.359* 0.556** 0.540** SR2A 0.256 0.212 0.371* SR3 0.310* 0.180 SR3A 0.602** SR4 SL1 0.465** 0.517** 0.493** 0.385** 0.359* 0.309* SL1A 0.286 0.790** 0.356* 0.791** 0.839** SL2 0.383** 0.258 0.280 0.182 SL2A 0.275 0.713** 0.624** SL3 0.209 0.291 SL3A 0.794**

Note: ** = Correlation is significant at the 0.01 level * = Correlation is significant at the 0.05 level

RUN = run time RNK_RUN = run time in rank order TOTAL= total triathlon time RNK_TOT = final finishing position

104

Discussion Study four sought to determine self-selected cycling cadences and running stride rates and stride lengths of senior elite female and male triathletes during competition.

Female cycle The cadences used at the commencement and conclusion of the cycle were significantly greater than those used at any other time. Currently there are no results with which these findings can be compared. It is known that self-selection of cycling cadence is usually the most economical pedalling frequency (Hagberg et al., 1981; Hausswirth et al., 1999; Moritani et al., 1993; Takaishi et al., 1996). The same is true for other movement patterns such as upper body exercise (Marais et al., 1999) and wheel chair propulsion (van der Woude et al., 1989).

As with running, cycling velocity is influenced by leg speed (rate). Typically, cyclists try to maintain a similar cadence throughout the entire event, as it is more efficient. Smith et al. (1998) reported an average cadence of 76.3 rpm for three elite triathletes during competition over the Sydney Olympic course during a trial event. Cadence was measured continuously through the use of SRM® cranks and the range recorded was 0- 114 rpm. This highlights the varying nature of when an event is draft legal and the technical nature of the cycling discipline.

However, rather than alter the length of stride as in running, gears are used in cycling to change velocity while maintaining the selected cadence. Without further investigation it is not possible to determine what is happening physiologically at the commencement of the cycle in terms of efficiency. However, the importance of being in the first pack during the ride suggests that all cyclists are making a concerted effort to pedal rapidly, albeit not efficiently, to achieve a greater velocity and move into a pack closer to the front for drafting benefits. The reduction of cadence at the end of the first lap could be related to the sorting out of the packs and the triathletes no longer being able to continue cycling at such high cadences because they create greater physiological demand.

It has been reasoned subjectively by coaches and triathletes, that an increase in cadence at the end of the cycle, through a change in gears, 'relaxes' the legs in order to prepare 105

for the subsequent run. Research has shown that using a higher cadence elevates the associated oxygen cost of cycling and is less efficient (Gotshall et al., 1996; Hagan et al., 1992; Hagberg et al., 1981). However, an increased cadence in a lower gear requires less force per pedal revolution, thus reducing muscle tension (Coast & Welch, 1985; Cox et al., 1994; Faria et al., 1982; Hull et al., 1988; Takaishi et al., 1996 & 1998). The same is true for swimming where distance swimmers use higher SRs than sprint swimmers as it requires less muscle force per stroke which reduces muscle fatigue (Craig & Pendergast, 1981). More recent research (Gotshall et al., 1996) has suggested that there is an increased ability of venous return at higher cadences through a reduced vascular resistance and improved muscle-blood pumping action. Thus, the increased cadence at the end of the ride might physiologically improve blood flow and removal of lactate with the associated increased ventilation. This would reduce muscle stress and improve the ability to run 'off the bike'. This increased venous blood flow may also be important at the commencement of the ride in order to help redirect peripheral blood distribution from the upper body after swimming and into the legs for cycling.

Male cycle When all male triathletes were analysed as a group, few significant differences were noted in comparison with the women. As a group, only the second last cadence recording (lap 6) was different from the second measured cadence. The pattern of a marked increase in cadence at the end of the ride was not as apparent with the men as with the women. However, the mean male cadence was, higher than the mean female cadence for the duration of the ride. This suggested that the male triathletes maintained a slightly higher cadence with less decrement in performance. This may be a result of greater amounts of training or related to body size (Gonzalez and Hull, 1989).

Female run Descriptive statistics It was hypothesised that SR would change during the run because prior cycling, typically done at a slightly higher frequency (1.5-2 Hz) than running (1-1.5 Hz) (Quigley & Richards, 1996), could cause a slight SR increase early in the run via changes in muscle firing patterns. It was noted that mean cycle cadence (94.8 rpm) was slightly greater than mean run cadence (93.3 rpm) but this was not significant. The 106

final cycle cadence (99.7 rpm) was significantly greater than the initial run cadence (93.9 rpm). The SR for women triathletes did not change throughout the entire run. Possibly, the primary data collection point for the run was too far from the transition of the bike to run (1200 m) to ascertain any effect that the cycle discipline had on the initial stages of the run. From another perspective, it might suggest that the SR used is ‘entrenched’ in the athlete, and does not vary throughout the run.

Table 5.9 also presents the SLs used by female triathletes and suggests that longer strides were taken at the commencement of the run. This contrasts with previous research which has suggested that, at the commencement of the run leg following a cycle during a triathlon, there is a significant decrease in SL (Ackland, 1997; Hausswirth et al., 1996 & 1997). This is probably caused by local muscle fatigue (Hausswirth et al., 1997). Hausswirth et al. (1997) also noted a subsequent increase in SR which helped maintain running velocity in elite male triathletes. However, Witt (1993) found a decrease in running SR following cycling. This disagrees with the findings of the present study where no changes in SR were noted throughout the entire run. Hence, results are still equivocal and further investigation is necessary.

These unvaried SR results of the present study could be due to collecting data too far from the transition where changes might have occurred within the first 1200 m. The remainder of the data suggest that, at the mid-point of the race, there was a marked decrease in the SL adopted by the female triathletes. Perhaps fatigue and athletes pushing too hard during the initial stages of the run created this result. Ackland (1997) collected data from a single triathlon World Cup event, whereas Hausswirth et al. (1996; 1997) collected data from simulated events within the laboratory. At a World Championships, athletes might sacrifice efficiency for speed. That is, although there is local muscle fatigue, the triathlete pushes hard to maintain a high velocity to maintain position. If this can not be sustained, the athlete then succumbs during the middle stages as revealed by a marked SL decrease. Hence, data collection in the first 1200m might not show any differences in SL, but could show SR variation.

The increased SL at the penultimate collection point could result from motivation at hearing the bell which signalled the last lap. Hence, this was ones last chance to improve race position. Finally, SL decreased between the last two measurement points. 107

This could be due to either the realisation that race position will not alter and the athlete slows down, or the surge in the previous half lap led to fatigue and a decreased SL.

The resultant fluctuation in SL might also be caused by the course design. At the commencement of each lap (SL2, SL3, SL4), SL was significantly greater than at the end of the previous lap. This could be due to the U-turn the triathletes had to make 300 m prior to data collection. Prior to testing, it was assumed that the cameras were sufficiently far from the turn to remove its effects. On the return trip, the camera was 900 m from the turn. Also, the triathletes had just passed the largest section of the crowd which may have enhanced motivation and 'striding out', possibly causing fatigue. When turning again, the fatigue, U-turn and diminished crowd motivation could have influenced the reduced SL.

Correlations between performance measures Positive correlations would be expected between time and ranking as they measure similar variables, with one being continuous (time) and the other a step integer (rank). The relationship between final time and run time could be due largely to the fact that most triathletes alighted their bikes in one pack. Hence, there were only small differences in their elapsed times prior to the start of the run and the better runners on the day would have placed higher at the finish. The significant relationship between run rank and final place (rank) suggested that those who climbed off the bike first had a greater chance of finishing higher up, and those who dismounted last typically stayed at the back of the field. With the above information, the importance of not only being in the first pack at the end of the cycle stage, but being in the front of that pack and then running well, is highlighted.

This agrees with data presented by Landers et al. (2001) which highlighted the importance of being in the first pack out of the water to have a chance of winning the race. The better runners in that first pack of swimmers typically placed higher at the finish than those who did not swim as well. The results of correlations between SR measures during the run discipline suggests that those with a higher SR at the 3000 m point of the run (SR1A) maintained a higher SR for the remainder of the run. The lack of correlation between the first data point at 1200 m (SR1) and any other point may be due to the effect of prior cycling causing the triathletes difficulty in selecting their 108

preferred running cadence. It could be that all the triathletes are either high or low in terms of their leg speed. The relationship at SR1A with run time (and rank) might suggest that stronger runners have chosen a higher cadence at this point which, in turn, allowed a greater running velocity. As there was no correlation between run time (or rank) with subsequent cadence measures, it could be said that the absolute SR did not disadvantage performance.

Unlike the results for SR, it appears that SL could play an important role in determining the velocity of the triathlete as seen in the run time, run rank and, in most instances, the final finishing position. Research on animals also has shown that an increase in speed is associated with either a change in gait or an increased SL, rather than SR, in order to maintain the best economy of motion (Alexander, 1984; Taylor, 1985). The present results indicate that those triathletes who ran faster, utilised a longer SL.

Results showed also that most triathletes who chose a longer SL, maintained this throughout the run. In the final stages of the run, the correlation between SL at this point, and the points prior, began to diminish. It is possible that fatigue affects different triathletes differently. Research generally has shown that increased fatigue during running results in a decreased SL and, if velocity is to be maintained, an increased SR (Ackland, 1997; Elliott et al., 1981; Elliott and Ackland, 1980; Elliott and Roberts, 1980; Hausswirth et al., 1997, Williams et al., 1991). When this is related to the data of final place and time, it suggests that those females who finished faster (top 10) had significantly longer strides at the beginning, middle and end of the run with relatively unchanged SRs.

Male run Descriptive data for all males Data revealed either an elevated SR at the start of the run, or a decreased SR at the end of the run, as the last three measured SRs were significantly lower than the initial measure. An initial elevated SR could possibly result from the increased muscle firing rates after cycling the last 2-5 km at a significantly faster cadence being carried over to the initial stages of the run. The average cycle cadence of 96.6 rpm was greater than the average run SR (91.0 rpm), and the final run cadence measured (97.5 rpm) was significantly greater than the initial run cadence (92.0 rpm). It could also be that, during 109

the final stages of the run, triathletes are fatigued and can no longer sustain the rapid leg turn over. However, previous research on running fatigue has indicated that SL declines prior to decreases in SR (Elliott & Ackland, 1981; Williams, 1985).

The lower SL found at the commencement of the run is in agreement with other studies. Hausswirth et al. (1996 & 1997 & 1999) suggested that, following the cycle, there is local muscle fatigue at the hip flexors which reduces the range of motion (ROM) and the length of stride attainable at this joint. However, the data revealed that the greatest SL was recorded at the end of the first lap and then decreased at the commencement of the second lap, 640 m later. Hence, it might not only be fatigue that caused the shorter, initial SL, but the prior cycling could have influenced the ability to adopt the most appropriate SL. It is also important to note the current differences in research findings in relation to biomechanical alterations which do not support the decreased SL results of Hausswirth et al. (1996, 1997 & 1999) (Hue et al., 1998; Marino & Goegan, 1993; Quigley & Riichards, 1996). During the middle stages of the run, it seems that the triathletes achieved a constant rhythm because SL did not alter. However, the above SR data demonstrates that there was a slight decrease in cadence during this time. The greatly reduced SL during the final 1.5 km is a key indicator of fatigue (Elliott et al., 1981; Elliott and Ackland, 1981; Elliott and Roberts, 1980; Hausswirth et al., 1996 & 1997; Williams et al., 1991). It could also be due, in part, to the acceptance by the triathletes of the inevitability of their final finishing position.

Correlations between performance measures It is important to note that there is a significant and positive relationship between the run variables (run time & run rank) and the 'final' variables (final time & final finishing position). This indicates the importance of the run in the total triathlon performance. That is, those who run faster, finish higher up in the placings. This is exaggerated by the fact that most of the successful male triathletes start the run at almost the same time due to pack formations created in the swim and cycle legs.

Typically, SR measures were significantly and positively related to each other throughout the run. This indicated that those who chose a higher cadence at the commencement of the run maintained it throughout. The only exception was at the second data collection point (SL1A) where no correlations were found with any other 110

SR measure. At this stage the triathletes could be running at borderline threshold to try and maintain position and some exceed their thresholds and succumb to the high intensity. The SR used at the start of the second lap (SR2) showed significant (p < 0.01) and positive relationships with all subsequent SRs. This possibly highlighted the fact that, by the start of the second lap, the male triathletes had found their 'running cadence'.

As with the female triathletes, it was the SL that correlated significantly with performance measures, and not SR. The greatest strength of the association appeared to be during the last two measurement points (SL3A & SL4) and at the second point (SL1A) (p < 0.0001). Those who managed the largest strides at 3.6 km and during the final 5-2.5 km, recorded the fastest run times and finished the triathlon in a better position. As SL is important in determining running velocity (Hay, 1993) it would be expected that, if SR is relatively similar, those with longer strides will run faster and finish earlier. Combining this with the non-relationship of SR and final position, SL might play a greater role in determining final finishing position and run times than the leg speed of the triathletes.

Summary Cycling cadence, and running SRs and SLs used during triathlon competition by 51 senior elite male and 46 senior elite female competitors in the 2000 Triathlon World Championships were determined via video and analysed using Video Expert II Coach. Study 3 examined the cadence during the first and last km of the cycle leg of a classic distance triathlon. A significantly faster cadence was noted at the end of the cycle when compared with the start for both male and female age group competitors. The initial cycle cadence also was comparable with run SR.

Results from both males and females showed a relatively consistent cadence being used throughout the run and cycle disciplines. The female triathletes used a significantly faster cadence at the commencement and completion of the cycle than during the middle stages. Variations in running SL appear to also be associated with running ability and triathlon performance. That is, the better performers maintained a longer and more consistent SL. 111

Study 5 Cadence selection and performance of senior elite triathletes at the 2000 Triathlon World Championships Based on “Cadence Selection and Performance” International Triathlon Coaching Symposium, July 23-24 2001, Edmonton, Canada, page 13.

Introduction Study five examined the relationship of triathlon performance to the cadences adopted spontaneously for each discipline throughout the race. Secondly, study five observed how the cadences of each discipline related to each other.

Methodology Methodology and subjects were previously outlined in Study 4.

Data Analysis In order to investigate further the results of Study 4, competitors were split into groups based on their cycling and running ability. Cyclists were categorised according to the pack in which they rode. For the run component, competitors were grouped into three equally sized groups according to their run times as Superior, Good and Average. Means and standard deviations of all these variables were calculated. Subsequently, repeated measures ANOVA were utilised to compare between groups at important data collection points, namely at the commencement and completion of each discipline, and also within groups across the cycle and run, to determine whether any significant differences occurred. Only differences between the Superior and Average runners were compared as this allowed more distinction between running abilities.

Results Female cycle At the commencement of the cycle, three distinct packs, plus a fourth group of 'stragglers' were apparent. During the third lap, pack 3 caught pack 2, which formed a larger chase pack and reduced the number of packs to three. For clarity, pack 4 remains pack 4 even when the number of packs were reduced to three. The descriptive statistics of each of these packs can be seen in Table 5.13. 112

Differences between packs during the ride As stated above, the number of packs changed throughout the cycle portion of the triathlon, and some athletes slowed down and dropped back. Hence, differences between packs have been examined in two ways; from the original pack arrangement after the swim, and after packs 2 and 3 combined during lap 3.

Significant cadence differences were found between packs (p< 0.01) up until the start of lap 6. Pack 4 used a significantly lower cadence than pack 2 and pack 3 at the start of the ride and lower than pack 1 and pack 4 at the end of the first lap. Lap 2 revealed no differences between the cycling cadences used in the differing packs. From lap four onwards, three packs existed due to the combined 2 and 3. The last pack (pack 4) continued to record lower cadences than the leading packs during laps four and five (Table 5.13).

Female cycling data compared by pack The three cyclists in pack 1 cycled at a very consistent cadence, with only laps 4 and 5 recording significantly lower cadences (p < 0.01) than those used for the rest of the ride. Packs 2 and 3 used a higher cadence at the start of the cycle stage than during the subsequent three laps. After combining packs, the average cadence further declined during lap 4. The lowest cadence for the ride was recorded at the start of lap 6. At the end of lap 6 (pre-T2) a significant cadence increase, from the previous measurement, was noted.

The back markers results, pack 4, can be split into two parts. A significantly higher cadence was used in the first half of the ride than in the second half, except for the final collection point (pre-T2) where the cadence was increased again.

113

Table 5.13 Descriptive cycle cadence statistics for female triathletes split by pack.

difference post hoc pack1 pack2 pack2&3 pack3 pack4 between groups difference Lap N Mean sd sig N Mean sd sig N Mean sd sig N Mean sd sig N Mean sd sig F ratio sig Pack No. Post- T1 3 98.21 7.73 ns 24 99.30 3.71 ns 18 101.38 4.08 ns 15 91.91 5.16 ns 13.707 0.0001 2 & 4, 3 & 4 1 3 102.38 4.12 ns 11 94.50 4.64 a 13 97.51 5.13 a 12 88.49 7.52 ns 7.221 0.001 1 & 4, 3 & 4 2 3 102.38 4.12 ns 18 95.16 6.53 a 15 93.46 7.40 ns 13 92.55 7.43 ns 1.778 0.165 ns 4 3 102.38 4.12 ns 14 94.37 6.49 a 28 93.65 5.59 a 16 93.59 4.51 a 12 89.93 3.90 a,b,c 4.911 0.012 1 & 2, 1 & 4 5 3 95.83 3.61 c 10 95.20 6.58 b 19 96.02 5.45 ns 11 94.33 5.63 a,b,c 10 89.90 5.93 a,b,c 7.848 0.002 2 & 4 6 3 93.92 6.00 c 11 88.99 7.36 a,c,e,f 24 90.45 6.55 a 16 91.73 5.31 a 9 88.48 4.95 a,b,c 0.699 0.504 ns Pre-T2 2 107.14 0.00 ns 8 96.15 7.54 ns 15 98.72 6.81 ns 9 101.66 4.18 e,f,g 10 99.41 8.56 ns 1.231 0.308 ns

Note: sd = standard deviation sig = significant difference between variables within groups (p < 0.01) a = significantly different than at post-T1 b = significantly different than at lap1 c = significantly different than at lap2 d = significantly different than at lap3 e = significantly different than at lap4 f = significantly different than at lap5 g = significantly different than at lap6 ns = not significant

114

Figure 5.6 Female cycle cadence by pack.

110

105

100 Mean Pack 1 95 Pack 4

cadence (rpm) 90

85 post-T 1 2 4 5 6 pre-T lap

Male cycle The chosen cadences of male triathletes during the cycle leg and significant differences between these variables are presented in Table 5.14. During the male cycle, five packs were evident until during the third lap when packs 2 & 3 combined.

Differences between packs during the ride During lap one (post-T1), the fastest swimmers who were in pack 1 revealed a significantly greater cadence than the slowest swimmers who were in packs 4 and 5. During lap four, the chase pack (pack 2) was using a significantly higher cadence than those in pack 4. In the final 2 km of the cycle (pre-T2), the chase pack cycled at a significantly higher cadence than the competitors in pack 5.

Although most finished the cycle leg, no competitors in pack five commenced the run, as they withdrew from the race.

Male triathlete data compared by pack When all competitors were sub-divided into the packs in which they initially exited the water and commenced cycling, significant differences were noted between the cadences selected at various points of the cycle for packs 1, 2 and 3. Pack 1 cyclists used a significantly higher cadence immediately after swimming (post-T1 = 103.15 rpm) when compared with lap 5 (96.76 rpm). Pack 2 cyclists used a significantly greater cadence

115 during the initial lap (100.01 rpm) when this was compared with the start of the final lap (lap 6 = 93.60 rpm). Pack 3 triathletes had a significantly greater cadence during the final collection point (pre-T2 = 101.77 rpm) than the preceding lap (lap 6 = 93.88 rpm) and the second lap (94.96 rpm). Packs 4 and 5 showed no significant differences during the ride.

As packs 2 and 3 merged during lap 3, and some other triathletes changed packs during the ride, this analysis process was repeated to determine changes of pack cadence over time, but not necessarily related to individual changes in cadence. In this case, significant differences were noted in packs 1, 2 and 5.

Post hoc testing did not reveal where the differences were in pack 1. Pack 2 cyclists, which was comprised of the initial pack 2 and 3 cyclists, used a lower cadence at the start of lap 6 (5.5 km from the end of the cycle) than that used at post-T1, lap 1, lap 4 and pre-T2. Post hoc testing did not reveal where the significant differences existed in pack 5 (p < 0.01).

Figure 5.7 Male cycle cadence by pack.

104 102 100 98 Mean 96 Pack 1 94 Pack 5 92 cadence (rpm) 90 88 post- 123456pre-T T lap

116

Table 5.14 Descriptive cycle cadence statistics for male triathletes split by pack.

pack1 pack2 pack2&3 pack3 pack4 pack5 difference between groups lap N Mean Sd sig N Mean sd sig N Mean sd sig N Mean sd sig N Mean sd sig N Mean sd sig F ratio sig post hoc Post- T1 7 103.15 4.94 17 100.01 4.67 21 95.74 8.78 15 94.10 7.01 10 92.91 5.61 4.009 0.006 1 & 4, 1 & 5 1 7 98.06 6.78 17 99.18 3.90 21 97.86 5.80 10 96.67 4.66 14 98.54 6.35 0.900 0.598 2 7 97.32 3.34 13 97.61 7.14 24 94.96 7.66 8 93.06 6.53 15 96.22 5.41 0.775 0.546 3 5 100.73 2.81 17 95.65 5.78 23 95.82 5.02 8 96.52 3.83 13 96.97 4.55 1.196 0.322 4 5 95.60 2.78 12 99.41 35 98.79 4.50 21 98.32 6 91.56 4.68 11 97.31 5.20 4.632 0.006 2 & 4 5 3 94.76 1.74 a 14 98.17 37 97.14 5.32 21 96.49 6 96.88 3.42 10 92.08 9.34 1.961 0.131 6 5 97.46 2.63 16 93.60 a 37 94.00 6.33 a,b,e 19 93.88 5 96.94 4.54 10 92.40 9.57 0.923 0.436 Pre-T2 5 98.75 2.80 9 96.28 28 99.94 6.47 g 18 101.77 c,g 6 97.92 3.23 10 89.64 9.24 5.997 0.002 2 & 5

Note: sd = standard deviation sig = significant diffence between variables within groups (p< 0.01)

A = significantly different than at lap1 B = significantly different than at lap1.1 C = significantly different than at lap2 D = significantly different than at lap3 E = significantly different than at lap4 f = significantly different than at lap5 G = significantly different than at lap6 ns = not significant

117

Female run Descriptive statistics for Superior vs Average Runners These results pertain to splitting the female field into three groups of equal numbers, based on run time. In order to highlight any differences, comparisons were made only between the top third (superior) and bottom third (average) runners. This was to try and determine whether the superior runners were more consistent with SR/SL, the average were worse, or if there were no differences in the patterns exhibited by the different groups (Table 5.15). A two-way repeated measures ANOVA revealed no significant differences between the two groups in terms of the SR used throughout the run.

Table 5.15 Descriptive stride rate and stride length run statistics for female triathletes, split by run time.

Superior Good Average Mean sd sig Mean sd sig Mean sd sig sig diff N=15 n=17 n=14 btw groups SR1 93.72 3.96 ns 94.96 4.86 ns 93.92 4.14 ns ns SR1A 93.36 6.29 ns 93.23 4.07 ns 90.69 3.99 ns ns SR2 94.02 6.87 ns 93.05 3.93 ns 92.62 3.33 b ns SR2A 93.62 5.04 ns 93.97 4.65 ns 93.16 4.90 ns ns SR3 93.88 6.28 ns 93.31 3.65 ns 91.02 2.88 a, c, d ns

SL1 2.99 0.19 ns 2.94 0.15 ns 2.73 0.21 ns y,z SL1A 3.03 0.23 ns 2.94 0.21 ns 2.85 0.13 ns y SL2 2.92 0.27 ns 2.61 0.13 a,b 2.57 0.12 a,b x,y SL2A 2.93 0.19 b 2.84 0.17 b,c 2.79 0.13 c ns SL3 2.95 0.43 ns 2.65 0.10 a,b,d 2.42 0.33 a,b,d x,y,z

Note: sd = standard deviation sig = significant difference between variables within groups (p<0.01) a = significantly different than at SR/SL1 b = significantly different than at SR/SL1A c = significantly different than at SR/SL2 d = significantly different than at SR/SL2A x = Superior runners significantly different from good runners y = Superior runners significantly different from average runners z = Good runners significantly different form average runners ns = not significant

A repeated measures ANOVA was performed to determine whether any significant SR differences over distance existed within each of the groups. No significant differences were found between the selected cadence at any point of the run for the superior runners. However, the average runners showed a significant increase in SR between SR1A and SR2. This coincides with a significant decrease in SL. The final cadence of

118 these average runners (SR3) was significantly less than that used earlier in the run (SR1, SR2 & SR2A). In addition, a significant decrease in SL was noted.

A repeated measures ANOVA conducted on the SL of the triathletes over the course of the run revealed significant differences (p < 0.01) at the start of each lap; SL1, SL2 and SL3. Post hoc testing clarified that these significant differences were greatly influenced by the differences between the superior and average runners. The SL used by the average runners 1200 m (SL1) into the run was significantly less than that of the superior runners. At 3000 m (SL1A), a significant difference between superior and average runners existed. At 3640 m (SL2), the SL of the superior runners was significantly greater than average runners. Finally, the superior runners used significantly longer strides during the last 2 km (SL3).

Within-group variation of the superior runners revealed that only the SL at 3000 m (SL1A) was significantly greater than that 1 lap later (5440 m, SL2A). All other SLs were the same for the superior runners. Significant differences were noted between the SLs for the average runners at different data collection points. The SLs at 1200 m (SL1) and 3000 m (SL1A) were significantly greater than those recorded at 3640 m (SL2) and 6080 m (SL3). The average runners recorded a significantly longer stride at SL2A than that immediately prior (SL2) or following (SL3).

Male run Descriptive Statistics for Superior vs Average Runners Splitting the pool of male triathletes into three groups of equal numbers allows further investigation into the SR and SL characteristics employed (Table 5.16). As with the female triathletes, only comparisons of the superior runners (top third) and average runners (bottom third) are reported and discussed.

A two-way, repeated measures ANOVA revealed no significant differences between the three groups for SR at each of the measurement points. However, a within-group variation over time was revealed. The superior male runners showed only one difference in average cadence, that being the SR at 7880 m (SR3A) being significantly less than the initial cadence used by the triathletes.

119

Table 5.16 Descriptive stride rate and stride length run statistics for male triathletes, split by run time.

Superior Good Average Mean sd sig Mean sd sig Mean Sd sig sig diff n=20 n =15 n = 16 btw groups SR1 92.63 3.05 ns 91.43 2.91 ns 92.14 2.60 ns ns SR1A 90.51 3.82 ns 91.63 2.79 ns 92.58 3.41 ns ns SR2 92.20 2.83 ns 90.54 3.30 ns 90.64 3.42 ns ns SR2A 91.85 3.81 ns 91.23 3.39 ns 90.55 4.38 ns ns SR3 91.59 3.46 ns 90.79 4.19 ns 89.30 3.52 a ns SR3A 90.43 3.82 a 89.79 3.62 ns 89.35 4.16 a ns SR4 91.56 2.99 ns 90.44 3.67 ns 89.03 2.84 a,b,c ns

SL1 3.04 0.19 3.05 0.21 ns 2.93 0.23 ns ns SL1A 3.48 0.14 a 3.41 0.07 a 3.14 0.17 a y,z SL2 2.98 0.24 b 2.97 0.22 b 2.90 0.25 b ns SL2A 3.31 0.15 a, b, c 3.28 0.15 a,b,c 3.09 0.23 a,b,c y,z SL3 3.28 0.15 a, b, c 3.23 0.15 a,c 3.12 0.24 a,c y SL3A 3.36 0.14 a, b, c 3.17 0.10 a,b,c 3.06 0.23 a,b,c x,y SL4 3.29 0.12 a, b, 3.13 0.13 b,c,d 2.89 0.16 b,d,e,f x,y,z c, f

Note: sd = standard deviation sig = significant difference between variables within groups (p<0.01)

a = significantly different than at SR/SL1 b = significantly different than at SR/SL1A c = significantly different than at SR/SL2 d = significantly different than at SR/SL2A e = significantly different than at SR/SL3 f = significantly different than at SR/SL3A

x = Superior runners significantly different from good runners y = Superior runners significantly different from average runners z = Good runners significantly different form average runners ns = not significant

The average group of runners demonstrated significant decline in SRs during the second half of the run. Stride rates at SR3, SR3A and SR4 were significantly smaller than the initial cadence used by this group. The final cadence measured (SR4) was also statistically smaller than the SR at SR1A and SR2.

The SL of the three groups was compared via a two-way, repeated measures ANOVA for each of the data collection points throughout the run. Significant differences were apparent between the groups at all stages other than the initial SL (SL1) and the third measurement (SL2).

120

Using a post hoc test, the SLs of the superior runners were significantly greater than the average runners at 3000 m (SL1A) and then at every point from 5440 m (SL2A) to the finish (SL2A, SL3, SL3A, SL4).

A two-way, repeated measures ANOVA was carried out to determine how constant the SL was maintained throughout the run in each of the previously defined groups. The superior runners significantly increased in their SL from 3.04 m at SL1 to 3.48 m at SL1A. The SL at SL1A was significantly the longest stride used by the superior male runners during the triathlon. The SL2 was significantly less than all SLs other than the initially recorded SL (SL1), where there were no significant differences. There were no significant differences between the SL of the final four SLs measured, other than a significant decrease between SL3A and SL4, the final two measurement sites.

When analysing the average runners, SL1 was significantly less than all other SLs except at SL2 and SL4 where there were no significant differences. The SL at SL2 (3640 m) was not significantly different from the SL used at the start (SL1) or the end (SL4) of the run section of the triathlon by the average male runners, but was significantly smaller than at any other time. The SL used at SL1A was significantly greater (p < 0.01) than that used throughout the run with the exception of that at SL3 where there were no differences. No significant differences were noted between the subsequent measures from SL2A to SL3A but a significant decline in SL was noted between SL3A and SL4.

Discussion Female cycle Difference between packs during the ride As stated above, the number of packs changed throughout the cycle portion of the triathlon. Some of the athletes slowed down and dropped back. Hence, differences between packs has been examined in two ways; from the original pack arrangement after the swim, and after packs 2 and 3 combined during lap 3.

121

Significant cadence differences were found between packs up until the start of lap 6. It was during lap 6 when all the female triathletes significantly increased their cadences prior to the run. Hence, it was a strategy used by all triathletes and not just one group. However, the results highlight that the cyclists closer to the front typically used higher cadences than those in the last pack. The reason for this could be that the triathletes in the last pack were not as accomplished cyclists and used a lower cadence to enable continuing. On the other hand, it might relate to speed because those in the final pack also recorded slower cycle times.

Female triathlete data compared by pack When analysing the individual groups throughout the ride, it became apparent that the lead pack had greater performance homogeneity in the group and there was a greater structure in sharing the workload. Higher cadences were used at the start of the ride for pack 2 and 3. This could be related to the fact that, at this time, these athletes were aiming to catch the group in front and tried to ride faster by pedalling faster. The decline in average cadence after the combining of packs 2 and 3 may be related to the size of the pack. That is, those who are not at the front of the peloton are not required to pedal as quickly or as forcefully as they have reduced drag. The back markers results can be split into two parts with the first half of the ride revealing a significantly higher cadence than the second half. The exception was at the final collection point (pre-T2) where cadence again was increased.

Male cycle Differences between packs during the ride. The results show that during lap one (1500 m from transition), the fastest swimmers (pack 1) had a significantly greater cadence than the slowest swimmers who were in packs 4 and 5. This might be due to both the first group of swimmers pedalling rapidly in order to increase velocity and hold off the chasing pack. Perhaps the last pack of swimmers were excessively fatigued after the swim due to their lesser ability and were not able to pedal at such a high cadence.

During lap 4, the chase pack (pack 2) used a significantly higher cadence than those in pack 4. During the final 2 km of the cycle (pre-T2), the chase pack cycled at a significantly higher cadence than the competitors in pack 5. Perhaps the chase pack was

122 trying to make up lost ground on the leaders and this was a last ditch effort to bridge the gap before the commencement of the run. Possibly, the chase pack increased cadence, as noted with the female triathletes, in order to reduce muscle stress (Coast & Welch, 1985; Cox et al., 1994; Faria et al., 1982; Hull et al., 1988; Takaishi et al., 1996 & 1998) and remove muscle waste products (Gotshall et al., 1996) prior to the run in order to improve run performance. Finally, the differences in cadence between the two groups could reflect the realisation by the competitors in pack 5 that their race was over and they reduced effort via a lower cadence and velocity. As reported, above no competitor in pack 5 completed the event as they were forced by officials to withdraw at the commencement of the run after the lead runners had already finished their first lap.

Male triathlete data compared by pack When all competitors were split into the packs in which they exited the water and commenced cycling, significant differences were noted between the cadence selected at various points during the cycle for packs 1, 2 and 3. It is possible this examined cycling cadence relationship to swimming performance, because the superior swimmers were grouped, as were the average swimmers.

Pack 1 cyclists used a significantly higher cadence during post-T1 (103.15 rpm) when compared with lap 5 (96.76 rpm). Pack 2 cyclists used a significantly greater cadence during the initial lap when compared with the start of the final lap. Pack 3 triathletes revealed a significantly greater cadence during the final collection point than the penultimate lap and the second lap. However, these results pertain to when packs 2 & 3 were combined and could be misinterpreted. After combining the two packs at the end of lap 3, the cadence used at the commencement of the final lap was significantly less than at any other point of the ride. This followed a similar trend to that of the women. That is, a lower cadence was found at the commencement of lap 6 and a higher cadence at the end of lap 6 (pre-T2). However, this final cadence was not greater than at any other point, whereas females recorded a significant increase in cadence during the final stages of the cycle discipline. It might be that the male triathletes at the start of lap 6 were jockeying for position because no one wanted to lead and face the wind resistance and the resultant increased energy expenditure (Faria, 1992). Hence, cadence and subsequent velocity were reduced. By the end of the last cycle lap, a position at the head of the bunch is important for a faster transition to the run (Millet & Vleck, 2000)

123 and an opportunity to attempt to both relax the legs (Coast & Welch, 1985; Cox et al., 1994; Faria et al., 1982; Hull et al., 1988; Takaishi et al., 1996 & 1998) and clear out waste products (Gotshall et al., 1996) before the final run leg. Packs 4 and 5 showed no significant differences during the ride and it could be that these athletes found a rhythm or comfort zone, and cycled at their most efficient cadences.

Female run Superior vs Average Runners The run results obtained from the female triathletes indicated that the superior runners maintained a consistent SR and SL throughout the entire run. However, the average female runners recorded large variations in SL during the run. In a number of instances these were significantly less than were recorded by the superior runners. Running velocity is the product of both SR and SL (Hay, 1993) and any decrease in either, while maintaining the other, leads to decreased running speed and increased running time.

This evidence, along with the consistent SR data, shows that the superior female triathlete runners were able to maintain a constant velocity for the entire duration of the run. It also shows that, as there were no differences in SR between the groups, it was the longer SL of the superior runners that provided them with the greater running speed for a better finishing performance. Elliott and Ackland (1980) showed that, at the end of a 10 km track run, fatigue was related to a decline in SL but there was no change in SR.

The present data could be interpreted to mean that the average runners attempted to stay with the superior runners for as long as possible. However, by the mid-way point they had overexerted themselves which resulted in a marked decrease in SL. Again, possibly spurred on by the bell, there was an increase in SL at the 5440 m but then a subsequent decrease 600 m later.

The data from the average runners fluctuated widely from one collection point to the next. Hence, these runners had difficulty maintaining a constant SL and velocity which used more energy. Although the results vary, the trend was for a declining SL. This suggests a greater level of fatigue was present and led to a reduction in SL.

124

The data could suggest that the selected cadence is ‘entrenched’ in an athlete because it varied little throughout the run. Only the average runners showed a decline in SR during the last 2 km. This decline may not just be fatigue related. It could include an acknowledgement that success in the race was reduced and they decreased the level of effort in the final stages.

The effects of the previous activity (cycle) might affect athletes differently. The results showed also that the superior runners ‘find’ their cadence earlier than the average runners. Hence, this could enable them to run more efficiently/economically from an earlier stage and perform better.

A two-way, repeated measures ANOVA conducted on the SLs used by the three groups over the course of the run revealed significant differences at SL1, SL2 and SL3. Tukey’s HSD post hoc testing clarified that these significant differences were greatly influenced by the differences between the superior and average runners. The SL used by the average runners 1200 m into the run was significantly less than that of the superior runners. At 3000 m, 3640 m and in the final 2 km, a significant difference existed between superior and average runners with the superior runners using a significantly longer stride.

Male run Superior vs Average Runners The superior male runners showed little or no change in their selected cadences over time. No significant differences were noted between each of the three groups at any stage of the run. These results of the complete male sample were comparable with the entire female sample in that there were no significant differences in SR between the ability levels of runners.

However, during the second half of the run, the average male runners recorded significantly lower SRs in comparison with the onset of the run. It could be assumed that the average runners either attempted to maintain cadences above their ability levels for the initial stages of the run to try and stay with the rest of the field. Another possibility is that fatigue in these runners created a decreased SR during the latter

125 stages, or they may have ‘given up’. Again, this was similar to the women triathletes where only the average runners decreased SR during the second half of the run.

Stride Length A SL variation was noted when including all male triathletes in one group (Study 4). This pattern continued through each of the sub-groups where the first three measurements of SL were significantly different from the SL selected during the rest of the event. Both groups recorded significantly shorter SLs during the final measurement when compared with the one immediately prior. At this stage, the average runners had shorter SLs than at any other point in the run. Again, this highlights the fact that all triathletes experienced fatigue during the run and resulted in a reduced SL (Elliott et al., 1981; Elliott and Ackland, 1980; Elliott and Roberts, 1980; Hausswirth et al., 1997; Williams et al., 1991). The poorer runners either experienced fatigue to a greater extent, or realised that they were unable to win and reduced their efforts, and thus reduced SL.

As with the female triathletes, the SLs used by each of the groups changed significantly throughout the run. This was true except at the initial collection point 1200 m from the cycle to run transition and at the start of the second lap. With no differences in SR or SL, the data suggested that all triathletes were travelling at the same speed at the commencement of the run. However, the average runners would be working at a proportionally higher level of their maximum energy expenditure capacity at this stage (Alexander, 1984) which could explain the greater decrease in SL in the latter stages of the run.

As there were few SR differences between the groups, the different SLs as the run progressed was expected because running velocity is a product of SR and SL (Hay, 1993). A faster time for the run would require an increased SR or SL. In the case of the male triathletes, a significantly longer stride appears to be the choice of the better performers. The greatest determinant of running ability and final triathlon outcome could be the ability to maintain a consistent SL throughout the run.

Because cycling uses both flexion and extension of the lower limbs, it is possible that cycling caused fatigue of the same musculature which is associated with running. This could result in a lower than average SL at the commencement of the run. Then, if the

126 superior runners attempted to increase their average running speeds by increasing SLs over the next kilometre, it could cause fatigue and a significant decrease in SL. During the second half of the run it appeared that the superior runners maintained a constant SL which was slightly shorter than that used at SL1A but greater than the initial SL. The significant decrease in SL during the last 2 km between SL3A and SL4 shows the effects of fatigue.

A similar pattern emerged for the average runners who recorded a significantly shorter SL at the start of the run than that used throughout the main body of the 10 km. In this case, the final SL measured was just as small as the initial SL. This could indicate that fatigue had occurred to a larger extent at the start of the run. Hence, they were unable to develop a long SL at the start of the run. Perhaps fatigue is much greater for the average runners at the end of the run and caused a decrease in SL to revert to a level equivalent to that used at the start.

Summary This study sought to determine if different cycling cadences and running stride rates and lengths used during triathlon competition could lead to better performances or predict race outcome. Results of both males and females show a relatively consistent cadence being used throughout the run and cycle disciplines with the female triathletes using a significantly faster cadence at the commencement and completion of the cycle. Variations in running SL appear to be related to running ability and triathlon performance. Those who maintained a longer and more consistent SL, performed better.

127

Chapter 6

Study 6 Running stride rate, stride length and body size relationship for senior elite triathletes at the 2000 Triathlon World Championships

Abstract Previous research has suggested a somewhat predictable relationship between self- selected run stride rate (SR) and stride length (SL) with whole body mass (Taylor, 1985). Earlier studies in this thesis have examined and revealed relationships between body size measures such as limb lengths and adiposity with performance (Chapter 3), and SL with performance (Chapter 5). Hence, the aim of this investigation was to examine whether there were any correlations between SR and SL, with body mass and height in senior elite triathletes during the final run stage of a triathlon.

The SRs and SLs of 30 female and 28 male senior elite competitors at the 2000 TWC were calculated from videography. These values were correlated with the athletes’ body masses and heights. The results for females and males respectively, indicated significant and negative correlations between SR with body height (r = -0.46 & -0.42) and mass (r = -0.44 & -0.39); and a significant, positive correlation between SL with height (r = 0.41 & 0.52). That is, those who were taller and heavier, typically used longer strides and a lower stride rate during the run phase of a triathlon. 128

Introduction People normally change from walking to running at the speed which produces the most efficient gait for the chosen velocity (Alexander, 1984). Hoyt and Taylor (1981) showed that horses avoided the awkward, inefficient speeds at the borderline between gaits, and preferred speeds for walk, trot and gallop. In these cases of borderline velocities, animals tend to speed up, slow down or change between the two gaits. Altering cadence during cycling could be considered in the same category as altering gait because cycling biomechanics change with cadence changes (Neptune et al., 1997). Although there is no change in gait during the run section of a triathlon, the selected SL/SR can vary considerably and selecting the best combination is important for efficient locomotion and successful performance (Cavanagh and Williams, 1982).

The importance of the run discipline during draft legal triathlons has become evident (Landers et al., 2001b). Those who run well after cycling, typically finish in a higher position than the less skilled runners. Hence, there is a need for more information regarding running when it follows cycling, to assist athletes and coaches to improve performance.

Self-selected cadence Self-selected cadence occurs spontaneously and could be a successful or unsuccessful choice. It also could be related to a natural frequency of the body, muscle firing pattern, or muscle contraction efficiency. The self-selected cadence by athletes in a variety of sports appear to be closely related to the optimal frequency for that movement (Gotshall et al., 1996; Hagan et al., 1992; Marais et al., 1999; Marsh & Martin, 1993 & 1997; Millet et al., 1997; Takaishi et al., 1996 & 1998; van der Woude et al., 1989). Through analysis of research data and observation of athletes, it has been noted that elite athletes choose similar leg frequencies in a variety of individual sports irrespective of limb length, fibre type, O2max or anaerobic threshold (Faulkner et al., 1971).

Most data related to SR and SL during running has shown that both parameters change in proportion to speed. Initially, there is an increase in SL, which is followed by an increase in SR (Williams, 1985). However, what triggers these increases is not clearly understood, but it appears to be related to optimal efficiency. Endurance athletes seek 129 an optimal SR/SL relationship that allows the greatest sustainable running speed for least effort. Some authors have shown that SL is related to lower limb length (LLL) or stature (Hay, 1993).

Elite cyclists choose a higher cadence than non-cyclists (90 rpm verses 60 rpm respectively) which was considered to be a result of cycling skill or experience (Takaishi et al., 1996 & 1998). However, Marsh and Martin (1993) also studied high class runners as non-cyclists in order to maintain similar O2max values between groups (67.8 ml·kg-1·min-1 verses 66.2 ml·kg-1·min-1). The runners chose a similar cadence to that of the cyclists during 200 W cycling. This observation suggests that cycling skill might not be the only factor, and that the fitness level also could be important. Fitness level was also offered as an explanation by Marsh and Martin (1997) in their follow-up study comparing skilled cyclists, runners and less trained non-cyclists. They suggested that the history of running training was similar to that of cycling (high number of repetitions, relatively low forces and relatively fast joint angle velocities), and resulted in similar force-velocity properties of the muscles of the lower limb.

Takaishi et al. (1998) selected non-cyclists from team sports to try to nullify the possible lower limb training effect and compared cycling cadence between equally fit groups, based on O2max levels. He found that cyclists had greater pedalling skill and chose higher cadences (75 or 90 rpm) than non-cyclists (60 or 75 rpm) at equivalent workloads.

Generally, swimming data are presented in terms of stroke rate and length (Craig and Pendergast, 1981; Hay, 1993). A quick analysis of two middle distance swimmers suggests that swimmers also choose a leg speed similar to that of cyclists (90 rpm). Eg. 44 strokes/50 m in 30 s, with a 4 beat kick is ~ 88 kicks per minute 50 strokes/50 m in 50 s, with a 6 beat kick is ~ 90 kicks per minute

Hay (1993) showed that the male stroke length was longer than that of females but there was a similar frequency. Male swimmers have longer limbs than females. Therefore, the difference in velocity between the two genders could be attributed to limb length, which directly relates to stroke length, and is not influenced by stroke rate (Carter & 130

Ackland, 1994). Grimston and Hay (1986) reported anthropometric characteristics accounted for 89% of the variation in stroke length and 49% of stroke rate in swimmers.

Limb length & anthropometric relationships with cadence Ackland et al. (1998a & 1998b), Landers (1998) and Landers et al. (1999) found a strong relationship between lever lengths and swim performance during a triathlon. Anthropometry was found to account for over 50% of the variance in total performance. Other studies have found a relationship between LL and SL during running but usually only at higher speeds (Hay, 1993). Thigh length correlated negatively with limb frequency during running (Tittle & Wutscherk, 1988). That is, a high crural index was related to higher stride frequencies at a similar SL; thus, greater speed for the same effort. Studies relating cycling cadence to anthropometry measures have shown equivocal results. However, there appears to be a relationship between LLL and seat height (Price & Donne, 1997) which requires greater clarification.

Increased speed results from an increase in both SL and SR. Initially, the most rapid increase is in SL. Then, the SL plateaus while the initially low SR then increases greatly once the freely chosen SL is selected (Elliott & Blanksby, 1976; Williams, 1985). Therefore, at lower speeds, the relationship between SL and LLL may not be as strong. Cavanagh and Williams (1982) found no relationship between SL and LLL while running at a moderate speed of 3.83 m/s (13.8 km/h).

However, body size might not have as great an influence on rates of contraction if body shape is similar (Dern et al., 1947; Wilkie, 1950). Following a singular contraction of biceps brachi, where the maximal force varied between subjects (12.5 - 20 megadynes), Wilkie (1950) found maximal velocity was relatively constant across all subjects. Wilkie (1950) also suggested that it followed similar lines of geometrical similarity in animals and noted the different sizes, but similar shapes, of greyhounds and race-horses, which generate similar running speeds. This may not be applicable to the present study as it relates to a maximal effort rather than economical movements. This was first presented by Hill (1936), who suggested that the most efficient muscle movements are those conducted at the product of one third times the maximal contraction velocity and force. In contrast, Taylor (1985) noted the strong relationship between stride frequency and whole body mass in bipedal hopping, quadruped galloping and human hopping. 131

This suggested that the entire body mass determines stride frequency, a frequency where muscle-tendon spring efficiency is optimised via minimum metabolic cost and maximal mechanical advantage through the cyclic action. Taylor (1985) also noted that animals (bipeds and quadrupeds) maintain the same frequency across a two to three fold increase in speed.

The research is equivocal as to the relationship between economy of motion and SL/SR. It is still not clear if taller/long limbed athletes choose a lower cadence when at race pace, or choose the same SR as their shorter counterparts and their longer strides ensure that they go faster. Perhaps taller people take relatively shorter running strides or swimming strokes and attain the same speed as shorter competitors. Hence, this study sought to determine whether any correlation existed between the body masses and heights, and the SRs and SLs used by senior elite male and female triathletes during competition. Based on results from previous research, it was hypothesised that there would be a significant and positive correlation between SL and height.

Stride rate, stride length and performance Preceding studies (chapter 5) have highlighted the interaction between SR and SL during the run portion of a triathlon and the effect on performance. No significant differences were reported between the SRs of the male and female triathletes. However, the male triathletes used a significantly longer SL than their female counterparts and, subsequently, had a higher running velocity. The variation in running SL during the event was significantly associated with running ability and final finishing position (Landers et al., 2001a).

Hence, the purpose of this study was to determine any relationships between SR and SL, with body mass and height in elite triathletes during the run stage of a triathlon.

Methodology Sample Subjects comprised 30 female and 28 male senior elite triathletes of the 80 triathletes who competed at the 2000 TWC. The sample included only those subjects for whom both running mechanics and body size data were available. 132

Testing Protocol Prior to competition, mass and height were collected following the procedures set out in Appendix F. Each competitor was marked with a number code to aid identification and then video taped during competition to determine individually selected SR and SL while running under race conditions. One camera was placed in a relatively flat section during the run phases in order to minimise the effect of terrain on the choice of cadence (Figure 5.2).

The event was the classic distance triathlon commencing with the 1500 m, one lap swim in a salt-water river after a dive start from a pontoon. Competitors were permitted wetsuits during the swim as water temperature measured 18.8oC. This was followed by a six lap, 40 km cycle. Each cycle lap consisted of four 90o right hand turns, two 90o left hand turns and one 180o right hand U-turn. In addition, one large ascent and descent covering 50 vertical metres, and one smaller climb and descent of 10 vertical metres were included. Drafting was permitted during the cycle leg, and the triathlon finished with a 10 km run conducted over a flat three and a half lap course.

On the run course, a camera (25 Hz) was placed 1200 m after the cycle-to-run transition. This allowed determination of SR and SL at 1.2 km (lap 1), 3.0 km (lap 1A), 3.64 km (lap 2), 5.44 km (lap 2A), 6.08 km (lap 3), 7.88 km (lap 3A) and 8.52 km (lap 4). Due to unforeseen circumstances (human error), the women triathletes only completed 7.8 km (2.5 laps) rather than the standard 10 km. This limited the data for the women to five data points and could have skewed the results by the females misjudging their energy expenditures due to thinking they had to run 10 km.

Data analysis The video was replayed after the event, and SRs and SLs for each triathlete were calculated at each of the points described above from the run (Table 5.6). Stride rate was determined as the number of strides (or half the number of steps), completed each minute. The SLs and instantaneous velocities were also calculated. The data collection used digitising via the computer program, Video Expert II Coach.

133

Each data point during the run was coded either SR for stride rate or SL for stride length. The subsequent 1, 1A, 2, 2A, 3 or 3A explains during which lap the data were measured. SR1A refers to stride rate data collected at the end of the first lap of the run, whereas SL3 is the stride length measured at the commencement of lap 3. For full details, see Table 5.6.

Data were entered onto an SPSS spreadsheet where the means and standard deviations for each collection point were determined for male subjects, female subjects and the combined sample. Pearson Product Moment correlations were then undertaken to determine the extent of the relationship between SR and SL with body mass and height. This process was conducted for the male and the female samples.

Results Anthropometric details of body mass and height for the triathletes can be seen in Table 6.1. The male triathletes were significantly heavier and taller than their female counterparts.

Table 6.1 Anthropometric and t-test data male (n = 28) and female (n = 30) senior elite triathletes. Female Male sig mean sd mean sd Mass (kg) 55.5 4.4 69.3 6.1 0.0001

Height (cm) 165.9 4.5 179.9 6.4 0.0001

Note: sd = standard deviation sig = significant difference between males and females

Self-selected SR and SL measures for each data collection point, and their subsequent correlation with body mass and height, are presented in Table 6.2 for both male and female competitors. Both male and female triathletes revealed weak, negative correlations between mean SR and body mass, and mean SR and height. No significance was apparent between body mass and SL for either gender. However, significant and positive correlations were recorded by both male and female triathletes for SL and height. 134

It was first envisaged that the data of all male and female subjects would be combined in order to increase the variation in body sizes. However, due to significant differences between all the male and female variables that were measured, this was deemed inappropriate.

Table 6.2 Stride rate, stride length data and correlation with mass and height for male and female senior elite triathletes. Female Male mean sd mass height mean sd mass height correlation correlation correlation correlation SR1 94.5 3.7 -0.07 -0.34 91.8 3.5 -0.48* -0.40* SR1A 92.1 5.0 -0.42* -0.24 92.0 3.6 -0.43* -0.60* SR2 93.4 5.4 -0.36* -0.32 91.5 2.3 -0.30* -0.30 SR2A 93.6 5.0 -0.36* -0.38* 91.8 3.5 -0.30* -0.37* SR3 92.8 4.5 -0.41* -0.49* 90.8 3.5 -0.06 -0.03 SR3A 90.2 3.5 -0.06 -0.09 SR 4 91.4 3.0 -0.03 -0.02 mean 93.3 3.6 -0.44 -0.46 91.3 2.2 -0.39 -0.42

SL1 2.88 0.24 0.11 0.40* 3.0 0.21 0.09 0.21 SL1A 2.96 0.21 -0.04 0.24 3.4 0.18 0.24 0.47* SL2 2.72 0.27 -0.04 0.50* 3.0 0.24 0.28 0.31 SL2A 2.87 0.19 0.00 0.36 3.2 0.19 0.29 0.41* SL3 2.72 0.43 -0.06 0.19 3.2 0.19 0.02 0.14 SL3A 3.2 0.23 0.34 0.57* SL4 3.1 0.21 0.29 0.53* mean 2.83 0.21 -0.02 0.41 3.2 0.15 0.31 0.52 Note: sd = standard deviation * = significant (p<0.01)

Discussion

Anthropometry The stature of the senior elite male (179.9 + 6.4 cm) and female (165.9 + 4.5 cm) triathletes of this study was compared with that presented by Ackland et al. (1998a) for similar competitors of the 1997 TWC (179.8 + 6.2 cm & 168.3 + 4.6 cm, respectively). Body mass of the 2000 TWC competitors (69.3 + 6.1 kg & 55.5 + 4.4 kg) was slightly less than those of 1997 for males and females, respectively (72.3 + 6.0 kg & 59.5 + 4.8 kg). However, as determined via t-tests (p <0.01), the difference was not significant, and further research is required to ascertain whether the triathlete body shape is 135 continuing to evolve as the sport of triathlon develops beyond its infant stages. Better performing triathletes tended to have lower, although not significantly so, levels of adiposity which equate to a lower body mass (Landers et al., 2000). It is possible that, if other anthropometric variables were measured and compared with the 1997 triathletes, there could be some variables such as skinfolds that have evolved slightly over time. As adipose mass accounted for 20-25% of total body mass (Landers et al., 1998), a significant change in skinfold measures might not divulge a significant change in total body mass. Over time, the body shapes of athletes, swimmers, runners and field game athletes have altered (Carter & Heath 1990; Štepnicka, 1986). The results indicate a need to follow up the initial data in order to compare the current body shapes and compositions of elite triathletes with their predecessors.

Female Triathletes The female triathletes recorded weak, negative correlations between SR and body mass at all data collection points except the SR1 where no correlation was evident with the mean (r = -0.44). This reveals that heavier athletes adopted a slower SR. A similar significant correlation was recorded between mean SR and height (r = -0.46) suggesting that the taller female triathletes chose a slower SR. Combining both sets of data, the taller and heavier triathletes used a lower SR.

No significant correlation was evident between the SL used by the senior female triathletes and body mass. However, a significant and positive correlation (r = 0.41) existed between mean SL and height. That is, the taller female triathletes took longer strides during the race. It is possible that the taller athletes with longer legs could take longer strides at a moderate SR which did not make excessive metabolic demands (Landers et al., 1998).

Male Triathletes The male triathletes recorded similar results as their female counterparts with regard to SR and body size. The heavier and taller male triathletes used a slower SR (r = -0.39 & -0.42, respectively (p < 0.01)). A weak but positive correlation existed between SL and mass (r = 0.31). The taller male triathletes also used longer strides than their shorter competitors (r = 0.52). Again, the taller triathletes typically have longer legs and the ability to take longer strides to cover more distance at a moderate energy efficient SR. 136

Instinctively there is a significant, positive correlation between height and mass. That is, those who are taller take longer strides and are generally heavier. However, it can not be assumed that body mass causes the selected SL.

The results indicated a stronger correlation between SL and body size, than SR with body size. Data from Landers et al. (2001a, study 4) showed that the SL used during competition was a better predictor of run time and final finishing position than SR. The better performers were those who maintained a longer and more consistent SL (Table 5.15 & 5.16). Results from study 1 indicated that low adiposity and longer lever lengths were the greatest anthropometric predictors of triathlon performance (Table 3.4). Hence, the link between body size and shape with selected SL, and thus performance, becomes clearer. It could be coincidental that longer strides are taken by the heavier athletes. That is, it is more likely that the longer strides result from taller athletes being larger and having a greater mass. However, Landers (1998) showed that proportional limb lengths of athletes were better predictors of performance than absolute size. That is, those who had longer levers, in the form of legs and arms, appeared to have a mechanical advantage as they performed better during the event. As the results of this study do not provide a perfect correlation between height and SL, it could be that the RLLL is a better predictor of SL. That is because some of the shorter triathletes might have absolute or relatively longer legs than the taller competitors. Further study is required to clarify this assumption.

Landers et al. (2001a, study 4) noted the small variation in SRs used throughout the event. No significant differences were recorded between the superior and average runners in both the male and female categories. The lack of significant differences between SRs used by athletes in differing finishing positions might suggest that the weak negative correlation between SR and body size is the result of the individual triathletes selecting the most efficient SR/SL combination. This is shown by the negative relationship between body size and SR and the positive relationship with SL. That is, those who are taller and heavier use longer strides and a slower cadence.

As all subjects in this study were elite competitors in a world championship, it could be that the homogeneity of the group masks the influence of body size and shape. If an ideal kinanthropometry model exists, these athletes would be close to a representative 137 cluster. Hence, further analysis of these parameters from a larger and wider range of abilities could help to clarify the matter.

Similar results have been obtained from swimmers, where anthropometric variables accounted for a greater variation in stroke length (89%) than stroke rate (41%) (Grimston and Hay, 1986).

Summary This study began to link together the physical morphological structure of triathletes with running mechanics and their performances. The running SRs and SLs were measured during triathlon competition, and body mass and height measured prior to competition. Senior, elite triathletes competing in the 2000 TWC were investigated and a correlation matrix developed to determine relationships between the variables. The SRs were negatively related to body size, whereas SLs were positively related. That is, the taller and heavier triathletes used longer strides and a lower SR. It could be hypothesised that the triathletes self-selected themselves into the sport of triathlon based on optimal body height and mass measures for males and females. This is because each gender, though homogeneous in height and mass, showed significant correlations between anthropometry and performance. Further research is warranted to examine the effects of cycling on the subsequent run discipline during triathlon. 138

Chapter 7

Studies 7 & 8 The effect of cycle cadence on subsequent running

Abstract Previous studies in this thesis have indicated that the run portion of the triathlon could be the most significant discipline in determining final race outcome (study 2). Physical characteristics (study 1) are important along with the selection of the correct SR/SL combination (study 5) because SL has been shown to be significantly related to run performance (study 4). However, studies 3, 4, 5 and 6 did not define conclusively how the first kilometre of running impacted on the total running performance based on possible variations in running mechanics and energy cost, or the effects of prior cycle cadence. Previous research has provided some suggestions regarding the effects of cycling on the subsequent run during triathlon competition.

Hence, this series of studies sought to examine laboratory simulated physiological and biomechanical changes that occur during the cycle-to-run transition and cycling cadence on initial run performance. In order to remove fatigue as a factor, cycling was completed in an unloaded fashion for a period of 6 min. Six cadences were presented to the triathletes in random order (0, 50, 70, 90, 110, 130 rpm) and this was followed by a 5 min run at 75% maximal aerobic speed (MAS). The average transition time was less than 35 s. During the cycle and run HR and O2 were measured and SR, SL, hip and knee ROM were measured during the run only.

A linear increase in oxygen cost and HR were recorded for the cycle in parallel with increasing cadence. The oxygen cost during the first minute of running following cycling at 130 rpm was significantly greater than after any other cycle cadence. A significant increase in muscle activity (up to 6 times) during cycling was recorded via EMG traces.

139

The average SR was significantly greater during the first 60 s of running after unloaded cycling at 110 and 130 rpm (88 rpm) when compared with running without prior cycling (82 rpm). Data from EMG showed greater muscle activity during running after cycling at higher cadences (20% - 85% EMG activity). No change in hip and knee ROM while running was evident with cadence variation. Due to the large amount of data, this study has been presented in two parts. Chapter 7 investigates the physiological variables and Chapter 8 examines the biomechanical data following a common introduction.

140

Introduction The art of changing from one physical activity to another with minimal biomechanical, physiological or mental disturbance on the new activity creates the greatest challenge in the sport of triathlon. Previous literature is unclear as to individual responses during running following a cycling (Gohlitz & Witt, 1993; Hausswirth et al., 1996 & 1997; Hue et al., 1998; Landers et. al, 2001a; Pfutzner et al. 1997; Quigley & Richards, 1996).

With the relatively new draft-legal format during classic distance events (1.5 km swim, 40 km cycle, 10 km run), race tactics and physiological demands on the athlete have changed. Recent research (Landers et al., 2001b) has shown that swim ability and placement in the first pack of swimmers when exiting the water greatly enhances the probability of a triathlete finishing in the top 10. The eventual race winner will come from this lead pack 80% of the time. Others (Landers, 1998; Landers et al., 2000; Millet & Vleck, 2000) have shown significant correlations between position at the end of the cycle discipline with the final finishing position. Hence, with the arrival of the first pack of swimmers into the final transition of cycle to run, those who are least affected by the previous exercise and attain an efficient running pattern earlier will increase their chances of winning (Landers et al., 2001b).

The limited data available suggest that a decreased SL at the start of a run following cycling, with no change of SR is that most commonly adopted by triathletes (Hausswirth et al., 1996 & 1997). Pfutzner et al. (1997) showed that 70% of triathletes were approximately 10% slower than their normal 10 km running velocity for the initial 500 - 1000 m but did not stipulate if this was the result of varying SR, SL or both. However, others have shown no change in either SR or SL (Hue et al., 1998; Quigley & Richards, 1996) or a decrease in SR (Witt, 1993). Landers et al. (2001a) analysed both male and female senior elite competitors in the 2000 TWC and found little variation in SR during the run portion of the race. They concluded that the initial data collection point might have been too far from the bike to run transition in order to precisely compare results from previous studies. In the same study (Landers et al., 2001a), SL was found to be associated with running performance and overall triathlon finishing position. That is, those who maintained a longer SL, ran faster.

141

Research generally has shown that, with increasing fatigue during running, there is a decrease in SL and, if velocity is to be maintained, SR must be increased (Ackland, 1997; Elliott et al., 1981; Elliott & Ackland, 1980; Elliott & Roberts, 1980; Hausswirth et al., 1997, Williams et al., 1991). This suggests that the cycling leg prior to running also induces fatigue as a factor and is a possible cause of the decrease in SL. However, as this decrement in SL is generally not for the entire run, but rather only for the first stages (Pfutzner et al., 1997), fatigue might not be the only factor contributing to these biomechanical changes. Landers et al. (2001a) also found a decline in SL for the slower runners, both male and female, in the latter stages of the triathlon run. This was attributed to both physiological fatigue and a mental acceptance that their final finishing positions would not be in the first half of the field.

Several studies (De Vito et al., 1995; Guezennec et al., 1996; Hausswirth et al., 1996, 1997, 1999; Hue et al., 1998; Kreider et al., 1988; Lehénaff et al., 1998) have shown that, after cycling, there is a decrease in efficiency, or greater oxygen cost of running following cycling, when compared with just running. This increase in oxygen cost could be the result of fatigue (Hausswirth et al., 1996 & 1997) or possibly due to using an inefficient SR/SL relationship for the given velocity (Cavanagh & Williams, 1982; Holt et al., 1991).

Hausswirth et al. (1999b), Hue et al. (1998) and Lehénaff et al. (1998) also have shown that, if drafting is permitted, there is less decrement in run performance. Drafting can reduce the energy cost of cycling by up to 30% (Faria, 1992). This would conserve energy and enable a faster run, or permit an increase in cycle speed, or a combination of both.

Hausswirth et al. (1997) showed that, after cycling, a significant increase in trunk flexion angle was noted at foot strike (3.5o - 2.1o) when compared with just running, and running in a fatigued state at the end of a marathon (0o). Average trunk flexion during the first five minutes of running was also significantly different from that of just running (6.6o), and running at the end of a marathon (10.3o). Other kinetic variables such as hip or ankle vertical oscillations, and thigh and knee angles were found to be different between running, and running preceded by cycling (Hausswirth et al., 1997; Quigley & Richards, 1996).

142

Increasing oxygen cost is associated with an increase in cadence at unloaded or low workloads (Hagberg et al., 1981). Faria et al. (1982) reported similar findings with national team cyclists. At low power outputs (800 kg·m-1·min-1), a linear increase in oxygen consumption occurred with an increase in cycle cadence (60, 100 & 130 rpm). However, this increase was not evident at high power outputs (1800 kg·m-1·min-1) where no significant differences in oxygen cost were reported.

Several authors (Marsh & Martin, 1995; Neptune et al., 1997; Suzuki et al., 1982; Takaishi et al., 1998) have reported changes in EMG patterns of lower limb musculature during cycling at various cadences. Typically, there is a forward shift in the muscle activation during the pedal revolution as cadence increases. The EMG patterns usually maintain their shape but become shorter in duration.

Therefore, this study set out to determine whether the fatigue induced by prior cycling causes changes in subsequent running, or whether some other underlying factors such as the prior movement patterns and/or neurological firing are responsible. Hence, only short periods of unloaded cycling were used to try and induce specific firing patterns without any associated fatigue.

The first part of the study sought to confirm whether there is a linear relationship between cadence and oxygen cost in unloaded cycling. In addition, the aim was to show whether, although there is an increase in oxygen consumption per minute with higher cadences, the cost per pedal stroke remained unchanged. The study also examined whether any relationship existed between SR or SL with height or body mass for age-group triathletes as this previously was shown to be so for senior elite triathletes (study 6). Finally, the effect of cycling cadence on the chosen SR/SL combination during running was examined where no prior fatigue was present as a result of the cycling.

Williams (1985) summarised the differences between over-ground and treadmill running by collating all previous research. At speeds of less than 5 m/s the mechanics of running did not significantly differ. Five metres per second corresponds to a performance time of 33:20 for 10 km. Data from the 1997 triathlon world

143 championships revealed that most elite males would run slightly faster (32:02 + 1:36). However, most elite females are slightly slower (36:37 + 2:10) (Landers, 2000). In this study of age-group triathletes, none of the subjects ran at speeds greater than 5 m/s. It was hypothesised that the extended time period of repetitive cycling at any given rate would influence the SR selected during the initial stages of the subsequent run.

144

Study 7 Physiological effects of cadence alteration on cycling and subsequent running

Methodology

Sample The subjects in this study included 20 competitive triathletes (14 males and 6 females), with a mean age of 21 y. All subjects had a minimum of 2 years triathlon experience and had completed a classic distance triathlon in less that 2 h 20 min for males and 2 h 30 min for females. Subject details can be viewed in Tables 7.1, 7.2 and 7.3.

Testing protocol Testing was conducted in two sessions, separated by a minimum of two days. Upon arrival, subjects signed an informed consent form (Appendix L) where all possible effects of the study were outlined. Then, mass, height and age were recorded, and lengths of lower limb, leg, upper limb, arm and forearm were measured (Appendix F). Training history and triathlon racing involvement were also recorded (Appendix M).

A continuous, incremental running test was performed to exhaustion on a motorised treadmill. The test began at 12 km/h for 9 min and velocity was increased by 1 km/h every 2 min. Female triathletes commenced at 10 km/h for the first 4 min then increased 1 km/h every 2 min. Measurements of O2 were undertaken every 15 s, using a Morgan Ventilation Monitor and Applied Electrochemistry Oxygen Analyzer S-3A/I and Carbon Dioxide Analyzer CD-3A, to allow determination of maximal oxygen consumption and maximal aerobic speed (MAS) (Hausswirth et al., 1997).

During the second testing session, subjects were marked to allow easy identification of ankle, knee, hip and shoulder landmarks. The location of these landmarks were based on ISAK standards (Norton and Olds, 1996).

The EMG electrodes were placed over the muscle bellies of biceps femoris, vastus medialis, medial gastrocnemius and tibialis anterior. An earth wire was placed over the

145 head of fibula. All electrodes were placed on the left side of the subject. Each site was first shaven to remove hair and then abraded to remove dry skin. Finally, the site cleaned using an alcohol swab. An inter-electrode distance of 2 cm was maintained for each muscle. The impedance was then tested and, if greater than 10 kΩ, the site was prepared again. Knee joint angles were monitored via an electro-goniometer (1000 Hz) to compare muscle activation at various portions of the cyclic movements.

The electrodes and the goniometer were passed through a grass amplifier before being recorded on the computer via a data acquisition and analysis software (Pearce, 1996). An 8 s epoch at a sampling rate of 200 Hz was used to collect EMG data. This was undertaken at the end of the third, fourth and fifth minute during the cycle and then at 0, 30 s, 2 min, 3 min and 4 min during the run.

Following a quiet period of 5 min seated on the bicycle, subjects ran at 75% MAS for 5 min to determine baseline measures. Each subject then undertook a series of five cycle- run transitions presented in a random order. The cycle was performed without any load, in order to determine the effect prior movement rates without the possible effect of fatigue due to cycling with a set load. This was achieved by removing the chain of the subject's own bike mounted on a wind trainer. The cycle was for 6 min at one of five cadences (50, 70, 90, 110 & 130 rpm). Cadence was maintained within +/- 2 rpm during the cycle. This was ensured via visual feed back from a cycle computer mounted on the subjects’ handlebars (Cateye® Astrale, model CC-CD100 N). Immediately following the cycle, the subject performed a 'transition' within 35 s to prepare for a 5 min run on a treadmill at a pace equivalent to 75% MAS. A 6 min rest period followed (or until HR returned to resting levels) before attempting subsequent trials. A video camera (25 Hz) was placed laterally 5 m from the subject to allow continuous visual recording of cycling, transition and running.

Data collection Oxygen consumption, carbon dioxide expiration, minute ventilation and heart rate were measured throughout the tests via a metabolic cart. The EMG of four muscles was also measured to determine when the muscles were active during both cycling and running. Cadence during the cycle portion was set and subjects had a visual display in order to maintain the appropriate cadence for the required time. SR was determined by

146 measuring the time to take six strides. As the velocity of running is known, the SL also can be determined as velocity/SR.

Video cameras were used to record data during both cycle and run sections of the test. All biomechanical variables were measured at 0, 30 s, 1 min, 2 min and 4 min of the run. The digitising was undertaken via the computer program Silicon Coach (2001).

Data analysis Cycle The average oxygen cost for each subject was determined over the final 2 min of the cycle. Using a Pearson product moment correlation, this was correlated with cadence to determine whether a linear relationship existed between cadence and oxygen cost. A second step was to determine the oxygen cost per pedal stroke across each of the five cycle cadences to determine any relationships.

Run Average oxygen cost was determined across each run and correlated with the cycle cadence used prior to the run. The oxygen cost during the first minute of running after the transition was also determined and correlated with prior cycle cadence. Significance was set at p < 0.01.

Results

Subject anthropometric, physiological and training characteristics can be viewed in tables 7.1, 7.2 and 7.3. The height and mass of the subjects are comparable with those of the 1997 TWC (Ackland et al., 1998a; Landers, 1999). The crural index (CI) of male triathletes (110.4) were significantly greater than those reported by Landers (1999) (102.0, p < 0.01 ). All other variables were similar to senior elite triathletes.

147

Table 7.1 Descriptive anthropometric data for male, female subjects.

Male Female Combined Mean sd Mean sd Mean sd Mass Kg 73.46 5.68 60.63 2.83 69.61 7.85 Height cm 181.64 5.64 169.3 1.92 177.94 7.30 Trochanteric height cm 94.57 4.39 91.07 3.11 93.52 4.24 Tibiale laterale height cm 49.60 2.11 45.67 1.34 48.42 2.59 Trochanterion-tibiale cm 44.97 2.37 45.40 1.78 45.10 2.15 Relative Lower Limb Length % 52.05 1.46 53.78 1.23 52.57 1.56 Crural Index % 110.37 2.37 100.61 1.01 107.44 4.89

Table 7.2 Physiological characteristics of males, females and all triathletes combined. Male Female Combined Mean sd Mean sd Mean sd ml·kg- O2max 1·min-1 64.36 7.52 52.05 2.9 60.67 8.66 MAS km/h 19.71 1.38 15.63 1.00 18.47 2.31 75% MAS km/h 14.69 0.60 11.70 0.52 13.79 1.54 HR max bpm 193.86 5.24 202.33 6.51 196.4 6.67

Senior elite female triathletes during the 2000 TWC (study 4 & 5) ran at a velocity between 13.7 to 18.7 km/h and the senior elite male triathletes reported a range a speeds from 14.4 to 19.8 km/h. These velocities are greater than the 75% MAS attained by the subjects of this study.

Table 7.3 Weekly training. Swim Cycle Run Distance km 6.64 153.33 28.79 Time h 3.14 5.58 2.79 Sessions n 2.57 3.17 3.28

Observation of training data indicates less training was undertaken when compared with senior elite triathletes.

Physiological response to unloaded cycling Figure 7.1 shows that, with increasing cadence, there is an associated increase in energy expenditure based on a higher oxygen cost. The increase remains linear between 50 and 90 rpm but then it increases steeply between 90 and 130 rpm. Significant increases in oxygen consumption were recorded between all measures as cadence increased (p<0.002) (Table 7.4). This is supported by subsequent analysis of oxygen cost for each

148 pedal revolution. No significant differences were recorded between the oxygen cost for each pedal revolution between 50 rpm and 90 rpm. Above 90 rpm, the oxygen cost for each revolution of the pedals significantly increases (Table 7.4). Figure 7.2 reveals similar HR trends to oxygen cost but less significant differences exist. The mean HR at rest was significantly less than for any other trial. The HR at 110 rpm was significantly higher than all other HR measures except at 130 rpm. At 130 rpm, the mean HR was significantly greater than the HR at 110 rpm (Table 7.4).

Table 7.4 Mean oxygen cost and heart rate values for cycling at various cadences. Oxygen Cycle cadence Oxygen cost sig dif cost/rev sig dif HR sig dif ml·kg-1·min- ml·kg-1·rev- rpm 1 1 bpm 0 5.06 Na 65.0 50 7.63 a 0.153 70.0 a 70 9.32 a,b 0.133 77.0 a 90 12.09 a,b,c 0.134 80.0 a 110 17.89 a,b,c,d 0.163 b,c,d 96.0 a,b,c,d 130 28.43 a,b,c,d,e 0.219 b,c,d,e 123.0 a,b,c,d,e Note: a = significantly different from 0 rpm b = significantly different from 50 rpm c = significantly different from 70 rpm d = significantly different from 90 rpm e = significantly different from 110 rpm p < 0.01

Figure 7.1 Oxygen cost of unloaded cycle at different cadences.

5 oxygen cost (ml/kg/min)

0 0 50 70 90 110 130 cadence (rpm)

149

Figure 7.2 Heart rate response to different unloaded cycle cadences.

140

120

100

60 HR (bpm) 40

0 0 50 70 90 110 130 cadence (rpm)

Physiological response to running following cycling Oxygen consumption values were recorded during the first minute of running after cycling and again during the fifth minute. The HRs were recorded at the commencement of each minute. The data are presented in Tables 7.5 and 7.6, and Figures 7.3 and 7.4. The oxygen cost of running, directly after cycling at 110 rpm and 130 rpm, was significantly greater than at any other time. The oxygen cost during the first minute of running was significantly less when no prior cycling was undertaken when compared with unloaded cycling at 90, 110 and 130 rpm.

Table 7.5 Mean oxygen cost at the start and end of run after different cycle cadences. Cycle cadence Oxygen cost sig dif Oxygen cost sig dif run 1 min run 5 min ml·kg-1·min- ml·kg-1·min- rpm 1 1 0 15.59 45.01 50 15.42 46.27 70 18.86 46.73 a 90 20.84 a,b 46.93 110 21.65 a 46.20 130 26.95 a,b,c,d 47.07 Note: a = significantly different from 0 rpm b = significantly different from 50 rpm c = significantly different from 70 rpm d = significantly different from 90 rpm e = significantly different from 110 rpm

150

Figure 7.3 Oxygen cost of running after unloaded cycling at various cadences.

50.00 45.00 40.00 35.00 30.00 run 1 min 25.00 run 5 min 20.00 15.00 10.00

oxygen cost (ml/kg/min) 5.00 0.00 0 50 70 90 110 130 cycle cadence (rpm)

Table 7.6 Mean heart rate values during running after unloaded cycling at various cadences. Cycle cadence HR HR HR HR HR HR run 0 min run 1 min run 2 min run 3 min run 4 min run 5 min Rpm bpm bpm bpm bpm bpm bpm 0 108.5 151.2 156.6 159.1 161.2 161.3 50 119.4 153.1a 159.8 163.5 165.6 165.5 70 121.4 154.8 160.6 163.4 165.7 166.3 90 123.6 155.0 161.0 164.1 166.4 166.8 110 131.8a 155.2 159.5 162.3 164.7 165.0 130 148.3a,b,c,d,e 159.3 163.7 165.3 168.5 169.8 Note: a = significantly different from 0 rpm b = significantly different from 50 rpm c = significantly different from 70 rpm d = significantly different from 90 rpm e = significantly different from 110 rpm p < 0.01

Directly after the unloaded cycle at 130 rpm, the running HR values were significantly greater than at any other cadence. The initial HR following cycling at 110 rpm was also significantly higher than that recorded at the commencement of the run without prior cycling. No significant differences existed at between HRs at any other time point.

151

Figure 7.4 Heart rate values at each minute after different unloaded cycling cadences.

180

170

160 0 min 150 1 min 2 min 140 3 min

HR (bpm) 130 4 min 5 min 120

110

100 0 50 70 90 110 130 cycle cadence (rpm)

Discussion Physiological response to unloaded cycling

The linear trend of an increasing O2 and HR between 50 rpm and 90 rpm was expected because, although there was no change in workload, there was an increase in the number of movements performed over the set time period of one minute. Analysing the relationship between oxygen consumed for each pedal revolution revealed that, up to 90 rpm, a similar oxygen cost is associated with each revolution. Above this cadence, greater energy is required for each pedal revolution while pedalling at 110 rpm and 130 rpm. This could be due to a lack of training at cadences greater than 90 rpm and that the triathletes were inefficient at these leg speeds. Also, perhaps there was insufficient time for the muscle to relax between contractions. This might be evident when viewing EMG traces at various cadences.

Faria et al. (1982) compared the efficiency of cycling at various cadences (60, 100 & 130 rpm) across two workloads (800 & 1800 kg·m-1·min-1). The oxygen consumption at 800 kg·m-1·min-1 revealed similar results to the present study; showing a linear

152 increase in oxygen cost with increasing cadence. The increase with unloaded cycling was greater than that with a workload. This is highlighted at high workloads where cadence has negligible effect on efficiency.

The HR data shows very little change from 50 rpm to 90 rpm. At these times, O2 is less than 25% of maximum oxygen consumption. Hence, the required increase in cardiac output might be provided as a result of a larger increase in SV rather than HR. The HR at 110 rpm and 130 rpm show significant increases when compared with cycling at other cadences. This could be similar to the results of oxygen cost where lack of skill at these leg speeds creates inefficiency. Alternatively, the SV could reach maximal volumes and an increased HR is required to increase cardiac output.

Physiological response to running following cycling Oxygen consumption was measured during the first minute of running following cycling and during the final minute. The oxygen cost at 5 minutes with no prior cycling of 45 ml·kg-1·min-1 was 75% of O2max (60.7 ml·kg-1·min-1) thereby indicating that the running speed selected was 75% MAS.

At the commencement of the run, O2 was lower than that measured during the final minute of the run. Greater oxygen consumptions were recorded between running after cycling at 90 rpm, 110 rpm and 130 rpm than running with no prior activity. The triathletes also consumed significantly more oxygen while running after cycling at 110 rpm and 130 rpm than after any other cadence. There was a trend that, with increasing cycle cadence, there was an increase in O2 at the start of the run. This is possibly due to the greater energy expenditure during cycling at higher cadences and insufficient recovery time during the transition. Similar results were evident for HR data with HR at the commencement of the run, after cycling at 130 rpm, being significantly greater than after cycling at any other cadence.

No significant differences were recorded at the completion of the run except between 90 rpm and rest. This suggested that, within 5 min, the triathletes reached a steady state equivalent to that obtained when running without prior cycling. Again, the HR data revealed similar results to the gas analysis. After the first 60 s of running, no significant

153 differences were evident at any time for each of the trials. This might be a result of a trained response of the athletes who are able to judge the workload quickly and set a required HR within 60 s. Further analysis of this issue would be valuable if comparisons were made with untrained subjects. Previous investigations have shown that running after cycling is not as efficient as running without prior activity. It could be that this is due to fatigue as similar results are reported for running comparing the beginning and end of a marathon (Hausswirth et al., 1997).

Although the findings of the present study do not support the theory of fatigue causing decreased efficiency, it does suggest other factors such as, motor control of running movements, that could contribute to changes in physiological response to running because subjects commenced running with minimal fatigue. It is possible that the different movement patterns of cycling has its own effect on running, irrespective of workload. It also adds support to the importance of the cycle to run adaptation on final performance. Researchers who had collected data after the first kilometre during the run may not have been successful in finding significant differences in variables as the athlete probably achieved normal running patterns within 3-5 minutes.

Summary This study sought to determine the physiological changes occurring with triathletes during a cycle to run transition when the cadence during the cycle was varied. The cycling was also performed in an unloaded state in order to remove the possible effects of fatigue and to view more closely the nature of prior movement patterns on subsequent movement efficiency.

The results of this study showed that with increasing cycle cadence there is a linear increase in oxygen cost which when related to work done shows little difference between oxygen cost per pedal stroke at cadences 50 to 90 rpm. Oxygen consumption was significantly greater at the commencement of the run after cycling at higher cadences (110 & 130 rpm). However, after 5 minutes of running there were no differences in HR or O2 measures. This indicated the significant, but short-lived effect, of prior cycling on subsequent running.

154

Chapter 8

Study 8 The effect of cycle cadence on running mechanics

Methodology See study 7, chapter 7.

Sample The subjects included 14 male and 6 female triathletes. Subject details can be viewed in Tables 7.1, 7.2 and 7.3.

Data Analysis Cycle Muscle activation and inactivation times were determined via EMG for tibialis anterior, gastrocenemius, vastus medials and biceps femoris muscles at each of five pedal frequencies during cycling. Resting EMG was recorded during the first trial while subject remained seated on the bicycle. Muscle activation was deemed to have occurred when EMG measures exceeded three times resting levels. Comparisons were made between cadences based on the actual times of activity and the relative percentage time of activity and inactivity. This is because, as cadence increases, the total time for each revolution will decrease and lead to a decrease in muscle activity time, recovery time, or both.

Run The SR and SL were analysed during each run to determine whether any changes occurred during the first 1500 m after cycling. A similar analysis was also conducted at each data collection point in order to determine if, as a result of using different cycling cadences, any other biomechanical changes in gait occurred such as hip or knee ROM. These angles were determined with respect to the vertical and calculated from a mean of five strides.

155

In order to compare the results from study 6, SR and SL data were correlated (Pearson Product Moment, p < 0.01) with triathlete anthropometric variables. Thus, conclusions could be drawn regarding the consistency of results for various ability levels and to further divulge any other physical characteristics that may contribute to running mechanics. The muscle activation patterns for each muscle during running were compared over time and across cycle cadence. These times were converted to percentages of each stride because with a change in SR there will be a change in either the absolute activation time, relaxation time or both. This comparison was done via a two-way, repeated measures ANOVA (p < 0.01).

Results

Stride rate and stride length The SR selected by the triathletes following cycling at 110 and 130 rpm was significantly greater during the first 60 s of running. This coincided with a decrease in SL as velocity remained constant. The SR was significantly greater at 0 min and 0.5 min after cycling at 110 rpm and at 0 min and 1 min after cycling at 130 rpm. At 2 min and 4 min, there were no significant differences between the SRs for any condition (Table 7.7, Figure 7.5).

Table 8.1 Selected running stride rate of triathletes following unloaded cycling at various cadences. Time Cadence (rpm) (min) 0 50 70 90 110 130 0 82.4 83.3 83.3 81.1 87.2a,b,c 88.2a,b,c 0.5 82.4 82.4 83.3 83.3 85.7a,b,c,f 83.3 1 82.4 82.4 80.6 82.0 84.3 85.7a,b,c,d 2 82.4 84.3 83.3 83.3 83.3 83.3 4 82.4 82.0 81.5 82.0 82.0 82.0 p < 0.01 a = significantly different from 0 rpm b = significantly different from 50 rpm c = significantly different from 70 rpm d = significantly different from 90 rpm e = significantly different from 110 rpm

156

The relationship between body size with SR and SL are presented in Table 8.2. Significant and negative correlations were reported between SR and each of height, trochanteric height, tibiale laterale height and femur length (p < 0.01). Stride length was significantly and negatively correlated with body mass and height (p <0.05).

Figure 8.1 Running stride rate following unloaded cycling at various cadences.

90.0

88.0

86.0 0 min 84.0 0.5 min 1 min 82.0

Run SR 2 min 80.0 4 min 78.0

76.0 0 50 70 90 110 130 unloaded cycle cadence (rpm)

Table 8.2 Correlations of stride rate and stride length at the end of the run with anthropometric variables. Trochanteric Tibiale Femur Mass Height height laterale ht length RLLL CI SR -0.13 -0.86** -0.79** -0.79** -0.76** -0.39 0.29 SL -0.59* -0.61* -0.38 -0.50 -0.25 0.02 -0.38 Note: * = significant correlation (p < 0.05) ** = significant correlation (p < 0.01)

Significant differences were noted between thigh and leg angles during the first minute of running following unloaded cycling (Table 8.3). However, when these two segments were combined to form the knee angle, no differences were recorded at heel strike or toe off. Both hip and knee angles, at the commencement of the run after cycling at various cadences, were not significantly different (Figure 8.2 and 8.3). 157

Table 8.3 Trunk, thigh and leg angles during the first minute of running following unloaded cycling at various cadences. trunk angle thigh angle leg angle Cadence heel strike toe off heel strike toe off heel strike toe off 0 4.50 5.93 23.17 -23.92 12.20 -45.33 50 3.39 5.52 25.92 -19.33 10.39 -45.94 70 3.92 5.61 24.70 -21.87 12.53 -43.00 90 4.38 5.63 25.00 -22.75b,c 11.33 -45.08b,c 110 3.50 4.67 24.56 -19.11b,c,d 9.97 -46.67b,c,d 130 3.25 4.50 23.53 -18.85b,c,d,e 10.33 -48.02b,c,d,e Note: p < 0.01 a = significantly different from 0 rpm b = significantly different from 50 rpm c = significantly different from 70 rpm d = significantly different from 90 rpm e = significantly different from 110 rpm

Figure 8.2 Average knee angle during the first minute of running following unloaded cycling at various cadences.

180

160

140 0 rpm 50rpm 70 rpm 120 90 rpm angle 110 rpm 100 130 rpm

60 1 2 3 4 5 6 7 8 9 10111213141516171819 hs time to

Note: hs = heel strike to = toe off

158

Figure 8.3 Average hip angle during the first minute of running following unloaded cycling at various cadences.

195.00

185.00 0 rpm 175.00 50 rpm 70 rpm 165.00 90 rpm angle 110 rpm 155.00 130 rpm

145.00

135.00 12345678910111213141516171819 hs time to

Note: hs = heel strike to = toe off

EMG during cycling The muscle activation during cycling can be seen in Table 8.4 and Figures 8.4, 8.5, 8.6, 8.7. The results indicate no change in gastrocnemius activity (~35%) during the pedal cycle. However, significant increase is noted at 110 and 130 rpm for tibialis anterior and vastus medialis. A linear increase is evident for biceps femoris. Tibialis anterior vastus medialis and biceps femoris had very short activation times at lower cadences (less than 10% of revolution).

159

Table 8.4 Muscle activity during unloaded cycling at various cadences. Gastrocn Tibialis Vastus Biceps Cadence emius Anterior Medialis Femoris % act %inact % act %inact % act %inact % act %inact 50 33.4 66.6 6.7 93.3 28.4 71.6 6.6 93.4 70 33.1 66.9 4.2 95.8 7.1 92.9a 5.6 94.4 90 30.6 69.4 3.8 96.2 3.9 96.1a 12.5 87.5 110 30.4 69.6 20.0 80.0a,b,c 30.1 69.9b,c 14.8 85.2 130 45.0 55.0 38.4 61.6a,b,c,d 31.1 68.9b,c 19.0 81.0 Note: p < 0.01 a = significantly different from 50 rpm b = significantly different from 70 rpm c = significantly different from 90 rpm d = significantly different from 110 rpm % act = percent of time active %inact = percent of time inactive

Figure 8.4 Tibialis Anterior activity during unloaded cycling at various cadences.

100%

80%

60% ta %inact ta % act 40%

20%

0% 50 70 90 110 130

160

Figure 8.5 Gastrocnemius activity during unloaded cycling at various cadences.

100% 90% 80% 70% 60% gas %inact 50% gas % act 40% 30% 20% 10% 0% 50 70 90 110 130

Figure 8.6 Vastus Medialis activity during unloaded cycling at various cadences.

100%

80%

60% vm %inact vm % act 40%

20%

0% 50 70 90 110 130

161

Figure 8.7 Biceps Femoris activity during unloaded cycling at various cadences.

100% 90% 80% 70% 60% bf %inact 50% bf % act 40% 30% 20% 10% 0% 50 70 90 110 130

EMG during running No significant differences were noted at any time during the first four minutes of running during any cadence condition for muscle activity time. Hence, muscle activation for each condition was composed of the entire data collection for each cadence (Table 8.5, Figures 8.8, 8.9, 8.10, 8.11). Greater muscle activity was evident during the run after cycling at higher cadences. Tibialis anterior activity level remained relatively constant (40-50%) after cycling at cadences 0-110 rpm. However, a significant increase (p < 0.01) was recorded after cycling at 130 rpm with this muscle being active for 75% of each stride. Gastrocnemius, revealed a tendancy of increasing muscle activity during running after cycling with increasing cadence. Gastrocnemius, like tibialis anterior, was active for a significantly greater time during each stride after cycling at 130 rpm. Vastus medialis revealed an increasing tendancy of muscle activation time up to 70 rpm before significantly increasing after cycling at 110 and 130 rpm. Finally, biceps femoris also showed a tendency for increasing muscle activation time with data collected after cycling at higher cadences of 110 and 130 rpm (p < 0.01). Data for all muscles show that, after cycling at 162

130 rpm, the time each muscle is active during each stride is approximately double that of running with no prior cycling.

Table 8.5 Muscle activation during running of biceps femoris, tibials anterior, gastrocnemius and vastus medialis after unloaded cycling at various cadences. Biceps Femoris Tibialis Anterior Gastrocnemius Vastus Medialis Cadence Active Inactive Active Inactive Active Inactive Active Inactive 0 17.4 82.6 43.8 56.2 32.4 67.6 28.5 71.5 50 20.5 79.5 a 46.8 53.2 34.7 65.3 41.7 58.3 a 70 17.4 82.6 b 51.3 48.7 a 36.6 63.4 53.2 46.8 a,b 90 18.2 81.8 43.5 56.5 c 50.6 49.4 47.4 52.6 a,b 110 29.9 70.1 a,b,c,d 41.0 59.0 c 48.4 51.6 62.7 37.3 a,b,c,d 130 39.6 60.4 a,b,c,d 74.6 25.4 a,b,c,d,e 60.4 39.6 a,b,d 82.6 17.4 a,b,c,d Note: p < 0.01 a = significantly different from 0 rpm b = significantly different from 50 rpm c = significantly different from 70 rpm d = significantly different from 90 rpm e = significantly different from 110 rpm

Figure 8.8 Tibialis Anterior activity during running after unloaded cycling at various cadences.

100% 90% 80% 70% 60% inactive 50% active 40%

% time active 30% 20% 10% 0% 0 50 70 90 110 130 cycle cadence (rpm)

163

Figure 8.9 Gastrocnemius activity during running after unloaded cycling at various cadences.

100% 90% 80% 70% 60% inactive 50% active 40%

% time active 30% 20% 10% 0% 0 50 70 90 110 130 cycle cadence (rpm)

Figure 8.10 Vastus Medialis activity during running after unloaded cycling at various cadences.

100% 90% 80% 70% 60% inactive 50% active 40%

% time active 30% 20% 10% 0% 0 50 70 90 110 130 cycle cadence (rpm)

164

Figure 8.11 Biceps Femoris activity during running after unloaded cycling at various cadences.

100% 90% 80% 70% 60% inactive 50% active 40%

% time active 30% 20% 10% 0% 0 50 70 90 110 130 cycle cadence (rpm)

Discussion

Stride rate and stride length SR was significantly elevated (84-89 rpm vs 82 rpm) during the first 30 – 60 s of running following unloaded cycling at 110 and 130 rpm. Within two minutes, no differences were evident between any cadence conditions. Perhaps after cycling, where the legs were rotating at a rate significantly greater (110 and 130 rpm) than the preferred run SR (82 rpm), that the movement patterns were somewhat distorted. Therefore, the legs inevitably moved faster at the commencement of the run. These results tend to support the findings of Scholz and Kelso (1989) who reported less stable hand coordination movements after higher frequency movements and also when movements were out of phase.

Analysis of the segments that comprise the SL, thigh and leg mechanics, did not clearly explain the changes in SL and SR. No changes in the running trunk flexion angle were recorded after cycling at any cadence. This was in contrast to Hausswirth et al. (1997) who reported a significant increase in trunk angle at foot strike after cycling when compared 165 with running without prior activity. Hausswirth et al. (1997) also found that average trunk flexion was greater at the end of an extended run when compared with the commencement of the run, thereby suggesting fatigue was a factor. In the present study, the triathletes were cycling for 6 min in an unloaded fashion and fatigue should not have been a factor. It also could demonstrate that cycling posture did not affect running mechanics. The results also supported previous findings that no significant differences were noted for knee and hip angles (Figure 8.2 & 8.3) (Hausswirth et al., 1997; Quigley & Richards, 1996). However, the combinations of leg and thigh angles at the commencement of running did significantly change at toe off as the unloaded cycle cadence increased.

EMG The EMG data during cycling did not follow results of earlier studies which had suggested a forward shift in the muscle activation; that is, muscles becoming active earlier during the pedal revolution. However, the findings of a shorter absolute duration of EMG and maintenance of activation patterns were in agreement with previous studies (Marsh & Martin, 1995; Neptune et al., 1997; Suzuki et al., 1982; Takaishi et al., 1998). This could be the result of the cycling in this study being conducted in an unloaded fashion and muscles were not required to be as active due to the lack of resistance. Hence, normal cycling skill was not necessary as subjects were only ‘pushing’ down on the pedals rather than actually cycling. Significantly greater vastus medialis muscle activation time was revealed when cycling at 50 rpm. Perhaps this muscle is used to control the speed of the cranks and is co-contracting.

The EMG data supported the suggestion that the prior movement pattern of cycling affects the subsequent movement pattern of running. After cycling at higher cadences of 110 and 130 rpm, absolute and percent muscle activation time was increased and tended not to follow normal firing patterns for running without prior cycling. This could be ‘confusion’ at either the central or peripheral nervous system level when changing from the discipline of cycling to that of running because the muscles of the LL are used differently. Also, it could result from the increased firing rate during cycling at higher cadences which could have a residual effect on the run. If the first 60 s of running without prior cycling (0 rpm) 166 are compared with the data from 50 – 90 rpm there are no significant differences in EMG readings. This could negate the possibility that it is the different activity (cycling) which is conducted prior to running, that causes the increased EMG activity after cycling at 110 and 130 rpm. It could be that a cadence greater than the preferred rate causes greater disturbance to subsequent movement patterns than initial rates that are slower. This is possibly supported by the EMG data during unloaded cycling at various cadences because very few differences were recorded in muscle activation time at cadences 90 rpm and below. Hence, the possible subsequent neural disturbance would not occur. It may be that if a loaded cycle was presented to the triathletes at various cadences, different EMG data would have been recorded during the initial run. Another way to test the hypothesis that it is the movement velocity, and not the prior activity, causing the initial increased SR and EMG activity, would be to subject the triathletes (or runners) to a bout of running at various pre-selected SRs before allowing them to freely choose their SR/SL combination.

Stride rate, electromyography and oxygen cost In the present study, the increase in SR during the first minute after unloaded cycling at 110 and 130 rpm is likely a result of the higher cadence during the unloaded cycle. It might not be due solely to a decrease in SL due to fatigue as others have reported (Hausswirth et al., 1996 & 1997). This is supported by the increased activity time of LL muscles during unloaded cycling at higher cadences and then during subsequent running. However, the continuing high activity time of muscle activity after the triathletes have reached their preferred SRs could be the result of possible lagging.

Similar results were reported for the physiological data where significant increases in the oxygen cost of cycling at higher cadences and subsequent running were evident. This could be the result of a number of factors. Firstly, a lack of cycling skill has been noted as a reason for increased oxygen cost when cycling with increasing cadence (Marsh & Martin, 1993). A lack of skill could be supported by the increased activity time of LL muscles (Marsh & Martin, 1995 &1997). That is, the LL muscles are active for a greater period of time during each pedal revolution at higher cadences due to a lack of skill. However, this 167 also helps to account for the greater oxygen cost as more oxygen is required to maintain increased muscle activity.

The effects of cycling are then carried across to the subsequent run. The oxygen cost of running during the first minute of running was significantly greater after cycling at higher cadences. So too was the level of muscle activity time revealed by EMG and an increased SR. However, the rate at which these variables returned to levels attained without prior cycling were different. The variation in SR returned to the preferred rate after running for

60 s. The O2 remained elevated after the first minute but within 4 minutes, no significant differences were evident across trials. This was probably due to a lag in O2 response. Finally, EMG data did not reveal a significant change across the data collection points and further research is required to pursue this question further. Only then could one determine whether EMG would return similar values to those obtained without prior cycling.

Summary The results of this final study highlight some of the effects of changing from one physical activity to another. More specifically, the cycling cadence which is used does affect running mechanics when performed sequentially. It can be concluded that cycling at higher cadences induces an increased running SR during the first 60 s. The reasons behind this finding are not clear-cut but perhaps the changing motor patterns or the increased firing rate of muscles plays a role. The EMG results showed greater muscle activity during cycling at higher cadences and during the initial stages of the run after cycling at higher cadences.

168

Chapter 9

Summary, Conclusions and Recommendations for Further Study

Summary The sport of triathlon has been in existence for 30 years and has undergone many changes in that time. These changes have included alterations to race distances, discipline order and event rules which have impacted on competitors. This thesis examined the body shapes of senior and junior elite competitors; the importance of each discipline within triathlon; cycling cadence, running SR and SL during competition; the relationship between SR, SL and body size; and the effect of altering cycling cadence on subsequent running.

Study 1 The opening study of this thesis sort to determine how the physical characteristics of elite triathletes related to triathlon performance by examining senior and junior, male and female triathletes. Elite triathletes were faster than their junior counterparts and showed less variation in performance times. Run time variation was the largest of the component disciplines and tended to show the importance of this discipline to the final outcome.

Following a factor analysis, the four distinguishable morphological factors that emerged were; robustness, adiposity, segmental lengths and skeletal mass. Relating these factors to the total time obtained by the triathletes in this study yielded a regression equation that correlated significantly with all triathletes and accounted for 47% of the variance in total triathlon duration. The regression equations illustrated the importance of low levels of adiposity for elite triathletes for total time and most of the subdiscipline times. The other factor that showed importance was that proportionally longer segmental lengths contributed to successful swimming outcome.

169

Study 2A The purpose of the second study (2A and 2B) were to investigate the relative importance of each of the three sub-disciplines in draft legal triathlon and the influence of pack formation in the cycle leg. The results of this study tend to confirm the subjective views of triathlon coaches, that the swim leg of the triathlon is more important than previously thought, especially since the introduction of draft legal events. In 90% of elite male and 70% of elite female races, the eventual winner exited the water in the first pack. Results also showed that, if a small group of swimmers (less than 10) do break away during the swim, the possibility of these triathletes finishing in the top 10 is even greater (47% male & 61% female). This is compared with the top 10 prospects of triathletes in larger packs exiting the water (21% male). It was found that 50% of male and 8% of female starters exited the water with the first pack of swimmers.

These results have training implications for both swimming and running. A triathlete must develop swimming ability to be able to maintain a pace fast enough to swim with the front pack. There is also a need to develop early speed and lactate tolerance to start the swim hard and be positioned with the lead pack from the outset. Both aspects can be worked on relatively easily through traditional swimming training. Finally, the ability to run well 'off the bike' in order to improve a top 10 position into winning, is an area that is not well understood and invites further investigation.

Study 2B The greatest amount of variation in performance is found in running times when examining both the raw SD and the percentage SD. The swim leg appears to be the second most important discipline based on findings from study 2A and the large variation percentage SD in swim times. There are few changes in position during the cycle leg, hence the importance of being in the first pack to exit the water and being able run quickly off the bike is increased.

Study 3 Research has shown that prior cycling has a physiological effect on a subsequent run caused by fatigue. Similarly, there could be a biomechanical effect such as body posture SR or SL. This research sought to determine whether running SR was affected 170 by prior cycling during a triathlon competition and what relationship cadence had on triathlon performance. The sample consisted of 20 male and 10 female finishers of an Australian selection race for the 2000 Triathlon World Championships.

Cycle cadence was measured at the commencement and completion of the 40km cycle discipline and SR at three points, 0.13 km, 3.7 km and 9.95 km during the 10 km run. Split times for each of the disciplines and total time was also recorded. Results indicated that significantly higher cycle cadences occurred at the end of the cycle when compared with the start, and higher run SRs occurred at the beginning of the run for both male and female competitors. Initial SR was similar to the initial cycling cadence for both male and females. Typically, the faster triathletes were those who recorded higher cadences and SRs. Further study is required to clarify the role that the cycle leg plays in influencing running mechanics and if the same pattern is utilised by the elite level triathletes.

Study 4 The previous study (3) examined the cadence during the first and last km of the cycle leg of a classic distance triathlon. A significantly faster cadence was noted at the end of the cycle when compared with the start for both male and female age group competitors. The initial cycle cadence also was comparable with run SR.

This study sought to determine the cycling cadence and running SR and SL used during triathlon competition by senior elite male and female competitors. Data were collected from 46 female and 51 male senior elite finishers of the 2000 Triathlon World Championships via video and analysed using Video Expert II Coach.

Results for both males and females showed a relatively consistent cadence being used throughout the run and cycle disciplines. The female triathletes used a significantly faster cadence at the commencement and completion of the cycle than during the middle stages. Variations in running SL appear also to be associated with running ability and triathlon performance. That is, the better performers maintained a longer and more consistent SL.

171

Study 5 The aim of this study was to determine if difference cycling cadence and running SR and SL used during triathlon competition could lead to better performances or predict race outcome. Data were collected from 46 female and 51 male senior elite finishers of the 2000 Triathlon World Championships via video and analysed using Video Expert II Coach.

Results for both males and females showed a relatively consistent cadence being used throughout the run and cycle disciplines. The female triathletes used a significantly faster cadence at the commencement and completion of the cycle. Variations in running SL appear to be correlated with running ability and triathlon performance. Those who maintained a longer and more consistent SL performed better.

Study 6 The results of this study have begun to link together the physical structure of triathletes with running mechanics and their performances. The SR and SL were measured during competition, and body mass and height prior to the competition. Both male and female senior elite triathletes of the 2000 TWC were investigated. Correlations were undertaken in order to determine the relationship between the variables. The SRs were negatively related to body size, whereas SLs were positively related. That is, the taller and heavier triathletes used longer strides and a lower SR.

This study also could add weight to the argument that athletes are self-selecting into the sport of triathlon based on optimal body height and mass measures for males and females. This is because each gender, though homogeneous in height and mass, showed significant relationships which became stronger when the two groups were combined. Further research is warranted to clarify the effects of cycling on the subsequent run discipline during triathlon.

Study 7 The first part of this study sought to determine the physiological changes occurring with triathletes during a cycle to run transition when the cadence during the cycle was varied. The cycling was also performed in an unloaded state in order to remove the possible effects of fatigue and to view more closely the nature of prior movement patterns on 172 subsequent movement efficiency. Results showed that increasing cycle cadence caused a linear increase in oxygen cost. When this was related to work done, it showed little difference between oxygen cost per pedal stroke at cadences of 50 to 90 rpm. Oxygen consumption was significantly greater at the commencement of the run after cycling at higher cadences (110 & 130 rpm). However, after 5 min of running, there were no differences in HR or O2 measures. This indicated a significant but short-lived effect of prior cycling on subsequent running.

Study 8 The final study of this thesis highlighted some of the effects of changing from one activity to another. More specifically, the cycling cadence that was used affects running mechanics when performed sequentially.

It was concluded that cycling at higher cadences induced an increased running SR during the first 60 s. The reasons behind this finding are not clear but it could be that the changing motor patterns and/or the increased firing rate plays a role. This is supported by the EMG results which showed greater muscle activity during cycling at higher cadences and during the initial stages of the run after cycling at higher cadences.

Conclusions During the course of this research several new findings, which were directly related to triathlon and triathletes, were revealed. These may apply to the wider areas of other endurance activities where subsequent movement patterns may be altered or affected by prior movements. For example, during an individual medley in swimming, changing from backstroke to breaststroke or breaststroke to freestyle. Not only are the movements very different, but they are also performed at different frequencies.

Initially the triathletes were characterised physically and how morphology related to performance. Size, shape and composition were determined for senior and junior elite triathletes. As with other endurance-based activities excess body mass in the form of adiposity was found to be detrimental to performance. Those triathletes with proportionally longer limbs appeared to have an advantage. 173

With the physical structure known, the important facets of the event were determined. Through the use of Mathmatica® the importance of each discipline were determined. Unlike previous research the swim was highlighted as being very important in draft legal events in order to give the athlete a chance of winning. However, it was concluded that the run is the most important discipline in the present draft legal format of triathlon.

Knowing the above, the aspects of the cycle to run transition were next investigated. The cadence during the cycle and the subsequent running stride rate and stride length were measured during competiton, compared to performance and then altered in the laboratory. It was found that triathletes utilise a similar cycle cadence to running stride rate and that during the final few kilometres of cycling, cadence is increased. Those who maintain a consistent SL performed better. This relationship is important as it prompted further investigation into triathlete size and SR/SL selection, which revealed larger triathletes used longer strides.

Finally it was found that the increases in cadence at the end of cycle discipline does have an affect on the subsequent run both physiologically and biomechanically. And leads to further investigation outlined below.

Recommendations for further study • Repeat study 1 as body shapes and composition could have changed in the short period of time. Other sports have shown changes in athlete’s kinanthropometry over time (Carter & Heath, 1990; Stepnicka, 1986). As triathlon is a new sport with developing rules and training principles this may lead to changes in athlete size and shape. • Find results of elite athletes pre-drafting and compare to the results of study 2 or compare the results of the best age group triathletes of today, which are not permitted drafting, with draft legal events. This could help to highlight how the sport has changed and gain knowledge into the important aspect of the race. 174

• Repeat study 6 comparing trained and untrained subjects or elite triathletes. It may be that constant repetition of the cycle to run transition has altered EMG signals. This may be highlighted with those who have not undertaken such a sport or in those who have done it many more times. • Alter bike setup to change the cycling mechanics in order to improve running following cycling. Some coaches suggest a cycle posture similar to running so movement patterns are the same for an easier transition, while others oppose this thought as it may cause undue fatigue to running muscles. • Conduct similar studies on the transition between swimming and cycling. As noted previously (chapter 2) some coaches suggest an increase in kicking speed at the conclusion of the swim to increase blood flow to the legs. Will this improve the beginning of the cycle or probably more importantly the transition? Or will it cause a decrement in performance? 175

References

Ackland, J. (1997). Swim Bike Stride. NZ Triathlete 61, 9-11.

Ackland, T.R. (1994). Techniques in the measurement and assessment of body composition. In: Exercise and Obesity, A.P. Hill and M.L. Wahlquist (Eds.), Smith and Gordon, London, pp. 165-166.

Ackland, T.R., Blanksby, B.A., Landers, G. and Smith, D. (1998a). Anthropometric profiles of elite triathletes. Journal of Science and Medicine in Sport, 1(1), 53-56.

Ackland, T., Blanksby, B., Landers, G. and Smith, D. (1998b). Anthropometric correlates with performance among world championship triathletes. In: K. Norton, T. Olds, and J. Dollman, (Eds.) Kinanthropometry VI, Proceedings of the Sixth Scientific Conference of the International Society for the Advancement of Kinanthropometry, 92- 104.

Ahlquist, L.E., Bassett, D.R., Sufit, R., Nagle, F.J. and Thomas, D.P. (1992). The effect of pedalling frequency on glycogen depletion in type I and type II muscles fibres during submaximal cycling exercise. European Journal of Applied Physiology, 65, 360- 364.

Alexander, R.M. (1984). Walking and running. American Scientist, 72, 348-354.

Atwater, A.E. (1990). Gender differences in running. In Biomechanics of Distance Running. P.R. Cavanagh (Ed.) Human Kinetics Publishers, Champaign, IL, pp. 321- 362.

Berry, M.J., Pollock, W.E., van Nieuwenhuizen, K. and Brubaker, P.H. (1994). A comparison between aero and standard racing handlebars during prolonged exercise. International Journal of Sports Medicine, 15(1), 16-20.

Bloomfield, J., Ackland, T.R. and Elliott, B.C. (1994). Applied Anatomy and Biomechanics in Sport. Melbourne, Australia: Blackwell Scientific Publications, pp 40- 92 and 267-109.

Bloomfield, J. and Sigerseth, P.O. (1965). Anatomical and physiological differences between sprint and middle distance swimmers at the university level. Journal of Sports Medicine and Physical Fitness, 5, 76-81.

Brisswalter, J., Hausswirth, C., Vallier, J.M., Smith, D., Vercruyssen, F. and Vallier, J.M. (2000). Energetically optimal cadence vs freely chosen cadence during cycling: effect of exercise duration. International Journal of Sports Medicine 20(5), 1-5.

Carter, J.E.L. (Ed). (1982). Physical Structure of Olympic Athletes part 1. Montreal Olympic Games Anthropological Project. Karger: Basel, pg, 25-52.

Carter, J.E.L. and Ackland, T.R. (Eds). (1994). Kinanthropometry in Aquatic Sports; a study of world class athletes. USA: Human Kinetics Publications Inc.

176

Carter, J. and Heath, B. (1990). Somatotyping – Development and Applications. Cambridge University Press, Cambridge, UK.

Cateye ® Astrale, model CC-CD100 N, Cateye co.ltd., Osaka, Japan.

Cavanagh, P.R. and Lafortune, M.A. (1980). Ground reaction force in distance running. Journal of Biomechanics, 13, 397-406.

Cavanagh, P.R. and Williams, K.R. (1982). The effect of stride length variation on oxygen uptake during distance running. Medicine and Science in Sport and Exercise, 14(1), 30-35.

Chapman, A.E and Caldwell, G.E. (1983). Kinetic limitations of maximal sprinting speed. Journal of Biomechanics, 16, 79.

Chatard, J., Chollet, P. and Millet, G. (1998). Performance and drag during drafting swimming in highly trained triathletes. Medicine and Science in Sports and Exercise, 30(8), 1276-1280.

Chatard, J., Senegas, X., Selless, M., Dreanot, P. and Geyssant, A. (1995). Wetsuit effect: a comparison between competitive swimmers and triathletes. Medicine and Science in Sports and Exercise, 27(4), 580-586.

Coast, J.R. and Welch, H.G. (1985). Linear increase in optimal pedal rate with increased power output in cycle ergometry. European Journal of Applied Physiology, 53, 339-342.

Cox, M.H., Miles, D.S., Verde, T.J., Nessenthaler, G. and Heinze, J. (1994). Influence of pedal frequency on the lactate threshold of elite cyclists Medicine and Science in Sports and Exercise, 26(5), S67.

Craig, A.B. and Pendergast, D.R. (1981). Relationship of stroke rate, distance per stroke and velocity in competitive swimming. In: Flavell, E.R. (1981) Biokinetics Strength Training, volume one; Application notes. Isokinetics Publications, Albany, CA, USA. pp79-84.

Davoren, B. and Gregory, J. (1999). The effects of three varied cycle leg protocols on self-paced run time trial performance. Proceedings from; The Gatorade International Triathlon Science II Conference, Maximising Olympisc Distance Performance, Noosa, Australia, 88-98. de Lucas, R.D., Balikian, P., Neiva, C.M., Greco, C.C. and Denadai, B.S. (2000). The effects of wetsuits on physiological and biomechanical indices during swimming. Journal of Science and Medicine in Sport, 3(1), 1-8.

Deitrick, R.W. (1991). Physiological responses of typical versus heavy weight triathletes to treadmill and bicycle exercise. The Journal of Sports Medicine and Physical Fitness, 31(3), 367-375.

177

Dengel, D., Flynn, M., Costill, D. and Kirwin, J. (1989). Determinants of success during triathlon competition. Research Quarterly for Exercise and Sport, 60(3), 234-238.

De Vito, G., Bernardi, M., Sproviero, E. and Figura, F. (1995). Decreased endurance performance during Olympic distance triathlon. International Journal of Sports Medicine, 16, 24-28

Dern, R.J. Levene, J.M. and Blair, H.A. (1947). Forces exerted at different velocities in human arm movements. American Journal of Physiology, 151, 415-437.

Dolan, K.P. (1987). The physical characteristics of elite triathletes Unpublished Masters Thesis, San Diego State University, San Diego, CA.

Drinkwater, B.L. (1986). (Ed). Female Endurance Athletes. Illinois: Human Kinetics Publications Inc.

Drinkwater, D.T. and Ross, W.D. (1980). The anthropometric fractionation of body mass. In: Kinanthropometry II, M. Ostyn, G. Beunen and J. Simons, (Eds.) University Park Press, Baltimore.

Edwards, R.H.T., Young, A., Hosking, G.P. and Jones, D.A. (1977). Human skeletal muscle function: description of tests and normal values. Clinical Science and Molecular Medicine, 52, 283-290.

Elliott, B. C. and Ackland, T. R. (1980). Biomechanical effects of fatigue on 10,000 meter running technique. Research Quarterly for Exercise and Sport, 52(2), 160-166.

Elliott, B.C. and Blanksby, B.A. (1976). A cinematographic analysis of overground and treadmill running by males and females. Medicine and Science in Sports and Exercise, 8(2), 84-87.

Elliott, B.C., Le Rossignol, P. and Pyke, F.S. (1981). Biomechanical adjustments of runners to prolonged continuous running and interval games training. Journal of Human Movement Studies, 7, 99-107.

Elliott, B.C. and Roberts, A.D. (1980). Biomechanical evaluation of the role of fatigue in middle distance running. Canadian Journal of Applied Science, 5, 203-207.

Faria, I.E. (1992). Energy expenditure, aerodynamics and medical problems in cycling – an update. Sports Medicine, 14(1), 43-63.

Faria, I, Dix, C. and Frazer, C. (1978). Effect of body position during cycling on heart rate, pulmonary ventilation, oxygen uptake and work output. Journal of Sports Medicine, 18, 49-56.

Faria, I., Sjojaar, G. and Bonde-Petersen, F. (1982) Oxygen cost during different pedalling speeds for constant power output. Journal of Sports Medicine and Physical Fitness 22(3), 295-299.

178

Faulkner, J.A., Roberts, D.E., Elk, R.L. and Conway, J. (1971). Cardiovascular responses to submaximum and maximum effort cycling and running. Journal of Applied Physiology, 30(4), 457-461.

Flynn, M.G., Costill, D.L., Kirwin, J.P., Mitchell, J.B., Hournard, J.A., Fink, W.J., Beltz, J.D. and D’Acquisto, L.J. (1990). Fat storage in athletes: Metabolic and hormonal response to swimming and running. International Journal of Sports Medicine, 11(6), 433-440.

Foley, J.P., Bird, S.R. and White, J.A. (1989). Anthropometric comparisons of cyclists from different events. British Journal of Sports Medicine, 23, 30-33.

Formby, P. (1989). The somatotype of triathletes as related to performance. The Australian Journal of Physiotherapy, 35(3), 195-196.

Garside, I. and Doran, D.A. (2000). Effect of bicycle frame ergonomics on triathlon 10- km running performance. Journal of Sports Sciences, 18, 825-833.

Gnehm, P., Reichenbach, S., Altpeter, E., Widmer, H. and Hoppeler, H. (1997). Influence of different racing positions on metabolic cost in elite cyclists. Medicine and Science in Sports and Exercise, 29(6), 818-823.

Gonzalez, H. and Hull, M.L. (1988). Multivariable optimization of cycling biomechanics. Journal of biomechanics 22, 1151-1161

Gotshall, R.W., Bauer, T.A. and Fahrner, S.L. (1996). Cycling cadence alter exercise hemodynamics. International Journal of Sports Medicine, 17(1), 17-21.

Grimston, S.K. and Hay, J.G. (1986). Relationships among anthropometric and stroking characteristics of college swimmers. Medicine and Science in Sports and Exercise, 18(1), 60-68.

Guezennec, C.Y., Vallier, J.M., Bigard, A.X. and Durey, A. (1996). Increased energy cost of running at the end of a triathlon. European Journal of Applied Physiology, 73, 440-445.

Hagan, R.D., Weis, S.E. and Raven, P.B. (1992). Effect of pedal rate on cardiorespirtory responses during continuous exercise Medicine and Science in Sports and Exercise, 24(10), 1088-1095

Hagberg, J.M., Mullin, J.P., Giese, M.D. and Spitznagel, E. (1981). Effect of pedalling rates on submaximal, exercise responses of competitive cyclists. Journal of Applied Physiology, 51, 447-451.

Hausswirth, C. (1996). Variability in energy cost of running at the end of a triathlon and a marathon. International Journal of Sports Medicine, 17(8), 572-579.

Hausswirth, C., Bigard, A.X. and Guezennec, C.Y. (1997). Relationship between running mechanics and energy cost of running at the end of a triathlon and marathon. International Journal of Sports Medicine, 18, 330-339.

179

Hausswirth, C. (1999). Effect of exercise duration on the self selected cadence in triathletes: interpretation from neuromuscular fatigue. (personal communication, November 1999).

Hausswirth C., Lehénaff, D., Dreano, P and Savonen, K. (1999b). Effects of cycling alone or in a sheltered position on subsequent running performance during a triathlon. Medicine and Science in Sports and Exercise, 31(4), 599-604.

Hausswirth C., Vallier, J.M., Lehénaff, D., Smith D., Millet, G., Dreano, P and Patou, E. (1999c). Effects of an alternate or continuous sheltered position in cycling on consecutive running performance during a sprint triathlon. Preceedings of: 4th Annual Congress of the European College of Sports Science.

Hay, J.G. (Ed.) (1993). Biomechanics of Sports Technique, 4th Edition. Prentice Hall: Englewood Cliffs, New Jersey. pp. 356-357, 396-402, 412.

Heil, D.P., Wilcox, A.R. and Quinn, C.M. (1995). Cardirespiratory responses to seat- tube angle variation during steady-state cycling. Medicine and Science in Sports and Exercise, 27(5), 730-735.

Holly, R.G., Barnard, R.J., Rosenthal, M., Applegate, E. and Pritikin, N. (1986). Triathlete characterisation and response to prolonged strenuous competition. Medicine and Science in Sports and Exercise, 18(1), 123-127.

Holt, K.G., Hamill, J. and Andres, R.O. (1991). Predicting minimum energy cost of human walking. Medicine and Science in Sports and Exercise, 25, 491-498.

Hoyt, D.F. and Taylor, C.R. (1981). Gait and the energies of locomotion in horses. Nature, 292, 239-240.

Hue, O., LeGallais, D., Chollet, D., Boussana, A. and Prefaut, C. (1998). The influence of prior cycling on the biomechanical and cardiorespirtory response profiles during running in triathletes. European Journal of Applied Physiology, 77, 98-105.

Hull, M.C., Gonzalez, H.K. and Redfield, R. (1988). Optimisation of pedalling rate in cycling using a muscle stress-based objective function. International Journal of Sports Biomechanics, 4, 1-20.

Ito, A., Komi, P.V., Sjodin, B., Bosco, C. and Karlsson, J. (1983). Mechanical efficiency of positive work in running at different speeds. Medicine and Science in Sports and Exercise, 15(4), 299-308.

International Triathlon Union Competition Rules (1999). www.worldsport.com/worldsport/sports/triathlon/comp_rules/comprules99_3htm1#cycl 1999 Worldsport Networks Ltd.

Kerr, D.A. and Ross, W.D. (1994). An anthropometric method for a five-way fractionation of human body mass. Unpublished Masters Thesis, Simon Fraser University, Canada.

180

Khort, W., Morgan, D., Bates, B. and Skinner, J.S. (1987). Physiological response of triathletes to maximal swimming, cycling and running. Medicine and Science in Sports and Exercise, 19(1), 51-55.

Khort, W., O’Connor, J.S. and Skinner, J.S. (1989). Longitudinal assessment of responses by triathletes to swimming, cycling and running. Medicine and Science in Sports and Exercise, 21(5), 569-575.

Knuttgen, H.G. (1961). Oxygen uptake and pulse rate while running with undetermined and determined stride lengths at different speeds. Acta Physiol. Scanda., 52, 366-371.

Kreider, R.B., Boone, T., Thompson, W.R., Burkes, S. and Cortes, C.W. (1988). Cardiovascular and thermal responses of triathlon performance. Medicine and Science in sports and exercise 20(4), 385-390.

Landers G.J. (1998). Kinanthropometry and performance of elite triathletes. Unpublished Honours Thesis, The University of Western Australia.

Landers, G.J., Blanksby, B.A. and Ackland, T.R. (2001a). Cadence selection and performance. Submitted: International Triathlon Coaching Symposium, July 23-24 2001, Edmonton, Canada, p.13

Landers, G.J., Blanksby, B.A., Ackland, T.R. and Smith, D. (1999). Kinanthropometry and performance of elite triathletes. The Book of Abstracts: Fith IOC World Congress, 31st October -5th November 1999, Sydney Australia, p.129.

Landers, G.J., Blanksby, B.A., Ackland, T.R. and Monson, R. (2001b). Swim Postion and its effect on triathlon outcome. Submitted: International Triathlon Coaching Symposium, July 23-24 2001, Edmonton, Canada, p.11.

Landers, G.J., Blanksby, B.A., Ackland, T.R. and Smith, D. (2000). Morphology and performance of world championship triathletes. Annals of Human Biology, 27(4), 387- 400.

Laurenson, N.M., Fulcher, K.Y and Korkin, P. (1993). Physiological characteristics of elite and club level female triathletes during running. International Journal of Sports Medicine, 14(8), 455-459.

Leake, C.N. and Carter, J.E.L. (1991). Comparisons of body compositions and somatotypes of trained female triathletes. Journal of Sports Sciences, 9, 125-135.

Lehénaff, D., Hausswirth, C., Dreano, P. and Savonen, K. (1998). The impact of bike- riding in a drafting position on the subsequent run performance during the triathlon. 11 Congres International De Triathlon De L’Insep. (Conception et realisation: Insep- publications).

Mathmatica (2000)

181

Marais, G., Weissland, T., Robin, H., Vanvelcenaher, J.M., Lavoie, J.M. and Pelayo, P. (1999). Physiological effects of variation in spontaneously chosen crank rate during submaximal and supramaximal upperbody exercise. International Journal of Sports Medicine, 20, 239-245.

Marino, G.W and Goegan, J. (1993). Work-energy analysis of triathletes running under bike/run and run only conditions. Proceedings of the XIth symposium of the International Society of Biomechanics in Sport, Amherst, MA, pg 86-88.

Marsh, A.P. and Martin, P.E. (1997). Effect of cycling experience, aerobic power, and power output on preferred and most economical cycling cadences. Medicine and Science in Sports and Exercise, 29(9), 1225-1232.

Marsh, A.P. and Martin, P.E. (1995). The relationship between cadence and lower extremity EMG in cyclists and non-cyclists. Medicine and Science in Sports and Exercise, 27(2), 217-225.

Marsh, A.P. and Martin, P.E. (1993). The association between cycling experience and preferred and most economical cadences. Medicine and Science in Sports and Exercise, 25(11), 1269-1274.

Martin, P.E. and Marsh, A.P. (1992). Step length and frequency effects on ground reaction forces during walking. Journal of Biomechanics, 25, 1237-1239.

Mero, A. and Komi, P.V. (1986). Force-,EMG-, and elastic-velocity relationships at submaximal, maximal and supramaximal running speeds in sprinters. European Journal of Applied Physiology, 55, 553-561.

Millet, G.Y., Candau. R. and Rouillon, J.D. (1997). Cycle length and crank rate in roller skiing: relationship with performance and maximal lower limb power. Journal of Human Movement Studies, 32, 267-281.

Millet, G.P. and Vleck, V.E. (2000). Physiological and biomechanical adaptations to the cycle and run transistion in Olympic triathlon: review and practical recommendations for training. British Journal of Sports Medicine 34, 384-390.

Milner-Brown, H.S. and Stein, R.B. (1975). The relationship between the surface electromyogram and muscular function. Journal of Physiology, 246, 549-569.

Miura, H., Kitagawa, K. and Ishika, T. (1997). Economy during a simulated laboratory test triathlon is highly related to Olympic distance triathlon. International Journal of Sports Medicine, 18, 276-280.

Moritani, T., Takaishi, T. and Matsumoto, T. (1993). Determination of maximal power output at neuromuscular fatigue threshold. Journal of Applied Physiology, 74(4), 1729- 1734.

Moulds, R.F.W, Young, A., Jones, D.A. and Edwards, R.H.T. (1977). A study of the contractility, biochemistry and morphology of an isolated human skeletal muscle Clinical Science and Molecular Medicine, 52, 291-297.

182

Napier (1967) Chapter 3. The Antiquity of Human Walking. In: Vertebrate Adaptations, Scientific America, pg 38-48.

Neptune, R.R., Kautz, S.A. and Hull, M.L. (1997). The effect of pedal rate on coordination in cycling. Journal of Biomechanics, 30(10), 1051-1058.

Norton, K. and Olds, T. (Eds). (1996). Anthropometrica. University of New South Wales Press, Sydney, Australia. pp 27-75, 409.

Olds, T.S., Norton, K.I., Lowe, E.L.A., Olive, S., Reay, F. and Ly, S. (1995). Modeling road cycling performance. Journal of Applied Physiology, 78, 1596-1161.

Origenes IV, M.M., Blank, S.E. and Schoene, R.B. (1993). Exercise ventilatory response to upright and aero-posture cycling. Medicine and Science in Sports and Exercise, 25(5), 608-612.

O’Toole, M.L., Douglas, E., Hiller, B., Crosby, L.O. and Douglas, P.S. (1987). The ultraendurance triathlete: a physiological profile. Medicine and Science in Sports and Exercise, 19(1), 45-50.

O’Toole, M.L., Douglas, P.S. and Hiller, W.B.D. (1989a). Lactate, oxygen uptake and cycling performance in triathletes. International Journal of Sports Medicine, 10(6), 413- 418.

O’Toole, M.L., Douglas, P.S. and Hiller, W.D.B. (1989b). Applied physiology of a triathlon. Sports Medicine, 8(4), 201-225.

Pearce, A. (1996). AP30 program. Department of Human Movement and Exercise Science, The University of Western Australia.

Pfutzner, A. and Grosse, S. (1997). Linking training 1er Colloque International Sur L’entrainment en Triathlon Le Insep, France.

Price, D. and Donne, B. (1997). Effect of seat tube angle at different seat heights on submaximal cycling performance in man Journal of Sports Sciences, 15, 395-402.

Pugh, L.G.C.E. (1971). The influence of wind resistance in running and walking and the mechanical efficiency of work against horizontal and vertical forces. Journal of Physiology, 213, 255-276.

Quigley, E.J. and Richards, J.G. (1996). The effects of cycling on running mechanics. Journal of Applied Biomechanics, 12(4), 470-479.

Ross, W.D. and Kerr, D.A., (1991). Fraccionament de la massa corporal: un nou methode per utilitzar en nutricio clinica i medicina esportiva. APUNTS, 18, 175-187.

Ross, W.D. and Marfell-Jones, M. (1991). Kinanthropometry. In: Physiological Testing of the High Performance Athlete, (second edition) J.D. MacDougall, H.A. Wenger and H.J. Green (Eds.) Human Kinetics Books, Champaign, IL. pp 223-308.

183

Ross, W.D. and Ward, R. (1986). Scaling anthropometric data for size and proportionality. In: Kinanthropetry III, Proceedings of the VIII Commonwealth and International Conference on Sport, Physical Education, Dance, Recreation and Health, T. Reily, J. Watkins and R. Borms (Eds.) E. & F.N. Spon. New York, USA. pp 85-91.

Rowbottom, D.G., Keast, D., Garcia-Webb, P. and Morton, A.R. (1997). Training adaptation and biological changes among well-trained male triathletes. Medicine and Science in Sports and Exercise, 29(9), 1233-1239.

Ryschon, T.W. and Stray-Gundersen, J. (1991). The effect of body position on the energy cost of cycling. Medicine and Science in Sports and Exercise, 23(8), 949-953.

Schneider, D.A., Lacroix, K.A., Atkinson, G.R., Troped, P.J. and Pollack, J. (1990). Ventilatory threshold and maximal oxygen uptake during cycling and running in triathletes. Medicine and Science in Sports and Exercise, 22(2), 257-264.

Scholz, J.P. and Kelso, J.A.S. (1989). A quantitative approach to understanding the formation and change of coordinated movement patterns. Journal of Motor Behavior, 21(2), 122-144.

Sheel, A.W., Lama, I., Potvin, P., Coutts, K.D. and McKenzie, D.C. (1996). Comparison of aero-bars versus traditional cycling postures on physiological parameters during submaximal cycling. Canadian Journal of Applied Physiology, 21(1), 16-22.

Sjogaard, G. (1977). Force-velocity curve for bicycling. In: E. Asmussen and K. Jorgensen (Eds.) Biomechanics VI-A. University Park Press, Baltimore, pp 93-98.

Sleivert, G.G. and Rowlands, D.S. (1996). Physical and physiological factors associated with success in triathlon. Sports Medicine, 22(1), 8-18.

Sleivert, G.G. and Wenger, H.A. (1993). Physiological predictors of short course triathlon performance. Medicine and Science in Sports and Exercise, 25(7), 871-876.

Sloane E. (1980) Biology of Women New York: John Wiley and Sons.

Smith D., Lee, H., Pickard, R., Sutton, B. and Hunter, E. (1998). Competition analysis and its implications for coaching and training. Proceedings of The Gatorade Triathlon Science I Seminar, Central Queensland University, Noosa Australia, pg 24-28.

SPSS for Windows Release 8.0.0 (1997). SPSS Inc., USA.

Stepnicka, J. (1986) Somatotype in relation to physical performance, sports, and body posture. In: Kinanthropometry III: VIII Commonwealth and international conference on sport, physical education, dance, recreation and health. ROYAUME-UNI, Glasgow, pp 39-52

Suzuki, S., Wantanabe, S. and Homma, S. (1982). EMG activity and kinematics of human cycling movements at different constant velocities. Brain Research, 240, 245- 258.

184

Takaishi, T., Yasuda, Y., Ono, T. and Moritani, T. (1996). Optimal pedalling rate estimated from neuromuscular fatigue for cyclists. Medicine and Science in Sports and Exercise, 28(2), 1492-1497.

Takaishi, T., Yamamoto, T., Ono, T., Ito, T. and Moritani, T. (1998). Neuromuscular, metabolic and kinetic adaptations for skilled pedaling performance in cyclists. Medicine and Science in Sports and Exercise, 30(2), 442-449.

Taylor, C.R. (1985). Force development during sustained locomotion: A determinant of gait, speed and metabolic power. Journal of Experimental Biology, 115, 253-262.

Tittle, K. and Wutscherk, H. (1988). Anatomical and anthropometric fundamentals of endurance. In: Endurance in Sport, the Encyclopedia of Sports Medicine, R.J. Shephard and P.O. Astrand (Eds.) Blackwell Scientific Publications, London, UK. pp. 35-45.

Travill, A.L., Carter, J.E.L. and Dolan, K.P. (1994). Anthropometric characteristics of elite male triathletes. In, F.I. Bell and G.H. Van Gyn (eds.) Proceedings for the 10th Commonwealth and International Scientific Congress: access to active living. University of Victoria, Victoria, 340-343.

Triathlon Australia, National Coaching Accreditation, Level One Manual. (1997). van der Woude, L.H.V., Veeger, H.E.J., Rozendal, R.H. and Sargeant, A.J. (1989). Optimal cycle frequency in hand rim wheelchair propulsion - wheelchair propulsion technique. European Journal of Applied Physiology, 58, 625-632.

Vercruyssen, F., Brisswalter, J., Hausswirth, C., Bernard, T., Bernard, O. and Vallier, J. (2002). Influence of cycling cadence on subsequent running performance in triathletes. Medicine and Science in Sports and Exercise, 34(3), .

Wilkie D.R. (1950). The relationship between force and velocity in human muscle. Journal of Physiology, 110, 249-280.

Williams, K.R. (1985). Biomechanics of running. Exercise and Sport Science, 13, 389- 441.

Williams, K.R., Snow, R. and Agruss, C. (1991). Changes in distance running kinematics with fatigue. Journal of Sports Biomechanics, 7, 138-162.

Wilmore, J. and Brown, C. (1974). Physiological profiles of women distance runners. Medicine and Science in Sports, 6(3), 178-181.

Witt, M. (1993). Coordination of the leg muscles during cycling and running in triathlon. XVIIth Congress of International Society of Biomechanics, Paris, 1470-1471.

Woolford, S., Bourdon, P., Craig, N. and Staner, T. (1993). Body composition and it's effects on athletic performance. Sports Coach October-December, 24-30.

185

Zinkgraf, S.A., Jones, C.J., Warren, B. and Krebs, P.S. (1986). An empirical investigation of triathlon performance. Journal of Sports Medicine and Physical Fitness, 26, 350-356.

Zhou, S., Robson, S.J., King, M.J. and Davie, A.J. (1997). Correlations between short course triathlon performance and physiological variables determined in laboratory cycle and treadmill test. Journal of Sports Medicine and Physical Fitness, 37, 122-130.

Appendix A

Publications

Appendix B

Abbreviations

a- O2diff Arteriovenous oxygen difference BI Brachial Index CI Crural Index EMG Electromyography iEMG Intergrated Electromyography GRF Ground Reaction Force HLa Blood Lactate IAT Individual Anaerobic Threshold ITU International Triathlon Union LL Lower Limb LLL Lower Limb Length LT Lactate Threshold OBLA Onset Blood Lactate Accumulation Cardiac Output RER Respiratory Exchange Ratio RLLL Relative Lower Limb Length rpm revolutions per minute ROM Range Of Motion RSH Relative Sitting Height SL Stride/Stroke Length SR Stride/Stroke Rate STA Seat Tube Angle SV Stroke Volume SVR Systemic Vascular Resistance TWC Triathlon World Championships

O2 Volume of Oxygen Consumed - oxygen cost

O2max Maximum Oxygen Uptake e Ventilation Rate

Appendix C

Summary of triathlete characteristics

Study n sex age height mass y sd cm sd kg sd

Deitrick (1991) 7 m 30.6 5.2 176.6 4.6 66.6 5.9 short course, recreational 7 m 29.6 4.4 185.5 5.1 90.6 3.2 heavy weight, recreational

Dengel et al. (1989) 11 m 31.4 1.8 179.9 1.6 74.5 2.3 1/2 IM, ability vary

DeVito et al. (1995) 6 m 27 5 176 8 69 9 well trained

Guezennec et al (1996) 10 m 29 3 67 2.6 competitive

Hue et al. (1998) 7 m 20.8 2.9 180.4 3 69.7 4.5 nat. team champs

Lehenaff et al. (1998) 8 m 20.8 2.1 179 9.1 elite sprint distance

Keskinen et al. (1997) 8 m 28.7 3.6 179 6 73.1 5.8 national level

Kohrt et al. (1987) 13 m 29.5 4.8 69.8 5.6 1/2 IM, ability vary

Kohrt et al. (1989) 8 m 29.4 5.1 71.1 2.5 1/2 IM, competitive

Laurenson et al. (1993) 10 f 27.1 3.5 167 6.8 56.4 6.1 national elite 9 f 28.2 4.1 168.8 8.1 59.5 5.3 national competitive

Leak and Carter (1991) 16 f 24.2 4.25 162.1 6.25 55.2 4.58 highly trained somatotype 3.1-4.3-2.6

O'Toole et al. (1987) 8 m 30.5 8.8 178.8 6.5 74.7 10 IM 6 f 31.3 5.6 168.5 4.7 60.3 4.6 IM

O'Toole et al. (1989) 14 m 40 11 180.2 8.2 73.7 8 IM 10 f 31 8 171 9 58 7 IM

Rowbottom et al. 8 m 29.6 2.6 180 2.4 73.9 2.1 national level (1997)

Schneider et al. (1990) 10 m 27.6 6.3 179.3 5.6 72 5.4 highly trained 4 m 25 1.4 178.8 5.3 70.2 7.3 elite

Sleivert et al. (1993) 18 m 27.2 1.3 180 1.5 76.2 2.1 recreational - competitive 7 f 28.3 2.3 166.4 2.1 59.3 2.1

Toussaint (1990) 5 m 34.2 7.9 184 2.1 78.4 4.3 endurance trained

Wells et al. (1987) 7 m 30.2 7.3 175 8.2 68.88 7.98 well trained 5 f 29.6 1.7 168.9 6 63.11 5.72 well trained

Zinkgraf et al. (1986) 65 m 31.7 179.8 74.4 various

APPENDIX D

Department of Human Movement & Exercise Science Parkway Entrance No. 3 Nedlands, Western Australia 6907 Telephone +61 8 9380 2361 Facsimile +61 8 9380 1039 http://www.general.uwa.edu.au/~hmweb/index.htm THE UNIVERSITY OF WESTERN AUSTRALIA

CONSENT FORM

Grant Landers is undertaking this project as an Honours research student in the Department of Human Movement and Exercise Science at The University of Western Australia, under the supervision of Drs Brian A. Blanksby and Tim R. Ackland.

The study aims to profile elite and junior, male and female, classic distance triathletes. It is hoped to create a set of standards of what the performances of the best triathletes in the world currently achieve and to examine the relationship with body shape, length and proportionality, and triathlon ability. These measurements could increase the information available to athletes, coaches and the general public, thereby improving the profile of the sport.

One testing session of about one hour will be required. The 28 measurements will be taken by qualified, competent technichians who will endevour to ensure your comfort is maintained. These include height, weight, sum of skinfolds, girths, breadths, and lengths of limbs and trunk (see attached sheet). Each participant will receive a report outlining the individual profile, as well as the mean data of all triathletes measured. We hope hope you will be able to find time to assist with this kinanthropometry survey. If you can help, please fill in the attached consent form and bring it with you to the laboratory for testing.

Should you have any concerns regarding this study, please contact the secretary of the Human Rights Committee at The University of Western Australia (Mrs K. Kirk, 9380 3703). Should you have any organisational queries, please contact one of the research team at the address or phone numbers below. All study participants will be provided with a copy of the information/consent form for their personal records.

Thank you in anticipation for your co-operation,

Yours Sincerely,

......

Mr Grant J. Landers Professor Brian A. Blanksby A. Professor Tim R. Ackland Human Movement Human Movement Human Movement UWA UWA UWA Nedlands 6907 WA Nedlands 6907 WA Nedlands 6907 WA Ph. 9380 1385 Ph. 9380 2658 Ph. 9380 2668 0417 944 885 I ...... hereby agree to participate in the study outlined above. (full name of participant)

I understand that all individual information will be kept confidential and, if any data are published, then anonymity will be maintained.

I am at liberty to withdraw from the study at any stage during the measuring procedures without prejudice. Also, I can request that certain measures are not to be taken or that a coach or manager must be present. I fully understand the nature of the testing procedures and my participation is voluntary.

The nature of the project has been fully explained to me.

…………………………… Signature Date (parent or guardian if under 18)

Appendix E Study 1: Data Collection Card

Triathlete Project

Subject Name Subject ID

Counrty DOB

date

Trial 1 Trial 2 Trial 3 Score

Triceps sf Subscapular sf Biceps sf Illiac crest sf Supra spinale sf Abdominal sf Front thigh sf Medial calf sf

Acromiale-radiale Radiale-stylion Midstylion-dactylion Trochanterion-tibiale laterale Tibiale laterale ht

Head girth Arm girth relaxed Arm girth flexed and tense Forearm girth max

Chest girth mesosternale Waist girth minimum Hip (gluteal) girth Thigh girth Calf girth maximum

Biacromial breadth Bbiiliocristale breadth Transverse chest breadth A-P chest depth Humerus breadth Femur breadth Foot Length

Body Mass Stature (stretch) Sitting height Arm span

APPENDIX F

Landmarks and Techniques

Landmarks Landmarks are essential for test, retest reliability. All landmarks are identified and located before any measurements were made. Inaccurate location of sites has been found to be the greatest source of error among investigators, with skinfold thickness shown to vary 2-3 mm when the calipers were placed 2.5 cm from the correct site (Ruzi, Colley & Hamilton, 1971).

All landmarks and hence all measurements were taken on the right side of the body in order to standardise the process.

Acromiale: The point at which the superior and lateral boarder of the acromion process, midway between posterior and anterior boarders of the deltoid muscle.

Radiale: The groove at the proximal and lateral boarder of the head of the radius.

Mid-acromiale-radiale: The point equidistant from the acromiale and radiale.

Stylion: The most distal point on the lateral margin of the styloid process of the radius.

Midstylion: The mid point, on the anterior surface of the wrist, of the horizontal line at the level of the stylion.

Dactylion: The most distal point of the third finger.

Subscapulare: The most inferior point of the inferior angle of the scapular.

Mesosternale: The mid point of the sternum at the level of the center of the articulation of the fourth rib with the sternum.

Iliocristale: The most laterale and superior aspect of the iliac crest.

Iliospinale: The most inferior tip of the anterior superior iliac spine.

Ilio-axilla line: The line joining the observed mid point of the armpit with the iliospinale.

Trochanterion: The most superior point of the greater trochanter of the femur.

Tibiale laterale: The most superior point on the lateral boarder of the head of the tibia.

Skinfolds The skinfold is picked up at the marked site. It is grasped so that a double fold of skin plus underlying subcutaneous tissue is held between the thumb and forefinger. Muscle tissue must not be grasped.

The nearest edge of the contact forces of the calipers, are applied 1 cm lateral or inferior to the thumb and the forefinger. The calipers should be placed at a depth of approximately mid-fingernail. The calipers are held at 90 degrees to the surface of the skinfold site and the hand grasping the skin remains throughout the measure.

The measurement is recorded 2 s after the full pressure of the calipers are applied. Sites are measure in succession to remove tester bias and allow tissue to return to resting state, ie. It reduces the effects of skinfold compressibility, or avoids depletion of fluids from the tissue.

Measures are recorded to the nearest 0.1 mm. If the first two measures of a site differ by more than 0.2 mm, a third is taken and the median recorded as the result.

In all instances Harpenden skinfold calipers were used.

Triceps: The subject should be standing in the anatomical position, with arms relaxed and externally rotated. This fold is taken at the most posterior point over the triceps muscle at the level of the mid-acromiale-radiale line. The fold is taken vertically and parallel to the upper arm.

Subscapular: The subject should be standing in the anatomical position. The skinfold is taken at a site 2 cm laterally and obliquely downwards from the subscapulare landmark. The fold is taken at an angle of approximately 45 degrees (following the natural fold line of the skin).

Biceps: Standing in the anatomical position, with the arm relaxed and slightly externally rotated. This fold is taken at the most anterior point over the biceps muscle, level with the mid-acromiale-radiale line. The fold is vertical; that is parallel to the upper arm.

Iliac crest: The subject places the right hand on the left shoulder. This skinfold is taken using the iliocristale landmark. The thumb is placed on the landmark and the fingers move sufficiently superior to grasp the fold. This fold may follow natural fold lines and, therefore run slightly downwards towards the medial aspect of the body.

Supraspinale: The subject stands with the right hand on the left shoulder. The site is the point of intersection of the ilio-axial line and a horizontal line projected from the iliocristale. This fold runs medially downwards at an angle of 45 degrees.

Abdominal: The fold is taken with the subject in a standing position. The fold is located 5 cm laterally and horizontally from the midpoint of the navel, and the fold is taken vertically.

Anterior thigh: The subject is seated with the knee flexed at right angles. The site is marked at the midpoint between the inguinal ligament and the superior boarder of the patella (while the subject is seated). The fold is then taken parallel to the long axis of the femur. If it is difficult to grasp a fold the subject should place their hands under the thigh and lift to relieve the tension of the skin. As a last resort, the scribe could use two hands to raise the skinfold, one hand on the landmark and the other about 6 cm distal.

Medial calf: The subject is standing and places the right foot on a box with the knee flexed at 90 degrees and the calf relaxed. The landmark is determined as the most medial aspect of the lower limb at the level of calf maximum girth. The fold is taken vertically.

Girths

Cross hand technique is used for measuring all girths. The tape is held at right angles to the limb or body segment being measured. Tension to the tape remains constant by ensuring no indentation of the skin.

Where possible, the tape is not removed from the subject but the arms are crossed to release the tension and the tape is slid to the next site. This speeds up the process of measurement and requires less invasion of the subjects’ personal space.

Measures are recorded to the nearest 0.1 cm.

At all times Lufkin steel metric anthropometric tapes were used for girth measurements.

Head: The subject can be either sitting or standing. The girth is measured in the Frankfort plane immediately above the glabella. It does not include the ears.

Arm relaxed: The upper arm is in a relaxed position by the subject’s side. The girth is measured at the level of the mid-acromiale-radiale.

Arm flexed and tensed: The subject raises the upper arm anteriorly until horizontal with the palm uppermost. The forearm is flexed at 70 degrees to the upper arm. The bicep is fully tensed and the maximum girth is measured.

Forearm: The subject stands in the anatomical position with the palm of the hand facing up. The maximum girth of the forearm is then measured.

Chest: While standing, the subject raises the arms to allow the tape to be passed around the body. The arms are then relaxed by the sides. The measurement is taken at the level of the mesosternale with the tape in a horizontal position. The value for the chest girth is recorded at the end of normal expiration (end tidal).

Waist: The subject stands with arms by the sides. The girth is taken at the narrowest point between the lower costal border and iliac crest. If no defined point exists, then the mid point between these two landmarks is used. The value is recorded at the end of a normal expiration.

Hip (gluteal): The subject stands with feet together and, for the ease of the investigator, is on a box. The girth is taken horizontally at the level of the greatest posterior protuberance of the buttocks.

Thigh: The subject stands on a box with feet slightly apart and equal distribution of mass on both feet. The girth is taken 1 cm below the level of the gluteal fold, perpendicular to the long axis of the thigh.

Calf: The subject stands on a box with equal distribution of mass on both feet. The girth is measured from the lateral aspect of the leg at the maximal diameter.

Lengths and Heights

A segmometer is used to measure body segments. The two end points of the segmometer are placed on the landmarks, and these are checked before a value is recorded. The end at which the measurement is read is located closest to the investigators eye level.

Measures are recorded to the nearest 0.1 cm.

Acromiale-radiale: The subject sands with the arms by the sides and palms facing the thighs. One end point of the segmometer is placed on the acromiale and the other on the radiale. The length of the upper arm is determined.

Radiale-stylion: The subject stands in the anatomical position with arms by the sides and palms facing the thighs. The distance between the radiale and stylion landmarks is measured. The segmometer should be positioned parallel to the radius, thus determining the length of the forearm.

Midstylion-dactylion: The subject sands with the arms by the sides and forearm flexed at 90 degrees with the hand in a supinated position and fingers are fully extended. The measurement is taken as the distance between the midstylion and dactylion.

Trochanterion height: The subject stands with feet together alongside a box of known height. One segmometer point is lined up with the trochanterion landmark and the other sits flush on the box. The distance is recorded and the height of the box add to obtain the distance from the ground to the trochanterion, or lower limb length.

Tibiale-laterale height: The subject stands on a box. The base of the segmometer is placed on top of the box and the other at the tibiale-laterale site. The distance in the vertical plane is measured to determine the shank length.

Breadths

When measuring breadths it is important to use enough force to compress the underlying soft tissue, so as to measure the bone breadths accurately.

Measures are recorded to the nearest 0.1 cm.

The anthropometer or a small bone caliper is used to measure the following breadths.

Biacromiale: The subject stands with arms by their sides and facing away from the investigator. The anthropometer blades are placed on the most lateral points of the acromion processes and the distance between these is recorded.

Biiliocristale: The subject stands on the ground, or on a box, facing the investigator. The anthropometer blades are placed on the most lateral aspects of the iliac crest (the iliocristale landmark). The blades are angled at 45 degrees upwards.

Foot length: The subject stands on a raised box. The distance between the longest toe (which may be the first or second toe) and the most posterior point on the heel is measured. The anthropometer should be kept parallel to the long axis of the foot.

Transverse chest: The subject either sits or stands facing the investigator. The distance is measured between the most lateral aspects of the chest at the level of the mesosternale landmark. The blades of the anthropometer are positioned on the ribs and at angle of approximately 30 degrees down, to the horizontal. The measurement is taken at end normal expiration (end tidal).

Chest depth: The subject remains seated but erect. Wide spreding calipers are applied to the chest at the level of the mesosternale. That is, on the anterior landmark and kept horizontal to be located posteriorly on the spinous process of the corresponding vertebra. The value is recorded at end tidal.

Biepicondylar humerus: With the subject seated or standing in the anatomical position, the upper arm is flexed to horizontal and the forearm flexed at 90 degrees. Small bone calipers are used to measure the distance between the medial and lateral epicondyles of the humerus. The arms of the calipers are directed upward at an angle of 45 degrees and placed directly on the epicondyles. The calipers may not remain in a horizontal position due to the general trend of the medial epicondyle to be lower than the lateral epicondyle.

Biepicondylar femur: The subject is seated with the leg flexed (knee at 90 degrees). Small bone calipers are used to measure the distance between the medial and lateral epicondyles of the femur. The arms of the calipers are directed downward and at an angle of 45 degrees.

Others

Measures are recorded to the nearest 0.1 cm for heights and lengths and to the nearest 0.05 kg for mass.

Height: Stretch stature is used in order to account for any diurnal variation. Height is recorded as the distance between the floor and the vertex of the head when the head is in the Frankfort plane. The head is lifted from under the mastoid process and the subject takes a deep breath in. Heels are kept together and directly below the head. A wall- mounted stadiometer is used in the lab and a mobile stadiometer for other locations.

Mass: Mass is recorded with the subject wearing swimwear only. The subject stands on the center of the scales, without support, and evenly distributes mass on both feet.

An electronic balance, accurate to 50 g was used in the lab and a portable beam balance, accurate to 50 g for off site locations.

Sitting height: The subject sits on a box of known height that rests against a stadiometer. The distance between the box and vertex of the head is measured, again with the head in the Frankfort plane. Slight upward pressure is applied to the mastoid process by the investigator to reduce any diurnal variation. The total height is recorded and the height of the box is then subtracted to obtain sitting height.

Arm span: The distance between the dactylia of the left and right hands when standing against and facing a wall. The outstretched arms should be horizontal, with one hand in a corner of a room as the zero point and the other stretched out to a perpendicular rule or in the mobile laboratories a measurement graph.

Appendix G

Formulae

Somatotype

Endomorphy = -0.7182 + 0.1451 (∑SF) +0.00068 (∑SF)2 + 0.0000014 (∑SF)3

Mesomorphy = 0.858 (humerus breadth (cm)) + 0.601 (femur breadth (cm)) + 0.188 (corrected arm girth (cm)) + 0.161 (corrected calf girth (cm)) - 0.131 (stretch stature (cm)) + 4.5

Ectomorphy = 0.732 (HWR) - 28.58 if: HWR > 40.75 = 0.463 (HWR) - 17.63 if: 40.75 > HWR > 38.25

Where:

∑SF = (tricep skinfold + subscapular skinfold + supraspinale skinfold) x 170.18 height Height is measure in centimeters ∑SF is reported in millimeters corrected arm girth (cm) = arm girth (cm) - triceps skinfold (cm) corrected calf girth (cm) = calf girth (cm) - medial calf skinfold (cm)

HWR = height height is measure in centimeters and weight in kg 3√weight

Phantom Scores

Z = 1 [V (170.18/H)d - P] S

Where:

Z = proportionality score S = phantom standard deviation for variable (V) V = obtained value for variable (V) 170.18 = phantom stature constant H = obtained statured d = dimensional exponent d = 1; lengths, breadths, girths, skinfolds d = 2; areas d = 3; volumes and masses P = phantom size for variable (V)

Fiveway Fractionation

Note below: M? = mass of the unknown S? = sum of the components that determine M? Z? = the z score related to the phantom model

Prediction of skin mass

Ms = SA x Tsk x 1.05

SA = Csa x weight0.425 x height0.725

Where: Ms = skin mass (kg) SA = surface area (m2) 1.05 = density of skin Tsk = thickness of skin; males = 2.07 mm, females = 1.96 mm Csa = coeffecient of surface area; males = 68.308, females = 73.074

Prediction of adipose tissue

Sa = ∑(tricep skinfold + subscapular skinfold + supraspinale skinfold + abdominal skinfold + thigh skinfold + medial calf skinfold) mm

Za = (Sa x 170.18/height) - 116.41 34.79

Adipose mass (Ma) = (Za x 5.85) + 25.6 (170.18/height)3

Where: 116.41 = phantom ∑ skinfolds (mm) 34.79 = phantom ∑ standard deviations (mm) 25.6 = phantom adipose tissue mass (kg) 5.85 = phantom standard deviation for adipose tissue mass (kg)

Prediction of Muscle mass

Sm = ∑ (corrected arm girth + forearm girth + corrected thigh girth + corrected calf girth + corrected chest girth)

Zm = (Sm x 170.18/height) - 207.21 13.74

Mn = (Zm x 5.4) + 24.5 (170.18/height)3

Where: Corrected arm girth (cm) = relaxed arm girth (cm) - tricep skinfold (cm) Corrected thigh girth (cm) = thigh girth (cm) - front thigh skinfold (cm) Corrected calf girth (cm) = calf girth (cm) - medial calf skinfold (cm) Corrected chest girth (cm) = chest girth (cm) - subscapular skinfold(cm)

207.21 = phantom sum of corrected girths (cm) 13.74 = phantom sum of standard deviations for girths (cm) 24.5 = phantom muscle mass (kg) 5.4 = phantom standard deviation for muscle mass (kg)

Predicted bone mass

• Bone mass of the head

Zhb = (head girth (cm) - 56.0)/1.44

Mhb = (Zhb x 0.18) + 1.20

Where: 56.0 = phantom head girth (cm) 1.44 = phantom standard deviation for head girth (cm) 1.20 = phantom head bone mass (kg) 0.18 = phantom standard deviation of head bone mass (kg)

• Bone mass of the remainder of the body

Sbb = ∑ [(biacromial breadth + biiliocristale breadth) + (2 x biepicondylar femur breadth) + (2 x biepicondylar humerus breadth)] (cm)

Zbb = (Sbb x 170.18/height) - 98.88 5.33

Mbb = (Zbb x 1.34) + 6.70 (170.18/height)3

Where: 98.88 = phantom sum of bone breadths (cm) 5.33 = phantom sum of standard deviations for bone breadths (cm) 6.70 = phantom sum of bone masses (kg) 1.34 = phantom standard deviations for bone masses (kg)

• Total bone mass = Mhb + Mbb (kg)

Prediction of residual mass

Sr = ∑ (chest depth + transverse chest breadth + corrected waist girth) (cm)

Zr = (Sr x 89.2/sitting height) - 109.35 7.08

Mr = (Zr x 1.24) + 6.10 (89.92/sitting height)3

Where: 89.92 = phantom sitting height (cm) 7.08 = phantom standard deviation for sitting height (cm) 6.10 = phantom residual mass (kg) 1.24 = phantom standard deviation for residual mass (kg)

Prediction of total body mass

Body mass = (skin + adipose tissue + bone + muscle + residual) (kg)

Indicies

Relative Lower Limb Length (RLLL)

RLLL = Trochanterion height x 100 Stretch stature

Brachial Index (BI)

BI = radiale-stylion x 100 acromiale-radiale

Crural Index (CI)

CI = tibiale-laterale height x 100 (trochanterion height - tibiale-laterale height)

Appendix H

Study 2: Raw Data

Following is the raw data from study two of the 1999 ITU World Cup events. It includes split times with as much detail as could be attained from the ITU web site. This data was entered into mathmatica for subsequent calculations.

Female Triathlons

Lausanne, Switzerland Note: name, swim, trans 1, bike, trans 2, run

{lausanne switzerland {loretta harrop, 0:19:36, 0:01:29, 1:12:07, 0:00:53, 0:36:13}, {tracy hargreaves, 0:21:32, 0:01:28, 1:11:58, 0:00:52, 0:34:57}, {joelle franzman, 0:20:11, 0:01:28, 1:11:31, 0:00:45, 0:36:59}, {magali messmer, 0:20:06, 0:01:31, 1:11:34, 0:00:53, 0:37:22}, {julie ricketts, 0:20:12, 0:01:30, 1:11:29, 0:00:52, 0:37:37}, {susan bartholomew, 0:20:17, 0:01:33, 1:11:22, 0:00:50, 0:37:40}, {michelle dillon, 0:22:15, 0:01:38, 1:11:52, 0:00:45, 0:35:18}, {kathleen smet, 0:20:14, 0:01:38, 1:11:21, 0:00:53, 0:37:45}, {ines estedt, 0:21:20, 0:01:35, 1:11:54, 0:00:58, 0:37:09}, {jasmine haemmerle, 0:22:27, 0:01:30, 1:11:47, 0:00:52, 0:36:42}, {stephanie forrester, 0:21:22, 0:01:34, 1:12:49, 0:00:50, 0:36:58}, {maribel blanco, 0:22:24, 0:01:29, 1:11:50, 0:00:51, 0:37:10}, {anja dittmer, 0:20:57, 0:01:27, 1:12:28, 0:00:59, 0:37:57}, {isabelle mouthon, 0:20:40, 0:01:32, 1:12:36, 0:00:56, 0:38:14}, {marie overbye, 0:21:27, 0:01:33, 1:11:50, 0:00:55, 0:38:16}, {bridgitte mcmahon, 0:20:56, 0:01:41, 1:12:18, 0:00:59, 0:38:29}, {anja heil, 0:20:45, 0:01:33, 1:12:30, 0:00:55, 0:38:44}, {lucienn groenendijk, 0:21:30, 0:01:31, 1:12:43, 0:00:53, 0:37:58}, {ute mueckel, 0:20:10, 0:01:33, 1:13:05, 0:00:52, 0:39:05}, {kim carter, 0:20:17, 0:01:35, 1:13:03, 0:00:56, 0:39:12}, {jill newman, 0:21:25, 0:01:40, 1:11:42, 0:00:55, 0:39:28}, {edith cigana, 0:00:00, 0:00:00, 0:00:00, 0:00:00, 2:15:17}, {beth thompson, 0:20:37, 0:01:38, 1:12:36, 0:00:57, 0:39:41}, {manuela ianesi, 0:20:57, 0:01:44, 1:12:14, 0:00:56, 0:39:54}, {nancy kemp- arendt, 0:20:22, 0:01:38, 1:12:53, 0:00:53, 0:40:16}, {ingrid van lubek, 0:20:54, 0:01:33, 1:12:22, 0:00:58, 0:40:42}, {becky gibbs, 0:20:10, 0:01:34, 1:13:09, 0:00:56, 0:40:51}, {leanda cave, 0:21:35, 0:01:28, 1:12:44, 0:00:55, 0:40:13}, {sophie delemer, 0:22:24, 0:01:33, 1:11:35, 0:00:52, 0:40:51}, {loretta sollars, 0:22:24, 0:01:32, 1:14:56, 0:00:52, 0:38:25}, {renato berkova, 0:20:26, 0:01:36, 1:18:39, 0:00:52, 0:37:49}, {kristen armstrong, 0:22:21, 0:01:42, 1:11:44, 0:00:57, 0:43:10}, {silvia gemignani, 0:20:24, 0:01:51, 1:18:08, 0:01:00, 0:38:50}, {francisca russli, 0:22:25, 0:01:36, 1:11:31, 0:00:55, 0:44:36}, {pamela birsinger, 0:22:22, 0:01:36, 1:16:17, 0:00:56, 0:40:03}, {shelia taormina, 0:19:35, 0:01:55, 1:18:47, 0:01:04, 0:40:49}, {karen dehmel, 0:20:44, 0:01:39, 1:18:11, 0:01:01, 0:41:14}, {lena wahlqvist, 0:21:37, 0:01:35, 1:17:20, 0:01:09, 0:41:34}, {gina kehr, 0:20:47, 0:01:41, 1:20:55, 0:01:14, 0:40:29}, {sian brice, 0:20:45, 0:01:29, 1:12:35, 0:00:56, dnf}, {lucie zelenkova, 0:20:15, 0:01:38, 1:18:31, 0:01:00, dnf}, {nina anisimova, 0:20:25, 0:01:36, dnf, dnf, dnf}, {christiane pilz, 0:20:57, 0:01:28, dnf, dnf, dnf}, {louise soper, 0:20:50, 0:01:38, dnf, dnf, dnf}, {silvia pepels, 0:20:59, 0:01:35, dnf, dnf, dnf}, {katja schumacher, 0:21:28, 0:01:30, dnf, dnf, dnf}, {shanelle barrett, 0:21:36, 0:01:33, dnf, dnf, dnf}, {helen salmon, 0:22:21, 0:01:36, dnf, dnf, dnf}, {aniko gog, 0:22:25, 0:01:34, dnf, dnf, dnf}}

Noosa, Australia note: swim, cycle and run splits only

{noosa australia {michelle dillon, 0:20:49, 0:59:49, 0:34:25}, {emma carney, 0:20:48, 0:59:47, 0:35:14}, {stephanie forrester, 0:20:12, 1:00:17, 0:36:03}, {siri lindley, 0:20:12, 1:00:18, 0:36:10}, {haruna hosoya, 0:20:22, 0:59:57, 0:36:33}, {machiko nakanishi, 0:19:40, 1:00:46, 0:37:13}, {jill newman, 0:20:46, 0:59:47, 0:37:43}, {melissa ashton, 0:20:01, 1:00:33, 0:37:52}, {evelyn williamson, 0:20:06, 1:00:19, 0:38:16}, {rina hill, 0:19:23, 1:01:04, 0:38:24}, {karen dehmel, 0:20:03, 1:00:37, 0:38:35}, {leanda cave, 0:19:20, 1:01:12, 0:39:14}, {ute muckel, 0:19:31, 1:01:04, 0:39:38}, {jenny mann, 0:20:15, 1:00:19, 0:39:56}, {yukie koumegawa, 0:20:43, 1:03:01, 0:36:50}, {becky gibbs, 0:19:25, 1:01:14, 0:39:57}, {amanda pagon, 0:20:14, 1:00:26, 0:40:30}, {josie laone, 0:20:13, 1:00:10, 0:40:51}, {tamoko hiranaka, 0:19:23, 1:01:08, 0:40:52}, {marci steelman, 0:20:16, 1:00:25, 0:41:41}, {gina derksgardner, 0:19:33, 1:00:55, 0:42:28}, {jane kargotich, 0:20:08, 1:00:26, 0:42:53}, {kiyomi niwata, 0:20:10, 1:01:28, 0:41:57}, {isabelle turcottebaird, 0:20:26, 1:03:23, 0:40:05}, {carla moreno, 0:20:20, 1:06:46, 0:41:41}, {tracy hargreaves, 0:21:08, dnf, dnf}}

Corner Brook, Canada note: swim, trans 1, cycle, trans 2, run note swim = 3 laps, cycle = 4 laps, run = 4 laps

{corner brook canada {jackie gallagher, 0:20:39.0, 0:02:14.81:10:51.1, 0:00:32.9, 0:35:28.1}, {michellie jones, 0:19:48.1, 0:02:06.4, 1:11:46.1, 0:00:33.0, 0:36:37.0}, {mariana ohata, 0:20:34.2, 0:02:14.7, 1:10:57.2, 0:00:35.5, 0:36:31.9}, {ingrid van lubek, 0:20:39.4, 0:02:25.4, 1:10:40.4, 0:00:37.3, 0:36:44.7}, {jennifer gutierrez, 0:19:26.3, 0:02:22.0, 1:11:55.7, 0:00:34.1, 0:37:12.9}, {susan bartholomew, 0:19:30.7, 0:02:11.6, 1:11:57.6, 0:00:31.9, 0:37:36.6}, {marie overbye, 0:20:40.9, 0:02:14.9, 1:10:48.9, 0:00:39.6, 0:38:04.1}, {sharon donnelly, 0:20:02.5, 0:02:14.4, 1:11:29.5, 0:00:34.7, 0:38:20.0}, {jill newman, 0:20:39.6, 0:01:23.8, 1:11:39.0, 0:00:32.3, 0:38:54.8}, {rina hill, 0:19:19.6, 0:02:26.3, 1:12:15.0, 0:00:36.0, 0:38:56.4}, {christine hocq, 0:19:53.2, 0:02:22.5, 1:11:38.0, 0:00:32.4, 0:39:27.0}, {maribel blanco, 0:22:24.3, 0:02:15.7, 1:11:36.6, 0:00:34.5, 0:37:19.5}, {marci steelman, 0:20:09.1, 0:02:21.7, 1:11:21.5, 0:00:40.5, 0:39:47.4}, {manuela ianesi, 0:20:34.3, 0:02:29.2, 1:10:51.3, 0:00:34.5, 0:40:13.5}, {gail laurence, 0:21:05.1, 0:02:40.0, 1:09:58.2, 0:00:41.5, 0:40:51.1}, {lucienne groenendijk, 0:21:23.6, 0:02:21.7, 1:14:14.4, 0:00:38.1, 0:38:06.0}, {laura reback, 0:19:28.8, 0:02:15.2, 1:16:17.0, 0:00:36.6, 0:38:42.7}, {beth thomson, 0:21:42.0, 0:02:17.0, 1:14:04.1, 0:00:37.9, 0:39:01.9}, {amanda pagon, 0:20:35.8, 0:02:23.9, 1:15:03.9, 0:00:36.1, 0:39:55.2}, {ute mukel, 0:19:17.4, 0:02:28.8, 1:14:44.6, 0:00:38.7, 0:42:05.5}, {shanelle barrett, 0:21:30.3, 0:02:31.8, 1:13:59.7, 0:00:34.4, 0:41:26.5}, {loretta sollars, 0:22:19.8, 0:02:23.6, 1:17:17.00:00:34.7, 0:37:35.2}, {edith cigana, 0:20:36.2, 0:02:29.6, 1:14:56.5, 0:00:37.0, 0:41:45.3}, {isabelle gagnon, 0:22:28.5, 0:02:37.4, 1:16:55.7, 0:00:37.1, 0:40:39.3}, {natasha yaremczuk, 0:22:20.9, 0:02:30.0, 1:17:08.9, 0:00:32.4, 0:41:47.1}, {bernadita gras-thomas, 0:22:26.2, 0:02:37.9, 1:19:36.0, 0:00:40.1, 0:41:28.3}, {annick dufour, 0:22:39.3, 0:02:21.0, 1:23:15.5, 0:00:38.0, 0:41:30.2}, {katie web, 0:21:32.6, 0:02:29.8, 1:13:59.1, 0:00:35.8, dnf}, {carol montgomery, 0:19:57.8, 0:02:22.9, dnf, dnf, dnf}, {silvia pepels, 0:20:00.6, 0:02:22.5, dnf, dnf, dnf}, {gina derksgardner, 0:20:05.6, 0:02:37.5, dnf, dnf, dnf}}

Tiszaujvaros, Hungary note: swim, trans 1, cycle, trans 2, run note: swim 2 laps, cycle 8 laps, run 4 laps

{tiszaujvaros hungary {loretta harrop, 0:18:08.5, 0:00:39.7, 0:59:59.7, 0:00:42.0, 0:37:11.4}, {tracy hargreaves, 0:19:56.2, 0:00:41.8, 1:00:53.9, 0:00:38.1, 0:35:08.5}, {rina hill, 0:19:13.3, 0:00:43.7, 1:01:40.1, 0:00:38.2, 0:35:55.9}, {meike suys, 0:19:55.9, 0:00:37.1, 1:00:57.6, 0:00:38.7, 0:36:12.5}, {sian brice, 0:19:44.4, 0:00:38.4, 1:01:10.2, 0:00:39.6, 0:36:18.8}, {haruna hosoya, 0:19:31.4, 0:00:43.5, 1:01:15.9, 0:00:37.2, 0:36:26.8}, {evelyn williamson, 0:19:21.5, 0:00:40.7, 1:01:29.5, 0:00:38.9, 0:36:33.6}, {siri lindley, 0:19:24.7, 0:00:41.2, 1:01:28.5, 0:00:37.6, 0:36:53.6}, {julie ricketts, 0:19:30.8, 0:00:40.8, 1:01:22.9, 0:00:43.5, 0:37:18.2}, {michelle dillon, 0:20:42.6, 0:00:40.0, 1:02:21.6, 0:00:36.6, 0:35:38.6}, {carol montgomery, 0:20:00.3, 0:00:42.0, 1:03:10.5, 0:00:40.4, 0:35:34.7}, {anja heil, 0:19:23.0, 0:00:44.5, 1:01:26.0, 0:00:44.7, 0:37:51.4}, {jasmine haemmerle, 0:20:39.4, 0:00:40.0, 1:02:25.2, 0:00:38.5, 0:35:49.8}, {stephanie forrester, 0:20:34.2, 0:00:41.5, 1:02:27.1, 0:00:34.1, 0:35:57.8}, {isabelle mouthon, 0:19:20.2, 0:00:40.4, 1:01:31.8, 0:00:37.2, 0:38:07.8}, {lena wahlqvist, 0:19:54.8, 0:00:42.0, 1:01:02.0, 0:00:47.1, 0:38:10.0}, {nina anisimova, 0:19:15.7, 0:00:42.3, 1:01:38.3, 0:00:39.6, 0:38:38.6}, {ingrid van lubek, 0:19:28.7, 0:00:40.7, 1:01:26.5, 0:00:41.5, 0:38:53.4}, {lucie zelenkova, 0:19:15.2, 0:00:44.9, 1:01:36.5, 0:00:44.1, 0:38:53.1}, {beatrice mouthon, 0:19:25.3, 0:00:38.9, 1:01:29.2, 0:00:41.5, 0:39:01.5}, {renato berkova, 0:19:26.4, 0:00:42.4, 1:03:44.1, 0:00:35.0, 0:37:20.3}, {aniko gog, 0:20:07.1, 0:00:45.1, 1:02:47.5, 0:00:42.0, 0:37:33.4}, {lucienne groenendijk, 0:20:39.3, 0:00:42.5, 1:02:23.6, 0:00:41.5, 0:37:45.9}, {ute mueckel, 0:19:04.2, 0:00:41.4, 1:01:48.4, 0:00:41.2, 0:40:33.4}, {nora edocseny, 0:20:37.1, 0:00:45.0, 1:02:27.7, 0:00:39.2, 0:38:31.0}, {oksana zhuravkov, 0:20:16.6, 0:00:42.1, 1:02:47.4, 0:00:42.2, 0:38:40.4}, {sandra soldan, 0:19:31.6, 0:00:41.0, 1:01:24.8, 0:00:46.5, 0:41:08.9}, {brigitte mcmahon, 0:19:53.5, 0:00:45.1, 1:03:19.8, 0:00:39.4, 0:38:59.3}, {louise soper, 0:19:18.7, 0:00:40.8, 1:01:36.4, 0:00:38.9, 0:41:34.7}, {christiane pilz, 0:19:57.9, 0:00:38.2, 1:03:08.2, 0:00:39.6, 0:39:45.3}, {simone buerli, 0:19:29.9, 0:00:47.2, 1:03:34.8, 0:00:47.4, 0:40:09.1}, {leanda cave, 0:19:11.5, 0:00:44.0, 1:01:37.2, 0:00:36.0, 0:43:18.0}, {virginia berastegui, 0:20:35.4, 0:00:44.4, 1:02:28.4, 0:00:44.7, 0:41:02.2}, {adel molnar, 0:19:44.1, 0:00:43.2, 1:03:22.0, 0:00:40.0, 0:41:20.9}, {maria alikina, 0:20:08.8, 0:00:42.7, 1:02:55.9, 0:00:39.4, 0:41:43.5}, {machiko nakanishi, 0:19:41.9, 0:00:41.3, 1:03:23.3, 0:00:39.7, 0:42:04.7}, {olga danka, 0:19:43.8, 0:00:40.3, 1:03:27.7, 0:00:45.6, 0:42:17.1}, {andrea walko, 0:19:57.3, 0:00:46.4, 1:03:04.9, 0:00:45.8, 0:44:34.8}, { olga generalova, 0:20:14.8, 0:00:42.2, 1:02:54.1, 0:00:44.1, 0:45:46.3}, {larisa sviridendo, 0:19:59.2, 0:00:48.6, 1:03:06.1, 0:00:51.5, dnf}, {petra harangi, 0:19:16.9, 0:00:46.6, dnf, dnf, dnf}, {schuhmacher, 0:20:35.1, 0:00:43.5, dnf, dnf, dnf}, {joanne king, 0:20:44.4, 0:00:46.4, dnf, dnf, dnf}, {ewa dederko, 0:21:04.8, 0:00:43.0, dnf, dnf, dnf}, {michaela novakova, 0:21:12.6, 0:00:48.4, dnf, dnf, dnf}, {olga zavsaylova, 0:21:35.4, 0:00:43.9, dnf, dnf, dnf}}

Ishigaki, Japan note: swim, cycle and run splits

{ishigaki japan {loretta harrop, 0:17:05, 1:01:46, 0:40:54}, {nicole hackett, 0:17:06, 1:01:47, 0:41:15}, {barbara lindquist, 0:17:04, 1:01:50, 0:41:23}, {emma carney, 0:18:32, 1:03:14, 0:39:02}, {jackie gallagher, 0:19:26, 1:03:09, 0:38:30}, {carol montgomery, 0:18:13, 1:03:38, 0:39:37}, {michellie jones, 0:18:10, 1:03:32, 0:40:01}, {sian brice, 0:18:17, 1:03:33, 0:40:05}, {isabell mouthon, 0:18:14, 1:03:31, 0:40:38}, {siri lindley, 0:18:17, 1:03:32, 0:40:40}, {weike hoogzaad, 0:18:12, 1:03:40, 0:41:09}, {magali messmer, 0:17:33, 1:04:12, 0:41:22}, {ines estedt, 0:19:15, 1:03:24, 0:40:36}, {maruna hosoya, 0:18:13, 1:03:31, 0:41:37}, {sharon donnelly, 0:17:08, 1:05:28, 0:40:54}, {kiyomi niwata, 0:19:06, 1:03:28, 0:40:56}, {jennifer gutierrez, 0:17:36, 1:04:15, 0:41:45}, {marie overbye, 0:19:10, 1:03:29, 0:41:33}, {sian welch, 0:19:17, 1:03:20, 0:41:44}, {jasmine haemmerle, 0:19:18, 1:03:20, 0:41:48}, {megumi shigaki, 0:19:20, 1:03:15, 0:41:52}, {anja dittmer, 0:19:05, 1:03:30, 0:41:55}, {kathleen smet, 0:18:09, 1:04:33, 0:41:51}, {silvia pepels, 0:18:17, 1:03:30, 0:42:47}, {gina derks, 0:18:05, 1:03:43, 0:42:48}, {katja schumacher, 0:18:55, 1:03:42, 0:42:05}, {tracy hargreaves, 0:19:39, 1:04:49, 0:40:20}, {laura reback, 0:17:18, 1:04:33, 0:43:01}, {ingrid van lubek, 0:19:18, 1:05:41, 0:40:29}, {kim carter, 0:18:20, 1:03:29, 0:43:44}, {sibylle matter, 0:18:32, 1:03:20, 0:43:45}, {jill newman, 0:19:02, 1:03:35, 0:43:08}, {brigitte mcmahon, 0:18:29, 1:04:15, 0:43:03}, {yukie koumegawa, 0:18:18, 1:03:32, 0:44:09}, {michelle dillon, 0:20:39, 1:05:34, 0:39:48}, {evelyn williamson, 0:18:40, 1:03:59, 0:43:27}, {nancy kemp-arendt, 0:17:43, 1:04:56, 0:43:29}, {akiko hirao, 0:19:13, 1:05:36, 0:41:36}, {suzamme nielsen, 0:19:24, 1:06:53, 0:40:17}, {clare carney, 0:20:40, 1:05:35, 0:40:32}, {renato berkova, 0:18:22, 1:06:39, 0:41:48}, {nora edosceny, 0:19:14, 1:05:47, 0:42:04}, {natascha badmann, 0:20:32, 1:03:50, 0:42:48}, {rina hill, 0:17:30, 1:05:08, 0:44:48}, {maribel blanco, 0:19:40, 1:05:14, 0:42:48}, {lizel moore, 0:19:34, 1:04:56, 0:43:28}, {annie emmerson, 0:20:34, 1:05:45, 0:42:41}, {lucienne groenendijk, 0:19:10, 1:07:07, 0:42:59}, {beatrice mouthon, 0:19:05, 1:07:10, 0:43:16}, {dominique donner, 0:18:19, 1:03:35, 0:48:31}, {nina anisimova, 0:18:35, 1:06:26, 0:45:46}, {merav tarshish, 0:19:27, 1:07:05, 0:44:24}, {iona wynter, 0:19:21, 1:05:36, 0:46:11}, {loretta sollars, 0:18:58, 1:08:46, 0:43:44}, {nicole mertes, 0:19:36, 1:10:13, 0:41:50}, {isabelle baird, 0:19:08, 1:05:51, 0:46:45}, {mika sakane, 0:18:52, 1:07:20, 0:45:50}, {katie webb, 0:19:50, 1:06:33, 0:46:24}, {carmen ochoa, 0:18:59, 1:06:03, 0:48:05}, {helen salomon, 0:20:28, 1:09:13, 0:44:55}, {jane kargotich, 0:19:10, 1:09:11, 0:46:27}, {machiko nakanishi, 0:18:03, 1:11:44, 0:45:11}, {aline garza, 0:20:32, 1:09:11, 0:45:24}, {maria luisa martinez, 0:19:37, 1:10:14, 0:45:20}, {ewa dederko, 0:19:36, 1:10:25, 0:47:37}, {lisset olivera, 0:19:03, 1:10:46, 0:49:53}, {kristie otto, 0:19:09, 1:05:51, 0:56:23}, {tomoko hiranaka, 0:17:30, 1:13:54, 0:51:09}, {joanne king, 0:18:45, dnf, dnf}, {mieke suys, 0:18:42, 1:03:55, dnf}, {oksana zhuravkov, 0:19:08, dnf, dnf}}

Gamagori, Japan note: swim, cycle and run splits

{gamagori japan {loretta harrop, 0:18:00, 1:02:08, 0:34:18}, {nicole hackett, 0:17:28, 1:02:40, 0:35:13}, {michellie jones, 0:18:36, 1:03:02, 0:33:47}, {siri lindley, 0:18:46, 1:02:56, 0:33:44}, {barbara lindquist, 0:17:26, 1:02:59, 0:35:24}, {jackie gallagher, 0:19:25, 1:03:35, 0:33:11}, {isabell mouthon, 0:18:50, 1:02:44, 0:34:40}, {magali messmer, 0:18:07, 1:03:29, 0:34:41}, {haruna hosoya, 0:18:45, 1:02:49, 0:34:45}, {anja dittmer, 0:18:53, 1:02:46, 0:34:50}, {weike hoogzaad, 0:18:27, 1:03:16, 0:34:50}, {mieke suys, 0:18:52, 1:02:53, 0:34:49}, {evelyn williamson, 0:18:53, 1:02:38, 0:35:16}, {sian brice, 0:18:08, 1:03:51, 0:34:54}, {tracy hargreaves, 0:19:46, 1:03:10, 0:34:13}, {nancy kemp-arendt, 0:18:19, 1:03:20, 0:35:43}, {kathleen smet, 0:18:12, 1:03:36, 0:35:41}, {kiyomi niwata, 0:18:54, 1:02:44, 0:35:52}, {nina anisimova, 0:18:09, 1:03:41, 0:35:51}, {carol montgomery, 0:18:28, 1:04:43, 0:34:43}, {silvia pepels, 0:18:24, 1:03:11, 0:36:27}, {akiko hirao, 0:19:37, 1:04:08, 0:34:25}, {sharon donelly, 0:17:59, 1:04:14, 0:36:00}, {ingrid van lubek, 0:19:36, 1:04:05, 0:34:43}, {christine hocq, 0:18:04, 1:03:37, 0:37:02}, {marie overbye, 0:19:31, 1:03:32, 0:35:52}, {lucienne groenendijk, 0:19:33, 1:03:44, 0:35:49}, {sibylle matter, 0:18:51, 1:02:48, 0:37:29}, {joanne king, 0:19:45, 1:03:58, 0:35:27}, {jane kargotich, 0:18:54, 1:04:05, 0:36:17}, {oksana zhuravkov, 0:19:02, 1:04:02, 0:36:27}, {lizel moore, 0:19:52, 1:03:56, 0:35:53}, {yukie koumegawa, 0:18:33, 1:03:11, 0:38:09}, {natascha badmann, 0:19:50, 1:03:49, 0:36:23}, {gina derks, 0:18:29, 1:04:41, 0:37:11}, {nora edosceny, 0:19:32, 1:04:42, 0:36:24}, {katja schumacher, 0:19:51, 1:04:05, 0:36:55}, {rina hill, 0:18:08, 1:08:05, 0:34:42}, {renato berkova, 0:18:30, 1:07:50, 0:34:38}, {machiko nakanishi, 0:18:23, 1:04:50, 0:37:46}, {michelle dillon, 0:20:49, 1:07:25, 0:33:09}, {kim carter, 0:19:00, 1:04:12, 0:38:14}, {maribel blanco, 0:20:28, 1:05:39, 0:35:25}, {jasmine haemmerle, 0:19:56, 1:03:50, 0:37:56}, {brigitte mcmahon, 0:18:49, 1:04:23, 0:38:36}, {nicole mertes, 0:20:13, 1:06:26, 0:35:17}, {sophie delemer, 0:19:52, 1:03:58, 0:38:16}, {suzanne nielsen, 0:19:29, 1:07:00, 0:35:43}, {iona wynter, 0:19:27, 1:03:52, 0:38:59}, {dominique donner, 0:18:24, 1:04:51, 0:39:40}, {lorretta sollars, 0:19:48, 1:06:29, 0:36:41}, {carmen ochoa, 0:18:56, 1:04:18, 0:40:07}, {lin xing, 0:20:01, 1:06:45, 0:37:06}, {karen gorden, 0:19:43, 1:06:31, 0:37:44}, {isabelle baird, 0:19:49, 1:06:45, 0:37:51}, {ewa dederko, 0:19:53, 1:06:32, 0:38:31}, {merav tarshish, 0:20:21, 1:07:50, 0:36:49}, {maria luisa martinez, 0:19:54, 1:06:26, 0:38:49}, {laura reback, 0:18:03, 1:08:17, 0:39:38}, {miyuki biwata, 0:20:24, 1:07:36, 0:38:10}, {lisset olivera, 0:19:49, 1:06:38, 0:40:16}, {jennifer gutierrez, 0:18:23, 1:03:48, dnf}, {clare carney, 0:21:17, dnf, dnf}, {ines estedt, 0:19:50, dnf, dnf}, {aline garza, 0:20:14, dnf, dnf}, {mariko, yamazaki, 0:19:34, dnf, dnf}}

Monte Carlo, Monaco note: swim, trans 1, cycle+swim, trans 2 and run splits

{monte carlo monaco {magali messmer, 0:19:17.5, 0:00:46.2, 1:23:34.4, 0:00:35.4, 0:38:37.2}, {sharon donnelly, 0:19:13.2, 0:00:45.3, 1:23:33.3, 0:00:33.9, 0:38:57.3}, {loretta harrop, 0:19:15.5, 0:00:40.7, 1:23:33.2, 0:00:37.0, 0:39:16.8}, {karen smyers, 0:20:46.6, 0:00:52.7, 1:24:37.2, 0:00:37.5, 0:38:17.0}, {siri lindley, 0:20:11.9, 0:00:49.0, 1:24:40.3, 0:00:36.0, 0:38:22.3}, {carol montgomery, 0:20:15.8, 0:00:51.3, 1:24:45.7, 0:00:39.9, 0:38:36.7}, {anja dittmer, 0:20:08.2, 0:00:48.9, 1:24:33.9, 0:00:42.9, 0:38:54.7}, {ingrid van lubek, 0:20:22.4, 0:00:49.9, 1:24:39.9, 0:00:41.5, 0:38:50.2}, {akiko hirao, 0:20:45.3, 0:00:59.8, 1:24:42.8, 0:00:38.1, 0:38:52.5}, {nicole hackett , 0:19:13.7, 0:00:48.1, 1:23:33.9, 0:00:34.4, 0:40:09.0}, {mieke suys, 0:20:11.0, 0:00:51.9, 1:24:37.7, 0:00:39.3, 0:39:18.5}, {ines estedt, 0:20:19.4, 0:00:51.2, 1:24:38.9, 0:00:48.1, 0:39:22.4}, {isabelle mouthon, 0:20:03.5, 0:00:47.1, 1:24:36.0, 0:00:40.2, 0:39:38.4}, {melissa ashton, 0:20:02.9, 0:00:49.3, 1:24:46.9, 0:00:36.9, 0:39:50.3}, {evelyn willamson, 0:20:08.7, 0:00:53.3, 1:24:36.9, 0:00:44.3, 0:40:17.3}, {tracy hargreaves, 0:21:30.4, 0:00:47.0, 1:27:29.5, 0:00:36.3, 0:37:31.5}, {lucienne groenendijk, 0:20:46.3, 0:00:53.4, 1:24:41.0, 0:00:40.5, 0:40:25.2}, {jennifer gutierrez, 0:19:18.3, 0:00:51.0, 1:24:39.4, 0:00:41.3, 0:40:40.1}, {nina anisimova, 0:20:15.9, 0:00:53.1, 1:24:42.9, 0:00:39.4, 0:40:46.3}, {gail laurence, 0:19:26.4, 0:00:49.1, 1:24:35.9, 0:00:43.8, 0:41:09.4}, {kim carter, 0:19:17.6, 0:00:51.5, 1:24:44.6, 0:00:40.9, 0:41:14.2}, {katja schumacher, 0:20:44.0, 0:00:52.3, 1:24:38.0, 0:00:47.8, 0:41:26.5}, {maribel blanco, 0:21:16.7, 0:00:51.1, 1:27:28.5, 0:00:39.5, 0:38:40.4}, {silvia pepels, 0:20:04.5, 0:00:49.9, 1:24:36.6, 0:00:39.2, 0:41:41.4}, {ute muckel, 0:19:15.8, 0:00:45.7, 1:23:35.2, 0:00:39.9, 0:43:03.6}, {oksana zhuravkov, 0:20:15.5, 0:00:52.3, 1:24:38.7, 0:00:41.9, 0:42:25.5}, {natascha badmann, 0:21:50.9, 0:00:48.0, 1:27:26.9, 0:00:32.3, 0:40:00.6}, {jill newman, 0:20:12.5, 0:00:56.4, 1:24:44.2, 0:00:38.0, 0:42:53.2}, {beth thompson, 0:19:49.1, 0:00:52.5, 1:24:38.3, 0:00:40.1, 0:43:10.2}, {aniko gog, 0:21:35.1, 0:00:44.7, 1:27:30.9, 0:00:37.5, 0:41:12.1}, {mary ellen powers, 0:19:21.3, 0:01:00.0, 1:24:45.2, 0:00:47.1, 0:44:48.4}, {sophie delemer, 0:21:13.5, 0:00:59.7, 1:27:27.4, 0:00:35.5, 0:42:18.4}, {dominique donner, 0:20:22.4, 0:00:58.2, 1:24:43.0, 0:00:46.2, 0:45:15.5}, {jasmine haemmerle, 0:21:29.3, 0:00:50.4, 1:29:31.8, 0:00:40.0, 0:40:56.9}, {helen salomon, 0:21:53.8, 0:00:53.2, 1:29:30.0, 0:00:34.0, 0:41:01.8}, {densie mclaughlin, 0:21:53.0, 0:00:53.0, 1:29:29.9, 0:00:45.4, 0:41:32.4}, {nora edosceny, 0:20:41.4, 0:00:54.0, 1:33:02.2, 0:00:35.7, 0:40:46.7}, {marie overbye, 0:21:26:.3, 0:00:47.9, 1:27:29.7, 0:00:39.4, dnf}, {agnes eppers, 0:26:22.3, 0:00:47.6, dnf, dnf, dnf}, {mariana ohata, 0:20:46.2, 0:00:51.5, dnf, dnf, dnf}, {callahan hatfield, 0:20:49.7, 0:00:56.3, dnf, dnf, dnf}, {annie emmerson, 0:22:50.9, 0:00:54.6, dnf, dnf, dnf}}

Kona, USA note: swim, trans 1, cycle+swim, trans 2 and run splits note: swim 1 lap, cycle 7 laps, run 5 laps

{kona, usa {michellie jones, 0:17:40.2, 0:00:16.9, 1:33:09.6, 0:00:25.7, 0:37:57.0}, {barbara lindquist, 0:17:23.0, 0:00:16.5, 1:33:10.2, 0:00:25.0, 0:38:02.8}, {rina hill, 0:17:25.0, 0:00:18.9, 1:33:09.0, 0:00:29.1, 0:39:17.0}, {laura reback, 0:17:23.7, 0:00:16.5, 1:33:10.5, 0:00:26.2, 0:40:07.0}, {gail laurence, 0:17:45.3, 0:00:16.8, 1:33:09.3, 0:00:28.0, 0:40:25.7}, {jennifer gutierrez, 0:17:27.1, 0:00:22.2, 1:33:10.8, 0:00:30.0, 0:41:34.7}, {akiko hirao, 0:19:07.8, 0:00:19.5, 1:37:38.5, 0:00:34.0, 0:38:30.5}, {haruna hosoya, 0:17:44.6, 0:00:15.5, 1:37:34.9, 0:00:28.4, 0:39:22.5}, {yukie koumegawa, 0:19:06.5, 0:00:17.6, 1:37:36.9, 0:00:27.9, 0:39:49.8}, {katie webb, 0:19:07.0, 0:00:21.3, 1:37:37.4, 0:00:32.5, 0:40:08.1}, {brigitte mcmahon, 0:19:06.7, 0:00:18.7, 1:37:38.6, 0:00:32.3, 0:40:39.2}, {isabelle baird, 0:19:06.9, 0:00:19.5, 1:37:41.5, 0:00:28.3, 0:40:56.1}, {gina derks, 0:17:44.9, 0:00:18.0, 1:37:33.6, 0:00:36.7, 0:41:34.7}, {martha sorensen, 0:20:41.1, 0:00:18.1, 1:39:22.3, 0:00:30.3, 0:40:08.7}, {kiyomi niwata, 0:19:07.4, 0:00:16.2, 1:37:39.5, 0:00:35.1, 0:42:04.7}, {machiko nakanishi, 0:17:40.9, 0:00:16.3, 1:37:38.8, 0:00:28.2, 0:42:48.1}, {becky gibbs, 0:17:38.5, 0:00:16.1, 1:33:11.6, 0:00:31.5, 0:47:53.8}, {susanne martineau, 0:20:43.0, 0:00:14.3, 1:39:20.1, 0:00:31.2, 0:42:40.0}, {luise soper, 0:17:38.3, 0:00:19.5, 1:37:36.4, 0:00:26.6, 0:44:42.6}, {erin philp, 0:17:47.3, 0:00:17.6, 1:37:40.6, 0:00:24.7, 0:46:38.3}, {carol montgomery, 0:17:46.4, 0:00:19.4, 1:37:39.7, 0:00:34.2, dnf}, {melissa ashton, 0:17:44.7, 0:00:19.5, dnf, dnf, dnf}, {meagan evans, 0:18:32.6, 0:00:24.4, dnf, dnf, dnf}, {jackie gallagher, 0:19:13.0, 0:00:17.8, dnf, dnf, dnf}, {agnes eppers, 0:24:21.0, 0:00:19.0, dnf, dnf, dnf}, {heather jorris, 0:20:44.0, 0:00:19.7, dnf, dnf, dnf}, {sarah johnson, 0:20:42.0, 0:00:18.7, dnf, dnf, dnf}}

Montreal, Canada note: swim time, trans 1, cycle time and total time (no run split) note:1 lap swim, 9 lap cycle, 4 lap run

{montreal canada {loretta harrop, 0:18:37, 0:00:36, 0:59:00, 1:55:28}, {jakie gallagher, 0:19:52, 0:00:36, 0:59:11, 1:56:00}, {emma carney, 0:20:00, 0:00:38, 0:59:04, 1:56:19}, {michellie jones, 0:19:50, 0:00:37, 0:59:12, 1:56:19}, {joanne king, 0:19:54, 0:00:42, 0:59:05, 1:56:36}, {sian brice, 0:19:51, 0:00:38, 0:59:07, 1:56:53}, {jennifer gutierrez, 0:18:39, 0:00:40, 0:58:56, 1:57:06}, {mariana ohata, 0:19:16, 0:00:40, 0:59:38, 1:57:21}, {barbara lindquist, 0:18:36, 0:00:39, 0:58:59, 1:57:24}, {mieke suys, 0:19:55, 0:00:39, 0:59:02, 1:57:28}, {isabelle mouthon, 0:19:51, 0:00:37, 0:59:08, 1:57:34}, {joelle franzmann, 0:18:40, 0:00:34, 0:58:59, 1:57:39}, {akiko hirao, 0:20:03, 0:00:40, 0:58:57, 1:57:50}, {michelle dillon, 0:20:42, 0:00:40, 0:59:56, 1:57:57}, {joanna zeiger, 0:19:10, 0:00:41, 0:59:53, 1:58:04}, {evelyn williamson, 0:19:19, 0:00:43, 0:59:36, 1:58:10}, {haruna hosoya, 0:19:49, 0:00:39, 0:59:07, 1:58:11}, {magali messmer, 0:18:39, 0:00:38, 0:58:58, 1:58:11}, {anja dittmer, 0:19:56, 0:00:37, 0:59:04, 1:58:15}, {susan batholomew, 0:19:15, 0:00:41, 0:59:40, 1:58:20}, {nicole hackett, 0:18:37, 0:00:38, 0:58:59, 1:58:28}, {sharon donnelly, 0:19:51, 0:00:38, 0:59:06, 1:58:34}, {stephanie forrester, 0:20:01, 0:00:44, 1:00:31, 1:58:48}, {ingrid vanlubek, 0:20:02, 0:00:37, 0:59:02, 1:58:57}, {anja heil, 0:19:45, 0:00:38, 0:59:12, 1:59:08}, {beatrice mouthon, 0:19:57, 0:00:42, 0:59:00, 1:59:17}, {nora edosceny, 0:20:03, 0:00:45, 1:00:31, 1:59:17}, {kathleen smet, 0:19:22, 0:00:47, 0:59:29, 1:59:22}, {natalie dumas, 0:20:00, 0:00:41, 0:58:59, 1:59:24}, {erika molnar, 0:20:22, 0:00:41, 1:00:12, 1:59:43}, {kiyomi niwata, 0:20:07, 0:00:42, 1:00:30, 1:59:56}, {sandra soldan, 0:19:47, 0:00:42, 0:59:10, 1:59:57}, {renato berkova, 0:19:58, 0:00:39, 1:00:37, 2:00:12}, {jasmine haemmerle, 0:21:49, 0:00:40, 1:00:13, 2:00:26}, {carmenza morales, 0:21:52, 0:00:39, 1:00:10, 2:00:27}, {yukie koumegawa, 0:19:45, 0:00:44, 0:59:05, 2:00:39}, {manuela ianesi, 0:19:59, 0:00:44, 0:58:57, 2:00:41}, {karen smyers, 0:20:41, 0:00:45, 1:01:16, 2:01:00}, {marie overbye, 0:20:25, 0:00:39, 1:00:13, 2:01:10}, {maribel blanco, 0:20:39, 0:00:40, 0:59:58, 2:01:13}, {brigitte mcmahon, 0:19:54, 0:00:43, 0:59:06, 2:01:16}, {machiko nakanishi, 0:19:14, 0:00:38, 0:59:44, 2:01:18}, {isabelle baird, 0:20:03, 0:00:42, 1:00:28, 2:01:53}, {nancy kemp-arendt, 0:19:14, 0:00:43, 0:59:48, 2:01:59}, {beth thompson, 0:20:01, 0:00:40, 0:59:01, 2:02:32}, {edith cigana, 0:20:27, 0:00:37, 1:00:14, 2:02:56}, {natascha badmann, 0:21:53, 0:00:36, 1:00:12, 2:03:58}, {oksana zhuravkov, 0:19:50, 0:00:40, 0:59:12, 2:04:26}, {megan evans, 0:19:49, 0:00:40, 0:59:12, 2:05:36}, {iona wynter, 0:21:09, 0:00:44, 1:00:48, 2:06:09}, {francisca russli, 0:20:39, 0:00:40, 0:59:56, 2:07:26}, {shanelle barrett, 0:20:26, 0:00:38, 1:00:15, 2:08:10}, {christine hocq, 0:19:15, 0:00:42, 0:59:41, dnf}, {julie ricketts, 0:19:52, 0:00:41, 0:59:06, dnf}, {ines estedt, 0:20:25, 0:00:45, 1:00:07, dnf}, {driana piaseck, 0:20:42, 0:00:41, 1:01:20, dnf}, {silvia gemignani, 0:19:18, 0:00;42, 1:04:24, dnf}, {nina anisimova, 0:19:15, 0:00:43, dnf, dnf}, {siri lindley, 0:19:56, 0:00:39, dnf, dnf}, {louise soper, 0:20:05, 0:00:39, dnf, dnf}, {lizel moore, 0:21:51, 0:00:39, dnf, dnf}, {agnes eppers, 0:26:52, dnf, dnf, dnf}} Sydney, Australia note: swim, trans 1, swim+bike, trans 2, run splits note: swim 1 lap, cycle 6 laps, run 2 laps

{sydney australia {michellie jones, 0:19:34.9, 0:00:24.1, 1:27:04.3, 0:00:24.7, 0:36:32.6}, {loretta harrop, 0:18:42.6, 0:00:28.5, 1:27:00.1, 0:00:25.1, 0:36:37.2}, {erika molnar, 0:19:29.3, 0:00:29.7, 1:28:21.9, 0:00:32.6, 0:36:00.4}, {siri lindley, 0:19:31.7, 0:00:34.4, 1:27:10.0, 0:00:29.6, 0:37:14.2}, {rina hill, 0:18:47.3, 0:00:30.7, 1:27:10.3, 0:00:32.5, 0:37:30.4}, {barbara lindquist, 0:18:23.1, 0:00:28.9, 1:26:16.2, 0:00:26.0, 0:38:40.0}, {mieke suys, 0:19:42.5, 0:00:27.5, 1:27:07.8, 0:00:27.4, 0:37:56.0}, {sian brice, 0:19:29.2, 0:00:35.5, 1:27:08.6, 0:00:29.2, 0:37:59.9}, {sharon donnelly, 0:19:25.0, 0:00:36.6, 1:27:06.5, 0:00:25.1, 0:38:16.7}, {isabelle mouthon, 0:19:35.8, 0:00:22.0, 1:27:02.3, 0:00:28.7, 0:38:28.0}, {akiko hirao, 0:20:26.8, 0:00:36.8, 1:28:25.6, 0:00:25.8, 0:37:05.4}, {tracy hargreaves, 0:20:31.1, 0:00:31.8, 1:28:27.7, 0:00:24.0, 0:37:05.7}, {nicole hackett, 0:19:27.2, 0:00:34.1, 1:27:02.0, 0:00:23.7, 0:38:45.4}, {melissa ashton, 0:19:37.0, 0:00:43.8, 1:27:13.0, 0:00:14.5, 0:38:38.5}, {gina derks, 0:19:38.5, 0:00:33.8, 1:27:09.4, 0:00:29.1, 0:38:50.6}, {laura reback, 0:19:32.3, 0:00:32.6, 1:27:11.0, 0:00:27.8, 0:38:57.1}, {machiko nakanishi, 0:19:27.5, 0:00:30.2, 1:27:08.0, 0:00:26.2, 0:39:10.2}, {evelyn williamson, 0:21:06.3, 0:00:49.3, 1:27:05.0, 0:00:24.1, 0:39:18.4}, {christine hocq, 0:19:26.6, 0:00:34.6, 1:27:11.2, 0:00:28.1, 0:39:15.5}, {natascha badmann, 0:20:34.2, 0:00:36.3, 1:28:21.1, 0:00:22.2, 0:38:17.8}, {jenny rose, 0:20:25.2, 0:00:31.4, 1:28:22.2, 0:00:32.2, 0:38:59.3}, {yukie koumegawa, 0:19:37.8, 0:00:31.5, 1:27:08.1, 0:00:28.2, 0:40:15.2}, {kathleen smet, 0:19:38.6, 0:00:38.2, 1:27:06.4, 0:00:36.3, 0:40:31.8}, {shanelle barrett, 0:20:24.1, 0:00:34.5, 1:28:26.7, 0:00:25.5, 0:39:22.8}, {marie overbye, 0:20:25.0, 0:00:28.6, 1:28:23.7, 0:00:27.0, 0:39:41.0}, {haruna hosoya, 0:20:24.8, 0:00:36.5, 1:28:21.3, 0:00:25.6, 0:39:51.4}, {megumi shigaki, 0:20:40.9, 0:00:30.4, 1:30:39.1, 0:00:29.0, 0:37:49.0}, {jill newnam, 0:19:48.5, 0:00:35.3, 1:27:09.8, 0:00:27.6, 0:41:36.0}, {nancy kemp-arendt, 0:19:36.2, 0:00:32.7, 1:27:08.9, 0:00:28.1, 0:41:41.0}, {maribel blanco, 0:21:03.8, 0:00:37.7, 1:30:38.5, 0:00:28.1, 0:39:03.3}, {kiyomi niwata, 0:20:26.3, 0:00:28.3, 1:30:40.9, 0:00:25.0, 0:39:06.8}, {gail laurence, 0:19:36.1, 0:00:32.9, 1:28:20.5, 0:00:30.7, 0:41:34.2}, {jasmine haemmerle, 0:20:34.0, 0:00:34.3, 1:30:37.7, 0:00:26.5, 0:39:21.8}, {clare carney, 0:22:06.3, 0:00:27.8, 1:30:40.2, 0:00:26.6, 0:39:24.9}, {dominique donner, 0:19:42.1, 0:00:33.3, 1:27:11.8, 0:00:28.5, 0:42:59.2}, {manuela ianesi, 0:20:29.5, 0:00:31.3, 1:28:24.9, 0:00:28.0, 0:42:20.9}, {ingrid van lubek, 0:20:33.8, 0:00:37.8, 1:27:03.4, 0:04:11.3, 0:44:30.3}, {helen salomon, 0:21:51.0, 0:00:33.3, 1:27:18.1, 0:01:32.3, 0:44:30.9}, {kim carter, 0:19:52.9, 0:00:31.1, 1:30:43.0, 0:00:32.5, 0:44:58.2}, {karen gordon, 0:21:06.5, 0:00:43.5, 1:35:57.5, 0:00:27.1, 0:40:58.2}, {sibylle matter, 0:19:42.8, 0:00:24.9, 1:27:06.8, 0:00:25.5, dnf}, {sian welch, 0:21:51.7, 0:00:41.0, 1:27:06.8, 0:00:25.5, dnf}, {ewa dederko, 0:21:46.4, 0:00:34.4, 1:35:58.1, 0:00:29.0, dnf}, {weike hoogzaad, 0:19:34.0, 0:00:37.5, dnf, dnf, dnf}, {jane kargotich, 0:19:43.3, 0:00:37.4, dnf, dnf, dnf}, {nicole mertes, 0:20:32.6, 0:00:15.6, dnf, dnf, dnf}, {carla morena, 0:20:26.4, 0:00:31.4, dnf, dnf, dnf}, {isabelle baird, 0:21:05.5, 0:00:39.9, dnf, dnf, dnf}, {nicole andronicus, 0:21:41.9, 0:00:24.6, dnf, dnf, dnf}, {annie emmerson, 0:21:42.0, 0:00:33.0, dnf, dnf, dnf}, {louise soper, 0:22:00.5, 0:00:22.0, dnf, dnf, dnf}, {francesca tibaldi, 0:21:57.1, 0:00:39.3, dnf, dnf, dnf}}

Cancun, Mexico note: swim, cycle and run splits

{cancun mexico {jackie gallagher, 0:20:56, 0:58:29, 0:36:47}, {michellie jones, 0:20:44, 0:58:26, 0:37:08}, {karen smyers, 0:21:10, 0:00:00, 1:37:09}, {stephanie forrester, 0:20:54, 0:58:19, 0:38:20}, {mariana ohata, 0:20:48, 0:58:27, 0:38:24}, {aruna hosoya, 0:20:48, 0:58:19, 0:38:59}, {shelia tamorina, 0:19:53, 0:59:14, 0:39:07}, {yukie koumegawa, 0:20:45, 0:58:23, 0:39:15}, {carla moreno, 0:20:47, 0:58:28, 0:39:32}, {susan bartholomew, 0:20:58, 0:58:17, 0:39:43}, {barbara lindquist, 0:20:03, 0:59:12, 0:39:44}, {machiko nakanishi, 0:20:41, 0:58:29, 0:40:19}, {ute mueckel, 0:20:44, 0:58:26, 0:41:25}, {sandra solden, 0:20:47, 0:58:31, 0:41:32}, {renato berkova, 0:21:03, 1:00:13, 0:39:40}, {aniko gog, 0:22:26, 0:58:48, 0:39:55}, {ingrid van lubek, 0:22:24, 0:58:50, 0:40:01}, {lisset olivera, 0:21:08, 0:58:08, 0:42:09}, {natalie daumas, 0:22:20, 0:58:51, 0:40:26}, {silvia gemignani, 0:20:39, 0:58:33, 0:42:37}, {gina derks, 0:20:55, 0:58:43, 0:43:11}, {katja schumacher, 0:22:21, 0:58:53, 0:41:29}, {maria luisa martinez, 0:22:23, 0:58:47, 0:41:42}, {laura reback, 0:20:42, 0:58:26, 0:44:07}, {kiyomi niwata, 0:21:18, 0:59:58, 0:42:14}, {isabelle turcotte-aird, 0:21:38, 0:59:31, 0:42:51}, {katie webb, 0:22:33, 1:02:22, 0:42:22}, {marci steelman, 0:22:35, 1:02:30, 0:43:04}, {natasha yaremczuk, 0:22:37, 1:02:30, 0:43:44}, {marie overbye, 0:22:26, 0:58:51, dnf}, {gail laurence, 0:21:00, 0:58:15, dnf}, {carol montgomery, 0:20:57, dnf, dnf}, {iona wynter, 0:22:29, 1:02:18, dnf}, {manuela ianesi, 0:21:17, 1:00:01, dnf}, {aline garza, 0:22:31, 1:02:36, dnf}, {heidi jesberger, 0:25:17, dnf, dnf}}

Male Triathlons

Cancun, Mexico Has splits for swim cycle and run (h:mm:ss)

{cancun mexico, {volodymyr polikarpenko, 0:19:00, 0:54:30, 0:32:49}, {ivan rana, 0:19:24, 0:53:50, 0:33:10}, {dimitry gaag, 0:19:01, 0:53:44, 0:33:39}, {filip osplay, 0:19:05, 0:54:18, 0:33:30}, {andriy glushchenko, 0:19:11, 0:53:55, 0:33:37}, {jan hansen, 0:19:41, 0:53:40, 0:33:43}, {nick radkewich, 0:19:20, 0:53:51, 0:33:37}, {gilberto gonzalez, 0:19:25, 0:53:51, 0:33:51}, {leandro macedo, 0:19:42, 0:53:34, 0:33:46}, {juracy moreira, 0:19:35, 0:53:41, 0:34:04}, {joe umphenour, 0:19:18, 0:53:52, 0:34:13}, {philippe fattori, 0:19:28, 0:53:31, 0:34:12}, {uzziel valderrabano, 0:19:32, 0:00:00, 1:28:27}, {greg bennett, 0:19:22, 0:53:43, 0:34:31}, {takumi obara, 0:00:00, 0:00:00, 1:49:26}, {norbert domnik, 0:19:58, 0:53:25, 0:34:34}, {alessandro bottoni, 0:19:50, 0:53:52, 0:34:40}, {eligio cervantes, 0:19:30, 0:53:38, 0:34:50}, {tim don, 0:19:36, 0:53:48, 0:34:58}, {rob barel, 0:19:19, 0:53:46, 0:35:21}, {oscar galindez, 0:19:48, 0:51:59, 0:36:39}, {johannes enzenhofer, 0:19:29, 0:53:46, 0:35:23}, {hunter kemper, 0:19:21, 0:53:54, 0:35:19}, {tony deboom, 0:00:00, 0:00:00, 1:50:25}, {matias brain, 0:19:32, 0:53:46, 0:35:19}, {jose luis zepeda, 0:19:03, 0:54:18, 0:35:45}, {andrew johns, 0:19:00, 0:53:34, 0:36:12}, {hideo fukui, 0:19:05, 0:54:11, 0:36:13}, {david haines, 0:19:29, 0:53:45, 0:36:06}, {jose rodriguez, 0:20:34, 0:00:00, 1:30:34}, {armando barcellos, 0:19:40, 0:53:35, 0:36:11}, {luiz preta, 0:19:27, 0:53:53, 0:36:11}, {raul lemir, 0:19:44, 0:53:37, 0:36:14}, {marcin wedlarski, 0:19:29, 0:53:51, 0:36:26}, {roland melis, 0:19:58, 0:53:26, 0:36:26}, {jose merchan, 0:19:31, 0:53:46, 0:36:34} ,{joshua, dapice, 0:19:30, 0:53:54, 0:36:34}, {vassilis krommidas, 0:19:12, 0:54:00, 0:36:37}, {wes hobson, 0:19:53, 0:53:27, 0:36:44}, {brad bevan, 0:19:06, 0:54:14, 0:36:59}, {xavier llobet, 0:19:54, 0:53:27, 0:36:55}, {jun-ichi yamamoto, 0:19:22, 0:53:48, 0:36:59}, {stuart hayes, 0:19:07, 0:54:02, 0:37:02}, {carlos probert, 0:19:21, 0:53:29, 0:37:24}, {dennis looze, 0:19:04, 0:00:00, 1:32:06}, {brian rhodes, 0:19:26, 0:53:45, 0:37:53}, {arturo garza, 0:19:37, 0:53:36, 0:38:04}, {hideo nakagome, 0:19:46, 0:53:46, 0:37:52}, {michael smedley, 0:19:10, 0:53:59, 0:38:21}, {leonardo casadio, 0:19:29, 0:53:49, 0:38:22}, {richard stannard, 0:18:58, 0:00:00, 1:35:13}, {peter robertson, 0:20:14, 0:53:12, 0:39:39}, {camilo gonzalez, 0:20:49, 0:56:55, 0:35:36}, {eugenio chimal, 0:00:00, 0:00:00, 1:54:31}, {felix molina, 0:20:37, 0:53:55, 0:41:03}, {alvaro martinez, 0:20:38, 0:56:53, 0:38:14}, {francisco serrano, 0:20:36, 0:56:59, 0:38:44}, {allan villaneva, 0:21:00, 0:56:51, 0:39:06}, {preben jeppesen, 0:19:57, 0:53:31, 0:44:22}, {chris mccormack, 0:20:00, dnf, dnf}, {chris hill, 0:19:05, dnf, dnf}, {martin krnavek, 0:19:10, 0:54:14, dnf}, {eneko llanos, 0:19:27, 0:53:42, dnf}, {stephane poulat, 0:19:08, 0:54:19, dnf}, {alec rukosuev, 0:19:05, 0:54:20, dnf}, {stefano belandi, 0:19:09, 0:54:33, dnf}, {richard allen, 0:19:55, 0:53:29, dnf}, {lach vollmerhaus, 0:19:29, 0:53:48, dnf}, {fabrizio ferraresi, 0:21:30, dnf, dnf}, {gianpietro de faveri, 0:19:19, 0:53:58, dnf}, {garrett mccarty, 0:19:28, 0:53:48, dnf}, {stefan timms, 0:19:20, 0:53:55, dnf}, {kevin carter, 0:20:10, dnf, dnf}, {nigel wynter, 0:25:00, dnf, dnf}, {daniel barrera, 0:21:35, dnf, dnf}, {carlo catulini, 0:19:57, dnf, dnf}, {rodolfo gonzalez, 0:25:03, dnf, dnf}}

Kapelle-op-den bos, Belgium Has splits for swim, trans 1, cycle, trans 2, run (h:mm:ss.0) note swim 1 lap, cycle 6 laps, run 3 laps

{kapelle-op-den bos belgium, {simon lessing, 0:18:50.7, 0:00:40.3, 0:53:51.3, 0:00:39.7, 0:31:31.8}, {reto hug, 0:18:58.1, 0:00:36.8, 0:53:43.5, 0:00:37.8, 0:31:46.5}, {andrew johns, 0:19:11.2, 0:00:38.6, 0:53:30.5, 0:00:41.2, 0:31:41.4}, {philippe fattori, 0:19:36.6, 0:00:37.6, 0:53:14.2, 0:01:31.5, 0:31:09.9}, {filip osplay, 0:18:57.3, 0:00:46.9, 0:53:39.0, 0:00:45.5, 0:32:04.5}, {jan hansen, 0:19:33.3, 0:00:36.8, 0:53:12.8, 0:00:41.7, 0:32:11.5}, {miles stewart, 0:19:08.6, 0:00:40.1, 0:53:29.4, 0:00:37.3, 0:32:25.3}, {tim don, 0:19:06.6, 0:00:41.4, 0:53:37.2, 0:00:41.5, 0:32:18.8}, {stefano belandi, 0:19:27.1, 0:00:39.3, 0:53:20.2, 0:00:45.3, 0:32:22.8}, {vladimir polikarpenko, 0:18:55.9, 0:00:38.4, 0:53:53.3, 0:00:43.9, 0:32:24.9}, {lothar leder, 0:19:44.2, 0:00:35.1, 0:53:01.8, 0:00:42.5, 0:32:38.6}, {markus keller, 0:19:14.3, 0:00:38.2, 0:53:26.8, 0:00:47.0, 0:32:49.7}, {dimitri gaag, 0:19:12.3, 0:00:37.7, 0:53:36.9, 0:00:44.5, 0:32:50.3}, {stefan vuckovic, 0:19:23.9, 0:00:37.2, 0:53:22.9, 0:00:41.6, 0:32:58.4}, {richard allen, 0:19:30.8, 0:00:37.8, 0:53:08.9, 0:00:40.6, 0:33:10.3}, {luc van lierde, 0:19:23.2, 0:00:42.0, 0:53:16.4, 0:00:48.0, 0:33:03.3}, {samuel peirreclaud, 0:19:09.7, 0:00:39.9, 0:53:38.2, 0:00:42.6, 0:33:03.6}, {arnd schomburg, 0:19:09.9, 0:00:37.2, 0:53:34.0, 0:00:42.1, 0:33:28.5}, {dennis looze, 0:18:56.8, 0:00:40.6, 0:53:29.7, 0:00:43.2, 0:33:56.0}, {martin matula, 0:19:03.7, 0:00:42.8, 0:53:44.0, 0:00:45.9, 0:33:30.1}, {stephane poulat, 0:18:52.4, 0:00:41.1, 0:53:46.9, 0:00:43.6, 0:34:03.3}, {marc jenkins, 0:19:09.0, 0:00:40.0, 0:53:40.6, 0:00:45.4, 0:34:01.6}, {franck bignet, 0:19:05.3, 0:00:39.6, 0:53:37.5, 0:00:39.7, 0:34:14.7}, {jonas djurback, 0:19:19.9, 0:00:40.0, 0:53:23.7, 0:00:40.7, 0:34:17.9}, {peter alder, 0:19:04.2, 0:00:37.8, 0:53:44.4, 0:00:41.9, 0:34:20.0}, {richard stannard, 0:18:50.8, 0:00:42.1, 0:53:46.7, 0:00:43.3, 0:34:39.7}, {gianpietro de faveri, 0:19:16.1, 0:00:39.1, 0:53:33.6, 0:00:45.3, 0:34:29.0}, {yves coura, 0:19:16.7, 0:00:49.2, 0:53:32.2, 0:00:45.6, 0:34:31.3}, {andriy glushchenko, 0:19:06.9, 0:00:41.1, 0:53:40.6, 0:00:41.4, 0:34:51.2}, {ralf eggert, 0:19:47.9, 0:00:37.4, 0:54:27.3, 0:00:38.7, 0:33:30.3}, {juracy moreira, 0:19:44.8, 0:00:38.9, 0:54:30.7, 0:00:46.1, 0:33:45.4}, {eric van der linden, 0:00:00.0, 0:00:00.0, 1:14:55.1, 0:00:46.4, 0:34:02.5}, {pierre-alain frossard, 0:21:00.7, 0:00:38.8, 0:54:15.3, 0:00:39.8, 0:33:12.2}, {markus forster, 0:20:51.1, 0:00:38.2, 0:54:26.3, 0:00:41.1, 0:33:18.1}, {david hyam, 0:19:22.8, 0:00:44.0, 0:53:23.1, 0:00:43.9, 0:35:42.1}, {alessandro bottoni, 0:19:55.8, 0:00:40.3, 0:55:20.9, 0:00:46.1, 0:33:20.3}, {fedor filipov, 0:19:25.1, 0:00:42.6, 0:54:48.2, 0:00:43.4, 0:34:49.0}, {johannes enzenhofer, 0:19:52.3, 0:00:38.2, 0:55:24.0, 0:00:42.5, 0:34:15.1}, {oscar galindez, 0:22:03.9, 0:00:39.7, 0:53:13.3, 0:00:44.0, 0:34:15.5}, {luc huntjens, 0:19:01.5, 0:00:41.8, 0:53:24.5, 0:00:54.7, 0:37:05.8}, {leandro macedo, 0:20:56.6, 0:00:36.8, 0:54:22.7, 0:00:40.4, 0:34:47.1}, {craig walton, 0:18:46.3, 0:00:37.0, 0:53:53.2, 0:00:39.4, 0:37:33.6}, {joe umphenour, 0:19:53.6, 0:00:42.9, 0:55:21.6, 0:00:44.0, 0:34:58.2}, {roland melis, 0:20:57.9, 0:00:40.6, 0:54:16.6, 0:00:46.5, 0:35:16.4}, {vassilis krommidas, 0:19:52.6, 0:00:38.9, 0:55:26.2, 0:00:46.0, 0:35:16.5}, {ricky jorgensen, 0:19:17.6, 0:00:39.1, 0:53:29.3, 0:00:55.6, 0:38:39.8}, {frederik van lierde, 0:19:46.8, 0:00:37.7, 0:54:28.8, 0:00:38.9, 0:38:29.6}, {andrew tarry, 0:22:11.2, 0:00:48.8, 0:56:53.8, 0:00:44.9, 0:33:51.2}, {james quann, 0:19:56.2, 0:00:38.8, 0:55:24.1, 0:00:44.0, 0:38:23.1}, {simon finch, 0:22:12.3, 0:00:41.0, 0:56:50.5, 0:00:53.8, 0:34:55.5}, {axel zeebroek, 0:19:44.9, 0:00:37.1, 0:54:28.5, 0:00:39.5, dnf}, {david koppensteiner, 0:22:05.7, 0:00:41.0, 0:57:06.0, 0:00:50.0, dnf}, {martin krnavek, 0:19:43.5, 0:00:44.9, dnf, dnf, dnf}}

Montreal, Canada Has splits for swim, trans 1, cycle and total time (h:mm:ss) note swim 1 lap, cycle 9 laps, run 4 laps note 1999 world champs

{montreal canada {dimitry gagg, 0:18:27, 0:00:36, 0:54:27, 1:45:25}, {simon lessing, 0:18:07, 0:00:37, 0:54:46, 1:45:31}, {miles stewart, 0:18:15, 0:00:36, 0:54:35, 1:45:47}, {andrew johns, 0:18:23, 0:00:33, 0:54:30, 1:45:56}, {martin krnavek, 0:18:06, 0:00:42, 0:54:46, 1:45:59}, {reto hug, 0:18:38, 0:00:35, 0:54:14, 1:46:03}, {simon whitfeild, 0:18:37, 0:00:35, 0:54:22, 1:46:03}, {hunter kemper, 0:18:10, 0:00:34, 0:54:47, 1:46:04}, {filip osplay, 0:18:09, 0:00:40, 0:54:44, 1:46:06}, {markus keller, 0:18;26, 0:00:39, 0:54:25, 1:46:08}, {hamish carter, 0:18:01, 0:00:37, 0:54:49, 1:46:10}, {greg welch, 0:18:11, 0:00:39, 0:54:37, 1:46:13}, {craig watson, 0:18:05, 0:00:42, 0:54:45, 1:46:14}, {lothar leder, 0:18:42, 0:00:32, 0:54:19, 1:46:19}, {vladimir polikarpenko, 0:18:22, 0:00:38, 0:54:36, 1:46:32}, {jan rehula, 0:18:17, 0:00:38, 0:54:36, 1:46:37}, {jamie hunt, 0:18:30, 0:00:34, 0:54:35, 1:46:40}, {chris mccormack, 0:18:36, 0:00:36, 0:54:20, 1:46:43}, {nick radewich, 0:18:16, 0:00:36, 0:54:38, 1:46:45}, {shane reed, 0:18:04, 0:00:36, 0:54:48, 1:46:53}, {tim don, 0:18:19, 0:00:32, 0:54:43, 1:47:00}, {greg bennett, 0:18:16, 0:00:38, 0:54:38, 1:47:07}, {sebastien, berlier, 0:18:28, 0:00:35, 0:54:24, 1:47:09}, {jan hansen, 0:18:40, 0:00:35, 0:54:20, 1:47:10}, {ralf eggert, 0:18:30, 0:00:32, 0:54:27, 1:47:11}, {johannes enzenhofer, 0:18:12, 0:00:38, 0:54:43, 1:47:17}, {kris gemmell, 0:18:07, 0:00:36, 0:54:45, 1:47:26}, {philippe fattori, 0:18:17, 0:00:40, 0:54:15, 1:47:27}, {eneko llanos, 0:18:39, 0:00:35, 0:54:17, 1:47:28}, {carl blasco, 0:18:12, 0:00:45, 0:54:33, 1:47:30}, {fedor filipov, 0:18:17, 0:00:41, 0:54:38, 1:47:33}, {hector llanos, 0:18:18, 0:00:40, 0:54:40, 1:47:38}, {brad bevan, 0:18:06, 0:00:36, 0:54:47, 1:47:47}, {stefan vuckovic, 0:18:07, 0:00:29, 0:54:54, 1:48:00}, {dennis looze, 0:18:24, 0:00:40, 0:54:31, 1:48:04}, {conrad stoltz, 0:18:14, 0:00:34, 0:54:42, 1:48:07}, {greg von holdt, 0:18:34, 0:00:36, 0:54:21, 1:48:13}, {rob barel, 0:18:43, 0:00:34, 0:54:10, 1:48:17}, {matthew reed, 0:18:06, 0:00:37, 0:54:46, 1:48:19}, {alec rukosuev, 0:18:04, 0:00:37, 0:54:57, 1:48:21}, {jonas djurback, 0:18:37, 0:00:39, 0:54:19, 1:48:22}, {oscar galindez, 0:19:23, 0:00:39, 0:53:31, 1:48:27}, {michael smedley, 0:18:19, 0:00:35, 0:54:40, 1:48:33}, {takumi obara, 0:18:37, 0:00:39, 0:54:17, 1:48:33}, {marc jenkins, 0:18:18, 0:00:37, 0:54:41, 1:48:40}, {laurent jeanselme, 0:17:56, 0:00:40, 0:55:02, 1:48:46}, {stefan timms, 0:18:31, 0:00:36, 0:54:29, 1:48:49}, {csaba kuttor, 0:18:20, 0:00:36, 0:54:43, 1:48:54}, {uzziel valderrabano, 0:18:47, 0:00:35, 0:54:17, 1:48:56}}

Ishigaki, Japan Has splits for swim, cycle and run

{ishigaki japan {greg welch, 0:17:00, 0:56:40, 0:33:16}, {greg bennett, 0:16:59, 0:56:43, 0:33:26}, {andrew johns, 0:17:02, 0:56:39, 0:33:50}, {hamish carter, 0:16:53, 0:56:48, 0:34:07}, {chris hill, 0:16:51, 0:56:46, 0:34:29}, {chris mccormack, 0:16:52, 0:56:48, 0:34:29}, {jan rehula, 0:17:03, 0:56:39, 0:34:30}, {craig watson, 0:16:59, 0:56:46, 0:35:05}, {shane reed, 0:16:59, 0:56:44, 0:35:26}, {dennis looze, 0:17:00, 0:56:47, 0:35:45}, {miles stewart, 0:17:01, 0:56:41, 0:35:53}, {jose luiz zepeda, 0:16:55, 0:56:51, 0:35:49}, {laurent, jeanselme, 0:16:54, 0:56:55, 0:35:47}, {lach vollmerhaus, 0:17:08, 0:56:38, 0:35:53}, {sebastien berlier, 0:17:31, 0:57:50, 0:34:39}, {dimitry gaag, 0:17:11, 0:58:12, 0:34:38}, {ralf eggert, 0:17:41, 0:57:36, 0:34:45}, {hideo fukui, 0:17:02, 0:56:40, 0:36:29}, {lothar leder, 0:17:30, 0:57:50, 0:34:57}, {nick radkewich, 0:17:06, 0:58:23, 0:34:49}, {markus forster, 0:17:42, 0:57:38, 0:35:14}, {reto hug, 0:17:27, 0:57:54, 0:35:17}, {eligio cervantes, 0:17:05, 0:58:18, 0:35:17}, {jean christoph guinchard, 0:17:29, 0:57:50, 0:35:22}, {luc van lierde, 0:17:13, 0:58:10, 0:35:27}, {hayato suzuki, 0:17:32, 0:57:49, 0:35:30}, {domnik norbert, 0:17:50, 0:58:06, 0:35:01}, {eneko llanos, 0:17:27, 0:57:57, 0:35:36}, {rob barel, 0:17:40, 0:58:06, 0:35:17}, {martin krnavek, 0:17:09, 0:58:19, 0:35:41}, {craig alexander, 0:17:02, 0:56:46, 0:37:28}, {peter alder, 0:16:47, 0:57:01, 0:37:31}, {conrad stoltz, 0:17:31, 0:57:51, 0:36:02}, {emmanuel dubreuil, 0:16:59, 0:58:32, 0:35:56}, {eric van der linden, 0:17:42, 0:58:07, 0:35:44}, {joshua dapice, 0:17:19, 0:58:03, 0:36:13}, {luc huntjens, 0:17:05, 0:56:42, 0:37:52}, {marc jenkins, 0:17:01, 0:58:25, 0:36:17}, {markus keller, 0:17:11, 0:58:12, 0:36:21}, {arnd schomburg, 0:17:08, 0:58:16, 0:36:22}, {jan hansen, 0:17:43, 0:58:12, 0:35:55}, {jose merchan, 0:17:44, 0:58:10, 0:35:57}, {garrett mccarthy, 0:17:15, 0:58:11, 0:36:30}, {csaba kuttor, 0:17:04, 0:58:25, 0:36:34}, {greg von holdt, 0:17:41, 0:58:18, 0:36:08}, {stefan vuckovic, 0:17:41, 0:57:35, 0:36:59}, {andreas raelert, 0:16:55, 0:56:54, 0:38:28}, {takumi obara, 0:17:15, 0:58:09, 0:36:54}, {uziel valderrabano, 0:17:33, 0:57:55, 0:36:55}, {simon whitfeild, 0:17:32, 0:57:51, 0:37:10}, {johannes enzebhofer, 0:17:30, 0:58:24, 0:36:42}, {richard allen, 0:17:46, 0:58:06, 0:36:48}, {dirk bockel, 0:17:03, 0:58:17, 0:37:29}, {arturo garza, 0:17:47, 0:58:05, 0:37:00}, {mark bates, 0:17:41, 0:58:14, 0:36:59}, {tim don, 0:17:00, 0:58:19, 0:37:43}, {lieuw boonstra, 0:17:42, 0:58:14, 0:37:16}, {ricky jorgensen, 0:17:07, 0:58:18, 0:37:51}, {michael smedley, 0:17:09, 0:58:18, 0:38:02}, {jun-ichi yamamoto, 0:17:01, 0:58:26, 0:38:21}, {roland melis, 0:17:48, 0:58:16, 0:38:03}, {tomoo chiba, 0:17:08, 0:58:16, 0:39:02}, {carlos probert, 0:17:47, 0:58:09, 0:38:58}, {guido gosselink, 0:17:48, 0:58:09, 0:39:07}, {ryan bolton, 0:17:34, 0:58:25, 0:39:42}, {teppei takeuchi, 0:17:25, 0:58:35, 0:39:48}, {matthew reed, 0:17:12, 0:58:07, 0:41:13}, {peter hobor, 0:17:33, 0:58:26, 0:41:00}, {kazuo sudo, 0:18:43, 1:02:14, 0:37:48}, {hirteru saito, 0:17:32, 1:03:19, 0:39:02}, {ido maori, 0:18:24, 1:02:35, 0:39:27}, {hiroyuki nishiuchi, 0:17:49, 1:03:04, 0:40:12}, {jamie hunt, 0:17:37, 0:57:52, dnf}, {javier rasas, 0:17:33, dnf, dnf}, {joachim willen, 0:17:14, 0:58:15, dnf}, {hector llanos, 0:17:15, dnf, dnf}, {jonas djurback, 0:17:35, dnf, dnf}, {leonardo nolasco, 0:18:37, 1:02:22, dnf}, {jared letica, 0:18:41, dnf, dnf}, {stefan timms, 0:17:42, 1:03:12, dnf}}

Gamagori, Japan Has splits for swim, cycle and run

{gamagori japan {chris mccormack, 0:17:20, 0:56:36, 0:29:42}, {craig watson, 0:17:08, 0:56:53, 0:29:44}, {reto hug, 0:17:17, 0:56:39, 0:30:01}, {eric van der linden, 0:17:27, 0:56:24, 0:30:18}, {matthew reed, 0:17:28, 0:56:28, 0:30:18}, {brad bevan, 0:17:11, 0:56:52, 0:30:21}, {markus keller, 0:17:36, 0:56:24, 0:30:29}, {nick radkewich, 0:17:23, 0:56:41, 0:30:30}, {dimitry gaag, 0:17:14, 0:57:03, 0:30:25}, {greg welch, 0:17:12, 0:56:43, 0:30:50}, {laurent jeanselme, 0:17:18, 0:56:44, 0:30:52}, {tim don, 0:17:19, 0:56:46, 0:30:50}, {javier rosas, 0:17:25, 0:56:41, 0:30:51}, {csaba kuttor, 0:17:06, 0:57:07, 0:30:48}, {conrad stoltz, 0:17:37, 0:56:31, 0:30:54}, {eneko llanos, 0:17:42, 0:56:22, 0:31:07}, {marc jenkins, 0:17:20, 0:56:54, 0:31:02}, {andreas realert, 0:17:23, 0:56:43, 0:31:16}, {stefan vuckovic, 0:17:44, 0:56:18, 0:31:37}, {jean christoph guinchard, 0:17:38, 0:56:31, 0:31:41}, {lach vollmerhaus, 0:17:14, 0:56:51, 0:31:53}, {jose luiz zepeda, 0:17:16, 0:57:11, 0:31:46}, {peter alder, 0:17:38, 0:56:30, 0:32:08}, {richard allen, 0:17:44, 0:57:42, 0:30:52}, {hideo fukui, 0:17:08, 0:56:48, 0:32:24}, {craig ball, 0:17:39, 0:56:27, 0:32:15}, {ralf eggert, 0:17:32, 0:56:31, 0:32:19}, {sebastien berlier, 0:17:39, 0:56:24, 0:32:27}, {martin krnavek, 0:17:11, 0:57:07, 0:32:13}, {rob barel, 0:17:20, 0:56:41, 0:32:31}, {markus forster, 0:18:08, 0:57:18, 0:31:07}, {joachim willen, 0:17:23, 0:56:43, 0:32:28}, {craig alexander, 0:17:43, 0:57:53, 0:31:03}, {jan hansen, 0:17:41, 0:56:31, 0:32:29}, {lothar leder, 0:17:50, 0:57:44, 0:31:13}, {ricky jorgensen, 0:17:18, 0:56:56, 0:32:39}, {garrett mccarthy, 0:17:17, 0:56:56, 0:32:59}, {shane reed, 0:17:13, 0:56:48, 0:33:26}, {vladimir polikarpenko, 0:17:25, 0:58:10, 0:31:57}, {hector llanos, 0:17:22, 0:58:34, 0:31:38}, {jose merchan, 0:17:41, 0:58:01, 0:32:26}, {simon whitfeild, 0:17:52, 0:59:21, 0:30:59}, {dirk bockel, 0:18:04, 0:57:35, 0:32:36}, {uziel valderrabano, 0:17:46, 0:59:32, 0:30:58}, {carlos probert, 0:17:47, 0:57:53, 0:32:53}, {luc huntjens, 0:17:24, 0:56:48, 0:34:32}, {takumi obara, 0:17:36, 0:56:27, 0:34:54}, {eligio cervantes, 0:18:01, 0:59:10, 0:31:50}, {jun-ichi yamamoto, 0:17:45, 0:57:43, 0:33:34}, {johannes enzenhofer, 0:17:43, 0:58:05, 0:33:17}, {arnd schomburg, 0:17:44, 0:58:11, 0:33:21}, {michael smedley, 0:17:29, 0:59:54, 0:31:57}, {jamie hunt, 0:18:14, 1:00:28, 0:30:39}, {teppei, takeuchi, 0:17:28, 0:56:37, 0:35:23}, {greg von holdt, 0:17:43, 0:59:42, 0:32:09}, {guido gosselink, 0:18:14, 0:59:04, 0:32:20}, {tomoo chiba, 0:17:37, 0:56:46, 0:35:24}, {mark bates, 0:17:45, 1:00:48, 0:31:54}, {peter hobor, 0:18:01, 1:00:43, 0:32:32}, {stefan timms, 0:17:46, 1:00:51, 0:33:12}, {roland melis, 0:18:29, 1:00:03, 0:33:41}, {domnik norbert, 0:17:49, 0:58:00, 0:36:43}, {chundi lu, 0:18:02, 1:00:48, 0:34:10}, {andriy glushchenko, 0:17:26, 1:01:23, 0:34:20}, {jared letica, 0:18:27, 1:02:55, 0:32:09}, {hayato suzuki, 0:18:34, 0:59:57, 0:35:58}, {hamish carter, 0:17:15, dnf, dnf}, {chris hill, 0:17:07, 0:56:55, dnf}, {andrew johns, 0:17:21, dnf, dnf}, {jan rehula, 0:17:16, 0:57:22, dnf}, {dennis looze, 0:17:17, dnf, dnf}, {artur garza, 0:18:15, dnf, dnf}, {lieuw boonstra, 0:17:25, dnf, dnf}, {leonardo nolasco, 0:18:39, 1:02:50, dnf}, {emmanuel dubreuil, 0:17:35, dnf, dnf}, {joshua dapice, 0:17:22, dnf, dnf}, {luc van lierde, 0:17:47, dnf, dnf}, {ido maori, 0:18:37, dnf, dnf}, {ryosuke hoshino, 0:17:55, 1:03:26, dnf}, {kazuo sudo, 0:18:31, dnf, dnf}}

Kona, USA Has splits for swim, transition 1, swim+t1+cycle, transition 2 and run note swim 1 lap, cycle 7 laps and run 2 laps

{kona usa {greg welch, 0:17:00.5, 0:00:16.8, 1:30:11.7, 0:00:23.2, 0:33:03.6}, {chris hill, 0:17:06.1, 0:00:14.7, 1:30:13.7, 0:00:19.4, 0:33:24.5}, {philippe fattori, 0:17:08.7, 0:00:19.6, 1:30:14.4, 0:00:20.9, 0:34:05.1}, {alessandro bottoni, 0:17:07.7, 0:00:20.2, 1:30:14.3, 0:00:21.9, 0:34:07.9}, {miles stewart, 0:17:03.8, 0:00:14.7, 1:30:12.8, 0:00:21.2, 0:34:34.2}, {chris mccormack, 0:16:59.8, 0:00:15.3, 1:30:13.9, 0:00:20.3, 0:35:02.2}, {lach vollmerhaus, 0:17:04.6, 0:00:18.2, 1:28:36.5, 0:00:24.7, 0:36:43.9}, {tony deboom, 0:17:03.6, 0:00:18.3, 1:30:16.4, 0:00:27.7, 0:35:30.1}, {shane reed, 0:16:59.1, 0:00:15.6, 1:30:13.3, 0:00:22.0, 0:35:35.3}, {takumi obara, 0:17:02.0, 0:00:16.9, 1:30:14.9, 0:00:21.9, 0:35:34.8}, {joshua dapice, 0:17:06.4, 0:00:17.2, 1:30:17.5, 0:00:23.0, 0:35:32.4}, {abe rogers, 0:17:07.3, 0:00:16.6, 1:30:16.6, 0:00:24.9, 0:36:42.4}, {wes hobson, 0:17:05.3, 0:00:20.2, 1:30:15.7, 0:00:23.9, 0:37:18.1}, {clemente alonso, 0:17:07.2, 0:00:18.7, 1:30:17.7, 0:00:30.2, 0:38:00.9}, {kevin carter, 0:17:07.4, 0:00:17.3, 1:31:20.9, 0:00:27.4, 0:37:06.3}, {brian rhodes, 0:17:02.7, 0:00:19.7, 1:30:16.8, 0:00:26.3, 0:38:33.1}, {kris gemmell, 0:17:03.1, 0:00:19.8, 1:30:16.9, 0:00:24.3, 0:38:57.9}, {andy kelsey, 0:17:01.9, 0:00:16.0, 1:31:23.3, 0:00:32.0, 0:38:07.8}, {joe umphenour, 0:17:04.9, 0:00:19.3, 1:33:42.3, 0:00:24.7, 0:36:37.6}, {nathan richmond, 0:17:03.5, 0:00:16.3, 1:30:18.7, 0:00:26.9, 0:41:03.7}, {stephen sheldrak, 0:16:56.6, 0:00:15.6, 1:30:15.5, 0:00:24.2, 0:42:17.1}, {mike garcia, 0:17:11.9, 0:00:20.4, 1:33:43.2, 0:00:25.6, 0:42:57.2}, {troy mckinna, 0:18:00.5, 0:00:17.6, 1:35:22.8, 0:00:25.2, 0:42:45.2}, {peter hursty, 0:17:08.0, 0:00:17.3, 1:33:48.8, 0:00:37.4, 0:55:45.0}, {jean sebastian leuba, 0:17:22.7, 0:00:16.4, 1:31:22.2, 0:00:33.6, dnf}, {jose luiz zepeda, 0:17:02.6, 0:00:19.1, dnf, dnf, dnf}, {nick radkewich, 0:17:03.9, 0:00:19.5, dnf, dnf, dnf}, {simon whitfield, 0:17:06.1, 0:00:18.5, dnf, dnf, dnf}, {gilberto gonzalez, 0:17:07.1, 0:00:18.1, dnf, dnf, dnf}, {anthony parish, 0:17:06.0, 0:00:20.1, dnf, dnf, dnf}, {hunter kemper, 0:17:06.8, 0:00:20.2, dnf, dnf, dnf}, {preben jeppensen, 0:17:11.1, 0:00:17.3, dnf, dnf, dnf}, {carlo catulini, 0:17:49.5, 0:00:17.8, dnf, dnf, dnf}, {fabrizio ferraresi, 0:17:54.4, 0:00:17.3, dnf, dnf, dnf}, {freddy tibocha, 0:17:51.6, 0:00:20.6, dnf, dnf, dnf}, {daniel lee chi wo, 0:17:54.8, 0:00:20.1, dnf, dnf, dnf}, {brent imonen, 0:18:12.1, 0:00:16.6, dnf, dnf, dnf}, {camilo gonzalez, 0:18:20.3, 0:00:18.2, dnf, dnf, dnf}, {matias opitz, 0:19:05.9, 0:00:23.7, dnf, dnf, dnf}, {allan villaneuva, 0:19;42.6, 0:00:19.2, dnf, dnf, dnf}, {johnathan herring, 0:19:39.7, 0:00:32.7, dnf, dnf, dnf}, {fernando gomez, 00:20:07.7, 0:00:18.2, dnf, dnf, dnf}, {rory mackie, 0:20:14.7, 0:00:17.4, dnf, dnf, dnf}, {jeff sanders, 0:20:08.8, 0:00:29.2, dnf, dnf, dnf}, {louis delatorre, 0:20:59.8, 0:00:25.2, dnf, dnf, dnf}}

Lausanne, Switzerland Has splits for swim, transition 1, cycle, transition 2 and run

{lausanne switzerland {andrew johns, 0:19:10, 0:01:21, 1:04:34, 0:00:44, 0:31:33}, {dimitry gaag, 0:19:28, 0:01:25, 1:04:13, 0:00:45, 0:31:39}, {jan rehula, 0:19:13, 0:01:28, 1:04:26, 0:00:45, 0:31:41}, {markus keller, 0:19:21, 0:01:27, 1:04:18, 0:00:43, 0:32:05}, {johannes enzenhofer, 0:19:16, 0:01:24, 1:04:30, 0:00:48, 0:32:03}, {olivier marceau, 0:19:29, 0:01:24, 1:04:15, 0:00:45, 0:32:26}, {reto hug, 0:19:24, 0:01:20, 1:04:21, 0:00:39, 0:32:37}, {craig walton, 0:18:31, 0:01:24, 1:04:20, 0:00:48, 0:33:22}, {stephane bignet, 0:19:13, 0:01:30, 1:04:25, 0:00:49, 0:32:30}, {jean christoph guinchard, 0:19:26, 0:01:21, 0:00:00, 0:00:00, 1:58:45}, {stefan vuckovic, 0:19:29, 0:01:20, 1:04:14, 0:00:41, 0:33:09}, {martin matula, 0:19:18, 0:01:29, 1:04:25, 0:00:50, 0:32:59}, {craig watson, 0:19:33, 0:01:24, 1:05:36, 0:00:48, 0:31:46}, {franck bignet, 0:19:13, 0:01:26, 1:04:30, 0:00:48, 0:33:44}, {gilberto gonzalez, 0:19:20, 0:01:26, 1:04:26, 0:00:47, 0:34:01}, {ben sanson, 0:18:34, 0:01:28, 1:05:03, 0:00:47, 0:34:11}, {jonas djurback, 0:19:23, 0:01:24, 1:04:23, 0:00:48, 0:34:08}, {dennis looze, 0:18:55, 0:01:27, 1:04:47, 0:00:48, 0:34:19}, {gianpietr de faveri, 0:19:15, 0:01:26, 1:04:30, 0:00:51, 0:34:17}, {stuart hayes, 0:19:14, 0:01:24, 1:04:31, 0:00:48, 0:34:32}, {eneko llanos, 0:19:52, 0:01:25, 1:05:14, 0:00:48, 0:33:40}, {nathan richmond, 0:18:42, 0:01:28, 1:05:02, 0:00:52, 0:35:24}, {peter alder, 0:19:20, 0:01:25, 1:04:19, 0:00:44, 0:35:58}, {richard stannard, 0:18:32, 0:01:24, 1:05:09, 0:00:49, 0:35:57}, {fabrizio ferraresi, 0:20:05, 0:01:23, 1:06:32, 0:00:52, 0:33:16}, {eric van der linden, 0:19:07, 0:01:24, 1:04:36, 0:00:44, 0:36:22}, {dirk bockel, 0:19:18, 0:01:27, 1:04:18, 0:00:47, 0:36:29}, {rob barel, 0:19:37, 0:01:22, 1:06:50, 0:00:38, 0:33:57}, {jan hansen, 0:19:42, 0:01:27, 1:06:51, 0:00:45, 0:33:42}, {csaba kuttor, 0:19:17, 0:01:25, 1:07:19, 0:00:50, 0:33:59}, {xavier llobet, 0:19:55, 0:01:25, 1:06:28, 0:00:52, 0:34:20}, {darren carnell, 0:19:33, 0:01:23, 1:07:05, 0:00:50, 0:34:15}, {martin biehler, 0:20:30, 0:01:28, 1:07:39, 0:00:46, 0:33:24}, {doug firman, 0:19:12, 0:01:24, 1:07:24, 0:00:48, 0:35:08}, {marcel vifian, 0:20:16, 0:01:29, 1:07:49, 0:00:51, 0:33:39}, {greg von holdt, 0:19:49, 0:01:27, 1:08:23, 0:00:46, 0:34:13}, {kevin carter, 0:19:56, 0:01:20, 1:06:44, 0:00:49, 0:36:00}, {jose merchan, 0:19:47, 0:01:25, 1:08:19, 0:00:58, 0:34:21}, {joe umphenour, 0:19:25, 0:01:25, 1:08:51, 0:00:52, 0:34:22}, {andreas raelert, 0:19:27, 0:01:22, 1:07:10, 0:00:49, 0:36:31}, {dominik norbert, 0:20:55, 0:01:28, 1:08:43, 0:00:49, 0:33:38}, {fedor filipov, 0:19:11, 0:01:28, 1:09:01, 0:00:45, 0:35:09}, {alessandro bottoni, 0:19:34, 0:01:25, 1:08:39, 0:00:52, 0:35:09}, {petteri kosonen, 0:20:26, 0:01:29, 1:07:44, 0:00:51, 0:35:28}, {roland melis, 0:20:51, 0:01:28, 1:08:46, 0:00:50, 0:34:08}, {robin simpson, 0:19:58, 0:01:26, 1:08:15, 0:00:48, 0:36:01}, {luc huntjens, 0:19:09, 0:01:26, 1:09:02, 0:00:51, 0:37:45}, {stephen sheldrake, 0:19:11, 0:01:32, 1:07:55, 0:00:56, 0:40:26}, {ryan bolton, 0:22:40, 0:01:24, 1:12:59, 0:00:49, 0:34:07}, {mark marabini, 0:23:05, 0:01:35, 1:16:16, 0:01:08, 0:39:47}, {conrad stoltz, 0:19:58, 0:01:20, 1:05:13, 0:00:47, dnf}, {hector llanos, 0:19:28, 0:01:24, 1:07:33, dnf, dnf}, {hamish cater, 0:19:09, 0:01:24, dnf, dnf, dnf}, {ricky jorgensen, 0:19:05, 0:01:30, dnf, dnf, dnf}, {anthony parish, 0:19:21, 0:01:20, dnf, dnf, dnf}, {tim don, 0:19:25, 0:01:19, dnf, dnf, dnf}, {marc jenkins, 0:19:22, 0:01:24, dnf, dnf, dnf}, {arnd schomburg, 0:19:18, 0:01:31, dnf, dnf, dnf}, {greg bennett, 0:19:26, 0:01:26, dnf, dnf, dnf}, {stefano belandi, 0:19:30, 0:01:23, dnf, dnf, dnf}, {ralf eggert, 0:19:54, 0:01:24, dnf, dnf, dnf}, {clemente alonso, 0:19:53, 0:01:24, dnf, dnf, dnf}, {akos gyorffy, 0:19:53, 0:01:32, dnf, dnf, dnf}, {richard allen, 0:20:17, 0:01:25, dnf, dnf, dnf}, {daniel barrera, 0:20:18, 0:01:27, dnf, dnf, dnf}, {markus forster, 0:21:47, 0:01:25, dnf, dnf, dnf}, {grega hocvar, 0:21:40, 0:01:34, dnf, dnf, dnf}, {bruno arochi, 0:22:41, 0:01:24, dnf, dnf, dnf}}

Corner Brook, Canada Has splits for swim, transition 1, cycle, transition 2 and run note 3 lap swim, 4 lap cycle, 4 lap run

{corner brook canada {craig walton, 0:17:18.6, 0:01:58.3, 1:01:13.1, 0:00:30.0, 0:33:11.8}, {laurent jeanselme, 0:17:18.6, 0:02:01.9, 1:01:10.1, 0:00:31.2, 0:33:41.3}, {chris mccormack, 0:18:33.6, 0:02:00.4, 1:02:53.6, 0:00:30.1, 0:31:43.2}, {shane reed, 0:17:54.5, 0:01:59.5, 1:03:35.9, 0:00:30.1, 0:31:42.1}, {hunter kemper, 0:18:23.9, 0:02:04.5, 1:03:02.0, 0:00:33.6, 0:31:57.3}, {gilberto gonzalez, 0:18:27.9, 0:02:07.6, 1:02:56.8, 0:00:28.9, 0:32:02.9}, {eric van der linden, 0:18:11.0, 0:02:11.2, 1:03:06.9, 0:00:30.5, 0:32:49.4}, {chris hill, 0:17:51.5, 0:02:06.1, 1:03:28.2, 0:00:28.2, 0:33:01.6}, {eneko llanos, 0:18:26.3, 0:02:04.1, 1:02:59.3, 0:00:36.1, 0:33:15.7}, {joshua dapice, 0:18:38.1, 0:01:56.3, 1:02:56.1, 0:00:30.0, 0:33:25.1}, {hector llanos, 0:18:21.2, 0:02:07.2, 1:02:59.9, 0:00:46.4, 0:33:17.0}, {tim don, 0:18:19.3, 0:02:03.4, 1:03:08.8, 0:00:29.6, 0:33:39.8}, {olivier marceau, 0:17:49.6, 0:01:58.5, 1:03:43.8, 0:00:34.4, 0:33:52.5}, {gianpietro de faveri, 0:17:41.8, 0:02:05.6, 1:03:45.6, 0:00:41.2, 0:34:01.9}, {dennis looze, 0:18:13.1, 0:02:01.8, 1:03:15.7, 0:00:34.8, 0:34:22.2}, {richard stannard, 0:17:42.8, 0:02:06.8, 1:03:39.0, 0:00:32.2, 0:34:32.6}, {rob barel, 0:19:06.1, 0:02:00.9, 1:04:22.6, 0:00:29.2, 0:33:08.5}, {nick radkewich, 0:18:46.2, 0:01:56.3, 1:04:49.4, 0:00:33.5, 0:33:04.6}, {tony deboom, 0:18:56.7, 0:02:00.4, 1:04:35.3, 0:00:35.0, 0:33:21.1}, {vassilis krommidas, 0:18:53.5, 0:02:10.4, 1:04:28.1, 0:00:40.0, 0:33:30.4}, {greg von holdt, 0:19:02.0, 0:01:58.2, 1:04:31.3, 0:00:35.1, 0:34:01.8}, {petteri kosonen, 0:19:17.2, 0:02:10.1, 1:04:54.2, 0:00:35.3, 0:34:38.1}, {kevin carter, 0:18:41.2, 0:01:57.5, 1:04:51.4, 0:00:35.1, 0:35:50.6}, {lach vollmerhaus, 0:19:01.6, 0:02:02.6, 1:04:26.5, 0:00:33.9, 0:36:00.5}, {clemente alonso, 0:18:50.8, 0:02:06.9, 1:04:37.4, 0:00:38.3, 0:36:00.3}, {stefan timms, 0:19:13.9, 0:02:11.1, 1:06:25.5, 0:00:34.6, 0:34:07.6}, {jose rodriguez, 0:19:14.5, 0:02:16.4, 1:06:20.7, 0:00:35.5, 0:36:40.2}, {simon whitfield, 0:19:19.0, 0:02:03.9, 1:06:36.4, 0:00:28.4, 0:37:44.3}, {sebaten laflamme, 0:19:16.2, 0:02:05.4, 1:06:28.3, 0:00:32.9, 0:37:58.2}, {julian dorais, 0:19:04.1, 0:02:07.5, 1:06:39.2}, 0:00:35.8, 0:40:12.6}, {jocelyn gascon-giroux, 0:19:58.4, 0:02:20.1, 1:12:20.3, 0:00:41.7, 0:36:55.7}, {samuel pierreclaud, 0:18:26.9, 0:02:04.6, 1:05:05.0, 0:00:40.2, dnf}, {richard allen, 0:19:00.7, 0:02:05.1, dnf, dnf, dnf}, (jamie hunt, 0:19:02.3, 0:01:53.5, 1:05:28.8, 0:00:30.4, dnf}, {jose barbany, 0:19:21.8, 0:02:11.6, 1:11:30.1, dnf, dnf}, {franck bignet, 0:17:47.2, 0:02:06.6, dnf, dnf, dnf}, {marc jenkins, 0:18:30.9, 0:02:02.6, dnf, dnf, dnf}, {jean sebastien rioux, 0:19:15.7, 0:02:22.3, dnf, dnf, dnf}, {toby colombe, 0:23:13.6, 0:02:12.8, dnf, dnf, dnf}, {luc huntjens, 0:18:51.3, dnf, dnf, dnf, dnf}}

Tiszaujvaros, Hungary Has splits for swim, trans 1, cycle, trans 2, run note 2 lap swim, 8 lap cycle, 4 lap run

{tiszaujvaros hungary {hamish carter, 0:17:24.6, 0:00:40.2, 0:54:22.9, 0:00:32.2, 0:31:50.9}, {shane reed, 0:17:23.1, 0:00:35.5, 0:54:29.9, 0:00:33.7, 0:31:58.0}, {vladamir polikarpenko, 0:17:23.8, 0:00:38.1, 0:54:29.1, 0:00:34.5, 0:32:15.4}, {reto hug, 0:17:30.6, 0:00:36.5, 0:54:20.8, 0:00:27.8, 0:32:39.4}, {spencer smith, 0:17:23.7, 0:00:35.8, 0:54:29.2, 0:00:30.2, 0:32:42.5}, {matthew reed, 0:17:27.2, 0:00:37.5, 0:54:24.6, 0:00:34.4, 0:32:46.4}, {dimitri gaag, 0:17:53.8, 0:00:36.3, 0:55:43.2, 0:00:28.8, 0:31:14.7}, {lother leder, 0:18:03.1, 0:00:36.5, 0:55:35.4, 0:00:31.3, 0:31:10.9}, {laurent jeanselme, 0:17:14.0, 0:00:40.8, 0:54:36.0, 0:00:33.3, 0:33:01.6}, {greg welch, 0:17:55.1, 0:00:39.8, 0:55:39.2, 0:00:36.0, 0:31:37.8}, {greg bennett, 0:17:55.8, 0:00:37.7, 0:55:40.4, 0:00:32.9, 0:31:58.2}, {sebastien berlier, 0:17:57.7, 0:00:39.3, 0:55:38.3, 0:00:34.5, 0:31:59.2}, {jan rehula, 0:18:02.4, 0:00:42.2, 0:55:30.1, 0:00:34.7, 0:32:02.5}, {craig watson, 0:17:50.9, 0:00:41.2, 0:55:45.2, 0:00:35.5, 0:32:02.6}, {andriy glushchenko, 0:17:50.0, 0:00:35.9, 0:55:50.4, 0:00:38.2, 0:32:20.9}, {conrad stoltz, 0:17:56.5, 0:00:39.5, 0:55:38.5, 0:00:30.7, 0:32:41.4}, {luke harrop, 0:17:27.9, 0:00:36.1, 0:54:25.1, 0:00:36.8, 0:34:25.7}, {joachim willen, 0:17:26.5, 0:00:39.8, 0:54:25.3, 0:00:33.6, 0:34:34.5}, {simon lessing, 0:18:01.7, 0:00:40.0, 0:55:35.2, 0:00:37.1, 0:32:56.6}, {csaba kuttor, 0:17:30.7, 0:00:38.5, 0:56:11.2, 0:00:39.5, 0:32:54.6}, {johannes enzenhofer, 0:17:57.4, 0:00:40.5, 0:55:38.1, 0:00:37.2, 0:33:06.0}, {richard stannard, 0:17:07.9, 0:00:40.8, 0:54:39.5, 0:00:32.9, 0:35:03.1}, {ralf eggert, 0:18:07.5, 0:00:36.1, 0:55:28.7, 0:00:31.9, 0:33:21.3}, {dennis looze, 0:17:49.4, 0:00:38.6, 0:55:52.0, 0:00:35.0, 0:33:15.3}, {andrew johns, 0:17:57.8, 0:00:37.2, 0:55:40.8, 0:00:36.5, 0:33:24.1}, {fedor filipov, 0:17:59.7, 0:00:40.6, 0:55:38.1, 0:00:37.2, 0:33:23.0}, {andreas raelert, 0:17:25.5, 0:00:35.2, 0:54:29.3, 0:00:32.1, 0:35:30.7}, {samuel piereclaud, 0:17:30.8, 0:00:40.9, 0:56:08.9, 0:00:41.0, 0:33:41.3}, {richard allen, 0:18:00.1, 0:00:38.8, 0:55:36.7, 0:00:35.9, 0:34:17.6}, {jonas djurback, 0:17:52.0, 0:00:40.8, 0:55:44.8, 0:00:35.5, 0:34:26.1}, {carl blasco, 0:17:36.1, 0:00:44.3, 0:55:57.9, 0:00:37.6, 0:34:40.3}, {david haines, 0:17:51.2, 0:00:40.0, 0:55:47.8, 0:00:40.6, 0:34:45.8}, {chris mccormack, 0:17:50.8, 0:00:35.9, 0:55:46.2, 0:00:35.9, 0:35:00.5}, {ben sanson, 0:17:07.4, 0:00:38.4, 0:54:40.6, 0:00:44.2, 0:36:39.5}, {garrett mccarthy, 0:17:34.3, 0:00:42.3, 0:56:04.6, 0:00:49.6, 0:34:56.0}, {rob barel, 0:18:27.4, 0:00:37.7, 0:57:47.9, 0:00:27.0, 0:33:02.5}, {peter alder, 0:18:04.2, 0:00:40.1, 0:55:38.1, 0:00:36.2, 0:35:29.7}, {martin krnavek, 0:17:53.0, 0:00:37.2, 0:55:49.2, 0:00:32.0, 0:35:47.1}, {jan hansen, 0:18:35.8, 0:00:36.1, 0:57:42.1, 0:00:34.7, 0:33:21.9}, {andreas heczey, 0:18:03.5, 0:00:39.1, 0:55:38.3, 0:00:42.6, 0:36:01.8}, {peter hobor, 0:18:33.2, 0:00:40.9, 0:57:40.2, 0:00:41.9, 0:34:18.1}, {luc huntjens, 0:18:02.9, 0:00:40.6, 0:55:36.6, 0:00:38.1, 0:37:03.4}, {peter tiernam, 0:18:34.1, 0:00:39.0, 0:57:43.0, 0:00:38.7, 0:34:41.0}, {gabor karpati, 0:18:05.7, 0:00:40.0, 0:58:16.1, 0:00:35.5, 0:34:59.6}, {nathan richmond, 0:17:31.4, 0:00:37.1, 0:56:15.0, 0:00:37.4, 0:37:54.1}, {markus forster, 0:18:39.6, 0:00:36.6, 0:57:37.2, 0:00:35.4, 0:35:56.9}, {ede rutkovszky, 0:18:15.3, 0:00:41.7, 0:57:58.5, 0:00:34.5, 0:36:10.4}, {carlos gil, 0:18:01.2, 0:00:43.5, 0:58:15.9, 0:00:43.2, 0:36:20.7}, {szabolcs agoston, 0:17:35.2, 0:00:37.6, 0:58:00.5, 0:00:37.6, 0:37:51.9}, {ulo ojava, 0:18:32.3, 0:00:35.4, 0:57:52.3, 0:00:37.0, 0:38:59.2}, {brad bevan, 0:17:18.7, 0:00:36.9, 0:54:29.7, 0:00:34.1, dnf}, {simon whitfield, 0:18:02.1, 0:00:37.2, 0:55:42.4, 0:00:39.5, dnf}}

Noosa, Australia Has splits for swim, cycle and run time

{Noosa, australia {shane reed, 0:18:07, 0:52:47, 0:32:44}, {andrew johns, 0:18:13, 0:52:40, 0:33:28}, {craig walton, 0:17:54, 0:53:01, 0:33:30}, {jamie hunt, 0:19:04, 0:54:37, 0:31:12}, {gilberto gonzalez, 0:18:39, 0:55:10, 0:31:17}, {paul amey, 0:18:19, 0:55:29, 0:31:19}, {ben bright, 0:18:06, 0:52:52, 0:34:29}, {Juracy moreira, 0:19:07, 0:54:45, 0:31:47}, {simon whitfeild, 0:19:05, 0:54:38, 0:32:03}, {greg bennett, 0:18:27, 0:55:13, 0:32:07}, {dimitry gaag, 0:18:31, 0:55:13, 0:32:07}, {filip osplay, 0:18:35, 0:55:13, 0:32:09}, {Andriy glushchenko, 0:18:46, 0:55:07, 0:32:20}, {tim don, 0:19:08, 0:54:42, 0:32:30}, {kris gemmell, 0:18:33, 0:55:06, 0:32:55}, {ben sanson, 0:18:05, 0:52:56, 0:35:41}, {troy mckinna, 0:18:54, 0:54:43, 0:33:36}, {joe umphenour, 0:18:53, 0:54:57, 0:33:33}, {luke harrop, 0:18:38, 0:55:16, 0:33:32}, {victor plata, 0:18:37, 0:55:18, 0:33:45}, {johannes enzenhofer, 0:18:47, 0:55:04, 0:33:56}, {hideo fukui, 0:18:32, 0:55:15, 0:34:08}, {vassilis krommidas, 0:18:59, 0:55:01, 0:34:08}, {courtney atkinson, 0:18:17, 0:55:39, 0:34:23}, {takumi obara, 0:18:30, 0:55:31, 0:34:19}, {andy kelsey, 0:18:50, 0:55:10, 0:34:22}, {jose merchan, 0:18:34, 0:00:00, 1:48:37}, {nick Radkewich, 0:18:44, 0:55:11, 0:35:03}, {anthony parish, 0:19:10, 0:54:47, 0:36:25}, {jun-ichi tamamoto, 0:19:31, 0:58:07, 0:32:58}, {joshua dapice, 0:18:52, 0:55:07, 0:36:50}, {leandro macedo, 0:19:30, 0:58:07, 0:33:33}, {norbert domnik, 0:19:16, 0:57:46, 0:34:13}, {ivan rana, 0:19:15, 0:54:31, 0:37:55}, {rasmus henning, 0:19:14, 0:57:51, 0:34:43}, {craig alexander, 0:19:04, 0:54:54, 0:38:24}, {ricky jorgensen, 0:18:02, 0:55:51, 0:38:39}, {hamish carter, 0:19:02, 0:54:41, 0:38:55}, {stean timms, 0:19:12, 0:58:31, 0:35:25}, {hideo nakagome, 0:20:00, 0:57:57, 0:35:25}}

Appendix I Study 2: Group sizes

Table I.1 Female group sizes with a 4 s separation distance at the end of the swim, cycle and run for each of the investigated ITU triathlons.

Table I.2 Male group sizes with a 4 s separation distance at the end of the swim, cycle and run for each of the investigated ITU triathlons.

Note: S = swim C = cycle R = run

Female event 1 2 3 4 5 6 7 8 9 10 group s c r sc r sc r sc r sc r sc r sc r sc r sc r sc r 1 211 411 2131 111 431 221 851 471 661 112 2 1 5 1 2 20 1 3 1 2 1 19 1 1 17 1 9 1 1 1 23 1 8 11 1 1 29 1 1 21 2 3 7 15 1 1 1 1 1 1 1 16 19 1 4 16 1 1 17 2 1 5 1 6 2 1 6 12 2 1 1 1 4 331 1111 111 4 1 111 911 1431 2 1 2621 18101 5 242 211 161 7 2 211 1114 511 1 131 161 6 32211113121131125121 141111 7 411418112149189222 142111 8 811 111161171137111 11113 9 121 111 1251111131 13131 10 721 111 5611111 1 1 111 11 11 1 1 1 1 111 211 1 1 1 1 1 12 1215111111621111 13 1 111 3321131 1 1 1 1 14 1 1 1 1 2 4 1 1 1 1 1 3 2 15 121112111111 16 1 1 1 1 1 1 2 1 1 1 2 17 1 1 1 1 1 1 1 1 1 1 1 18 1 1 1 1 1 2 1 1 1 1 1 19 1111611111 20 1111131111 21 1111111 11 22 1111111 21 23 1111111 21 24 1111111 11 25 1 1 1 1 1 1 2 1 26 1 1 1 2 1 1 1 1 27 1 1 2 1 1 2 1 28 1 1 2 1 1 2 1 29 1 1 1 3 1 1 1 30 1 1 1 2 1 4 1 31 1 1 2 1 1 1 1 32 1 1 1 1 1 1 1 33 1 11 11 34 1 11 11 35 1 11 1 36 1 11 1 37 11 1 38 12 1 39 11 1 40 11 1 41 11 1 42 11 43 11 44 11 45 11 46 11 11 11 1 1 1 1 1 1 1 1 1 1

Male event 1 2 3 4 5 6 7 8 9 group s c r sc r sc r sc r sc r sc r sc r sc r sc r 1 111 111 1191 3111 2311 311 221 2111 151 2 32292 46481 6321 2551 1172 122 2121 111 4281 3 423 1 1 31161 4301 21 1201 322 15271 121 4 511 1 2 111 111 21 2521 212 2511 221 5 121 8 1423 211 11 412 551 111 912 6 322 1 1721 231 14 112 121 111 711 7 1101 1 1 1 161 1 671 211 532 12 1 8 221 415 21 11111514122 9 21322111013111111 10 1112311214111 11 2422211135131 12 1112111121111 13 25311111111 14 11221111111 15 21321111211 16 1111111122 17 122111111 18 212611121 19 142511111 20 1112 1121 21 2231 2113 22 1121 1111 23 1211 1111 24 1211 1111 25 1 1 2 1 1 1 1 26 1 3 1 1 1 3 1 27 1 2 3 1 1 1 1 28 1 1 1 1 1 1 1 29 111211 30 231111 31 113111 32 111211 33 112111 34 111111 35 11111 36 1 1 1 1 37 1 1 1 1 38 1 1 1 1 39 1 1 1 1 40 1 1 1 1 41 1 1 1 42 1 1 1 43 1 1 1 44 1 1 45 1 1 46 1 1

Appendix J. Study 3: Raw Data No. name sex cat h'cap swim swim swim bike swim + cycle swim + run finish final otal elapsed total cadence cadence SR SR SR time elapsed position time cycle time elapesd cycle time time position time position c1 c2 r1 r2 r3 (h:m:s) (h:m:s) position (h:m:s) (h:m:s) (rpm) (rpm) (rpm) (rpm) (rpm) 7 co 1 2 0:05:00 0:18:03 0:23:03 18 1:03:45 1:21:48 1:26:48 1 0:36:23 1:58:11 1 2:03:11 1 90.7 112.5 90.6 88.7 88.9 36 cn 1 2 0:05:00 0:18:01 0:23:01 16 1:04:45 1:22:46 1:27:46 2 0:38:44 2:01:30 3 2:06:30 2 93.7 122 88.5 84.7 84.4 74 pg 1 2 0:05:00 0:19:56 0:24:56 38 1:05:31 1:25:27 1:30:27 9 0:36:55 2:02:22 5 2:07:22 3 105 127 92.3 80.6 84.4 35 jm 1 2 0:05:00 0:19:52 0:24:52 37 1:02:59 1:22:51 1:27:51 3 0:40:16 2:03:07 7 2:08:07 4 97.7 122 96.2 91.1 90.7 24 hb 1 2 0:05:00 0:20:45 0:25:45 50 1:04:47 1:25:32 1:30:32 11 0:37:37 2:03:09 8 2:08:09 5 95.4 123 87.5 88.5 97 89 dw 1 2 0:05:00 0:20:47 0:25:47 51 1:04:32 1:25:19 1:30:19 4 0:38:31 2:03:50 9 2:08:50 6 94 112.5 93.4 87.1 94.3 29 bk 1 2 0:05:00 0:18:15 0:23:15 20 1:07:11 1:25:26 1:30:26 7 0:38:32 2:03:58 10 2:08:58 7 89.4 130 89.4 88.5 108 39 wp 1 2 0:05:00 0:20:47 0:25:47 52 1:05:43 1:26:30 1:31:30 14 0:37:31 2:04:01 11 2:09:01 8 96 123 95 84.9 90.1 114 mf 1 3 0:10:00 0:20:47 0:30:47 72 1:02:45 1:23:32 1:33:32 22 0:35:38 1:59:10 2 2:09:10 9 . 132 93.4 91.1 100 80 ah 1 2 0:05:00 0:20:05 0:25:05 41 1:05:18 1:25:23 1:30:23 5 0:39:15 2:04:38 13 2:09:38 10 100 123 96.4 96 103.8 88 tt 1 2 0:05:00 0:20:38 0:25:38 47 1:05:36 1:26:14 1:31:14 13 0:38:30 2:04:44 14 2:09:44 11 100 130 92.3 86.8 93.1 83 cm 1 2 0:05:00 0:21:44 0:26:44 57 1:04:28 1:26:12 1:31:12 12 0:39:07 2:05:19 15 2:10:19 12 98 127 89.1 91.1 87.1 37 jn 1 2 0:05:00 0:18:25 0:23:25 25 1:07:02 1:25:27 1:30:27 8 0:40:00 2:05:27 16 2:10:27 13 94.3 122 93.8 85.1 86.8 75 dg 1 2 0:05:00 0:18:26 0:23:26 26 1:09:07 1:27:33 1:32:33 18 0:39:21 2:06:54 18 2:11:54 14 82.4 123 96.6 90 91.7 116 ah 1 3 0:10:00 0:19:30 0:29:30 67 1:05:08 1:24:38 1:34:38 28 0:37:30 2:02:08 4 2:12:08 15 . 118 . . 94.3 5 sb 1 2 0:05:00 0:19:50 0:24:50 36 1:06:57 1:26:47 1:31:47 15 0:40:48 2:07:35 19 2:12:35 16 92.3 123 84.4 85.7 84.4 18 mk 2 1 0:00:00 0:22:08 0:22:08 10 1:09:47 1:31:55 1:31:55 16 0:40:58 2:12:43 40 2:12:43 17 88.5 123 92.7 91.3 92 166 tr 1 3 0:10:00 0:20:43 0:30:43 71 1:02:46 1:23:29 1:33:29 21 0:39:21 2:02:50 6 2:12:50 18 . 127 84.2 85.7 96 76 jg 1 2 0:05:00 0:21:42 0:26:42 56 1:07:13 1:28:55 1:33:55 24 0:38:57 2:07:52 20 2:12:52 19 109 117 96.6 97.3 90 27 ti 1 2 0:05:00 0:20:04 0:25:04 40 1:08:17 1:28:21 1:33:21 20 0:40:01 2:08:22 23 2:13:22 20 88.7 125 88.2 94.2 . 28 bk 1 2 0:05:00 0:20:29 0:25:29 43 1:05:01 1:25:30 1:30:30 10 0:43:34 2:09:04 25 2:14:04 21 100 118 85.7 83.7 . 129 ap 1 2 0:10:00 0:20:42 0:30:42 70 1:04:45 1:25:27 1:35:27 38 0:38:43 2:04:10 12 2:14:10 22 . . . . . 26 jh 1 2 0:05:00 0:20:32 0:25:32 45 1:09:49 1:30:21 1:35:21 36 0:39:30 2:09:51 28 2:14:51 23 . . . . . 85 ws 1 2 0:05:00 0:20:41 0:25:41 48 1:09:43 1:30:24 1:35:24 37 0:43:06 2:09:54 43 2:14:54 24 . . . . . 40 bs 1 2 0:05:00 0:22:16 0:27:16 62 1:08:22 1:30:38 1:35:38 39 0:39:28 2:10:06 30 2:15:06 25 . . . . . 90 lw 1 2 0:05:00 0:19:43 0:24:43 35 1:05:41 1:25:24 1:30:24 6 0:44:53 2:10:17 31 2:15:17 26 97.9 130 84.7 90.6 . 73 bf 1 2 0:05:00 0:22:08 0:27:08 60 1:08:09 1:30:17 1:35:17 34 0:40:17 2:10:34 32 2:15:34 27 . . . . . 72 sd 1 2 0:05:00 0:22:38 0:27:38 64 1:08:02 1:30:40 1:35:40 40 0:40:02 2:10:42 33 2:15:42 28 . . . 83.3 . 33 dm 1 2 0:05:00 0:21:58 0:26:58 58 x 2:10:46 34 2:15:46 29 . . . 95 . 38 so 1 2 0:05:00 0:20:43 0:25:43 49 1:08:03 1:28:46 1:33:46 23 0:42:09 2:10:55 35 2:15:55 30 . 127 96.6 96 . 123 cm 1 3 0:10:00 0:21:10 0:31:10 79 1:07:02 1:28:12 1:38:12 53 0:37:45 2:05:57 17 2:15:57 31 . . . . . 42 bw 1 2 0:05:00 0:19:19 0:24:19 34 1:08:03 1:27:22 1:32:22 17 0:44:04 2:11:26 38 2:16:26 32 94.1 123 81.8 . . 17 ek 2 1 0:00:00 0:23:31 0:23:31 30 1:11:30 1:35:01 1:35:01 32 0:42:07 2:17:08 55 2:17:08 33 90.8 89 90.9 100 90 11 bc 2 1 0:00:00 0:23:29 0:23:29 28 1:10:52 1:34:21 1:34:21 27 0:43:39 2:18:00 59 2:18:00 34 91.3 96.4 98.3 96.7 90.7 160 th 1 3 0:10:00 0:23:07 0:33:07 88 1:06:48 1:29:55 1:39:55 67 0:38:23 2:08:18 21 2:18:18 35 . . . . . 146 cb 1 3 0:10:00 0:20:47 0:30:47 73 1:08:11 1:28:58 1:38:58 58 0:39:23 2:08:21 22 2:18:21 36 . . . . . No. name sex cat h'cap swim swim swim bike swim + cycle swim + run finish final otal elapsed total cadence cadence SR SR SR time elapsed position time cycle time elapesd cycle time time position time position c1 c2 r1 r2 r3 (h:m:s) (h:m:s) position (h:m:s) (h:m:s) (rpm) (rpm) (rpm) (rpm) (rpm) 61 mv 2 1 0:00:00 0:23:05 0:23:05 19 1:12:13 1:35:18 1:35:18 35 0:43:14 2:18:32 62 2:18:32 37 91.3 105 88.2 100 86.4 8 ap 1 2 0:05:00 x #VALUE! x 1:29:20 1:34:20 26 0:44:17 2:13:37 44 2:18:37 38 . 104 . . . 1 lb 2 1 0:00:00 0:20:45 0:20:45 2 1:12:13 1:32:58 1:32:58 19 0:45:55 2:18:53 64 2:18:53 39 91.5 123 96.7 97 84.6 156 rg 1 3 0:10:00 0:20:05 0:30:05 68 1:06:45 1:26:50 1:36:50 45 0:42:08 2:08:58 24 2:18:58 40 . . . . . 117 gh 1 3 0:10:00 0:20:47 0:30:47 74 1:08:20 1:29:07 1:39:07 60 0:40:12 2:09:19 26 2:19:19 41 . . . . . 113 gd 1 3 0:10:00 0:20:38 0:30:38 69 1:08:42 1:29:20 1:39:20 61 0:44:08 2:09:32 42 2:19:32 42 . . . . . 108 ib 1 3 0:10:00 0:22:24 0:32:24 85 x 2:09:49 27 2:19:49 43 . . . . . 6 km 1 2 0:05:00 0:20:30 0:25:30 44 1:09:28 1:29:58 1:34:58 31 0:44:59 2:14:57 48 2:19:57 44 95 . . . . 148 db 1 3 0:10:00 0:22:21 0:32:21 84 1:07:08 1:29:29 1:39:29 64 0:40:32 2:10:01 29 2:20:01 45 . . . . . 41 pt 1 2 0:05:00 0:21:14 0:26:14 54 1:10:48 1:32:02 1:37:02 47 0:43:41 2:15:43 50 2:20:43 46 . . . . . 32 lm 1 2 0:05:00 0:22:10 0:27:10 61 1:09:53 1:32:03 1:37:03 48 0:39:47 2:15:44 39 2:20:44 47 . . . . . 145 pb 1 3 0:10:00 0:21:05 0:31:05 78 1:08:20 1:29:25 1:39:25 62 0:41:32 2:10:57 36 2:20:57 48 . . . . . 22 gw 2 1 0:00:00 0:23:32 0:23:32 31 1:11:31 1:35:03 1:35:03 33 0:45:56 2:20:59 71 2:20:59 49 91.9 106 94.3 97.5 86.7 97 nl 2 1 0:00:00 0:23:27 0:23:27 27 1:11:18 1:34:45 1:34:45 29 0:46:28 2:21:13 72 2:21:13 50 100 92.3 96.6 96.4 86.4 202 ds 1 3 0:10:00 0:24:42 0:34:42 91 1:06:46 1:31:28 1:41:28 72 0:39:57 2:11:25 37 2:21:25 51 . . . . . 25 bf 1 2 0:05:00 0:22:01 0:27:01 59 1:11:53 1:33:54 1:38:54 57 0:42:39 2:16:33 54 2:21:33 52 . . . . . 179 rs 2 1 0:00:00 0:22:10 0:22:10 11 1:14:01 1:36:11 1:36:11 42 0:45:30 2:21:41 75 2:21:41 53 93.3 122 102 104 86.8 269 jc 2 1 0:00:00 0:20:47 0:20:47 3 1:13:24 1:34:11 1:34:11 25 0:47:58 2:22:09 78 2:22:09 54 82.3 100 96.7 97.3 87.1 272 kw 1 2 0:05:00 0:20:35 0:25:35 46 1:11:44 1:32:19 1:37:19 50 0:44:51 2:17:10 57 2:22:10 55 . . . . . 143 lw 2 1 0:00:00 0:25:09 0:25:09 42 1:13:27 1:38:36 1:38:36 55 0:44:07 2:22:43 82 2:22:43 56 . . . . . 19 tm 2 1 0:00:00 0:23:02 0:23:02 17 1:14:08 1:37:10 1:37:10 49 0:45:59 2:23:09 84 2:23:09 57 90 . . . . 46 kb 2 1 0:00:00 0:22:04 0:22:04 7 1:14:08 1:36:12 1:36:12 43 0:47:10 2:23:22 86 2:23:22 58 . 122 . . . 77 lg 1 2 0:05:00 0:26:35 0:31:35 81 1:08:13 1:34:48 1:39:48 66 0:43:43 2:18:31 61 2:23:31 59 . . . . . 122 rl 1 3 0:10:00 0:21:15 0:31:15 80 1:07:49 1:29:04 1:39:04 59 0:44:29 2:13:33 47 2:23:33 60 . . . . . 135 kc 2 1 0:00:00 0:23:38 0:23:38 33 1:14:31 1:38:09 1:38:09 52 0:45:29 2:23:38 88 2:23:38 61 . . . . . 167 rs 1 3 0:10:00 0:27:09 0:37:09 100 x 2:13:43 45 2:23:43 62 . . . . . 15 cc 2 1 0:00:00 0:23:21 0:23:21 23 1:17:42 1:41:03 1:41:03 70 0:43:02 2:24:05 90 2:24:05 63 91.8 . . . . 165 sp 1 3 0:10:00 0:21:02 0:31:02 77 1:10:37 1:31:39 1:41:39 74 0:42:41 2:14:20 46 2:24:20 64 . . . . . 84 ar 1 2 0:05:00 0:22:36 0:27:36 63 1:11:13 1:33:49 1:38:49 56 0:45:42 2:19:31 66 2:24:31 65 . . . . . 4 csc 2 1 0:00:00 0:22:07 0:22:07 9 1:17:20 1:39:27 1:39:27 63 0:45:17 2:24:44 93 2:24:44 66 80 . . . . 20 km 2 1 0:00:00 0:25:03 0:25:03 39 1:15:43 1:40:46 1:40:46 69 0:44:41 2:25:27 96 2:25:27 67 . . . . . 163 gl 1 3 0:10:00 0:23:18 0:33:18 89 1:10:48 1:34:06 1:44:06 82 0:41:25 2:15:31 49 2:25:31 68 . . . . . 78 sh 1 2 0:05:00 0:26:47 0:31:47 82 1:12:12 1:38:59 1:43:59 81 0:41:45 2:20:44 69 2:25:44 69 . . . . . 95 jg 2 1 0:00:00 0:25:48 0:25:48 53 1:17:01 1:42:49 1:42:49 77 0:43:01 2:25:50 99 2:25:50 70 . . . . . 86 mt 1 2 0:05:00 0:21:29 0:26:29 55 1:20:22 1:41:51 1:46:51 89 0:39:06 2:20:57 70 2:25:57 71 . . . . . 104 da 1 3 0:10:00 x x 1:36:15 1:46:15 85 0:40:13 2:16:28 53 2:26:28 72 . . . . . 140 kl 2 1 0:00:00 0:22:12 0:22:12 13 1:20:24 1:42:36 1:42:36 76 0:44:27 2:27:03 105 2:27:03 73 79.6 . . . . 235 gc 1 4 0:15:00 0:20:59 0:35:59 98 1:09:47 1:30:46 1:45:46 83 0:42:16 2:15:13 41 2:30:13 74 . . . . . 51 kg 2 1 0:00:00 0:23:30 0:23:30 29 1:13:27 1:36:57 1:36:57 46 0:50:17 2:27:14 106 2:27:14 75 . . . . . No. name sex cat h'cap swim swim swim bike swim + cycle swim + run finish final otal elapsed total cadence cadence SR SR SR time elapsed position time cycle time elapesd cycle time time position time position c1 c2 r1 r2 r3 (h:m:s) (h:m:s) position (h:m:s) (h:m:s) (rpm) (rpm) (rpm) (rpm) (rpm) 96 sj 2 1 0:00:00 0:22:54 0:22:54 14 1:12:00 1:34:54 1:34:54 30 0:52:24 2:27:18 107 2:27:18 76 92.3 107 91.5 94.3 . 34 pm 1 2 0:05:00 0:23:56 0:28:56 66 1:10:37 1:34:33 1:39:33 65 0:47:53 2:22:26 80 2:27:26 77 . . . . . 270 mp 1 3 0:10:00 0:20:48 0:30:48 75 1:10:43 1:31:31 1:41:31 73 0:46:48 2:18:19 60 2:28:19 78 . . . . . 124 nm 1 3 0:10:00 0:20:56 0:30:56 76 1:05:38 1:26:34 1:36:34 44 0:52:05 2:18:39 63 2:28:39 79 . . . . . 68 cb 1 2 0:05:00 0:23:50 0:28:50 65 1:14:34 1:38:24 1:43:24 79 0:45:28 2:23:52 89 2:28:52 80 . . . . . 53 mh 2 1 0:00:00 0:19:43 0:19:43 1 1:16:25 1:36:08 1:36:08 41 0:52:49 2:28:57 117 2:28:57 81 85.1 100 . . . 99 sp 2 1 0:00:00 0:23:24 0:23:24 24 1:18:02 1:41:26 1:41:26 71 0:47:32 2:28:58 118 2:28:58 82 85.1 . . . . 105 ja 1 3 0:10:00 0:27:10 0:37:10 101 1:09:07 1:36:17 1:46:17 86 0:42:43 2:19:00 65 2:29:00 83 . . . . . 60 ts 2 1 0:00:00 0:20:52 0:20:52 6 x 2:29:37 120 2:29:37 84 83.1 . . . . 234 cd 2 1 0:00:00 0:23:00 0:23:00 15 1:14:39 1:37:39 1:37:39 51 0:52:05 2:29:44 122 2:29:44 85 89.1 . . . . 106 ga 1 3 0:10:00 0:23:06 0:33:06 87 1:12:58 1:36:04 1:46:04 87 0:44:00 2:20:04 67 2:30:04 86 . . . . . 109 db 1 3 0:10:00 0:25:32 0:35:32 95 1:10:48 1:36:20 1:46:20 88 0:44:12 2:20:32 68 2:30:32 87 . . . . . 185 db 1 4 0:15:00 0:25:12 0:40:12 104 1:09:54 1:35:06 1:50:06 97 0:40:51 2:15:57 51 2:30:57 88 . . . . . 228 mt 1 4 0:15:00 0:25:22 0:40:22 107 1:09:36 1:34:58 1:49:58 95 0:41:04 2:16:02 52 2:31:02 89 . . . . . 16 kk 2 1 0:00:00 0:23:33 0:23:33 32 x 2:31:15 128 2:31:15 90 . . . . . 64 tz 2 1 0:00:00 0:20:49 0:20:49 4 1:17:24 1:38:13 1:38:13 54 0:53:15 2:31:28 129 2:31:28 91 83.3 . . . . 119 ck 1 3 0:10:00 0:25:53 0:35:53 96 1:11:57 1:37:50 1:47:50 91 0:43:46 2:21:36 74 2:31:36 92 . . . . . 159 mh 1 3 0:10:00 0:25:08 0:35:08 92 1:12:00 1:37:08 1:47:08 90 0:44:35 2:21:43 76 2:31:43 93 . . . . . 191 gh 1 4 0:15:00 0:20:56 0:35:56 97 1:11:59 1:32:55 1:47:55 92 0:44:13 2:17:08 56 2:32:08 94 . . . . . 149 pb 1 3 0:10:00 0:22:20 0:32:20 83 1:16:14 1:38:34 1:48:34 93 0:44:01 2:22:35 81 2:32:35 95 . . . . . 184 ra 1 4 0:15:00 0:25:10 0:40:10 103 1:13:34 1:38:44 1:53:44 103 0:39:09 2:17:53 58 2:32:53 96 . . . . . 2 dc 2 1 0:00:00 0:22:11 0:22:11 12 1:18:32 1:40:43 1:40:43 68 0:52:16 2:32:59 140 2:32:59 97 90 . . . . 155 bf 1 3 0:10:00 0:25:17 0:35:17 93 1:15:09 1:40:26 1:50:26 99 0:42:37 2:23:03 83 2:33:03 98 . . . . . 147 gb 1 3 0:10:00 0:23:25 0:33:25 90 1:16:38 1:40:03 1:50:03 96 0:44:06 2:24:09 91 2:34:09 99 . . . . . 168 cs 1 3 0:10:00 0:25:26 0:35:26 94 1:14:59 1:40:25 1:50:25 98 0:44:22 2:24:47 94 2:34:47 100 . . . . . 3 kk 2 1 0:00:00 0:22:05 0:22:05 8 1:19:52 1:41:57 1:41:57 75 0:53:20 2:35:17 155 2:35:17 101 . . . . . 154 sf 1 3 0:10:00 0:26:56 0:36:56 99 1:12:19 1:39:15 1:49:15 94 0:46:11 2:25:26 95 2:35:26 102 . . . . . 127 tn 1 3 0:10:00 0:22:34 0:32:34 86 1:10:30 1:33:04 1:43:04 78 0:52:44 2:25:48 98 2:35:48 103 . . . . . 211 gg 1 4 0:15:00 0:25:08 0:40:08 102 1:11:47 1:36:55 1:51:55 100 0:44:20 2:21:15 73 2:36:15 104 . . . . . 188 rc 1 4 0:15:00 0:27:41 0:42:41 110 1:12:16 1:39:57 1:54:57 106 0:41:57 2:21:54 77 2:36:54 105 . . . . . 21 lo 2 1 0:00:00 0:20:51 0:20:51 5 1:22:41 1:43:32 1:43:32 80 0:53:47 2:37:19 161 2:37:19 106 77.8 . . . . 195 pl 1 4 0:15:00 0:25:16 0:40:16 105 1:13:00 1:38:16 1:53:16 102 0:44:07 2:22:23 79 2:37:23 107 . . . . . 192 fk 1 4 0:15:00 0:25:22 0:40:22 108 1:14:16 1:39:38 1:54:38 105 0:43:41 2:23:19 85 2:38:19 108 . . . . . 193 rk 1 4 0:15:00 0:25:17 0:40:17 106 1:12:48 1:38:05 1:53:05 101 0:45:26 2:23:31 87 2:38:31 109 . . . . . 217 pm 1 4 0:15:00 0:28:48 0:43:48 111 1:10:22 1:39:10 1:54:10 104 0:45:03 2:24:13 92 2:39:13 110 . . . . . 219 cm 1 4 0:15:00 0:26:36 0:41:36 109 1:14:07 1:40:43 1:55:43 107 0:45:05 2:25:48 97 2:40:48 111 . . . . . 59 nr 2 1 0:00:00 0:23:17 0:23:17 22 1:22:53 1:46:10 1:46:10 84 0:54:03 2:40:13 172 2:40:13 112 78.6 . . . . 216 ck 1 4 0:15:00 0:30:41 0:45:41 112 1:11:02 1:41:43 1:56:43 108 0:44:43 2:26:26 100 2:41:26 113 . . . . . 43 ta 2 1 0:00:00 0:23:17 0:23:17 21 x 114 81.8 . . . . Appendix K. Study 4, 2000 Triathlon World Championships Race Results

Female triathletes Bib Category Sex SName FName Overall Overall Overall Swim Swim Swim Trans1 Trans1 Cycle Cycle Cycle Trans2 Trans2 Run Run Run No. Place Cat Place Place Cat Place Place Place Cat Place Place Place Cat Place 21 F ELITE F Hackett Nicole 1:54:43.3 1 1 0:18:53.2 2 2 00:00:52.80 27 1:07:09.9 1 1 00:00:34.70 11 0:27:12.7 15 15 20 F ELITE F Montgomery Carol 1:54:50.2 2 2 0:19:28.7 22 22 00:01:03.00 61 1:08:18.9 20 20 00:00:38.40 29 0:25:21.2 1 1 1 F ELITE F Jones Michellie 1:55:25.7 3 3 0:19:24.1 15 15 00:00:41.50 1 1:08:34.1 36 36 00:00:33.40 3 0:26:12.6 2 2 7 F ELITE F Lindquist Barb 1:55:41.0 4 4 0:18:54.6 3 3 00:00:50.00 12 1:07:11.8 2 2 00:00:34.40 7 0:28:10.2 31 31 14 F ELITE F Dittmer Anja 1:55:46.1 5 5 0:20:04.9 37 37 00:00:47.90 4 1:07:50.8 15 15 00:00:38.90 31 0:26:23.6 3 3 52 F ELITE F Taormina Sheila 1:55:50.2 6 6 0:18:52.5 1 1 00:00:51.80 20 1:07:12.0 3 3 00:00:39.00 32 0:28:14.9 32 32 4 F ELITE F Carney Emma 1:55:56.8 7 7 0:20:11.4 42 42 00:00:49.00 7 1:07:45.1 13 13 00:00:39.70 36 0:26:31.6 5 5 26 F ELITE F McMahon Brigitte 1:55:58.9 8 8 0:19:15.5 9 9 00:00:58.10 51 1:08:31.2 33 33 00:00:45.00 51 0:26:29.1 4 4 28 F ELITE F Hirao Akiko 1:56:01.4 9 9 0:19:30.0 24 24 00:00:51.70 19 1:08:24.8 26 26 00:00:39.90 39 0:26:35.0 6 6 18 F ELITE F Lindley Siri 1:56:02.1 10 10 0:19:28.3 21 21 00:00:48.90 5 1:08:27.4 28 28 00:00:36.70 22 0:26:40.8 8 8 5 F ELITE F Hill Rina 1:56:09.8 11 11 0:19:20.3 12 12 00:00:57.10 45 1:08:36.5 38 38 00:00:36.30 15 0:26:39.6 7 7 6 F ELITE F Messmer Magali 1:56:11.8 12 12 0:19:10.2 6 6 00:00:49.40 10 1:08:40.7 39 39 00:00:33.30 2 0:26:58.2 11 11 8 F ELITE F Hosoya Haruna 1:56:12.6 13 13 0:20:08.3 39 39 00:00:50.30 14 1:07:42.9 12 12 00:00:32.20 1 0:26:58.9 12 12 19 F ELITE F Niwata Kiyomi 1:56:14.0 14 14 0:19:41.5 35 35 00:00:50.00 11 1:08:14.7 17 17 00:00:36.40 17 0:26:51.4 9 9 23 F ELITE F Hoogzaad Wicke 1:56:16.2 15 15 0:19:32.1 28 28 00:00:49.30 8 1:08:21.3 22 22 00:00:36.30 15 0:26:57.2 10 10 41 F ELITE F Moreno Carla 1:56:28.6 16 16 0:19:31.0 27 27 00:00:48.90 5 1:08:27.4 28 28 00:00:35.80 13 0:27:05.5 13 13 69 F ELITE F Ashton Melissa 1:56:33.6 17 17 0:19:38.0 34 34 00:00:47.40 2 1:08:26.4 27 27 00:00:34.60 8 0:27:07.2 14 14 12 F ELITE F Suys Mieke 1:56:42.4 18 18 0:19:43.6 36 36 00:00:53.30 32 1:08:05.6 16 16 00:00:34.20 6 0:27:25.7 16 16 30 F ELITE F Nakanishi Machiko 1:56:52.2 19 19 0:19:25.4 17 17 00:00:51.10 17 1:08:31.3 34 34 00:00:34.60 9 0:27:29.8 17 17 10 F ELITE F Mouthon-Michellys Isabelle 1:56:55.9 20 20 0:19:24.4 16 16 00:00:55.20 39 1:08:23.0 24 24 00:00:36.20 14 0:27:37.1 20 20 17 F ELITE F Donnelly Sharon 1:56:58.0 21 21 0:19:20.0 11 11 00:00:55.00 38 1:08:27.6 30 30 00:00:36.40 17 0:27:39.0 22 22 9 F ELITE F King Joanne 1:57:01.9 22 22 0:20:23.0 51 51 00:00:49.30 8 1:07:36.1 5 5 00:00:36.90 24 0:27:36.6 19 19 43 F ELITE F Smet Kathleen 1:57:01.9 22 22 0:19:37.6 33 33 00:00:51.90 22 1:08:16.1 18 18 00:00:41.50 44 0:27:34.8 18 18 38 F ELITE F Reback Laura 1:57:04.5 24 24 0:19:27.3 19 19 00:00:53.80 34 1:08:27.9 31 31 00:00:33.40 3 0:27:42.1 23 23 42 F ELITE F Edocseny Nora 1:57:08.0 25 25 0:20:07.4 38 38 00:00:58.90 54 1:07:42.1 11 11 00:00:42.20 47 0:27:37.4 21 21 32 F ELITE F Koumegawa Yukie 1:57:10.6 26 26 0:19:16.1 10 10 00:00:53.00 30 1:08:35.5 37 37 00:00:35.00 12 0:27:51.0 25 25 61 F ELITE F Mueckel Ute 1:57:14.4 27 27 0:19:23.9 14 14 00:00:47.50 3 1:08:34.0 35 35 00:00:33.90 5 0:27:55.1 26 26 47 F ELITE F Gemignani Silvia 1:57:18.2 28 28 0:19:12.8 8 8 00:00:59.00 55 1:08:42.9 40 40 00:00:41.30 43 0:27:42.2 24 24 Bib Category Sex SName FName Overall Overall Overall Swim Swim Swim Trans1 Trans1 Cycle Cycle Cycle Trans2 Trans2 Run Run Run No. Place Cat Place Place Cat Place Place Place Cat Place Place Place Cat Place 25 F ELITE F Newman Jill 1:57:27.4 29 29 0:20:12.0 43 43 00:00:57.80 49 1:07:36.7 8 8 00:00:39.20 33 0:28:01.7 27 27 15 F ELITE F Brice Sian 1:57:29.8 30 30 0:19:29.8 23 23 00:00:52.00 24 1:08:22.0 23 23 00:00:36.70 22 0:28:09.3 30 30 66 F ELITE F Anisimova Nina 1:57:34.7 31 31 0:19:21.7 13 13 00:00:58.30 52 1:08:31.0 32 32 00:00:37.30 25 0:28:06.4 28 28 29 F ELITE F Soldan Sandra 1:57:39.6 32 32 0:19:26.0 18 18 00:00:52.80 27 1:08:24.5 25 25 00:00:49.00 53 0:28:07.3 29 29 24 F ELITE F Williamson Evelyn 1:58:16.5 33 33 0:20:13.2 46 46 00:00:51.10 17 1:07:37.6 9 9 00:00:36.40 18 0:28:58.2 38 38 22 F ELITE F Overbye Marie 1:58:19.0 34 34 0:20:12.3 44 44 00:00:52.90 29 1:07:41.7 10 10 00:00:39.80 37 0:28:52.3 37 37 57 F ELITE F Ianesi Manuela 1:58:20.4 35 35 0:20:12.7 45 45 00:00:51.90 22 1:07:46.6 14 14 00:00:39.50 34 0:28:49.7 36 36 48 F ELITE F Pepels Silvia 1:58:43.0 36 36 0:20:09.1 41 41 00:00:54.80 37 1:07:36.5 7 7 00:00:41.60 46 0:29:21.0 39 39 68 F ELITE F Gibbs Beccy 1:58:51.6 37 37 0:19:09.6 5 5 00:00:53.00 30 1:08:47.0 41 41 00:00:36.50 21 0:29:25.5 40 40 50 F ELITE F Donner Dominique 1:58:58.6 38 38 0:19:36.2 32 32 00:00:56.30 42 1:08:17.8 19 19 00:00:36.40 19 0:29:31.9 42 42 63 F ELITE F Matter Sibylle 1:59:07.8 39 39 0:19:30.5 26 26 00:00:54.70 36 1:08:20.4 21 21 00:00:38.30 28 0:29:43.9 43 43 37 F ELITE F Rose Jenny 1:59:16.0 40 40 0:20:14.3 47 47 00:00:53.60 33 1:07:36.1 5 5 00:00:42.90 48 0:29:49.1 44 44 45 F ELITE F Cigana Edith 1:59:35.4 41 41 0:20:15.0 48 48 00:00:57.40 47 1:07:34.3 4 4 00:00:39.60 35 0:30:09.1 45 45 31 F ELITE F Blanco Velasco Maribel 2:02:25.9 42 42 0:20:30.9 53 53 00:00:56.90 44 1:11:33.7 42 42 00:00:40.10 40 0:28:44.3 34 34 62 F ELITE F Berkova Renata 2:02:40.7 43 43 0:19:30.1 25 25 00:01:01.00 58 1:13:03.3 48 48 00:00:37.90 26 0:28:28.4 33 33 33 F ELITE F Moore Lizel 2:02:48.4 44 44 0:20:32.7 54 54 00:00:57.40 47 1:11:50.8 43 43 00:00:41.50 44 0:28:46.0 35 35 36 F ELITE F Mouthon Beatrice 2:03:39.0 45 45 0:20:34.9 55 55 00:00:58.60 53 1:11:54.3 44 44 00:00:40.30 41 0:29:30.9 41 41 54 F ELITE F Estedt Ines 2:04:28.4 46 46 0:20:18.3 50 50 00:00:57.30 46 1:12:15.8 46 46 00:00:43.10 49 0:30:13.9 46 46 APPENDIX L

Effect of cycle cadence on running mechanics

Study Infomation

Grant Landers is undertaking this project as a PhD research student in the Department of Human Movement and Exercise Science at The University of Western Australia, under the supervision of Professors Brian A. Blanksby and Tim R. Ackland.

The study aims to determine the possible effects of prior cycling on running economy and mechanics. It is hoped to gain a greater understanding of the inter-relationship between these two disciplines within triathlon performance. These measurements could increase the information available to athletes, coaches and the general public, thereby improving the profile of all triathlon.

Two testing sessions will be required. Firstly a running step test will be performed to exhaustion, together with measurement of anthropmetric charateristics. This will allow determination of maximal oxygen consumption. During the second testing session, following a warm-up of 6 minutes running at 75% Maximal Aerobic Speed (MAS), each subject will undertake a series of five cycle-run transitions presented in a random order. The cycle will be performed in an unloaded fashion, using the subjects own bike mounted on a wind trainer and removal of the chain, for six minutes at one of five cadences. Immediately following the cycle the subject shall run 1.5 km on a treadmill at a pace equivalent to 75% MAS. Each participant will receive a report outlining the individual profile as well as mean data of elite triathletes.

Any concerns regarding this study should be directed to the secretary of the Human Rights Committee at The University of Western Australia (Mrs K. Kirk, 9380 3703). For any organisational queries, please contact one of the research team at the address or phone numbers below. All study participants will be provided with a copy of the information/consent form for their personal records.

Thank you in anticipation for your co-operation,

Yours Sincerely,

......

I understand that all individual information will be kept confidential and, if any data are published, then anonymity will be maintained.

The nature of the project has been fully explained to me.

…………………………… Signature Date (parent or guardian if under 18)

Appendix M

Unloaded Cadence Data Collection Card

Name

Address dob

mass height trochanteric height tibiale laterale height upper limb length arm length forearm length average weekly training km hours sessions swim bike run number of competitive years

Number of races super sprint sprint 750:20:5 olympic 1.5:40:10 long course