Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Muscle function and control roles of the elbow flexion system in European Starlings Wood, Leo 2020

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Notice for Google Chrome users:
If you are having trouble viewing or searching the PDF with Google Chrome, please download it here instead.

Item Metadata


24-ubc_2020_november_wood_leo.pdf [ 15.66MB ]
JSON: 24-1.0391993.json
JSON-LD: 24-1.0391993-ld.json
RDF/XML (Pretty): 24-1.0391993-rdf.xml
RDF/JSON: 24-1.0391993-rdf.json
Turtle: 24-1.0391993-turtle.txt
N-Triples: 24-1.0391993-rdf-ntriples.txt
Original Record: 24-1.0391993-source.json
Full Text

Full Text

MUSCLE FUNCTION AND CONTROL ROLES OF THE ELBOW FLEXION SYSTEM IN EUROPEAN STARLINGS by  Leo Wood  B.Sc., Texas A&M University, 2017  A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF  MASTER OF SCIENCE in THE FACULTY OF GRADUATE AND POSTDOCTORAL STUDIES (Zoology)  THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver)  June 2020  © Leo Wood, 2020  ii  The following individuals certify that they have read, and recommend to the Faculty of Graduate and Postdoctoral Studies for acceptance, a thesis entitled:  Muscle function and control roles of the elbow flexion musculature in European Starlings  submitted by Leo Wood in partial fulfillment of the requirements for the degree of Master of Science in Zoology  Examining Committee: Doug Altshuler, Professor, Zoology, UBC Supervisor  Mike Gordon, Assistant Professor, Zoology, UBC Supervisory Committee Member  Phil Matthews, Assistant Professor, Zoology, UBC Additional Examiner   Additional Supervisory Committee Members: Robert Shadwick, Professor, Zoology, UBC Supervisory Committee Member iii  Abstract The muscular basis for dynamically changing wing shape in birds is not well understood, owing to the complex and redundant network of muscles in the avian wing as well as a poorly-understood and diverse avian musculature. In passerines the biceps, Tensor Propatagialis Brevis (TPB), and Tensor Propatagialis Longus (TPL) share redundant control of elbow flexion. The TPB and TPL are smaller than the biceps but their lever arms are considerably larger, which suggests that the three muscles should produce similar elbow torque. Another distinct anatomical feature is that the TPL and TPB has have long in-series tendons unlike the biceps. Collectively, the anatomy of theses muscles suggests a hypothesis for the control of elbow flexion in flight: The distally attached TPL and TPB with long in-series tendons act primarily in torque control, whereas the proximally-attached biceps with no in-series tendon acts primarily for position control. To test this hypothesis, I collected EMG and high-speed kinematics on European starlings (Sturnus vulgaris) flying in a wind tunnel under varied conditions. A treatment was applied of attached weights to the distal wing on some flights to alter wing inertia and observe corresponding changes in kinematics and muscle activity suggesting specific muscle control roles. Wing weights did not have a robust effect on elbow kinematics or muscle activity, but EMG, kinematics, and morphology data suggested different functional roles for the three muscles. All three elbow flexors showed peak activity at the beginning of the downstroke, and the biceps demonstrated timing consistent with a role stabilizing the elbow. The TPB and TPL had in vivo timing and morphology consistent with unique roles for wing morphing, with precise and short-duration activity from the TPB consistent with its role in manus extension as well as elbow flexion and a two-pulse response of the TPL consistent with storage and release of elastic energy.   iv  Lay Summary  Birds dramatically alter the shape of their wings during flight, but we don’t know how the muscles in their wings accomplish this. One part of the wing, where the muscles are responsible for flexing the elbow, is potentially important for controlling the overall shape of the wing, and has several muscles unique to birds that haven’t been studied closely. This study aims to learn when the muscles in this part of the wing are active in flapping flight, how the muscles may be used to control wing shape, and how the different muscles that flex the elbow are controlled together.  v  Preface All of the work presented in this manuscript were conducted in the University of British Columbia Flight Lab and the University of Montana Flight Laboratory under the guidance of Dr. D. Altshuler and Dr. B. Tobalske. All of the text within this manuscript is original and unpublished material. All experiments presented in this manuscript were covered under the Institutional Animal Care and Use Committee (IACUC) Animal Use Proposal 018-19BTDBS-041919, “Muscle contributions to wing morphing in bird flight”, as well as the Animal Care Committee (ACC) A19-0113 protocol. I devised the experiment design and data collection procedure with input from Dr. B. Tobalske and Dr. D. Altshuler, and I performed all experiments and data collection with the aid of Dr. B. Tobalske, A.B. Lapsansky, and H. Chase at the University of Montana Flight Laboratory. I and Dr. B. Tobalske performed all surgical procedures, and the surgical procedure was designed by Dr. B. Tobalske.  I performed all data analysis presented in this manuscript, with frequent discussions from Dr. D. Altshuler and my colleagues in the UBC Flight Lab.  vi  Table of Contents  Abstract ......................................................................................................................................... iii	Lay Summary ............................................................................................................................... iv	Preface .............................................................................................................................................v	Table of Contents ......................................................................................................................... vi	List of Tables .............................................................................................................................. viii	List of Figures ............................................................................................................................... ix	Acknowledgements ........................................................................................................................x	Dedication ..................................................................................................................................... xi	Chapter 1: General Introduction .................................................................................................1	1.1	 Background –Wing Morphing and the Role of Wing Musculature ................................ 1	1.2	 Muscular Control and Redundancy ................................................................................ 3	1.3	 The Propatagium ............................................................................................................. 6	1.4	 Study Goals and Design .................................................................................................. 9	1.5	 Hypotheses and Predictions .......................................................................................... 12	Chapter 2: Muscle Function and Control Roles In-Flight .......................................................14	2.1	 Introduction ................................................................................................................... 14	2.2	 Materials and Methods .................................................................................................. 15	2.2.1	 Animals and Flight Training ................................................................................. 15	2.2.2	 Data Collection Procedure .................................................................................... 16	2.2.3	 Electromyography and Surgical Implantation ...................................................... 16	2.2.4	 High-Speed Kinematics ........................................................................................ 18	vii  2.2.5	 Morphological Measurements .............................................................................. 18	2.2.6	 Data Analysis ........................................................................................................ 21	2.3	 Results ........................................................................................................................... 24	2.4	 Discussion ..................................................................................................................... 34	2.4.1	 Muscle Morphology and Energetic Cost .............................................................. 34	2.4.2	 Functional Implications of Muscle Timing ........................................................... 35	2.4.3	 Evaluation of Control Roles .................................................................................. 37	2.4.4	 Evaluation of Control Roles through Wingbeat Frequency .................................. 38	Chapter 3: Conclusion .................................................................................................................40	3.1	 Limitations .................................................................................................................... 41	3.2	 Future Work .................................................................................................................. 43	Bibliography .................................................................................................................................44	Appendix .......................................................................................................................................53	 viii  List of Tables  Table 2.1 Pearson correlation coefficients between mean pulse amplitude and cycle frequency. 32	Table 2.2 Pearson correlation coefficients between total duration active per cycle and cycle frequency. ...................................................................................................................................... 32	 ix  List of Figures Figure 1.1. Ventral view of European starling propatagial anatomy. ............................................. 9	Figure 1.2. Flow chart of wing weight experiment outcomes and implications. .......................... 12	Figure 2.1. Analysis process for EMG and kinematics data. ........................................................ 23	Figure 2.2. Analysis process for extracting wingbeat downstroke. .............................................. 24	Figure 2.3. Morphology measurements of starling propatagial muscles. ..................................... 25	Figure 2.4. Elbow angle and EMG activation of descending and level flights. ........................... 27	Figure 2.5. Effect of wing weights on kinematics and muscle activity. ....................................... 30	Figure 2.6. Wingbeat frequency and its effects on amplitude and duration of muscle activation. 31	Figure 2.7. Elbow angle and EMG activation grouped by wingbeat frequency. .......................... 33	  x  Acknowledgements I extend my gratitude to my supervisory committee: Dr. D. Altshuler, Dr. R. Shadwick, and Dr. M. Gordon for their dedication of time and effort to my graduate studies. I also owe a special thanks to my lab: S. Senthivasan, S. Wu, J. Wong, J. Theriault, A. Gaede, E. Press, V. Baliga, M. Armstrong, and S. Azargoon, with additional mention to K. Ashley, S. Scratch, and F. Ciocca.  I have a tremendous debt of thanks to Dr. B. Tobalske, whose selfless efforts of time, thought, and guidance made this work possible, and his students A.B. Lapsansky and H. Chase, for their aid and kindness in my time in Montana.  Special thanks to the students, faculty, and staff of the University of British Columbia Zoology Department, of which there are more people I would like to thank than there is space to name.     xi  Dedication   To all my loving family and friends, who supported me in my decision to walk away from a real career and instead move to Canada to study bird elbows.1  Chapter 1: General Introduction 1.1 Background –Wing Morphing and the Role of Wing Musculature The ability of the avian wing to dynamically alter its shape, often dubbed “wing morphing”, is thought to be fundamental to how avian wings function. Wing morphing enables the same flight surface to continuously alter its properties, enabling a diverse array of flight behaviors (Baliga et al., 2019) and allowing properties such as lift, drag, and stability to be highly variable and controlled to different conditions (Altshuler et al., 2015; Harvey et al., 2019). Even in flapping flight, the avian wing exhibits dramatic changes in shape which are highly controlled; birds will often modulate flight speed without changing flapping frequency, instead controlling aerodynamic forces largely through wing morphing (Schmidt-Wellenburg et al., 2007; B. Tobalske & Dial, 1996).  Some recent studies have examined external aspects of wing morphing, such as influences on aerodynamics to uses in flight behaviors (Chang et al., 2011; Lapsansky et al., 2019; Usherwood et al., 2020). In contrast, there is a fundamental lack of understanding of how internal mechanics of the musculoskeletal system of the wing actually produces and controls wing shape. In part this is due to the difficulties involved in measurement—basic information on muscle activity and strains in vivo requires invasive implantation of electromyography (EMG) wires and either sonomicrometry beads (Robertson & Biewener, 2012) or x-ray opaque markers (Konow et al., 2015). These are challenging methods which cannot be applied to a large number of muscles simultaneously, are often constrained by animal size, and provide an incomplete picture of how wing muscles are actually producing and modifying wing kinematics and dynamics. Measuring force production of individual wing muscles in-flight would be the most direct way to understand their contributions and roles in altering wing shape, but this is largely 2  infeasible for intrinsic wing muscles. While strain gauges attached to skeletal elements have often been used to measure pectoralis force in vivo (Biewener & Roberts, 2000; Jackson et al., 2011; B. W. Tobalske, 2007; B. W. Tobalske & Biewener, 2008), this requires skeletal surfaces largely free of stresses induced by inertial forces and other skeletal muscles to resolve force production from individual muscles. Alternatives such as tendon buckle methods used in mammalian locomotion studies have not been implemented on wing muscle, owing to constraints on size, procedure invasiveness, and introduction of additional mass on the outboard wing (Biewener & Roberts, 2000). In spite of these difficulties, prior literature has established important findings on intrinsic wing muscle activity during flight. Early EMG data on major intrinsic muscles in pigeons (Columbia livia) and starlings (Sturnus vulgaris) refined understanding of muscle function from simpler anatomical assumptions, providing hypotheses for the roles of individual wing muscles. These studies further noted that muscles in the proximal wing and forearm were active throughout the entire wingbeat, with concurrent activity of antagonistic pairs suggesting intrinsic muscles playing a major role in wing stability (Dial, 1992a; Dial et al., 1991). An experimental treatment on pigeons by the same authors found that when forearm muscles were denervated by severing the medianoulnaris and radialis nerves, birds were unable to take off or land but were able to maintain level flapping flight with no discernable differences in wing kinematics compared to controls (Dial, 1992b).  Robertson et al.’s 2012 study on pigeons is to date the only measurement of in vivo muscle strain of intrinsic wing muscles in birds, using sonomicrometry beads to measure fascicle strains in the biceps, scapulotriceps, and humerotriceps (Robertson & Biewener, 2012). From the data, the authors concluded that the biceps potentially produced negative and positive work 3  during a wingbeat cycle, and likely functioning to stabilize the elbow during the downstroke. They also found it likely that the two heads of the triceps acted together with net positive work to extend the elbow during the upstroke, but showed separate activity in stabilizing different parts of the upstroke. Their predictions and findings on the pigeon humerotriceps were later expanded on with an in situ work loop study by Theriault et al. (J. S. Theriault et al., 2019), who found that the humerotriceps exhibited predominantly net negative work when subjected to activation and strain patterns similar to Robertson et al.’s measurements. The study further concluded that while the humerotriceps predominantly acted as a “brake” by generating net negative work, it typically generated both positive and negative instantaneous power during each cycle and likely has a dynamically variable role in different modes of flight. These studies have served to develop a functional understanding of the role of wing musculature in flight, but understanding of the muscular basis for avian wing morphing is still quite limited. One major limitation is the problem of redundancy; the avian wing musculature is a highly complex network, with many more active elements interacting to produce motion than degrees of freedom.   1.2 Muscular Control and Redundancy  Avian wings are, at their core, vertebrate limbs evolved for use as wings, and a consistent pattern of this evolution has been the loss of underlying degrees of freedom. The carpal and metacarpal bones have fused over time into the single carpometacarpus (Norell & Clarke, 2001; Rick J Vazquez, 1992), the phalanges are near-vestigial with limited mobility (Xu & Mackem, 2013), and there is an extremely strong coupling of elbow and wrist motion in all flying birds (Baliga et al., 2019; Stowers et al., 2017; R J Vazquez, 1994). Despite the simplification in 4  structural degrees of freedom, however, there has been no associated simplification of wing musculature, with birds possessing 22-26 major wing muscles (Dial, 1992a; Hudson & Lanzillotti, 1955; Razmadze et al., 2018; Rick J Vazquez, 1995) compared to the ~22 analogous arm muscles found in humans. The benefits of reducing structural degrees of freedom are somewhat well understood, including reduction of muscle effort to maintain aspects of wing shape subject to high stress in flight (Rick J Vazquez, 1995) and reducing inertial costs in flight by concentrating muscle mass more proximally (Konow et al., 2015; R J Vazquez, 1994). This reduction in degrees of freedom, however, leaves a many-to-one problem when trying to understand the functional contributions of muscles to wing shape. The reduced structural degrees of freedom and lack of similar reduction in motor complexity makes the avian wing a particularly difficult case of an already well-established problem encountered when studying any neuromuscular locomotion system, often called the “degrees of freedom problem” or the “redundancy problem” (Herzog, 2017; Scott, 2004). To accomplish a desired task such as flapping a wing or grasping a glass involves many redundant kinematic degrees of freedom (many different trajectories, velocities, and accelerations accomplish the task), and to accomplish even a single one of these trajectories involves many redundant dynamic degrees of freedom (many possible combinations of muscle forces which will produce the same kinematics). This pattern of there being “too many” solutions to achieve a goal seemingly repeats at every length scale, with joints often composed of many redundant muscles, each of which are composed of many redundant muscle fibers, often innervated by redundant motoneurons. Why are animals set up with such bewildering redundancies, and how can a nervous system navigate infinite spaces to control muscles and produce motion? 5  Unified understanding is difficult, but there are many different theories and re-framings of the problem. Some of these theories, such as the hypothesis that the Central Nervous System (CNS) controls muscles in force control using a feedforward internal model, have all but been proven wrong (Ostry & Feldman, 2003). Other theories have simply proven difficult to falsify, like the concept of motor synergies, where muscles and motor units are thought to be co-activated by the CNS in groupings to simplify redundancies (Bernstein, 1966; Latash, 2010). Synergies are experimentally apparent, but it is difficult to attribute them to an underlying control schema when many possible control programs can independently demonstrate synergies simply as an emergent property (Tresch & Jarc, 2009). This is a good illustration of the limitations of current understanding of motor control—while data of external behavior abounds, a large set of different underlying control schemes could produce the observed behaviors. Said another way, there is a redundancy of theories on the redundancy problem! In spite of the difficulty of understanding how the CNS may organize control of redundant muscles, for the case of avian wing morphing some general principles based on muscle architecture can prove valuable. It is well documented, for instance, that the length of the in-series tendon in a Muscle-Tendon Unit (MTU) is a good predictor of whether that MTU will be used for elastic energy storage (Biewener, 2016; Pollock & Shadwick, 1994), as well as greatly defining the control authority of a muscle. A muscle with no in-series tendon will have a high position control gain (∆joint angle /∆muscle length) as muscle length will be directly coupled with joint angle (Biewener, 2016), whereas a long in-series tendon will decouple muscle length from joint position (Sawicki et al., 2015), reducing position control gain but improving an actuator’s ability to operate in force control by improving bandwidth and stability (G. A. Pratt & Williamson, 1995; J. Pratt et al., 2002). In general, too, in-series tendons enable dynamic roles 6  involving energy storage and power amplification that are unavailable to muscle tissue alone (Roberts & Azizi, 2011). Architectural trade-offs and specializations of MTUs can, in some cases, simplify understanding of how redundant muscles may be controlled in vivo. Redundant muscles are often not identical, and architectural differences can constrain the use of muscles to predictable specific timings and functional roles for different tasks. The three redundant ankle extensors in the cat hindlimb, for instance, have been shown to alter their relative contributions with movement speed in a manner consistent with the limitations of their architectures and fiber types (Prilutsky et al., 1996).  Understanding how redundant muscles contribute to wing morphing and are controlled, then, is a difficult task, but known patterns of how muscles function based on architecture can be beneficial. To address the shortcomings in understanding of wing morphing whilst dealing with the redundancy problem, the most knowledge can be gained by isolating an area of the wing which has specialized musculature, some significant importance to wing motion, and is understudied. One major area of the wing, the propatagium, is a promising candidate for this.  1.3 The Propatagium The propatagium is the forward flight surface of the avian wing, a membranous surface connecting the shoulder and wrist of all flighted birds. Bats possess a lightly analogous structure (Cheney et al., 2014), but the avian propatagium is wholly unique in its structure and variation, possessing muscular and tendinous elements unique to birds. Figure 1.1 illustrates the propatagial anatomy of the European starling (Sturnus vulgaris), a common species with anatomy roughly representative of Passeriformes, a large clade which comprises over half of all 7  bird species (Oliveros et al., 2019; Prum et al., 2015). In these species, the propatagium is comprised of a middle elastinous fascial sheet, bounded at the forward edge by the long, elastic tendon of the Tensor Propatagialis Longus (TPL) and suspending in its midsection the Tensor Propatagialis Brevis (TPB). Both of these muscles are unique to birds, with no analogous equivalents in any vertebrates. Qualitative studies on avian propatagial anatomy date back hundreds of years, with early scientific descriptions at least as early as 1773 (d’Azyr, 1773) and a wealth of anatomical studies throughout the 19th and 20th centuries (Fisher & Goodman, 1955; Hudson & Lanzillotti, 1955; Shufeldt, 1890; Swinebroad, 1954). One feature of the propatagium many authors have consistently noted is its abnormally high degree of variation between species: Columbiformes (the clade of pigeons and doves), for instance, have an entire muscle referred to as the “pars biceps” which is found in the propatagial anatomy of no other avian species (Dial, 1992a; Hieronymus, 2016). Aequornithes (the clade of core water birds including gulls, storks, and others) lack the TPB and TPL in any significant capacity, instead possessing a unique set of large, branching tendons in the propatagium (Berge, 1970; Meyers & Stakebake, 2005). Even the near-passerine clade of Psittaciformes (parrots) possesses propatagial musculature unique from what is observed in passerines (Razmadze et al., 2018). This thesis will focus purely on the propatagial anatomy in the European starling, as they are a relatively common model species and have propatagial anatomy which is roughly representative of all passerines, but it is important to note just how variable the muscular structures in the propatagium are.  The TPB and TPL are both elbow flexors, which makes them redundant—birds, like other vertebrates, have a sizeable biceps acting to generate elbow flexion. The TPB and TPL, however, have a number of unusual features which distinguish them from the biceps: The TPB, 8  for instance, does not distally attach to bone, instead connecting to the Extensor Metacarpi Radialis (EMR), a forearm muscle responsible for manus extension. This allows the TPB to potentially stretch the EMR and reduce its effective length, particularly at higher elbow angles, co-opting the EMR’s function and extending the manus as well as flexing the elbow. The TPL has a similarly convoluted attachment, in some species connecting to the manus as well as the radius, potentially influencing wing shape outside of elbow angle (Brown & Cogley, 1996). The midsection of the TPL’s extremely long tendon is highly elastinous in composition (Brown, Baumel, et al., 1994; Brown, Butler, et al., 1994), allowing it to repeatedly experience strains of 100% or greater each wingbeat (Brown et al., 1995). In spite of this, the TPL’s tendon demonstrates a very low hysteresis of 1-2%, unlike other tissues with a high elastin content which are typically highly dissipative (Lillie & Gosline, 1990; Mijailovich et al., 1994).  Outside of the previously mentioned anatomical studies, a single EMG study in pigeons (Dial, 1992a), and some mention in a series of studies on the passive mechanics of the propatagium by Brown et al. (Brown, Baumel, et al., 1994; Brown, Butler, et al., 1994; Brown et al., 1995; Brown & Cogley, 1996), very little study has been performed on the propatagial muscles. This may partially be due to the diverse muscular anatomy of the avian propatagium, and partially also due to the diminutive size of the TPB and TPL compared to the biceps, which is typically thought to be the only primary elbow flexor in birds. Both the TPB and TPL, however, attach at a considerable distance from the center of rotation of the elbow, providing large lever arms which result in much higher torques for the same force. Because of the strong elbow-manus present in avian wings, too, muscles which control elbow angle have an outsize influence on the shape of the wing, all of which warrants further study of these muscles.   9   Figure 1.1. Ventral view of European starling propatagial anatomy. Tensor Propatagialis Longus (TPL), Tensor Propatagialis Brevis (TPB), and Biceps muscles labeled. Not all wing musculature is shown. Hypotheses on the effect of increasing elbow flexor attachment point and tendon length on muscle function and control roles are written on wing.    1.4 Study Goals and Design The elbow flexion system in passerines exists at the intersection of several major gaps in understanding of the muscular basis of wing morphing. The highly variable and novel propatagial muscles are poorly understood, with little study beyond that of anatomy despite their 10  potentially outsize influence on wing shape. Simply collecting EMG data on the timing of the TPB and TPL would fill a gap in literature, as a 1992 study on pigeons constitutes the only measurement of these muscles in vivo (Dial, 1992a). These muscles are also well-positioned to study how attachment point and MTU architecture may determine control roles between redundant muscles, with three muscles performing the same role of flexing the elbow yet demonstrating markedly different architectures, with sequentially increasing lever arms and tendon lengths. To remedy these gaps and develop understanding of how these novel muscles contribute to and control wing morphing in vivo, I aimed to answer the following questions: 1. What is the timing and activity of the passerine elbow flexors in flight, and what are the functional implications of this timing and activity? 2. What are the effects of increasing lever arm and tendon length on the control roles and function of the elbow flexor muscles?  I chose to answer these questions using an EMG study on European starlings. Training starlings to fly in a wind tunnel under a variety of speeds and at different levels of inclination would allow a variety of flight behaviors to be repeatedly elicited in a stationary volume. This enables measurements of muscle timing, activity, and wing kinematics in a range of flight conditions necessary to answer the study questions.  A purely mensurative experiment can be sufficiently informative, but to study the control roles of the elbow flexors in detail a controlled experiment was warranted. I included in my study a treatment where 0.25g weights were attached to the wrists of starlings to increase the moment of inertia of the wing distal to the elbow. Increasing wing inertia would allow observation of the control at the elbow by inducing a change that has to be compensated for by 11  either muscle recruitment or an observable change in wing kinematics. Viewing the elbow as a hinge, from Newton’s 2nd law the dynamics of the system would be governed by the equation  Σ"($) = ' ∗ )($) (1) where Σ"($) is the sum of all individual torques "*($) about the elbow, ' is the moment of inertia of the forearm and wing distal from the elbow, and )($) is the rotational acceleration of the elbow over time. By adding wing weights, ' is increased such that either Σ"($) or )($) must change. If )($) does not change, observed elbow kinematics will be unchanged and a change in the muscle torques about the elbow must occur. Though in reality the wing is far more complex than a hinge and a 3-dimensional inertia tensor would have to be considered, these principles still stand.  This experiment design has the benefit of potentially parsing out whether muscles observed operate more in position control or torque/acceleration control. Muscles operating purely in torque control would only act to produce an expected torque over time "+,-.($), an output that can be maintained (to an extent) regardless of changes to moment of inertia and kinematics: As such, a pure torque-controlled muscle would not alter activity with the addition of wing weights. Muscles operating purely in position control would act to maintain an expected position )($) over time (which implies maintaining an expected acceleration )($) over time), which would require a commensurate change in the muscle’s torque output "($), observable through increased EMG amplitude. A summary of the possible outcomes and control implications for the wing weight trials is shown in Figure 1.2. 12   Figure 1.2. Flow chart of wing weight experiment outcomes and implications.  1.5 Hypotheses and Predictions A summary of my hypotheses on lever arm and tendon length is presented in Figure 1.1. I hypothesized that the increasing lever arm between the biceps, TPB, and TPL would result in the more distally attached TPL and TPB producing similar elbow flexion torque to the biceps from lower muscle force, reducing the effective cost of producing a given torque. I also hypothesized that the high position control gain resulting from a short lever arm and lack of an in-series tendon would make the biceps better-suited for position control, whereas the long in-series tendons of the TPB and TPL in particular would promote torque control-oriented activity.  If these hypotheses are true, we would see similar estimates of maximal torque from all three muscles despite the more proximally-attached muscles possessing a much higher cross-sectional area (proportional to force). These hypotheses alone do not predict the muscle timing 13  that would be observed, but it would be expected that flight modes where kinematics are likely more controlled than dynamics, such as descending flight or lower-frequency wingbeat cycles, proximally-attached muscles will show greater activity than distally-attached muscles.  There are different predicted outcomes for the effect of wing weights based on my hypothesis, shown in Figure 1.2. If the wing weights are large enough to be effective, I see the most likely outcome to be that elbow angle is controlled by the CNS and the biceps is the only muscle primarily tuned for position control—this would result in a notable increase in activity and duration of activity for the biceps, with no corresponding change in elbow kinematics or activity of the other flexors.     14  Chapter 2: Muscle Function and Control Roles In-Flight 2.1 Introduction Birds can alter the shape of their wing in flight, often called “wing morphing”, in a manner thought to be crucial avian flight such as flapping flight (B. W. Tobalske et al., 2003) and gliding flight (Harvey et al., 2019). The muscular underpinnings of how wing morphing is accomplished, however, is poorly understood. This lack of understanding derives from many difficulties in studying the wing, but a significant portion is due to both the large number of wing muscles relative to the number of wing degrees of freedom (Rick J Vazquez, 1995), their relative small size in most species, and also the highly variable and novel musculature found in the avian wing (Fisher & Goodman, 1955; Hudson & Lanzillotti, 1955; Razmadze et al., 2018).  The elbow flexion system of passerines is well-positioned to remedy these gaps in understanding. Because of the strong elbow-manus coupling in avian wings (Baliga et al., 2019; Stowers et al., 2017; R J Vazquez, 1994), muscles which control elbow angle have an outsize influence on wing shape. Beyond this, the elbow flexion system is a particularly promising subject for studying wing morphing as it is composed of three redundant muscles with unique morphologies, providing an effective case study on how muscle architecture may affect muscle control roles. Two of the elbow flexors, the TPB and TPL, are also wholly unique to birds and have a number of unusual features, including the attachment of the TPB to another distal muscle instead of bone (Hudson & Lanzillotti, 1955) and a highly elastinous tendon in-series with the TPL that defines the wing leading edge (Brown, Butler, et al., 1994).  I sought to collect information on the timing and activity of the passerine propatagial muscles, and study how the three redundant elbow flexors in passerines are controlled in flapping flight. To do this, I devised and performed an experiment where EMG and wing 15  kinematics were collected on European starlings in varied flight conditions in a wind tunnel, and where flights with additional inertial loading from wing weights were measured to observe effects of differing muscle attachment points and tendon lengths on the control roles between the redundant elbow flexors. By inducing changes in wing inertia, associated changes in wing kinematics and/or muscle activity could inform the likely control roles exhibited by the elbow flexors, and whether wing kinematics were closely controlled for by the CNS. Additional morphological measurements were taken, to better understand the potential maximal output and energetic implications of the morphology of the elbow flexors.   2.2 Materials and Methods 2.2.1 Animals and Flight Training Study subjects were adult European Starlings (Sturnus vulgaris), body mass 72.78±4.74g, mean±s.d., trapped from wild populations in Missoula, Montana area using agricultural ladder traps in accordance with all local laws and an IACUC-approved and ACC-approved protocol. All birds were housed in 1m x 1m x 0.5m outdoor cages with shaded areas and given baths, water, and insect suet ad libitum, as well as daily feedings of dry cat food and mealworms.  Initial training consisted of operating the wind tunnel at comfortable speeds of 6-10 m/s and allowing birds to acclimate on a perch in the tunnel for periods of 30 minutes each day. The perch was initially removed for short durations of 1-2 seconds to induce flight, and progressively the perch was repeatedly removed during training sessions for longer durations until birds could demonstrate repeatable sustained flights of at least 30 seconds or greater. Initial training sessions were performed with wind tunnel level, but later sessions were performed with wind tunnel inclination alternating between 20° and level to acclimate birds with gliding and descending 16  flight. Of the 12 birds trapped for the study, 7 demonstrated flight performance sufficient for EMG implantation and data collection. 2.2.2 Data Collection Procedure All experiments in this study were performed at the University of Montana’s Flight Laboratory facility using a 30”x30”x30” test section variable-speed wind tunnel. Wind speeds in the tunnel were measured and calibrated prior to experiments using a pitot-static tube mounted in the center of the test section and differential pressure sensor, and wind speeds during all animal flights were independently verified using a Kestrel 5500 handheld anemometer (Kestrel Instruments, PA, USA).  For birds with sufficient training to be included in study, the data collection process consisted of an implantation surgery followed by a day of recovery, after which data collection was performed. All individuals initially performed descending flights, with the wind tunnel angled down 20°, with a randomized sequence of wind speeds at either 8 or 10 m/s. Flights were then recorded with the wind tunnel leveled, again with a random sequence of speeds at either 8 or 10 m/s. Hemispherical lead weights of 0.25g each were then attached with cyanoacrylate adhesive to skin on both wings at the distal end of the ulna, and then a final sequence of flights was recorded with the tunnel level and wind speed at only 8 m/s (to avoid overexertion due to the addition of weights). Due to constraints from both the ease of removability of weights and the time required to angle the wind tunnel, the above sequence of treatments was not randomized or altered between individuals. 2.2.3 Electromyography and Surgical Implantation EMG implants consisted of four intramuscular bipolar EMG wires, a ground wire, and an EMG back plug. Bipolar EMG wires were 0.004-inch diameter enamel-coated 99.9% silver wire 17  (California Fine Wire Co., CA, USA), coaxially wound to reduce EM noise with 0.5mm exposed tips 1mm apart. A ground-reference 0.5mm diameter, 1cm long copper wire loop was placed adjacent to the implant underneath skin, and the implant itself consisted of a 2x3 grid of female strip connectors (Milspec West, NC, USA), 9.52x3.16mm, with an epoxy base for securing and fitting the implant underneath skin.  EMG implant surgeries were performed on each bird one day prior to data collection following a procedure similar to those found in (B. Tobalske, 1995). Birds were kept on a heated surface and anesthetized using an isoflurane-oxygen mixture with a 5% induction dose and maintained at a surgical plane using a 1-3% dose, depending on the individual. Once individuals demonstrated a surgical plane of anesthesia was reached through reduced reflexes to leg pull-back and pinching, two 1cm incisions were made into the skin, one dorsally above the upper spine and the other ventrally off-center above the pectoralis. The implant and wires were then inserted under the skin via the dorsal incision and the wires threaded to the ventral side through the pectoral incision. Two final incisions were made on the ventral surface of the wing, one 1cm incision between the biceps and the TPB, and one 0.5cm incision above the TPL, and the three wing-muscle EMG wires were threaded above their respective muscles. All wires were implanted intramuscularly with the aid of a hypodermic needle and sutured to muscle fascia with 5-10mm of slack between suture point and wire insertion. All ventral incisions were sutured closed, the back plug sutured to the spinal intervertebral ligaments, and the dorsal incision sutured closed snugly around the back plug.  During data collection, signals from the bipolar electrodes were carried to an AM Systems Model 1700 differential amplifier (AM Systems, WA, USA) via a trailing wire, attached to the upper rear portion of the wind tunnel test section to reduce weight impact on birds. All 18  channels were amplified with a 1,000x gain and filtered with a 10-5000Hz band-pass filter and a 60Hz notch filter, and a common amplifier ground for all four channels was established via the ground wire inside the bird’s implant. Amplifier output was recorded by an ADinstruments Powerlab analog-digital converter at 10,000Hz through Labchart software (ADinstruments, CO, USA). 2.2.4 High-Speed Kinematics To collect wing kinematics, the wind tunnel test section was filmed at 250Hz from four high-speed cameras positioned around the test section (Photron USA Inc., CA: FASTCAM PCI-1024, SA3, and Mini AX100 cameras). The position of markers painted on the shoulder, elbow, and wrist joints with white water-soluble paint was then manually tracked in the footage and converted into a 3-dimensional representation using Direct Linear Transformation (DLT) using the Matlab script DLTdv7 (Hedrick, 2008), with DLT coefficients and global coordinate frame generated using the Matlab program easyWand5 (D. H. Theriault et al., 2014). In these calibrations, the global Z axis was always oriented with gravity by tracking the motion of a dropped ball, though direction of X and Y axes varied between level and angled wind tunnel conditions. As the position of the wind tunnel and cameras changed when the tunnel was angled, calibrations were performed twice for each bird—one calibration for when the tunnel was angled, and one for when the tunnel was level.  2.2.5 Morphological Measurements All individuals used in the study were euthanized immediately following experimentation for measurements of static morphology. Subjects were weighed, and a series of wing morphology measurements such as wing span and chord length were made. Measurements of skeletal morphology such as the length of the humerus, radius, and ulna were then made, and a 19  series of measurements of the elbow flexor muscles were performed. The distance of muscle attachment points from the elbow center of rotation was measured, as well as the angle of the attachment point when the elbow was positioned at 90°. Muscles were then dissected, immediately weighed on a precision scale to the nearest 0.1 mg, and fiber lengths were measured. All measurements of distance and length were performed three times using digital calipers and then averaged. With these measurements of muscle mass, length, and attachment point and attachment angle, approximations of maximum possible force and torque output for each muscle can be calculated. Using the equation  / = 0 cos 45 ∗ 67  (2)  where A is muscle area, M is muscle mass, α is fiber pennation angle, 67 is fiber length, and 5 is skeletal muscle density, known to be 1,060 kg/m3 in most vertebrates (Pollock & Shadwick, 1994). From muscle area, maximal muscle force and torque can be calculated using the equations  8 = 9/ (3)  " = 9/: sin= (4) where F is muscle force, T is torque about the elbow, 9 is maximal muscle stress, and r is the distance of attachment from the elbow center of rotation, and = is the angle of attachment on the forearm. Maximal stress was assumed for all muscles to be the physiologically-common value of 300kPa (Josephson, 1993; Pollock & Shadwick, 1994). Note that all muscles were assumed to be fusiform and parallel-fibered, with a pennation angle of 0°. This simplification is accurate for the TPB and TPL, but likely underestimates maximal force for the unipennate biceps. The exact 20  degree of this underestimation is uncertain without precise measurement, but computed tomography scans of the starling biceps show fiber pennation angles of approximately 20°-40° and fiber lengths approximately 60-80% of main muscle belly length (Sullivan et al., 2019). From these estimates, the biceps measurements of area, force, and torque in this manuscript would be at most 1.5x less than true values.  From measurements of muscle fiber length, lever arm, and angle of attachment, terms proportional to the relative energetic cost of producing a given force or torque can be approximated. Not accounting for physiological differences in muscle fiber types, the energy usage of a muscle will scale with the number of cells in a muscle, which is proportional to both the muscle volume and mass. The energetic cost for a given force can then must be proportional to fiber length, given that  >?@A+B	DE@F,AG@ ∝ IJEG.@	K,.JL@IJEG.@	MA@- ∝ F*N@A	O@?+PQR . (5) A similar process can be used to approximate energy cost for a given torque:  >?@A+B	DE@S,ATJ@ ∝ IJEG.@	K,.JL@U∗MA@-∗A VWXY ∝ F*N@A	O@?+PQA VWXY . (6) It is important to note that both of these calculations rely on the assumption that volume/area is equal to fiber length, an assumption that is not true for pennate or non-parallel-fibered muscle. These calculations can also only comment on how energy use for a given force or torque varies based on architecture alone, as different physiological fiber types can have dramatically different energy consumption properties.  21  2.2.6 Data Analysis  Two out of the seven birds (birds #2 and #4) suffered implant failures that resulted in no observable signal on at least 2/4 muscles observed during the experiment, warranting the removal of their data from the experiment. These implant failures are likely due to either a loss of nonconductive enamel coating on the electrode wires, disconnection of electrodes from the implant, or migration of electrode tips outside of the muscle belly. Of the remaining 5 birds, 4 have had kinematics from a sufficient number of wingbeats digitized at the time of writing, so the following data presented in this thesis comprises data collected from these 4 birds (birds #1, #3, #5, and #6, numbered by data collection date). Data from wing weight treatments at time of writing has only been digitized in three of these individuals, birds #3, #5, and #6.   Data processing followed the process illustrated in Figure 2.1. Raw voltage measurements from each bipolar electrode were full-wave rectified (equivalent to the absolute value of the signal) and divided by the maximum rectified value observed for a given electrode in a given bird. Prior to rectification, the frequency spectra of all EMG signals were checked for high-power 60Hz harmonics caused by electromagnetic interference from power mains—trials showing this interference were filtered using a zero-phase IIR comb filter with a Q factor of 35 and attenuation notches every 60Hz to remove these artifacts with minimal change to the EMG signal.   After normalization and rectification, envelopes were fit to EMG signals to generate a useful metric of the approximate shape and amplitude of the EMG signals, and the resulting envelopes were passed through a double-threshold detector similar to that of (Bonato et al., 1998) to obtain a binary metric of whether a muscle was active at a given time. Each period of activity (which I will refer to as a “pulse”) required a minimum duration of 5ms to be counted, 22  and pulses adjacent by 5ms or less were connected following the methods outlined in (Bonato et al., 1998). Envelopes comprised a spline interpolation fit to local maxima separated by at least 1ms; this peak-based method has distinct differences from low-pass filtering traditionally used to obtain EMG envelopes, as in (Roberts & Gabaldón, 2008; Zajac, 1989). A peak-based envelope tracks the maximum amplitude and shape of the EMG trace instead of removing high-frequency components of the signal, which allows little to no lag or delay between signal and envelope onset and offset timing, but is more sensitive to high-frequency bursts and peaks.   Wingbeat cycles were distinguished based on cyclic changes in elbow angle using a peak-finding algorithm applied to cubic-spline interpolated elbow angle data smoothed using a 3rd order zero-phase discrete butterworth filter. This process of interpolation and smoothing of elbow angle produces an approximation of elbow angle at the same sample rate as the EMG data (10kHz), allowing the separation of wingbeat cycles to occur at points where there is no kinematic data present, resulting in a more accurate estimate of wingbeat cycles. With wingbeat cycles separated, time can be normalized as a fraction of wingbeat cycle and all cycles overlaid to observe larger trends in muscle timing and activity, as shown in Figure 2.1C.  Downstroke periods were determined by performing a Principal Components Analysis (PCA) on the vector PWS defined as the difference between the (x,y,z) positions of the wrist marker PW and the butterworth low-pass filtered (x,y,z) positions of the shoulder marker PSfilt. Filtering the shoulder marker with a low-pass filter (2nd order filter, 17.5Hz cutoff) roughly approximates the bird’s body position and motion by removing the per-wingbeat oscillations of the shoulder position. With the knowledge that the majority of the variance of wrist position is from upstroke-downstroke motion and a metric of wrist position relative to body motion PWS, the maximum and minimum values of a weighted sum of Principal Components (PCs) 1 and 2 of 23  PWS can be used to define the upstroke-downstroke (US-DS) and downstroke-upstroke (DS-US) transitions (see Figure 2.2). As the X and Y axes of the global coordinate system are not consistent between easyWand5 calibrations, a separate PCA was performed on data from each calibration.   During descending flights, intermittent short glides were occasionally observed. To prevent these glides from being counted as wingbeat cycles and analyzed with the flapping wing data, periods of 100ms or greater where elbow angle changed less than 20° and wing elevation (defined as the z position of the wrist marker) was similarly constant were set aside into a separate dataset. This dataset was deemed too small for analysis in this study, but data from all of these glide sections is presented in the appendix.   Figure 2.1. Analysis process for EMG and kinematics data. A. Raw EMG and elbow angle data taken from three concurrent wingbeats. B. The same data with EMG signals full-wave rectified, normalized by maximum value observed for a given bird and electrode, fit with a peak-based envelope shown in red, and with periods of activity demarcated using a double-threshold detector similar to (Bonato et al., 1998) indicated by shaded regions. Elbow 24  angle is colored by cycle using the methods described in the text. C. Normalizing time for each cycle allows all data to be overlaid, clarifying cyclic patterns of muscle activation relative to elbow angle.  Figure 2.2. Analysis process for extracting wingbeat downstroke. A. Example data from one individual of wrist position vector PWS in a 3D scatterplot. B. The same data plotted on PC1 and PC2 with downstroke region highlighted in blue.  2.3 Results Figure 2.3 displays the morphology measurements most relevant to this study, including muscle areas, fiber lengths, theoretical force and torque production, and the relative energetic costs of producing force and torque. Though measurements of biceps area (and thus force and torque) are likely to be slightly underestimated, it is evident from Figure 2.3 that despite large differences in force production, all three elbow flexors in the starling produce similar levels of maximal torque owing to the longer lever arms of the TPL and TPB, with the TPL producing the highest torque, followed by similar levels of torque production from the biceps and TPB.  Figure 2.3.C. displays the approximate energy cost per force and energy cost per torque, based on muscle architecture, for all three elbow flexors in the starlings measured. All three muscles have similar values of architectural energy cost per force, though the long-fibered TPB 25  demonstrates the highest cost per force values. Architectural energy cost per torque, however, is dramatically different between the three muscles, with the TPL displaying an order of magnitude lower cost per torque than the biceps and the TPB showing a consistently lower cost per torque. This is largely due to the long lever arms of the TPB and TPL relative to the biceps, as all three muscles have fiber lengths of similar magnitude.   Figure 2.3. Morphology measurements of starling propatagial muscles. A. Muscle fiber length plotted against cross-sectional area. B. Muscle maximum theoretical torque plotted against cross-sectional area, assuming stress of 300kPa. Maximum force and torque for both A and B is calculated assuming stress of 300kPa. C. The ratio of fiber length to angle-adjusted lever arm (proportional to the energy cost of torque production) plotted against fiber length (proportional to energy cost of force production). In all plots connecting lines denote individuals. 26   Figure 2.4 displays all of the EMG and elbow angle data collected from flights performed when the wind tunnel was angled 20° downwards or was level (level flights with wing weights are not included). Variations between different individuals are present, including different minimum and maximum elbow angle values for each individual: This is likely due to slightly different positioning of tracking markers per individual. Similarly, EMG activity and timing for each muscle shows variation between individuals, owing to several likely sources. Individual birds may demonstrate different muscle recruitment patterns, and simple differences in electrode noise floor and signal quality are likely (TPB recordings from bird 3, for instance, show a noticeably higher noise floor). Intramuscular EMG is relatively robust and repeatable to differences in electrode placement (Chapman et al., 2010), but varying electrode proximity to motor end-plates and potentially different fiber types within a muscle can also lead to signal variance.  Despite some variation between individuals, overall patterns of muscle timing and activity relative to elbow angle are markedly consistent. All three elbow flexors show their peak activity around the US-DS transition as elbow angle is increasing (although the biceps in bird 6 shows peak activity closer to the mid-downstroke). The TPL consistently demonstrates two pulses of activity per wingbeat, with the second pulse being more variable in timing between individuals, typically of lower magnitude, and coinciding with the DS-US transition as elbow angle is decreasing. The TPB shows the most consistent pattern of activity and timing across individuals and wingbeats, with a single two-lobed pulse near the US-DS transition as elbow angle is decreasing. The biceps shows the longest duration of activity, with an initial peak of 27  activity near the US-DS transition that gradually decreases in magnitude over the course of the downstroke.  Additional patterns of relative timing are consistent across individuals, most notably that the relative onset timing of the three elbow flexors always follows the sequence of biceps, TPB, and then TPL, separated by gaps of roughly 5% of cycle length.  Figure 2.4. Elbow angle and EMG activation of descending and level flights. Data from all wingbeat cycles (n=254 wingbeats) overlaid following process shown in Figure 2.1, separated into columns by individual. Colored lines of mean EMG activation for descending and level flights drawn by averaging EMG envelope values at 500 evenly-spaced bins. Gray shaded region indicates mean downstroke start and end for all trials, lighter gray shading indicates 1 standard deviation of downstroke start and end. To improve readability of events at the start and end of cycles, data is repeated twice.  28  The effects of wing weights on kinematics and muscle activity are summarized in Figure 2.5. Note that wing weight flights from bird #1, whose data is present in other figures, have not yet been digitized at time of writing. Of the three birds with sufficient wing weights data, only one (bird #5) shows notable differences in kinematics and muscle activity between flights with and without wing weights. This individual demonstrates increased pectoral activity and increased biceps activity throughout the downstroke and subsequent period of decreasing elbow angle in the upstroke, as well as demonstrating noticeably higher wrist elevation during the upstroke. All three individuals demonstrated no major changes in elbow angle when wing weights were added, and 2/3 individuals showed no major changes in muscle activity or wrist kinematics due to the addition of wing weights.  A summary of the wingbeat cycle frequencies observed from each individual and flight type, as well as the relationship between cycle frequency and parameters of muscle activity is shown in Figure 2.6. All individuals averaged showed a wingbeat frequency of 13.0±1.7 Hz in descending flights, 17.0±1.4 Hz in level flights without wing weights, and 16.9±1.0Hz in wing weight flights (values quoted as mean±s.d.), with bird #5 showing notably lower wingbeat frequencies for all level and wing weight flights.  Figure 2.6 also includes plots of two parameters against cycle frequency, the mean amplitude of each period of activity, or “pulse”, (Figure 2.6.B) and the total time spent active in a cycle (Figure 2.6.C), where periods of muscle activity are determined using the double-threshold detector described in the methods section. These plots visually indicate that the mean pulse amplitude of all muscles positively covaries with cycle frequency for all muscles, but total duration of activation does not show the same level of correlation for all muscles, with only the biceps and pectoralis demonstrating robust positive correlation with frequency. This is further 29  corroborated with Tables Table 2.1 and Table 2.2, which provide Pearson correlation coefficients between mean pulse amplitude and total duration active per cycle. All muscles show a consistently high positive correlation between mean pulse amplitude and cycle frequency of 0.4 or greater, but duration more weakly covaries with cycle frequency for all three elbow flexors, with the TPB and TPL showing the weakest correlations.    30   Figure 2.5. Effect of wing weights on kinematics and muscle activity. A. X and Z data of PWS wrist position vector for level and wing weight flights of all individuals with wing weight flights digitized (n=184 wingbeats). Negative X direction corresponds to anterior direction, and kinematic loops occur in counter-clockwise direction. B. Elbow angle mean ± s.d. for level and wing weight flights, with mean binned EMG activation for all muscles (n=184 wingbeats). Dark gray bar corresponds to mean downstroke start and end, with light gray shading denoting one standard deviation. Data is repeated twice to improve readability of events at the start and end of cycles.  31   Figure 2.6. Wingbeat frequency and its effects on amplitude and duration of muscle activation. A. Boxplots with background scatter of cycle frequency, in Hz, of wingbeats for all individuals, separated by flight type (n=322 wingbeats). Total for all flights is shown on top row. Box midline corresponds to median, box extents correspond to 1st and 3rd quartiles, and horizontal lines correspond to largest values no further than 1.5x Interquartile Range (IQR) B. Mean EMG envelope amplitude during each continuous period of muscle activity (pulse) plotted against cycle frequency. Active regions were determined using double-threshold detection. C. Total duration of muscle activity in each cycle, in milliseconds, plotted against cycle frequency. Columns are separated by individual for panels B and C.  32   Bird 1 Bird 3 Bird 5 Bird 6 All Data  0.618 0.567 0.616 0.681 0.665 Biceps 0.735 0.803 0.804 0.702 0.714 TPB  0.521 0.667 0.511 0.725 0.585 TPL 0.399 0.564 0.697 0.554 0.435 Pectoralis Table 2.1 Pearson correlation coefficients between mean pulse amplitude and cycle frequency.  Bird 1 Bird 3 Bird 5 Bird 6 All Data  -0.179 -0.558 0.719 0.796 0.280 Biceps 0.358 0.332 0.140 -0.111 0.142 TPB  0.847 0.310 -0.025 -0.606 0.026 TPL 0.613 0.220 0.755 0.418 0.466 Pectoralis Table 2.2 Pearson correlation coefficients between total duration active per cycle and cycle frequency.     Figure 2.7 displays elbow angle and mean binned EMG amplitude plotted for each individual in the same format as Figures Figure 2.4 and Figure 2.5 B, but with data grouped into three cycle frequency groups of roughly equal size. EMG amplitude for all muscles increases as cycle frequency increases, with cycles below 14Hz displaying both a lower amplitude and more diffuse timing. Notable shifts in muscle timing are observed between cycles 14-17Hz in frequency and cycles of frequency 17Hz or greater, with the onset of the biceps occurring earlier for all individuals and similar shifts to earlier activity for many individuals in the TPB and TPL.  33    Figure 2.7. Elbow angle and EMG activation grouped by wingbeat frequency.  Elbow angle and binned means of EMG activation, grouped into three groups of wingbeat frequency less than 14Hz (yellow), between 14 and 17Hz (blue), and 17Hz or greater (purple), wing weight trials excluded (n=254 wingbeats). Data from each individual is separated into columns. Gray shaded region indicates mean downstroke start and end for all trials, lighter gray shading indicates 1 standard deviation of downstroke start and end. To improve readability of events at the start and end of cycles, data is repeated twice.   34  2.4 Discussion 2.4.1 Muscle Morphology and Energetic Cost Based on morphology alone, from Figure 2.3 B it’s apparent that all three elbow flexors produce similar maximal torque about the elbow, with the diminutive size of the TPB and TPL compensated for by their long lever arms. In contrast to the previous literature, which has treated the TPL and its associated leading-edge tendon as likely for sensory and aerodynamic functions (Dial, 1992a), it’s clear based on morphology alone that the TPL, at least in the starlings measured here, can have as notable of an influence on wing dynamics as the biceps. The TPB has similarly been claimed in anatomical literature as primarily functioning to tense the propatagium or passively limit elbow extension (Brown et al., 1995; Dial, 1992a). My findings don’t conflict with these interpretations, but merely clarify that these muscles, by virtue of their longer lever arms, produce similar torque about the elbow to the biceps.   The long lever arms of the TPB and TPL have the additional effect of providing these muscles an extremely low cost of torque production compared to the biceps (Figure 2.3 C). While the metabolic expenditures of the intrinsic wing muscles is dwarfed by the energetic demand of the primary flight muscles (in most birds the pectoralis alone occupies ¼ of a bird’s bodyweight), the potential reduction in energetic cost per wingbeat enabled by the TPB and TPL does not appear to be insignificant. It is important to remember that an energetic cost is incurred purely by the mass of the wing (Hedrick et al., 2004)—producing the same torque with significantly less distal mass reduces both metabolic cost as well as the inertial costs of flapping. Authors have speculated that the evolution of vertebrate wings has coincided with the shift of muscle mass proximally to reduce inertial cost, either through longer tendons (Konow et al., 2015) or through coupling of proximal muscles with distal degrees of freedom (R J Vazquez, 35  1994). Both the TPB and TPL demonstrate these hallmarks as muscles with long in-series tendons, located proximally near the shoulder but acting on multiple degrees of freedom, with the TPB flexing the elbow and extending the manus. This suggests that the TPB and TPL may have a particular utility when the inertial costs of flapping flight are higher, but further work is required.   2.4.2 Functional Implications of Muscle Timing It’s clear from Figure 2.4 and Figure 2.5 that all three muscles are primarily active as elbow angle is increasing, despite being elbow flexors. While there are many disconnects between EMG and force that impair interpretation of muscle force output from EMG (Roberts & Gabaldón, 2008), the timing is suggestive that all three muscles are decelerating the elbow at the beginning of the downstroke. This fits with many other findings: Similar activity has been observed in the pigeon biceps (Robertson & Biewener, 2012). It is also known that centrifugal forces will naturally extend the elbow of a flapping wing during a downstroke (Stowers & Lentink, 2015). In flapping motions as dynamic as those observed in most birds, to avoid overextension elbow flexors would have to produce torque (and very likely negative work) during the first portion of the downstroke to counteract this outwards swing. This also fits in with a larger pattern that distal muscles tend more to contract near-isometrically, tuning kinematics more than producing work and power (Biewener, 2011; Gillis & Biewener, 2001).   Not all muscle activity suggested opposition to elbow motion, though, as the biceps and TPL in were active during the late downstroke as elbow angle decreases, with the biceps active throughout the downstroke with one continuous period of activation compared to the two focused bursts of activity shown by the TPL. That these two pulses of activity happened to 36  coincide with the start and end of the downstroke is likely not coincidental. Contraction at two points in a cycle when the MTU of the TPL is at nearly equal length is highly suggestive of elastic energy storage and release, with the first contraction serving to increase the force on the elastic tendon, biasing it to a greater length and storing a greater amount of energy, while the second contraction serves to modulate the timing and release of this stored elastic energy. This theory would explain why the timing and amplitude of the second pulse is so variable (see Figure 2.4); even small shifts in muscle timing and force could adjust the timing and rate of energy release from the leading-edge tendon. Previous work has shown that a significant inertial cost is incurred during the upstroke due to limited transfer of kinetic energy to surrounding air (Hedrick et al., 2004), and that this inertial cost could be substantially reduced with elastic energy storage mechanisms. The supracoracoideus muscle has been shown to demonstrate significant elastic energy storage during the upstroke (B. W. Tobalske & Biewener, 2008), and my findings further suggest that the leading-edge tendon of the TPL may also contribute to this elastic energy storage, as the TPL is most active at the US-DS and DS-US transitions where the most kinetic energy would be lost. Given that the majority of the strain energy absorbed by the leading-edge tendon would result from the motion of the wing and not from contraction of the TPL (fiber length of 10-15mm, compared to tendon length change of >20mm per wingbeat), the TPL’s role is likely one of energy conservation or power attenuation rather than power amplification (Roberts & Azizi, 2011).  The timing and activity of the TPB was particularly illuminating in its consistency, both between cycles and individuals. No secondary pulses of activity or noticeable shifts in timing were observed for any of the cycles recorded in this study, which is suggestive of the role the TPB plays in wing morphing. Given that the TPB is a multi-joint muscle capable of both elbow 37  flexion and manus extension, it makes sense that it would have highly constrained timing in flight. A muscle acting on several major degrees of freedom is constrained to activity only at times when motion in both of those degrees of freedom is warranted, and the time at which the TPB is active, the US-DS transition, is likely the only part of the flapping cycle when both elbow flexion torque and manus extension torque are required. Centrifugal forces require flexion torque to sustain elbow angle, and the manus is required to go from a fully flexed state in the upstroke to a fully extended state in the downstroke rapidly. Enabling a rapid unfurling of the distal wing is likely a major function of the TPB; from its morphology in Figure 2.3 it’s clear the TPB has the highest average fiber length, and thus contraction velocity, of the muscles measured in this study, and the in-series tendon of the TPB being shorter than the TPB’s fiber length suggests a power amplification role (Roberts & Azizi, 2011). The connection of the TPB with the EMR, too, should allow the TPB to stretch the EMR and decrease its effective length in addition to the EMR’s own contraction, effectively boosting the contraction velocity of the EMR and thus the speed at which the manus can be extended at the US-DS transition. Further work is needed to study the nature of the coupling of these two muscles and the relative timing of the TPB and EMR, however, to evaluate these theories.  2.4.3 Evaluation of Control Roles  The main hypothesis of this study was that the elbow flexors would demonstrate specific roles reflected by their attachment points and architecture, such that the biceps would largely act in position control, while the TPB and TPL would act more in torque control. The increase in wing inertia by addition of wing weights was intended to test this, by requiring either a change in elbow torque or a change in elbow kinematics due to the conservation of momentum. Apparent from Figure 2.5, however, is that for most individuals in this study neither elbow kinematics nor 38  muscle activity and timing (and, by implication, muscle torque) were significantly altered by addition of wing weights. One individual showed no change in elbow kinematics and an increase in only biceps activity, which supports the notion that the biceps is acting in position control while the TPB and TPL are guided by primarily torque control, but this was not repeated outside of this individual. This individual, it should be noted, also demonstrated generally different flight patterns, including higher average descending flight wingbeat frequencies and lower average level flight wingbeat frequencies (Figure 2.6).   As Figure 1.2 in the introduction illustrates, the most likely outcomes for no observed changes in kinematics and muscle activity are that the wing weights were too small to have influenced wing inertia, or some other muscle not measured in this study provided the additional torque and energy required to maintain elbow kinematics. This makes it difficult for the results of the wing weight treatment in this study to contribute meaningful knowledge to the control roles of the muscles studied. Though one individual’s results were suggestive of my initial hypothesis, the effect has no repeatability or statistical power.   2.4.4 Evaluation of Control Roles through Wingbeat Frequency Though observations of control using a controlled treatment were inconclusive, observing how muscle activity and timing changed with cycle frequency (Figure 2.6 and Figure 2.7, Table 2.1 and Table 2.2) can provide an indirect method to study how the three elbow flexors are recruited and controlled. From Figure 2.6 B and C, it is clear that all muscles increased mean activity with increasing cycle frequency, but duration of activity showed a less clean trend with cycle frequency. The duration of biceps and pectoralis activation demonstrated relatively strong positive correlation with cycle frequency, but the TPB and TPL showed no major correlations 39  between duration of activity and cycle frequency (Table 2.1 and Table 2.2). There are different explanations as to why mean activity correlates with cycle frequency, and it is important to note that this apparent correlation may simply be due to the reduced energetic output required for descending flight compared to level flight. But the lack of strong correspondence between changes in mean activity and duration indicates that changes in duration with cycle frequency and flight type are not simply due to higher amplitude signals altering the behavior of the threshold detection used.  It is well-understood that as cycle frequency increases, without increasing active duration or altering strain patterns work per cycle must decrease, as increased muscle strain rate results in reduced force production (Josephson, 1993). As the biceps and pectoralis have little in-series elasticity, their strain patterns are directly coupled with skeletal kinematics, so to maintain or increase their work production with increasing cycle frequency, their active duration must increase. Increase in active duration of the avian pectoralis with cycle frequency has been previously observed (Jackson et al., 2011), but observation of similar behavior in the biceps is a novel finding. The in-series elasticity of the TPB and TPL potentially allow limited decoupling of the connection between cycle frequency and muscle work output, which may explain their limited change in active duration with cycle frequency, but it is important to also consider that the TPB and TPL may have control roles that do not optimize for their work production. If the TPL is primarily tuning the strain energy storage of its long tendon, for instance, its activity would be more likely optimized for specific timing relative to its tendon length over work production.     40  Chapter 3: Conclusion The results of this study indicated that the propatagial musculature and elbow flexors of the European starling demonstrate unique roles in flight. While the primary goal of the study to parse out the effects of muscle architecture and attachment point on control roles between the biceps, TPB, and TPL proved inconclusive, new findings and information that suggest previously unsuggested functions of these muscles were uncovered.  Quantification of the morphology of the biceps, TPB, and TPL demonstrated that despite their dramatically different masses and cross-sectional areas, all three elbow flexors are capable of producing similar levels of maximal torque about the elbow. The long lever arms that allow the comparatively smaller propatagial muscles to produce similar torque to the biceps have the additional effect of reducing the potential cost of producing a given torque. This and the proximal concentration of mass permitted by the TPB and TPL suggest the propatagial musculature, among other potential functions, allow flapping flight to be more energetically favorable.  The biceps was found to show activity similar to that measured in pigeons (Dial, 1992a; Robertson & Biewener, 2012), decelerating and then stabilizing the elbow during the initial to mid- downstroke, with bouts of activity into the late downstroke. Our findings and the previous work studying the pigeon biceps form a consistent body of literature that suggests the muscle has a variable and stabilizing role in controlling wing shape, contracting near-isometrically for much of the wingbeat cycle but with potentially negative work output in the initial downstroke and positive work output in the late downstroke. As a new result, this study found that the duration of biceps activity increases with wingbeat cycle frequency, suggesting that the biceps is being controlled to maintain or increase work output as cycle frequency increases.  41   This study is the first to perform EMG measurements on the propatagial musculature of the TPB and TPL in a non-pigeon species, a notable contribution given the major differences between the wing anatomy of pigeons and other avian species. The timing of the TPB and TPL observed in the European starlings of this study was dramatically different from the timings observed in pigeons in (Dial, 1992a), and suggested the two muscles demonstrate unique roles in controlling wing shape. The long tendon of the TPL and its two pulses of activation at the US-DS and DS-US transitions suggest it has a role in storing and releasing energy from wing motion, with the activation of the TPL mainly serving to modulate the rate and timing of this energy release. The precise timing and coupled attachment of the TPB with a manus-extending muscle, the EMR, suggests it serves a dual function to decelerate the elbow and rapidly extend the manus at the start of the downstroke. This is further supported by the long fibers and medium-length tendon of the TPB, which suggest a high contraction velocity and potential for power amplification. Both the TPB and TPL demonstrated little change in duration of activity with increasing wingbeat cycle frequency, suggesting both muscles operate in roles that do not   3.1 Limitations While this work has provided new and interesting insights, this thesis has a number of limitations that must be addressed. As has been previously mentioned, it is difficult to accurately interpret muscle output from EMG (Roberts & Gabaldón, 2008), and while the EMG data collected in this work presents a number of interesting potential muscle roles and functions, all of these results require further validation and study of the actual force and torque outputs of the muscles studied. Given how little previous study has been dedicated to the propatagial muscles, 42  this study should serve primarily as a guide to the most interesting aspects of this muscle system to be studied in future work.  The morphology measurements and associated conclusions, too, should be approached with some caution. Maximal muscle forces and torques are likely underestimated for the biceps, due to no consideration of pennation, and estimates of relative energetic cost for a given force or torque are informative only to the effects of muscle architecture and attachment point on the energetic costs of force and torque production; different fiber types and recruitment patterns can dramatically alter energy consumption between otherwise similar muscles.  Finally, the attempt to observe muscle control roles using a treatment of attached weights on the wing proved inconclusive, potentially owing to weights that were not heavy enough to elicit a sufficient response or compensation of additional inertia from unobserved muscles. Further observation of potential control roles and shifts in muscle work production was performed through analysis of changes in muscle active duration with wingbeat frequency, but these results also warrant caution. Unobserved quantities such as differences in muscle fiber types, changes in fiber recruitment, and changes in muscle force output with cycle frequency can alter work output without any changes to active duration; though the observed results very likely indicate a shift in biceps work output with frequency and lack of similar shifts in the TPB and TPL, further work is needed.     43  3.2 Future Work Future work in understanding the control roles of these and other muscles should pursue either a similar experiment with more thorough controls, or a different approach to parse individual muscle contributions to dynamics. Performing the same experiment of this study but with treatments of varying weight in randomized order would systematically remove the possibility of weights being too small to elicit changes, as well as prevent the observation of effects due simply to acclimation or fatigue. Knockout methods, like the denervation treatment used in (Dial, 1992b), are a promising avenue for study. Nondestructive and impermanent chemical knockout of redundant muscles is particularly promising, as injections of Ca2+ channel blockers could provide temporary and involuntary relaxation of targeted muscles. This would enable observations of effects loss of targeted muscles has on kinematics and the control of redundant muscles with more robust, repeatable, and humane experimental designs.  The results of this study indicated that the TPL and TPB have potentially unique roles, with the TPL potentially acting to store and release elastic energy and the TPB acting to enable rapid extension of the manus with a proximally-located muscle. Future work should investigate these results in detail with a host of approaches. In situ or in vivo approaches to measure the length change of the TPL and the leading-edge tendon would further clarify its role in flight, and precise measurement of the amount of elastic energy stored by the leading-edge tendon would be necessary. The TPB-EMR coupling is highly unique, and whether the TPB serves to alter the effective contraction speed of the EMR could be studied through in situ work-loop approaches similar to (J. S. Theriault et al., 2019) or modeling studies similar to  (Rosario et al., 2016).   44  Bibliography  Altshuler, D. L., Bahlman, J. W., Dakin, R., Gaede, A. H., Goller, B., Lentink, D., Segre, P. S., & Skandalis, D. A. (2015). The biophysics of bird flight: functional relationships integrate aerodynamics, morphology, kinematics, muscles, and sensors. Canadian Journal of Zoology, 93(12), 961–975. Baliga, V. B., Szabo, I., & Altshuler, D. L. (2019). Range of motion in the avian wing is strongly associated with flight behavior and body mass. Science Advances, 5(10), eaaw6670. Berge, J. C. vanden. (1970). A comparative study of the appendicular musculature of the order Ciconiiformes. American Midland Naturalist, 289–364. Bernstein, N. (1966). The co-ordination and regulation of movements. The Co-Ordination and Regulation of Movements. Biewener, A. A. (2011). Muscle function in avian flight: achieving power and control. Philosophical Transactions of the Royal Society B: Biological Sciences, 366(1570), 1496–1506. Biewener, A. A. (2016). Locomotion as an emergent property of muscle contractile dynamics. Journal of Experimental Biology, 219(2), 285–294. Biewener, A. A., & Roberts, T. J. (2000). Muscle and tendon contributions to force, work, and elastic energy savings: a comparative perspective. Exerc Sport Sci Rev, 28(3), 99–107. Bonato, P., D’Alessio, T., & Knaflitz, M. (1998). A statistical method for the measurement of muscle activation intervals from surface myoelectric signal during gait. IEEE Transactions on Biomedical Engineering, 45(3), 287–299. 45  Brown, R. E., Baumel, J. J., & Klemm, R. D. (1994). Anatomy of the propatagium: the great horned owl (Bubo virginianus). Journal of Morphology, 219(2), 205–224. Brown, R. E., Baumel, J. J., & Klemm, R. D. (1995). Mechanics of the avian propatagium: Flexion-extension mechanism of the avian wing. Journal of Morphology, 225(1), 91–105. Brown, R. E., Butler, J. P., Rogers, R. A., & Leith, D. E. (1994). Mechanical connections between elastin and collagen. Connective Tissue Research, 30(4), 295–308. Brown, R. E., & Cogley, A. C. (1996). Contributions of the propatagium to avian flight. Journal of Experimental Zoology, 276(2), 112–124. Chang, Y.-H., Ting, S.-C., Liu, C.-C., Yang, J.-T., & Soong, C.-Y. (2011). An unconventional mechanism of lift production during the downstroke in a hovering bird (Zosterops japonicus). Experiments in Fluids, 51(5), 1231–1243. Chapman, A. R., Vicenzino, B., Blanch, P., Knox, J. J., & Hodges, P. W. (2010). Intramuscular fine-wire electromyography during cycling: repeatability, normalisation and a comparison to surface electromyography. Journal of Electromyography and Kinesiology, 20(1), 108–117. Cheney, J. A., Konow, N., Middleton, K. M., Breuer, K. S., Roberts, T. J., Giblin, E. L., & Swartz, S. M. (2014). Membrane muscle function in the compliant wings of bats. Bioinspiration & Biomimetics, 9(2), 25007. d’Azyr, V. (1773). Collection Académique. L’Academie Royale des Sciences de Paris. Dial, K. P. (1992a). Activity patterns of the wing muscles of the pigeon (Columba livia) during different modes of flight. Journal of Experimental Zoology, 262(4), 357–373. Dial, K. P. (1992b). Avian forelimb muscles and nonsteady flight: can birds fly without using the muscles in their wings? The Auk, 874–885. 46  Dial, K. P., Goslow Jr, G. E., & Jenkins Jr, F. A. (1991). The functional anatomy of the shoulder in the European starling (Sturnus vulgaris). Journal of Morphology, 207(3), 327–344. Fisher, H. I., & Goodman, D. C. (1955). The myology of the whooping crane, Grus americana 24. Illinois Biological Monographs; v. 24, No. 2. Gillis, G. B., & Biewener, A. A. (2001). Hindlimb muscle function in relation to speed and gait: in vivo patterns of strain and activation in a hip and knee extensor of the rat (Rattus norvegicus). Journal of Experimental Biology, 204(15), 2717–2731. Harvey, C., Baliga, V. B., Lavoie, P., & Altshuler, D. L. (2019). Wing morphing allows gulls to modulate static pitch stability during gliding. Journal of the Royal Society Interface, 16(150), 20180641. Hedrick, T. L. (2008). Software techniques for two-and three-dimensional kinematic measurements of biological and biomimetic systems. Bioinspiration & Biomimetics, 3(3), 34001. Hedrick, T. L., Usherwood, J. R., & Biewener, A. A. (2004). Wing inertia and whole-body acceleration: an analysis of instantaneous aerodynamic force production in cockatiels (Nymphicus hollandicus) flying across a range of speeds. Journal of Experimental Biology, 207(10), 1689–1702. Herzog, W. (2017). Skeletal muscle mechanics: questions, problems and possible solutions. Journal of Neuroengineering and Rehabilitation, 14(1), 98. Hieronymus, T. L. (2016). Flight feather attachment in rock pigeons (Columba livia): covert feathers and smooth muscle coordinate a morphing wing. Journal of Anatomy, 229(5), 631–656. 47  Hudson, G. E., & Lanzillotti, P. J. (1955). Gross anatomy of the wing muscles in the family Corvidae. The American Midland Naturalist, 53(1), 1–44. Jackson, B. E., Tobalske, B. W., & Dial, K. P. (2011). The broad range of contractile behaviour of the avian pectoralis: functional and evolutionary implications. Journal of Experimental Biology, 214(14), 2354–2361. Josephson, R. K. (1993). Contraction dynamics and power output of skeletal muscle. Annual Review of Physiology, 55(1), 527–546. Konow, N., Cheney, J. A., Roberts, T. J., Waldman, J. R. S., & Swartz, S. M. (2015). Spring or string: does tendon elastic action influence wing muscle mechanics in bat flight? Proc. R. Soc. B, 282(1816), 20151832. Lapsansky, A. B., Igo, J. A., & Tobalske, B. W. (2019). Zebra finch (Taeniopygia guttata) shift toward aerodynamically efficient flight kinematics in response to an artificial load. Biology Open, 8(6), bio042572. Latash, M. L. (2010). Motor synergies and the equilibrium-point hypothesis. Motor Control, 14(3), 294–322. Lillie, M. A., & Gosline, J. M. (1990). The effects of hydration on the dynamic mechanical properties of elastin. Biopolymers: Original Research on Biomolecules, 29(8–9), 1147–1160. Meyers, R. A., & Stakebake, E. F. (2005). Anatomy and histochemistry of spread-wing posture in birds. 3. Immunohistochemistry of flight muscles and the ?shoulder lock? in albatrosses. Journal of Morphology, 263(1), 12–29. 48  Mijailovich, S. M., Stamenovic, D., Brown, R., Leith, D. E., & Fredberg, J. J. (1994). Dynamic moduli of rabbit lung tissue and pigeon ligamentum propatagiale undergoing uniaxial cyclic loading. Journal of Applied Physiology, 76(2), 773–782. Norell, M. A., & Clarke, J. A. (2001). Fossil that fills a critical gap in avian evolution. Nature, 409(6817), 181–184. Oliveros, C. H., Field, D. J., Ksepka, D. T., Barker, F. K., Aleixo, A., Andersen, M. J., Alström, P., Benz, B. W., Braun, E. L., Braun, M. J., & others. (2019). Earth history and the passerine superradiation. Proceedings of the National Academy of Sciences, 201813206. Ostry, D. J., & Feldman, A. G. (2003). A critical evaluation of the force control hypothesis in motor control. Experimental Brain Research, 153(3), 275–288. Pollock, C. M., & Shadwick, R. E. (1994). Allometry of muscle, tendon, and elastic energy storage capacity in mammals. American Journal of Physiology-Regulatory, Integrative and Comparative Physiology, 266(3), R1022–R1031. Pratt, G. A., & Williamson, M. M. (1995). Series elastic actuators. Proceedings 1995 IEEE/RSJ International Conference on Intelligent Robots and Systems. Human Robot Interaction and Cooperative Robots, 1, 399–406. Pratt, J., Krupp, B., & Morse, C. (2002). Series elastic actuators for high fidelity force control. Industrial Robot: An International Journal. Prilutsky, B. I., Herzog, W., & Allinger, T. L. (1996). Mechanical power and work of cat soleus, gastrocnemius and plantaris muscles during locomotion: possible functional significance of muscle design and force patterns. Journal of Experimental Biology, 199(4), 801–814. 49  Prum, R. O., Berv, J. S., Dornburg, A., Field, D. J., Townsend, J. P., Lemmon, E. M., & Lemmon, A. R. (2015). A comprehensive phylogeny of birds (Aves) using targeted next-generation DNA sequencing. Nature, 526(7574), 569. Razmadze, D., Panyutina, A. A., & Zelenkov, N. v. (2018). Anatomy of the forelimb musculature and ligaments of Psittacus erithacus (Aves: Psittaciformes). Journal of Anatomy. Roberts, T. J., & Azizi, E. (2011). Flexible mechanisms: the diverse roles of biological springs in vertebrate movement. Journal of Experimental Biology, 214(3), 353–361. Roberts, T. J., & Gabaldón, A. M. (2008). Interpreting muscle function from EMG: lessons learned from direct measurements of muscle force. Integrative and Comparative Biology, 48(2), 312–320. Robertson, A. M. B., & Biewener, A. A. (2012). Muscle function during takeoff and landing flight in the pigeon (Columba livia). Journal of Experimental Biology, jeb–075275. Rosario, M. v, Sutton, G. P., Patek, S. N., & Sawicki, G. S. (2016). Muscle–spring dynamics in time-limited, elastic movements. Proceedings of the Royal Society B: Biological Sciences, 283(1838), 20161561. Sawicki, G. S., Robertson, B. D., Azizi, E., & Roberts, T. J. (2015). Timing matters: tuning the mechanics of a muscle-tendon unit by adjusting stimulation phase during cyclic contractions. Journal of Experimental Biology, jeb–121673. Schmidt-Wellenburg, C. A., Biebach, H., Daan, S., & Visser, G. H. (2007). Energy expenditure and wing beat frequency in relation to body mass in free flying Barn Swallows (Hirundo rustica). Journal of Comparative Physiology B, 177(3), 327–337. 50  Scott, S. H. (2004). Optimal feedback control and the neural basis of volitional motor control. Nature Reviews Neuroscience, 5(7), 532–545. Shufeldt, R. W. (1890). The myology of the raven (Corvus corax sinuatus.): A guide to the study of the muscular system in birds. Macmillan and Company. Stowers, A. K., & Lentink, D. (2015). Folding in and out: passive morphing in flapping wings. Bioinspiration & Biomimetics, 10(2), 25001. Stowers, A. K., Matloff, L. Y., & Lentink, D. (2017). How pigeons couple three-dimensional elbow and wrist motion to morph their wings. Journal of The Royal Society Interface, 14(133), 20170224. Sullivan, S. P., McGechie, F. R., Middleton, K. M., & Holliday, C. M. (2019). 3D muscle architecture of the pectoral muscles of European Starling (Sturnus vulgaris). Integrative Organismal Biology, 1(1), oby010. Swinebroad, J. (1954). A comparative study of the wing myology of certain passerines. American Midland Naturalist, 488–514. Theriault, D. H., Fuller, N. W., Jackson, B. E., Bluhm, E., Evangelista, D., Wu, Z., Betke, M., & Hedrick, T. L. (2014). A protocol and calibration method for accurate multi-camera field videography. Journal of Experimental Biology, 217(11), 1843–1848. Theriault, J. S., Bahlman, J. W., Shadwick, R. E., & Altshuler, D. L. (2019). Work loop dynamics of the pigeon (Columba livia) humerotriceps demonstrate potentially diverse roles for active wing morphing. Journal of Experimental Biology, 222(7), jeb195578. Tobalske, B. (1995). Neuromuscular control and kinematics of intermittent flight in the European starling (Sturnus vulgaris). Journal of Experimental Biology, 198(6), 1259–1273. 51  Tobalske, B., & Dial, K. (1996). Flight kinematics of black-billed magpies and pigeons over a wide range of speeds. Journal of Experimental Biology, 199(2), 263–280. Tobalske, B. W. (2007). Biomechanics of bird flight. Journal of Experimental Biology, 210(18), 3135–3146. Tobalske, B. W., & Biewener, A. A. (2008). Contractile properties of the pigeon supracoracoideus during different modes of flight. Journal of Experimental Biology, 211(2), 170–179. Tobalske, B. W., Hedrick, T. L., & Biewener, A. A. (2003). Wing kinematics of avian flight across speeds. Journal of Avian Biology, 34(2), 177–184. Tresch, M. C., & Jarc, A. (2009). The case for and against muscle synergies. Current Opinion in Neurobiology, 19(6), 601–607. Usherwood, J. R., Cheney, J. A., Song, J., Windsor, S. P., Stevenson, J. P. J., Dierksheide, U., Nila, A., & Bomphrey, R. J. (2020). High aerodynamic lift from the tail reduces drag in gliding raptors. Journal of Experimental Biology, 223(3). Vazquez, R J. (1994). The automating skeletal and muscular mechanisms of the avian wing (Aves). Zoomorphology, 114(1), 59–71. Vazquez, Rick J. (1992). Functional osteology of the avian wrist and the evolution of flapping flight. Journal of Morphology, 211(3), 259–268. Vazquez, Rick J. (1995). Functional anatomy of the pigeon hand (Columba livia): a muscle stimulation study. Journal of Morphology, 226(1), 33–45. Xu, X., & Mackem, S. (2013). Tracing the evolution of avian wing digits. Current Biology, 23(12), R538–R544. 52  Zajac, F. E. (1989). Muscle and tendon: properties, models, scaling, and application to biomechanics and motor control. Critical Reviews in Biomedical Engineering, 17(4), 359–411.   53  Appendix EMG and kinematics data from all glide sections recorded is shown on the following page, printed horizontally due to size. Each column corresponds to a single glide, elbow angle is colored by individual bird.   54  


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items