Locomotion, Morphology, and Habitat Use in Arboreal

Home , Animal locomotion, Arboreal locomotion, Forelimb, Microsciurus, Patagium, Red squirrel

LOCOMOTION, MORPHOLOGY, AND HABITAT USE IN ARBOREAL

SQUIRRELS (RODENTIA: SCIURIDAE)

A dissertation presented to

the faculty of

the College of Arts and Sciences of Ohio University

In partial fulfillment

of the requirements for the degree

Doctor of Philosophy

Richard L. Essner, Jr.

June 2003 This dissertation entitled

LOCOMOTION, MORPHOLOGY, AND HABITAT USE IN ARBOREAL

SQUIRRELS (RODENTIA: SCIURIDAE)

RICHARD L. ESSNER, JR.

has been approved for

the Department of Biological Sciences

and the College of Arts and Sciences by

Stephen M. Reilly

Associate Professor of Biological Sciences

Leslie A. Flemming

Dean, College of Arts and Sciences

ESSNER, JR., RICHARD L. Ph.D. June 2003. Biological Sciences

Locomotion, Morphology, and Habitat Use in Arboreal Squirrels (Rodentia: Sciuridae)

(135pp.)

Director of Dissertation: Stephen M. Reilly

Arboreal locomotion has not been well studied in mammals outside of primates

and mammalian gliding has received even less attention. While numerous studies have examined morphological variation in these forms, there is currently a lack of detailed

kinematic, behavioral, and ecological data to assist in explaining the patterns. Here, I

present three studies that focus on differing aspects of locomotion in arboreal squirrels.

These range from 3-D kinematics (Chapters 1 & 2) to morphology, locomotor behavior,

and habitat use (Chapter 3).

First, kinematics were quantified and compared among leaping, parachuting, and

gliding squirrels to test for differences during the launch phase. Only six out of 23

variables were found to differ significantly among the three species investigated. The six

significant variables were partitioned into morphological, behavioral, and performance

based differences. Remarkably, there were no differences attributable to hindlimb

kinematics indicating that propulsion is the same in leaping, parachuting, and gliding

squirrels.

Next, the initial airborne phase was investigated. Following the launch

nongliding squirrels initiated gradual and symmetrical movements of the limbs. Flying

squirrels initiated highly stereotyped fluttering movements characterized by a series of rapid asymmetrical rotations of the fore and hindlimbs. This behavior is hypothesized to

provide control over angular momentum and disrupt detrimental fluid dynamic effects.

The kinematics of the landing phase were also investigated in flying squirrels. Landing

was characterized by limb adduction, tail dorsiflexion, and billowing of the patagium which acted to increase angle of attack and slow descent. Flying squirrels reached maximum extension of their limbs and maximal flexion of the vertebral column as they contacted the landing platform. This was hypothesized as a mechanism for increasing deceleration time and reducing peak landing forces.

Finally, in order to determine which aspects of morphology, behavior, and ecology define evolutionary transitions from arboreality to semiarboreality/terrestriality or gliding in squirrels I investigated each of the levels using ordination (principal components analysis and correspondence analysis) and multivariate analysis of variance.

In addition, I examined causal linkages across levels by testing a number of biomechanical predictions. Flying squirrels and chipmunks occupied extreme regions of morphospace, ethospace, and ecospace, while red squirrels were generally located in intermediate positions consistent with the idea that they exhibit the ancestral condition.

Biomechanical predictions were supported in some instances but not in others, reinforcing the need for further study.

Approved: Stephen M. Reilly

Associate Professor of Biological Sciences

Acknowledgments

I extend my sincere gratitude to my advisor Steve Reilly. Steve initially inspired me through the book he co-edited entitled Ecological Morphology. The ragged pages of that book have provided me with a roadmap throughout my graduate career. Steve has continued to inspire me through his passion for science and his quest for “the nanosecond of discovery.” I am also deeply indebted to Clay Corbin, Pete Larson, and Lance

McBrayer. I have learned more from them than they will ever know. Clay in particular made the long commutes from Lancaster bearable. I will truly miss our conversations on that long and winding road.

I thank Audrone Biknevicius, Ron Heinrich, and Larry Witmer for being excellent mentors. Their guidance and encouragement through the years have been a valuable asset. In addition, Sue Bullard and Brigitte Demes were extremely helpful with their comments. I sincerely appreciate their efforts on my behalf. These outstanding individuals comprised my dissertation committee and I thank them all for their support and dedication.

I thank John Scheibe for giving me the “spark” of scientific curiosity. John was the first person to introduce me to the field of ecology and evolutionary biology. In addition, he did me the great service of passing along his enthusiasm for ecomorphology and gliding locomotion. I am forever in his debt. I am also indebted to Jim Robins and

A. J. Hendershott for the times spent discussing gliders ad infinitum over the years.

I thank the Reilly Lab past & present especially Jason Elias, Andy Parchman, and

Kristen Hickey for their help with many aspects of my research. I also thank the OU

Evomorph group, most notably Kay Earls and Nancy Stevens, for their welcome advice.

I have never been associated with a finer group of individuals.

I sincerely thank the army of undergraduates who helped with much of this research. Special thanks go to Jeremy Caudill, Tara Chapman, Kim Huber, Brian

Mussio, Colin Shelton, and Jonathan Vega. Their hard work made this project possible.

I thank Robert Dudley, Steve Edinger, Cliff Frohlich, Ted Goslow, Robert Hikida, and

Robert Srygley for helpful comments. I thank Jerry Svendsen for expanding my background in the fields of ecology, mammalogy, and statistics and for generously allowing me to borrow traps for an extended period of time. Dave Elliott was extremely generous with his time and materials on many occasions during my time at OU.

Scott Carpenter and the Lab Animal Resources Staff took excellent care of the animals used in this study. Their expertise and dedication are truly impressive.

I sincerely thank my family for their construction skills, but more importantly for their love and support over all of these years. They have gone above and beyond the call of duty on countless occasions and are in large part responsible for getting me to this point. Most of all I extend my deepest gratitude to my wife Jill for her inexhaustible love and patience. I cannot count the number of weekends she gave up to assist me with data collection. She has been a partner in everything I have done.

This research was supported by NSF Dissertation Improvement Grant (IBN

0076342), an Ohio University Student Enhancement Award, grants-in-aid from the

Society for Integrative and Comparative Biology and the American Society of

Mammalogists, and several Ohio University John Houk Memorial Research Grants.

Table of Contents

Abstract...... 3

Acknowledgments...... 5

List of Tables ...... 8

List of Figures...... 9

General Introduction ...... 11 References...... 15

Chapter 1 Three-Dimensional Launch Kinematics in Leaping, Parachuting, and Gliding Squirrels ...... 17 Materials and Methods ...... 20 Results ...... 25 Discussion ...... 28 References ...... 36

Chapter 2 Three-Dimensional Aerial Kinematics in Arboreal Squirrels...... 46 Materials and Methods ...... 52 Results ...... 57 Discussion ...... 64 References ...... 73

Chapter 3 Morphology, Locomotor Behavior, and Habitat Use in Arboreal Squirrels....93 Materials and Methods ...... 98 Results ...... 102 Discussion ...... 105 References ...... 114

List of Tables

Table Page

1.1 Repeated-measures analysis of variance (ANOVA) of launch kinematics in chipmunks, red squirrels, and flying squirrels ...... 40

2.1 Results of a repeated-measures analysis of variance of fluttering kinematics in flying squirrels ...... 78

3.1 Limb proportions in extant and fossil squirrels (Ra=radius, Hu=humerus, Ti=tibia, Fe=femur ...... 122

3.2 Variation in body size and six limb morphological variables in chipmunks, red squirrels, and flying squirrels ...... 123

3.3 Factor loadings for PCA of six limb morphological variables in chipmunks, red squirrels, and flying squirrels ...... 124

3.4 Factor loadings for correspondence analysis of six locomotor behavior variables in chipmunks, red squirrels, and flying squirrels ...... 125

3.5 Factor loadings for correspondence analysis of nine habitat use variables in chipmunks, red squirrels, and flying squirrels ...... 126

3.6 Biomechanical predictions from Thorington and Thorington (1989), Youlatos (1999), and Stafford et al. (2003)...... 127

List of Figures

Figure Page

1.1 Defining sciurid arboreal locomotion ...... 41

1.2 Landmarks used to describe limb, body, and tail movements in squirrels during the launch phase...... 42

1.3 Representative video frames portraying a single takeoff sequence during a launch in the flying squirrel...... 43

1.4 Mean kinematic profiles of the hind limb and tail in chipmunks, red squirrels, and flying squirrels ...... 44

1.5 Mean kinematic profiles of the forelimb in chipmunks, red squirrels, and flying squirrels...... 45

2.1 Three principal body axes for rotational motion...... 79

2.2 Fluid dynamic effects on flat objects...... 80

2.3 Landmarks used to describe limb, body, and tail movements in squirrels during airborne and landing phases...... 81

2.4 Angles of abduction and protraction measured for the entire limb during the airborne phase in squirrels ...... 82

2.5 Mean kinematic profiles for abduction and protraction of the right forelimb and hindlimb in chipmunks and red squirrels...... 83

2.6 Representative video frames (16 ms intervals) in posterior and dorsal views portraying the initial airborne phase in flying squirrels...... 84

2.7 Mean kinematic profiles of fluttering behavior ...... 85

2.8 Amplitudes and wavelengths during fluttering locomotion...... 86

2.9 Mean velocity plots demonstrating how gliding velocity changes throughout the glide path...... 87

2.10 Representative video frames (16 ms intervals) in posterior view portraying the rolling movement characteristic of the majority of trials in chipmunks...... 88

2.11 Representative video frames (16 ms intervals) in posterior view portraying the rolling movement characteristic of 8% of trials in flying squirrels ...... 89

2.12 Mean kinematic plot of leading edge and trailing edge lengths measured as the distance from the occiput to the wrist (leading edge) and base of the tail to the ankle (trailing edge) ...... 90

2.13 Landing sequence in flying squirrels traced from actual video frames (16 ms intervals) ...... 91

2.14 Mean kinematic profiles of the landing phase of flying squirrels ...... 92

3.1 Phylogenetic relationships of the sciurid taxa included in this study ...... 128

3.2 Enclosure constructed to study squirrel behavior and habitat use ...... 129

3.3 Size removed principal components analysis (PCA) of six limb morphological variables in chipmunks, red squirrels, and flying squirrels ...... 130

3.4 Proportions for the seven locomotor behaviors examined in chipmunks, red squirrels, and flying squirrels...... 131

3.5 Correspondence analysis (CA) of six locomotor behaviors in chipmunks, red squirrels, and flying squirrels...... 132

3.6 Proportions for the nine habitat use variables examined in chipmunks, red squirrels, and flying squirrels ...... 133

3.7 Correspondence analysis (CA) of nine habitat use variables examined in chipmunks, red squirrels, and flying squirrels ...... 134

3.8 Multilevel comparisons of sciurid morphospace, ethospace, and ecospace ...... 135

General Introduction

The research presented in this dissertation represents an initial foray into the ecological morphology of arboreal locomotion in squirrels. It was first conceived during my master’s research on shape variation in the scapula of flying and tree squirrels. That study uncovered interesting variation in size and shape among squirrels, but provided few answers regarding functional importance (Essner and Scheibe, 2000). It was at this point that I realized that morphological patterns described in my own research and in that of others (e.g. Thorington and Heaney, 1981; Scheibe et al., 1990; Runestad and Ruff, 1995;

Scheibe and Essner, 2000) might benefit from a detailed investigation of locomotion, morphology, and habitat use in gliding and nongliding squirrels. Specifically, this study compares three species that are thought to vary considerably in aspects of their morphology, behavior, and ecology. They include an arboreal glider, the southern flying squirrel (Glaucomys volans), an arboreal parachuter, the red squirrel (Tamiasciurus hudsonicus), and a semiarboreal leaper, the eastern chipmunk (Tamias striatus).

Much of the research presented in the following chapters necessarily focuses on quantification of the kinematics of arboreal locomotion in squirrels. The majority of descriptions regarding arboreal locomotion outside of primates have been anecdotal and relatively simplistic. The same holds for descriptions of gliding locomotion. I discovered relatively early during the course of this research that I was dealing with an impressive degree of complexity. I decided that the best course of action was to break down arboreal locomotion into discrete and manageable units for analysis by partitioning it into launch, airborne, and landing stages. The detailed kinematic data presented in the following chapters represent the first of its kind for squirrels or for gliders in general.

Preview

Chapter one focuses on comparing the kinematics of the three species during the initial launch phase of locomotion. There are notable differences among the species with respect to midair postures and aerodynamics but it was unclear whether variation should be expected during the launch. In order to address this question I filmed animals launching to the ground using high-speed video. Statistical comparisons among taxa indicated that only six out of 23 variables were significantly different among the three species. Two were associated with tail kinematics and were a consequence of tail morphology. Two were forelimb related and discriminated gliding from nongliding taxa.

The remaining two variables were performance attributes, indicating significant variation among the species in takeoff velocity and horizontal range. Remarkably, none of the differences involved hindlimb kinematics, indicating that propulsion was essentially identical in leaping, parachuting, and gliding squirrels.

Chapter two is a detailed comparison of the initial airborne phase of locomotion among flying squirrels, red squirrels, and chipmunks. In addition, landing kinematics are quantified in flying squirrels. Nongliding squirrels exhibited relatively gradual and symmetrical movements of the limbs following the launch. In contrast, flying squirrels initiated highly stereotyped fluttering movements, characterized by a series of rapid asymmetrical rotations, involving simultaneous rolling (abduction/adduction) and yawing

(protraction/retraction) limb movements. This complex sequence of rotations and counter-rotations is hypothesized to be a novel behavioral mechanism in flying squirrels that 1) overwhelms angular momentum imparted by the launch; and 2) controls the formation of vortices that form under flat objects moving through fluids. Landing

13 behavior in flying squirrels was characterized by adduction of the limbs, dorsiflexion of the tail, and billowing of the patagium, which together acted to increase the angle of attack in order to slow descent. Flying squirrels reached maximum extension of their limbs and maximal flexion of the vertebral column as they contacted the landing substrate. It is hypothesized that this represents a mechanism for increasing deceleration time and reducing peak landing forces.

Chapter three is an integrative study that examines flying squirrels, red squirrels, and chipmunks at multiple levels—specifically, morphology, locomotor behavior, and habitat use. This approach has been used with success in the past to describe how novel behaviors evolve (Reilly and Lauder, 1992; Lauder and Reilly, 1996). The goal of this study was to identify the evolutionary changes associated with transitions from the arboreal (ancestral) condition exhibited by modern tree squirrels (Thorington et al, 1998) to one of 1) semiarboreality/terrestriality in chipmunks; and 2) gliding in flying squirrels.

In addition, biomechanical predictions discussed in previous studies of squirrel positional behavior were tested to help determine their generality. The three species were significantly different in aspects of morphology, locomotor behavior, and ecology. In most respects, flying squirrels and chipmunks occupied extreme positions in multivariate space while red squirrels were generally intermediate.

These chapters contribute greatly to our understanding of arboreal locomotion outside of primates and provide the first detailed descriptions of mammalian gliding locomotion. While a number of important questions have been addressed, many more have been brought to light. It is hoped that these studies will result in new avenues of

14 research and provide a framework for future investigations of arboreal/gliding locomotion.

References

Essner, Jr., R. L. and Scheibe, J. S. (2000). A comparison of scapular shape in flying

squirrels (Rodentia: Sciuridae) using relative warp analysis. In: Biology of

Gliding Mammals. 213-228. Goldingay, R. L. and Scheibe, J. S. (Eds). Fürth:

Filander Verlag.

Lauder, G. V. and Reilly, S. M. (1996). The mechanistic bases of behavioral evolution:

A multivariate analysis of musculoskeletal function. In: Phylogenies and the

Comparative Method in Animal Behavior: pp. 105-137. Martins, E. P. (ed). New

York: Oxford University Press.

Reilly, S. M. and Lauder, G. V. (1992). Morphology, behavior, and evolution:

Comparative kinematics of aquatic feeding in salamanders. Brain Behav. Evol.

40, 182-196.

Runestad, J. A. and Ruff, C. B. (1995). Structural adaptations for gliding in mammals

with implications for locomotor behavior in Paromomyids. Am. J. Phys. Anthrop.

98, 101-119.

Scheibe, J. S., and Essner, Jr., R. L. (2000). Pelvic shape in gliding rodents:

Implications for the launch. In: Biology of Gliding Mammals. 167-184.

Goldingay, R. L. and Scheibe, J. S. (Eds). Fürth: Filander Verlag.

Scheibe, J. S., Figgs, D., and Heiland, J. (1990). Morphological attributes of gliding

rodents: A preliminary analysis. Trans. Missouri Acad. Sci. 24, 49-55.

Thorington, Jr., R. W., Miller, A. M. L., and Anderson, C. G. (1998). Arboreality in

tree squirrels (Sciuridae). In: Ecology and evolutionary biology of tree squirrels.

119-130. Steele, M. A., Merritt, J. F., and Zegers, D. A. (Eds). Martinsville, VA:

Virginia Museum of Natural History.

Thorington, R. W., Jr. and Heaney, L. R. (1981). Body proportions and gliding

adaptations of flying squirrels (Petauristinae). J. Mamm. 62, 101-114.

Chapter 1. Three-dimensional Launch Kinematics in Leaping, Parachuting, and

Gliding Squirrels

Introduction

Squirrels (Family Sciuridae) likely originated from an arboreal ancestry and first appear in the fossil record during the Eocene (Emry and Thorington, 1982; Thorington et

al., 1997). Since then they have undergone a remarkable radiation into a multitude of

arboreal and terrestrial environments, ranging from deserts to alpine meadows and

forests. With respect to arboreality, the radiation has resulted in an array of locomotor

modes used by squirrels for negotiating gaps in the canopy, principally through leaping,

parachuting, or gliding. These locomotor modes are of critical interest, since together

they comprise a considerable proportion of the behavioral repertoire of arboreal squirrels

(Essner, in prep). In addition, they may have important effects on fitness through reduced costs of transport (compared with descending to the ground and climbing up), predator avoidance, or foraging optimization (Rayner, 1981; Scheibe et al., 1990; Keith et al., 2000). While arboreal leaping has been well studied in primates, it remains unexamined in squirrels and other mammalian taxa. Similarly, detailed studies of mammalian parachuting and gliding locomotion are lacking.

The lack of comparative studies involving leaping, parachuting, and gliding may be attributable in part to the treatment of these locomotor modes as continuous rather than discrete behaviors (Pennycuick, 1986). For example, parachuting and gliding have traditionally been defined based upon the angle of descent from the horizontal (> 45º = parachute, < 45º = glide), rather than upon specific morphological or behavioral

18 characteristics (e.g., Oliver, 1951; Rayner, 1981). While this definition presents a useful way to classify locomotion in terms of basic aerodynamics, it is of limited utility for classifying behavior since many animals can actively choose their angles of descent and because these angles are dependent upon unpredictable air currents (Moffett, 2000).

Moreover, if a ballistic component is included it could result in “gliding” angles over short to moderate distances, making it difficult to apply the criterion universally. A more biologically relevant criterion would incorporate the features that enable an organism to control its descent, rather than strictly defining locomotion based on aerodynamic performance (Moffett, 2000).

Defining leaping, parachuting, and gliding based upon midair postural behavior is a useful criterion for examining functional attributes. Sciurid arboreal leaping is here considered to be a relatively unspecialized locomotor mode, accompanied by minor aerodynamic effects, where the limbs remain adducted during the airborne phase (Fig.

1.1A). In contrast, sciurid parachuting is defined as midair limb abduction with flexion of the distal elements in order to assume a flattened posture, resulting in significant amounts of drag (Fig. 1.1B). Finally, sciurid gliding is defined as midair limb abduction with full extension of the distal limb elements, generating relatively high amounts of lift

(Fig. 1.1C).

Despite key differences among leaping, parachuting, and gliding locomotor modes with respect to midair posture and associated aerodynamics, it is currently unclear whether such differences are apparent during the initial phase of locomotion, prior to the squirrel becoming airborne. Indeed, there are reasons for suspecting that the locomotor modes may initially be indistinguishable. For example, there seems to be a general

19 reliance on hind limb propulsion within sciurids. Keith et al (2000) demonstrated that the cost of active launching is relatively inexpensive for a gliding squirrel and suggested that a hind limb driven “leaping” launch may reduce the distance at which gliding becomes cost effective by improving glide velocity or glide angle. It is not surprising then that parachuting and gliding squirrels actively rely on the hind limbs for generating propulsion, rather than passively dropping into a parachute or glide (e.g., Keith et al.,

2000). Moreover, since the functional demands of takeoff are exceptionally high (e.g.,

Demes et al., 1995, 1999), they may act to limit the degree of variation in hind limb kinematics among the three locomotor modes. In addition, it is not known whether morphological elements such as the forelimbs and tail, which do not contribute as significantly to propulsion, are free to exhibit kinematic variation. If they are, such variation could help to further define these locomotor modes.

In order to test for functional differences in the launch phase during leaping, parachuting, and gliding locomotion, three-dimensional kinematic data were collected in the eastern chipmunk Tamias striatus, a semiarboreal leaper, red squirrel Tamiasciurus hudsonicus, an arboreal parachuter, and the southern flying squirrel Glaucomys volans, an arboreal glider. These three North American species represent major lines of divergence within squirrels and are a good sample of sciurid diversity (Fig. 1.1D).

Phylogenetic evidence provided by morphological, molecular, and immunological data points to a sister-group relationship between tree squirrels and flying squirrels, with chipmunks branching off relatively early in the history of the group (Fig. 1.1D; Hight,

1974; Oshida et al., 1996; Roth, 1996). The three taxa included in this study are of relatively similar body size (chipmunk 99 ± 1.1 g; red squirrel 181 ± 5.7 g; flying squirrel

107 ± 1.9 g; mean ± S.E.M.), compared with other sciurids, which range in size from 10 g to 7.5 kg (Nowak, 1991). In addition, there is some degree of proportional variation among the three species, presumably related to locomotor variation. In general, the fore and hind limbs are elongated relative to vertebral column length as the degree of arboreality increases. Thus, semiarboreal chipmunks possess relatively short limbs; while at the other extreme, highly arboreal flying squirrels possess relatively elongated limbs (Bryant, 1945). Hence, these three species provide a suitable test for functional differences associated with the launch phase of leaping, parachuting, and gliding.

In this study I define and compare the launch phases in three species that exhibit leaping, parachuting, and gliding locomotion, and relate launch movements to differences apparent during the airborne phase. Kinematic variation is then compared to morphological variation to examine the morphological, behavioral, and performance bases for differences in arboreal takeoffs in squirrels.

Materials and Methods

Kinematic analysis

Adult animals were collected from the wild and maintained in a colony at Ohio

University. Launching trials were filmed in the lab with two orthogonally placed JVC

GR-DVL9800U high-speed digital camcorders at 120 Hz with the aid of two Nova-

Strobe DA Plus stroboscopes (Monarch Instrument). A preliminary study of the three species under natural conditions indicated that horizontal takeoffs were the most frequently used method of launching. Therefore, animals were filmed in dorsal and lateral views as they launched from a 1.5-meter elevated horizontal platform to the

21 ground. The launch platform was constructed from a 5x20x30 cm pine board supported by metal shelf brackets attached to a vertical stand. The surface of the platform was covered with green indoor/outdoor carpet to prevent slipping during the launch.

Launches were part of an escape response elicited by tapping the platform, immediately behind the tail.

Animals were shaved on the right side of the body and markers (0.5 mm cotton pom poms) were glued over the joint centers in order to determine joint kinematics for the body, limbs, and tail (Fig. 1.2). Due to problems with skin movement, estimating the location of the knee and elbow via landmarks proved to be unreliable. Instead, these angles were estimated with trigonometry, using limb lengths measured from x-rays to construct two sides of a triangle and video measurement to construct the third side.

Additionally, markers (landmarks 3 and 7; Fig. 1.2) were placed slightly above the wrist and ankle, in line with the elbow and knee, in order to estimate the wrist and ankle angles.

Five individuals (of each species) were filmed and data from five trials per individual were used in the kinematic analysis. A total of 75 takeoffs were included in the analysis (25 per species). Horizontal distances were recorded for all of the trials and only the longest jumps for each individual where all landmarks were visible were included. Images were captured from both camera views using Ulead VideoStudio v.4.0 and imported into APAS motion analysis software (Ariel Dynamics) for 3-D kinematic analysis. The APAS trim module was used to synchronize the dorsal and lateral images based upon a shared kinematic event. The frame at which the toe was last in contact with the platform was used as the synchronization point. Launch sequences were digitized

22 using the autodigitizing function in the digitizing module. Once digitized, sequences were imported into the transformation module in order to convert the separate sets of 2-D coordinates into a unified set of 3-D coordinates. Data were unfiltered prior to their input into the display module, where 3-D angles were calculated and kinematic plots were recorded.

Kinematic variables

Hind limb and tail variables

A series of angular and timing variables were taken from each launch sequence in order to describe and compare statistically the three-dimensional movements of the limbs

(Table 1.1). Knee and ankle angles were measured in order to describe movement of the hind limb during the launch. Knee angles were calculated by measuring the lengths of the femur and tibia from x-rays and using video measurements to obtain the distance between the hip and ankle markers. Ankle angles were calculated using the angle formed by the tibia marker (placed slightly above the ankle, in line with the knee), ankle, and toe.

Minimum, maximum, and excursion values for the hind limb joints were determined and included in statistical comparisons. Tail movement (dorsiflexion/ventroflexion) was described by the angle formed by the midpoint of the tail, base of the tail, and a point projected directly beneath the base of the tail. Minimum and maximum values for the tail angle were also included in statistical comparisons.

Forelimb variables

Movement of the entire forelimb was described by measuring angles of forelimb protraction and forelimb abduction. Forelimb protraction was measured by the angle formed by the wrist, occiput, and base of the tail and describes the movement of the

23 entire limb with respect to the long axis of the body. Forelimb abduction was measured by the angle formed by the wrist, occiput, and a point that was projected directly beneath the occiput. It describes the movement of the limb with respect to an axis running dorsoventrally through the midline of the body. Adduction brings the forelimb closer to the midline of the body; whereas, abduction moves it further away.

During the initial part of the takeoff sequence the forelimbs remain in contact with the platform. At approximately the onset of the propulsive phase the forelimbs begin to lift from the platform and are brought forward toward the head. Forelimbs were not digitized until they began to lift off since the landmarks were not clearly discernable prior to that point. The timing of this event relative to the onset of a countermovement phase was measured as time to hand off. Elbow and wrist angles were also measured in order to describe the position of the forelimb joints. The starting and ending values for these angles were included in the statistical analysis.

Performance variables

Performance variables are those characteristics that can be related to takeoff performance. They include durations, takeoff velocity, and takeoff angle, all of which have effects on horizontal distance (Emerson, 1985). The durations of the preparatory and countermovement+propulsive phases as well as the entire takeoff event were measured from video. Takeoff velocities were measured by using the landmark located at the base of the tail (a relatively stable point during the launch sequence) to generate a displacement-time curve and obtaining the slope from the last five frames prior to loss of contact with the platform. Takeoff angles were measured using the angle formed by the occiput, toe and the horizontal and averaging the three frames prior to loss of contact with

24 the platform. Horizontal range was measured as the horizontal distance from the edge of the launch platform to the center of the landing site on the ground.

Multispecies comparisons

In order to graphically illustrate movement patterns for the forelimbs, hind limbs and tail, mean kinematic profiles were constructed. Individuals from each species were pooled and the means (± standard error) of five trials were calculated from trials exhibiting the same total duration. In order to compare differences among species statistically, a one-way repeated-measures analysis of variance was performed on a total of 23 kinematic variables, including timing, angle, and performance variables (Table 1.1).

For each variable the analysis was run on five trials each from each of the five individuals per species. A repeated-measures design has the advantage of testing differences in the main effects after variation within individuals has been extracted. The a priori choice to use the same individuals repeatedly was made to control for the problem of interindividual variation and because the within-subjects design provides more conservative tests for significance than standard analysis of variance tests since the F- ratios for the main effects and their interaction are calculated by dividing the mean square rather than the error mean square (Zolman, 1993). A sequential Bonferroni correction

(Rice, 1989) was used to reduce the risk of making a Type I Error due to multiple comparisons. Post hoc tests were performed on significantly different variables in order to identify differences among species. All statistical analyses were performed using

Systat v.6.1.

Results

Representative video frames portraying a single launch in a flying squirrel are presented in Figure 1.3. Mean kinematic profiles for the hind limbs, tail, and forelimbs

(all three species) are presented in Figures 1.4 and 1.5. Species kinematic data and analysis of variance results are presented in Table 1.1. Because hind limb kinematics were similar for all three species, mean kinematic values reported in the text are pooled for the knee and ankle. All other data represent values for particular species.

Phases of the launch

Three distinct phases were identified in the launch sequence of individuals from all taxa investigated (Fig.1.3, 1.4). The first phase was termed the preparatory phase. It was characterized by a preliminary hop that transported the hind limbs forward to the edge of the platform. The preliminary hop resulted from extension of the knee and ankle

(pooled mean ± S.E.M., knee 72.3 ± 1.9°; ankle 67.8 ± 2.2°; N=75), flexion during the swing phase (knee 48.5 ± 1.9°; ankle 36.8 ± 2.2°), and extension as the toe made contact with the platform (knee 71.1 ± 1.5°; ankle 56.3 ± 2.1°). In contrast, the tail and forelimbs remained relatively stationary during the preparatory phase (Fig. 1.3, 1.4). As mentioned previously, forelimb movement was not quantified until approximately the onset of the propulsive phase, due to the difficulty of discerning the landmarks. In general, the preparatory phase was remarkably stereotyped in all sciurid launches.

The second phase was termed the countermovement phase and was initiated immediately following the preparatory phase (Fig. 1.3). This phase began at toe down and was characterized by flexion of the knee and ankle (knee 47.6 ± 1.6°; ankle 24.5 ±

1.9°) producing a countermovement important for maximizing takeoff velocity (Zajac,

1993). The countermovement was followed by a propulsive phase characterized by rapid extension (knee 109.4 ± 2.8°; ankle 131.3 ± 8.0°), until the animal lost contact with the platform (Fig.1.4). The propulsive phase was typified by dorsiflexion of the tail and protraction of the forelimbs (Fig. 1.3).

Kinematic patterns of the hind limb and tail

In general, hind limb kinematics are virtually identical in all three species. The remarkable similarities observed in the kinematic profiles of the knee and ankle (Fig. 1.4) are further reinforced by a lack of significant differences among species in the hind limb joint angle variables included in the repeated-measures analysis of variance (Table 1.1).

In contrast to the hind limb kinematics, the profile of the tail during the propulsive phase indicates a divergence among the three species with respect to tail dorsiflexion.

Chipmunks dorsiflex the tail to the greatest extent (mean 193.5°), followed by flying squirrels (mean 139.5°) and red squirrels (mean 118.1°; Table 1.1). In addition, the three species differ with respect to minimum tail angle during the propulsive phase. Flying squirrels initiate the propulsive phase with the tail still in contact with the platform (Fig.

1.3) resulting in a low minimum tail angle (mean 70.9°), compared with red squirrels

(mean 103.7°) and chipmunks (mean 126.7°; Table 1.1). Analysis of variance revealed that minimum and maximum tail angles differ significantly among the three taxa (both P

< 0.001; Table 1.1).

Kinematic patterns of the forelimb

Forelimb protraction during the propulsive phase is similar in all three taxa.

They all gradually bring the forelimbs forward from approximately 45° of protraction at the onset of hand off to approximately 65° at toe off (Fig. 1.5). In contrast, there is a

27 divergence among taxa with respect to forelimb abduction. The angle of forelimb abduction at hand off does not differ significantly (P = 0.172); but by toe off, there is a significant difference among the taxa (P < 0.001; Table 1.1). Flying squirrels were the only species that abducted the forelimb prior to becoming airborne, indicated by an increasing forelimb abduction angle (from mean 64.4° at hand off to 71.5° at toe off;

Table 1.1, Fig. 1.5). The other two species show a decrease in the abduction angle (mean

67.6° to 56.5° in chipmunks; mean 54.1° to 41.3° in red squirrels), indicating adduction.

Besides forelimb abduction, the timing of hand off is the only other significant difference involving the forelimb. Time from the onset of the countermovement phase

(toe down) to the point when the hands are lifted from the platform was significantly longer in flying squirrels (mean 60 ms) than in red squirrels (mean 20 ms) or chipmunks

(mean 30 ms; Table 1.1).

Launch performance

All three species differed significantly with respect to takeoff velocity (P = 0.001;

Table 1.1). The mean takeoff velocity for chipmunks was 2.3 m/s, followed by flying squirrels with a mean takeoff velocity of 2.5 m/s. Red squirrels exhibited the best performance with a mean takeoff velocity of 3.0 m/s (Table 1.1). Likewise, all three species differed significantly with respect to horizontal range. The mean range for chipmunks was only 1.6 m, while, flying squirrels and red squirrels performed better with mean ranges of 1.9 m and 2.3 m, respectively. In contrast, takeoff angles, although higher in red squirrels (mean 21.0°), were not significantly different from flying squirrels

(mean 12.0°) or chipmunks (mean 9.7°; Table 1.1) due to a strong interaction effect

28 between species and individuals. No significant differences were found in any of the remaining performance variables.

Discussion

Hind limb kinematics

Despite relying upon different locomotor modes and considerable ecological and morphological differences, the three species in this study do not differ with respect to hind limb kinematics during the launch phase. This suggests that propulsion is relatively unspecialized in sciurids, irrespective of locomotor mode. While it is possible that one or more of the taxa included in this study have converged upon identical patterns for generating propulsion, the most parsimonious explanation is that the three species investigated have retained this pattern from a common ancestor. Given that chipmunks, tree squirrels, and flying squirrels are thought to have diverged in the late Oligocene;

Black, 1963), the high degree of conservatism seems especially remarkable and indicates considerable constraint on the launch.

Despite variation in limb proportions among chipmunks, red squirrels, and flying squirrels, they are classified as small-bodied leapers when compared with the range of size variation that has been examined in primates (e.g. Demes et al., 1996). In general, small-bodied leapers are limited by hind limb length; while, large-bodied leapers are limited by force generating capacity (Bennet-Clark, 1977; Emerson, 1985; Demes and

Günther, 1989; Demes et al., 1996; Preuschoft et al., 1996). This scaling phenomenon has resulted in a dichotomy in the leaping kinematics of small-bodied versus large-bodied primates, based upon a differing reliance on proximal versus distal limb segments for

29 generating propulsion (Demes et al., 1996). Based on the limb kinematics presented in this study, squirrels appear to be launching like small-bodied primates, relying more upon the ankle (mean ankle excursion = 106.8°) than the knee (mean knee excursion = 61.8°) for propulsion. More studies are needed over a range of body masses in order to determine the degree to which generalized arboreal mammals, and squirrels in particular, fit this allometric pattern. Nevertheless, it is noteworthy that the differences in hind limb morphology over the subset of sciurid size ranges (≈ 100-200 g) used in this study were not substantial enough to have an effect on hind limb kinematics.

Morphologically based kinematic differences

Both of the variables describing movement of the tail were significantly different.

Tail dorsiflexion is frequently observed in leaping animals and is an inherent response that balances the angular momentum generated by counterclockwise rotation (when viewed from the right) of the pelvis during the launch (Emerson, 1985; Günther et al.,

1991). However, the degree of response of the tail is dependent on its moment of inertia

(mass x radius of gyration2). Thus, a longer tail with the center of mass located further

from the axis of rotation will respond less than a shorter tail with the mass concentrated

closer to the axis of rotation (Hall, 1995). Consistent with this principle, red squirrels

have the longest tails, followed by flying squirrels, and then chipmunks (Essner, pers.

obs.). This fits the pattern identified by Scheibe et al., (1990) that arboreal nongliding

forms generally have the longest tails, followed by gliding and ground dwelling forms.

In general, tail movement in flying squirrels was more stereotyped than in the

other taxa. This likely explains their significantly lower minimum angle for the tail.

Flying squirrels always began the propulsive phase with the tail in contact with the

30 platform and in line with the long axis of the body. In contrast, chipmunks and red squirrels often began the propulsive phase with the tail elevated or directed to one side.

This is illustrated by the greater standard errors associated with their tail movements

(Table 1.1). A possible explanation may be that flying squirrels are constrained to move in a more controlled manner since their dorsoventrally flattened tails have aerodynamic properties that could initiate detrimental rotations of the body during the initial airborne phase. More data are needed to determine the exact role of the tail in leaping and gliding.

The significant difference in the timing of hand off can also be explained by morphological variation. This variable discriminates flying squirrels from the two nongliding taxa. In contrast to the more subtle differences in hind limb proportions, the forelimbs are extremely elongated in flying squirrels. Forelimb elongation is undoubtedly a gliding-related trait that acts to increase the width of the airfoil during the glide (Rayner, 1981; Thorington and Heaney, 1981). Delayed timing of hand off in flying squirrels probably results from their relatively long forelimbs maintaining contact with the platform for an extended period of time.

Behaviorally based kinematic differences

While tail kinematics and the timing of hand off are likely attributable to morphological variation, forelimb abduction is best considered as a behavioral difference.

Forelimb abduction is a gliding-related behavior that brings the forelimbs into the proper position to form an airfoil during the airborne phase. The process of abduction during the launch in flying squirrels appears to be relatively uncomplicated. In general, the wrists and elbows are flexed as the limbs are brought forward in all three species. Since the wrist and elbow angles are statistically indistinguishable among the three species at both

31 hand off and toe off (Table 1.1), we can conclude that assuming an abducted posture during the launch in flying squirrels only involves abduction of the forelimb at the shoulder and no reorientation of the forelimbs themselves.

The precise three-dimensional description of takeoff kinematics in this study has demonstrated conclusively that forelimb abduction in flying squirrels begins prior to the animals becoming airborne. This is likely done in order to initiate gliding sooner. To accomplish this, flying squirrels provide angular momentum to the forelimbs while still in contact with the platform. Forelimb abduction is still possible in the absence of angular momentum; however, it undoubtedly takes longer and is of greater complexity since it must be accompanied by rotations about other body axes in order for angular momentum to be conserved (Frohlich, 1979, 1980; Dunbar, 1988). In general, midair rotations are minimized in all but the most specialized arboreal leapers (e.g., prosimians), due to the danger of initiating detrimental rotations that could result in an improper landing posture (e.g. Dunbar, 1988).

There are a number of advantages to be gained from an early onset of gliding in flying squirrels. For example, beginning a glide early produces a flatter trajectory with less initial vertical drop, resulting in a more energetically efficient glide (Pennycuick,

1986; Scholey, 1986; Scheibe and Robins, 1998). In addition, it allows gliding over relatively short distances. For example, the animals in this study were reaching stable glides over distances as short as one meter. Finally, an early onset of gliding enables maneuverability sooner within the glide phase. Early maneuverability, in turn, allows for quicker changes in direction in order to avoid predators, obstacles or even to choose a different landing site.

Curiously, chipmunks and red squirrels adduct the limbs prior to becoming airborne. Observations of the airborne phase indicate that at some point red squirrels reverse this trend and begin to abduct the limbs in midair; while, chipmunks remain adducted. It is not clear why red squirrels do not abduct prior to becoming airborne in the same manner as flying squirrels. One possibility is that midair abduction of a flexed

“parachuting” limb, possessing a relatively low moment of inertia, is less problematic than midair abduction of an extended “gliding” limb with a relatively high moment of inertia. Another possibility is that the advantages gained by an early onset of gliding are not relevant to parachuting.

Performance based kinematic differences

The high takeoff velocities and horizontal ranges of red squirrels are consistent with previous observations of their leaping proficiency. Contributing to their launching ability are their absolutely longer hind limbs (mean femur + tibia, red squirrel: 8.1 cm; flying squirrel: 6.4 cm; chipmunk: 5.6 cm). The fact that takeoff velocities are significantly different between the similarly sized flying squirrels and chipmunks suggests that flying squirrels are taking advantage of their longer hind limbs to substantially increase takeoff velocity. This combined with the ability to glide, even over short distances, allows flying squirrels to significantly increase horizontal range relative to chipmunks.

Generally, sciurid takeoff angles were lower than expected. The optimum takeoff angle for maximizing horizontal range depends upon the height difference between takeoff and landing sites (Lichtenberg and Wills, 1993). In this study, where squirrels launched from an elevated platform to the ground, the optimal takeoff angle is not the 45°

33 expected with level takeoff and landing sites. Instead, the 1.5-meter height differential reduces the optimal angle well below 45º, due to an increased flight time. For example, the mean optimal takeoff angle for a ballistically moving chipmunk, determined using the approach of Lichtenberg and Wills (1978), which takes into account the relative height of takeoff and landing sites, is 23.4º compared with an observed angle of 9.7º (Table 1.1).

The low takeoff angles used by squirrels in this study differ dramatically from those reported from arboreal leaping primates, which generally approach optimum takeoff angles (Crompton et al., 1993; Demes et al., 1996). It is unclear why the takeoff angles preferred by primates should differ from those used by squirrels. One possibility may be that quadrupedal squirrels are not able to raise their centers of mass as high as bipedally leaping primates. Another explanation would be the existence of a differing tradeoff between takeoff angle and takeoff velocity between the two groups. The tradeoff between angle and velocity has been well documented in human athletes; whereby, attempts to optimize takeoff angles in the jumping or throwing events result in significantly reduced horizontal velocities (Hall, 1995). Consistent with this, Keith et al.

(2000) found a similar relationship during the launch in flying squirrels. Future research should attempt to determine if the tradeoff between takeoff angle and takeoff velocity in quadrupedal leapers, such as squirrels, is more substantial than it is in bipedally oriented primate leapers.

Another possible explanation for the disparity between squirrels and primates is that animals launching in an escape response (e.g., this study) launch differently than animals taking off for a food reward (e.g., many primate studies). Low takeoff angles may be preferred during escape responses since they place greater horizontal distance

34 between a predator and prey in a given amount of time, despite higher takeoff angles resulting in greater overall distance. Nevertheless, observations of squirrels launching in the wild as well as inside an enclosure used for studying locomotor behavior and habitat use, suggests that low takeoff angles are the norm for sciurids (Essner, in prep).

Evolutionary Implications

While this study investigated only a subset of sciurid diversity and more taxa are undoubtedly required before definitive conclusions can be made, it may be fruitful to explore some of the evolutionary implications of the launch in this group. The evolution of gliding in squirrels is generally perceived as having progressed through intermediate leaping and parachuting stages (e.g. Bock, 1965). While it is impossible to directly test such a model, inferences may be drawn based upon extant forms that exhibit these stages.

Based upon the three species investigated in this study it appears that the demands of hind limb propulsion have resulted in a single mechanism for generating thrust during horizontal takeoff. Furthermore, the lack of variation in hind limb kinematics implies that launch propulsion played a relatively minor role during the evolution of parachuting and gliding locomotion in squirrels, since no specialization appears to be necessary to enter the airborne phase. In contrast, we cannot infer this for the other morphological elements, since some degree of specialization related to gliding was evident in tail and forelimb kinematics.

In conclusion, it is surprising to find that only six out of 23 kinematic variables investigated differed among the three species. While there are key differences that discriminate gliders from nongliders, none of these fundamentally affect the launch itself.

Undoubtedly, movement patterns during latter phases of leaping, parachuting, and gliding

(e.g. airborne and landing phases) will prove to be more complex. A detailed investigation of these phases may reveal additional distinguishing characteristics that will further elucidate the functional importance of locomotor variation in this group.

References

Bennet-Clark, H. C. (1977). Scale effects in jumping animals. In Scale Effects in

Animal Locomotion (ed. T. J. Pedley), pp.185-201. London: Academic Press.

Black, C. C. (1963). A review of the North American Tertiary Sciuridae. Bull. Mus.

Comp. Zool. 130, 109-248.

Bock, W. J. (1965). The role of adaptive mechanisms in the origin of higher levels of

organization. Syst. Zool. 14, 272-287.

Bryant, M. D. (1945). Phylogeny of the Nearctic Sciuridae. Am. Midl. Nat. 33, 257-

390.

Crompton, R. H., Sellers, W. I., and Günther, M. M. (1993). Energetic efficiency and

ecology as selective factors in the salutatory adaptation of prosimian primates.

Proc. R. Soc. Lond. B 254, 41-45.

Demes, B. and Günther, M. M. (1989). Biomechanics and allometric scaling in primate

locomotion and morphology. Folia Primatol. 53, 125-141.

Demes, B. Jungers, W. L., Gross, T. S., and Fleagle, J. G. (1995). Kinetics of leaping

primates: influence of substrate orientation and compliance. Am. J. Phys.

Anthrop. 96, 419-429.

Demes, B., Jungers, W. L., Fleagle, J. G., Wunderlich, R. E., Richmond, B. G., and

Lemelin, P. (1996). Body size and leaping kinematics in Malagasy vertical

clingers and leapers. J. Hum. Evol. 31, 367-388.

Demes, B., Fleagle, J. G., and Jungers, W. L. (1999). Takeoff and landing forces of

leaping strepsirhine primates. J. Hum. Evol. 37, 279-292.

Dunbar, D. C. (1988). Aerial maneuvers of leaping lemurs: The physics of whole-body

rotations while airborne. Am. J. Phys. Anthrop. 16, 291-303.

Emerson, S. B. (1985). Jumping and leaping. In Functional Vertebrate Morphology

(eds. M. Hildebrand, D. M. Bramble, K. F. Liem, and D. B. Wake), pp. 58-72.

Cambridge: Belknap Press.

Emry, R. J., and Thorington, R. W., Jr. (1982). Descriptive and comparative

osteology of the oldest fossil squirrel, Protosciurus (Rodentia: Sciuridae).

Smithson. Contrib. Paleobiol. 47, 1-35.

Frohlich, C. (1979). Do springboard divers violate angular momentum conservation?

Am. J. Physics 47, 583-592.

Frohlich, C. (1980). The physics of somersaulting and twisting. Sci. Amer. 242, 154-

164.

Günther, M. M., Ishida, H., Kumakura, H., and Nakano, Y. (1991). The jump as a

fast mode of locomotion in arboreal and terrestrial biotopes. Z. Morph. Anthrop.

78, 341-372.

Hall, S. J. (1995). Basic Biomechanics. St. Louis: Mosby-Yearbook Inc.

Hight, M. E., Goodman, M., and Prychodko, W. (1974). Immunological studies of the

sciuridae. Syst. Zool. 23, 12-25.

Keith, M. M., Scheibe, J. S., and A. J. Hendershott. (2000). Launch dynamics in

Glaucomys volans. In Biology of Gliding Mammals (eds. R. L. Goldingay, and J.

S. Scheibe), pp. 185-198. Fürth: Filander Verlag.

Lichtenberg, D. B., and J. G. Wills. (1978). Maximizing the range of the shotput.

Amer. J. Phys. 46, 546-549.

Moffett, M. W. (2000). What’s “Up”? A critical look at the basic terms of canopy

biology. Biotropica 32, 569-596.

Nowak, R. M. (1991). Walker’s Mammals of the World (5th edition). Johns Hopkins

University Press: Baltimore, pp 563-564.

Oliver, J. A. (1951). “Gliding” in amphibians and reptiles, with a remark on an arboreal

adaptation in the lizard, Anolis carolinensis carolinensis Voigt. Am. Nat. 85, 171-

176.

Oshida, T., Masuda, R., and Yoshida, M. (1996). Phylogenetic relationships among

Japanese species of the family Sciuridae (Mammalia, Rodentia), inferred from

nucleotide sequences of mitochondrial 12s ribosomal RNA genes. Zool. Soc. Jap.

13, 615-620.

Pennycuick, C. J. (1986). Mechanical constraints on the evolution of flight. In The

Origin of Birds and the Evolution of Flight (ed. K. Padian), pp. 83-98. San

Francisco: California Academy of Sciences.

Preuschoft, H., Witte, H., Christian, A., and Fischer, M. (1996). Size influences on

primate locomotion and body shape, with special emphasis on the locomotion of

‘small mammals’. Folia Primatol. 66, 93-112.

Rayner, J. M. V. (1981). Flight adaptations in vertebrates. Symp. Zool. Soc. Lond. 48,

137-172.

Rice, W. R. (1989). Analyzing tables of statistical tests. Evolution 43, 223-225.

Roth, V. L. (1996). Cranial integration in the Sciuridae. Am. Zool. 36, 14-23.

Scheibe, J. S., and Robins, J. H. (1998). Morphological and behavioral attributes of

gliding mammals. In Ecology and Evolutionary Biology of Tree Squirrels (eds.

M. A. Steele, J. F. Merritt, and D. A. Zegers), pp. 131-144. Martinsville, VA:

Virginia Museum of Natural History.

Scheibe, J. S., Figgs, D. and Heiland, J. (1990). Morphological attributes of gliding

rodents: A preliminary analysis. Trans. Mo. Acad. Sci. 24, 49-56.

Scholey, K. D. (1986). The climbing and gliding locomotion of the Giant Red Flying

Squirrel, Petaurista petaurista (Sciuridae). BIONA-report 5, 187-204.

Thorington, R. W., Jr. and Heaney, L. R. (1981). Body proportions and gliding

adaptations of flying squirrels (Petauristinae). J. Mamm. 62, 101-114.

Thorington, R. W., Jr., Darrow, K., and Betts, A. D. K. (1997). Comparative

myology of the forelimb of squirrels (Sciuridae). J. Morph. 234, 155-182.

Zajac, F. E. (1993). Muscle coordination of movement: A perspective. J. Biomech. 26,

109-124.

Zolman, J. F. (1993). Biostatistics: Experimental Design and Statistical Inference.

New York: Oxford University Press.

Table1.1. Repeated-measures analysis of variance (ANOVA) of launch kinematics in chipmunks, red squirrels, and flying squirrels (mean ± S.E.M; in degrees unless noted). Significance following sequential Bonferroni correction indicated by *.

Joint timing and angles Chipmunk Red Squirrel Flying Squirrel P Knee minimum angle 49.7 ± 1.4 52.5 ± 2.0 40.5 ± 1.4 0.031 Knee maximum angle 110.0 ± 1.7 114.1 ± 3.7 104.0 ± 2.9 0.300 Knee excursion angle 60.3 ± 1.7 61.7 ± 3.8 63.4 ± 3.3 0.906 Ankle minimum angle 23.6 ± 1.9 33.1 ± 2.0 16.8 ± 1.7 0.019 Ankle maximum angle 127.2 ± 15.7 133.5 ± 5.4 133.3 ± 2.9 0.716 Ankle excursion angle 103.6 ± 2.6 100.4 ± 5.3 116.6 ± 2.9 0.144 Tail minimum angle 126.7 ± 8.6 103.7 ± 4.4 70.9 ± 3.8 0.001* Tail maximum angle 193.5 ± 6.4 118.1 ± 5.9 139.5 ± 3.2 0.001* Protraction angle at hand off 46.9 ± 1.0 43.0 ± 0.8 47.2 ± 1.1 0.142 Protraction angle at toe off 57.5 ± 11.9 60.3 ± 9.7 59.8 ± 2.0 0.746 Abduction angle at hand off 67.6 ± 2.1 54.1 ± 1.8 64.4 ± 2.6 0.172 Abduction angle at toe off 56.5 ± 2.3 41.3 ± 2.8 71.5 ± 2.1 0.001* Time to hand off (ms) 30 ± 2.0 20 ± 1.9 60 ± 5.0 0.001* Elbow angle at hand off 156.6 ± 5.6 168.1 ± 4.6 133.7 ± 7.2 0.197 Elbow angle at toe off 70.5 ± 3.1 68.5 ± 5.6 71.9 ± 4.0 0.947 Wrist angle at hand off 145.6 ± 3.5 156.9 ± 2.7 153.1 ± 2.8 0.195 Wrist angle at toe off 131.1 ± 5.1 138.1 ± 3.8 110.9 ± 3.9 0.084

Performance variables Chipmunk Red Squirrel Flying Squirrel P Takeoff velocity (m/s) 2.3 ± 0.05 3.0 ± 0.08 2.5 ± 0.06 0.001* Range 1.6 ± 0.04 2.3 ± 0.07 1.9 ± 0.05 0.002* Takeoff angle 9.7 ± 2.5 21.0 ± 2.2 12.0 ± 2.6 0.402 Preparatory duration (ms) 60 ± 4.0 70 ± 4.0 70 ± 0.005 0.808 CM+Propulsive duration (ms) 60 ± 2.0 50 ± 2.0 90 ± 0.01 0.377 Total duration (ms) 120 ± 4.0 130 ± 5.0 160 ± 0.006 0.061

Figure 1.1. Defining sciurid arboreal locomotion. Here, discrete airborne postures are used to define locomotor mode: (A) Chipmunks are relatively unspecialized semiarboreal leapers that exhibit an adducted limb posture in midair. (B) Red squirrels are arboreal parachuters that exhibit a flattened posture in midair characterized by abduction of the proximal limb elements and flexion of the distal limb elements. (C) Flying squirrels are arboreal gliders that exhibit an abducted posture in midair with extension of the distal limb elements. (D) Phylogenetic relationships of the sciurid taxa included in this study (Hight et al., 1974; Oshida et al., 1996; Roth, 1996).

A B C

Chipmunk Red squirrel Flying squirrel

Figure 1.2. Landmarks used to describe limb, body, and tail movements in squirrels during the launch phase: 1) occiput; 2) shoulder over the glenoid fossa; 3) point slightly above wrist; 4) wrist; 5) base of the fifth phalanx of the manus; 6) hip over the greater trochanter; 7) point slightly above ankle; 8) ankle at the lateral malleolus; 9) base of the fifth phalanx of the pes; 10) base of the tail; 11) midpoint of the tail; 12) tip of the tail.

11 1 6 2 10 7 3 4 5 8 9

Figure 1.3. Representative video frames portraying a single takeoff sequence during a launch in the flying squirrel. Three distinct phases were identified: 1) the preparatory phase is characterized by a stereotyped preliminary hop that transports the hind limbs forward to the edge of the platform; 2) the countermovement phase is characterized by flexion of the knee and ankle producing a countermovement that increases takeoff velocity; 2) the propulsive phase immediately follows the countermovement phase and is characterized by rapid extension of the knee and ankle, until the animal loses contact with the platform. Note that during the propulsive phase the tail is dorsiflexed and the forelimbs are protracted.

0 ms 8 ms 16 ms 24 ms

32 40 ms 48 ms 56 ms ms

72 ms 80 ms 64 ms

88 ms 96 ms 104 ms

112 ms 120 ms 128 ms

Figure 1.4. Mean kinematic profiles of the hind limb and tail in chipmunks (triangles), red squirrels (squares), and flying squirrels (circles). All three species exhibit similar hind limb kinematics and share a preparatory phase with a preliminary hop. During the preliminary hop the foot is initially extended, flexed during transport, and extended once again as it is set back down. Toe down (TD) marks the beginning of the countermovement phase, characterized by flexion of the knee and ankle. This is followed by extension during the propulsive phase. Tail kinematics indicate a divergence among the three species, with chipmunks exhibiting the greatest amount of dorsiflexion, followed by flying squirrels, and red squirrels.

TD Preparatory CM Propulsive 140 120 100 80

Knee 60 40 20 0 0 16 32 48 64 80 96 112 128

140 120 100 80 60 Ankle 40 20

Angle (degrees) 0 0 16 32 48 64 80 96 112 128

200 150

Tail 100 50 0 0 16 32 48 64 80 96 112 128 Time (ms)

Figure 1.5. Mean kinematic profiles of the forelimb in chipmunks (triangles), red squirrels (squares), and flying squirrels (circles). Protraction brings the forelimbs closer to the head and is indicated by an increasing angle. A protraction angle of 90º indicates that the forelimbs have been brought forward to the level of the occiput. All three species exhibit similar values for forelimb protraction, bringing the forelimbs from approximately 45º at the onset of hand off to approximately 65º at toe off. Forelimb abduction moves the forelimbs away from the midline of the body and is indicated by an increasing angle. Forelimb adduction moves the forelimbs closer to the midline of the body and is indicated by a decreasing angle. An abduction angle of 90º indicates the forelimbs are fully abducted to the level of the occiput; while, an angle of 0º indicates the forelimbs are fully adducted to the midline. Flying squirrels abduct to approximately 72º before losing contact with the platform. The other two species show a decrease in the abduction angle indicating that they are adducting the forelimbs during the propulsive phase.

90 80 70 60 50 40 30

Forelimb Protraction 0 8 16 24 32 40

80 Angle (degrees) 60

20 Forelimb Abduction 0 8 16 24 32 40 Time (ms)

Chapter 2. Three-Dimensional Aerial Kinematics in Arboreal Squirrels

Introduction

Aerial locomotion in mammalian gliders has never been fully investigated using high-speed video and kinematic analysis. Consequently, descriptions of gliding behavior have remained relatively simplistic. This has led to a generalized perception that gliders passively drop out of trees, utilizing gravity to increase velocity before deploying the patagium to utilize aerodynamic forces to maintain orientation (e.g. Balda et al., 1987).

As with other broad categorizations (e.g. primate vertical clinging and leaping; Napier and Walker, 1967), there is a large degree of unappreciated variation among gliding mammals with respect to locomotor strategies (Addington et al., 2000). Keith and

Scheibe (2000) pointed out that representatives from at least three independently derived gliding taxa utilize an active (ballistic) rather than passive (dropping) launch. Likewise,

Stafford et al. (2002) suggested that gliding mammals that vary in wing loading might rely on different gliding trajectories ranging from active launches to steep dives. Thus, while passive falls and delayed onset of gliding may occur in some gliding forms, these do not reflect the gliding mechanism of many groups, including flying squirrels, which generally rely on active launches and assume gliding postures early in the glide path

(Ando and Shiraishi, 1993; Keith and Scheibe, 2000; Stafford et al., 2002; Essner, 2002).

Angular Momentum

Active launching was likely retained by flying squirrels from their nongliding ancestry. Nevertheless, it presents several distinct advantages. First, it increases velocity and allows the animals to reach the minimum speed needed for gliding (see Norberg,

1990) earlier than if they were to passively wait for gravity to increase their velocity.

Second, gliding may be initiated as an escape response (Emmons and Gentry, 1983) and an active launch allows for a more rapid getaway. Finally, active launching allows flying squirrels to clear underlying obstacles, whereas, passive falls could be impossible in a complex habitat. Nevertheless, active launching does present a potentially dangerous problem to leaping or gliding animals in the form of angular momentum (Caple et al.,

1983).

Angular momentum is the rotational analogue of linear momentum and is the product of moment of inertia and angular velocity. Rotational motion occurs about three principal body axes that are orthogonal to one another (Fig. 2.1). Rotation about the longitudinal axis is referred to as roll. While rotations about the transverse and vertical axes are referred to as pitch and yaw, respectively. The moments of inertia about the three axes indicate the relative degree of resistance to changes in angular velocity about the axes and are calculated as the product of mass and the squared distance to the radius of gyration. In the case of leaping forms, the moment of inertia about the longitudinal axis is typically the lowest (Caple et al., 1983), indicating the ease with which rolling is initiated relative to pitch and yaw. For example, as an arboreal animal launches into the air its overall shape is elongated as the hindlimbs are extended during propulsion (e.g.

Demes et al., 1996). Consequently, mass is distributed in closer proximity to the longitudinal axis than to the transverse or vertical axes.

As with linear momentum, angular momentum remains constant in the absence of external forces. Therefore, if an animal leaves the substrate with angular momentum, there is the potential to rotate out of the proper airborne or landing orientation (e.g. Peters

48 and Preuschoft, 1984). Rotations may be shifted to other body axes either directly or indirectly via controlled movements of the limbs, body, or tail (Dunbar, 1988; 1994).

These segmental movements either increase or decrease moment of inertia, enabling control over rotational velocity (e.g. Niemitz, 1984; Peters and Preuschoft, 1984; Dunbar

1988). Many leapers, including arboreal primates, do this quite effectively in order to produce midair rotations that bring the limbs into proper landing position (Peters and

Preuschoft, 1984, Dunbar, 1988, Demes et al., 1996). Allometric differences among leaping primates result in differing strategies for accomplishing midair rotations. Large- bodied primates (e.g. indris) rely on forelimb movements to initiate airborne rotations due to the large size of their forelimbs compared to the relatively small size of their tails. In contrast, small-bodied primates have relatively small forelimbs and instead use their tails to initiate rotations (Demes et al., 1996). Gliding and flying forms are unique because they have the potential to use aerodynamic forces to control angular momentum. Lift and drag enable a gliding or flying animal to alter its angular momentum because they can be used to create torque about an axis (Caple et al., 1983). The extent to which angular momentum poses a problem for flying squirrels is currently unclear, as are mechanisms for controlling it. The red squirrel, a nongliding relative of flying squirrels, has been described anecdotally as using mid-air contortions of the limbs and tail to maintain body posture (MacClintock, 1970). Parsimony would predict that flying squirrels move in the same manner as their nongliding relatives. Alternatively, flying squirrels may rely on their tails to initiate rotations as evidenced by small-bodied primates, or they may instead use the forelimbs in the manner of large-bodied leapers. While the forelimbs of flying

49 squirrels are lightweight, they may be able to compensate for the lack of mass by the assistance of aerodynamics (Caple et al., 1983).

Perhaps due to an ability to use aerodynamic forces to control angular momentum, flying squirrels are able to launch from any position on a tree. In addition to the typical horizontal launch, they frequently make use of vertical clinging and leaping launches (Essner, pers. obs.). In order to place the body in the proper orientation for launching during vertical clinging and leaping, the flying squirrel must roll about its long axis with one foot acting as a pivot. This produces an asymmetrical launch since the pivot foot maintains contact with the substrate longer and generates higher force. A large proportion of horizontal launches may also be asymmetrical due to substrate variation or accidental slips (see Fig. 2.3 in Essner, 2002 for example of an asymmetrical takeoff). In addition, launching in any direction other than straight ahead requires one foot to maintain contact with the substrate for a longer period of time, producing asymmetry.

Thus, asymmetrical launches may be the rule rather than the exception under normal conditions. This is also the case in arboreal leaping primates, such as the galago, in which asymmetrical launches are more prevalent than symmetrical ones (Jouffroy et al.,

1974). Asymmetrical launches may result in high amounts of angular momentum, causing the body to continue rotating once airborne and putting the leaping or gliding animal at risk of getting out of their proper orientation (Caple et al., 1983).

Fluid Dynamics

Another potential obstacle to maintaining orientation is the interaction between the animal and the air. Objects moving through the air are affected to some extent by fluid flow. The degree to which this occurs is dependent upon the object’s shape. For

50 example, a feather and a metal ball descend at different rates on Earth due to the effects of air resistance. However, when air resistance is removed in a vacuum they fall at the same rate. This is because feathers are more susceptible to the effects produced by the surrounding fluid due to their greater surface area. In addition to producing drag and slowing descent, fluids cause objects to deviate from their trajectories. Feathers, leaves and other flat objects dropped in a fluid exhibit periodic side-to-side rolling oscillations

(Tanabe and Kaneko, 1994) that are coupled with their forward motion (Belmonte et al.,

1998). In addition, they may exhibit end-over-end tumbling behavior. One major factor in determining whether an object will undergo side-to-side versus tumbling rotations is its shape—elongated objects (i.e., rectangles) oscillate side-to-side while short objects (i.e., squares) tumble (Belmonte, 1998; Fig. 2.2). Fluid dynamicists have determined that vortices form under flattened shapes and that the shedding of these vortices is associated with the side-to-side rolling oscillations (Fig. 2.2A). As the shape becomes more square- like, tumbling occurs because square-like shapes produce higher angular momentum that results in the oscillations increasing in magnitude until they become so large that the object rolls over completely (Fig. 2.2B). This has important implications for gliding organisms since all non-spherical objects moving with a high enough velocity are prone to exhibit oscillations (Belmonte, 1999). This may be especially problematic in flying squirrels since the transition to a gliding posture involves moving from a rectangular shape at the launch to a square-like shape during gliding. The change in shape may signify a greater likelihood of rollover.

Gliding Posture

Assumption of a gliding posture in mammals involves abduction and protraction of the fore and hindlimbs. This stretches the patagium (gliding membrane) and creates an airfoil that is a necessary component of gliding. The limb movements associated with deploying the patagium have never been investigated in detail. It is known that the launch phase of locomotion in flying squirrels is characterized by relatively gradual and symmetrical movements of the fore and hindlimbs (Essner, 2002). Whether the airborne phase is also characterized by such movements is unclear. Unfortunately, there are few data regarding limb movement patterns associated with the onset of gliding posture and there is reason to suspect that they may be characterized by a high degree of complexity in flying squirrels. For example, patagial shape and area are known to be highly dynamic immediately following the launch (Keith and Scheibe, 2000). This may indicate that the limbs play an active role in controlling patagial shape in flying squirrels. Changes in patagial shape following the launch could have important aerodynamic consequences.

While a thorough examination of aerial locomotion requires a complete aerodynamic analysis, a detailed kinematic analysis of limb movements as animals transition to a glide under normal conditions may elucidate problems faced by flying squirrels and uncover potential mechanisms for overcoming them.

Landing

Likewise, the landing behavior of gliding mammals has been little studied. In general, descriptions have consisted primarily of gliding mammals performing a flare

(increased angle of attack resulting in an upwards swoop due to high lift generation), adducting the limbs, and using the patagium as a parachute for slowing descent prior to

52 impact (Scholey, 1986; Ando and Shiraishi, 1993). Slowing descents through postural adjustments may be an important means of reducing impact forces in gliding mammals since peak velocities occur during the final segment of the glide (e.g. 8.27 ms-1 in

Glaucomys volans; Scheibe and Robins, 1998) and preferred landing sites are generally

noncompliant large-diameter tree trunks (Ando and Shiraishi, 1993; Jackson, 1999;

Essner, Chapter 3). Nevertheless, the precise movements associated with landing, as well

as, potential mechanisms for reducing landing forces have yet to be fully investigated.

The primary goals of this study are: 1) to quantify limb movement patterns

during the initial airborne phase of locomotion in flying squirrels in order to determine

how they are able to make the transition to a gliding posture, in light of the potentially

confounding obstacles presented by angular momentum and fluid dynamics; and 2) to

quantify movement patterns during the landing phase of locomotion in flying squirrels to

determine the mechanisms by which they slow descent. This will add to a body of

knowledge that already includes detailed descriptions of launch kinematics (Essner,

2002).

Materials and Methods

Kinematic Analysis

Adult animals (eastern chipmunk Tamias striatus (Illiger, 1811), red squirrel

Tamiasciurus hudsonicus (Trouessart, 1880) and southern flying squirrel Glaucomys

volans (Thomas, 1908) were collected from the wild and maintained in a colony at Ohio

University. All procedures were approved by Ohio University’s Institutional Animal

Care and Use Committee. Whereas flying squirrels were the primary focus of the study,

53 two nongliding squirrel taxa, red squirrels and chipmunks, were included for the purpose of comparison. Airborne trials were filmed in the lab with two orthogonally placed JVC

GR-DVL9800U high-speed digital camcorders filming at 120 Hz with the aid of two

Nova-Strobe DA Plus stroboscopes (Monarch Instrument) that were synched with the cameras. Both camera views were calibrated with a rectangular cube spanning the length of the kinematic event. Animals were filmed as they launched from a 1.5-meter elevated horizontal platform either to the ground (airborne trials) or to a vertically oriented landing platform (landing trials; 20.0 cm diameter White Ash log). The launch platform was constructed from a 5x20x30 cm pine board supported by metal shelf brackets attached to a vertical stand. The surface of the platform was covered with green indoor/outdoor carpet to prevent slipping during the launch. Launches were part of an escape response elicited by tapping the platform, immediately behind the tail.

Surface markers (5.0 mm cotton pom-poms) were glued on the occiput, left and right wrists and ankles, and base of the tail, in order to quantify kinematic movements of the forelimbs and hindlimbs during the airborne phase (Fig. 2.3). Landing kinematics were investigated by shaving the animals on the right side of the body and placing markers on the occiput, shoulder, ulna, wrist, finger, midpoint of vertebral column, hip, ankle, tibia, toe, base of tail, and midpoint of the tail (Fig. 2.3). Due to problems with skin movement, estimating the location of the elbow and knee via landmarks directly over these joints proved to be unreliable. Instead, these angles were estimated with trigonometry, using limb lengths measured from x-rays to construct two sides of a triangle and video measurement to construct the third side. Additionally, markers

(landmarks 3 and 7; Fig. 2.3) were placed slightly proximal to the wrist and ankle, in line with the elbow and knee, in order to estimate the wrist and ankle angles.

Five individual flying squirrels were filmed and five trials per individual were examined in order to quantify the kinematics of the airborne phase of locomotion (N=25).

Sample size was three individuals per species and five trials per individual during the airborne phase for chipmunks (N=15) and red squirrels (N=15). It was not possible to conduct a kinematic analysis of landing trials for the nongliding species since they refused to land directly on the tree. Typically, the nongliding squirrels changed orientation prior to contacting the platform and landed to the side of the tree making it impossible to clearly define landmarks and compare them to flying squirrels landing on front. Therefore, only landing trials from flying squirrels were analyzed kinematically

(N=25).

Horizontal distances were recorded for all of the trials and only the longest jumps for each individual where all landmarks were visible were included. Images were captured from both camera views using Ulead VideoStudio v.4.0 and imported into

APAS motion analysis software (Ariel Dynamics) for 3-D kinematic analysis. The

APAS trim module was used to synchronize the two orthogonal images based upon a shared kinematic event. For airborne trials, the frame at which the toe was last in contact with the platform was used as the synchronization point. For landing trials, the frame at which forelimb contact was made with the landing platform was the synchronization point. Sequences were digitized using an autodigitizing function in the APAS digitizing module. Once digitized, sequences were imported into the transformation module in order to convert the separate sets of 2-D coordinates into a unified set of 3-D coordinates.

Data were unfiltered prior to their input into the display module, where 3-D angles were calculated and kinematic plots were recorded. Measurement error for kinematic data was estimated to be ± 1º.

Airborne movement variables

Movement of the limbs during the airborne phase was described by measuring angles of abduction and protraction. Forelimb abduction was measured by angle ABC formed by the wrist, occiput, and a point that was projected vertically beneath the occiput

(relative to the horizontal plane defined by the calibration cube; Fig. 2.4A). It describes the movement of the forelimb with respect to an axis running dorsoventrally through the midline of the body. Adduction brings the limbs closer to the midline of the body while abduction moves them further away. Forelimb protraction describes the movement of the entire forelimb with respect to the long axis of the body and was measured by angle ABC formed by the wrist, occiput, and base of the tail (Fig. 2.4B). Forelimb abduction and protraction both result in increases in the measured angles.

Hindlimb abduction was measured in the same manner as forelimb abduction except that the hindlimb abduction angle was formed by the ankle, base of the tail, and a point that was projected directly beneath the base of the tail. Again, this parameter describes the movement of the limb with respect to an axis running dorsoventrally through the midline of the body. Hindlimb protraction was measured by the angle formed by the occiput, base of the tail, and the ankle and describes the movement of the entire hindlimb with respect to the long axis of the body. Unlike the forelimb for which protraction resulted in an increased angle, hindlimb protraction resulted in a decrease in the measured angle.

To determine the extent to which the patagium was deployed following the launch the distance from the wrist to the occiput (approximating the length of the leading edge) and the ankle to the base of the tail (approximating the length of the trailing edge) were measured for the right side of the body. In addition, patagial deployment was quantified by estimating patagial surface area during a portion of the airborne phase. Video frames were obtained from the dorsal camera view and the outline of the squirrel (excluding the tail) was digitized using SigmaScan v.2.0 (SPSS Inc.). Surface area was calculated by using a surface marker as a scale and converting area from pixels to centimeters.

To determine if angular momentum was an important factor during the airborne phase, films were examined in chipmunks (N=25), red squirrels (N=25), and flying squirrels (N=50). Frequency data were collected regarding presence or absence of rotation following the launch. In addition, frequency data were collected for flying squirrels (N=50) to determine the degree to which asymmetry characterized their horizontal launches. A test of independence was used to determine if there was a relationship between asymmetry and limb movements following the launch (Sokal and

Rohlf, 1996). In addition repeated-measures analysis of variance was used to compare features associated with aerial limb movements in flying squirrels (Systat v.6.1).

Landing movement variables

Elbow and wrist angles were measured in order to describe movement of the forelimb during the landing sequence. Elbow angles were calculated by measuring the lengths of the humerus and ulna from x-rays of the same individuals and using video measurements to obtain the distance between the shoulder and wrist markers. Wrist angles were calculated using the angle formed by the ulna marker (placed proximal to the

57 wrist, in line with the elbow), wrist, and finger. Movement of the entire forelimb was described by forelimb adduction, measured by the angle formed by the wrist, occiput, and a point that was projected directly beneath the occiput in the same manner as during the airborne trials. Knee and ankle angles were measured in order to describe the position of the hindlimb joints in a manner similar to the elbow and wrist. The angle of the vertebral column was measured as the angle formed by the occiput, midpoint of vertebral column, and base of tail. Tail movement (dorsiflexion/ventroflexion) was described by the angle formed by the midpoint of the tail, base of the tail, and a point projected directly beneath the base of the tail.

Results

Airborne kinematics

Mean kinematic profiles in chipmunks and red squirrels (Fig. 2.5 A-D) indicate gradual protraction and abduction during the airborne phase. Chipmunks do not appear to abduct the forelimbs: 45.0±3.7º (mean ± S.E.M.) at toe off to 46.5±3.7º by the end of sampling (216 ms; Fig. 2.5A). In contrast, hindlimbs are abducted about 14º from

44.6±5.2º to 58.1±7.8º (N=15). Red squirrels abduct both the forelimbs and the hindlimbs and assume a parachuting posture by the end of the sampling period (Essner,

2002). Forelimbs are abducted about 15º (55.7±4.6º to 70.9±3.3º) and hindlimbs are abducted about 30º (49.3±3.8º to 79.5±3.7º; Fig. 2.5B)

Both species exhibit protraction of the forelimbs and hindlimbs. Chipmunks protract forelimbs about 12º (70.5±5.3º to 82.6±8.6º) and hindlimbs about 21º

(117.3±7.4º to 96.6±7.3º; Fig. 2.5C). Red squirrel hindlimbs are protracted about 16º

58 from 129.8±6.8º to 114.1±5.5º but forelimbs are protracted to a much greater extent than chipmunks (56º) from 52.7±6.2º to 108.6±6.3º (Fig. 5D).

In contrast to the gradual abduction and protraction observed in the forelimbs and hindlimbs of chipmunks and red squirrels, flying squirrels initiate dramatic bilaterally and dorsoventrally asymmetrical limb rotations, hereafter referred to as fluttering, while airborne. Figure 2.6 depicts representative video frames for the initial airborne phase of flying squirrels. Initially (e.g. 0-48 ms; Fig. 2.6), the flying squirrel leaves the launch platform with a relatively flat trajectory (mean 9.4±1.1º) behaving in a similar manner to nongliding squirrels by protracting and abducting the hindlimbs, although one foot usually precedes the other. Following this (e.g. 48-80 ms; Fig. 2.6A), the flying squirrel initiates highly stereotyped fluttering behavior (observed in all trials; N=25) by rolling the back half of the body counterclockwise (as viewed from the rear), while the front half of the body simultaneously rolls clockwise (e.g. 48–80 ms; Fig. 2.6A). These initial rotations are immediately followed by a counter-rotation with the hindlimbs rotating clockwise and the forelimbs rotating counterclockwise (e.g. 80–112 ms; Fig. 2.6A).

These movements are repeated several more times with decreasing amplitude until the posture stabilizes (≈320 ms). In addition, simultaneous yawing occurs with each flutter cycle (Fig. 2.6B) as first the limbs on the left side of the body are protracted by extension of the elbow and flexion of the knee while the limbs on the right side are retracted via flexion of the elbow and extension of the knee (e.g. 80–112 ms). In addition, the left wingtip, formed by a cartilaginous strut located at the wrist, is fully extended (112 ms).

This is followed by a counter-rotation in which the left side is retracted while the right side is protracted, accompanied by extension of the right wingtip (e.g. 120–144 ms).

These movements are repeated multiple times with decreasing magnitude until a static posture is reached. While fluttering behavior is highly stereotyped (see below), there is variation in the initial direction of rotation (i.e. rotations may start in either clockwise or counterclockwise direction).

Patagial surface area is dynamic during fluttering behavior. The difference between the left and right sides (not including tail) appears to be greatest at 112 ms, when the left limbs are in close proximity. This produces extensive folding of the patagium on the left side, reducing its surface area and resulting in a 14% reduction in planar surface area for the airfoil on the left side (Protraction=39.9 ± 5.5 cm2; Retraction=46.2 ± 3.0 cm2; N=5)

A mean kinematic profile of abduction/adduction (Fig. 2.7A) indicates that initially the left and right forelimb movements are approximately 180º out of phase. The

same holds for the left and right hindlimbs. As a result, contralateral fore and hindlimbs

are approximately in phase dorsoventrally as the limbs abduct/adduct. A mean kinematic

profile of protraction/retraction (Fig. 2.7B) indicates a limb movement pattern opposite to

abduction/adduction, namely, the ipsilateral limbs are in phase anterioposteriorally. In

general, phase relationships for abduction/adduction are less consistent than they are for

protraction/retraction.

Flying squirrel kinematic profiles had the general appearance of damped

oscillations. To verify this, amplitudes were determined for successive waves of

abduction and protraction (Fig. 2.8) and tested statistically using repeated-measures

analysis of variance (ANOVA; Table 2.1). Forelimb and hindlimb amplitudes were significant for both abduction and protraction and showed a clear reduction in the peak

60 angles over time. Scheffe’s Multiple Comparison tests revealed that means decreased significantly over the course of four waves by the onset of the fourth measured amplitude. Wavelengths were also measured by comparing the time interval between successive peaks. Three out of the four waves were significant (only hindlimb wavelength during protraction was not significant, P=0.158). Forelimb protraction was characterized by a significantly smaller second wavelength. Forelimb and hindlimb abduction were characterized by a third wavelength that was considerably larger than the fourth. Thus, wavelengths did not appear to dampen in the same manner as amplitudes.

A mean velocity plot (calculated from the displacement of the landmark located at the base of the tail) constructed by averaging instantaneous velocities from all of the trials in the study (N=25) demonstrates how velocity changes during the initial part of the airborne phase (Fig. 2.9). The low vertical component of velocity at takeoff indicates the

-1 predominantly horizontally-directed nature of the launch (vinitial=0.17 ms ). Vertical

velocity toward the ground increases steadily due to the forces of gravity until it reaches -

2.97 ms-1 by the end of filming (360 ms). The horizontal component of velocity remains

-1 relatively constant (mean vinitial, vfinal=2.41 ms ) and does not appear to be noticeably

affected by air resistance during this stage of locomotion.

Angular Momentum

Nongliding squirrels exhibited a remarkably high degree of rotation about the

longitudinal axis following the launch. Rolling was evident at some point in the majority

of leaps in chipmunks (15/25 trials ≈ 60%) and red squirrels (17/25 trials ≈ 68%). Tail

rotations accompanied all of the trials that exhibited rolling. In chipmunks the majority

of rolling trials were corrected by tail rotation (13/15 ≈ 87%). In these trials the body

61 began to roll following the launch (Fig. 2.10). The tail rotated in the same direction as the roll, often repeatedly, causing the body to rotate in the opposite direction presumably to conserve angular momentum (176–240 ms; first corrective tail rotation shown). Tail rotations were sometimes preceded by apparently unhelpful tail jerks or rotations before the repetitive pattern began (64—144 ms). In contrast, in red squirrels rolling was actually initiated by tail rotation in the majority of rolling trials (10/17 ≈ 59%). In these trials the tail rotated the body out of an initially horizontal position into a banked position.

Compared with nongliding species, rolling was relatively infrequent in flying squirrels, possibly due to the short time interval between the launch and the onset of fluttering behavior (≈ 32 ms). However, there were instances where rolling was present at the launch (4/50 trials ≈ 8%). In these trials, the roll was interrupted by fluttering behavior in which the hindlimbs and forelimbs repeatedly rotated in opposing directions

(Fig. 2.11; e.g., 80 ms hindlimbs CCW, forelimbs CW). Rotations of the tail in the manner of the nongliding squirrels were not observed in flying squirrels. While all three species exhibited dorsiflexion of the tail at the launch, flying squirrels immediately ventroflexed the tail following the launch and then moved it left and right repeatedly in a relatively horizontal plane (Fig. 2.11; 128-240 ms) in a manner that did not appear to be precisely coupled with fluttering movements. Nongliding species either initiated rotations with the tail in the dorsiflexed position or they ventroflexed the tail prior to inititating rotations (Fig. 2.10).

In 20/50 trials examined in flying squirrels (40%) there was evidence of an asymmetrical pushoff. A test of independence was used to determine if there was a

62 relationship between asymmetry at the launch (left foot versus right foot) and the direction of rotation of the hindlimbs during fluttering behavior (clockwise versus counterclockwise). The majority of asymmetrical launches involved the right foot maintaining contact with the substrate for a longer period of time than the left foot (LF=4,

RF=16). In addition, the majority of hindlimb rotations were in the clockwise direction

(CW=15, CCW=5). Despite this there was no significant association between asymmetry and direction of rotation (χ2=0.060, df =1, P=0.8143).

Mean kinematic plots of leading edge and trailing edge lengths (approximated as

distance from occiput to wrist) following the launch reveal how the patagium is

outstretched during fluttering in flying squirrels (Fig. 2.12; N=5). Leading edge length

increases (0–48 ms) and then decreases slightly (48–72 ms) due to flexion at the elbow at

the onset of fluttering behavior. It then increases rapidly (72–128 ms) and is followed by

a more gradual increase as the limbs are extended slightly further during each cycle of

fluttering (128–368 ms). A plot of trailing edge length (approximated as distance from

base of tail to ankle) reveals a large degree of flexion following the launch as the

hindlimbs are protracted from their extended launch position (0–64 ms). This is followed

by a rapid increase at the onset of fluttering (64–128 ms) eventually hitting a plateau

through the remainder of the trial (128–368 ms). By the end of the trial the leading edge is longer than the trailing edge (mean length LE=7.6±0.2 cm; mean length TE=6.8±0.1 cm) giving the patagium a slightly trapezoidal shape.

Landing kinematics

The landing phase of flying squirrels is characterized by a highly coordinated approach involving simultaneous movements of the tail, vertebral column, fore and

63 hindlimbs (Fig. 2.13). The landing phase involves a nose-up pitching rotation accompanied by a steady dorsiflexion of the tail from 81±8.1° to a peak of 113±2.8° at forelimb contact (FC; defined as time zero) when it begins to ventroflex (Fig. 2.14A).

The vertebral column initially appears to be extending (-48 to -24 ms; prelanding indicated by negative values). However, this appears to be caused by a rearward rotation of the head and does not appear to represent actual movement of the vertebral column.

Nevertheless, from -24–0 ms the vertebral column begins to go into flexion, resulting in a deep arching of the back and billowing of the patagium, which peaks at 148±6.0° at FC.

Following FC, the vertebral column again begins to extend. It should be noted that these values significantly underestimate the degree of flexion in the vertebral column since the occipital landmark is elevated relative to the vertebral column. Also, the maximum curvature of the vertebral column was frequently caudal to the landmark located at the midpoint of the back.

Arching of the vertebral column is accompanied by a gradual adduction of the forelimbs from 86.2±0.74° to 48.3±2.3° at FC (Fig. 2.14B) and hindlimbs from 79.5±6.6° to 54.6±1.1° at hindlimb contact (HC; Fig. 2.14C). Following HC, both adduction angles increase as momentum carries the body forward between the limbs. Joint angles for the forelimb gradually extend from 102.6±7.5° for the elbow and 113.6±8.6° for the wrist and reach peaks of 157.8±10.1° and 146.9±1.8°, respectively, at FC. Following FC, the elbow and wrist begin rapid flexion; however, the wrist joint begins to extend once more from its minimum of 91.7±13.5° since momentum swings the body downward away from the wrist until it ends at 116.9±7.4°. In contrast, the elbow continues to gradually flex to

52.6±2.3°.

Hindlimb movement during the landing phase involves rotation from a fully abducted gliding posture to a fully adducted landing posture. The knee is initially extended at 116±10.0°. As the knee is rotated into an adducted position for landing, it is flexed to 99.6±4.3° (-24 ms), reducing the moment of inertia of the hindlimb (Fig.2.14C).

Once in position, it is extended once again to a peak of 126.5±2.0° at HC. At HC it begins to flex until it reaches 57.4±7.0°. In contrast, the ankle joint stays relatively constant as the hindlimbs are adducted (ankle 84.7±7.5°). It begins simultaneous extension along with the knee at approximately -8 ms until HC when it begins to flex to

38.5±4.6°.

Gliding velocity prior to initiation of the nose-up pitching rotation was 3.56±0.19 ms-1, whereas gliding velocity immediately prior to forelimb contact with the platform

was 3.06±0.12 ms-1. Comparing the two velocities indicates that the landing maneuver

accounts for a reduction in velocity of approximately 13%.

Discussion

Fluttering Aerodynamics

A comparison of kinematic plots between nongliding and gliding squirrels (Fig.

2.5 and 2.7) illustrates the unique complexity associated with the airborne phase of

gliding in flying squirrels. While nongliding squirrels make only gradual changes in limb

position and use rotations of the tail to control body orientation (Fig. 2.10), flying

squirrels initiate a series of rapid, coordinated limb rotations (fluttering) with primarily

horizontal movements of the tail. In contrast, such complexity is not apparent during the

launch phase for which relatively minor variation in limb kinematics has been observed

(Essner, 2002). It is readily apparent from the kinematic data presented in this study that the two halves of the patagium work in an opposing manner, possibly in the same fashion as ailerons work in aircraft (Smith, 1992). For example, as the anterior half of the body rolls counterclockwise, the posterior simultaneously rotates clockwise. This is accompanied by protraction on the left side of the body in which the left wingtip is extended. Such limb movements result in differing angles of attack for the left versus the right sides of the patagium. It can be inferred that the right side of the body, with a positive camber, produces lift in the positive direction while the left side of the body, with a negative camber, produces lift in the negative direction (Fig. 2.6A, 112 ms). The end result would be a counter-clockwise rolling moment (when viewed from the posterior). In addition, the left forelimb is pronated on its way down resulting in the hand being angled down and pointed toward the midline, extending the wingtip (Thorington et al., 1998). Extension of the left wingtip produces a ventrally directed leading edge on the left airfoil, possibly increasing the rolling moment. The right forelimb is supinated on its way up as its tip retracts. In addition, the fore and hindlimbs on the protracting (left) side of the body are generally closer together than those on the retracting side (Fig. 2.6B, 112 ms) producing a folding of the patagium that results in a 14% reduction in planar surface area for the left airfoil. The difference in surface area on the two sides of the body may further contribute to the lift differential since lift is dependent on planar surface area

(Addington et al, 2000). These movements are immediately corrected by opposing limb movements that counteract the effect, producing no net change in the position of the center of mass. They are repeated as damped oscillations that eventually attenuate into a relatively static posture.

Control of Angular Momentum

Nongliding squirrels appear to roll frequently during leaping as a result of the low moment of inertia about the longitudinal axis. They are able to counteract rolling (or initiate it) via rotations of the tail. Flying squirrels would likely have difficulty with this approach due to the aerodynamic properties associated with their dorsoventrally flattened tails (Thorington and Heaney, 1981; Schaller, 1984, Addington et al., 2000). During landing, vertical movements of the tail are associated with pitching (Fig. 2.13).

Ventroflexion of the tail produces lift in the positive direction resulting in a nose-down pitching moment while dorsiflexion produces a nose-up pitching moment (Norberg,

1990). Vertically oriented movements could have a similar effect earlier in the glide, rotating the squirrel out of its optimal angle of attack. It is also possible that a vertically oriented tail would act as a rudder, producing an aerodynamic side force that would result in undesirable yawing (Smith, 1992). In light of this, flying squirrels seem to direct the majority of tail movement within the horizontal plane. The role of tail movements in flying squirrels remains unaddressed and should be examined further to elucidate the aerodynamic properties.

The differential angle of attack produced during fluttering behavior could conceivably be used to counteract roll produced at the launch. For example, a flying squirrel rolling clockwise following a launch could rotate both hindlimbs clockwise accompanied by a counterclockwise forelimb rotation to produce a high angle of attack on the right side of the patagium. This would produce an aerodynamically induced counterclockwise roll that would cancel out angular momentum from the launch. While no direct relationship between the handedness of fluttering behavior and launch

67 asymmetry could be determined, fluttering does interrupt rolling actions (Fig. 2.11). The extent to which the limb rotations themselves (i.e., increasing moment of inertia about the roll axis and slowing down angular velocity via limb extension) or the emergent aerodynamic effects are responsible for controlling roll is currently unclear. One possible reason that a link between launch asymmetry and limb rotation during fluttering could not be established is that angular momentum produced during asymmetrical horizontal launches may be insufficient to produce a corresponding change in fluttering behavior.

Examination of vertical clinging and leaping launches would be a more appropriate test since these are associated with higher angular momentum about the longitudinal axis (e.g.

Demes et al., 1996). It is also possible that the aerodynamic torques produced during fluttering are large enough to overwhelm any angular momentum imparted at the launch.

An aerodynamic analysis of the initial airborne phase will do much to resolve this question.

Control of Fluid Dynamic Effects

Fluid dynamic models suggest that flattened, elongated animals may incur rolling oscillations due to their shape (Tanabe and Kaneko, 1994; Belmonte et al., 1998,

Belmonte, 1999). The relevance of fluid dynamics to nongliding squirrels also requires aerodynamic testing. Nevertheless, the flattening of the body observed in nongliding squirrels (evidenced by hindlimb abduction in chipmunks and fore and hindlimb abduction in red squirrels), suggest that they may be affected by fluid flow. If so, the frequently observed tail rotations in these forms, may assist in maintaining orientation.

Fluid dynamic effects are predicted to be greatly magnified in flying squirrels due to their relatively large surface area. In addition, the side-to-side oscillations would be

68 expected to become worse, as they become less elongated and more square-like, eventually leading to tumbling (completely rolling over). Fluttering behavior may represent a way of avoiding these naturally occurring side-to-side oscillations by using the limbs to produce a lift differential on the two sides of the patagium—essentially placing fluid flow under the control of the flying squirrel. This may overwhelm any natural oscillatory pattern in the same manner as was hypothesized for angular momentum produced at the launch.

The direction of the oscillations would also be expected to change with the shape of the flying squirrel. As the flying squirrel transitions to a more square-like gliding posture, the width of the gliding membrane exceeds its length (e.g., Addington et al.,

2000). At the point where they begin to experience greater elongation about the transverse axis they would be expected to shift to fore-aft oscillations since fluid dynamics produce rotations about the axis with the lowest moment of inertia. These oscillations would cause a flying squirrel to rotate nose-up or nose-down. The protraction/retraction movements associated with fluttering may represent a mechanism for dealing with such pitching moments. For example, birds are known to use fore-aft movements of the wings to control pitch (Pennycuick, 1975; Caple et al., 1983; Norberg,

1990). Moving the wings forward produces a nose-up rotation because lift is generated in front of the center of mass. In contrast, moving the wings backwards produces a nose- down rotation. While the fore-aft movements in flying squirrels are asymmetrical, they could assist in controlling fluid dynamic oscillations as was suggested for rolling movements.

This is the only known kinematic description of fluttering behavior and represents a potentially novel mechanism for maintaining orientation. Although similar alternating rotations have been observed in the giant Mastiff bat Otomops martiensseni, these bats use alternating rotations in order to perform sideslips that result in rapid height loss when descending into caves, rather than for the sole purpose of controlling orientation

(Norberg, 1976). Still, the asymmetrical limb movement patterns in Otomops keep the center of mass from veering away from the desired trajectory, in the same manner as flying squirrels.

Landing

The landing maneuver (nose-up pitching and limb adduction) not only positions the animal’s limbs for landing, but also acts to reduce gliding velocities by 13% during the short glides (≈2.0 meters) examined in this study. This reduction in velocity is associated with a pitching rotation that increases the body’s angle of attack. This is accompanied by billowing of the patagium, which together act to increase lift and drag in order to slow descent. Undoubtedly, under natural conditions the reduction in velocity due to landing behavior would be greater since animals glide faster over longer glides

(Stafford, 2002) and have the ability to initiate pitching behavior earlier in the trajectory.

Still it is noteworthy that in this study landing velocities are higher than launch velocities recorded under similar conditions (Essner, 2002). The flying squirrels in the present study were filmed on noncompliant substrates. They often use large-diameter, noncompliant substrates when moving under natural conditions as well (Essner, Chapter

3). It has been noted that primates landing on noncompliant substrates generally exhibit shorter landing durations and higher peak forces than they do when taking off

(Preuschoft, 1985; 1990) although the opposite is true with compliant substrates (Demes et al., 1995; 1999). This suggests that peak landing forces may be greater than peak takeoff forces in flying squirrels, despite their unique ability to slow descents via aerodynamic braking. If landing forces exceed takeoff forces, it may explain why flying squirrels appear to be more stereotyped in their landing behavior (single type) than in their takeoff behavior (multiple types). Future research should attempt to quantify takeoff and landing forces in flying squirrels.

In addition to increasing the angle of attack and billowing the patagium in order to slow descents, the mean kinematic profiles indicate that flying squirrels reach peak extension in the limb joints and peak flexion in the vertebral column as they contact the landing platform. This would effectively lengthen the limb segments allowing the limbs to act as more effective shock absorbers via increased deceleration time (Preuschoft et al., 1996). This represents an effective means of reducing peak landing forces. Another possible mechanism for reducing peak loads on each limb would be to distribute landing forces over all four limbs simultaneously (Preuschoft, 1990; Terranova, 1995). While the flying squirrels in this study always landed forelimbs first, the time interval between forelimb and hindlimb contact was extremely short (≈8-16 ms). Also, since the body still has momentum following forelimb contact and because the elbow is still in flexion following hindlimb contact, substrate reaction forces are undoubtedly being shared by both fore and hindlimbs (Fig. 2.14B,C). Nevertheless, the stereotypical (forelimb first) landing behavior of flying squirrels is consistent with observations of their relatively long forelimbs when compared with nongliding squirrels (Thorington and Heaney, 1981;

Runestad and Ruff, 1995). Likewise, the comparatively short forelimbs of nongliding

71 squirrels may prevent them from landing on noncompliant tree trunks during long jumps, possibly explaining their avoidance of direct contact with the landing platform in this study.

Implications for the Origin of Flight

One consequence of fluttering behavior is a forelimb movement pattern resembling a flight stroke. In flying squirrels, the arms are repeatedly flexed, supinated and drawn backward on the upstroke and extended, pronated and brought forward on the downstroke. Indeed, the high degree of similarity between movements associated with roll and pitch control and flight strokes led some to suggest them as potential pathways for the evolution of flight (Caple et al., 1983; Dudley, 2000). For example, roll control involves producing a differential angle of attack on the two forelimbs via abduction/adduction and pronation/supination, while controlling pitch involves protraction and retraction of the forelimbs (Pennycuick, 1975; Caple et al., 1983;

Norberg, 1990). However, unlike previous models where a single active control movement is applied to correct for an inadvertent body rotation (e.g. a roll to the left to correct for a roll to the right), fluttering represents a mechanism for repeated stereotypical flapping movements characteristic of powered flight. The need to control orientation would have been equally important to protobats if, as is generally accepted, they were small, quadrupedal gliders (Norberg, 1990). Likewise, if the recently discovered four- winged gliding dromaeosaur, Microraptor gui (Xu et al., 2003), is indicative of a gliding ancestry for birds, then similar control mechanisms may have been necessary in this group as well. Thus, fluttering behavior may represent a critical exaptation for the evolution of flapping flight.

Whereas fluttering provides a mechanism for a highly stereotyped and repetitive flight stroke, it is asymmetrical in nature. Interestingly, the diagonal couplet synchrony characteristic of fluttering is strikingly similar to a terrestrial trotting gait and suggests that flying squirrels have co-opted a terrestrial trotting gait for a novel aerial use. By contrast, the evolution of flight would have necessitated a transition to symmetrical limb movements, requiring alteration of the neuromuscular pattern. Nevertheless, symmetrical movements of the forelimbs are present in flying squirrels during the launching (forelimb protraction) and landing (forelimb adduction) segments of the glide. Thus, the neuromuscular patterns associated with gliding may be sufficiently complex to provide the raw material necessary for further evolutionary specialization. Unfortunately, extant bats and birds are highly derived in their flight characteristics and discerning the ancestral features of flight may be difficult. However, if fluttering truly represents an exaptation for flight, then the asymmetrical limb movements used by bats such as Otomops, may hint at possible retention of an ancestral neuromuscular pattern. An investigation of control mechanisms should be undertaken in other gliding and flying taxa to determine the ubiquity of fluttering and its relevance to models for the origin of flight.

References

Addington, T. M., Scheibe, J. S., and Hendershott, A. J. (2000). Planar surface area

and launch performance in Glaucomys volans. In Biology of Gliding Mammals

(eds. R. L. Goldingay and J. S. Scheibe), pp 199-212. Fürth: Filander Verlag.

Ando, M. and Shiraishi, S. (1993). Gliding flight in the Japanese Giant Flying Squirrel

Petaurista leucogenys. J. Mamm. Soc. Jap. 18, 19-32.

Balda, R. P., Caple, G., and Willis, W. R. (1987). Comparison of the gliding to flapping

sequence with the flapping to gliding sequence. In Proceedings of the

International Archaeopteryx Conference (eds. M. K. Hecht, J. H. Ostrom, G.

Viohl, and P. Wellnhofer), pp. 267-277. Freunde des Jura-Museums Eichstatt:

Willibaldsburg.

Belmonte, A., Eisenberg, H., and Moses, E. (1998). From flutter to tumble: Inertial

drag and Froude similarity in falling paper. Phys. Rev. Lett. 81, 345-348.

Belmonte, A. (1999). Flutter and tumble in fluids. Phys. World, 12, 21-25.

Caple, G., R. Balda, and Willis, W. (1983). The physics of leaping animals and the

evolution of preflight. Am. Nat. 121, 455-467.

Demes, B., Fleagle, J. G., Jungers, W. L. (1999). Takeoff and landing forces of leaping

and strepsirhine primates. J. Hum. Evol. 37, 279-292.

Demes, B., Jungers, W. L., Fleagle, J. G., Wunderlich, R. E., Richmond, B. G., and

Lemelin, P. (1996). Body size and leaping kinematics in Malagasy vertical

clingers and leapers. J. Hum. Evol. 31, 367-388.

Demes, B., Jungers, W. L., Gross, T. S., and Fleagle, J. G. (1995). Kinetics of leaping

primates: influence of substrate orientation and compliance. Am J. Phys. Anthrop.

96, 419-429.

Dudley, R. (2000) The Biomechanics of Insect Flight: Form, Function, Evolution.

Princeton, NJ: Princeton University Press.

Dunbar, D. C. (1988). Aerial maneuvers of leaping lemurs: The physics of whole-body

rotations while airborne. Am. J. Phys. Anthrop. 16, 291-303.

Dunbar, D. C. (1994). The influence of segmental movements and design on whole-

body rotations during the airborne phase of primate leaps. Z. Morph. Anthrop. 80,

109-124.

Emmons, L. H. and Gentry, A. H. (1983). Tropical forest structure and the distribution

of gliding and prehensile-tailed vertebrates. Am. Nat. 121, 513-524.

Essner, R. L., Jr. (2002). Three-dimensional launch kinematics in leaping, parachuting,

and gliding squirrels. J. Exp. Biol. 205, 2469-2477.

Jackson, S. M. (1999). Glide angle in the genus Petaurus and a review of gliding in

mammals. Mammal Rev. 30, 9-30.

Jouffroy, F. K., Gasc, J. P., Décombas, M., and Oblin, S. (1974). Biomechanics of

vertical leaping from the ground in Galago alleni: A cinereadiographic analysis.

In Prosimian Biology (eds. R. D. Martin, G. A. Doyle, and A. C. Walker), pp.

817-827. Pittsburgh: University of Pittsburgh Press.

Keith, M. M., Scheibe, J. S., and Hendershott, A. J. (2000). Launch dynamics in

Glaucomys volans. In Biology of Gliding Mammals (eds. R. L. Goldingay, and J.

S. Scheibe), pp. 185-198. Fürth: Filander Verlag.

MacClintock, D. Squirrels of North America. New York: Van Nostrand Reinhold, 184

pp.

Napier, J. and Walker, A. C. (1967). Vertical clinging and leaping: a newly recognized

category of locomotor behaviour of primates. Folia primatol. 6, 204-219.

Niemitz, C. (1984). Locomotion and posture of Tarsius bancanus. In Biology of

Tarsiers (ed. C. Niemitz), pp. 191-225. New York: Gustav Fischer Verlag.

Norberg, U. M. (1976). Some advanced flight manoeuvres of bats. J. Exp. Biol. 64, 489-

495.

Norberg, U. M. (1990). Vertebrate Flight. Berlin: Springer-Verlag, 291 pp.

Pennycuick, C. J. (1975). Mechanics of flight. In Avian Biology, Vol. 5 (eds. D. S.

Farner and J. R. King), pp. 1-75. New York: Academic Press.

Peters, A., and Preuschoft, H. (1984). External biomechanics of leaping in Tarsius and

its morphological and kinematic consequences. In Biology of Tarsiers (ed. C.

Niemitz), pp. 227-255. New York: Gustav Fischer Verlag.

Preuschoft, H. (1985). On the quality and magnitude of mechanical stresses in the

locomotor system during rapid movements. Z. Morph. Anthrop. 75, 245-262.

Preuschoft, H. (1990). Gravity in primates and its relation to body shape and

locomotion. In Gravity, Posture, and Locomotion in Primates (eds. F. K.

Jouffroy, M. H. Stack, and C. Niemitz), pp. 109-127. Il Sedicesimo, Firenze.

Preuschoft, H., Witte, H., Christian, A., and Fischer, M. (1996). Size influences on

primate locomotion and body shape, with special emphasis on the locomotion of

‘small mammals’. Folia Primatol. 66, 93-112.

Runestad, J. A., and Ruff, C. B. (1995). Structural adaptations for gliding in mammals

with implications for locomotor behavior in paromomyids. Am. J. Phys. Anthrop.

98, 101-119.

Schaller, D. (1984). Wing evolution. In The Beginnings of Birds: Proceedings of the

International Archaeopteryx Conference, Eichstatt (eds. M. K. Hecht, J. H.

Ostrom, G. Viohl, and P. Wellnhofer), pp. 333-348. Willibaldsburg: Freunde des

Jura-Musums Eichstatt.

Scheibe, J. S., and Robins, J. H. (1998). Morphological and behavioral attributes of

gliding mammals. In Ecology and Evolutionary Biology of Tree Squirrels (eds.

M. A. Steele, J. F. Merritt, and D. A. Zegers), pp. 131-144. Martinsville, VA:

Virginia Museum of Natural History.

Scholey, K. D. (1986). The climbing and gliding locomotion of the Giant Red Flying

Squirrel, Petaurista petaurista (Sciuridae). BIONA-report 5, 187-204.

Smith, H. (1992). The Illustrated Guide to Aerodynamics 2nd Edition. Blue Ridge

Summit, PA: Tab Books Inc.

Sokal, R. R. and Rohlf, F. J. (1995). Biometry. San Francisco: Freeman Press.

Stafford, B. J., Thorington, R. W., Jr., and Kawamichi, T. (2002). Gliding behavior of

Japanese giant flying squirrels (Petaurista leucogenys). J. Mamm. 83, 553-562.

Tanabe, Y. and Kaneko, K. (1994). Behavior of a falling paper. Phys. Rev. Letters. 73,

1372-1375.

Terranova, C. J. (1995). Leaping behaviors and the functional morphology of

strepsirhine primate long bones. Folia Primatol. 65, 181-201.

Thorington, R. W., Jr. and Heaney, L. R. (1981). Body proportions and gliding

adaptations of flying squirrels (Petauristinae). J. Mamm. 62, 101-114.

Thorington, R. W., Jr., Darrow, K., and Anderson, C. G. (1998). Wingtip anatomy

and aerodynamics in flying squirrels. J. Mammal. 79, 245-250.

Xu, X., Zhou, Z., Wang, X., Kuang, X., Zhang, F., and Du, X. (2003). Four-winged

dinosaurs from China. Nature. 421, 335-340.

Table 2.1. Results of a repeated-measures analysis of variance of fluttering kinematics in flying squirrels. Means not significantly different from each other at the P<0.05 level are connected by underlining.

Wave 1 Wave 2 Wave 3 Wave 4 P Protraction Forelimb Amplitude 21.8±0.9 21.9±1.2 20.2±1.2 15.7±1.0 0.0001* Hindlimb Amplitude 29.6±0.7 27.0±0.8 21.0±0.9 15.7±1.0 0.0001* Forelimb Wavelength 10.1±0.3 9.0±0.2 9.9±0.2 9.4±0.3 0.006*

Hindlimb Wavelength 9.7±0.2 9.6±0.2 9.6±0.2 9.2±0.2 0.158

Abduction Forelimb Amplitude 14.8±0.9 13.5±0.8 11.7±0.8 10.2±0.9 0.0001*

Hindlimb Amplitude 19.3±1.2 12.1±1.0 8.2±0.9 5.6±0.7 0.0001*

Forelimb Wavelength 9.8±0.3 9.1±0.2 10.1±0.26 9.0±0.3 0.006*

Hindlimb Wavelength 10.0±0.4 10.2±0.4 10.7±0.4 9.1±0.4 0.018*

Values are means ± S.E.M. (N=5 trials per individual, total of 5 individuals; thus N=25 trials for each mean). Significance following Bonferroni correction indicated by an asterisk.

Figure 2.1. Three principal body axes for rotational motion. Axes are orthogonal to one another, meeting at the center of mass. Rotations are classified with respect to their principal axis of rotation: (1) longitudinal rotation produces rolling; (2) transverse rotation produces pitching; and (3) vertical rotation produces yawing.

Longitudinal axis

Transverse axis

Vertical axis

Figure 2.2. Fluid dynamic effects on flat objects. Objects dropped in a fluid such as air produce vortices due to their aerodynamic properties (A). Oscillatory movements coinciding with vortex shedding are dependent on an object’s shape: elongated shapes produce periodic side-to-side rolling oscillations (A); square-like shapes produce higher magnitude tumbling oscillations (B).

A B

Side-to-side Tumbling oscillations oscillations

Modified from Belmonte et al., 1998

Figure 2.3. Landmarks used to describe limb, body, and tail movements in squirrels during airborne and landing phases: 1) occiput; 2) shoulder over the glenoid fossa; 3) point proximal to the wrist; 4) wrist; 5) base of the fifth phalanx of the manus; 6) hip over the greater trochanter; 7) point proximal to the ankle; 8) ankle at the lateral malleolus; 9) base of the fifth phalanx of the pes; 10) midpoint of vertebral column; 11) base of the tail; 12) midpoint of the tail

1 10 2 3 12 4 5 11 6

8 9

Figure 2.4. Angles of A) abduction and B) protraction were measured for the entire limb during the airborne phase in squirrels. Forelimb abduction was measured by angle ABC (wrist, occiput, point projected beneath occiput); while the hindlimb was measured in the same manner (ankle, base of tail, point projected beneath base of tail). Forelimb adduction during landings was measured using the same angle. Note that projected points are mathematical projections in the vertical direction with respect to the horizontal plane defined by the calibration cube. Forelimb protraction was measured by angle ABC (wrist, occiput, base of tail) and describes the rotation of the entire forelimb with respect to the long axis of the body. As the forelimbs are protracted angle ABC increases. In contrast, hindlimbs were measured by angle BCD (ankle, base of tail, occiput) where protracting the hindlimbs reduces the angle of protraction.

A B Abduction Protraction

Figure 2.5. Mean kinematic profiles for abduction (A,B) and protraction (C,D) of the right forelimb (z) and hindlimb ({) in chipmunks, Tamias striatus, and red squirrels, Tamiasciurus hudsonicus. Note that abduction and protraction movements in the two species occur relatively gradually.

A 120 Chipmunk Abduction 100 80 60 40 20 0 0 40 80 120 160 200 B 100 Red squirrel Abduction 80

0 0 40 80 120 160 200 C 160 40 140 Chipmunk Protraction 60 Angle (degrees) 120 80 100 100 80 120 60 40 140 20 160 0 180 0 40 80 120 160 200 D 160 60 Red squirrel Protraction 140 80 120 100 100 80 120 60 140 40 20 160 0 180 0 40 80 120 160 200

Figure 2.6. Representative video frames (16 ms intervals) in posterior (A) and dorsal views (B) portraying the initial airborne phase in flying squirrels. Initially the squirrel moves in the manner of nongliding forms (0–48 ms). Following this phase, flying squirrels begin a series of coordinated fluttering movements involving abduction/adduction (A) and protraction/retraction (B; 48–112 ms). This series of rotations is followed immediately by counter-rotations that act to cancel out the first series. The entire sequence is repeated several more times in rapid succession with decreasing magnitude until it attenuates into a static gliding posture (≈320 ms). Note that the wingtip is fully extended on the left side possibly contributing to the rolling action (B; 112 ms).

Figure 2.7. Mean kinematic profiles of fluttering behavior. Left forelimb ({), right forelimb (z), left hindlimb (U), right hindlimb (S). During abduction/adduction (A) contralateral fore and hindlimbs are initially in phase. In contrast, during protraction/retraction (B) the ipsilateral limbs are in phase. Note that the y-axis on the right has been reversed in order to depict the proper phase relationships due to the different reference points used to calculate the angles. Both plots reveal the series of damped oscillations that inevitably precede a static gliding posture. Note also that amplitudes are somewhat reduced from actual values due to slight variation in timing across trials.

A Abduction 140 120 100 80 60 40

Angle (degrees) Angle (degrees) ) 20 0 40 80 120 160 200 240 280 320 360 rees g

de B Protraction ( 140 40 Forelimb Angle 120 60 100 80 80 100 120 60

140 (degrees) 40 16 20 180 Hindlimb Angle 0 40 80 120 160 200 240 280 320 360 Right forelimb Left forelimb Time (ms) Right hindlimb Left hindlimb

Figure 2.8. Amplitudes and wavelengths during fluttering locomotion. A total of four amplitudes (A1-A4) and wavelengths (λ1-λ4) were measured during fluttering to determine the degree to which they damp out over time. Amplitudes were measured as the distance from a peak or trough to the midpoint (dotted line). Midpoints were determined by taking the average angle on either side of a peak or trough and then taking the average of those two values. The average midpoint was subtracted from the minimum or maximum (peak or trough) values to determine the amplitude. Wavelengths were measured as the time between adjacent peaks or troughs.

180

160

140 A1 A3 λ2 λ4 120

100 λ3

Angle (degrees) λ1 80 A4 60 A2 0 40 80 120 160 200 240 280 320 360 400 Time (ms)

Figure 2.9. Mean velocity plots demonstrating how gliding velocity changes throughout the glide path. The horizontal component of velocity (■) stays relatively constant following the launch. The vertical component of velocity (z) increases steadily in the negative direction due to gravitational acceleration.

4.0

) 2.0 -1 Horizontal

0 Vertical

-2.0 Velocity (ms

-4.0 0 40 80 120 160 200 240 280 320 360 Time (ms)

Figure 2.10. Representative video frames (16 ms intervals) in posterior view portraying the rolling movement characteristic of the majority of trials in chipmunks. The chipmunk is rotating in the clockwise direction following the launch. Immediately following the launch this individual rotated the tail counterclockwise (presumably making the roll worse). Rotation is corrected by a series of three tail rotations also in the clockwise direction (176–240 ms; only first clockwise tail rotation depicted). These movements act to correct the orientation of the animal and conserve angular momentum.

0 ms 16 ms 32 ms 48 ms

64 ms 80 ms 96 ms 112 ms

128 ms 144 ms 160 ms 176 ms

192 ms 208 ms 224 ms 240 ms

Figure 2.11. Representative video frames (16 ms intervals) in posterior view portraying the rolling movement characteristic of 8% of trials in flying squirrels. Rolling is interrupted by the onset of fluttering behavior. In this example, hindlimbs rotate clockwise (48-80 ms), counterclockwise (80-128 ms), and then repeat. Forelimbs exhibit opposing rotations.

0 ms 16 ms 32 ms 48 ms

64 ms 80 ms 96 ms 112 ms

128 ms 144 ms 160 ms 176 ms

192 ms 208 ms 224 ms 240 ms

Figure 2.12. Mean kinematic plot of leading edge and trailing edge lengths measured as the distance from the occiput to the wrist (leading edge) and base of the tail to the ankle (trailing edge).

10 9 Leading Edge 8 7 6 Trailing Edge 5 4 Length (cm) 3 2 0 40 80 120 160 200 240 280 320 360 Time (ms)

Figure 2.13. Landing sequence in flying squirrels traced from actual video frames (16 ms intervals). Landing behavior in flying squirrels is characterized by a highly coordinated approach involving simultaneous movements of the tail, vertebral column, fore and hindlimbs. Note that the frame at which the forelimbs contact the landing platform is designated as time zero.

-48 ms 0 ms

-32 ms 16 ms

-16 ms 32 ms

Figure 2.14. Mean kinematic profiles of the landing phase of flying squirrels. Tail (¥), vertebral column (¡), wrist (T), elbow (), forelimb adduction (²), knee (z), ankle (S), hindlimb adduction (O). Landing behavior involves a nose-up pitching motion induced by a steady dorsiflexion of the tail and arching of the vertebral column (A), accompanied by adduction and extension of the limbs (B,C) prior to their contact with the platform. Forelimbs make initial contact with the platform (FC), followed immediately (duration=8–16 ms) by the hindlimbs (HC). After contact with the platform the tail ventroflexes, vertebral column straightens out, and the limbs go into flexion (with the exception of the wrist which initially flexes before returning to extension).

A FC HC 180 Tail/Vertebral Column 160 140 Vertebral Column 120 100 80

Angle (degrees) 60 Tail 40 B 180 Forelimb 160 140 Wrist 120 100 X X 80 X X

Angle (degrees) X 60 X Elbow X X Adduction X X X 40 Hindlimb C 140 120 Knee 100 80 O O O Ankle O O O 60 O O O O Adduction O Angle (degrees) 40 20 -48 -40 -32 -24 -16 -8 0 8 16 24 32 Time (ms)

Chapter 3. Morphology, Locomotor Behavior, and Habitat Use in North American

Squirrels

Introduction

The modification of design with evolutionary novelty often results in corresponding changes in the other features that define an organism. Bats provide a good example of this complex hierarchical relationship. Wing shape has been shown to have major implications for bat ecology and behavior (Norberg, 1990). For example, high- aspect ratio/low wing loading designs enable relatively inexpensive flight over long distances. However, long, narrow wings do not allow for flight through dense vegetation and limit foraging to relatively open areas (Norberg, 1994). Thus, a thorough investigation of design requires an approach that integrates the multiple levels that define an organism and reveals the key interactions across levels (see Reilly and Wainwright,

1994). This technique has been used successfully in a multitude of interspecific comparisons including: morphology, kinematics, motor patterns, performance, and habitat use (e.g., Wainwright, 1987; Losos, 1990; Reilly and Lauder, 1992;

Vanhooydonck et al., 2002). Such studies have enabled the discovery of causal links across levels (i.e., biomechanical causation; Lauder, 1991) and provided compelling evidence of adaptation.

Gliding locomotion is an evolutionary novelty that has also benefited from the multilevel perspective. Emerson and Koehl (1990) and Emerson et al. (1990) examined interactions among morphology, posture, and performance in order to reconstruct the sequence of events leading to gliding in frogs. They determined that manipulating

94 morphology or posture independently had dramatic effects on locomotor performance.

The functional demands of gliding are directly linked to morphological traits via biomechanics/aerodynamics and should have important ramifications for behavioral and ecological (habitat) traits as well. Nevertheless, fundamental aspects of character variation among gliding forms remain unexplored, to varying degrees, at each of these levels.

The rodent family Sciuridae (squirrels) provides an ideal system for examining the evolution of locomotor novelty. Ancestrally, sciurids were arboreal, a condition retained by modern tree squirrels (Thorington et al., 1998). The morphology and behavior of modern tree squirrels are thought to be characterized by a high degree of evolutionary conservatism (Emry and Thorington, 1984; Roth, 1996; Youlatos, 1999).

Morphologically, the oldest known squirrel, Douglassciurus (formerly Douglassia, Emry and Korth, 1996, 2001) from the Early Oligocene, differs little in its postcranial anatomy from the modern fox squirrel (Sciurus niger) except in minor aspects of joint construction. In other respects (e.g., size and limb proportions) they are strikingly similar

(Emry and Thorington, 1984; Table 3.1). This led Emry and Thorington (1984) to suggest that the ecology of Douglassciurus is best represented by modern tree squirrels.

Two major transitions from the ancestral arboreal condition have occurred in the sciurid clade. In addition to a hypothesized evolutionary transition from arboreal leaping (or parachuting) to gliding in flying squirrels (Savile, 1962) chipmunks and ground squirrels have become secondarily semiarboreal/terrestrial (Emry and Thorington, 1982;

Thorington et al., 1998). In both cases, the suite of morphological, behavioral, and ecological features associated with these novelties has yet to be fully examined.

Variation at the morphological level has received the most research attention in sciurids. Accordingly, a number of studies have focused on comparisons of morphological variation among gliding, arboreal, and semiarboreal/terrestrial forms.

These studies have revealed similarities between gliding and arboreal squirrels in some features and between arboreal and semiarboreal/terrestrial squirrels in others (Bryant,

1945; Scheibe et al., 1990; Thorington et al., 1997). However hypotheses regarding causal links between form and function remain untested due in part to a paucity of data at the behavioral and ecological levels.

Evidence exists in many arboreal forms (e.g. primates) that differential habitat use among taxa is consistent with morphological variation (e.g., Fleagle, 1976; Ward and

Sussman, 1979; Fleagle and Mittermeier, 1980). Likewise, relationships between habitat structure and locomotor behavior have been well documented (e.g., Moermond, 1979;

Pounds, 1988, 1991; Gebo and Chapman, 1995; Warren and Crompton, 1997). Recent studies have begun to examine the relationship between morphology, behavior, and ecology in squirrels, giving reason to suspect that biomechanical relationships observed in primates and other arboreal taxa may not be applicable to the sciurid clade (Youlatos,

1999; Stafford et al., 2003).

Youlatos (1999) compared locomotor behavior and habitat use in two tropical tree squirrels, Sciurus igniventris and Microsciurus flaviventer, and found a lack of concordance with a priori predictions based upon biomechanical models of limb morphology. Small squirrels such as Microsciurus were predicted to leap more than larger squirrels due to their relatively longer hindlimbs and use smaller diameter supports due to their smaller arm spans. Biomechanical theory predicts that elongated hindlimbs

96 will increase the distance through which acceleration occurs during leaping, resulting in increased takeoff velocity and horizontal distance (Emerson, 1985). Moreover, increased arm span is predicted to increase the normal component of the adductive force generated by the forelimbs, helping to drive claws into bark during climbing, and enabling negotiation of relatively large diameter supports (Cartmill, 1974, 1985; Thorington and

Thorington, 1989). Neither of these predictions were supported in squirrels.

Microsciurus leapt significantly less and used large diameter supports more frequently than the much larger Sciurus (Youlatos, 1999). However a third hypothesis predicting that Microsciurus would climb more frequently due to relatively long forearms that are thought to aid in forelimb protraction (Thorington and Thorington, 1989) was supported.

Theoretical biomechanical predictions were also not supported in a study by

Stafford et al. (2003) who determined that Petaurista leucogenys (Japanese giant flying squirrel) made more frequent use of smaller supports than either Microsciurus or Sciurus, despite their larger size and arm span. This led to the suggestion that ecological differences (e.g. foraging) may have a greater influence on habitat use than morphology

(Stafford et al., 2003). Moreover, they determined that Petaurista rarely used vertically oriented supports, contradicting the presumed relationship between gliding and vertical substrates that has been incorporated into models of gliding energetics (Scholey, 1986;

Scheibe and Robins, 1998). Thus, the relationships across morphological, behavioral, and ecological levels in sciurids appear to be complex and require additional investigation.

In order to address multilevel relationships in sciurids, this study will examine morphology, behavior, and habitat use under controlled laboratory conditions in three

North American sciurids: eastern chipmunk, Tamias striatus; red squirrel, Tamiasciurus hudsonicus; and southern flying squirrel, Glaucomys volans. Ideally, laboratory studies of behavior and habitat use are accompanied by data from the field (Pounds, 1991).

However, a pilot study designed to observe flying squirrels at night using night vision equipment proved unfeasible. Controlled laboratory studies provide data that would otherwise remain anecdotal and present the advantage of comparing species under identical conditions, controlling for the proximate effects of habitat structure (Pounds,

1988, 1991; Dagosto and Gebo, 1998). Field studies have uncovered many examples of substantial intraspecific variation in behavior and habitat use due to local habitat conditions, seasonality, etc. (see Dagosto and Gebo, 1998 for a review). The present study is not intended to reflect the behavior and habitat use that would be observed under the diverse array of conditions in which these species are found. Rather it is intended to identify interspecific patterns in species moving under identical conditions.

The primary goal of this study is to compare locomotor characteristics of these three species by defining the morphospace, ethospace, and ecospace in which they exist.

1974; Oshida et al., 1996; Roth, 1996). Given that chipmunks and flying squirrels represent two major evolutionary radiations from the ancestral tree squirrel condition, they are predicted to define extreme regions of multivariate space. Red squirrels are included in the study as representatives of the modern tree squirrel clade. They are smaller than other tree squirrels but share similar limb proportions (Thorington and

Heaney, 1981; Table 3.1). As tree squirrels, they are considered to retain ancestral features of morphology, behavior, and ecology, and are predicted to occupy intermediate positions in these spaces (Fig. 3.1). The multilevel approach will also assist in establishing ecological relevance, elucidating causal links between morphology, behavior, and ecology, and enabling the application of biomechanical criteria (Lauder,

1991; Wainwright, 1994) to test whether a diverse group of sciurids varying in limb morphology fit the biomechanical predictions discussed previously.

Materials and Methods

Behavioral and Ecological Data

Chipmunks (Tamias striatus), red squirrels (Tamiasciurus hudsonicus), and flying squirrels (Glaucomys volans) were collected from the wild and maintained in a colony at

Ohio University. A 9 m x 7 m x 7 m enclosure was constructed inside a handball court using trees and branches of various heights, diameters, and orientations (Fig. 3.2).

Supports were arranged to represent the typical regrowth deciduous forest habitat in which the three species were originally collected (Essner, pers. obs.). A 2.5-cm layer of hardwood mulch and dried leaves were scattered on the floor of the enclosure in order to mimic natural conditions. The walls of the handball court were painted black and a barrier was constructed from 1.2-meter high sheet rock and erected across the width of the enclosure to prevent the animals from escaping the study area. Two Sony Hi-8 camcorders were connected to VCRs and arranged orthogonally so that the field-of-view covered the entire enclosure. Ten individuals from each species were placed in the enclosure independently and allowed to habituate for a six-hour period. Following the

99 habituation period, VCRs were timed to record their movements for an additional six hours. Diurnal squirrels were filmed with the aid of halogen lamps while the nocturnal flying squirrels were filmed using red-filtered lamps so that their behavior would be unaffected. Food and water were not placed in the enclosure since their presence may have affected movement. Therefore, animals were given food and water immediately before and after trials.

Behavioral and ecological data were collected by reviewing videotapes in the lab and using the bout method of continuous sampling. Bouts were defined as: 1) discrete movements where the animal travels at least one body length, 2) movements separated by at least a three-second interval from other movements, and 3) movements involving a change in support type (Rosenberger and Stafford, 1994). In order for an individual to be included in the statistical analyses it had to provide at least 200 bouts during the six-hour filming period. Animals that failed to meet this cutoff were excluded from further analysis. This reduced the number of individuals included in the study to six per species resulting in a total of 1200 bouts per species (3600 bouts total). Bouts were then partitioned into discrete locomotor behavior and habitat use categories for analysis.

The examination of locomotor behavior focused on seven discrete locomotor variables: climb up, climb down, run ground, run branch, run incline, leap, and parachute/glide. Climbing was defined by movement up or down a vertically oriented trunk. Locomotor gaits (e.g. bounding) could not be determined reliably due to the small size of the animals and their distance from the cameras. Therefore, movements on horizontally oriented substrates were classified as either run ground or run branch, depending upon height, while movements on obliquely oriented substrates were classified

100 as run incline. Leaping was defined as negotiating a gap without fully abducting the limbs. Leaps were typically performed over short horizontal distances and involved only a minor loss in height. In contrast, parachuting and gliding were defined as negotiating gaps with limb abduction. These movements were typically used over long horizontal distances and involved a significant loss in height (Essner, 2002). Both parachuting and gliding make use of lift and drag to control descents, however, a distinction is generally made based upon the steepness of the angle (e.g.., Oliver, 1951; Rayner, 1981). In this study, no attempt was made to discriminate between the two locomotor modes.

A total of nine habitat use variables were investigated by marking the location of the individual during each locomotor bout. Prior to filming, measurements were recorded for each substrate type (i.e. height, diameter, and orientation) within the enclosure.

Following this, a photograph of the entire enclosure was color-coded to assist with habitat categorization during tape review. Height was partitioned into ground (0 m), low (0 – 4 m) and high (> 4 m); substrate diameter into small (< 5 cm), medium (5 – 15 cm), and large (> 15 cm); and substrate orientation into horizontal (0 ± 30°), oblique (30 - 60°), and vertical (90 ± 30°; Fig.3.2).

Morphological Data

Morphological data were collected from caliper measurements on external limb elements taken from 48 individuals (chipmunks, n=13; red squirrels, n=15; flying squirrels, n=20). A total of six morphological variables measured from the right side of the body were investigated including: forefoot length (radiocarpal joint to end of longest phalanx), forearm length (olecranon to radiocarpal joint), upper arm length

(glenohumeral joint to lateral epicondyle), hindfoot length (calcaneus to end of longest

101 phalanx), shank length (tibiofemoral joint to lateral malleolus), and thigh length (greater trochanter to tibiofemoral joint).

Statistical Analysis

Proportions were calculated for each behavioral and habitat use category in order to facilitate multispecies comparisons. All proportions were arc sine transformed prior to statistical analysis in order to meet statistical requirements (Sokal and Rohlf, 1995).

Statistical analysis involved ordination of species in multidimensional space, reducing the dimensionality of the data. In order to define morphospace without the confounding effects of body size, size was removed from the series of six linear morphological measurements using the technique of Mosimann and James (1979). The six raw morphological measures of each individual were log transformed, summed, and divided by the total number of measurements in order to produce a log-size component. Each measurement was divided by its log-size value to yield a size-free estimate of that measurement.

Size-free measurements were ordinated using principal components analysis

(PCA) run on the covariance matrix (NCSS2000). Multivariate analysis of variance

(MANOVA) was applied to the factor scores to test for significant differences in character states among taxa. Post hoc tests (Scheffe’s Multiple-Comparison Test) were used to determine specific taxa responsible for significance. In contrast, the ethospace and ecospace were defined by collecting locomotor behavior and habitat use data based upon frequencies of occurrence and ordinating using correspondence analysis (CA;

NCSS2000). Correspondence analysis was designed for analyzing contingency tables to measure the degree of correspondence between rows and columns (Greenacre, 1984). It

102 is similar to PCA in that it reduces the dimensionality of a dataset and produces a graphical display that can be a valuable interpretative tool. MANOVA and post hoc tests were applied to the factor scores from CA to test for significant differences in character states among taxa in the same manner as the morphological data.

Results

Morphology

Morphological measurements for the three species of squirrels are presented in

Table 3.2. The size removed principal components analysis of the six limb variables

(Fig. 3.3) indicates that factor 1, which explains 72% of the variance, is describing a contrast between proximal and distal limb elements (Table 3.3). Flying squirrels possess long upper arms, forearms and shanks and short fore and hind feet. In contrast, chipmunks possess short upper arms, forearms, and shanks and long fore and hind feet.

Red squirrels are located in an intermediate position, but are most similar to chipmunks.

Factor 2 loads for thigh length, explaining an additional 10% of the variance but failing to discriminate among species. A multivariate analysis of variance (MANOVA) of factor scores, along with post hoc tests indicate that the three species differ significantly (Wilks’

λ=0.056, P<0.00001). Factor 1 discriminates the three species based primarily upon forearm and hindfoot length (F = 335.9, d.f. = 47, P < 0.0001). In contrast, the three species do not differ significantly on Factor 2 (F = 0.21, d.f. = 47, P = 0.815).

Locomotor Behavior

Proportions for each locomotor behavior (Fig. 3.4) show that the greatest differences among the species are in run ground, climb up, leap, and parachute/glide.

103

The predominant behavior of chipmunks was ground running (29.9%) when compared with red squirrels (15.2%) and flying squirrels (6.8%). Red squirrels used predominantly run branch (27.9%). Flying squirrels exhibited the greatest reliance on climbing up

(28.4%), followed by red squirrels (19.1%) and chipmunks (17.0%). Leaping was observed most often in flying squirrels (9.3%) and red squirrels (8.7%), and less frequently in chipmunks (3.9%). Finally, flying squirrels were the only species to parachute/glide (7.2%). One striking observation was that red squirrels did parachute from the highest point on the tree to the ground (≈ 7 meters) on multiple occasions during recapture attempts. However, this was never observed as part of the normal behavioral repertoire and was not included in the analysis.

Since parachuting or gliding was only observed in flying squirrels, it would have explained almost all of the variance in locomotor behavior among the taxa. Therefore, this category was combined with leaping to form a new category referred to as “aerial” for further analysis. An ordination plot of locomotor behavior obtained from the correspondence analysis (empty symbols equal species means; Fig. 3.5) indicates that factor 1, which accounts for 66% of the variance, loads for run ground and aerial (Table

3.4). A MANOVA of factor scores and post hoc tests indicate that the three species are significantly different (Wilks’ λ=0.155, P<0.0001). The three species differ significantly from one another on Factor 1 (F = 22.14, d.f. = 17, P < 0 .001). Factor 2 explains an additional 16% of the variance and loads for aerial and climb down. Thus, run ground is associated with chipmunks, aerial is associated with flying squirrels, and red squirrels are located in an intermediate position, but there is no significant difference among species on factor 2 (F = 1.56, d.f. = 17, P = 0.243).

104

Habitat use

Proportions for habitat use show that chipmunks (30.7%) move on the ground more frequently than red squirrels (16.1%) or flying squirrels (8.2%; Fig. 3.6A). In contrast, flying squirrels (33.2%) and red squirrels (32.2%) move high in the trees more often than chipmunks (15.1%). There is a similar use of large diameter supports among flying squirrels (48.1%) and chipmunks (48.3%) while red squirrels make more frequent use of medium sized supports (36.6%; Fig. 3.6B). Finally, there is a tendency for flying squirrels to use vertical substrates (51.0%) while horizontal substrates are used more frequently by chipmunks (53.0%) and red squirrels (47.9%; Fig. 3.6C).

The correspondence analysis of habitat use (empty symbols equal species means;

Fig. 3.7) indicates a height-based discrimination along factor 1 and a limb diameter-based discrimination along factor 2 (Table 3.5). Factor 1 explains 60% of the variance and indicates a correspondence between ground and chipmunks, and between high and both flying squirrels and red squirrels. Factor 2 explains 18% of the variance indicating a correspondence between large diameter supports and chipmunks and flying squirrels, while medium diameter supports correspond with red squirrels. A MANOVA of factor scores indicates a significant difference among species (Wilks’ λ=0.241, P<0.0005). Post hocs indicate chipmunks (ground) differ significantly from both flying squirrels and red squirrels (high) on factor 1 (F = 12.06, d.f. = 17, P < 0.001). Red squirrels (medium diameter supports) are significantly different from both flying squirrels and chipmunks

(large diameter supports) on factor 2 (F = 4.40, d.f. = 17, P < .05). In multivariate space, red squirrels make greater use of medium supports high in the canopy, flying squirrels make greater use of large supports high in the canopy, and chipmunks make greater use

105 of large supports near the ground. There are no significant differences among the species on factor 3 (F = 0.62, d.f. = 17, P = 0.5489).

Discussion

Ecological Relevance

The classification of chipmunks as semiarboreal appears to fit (e.g. Bryant, 1945).

A considerable proportion of their bouts occurred on trees (69.2%) compared with red squirrels (83.4%) and flying squirrels (91.8%; Fig. 6A). In addition, most of their movements occurred on supports close to the ground (e.g. downed logs). Only 15.1% of their movements were high in the canopy. Their use of arboreal habitat may have been affected by unfamiliarity possibly leading them to spend more time in trees than they would normally in order to aid in vigilance. Moreover, chipmunks did not have access to their burrow systems where they are known to spend a substantial portion of their time in the wild (Elliott, 1978). Nevertheless, they were significantly different from the other two species due to their extensive use of ground-based locomotion and habitat.

The description of flying squirrels as highly arboreal also appears to be appropriate (e.g. Chickering, 1996). The flying squirrels in this study visited the ground only occasionally (8.2%; Fig. 6A), typically at the termination of a glide. While on the ground, they usually moved directly to the nearest tree. This proportion may be somewhat greater than what would occur naturally due to the relative unfamiliarity of animals with their surroundings. It has been suggested that flying squirrels utilize learned glide paths (Wells-Gosling, 1985; Ando and Shiraishi, 1993) that take time to establish.

The flying squirrels in the present study landed on the ground the majority of the time.

106

Terminating a glide on the ground would seem to offer greater initial predictability until glide routes ending on vertical trunks are established. Nevertheless, a close relative, the northern flying squirrel (Glaucomys sabrinus) visits the ground frequently to feed on fungi (Goldingay, 2000). In contrast, Japanese giant flying squirrels (Petaurista leucogenys) were never observed on the ground by Stafford et al. (2003). Southern flying squirrels (Glaucomys volans) probably lie somewhere between these extremes.

Red squirrels were intermediate relative to chipmunks and flying squirrels in their degree of arboreality. While they do fit their description as arboreal, using trees in 83.4% of bouts, they are not as arboreal as flying squirrels (91.8%; Fig. 3.6A). The description of red squirrels as parachuters also seems to fit (Savile, 1962; Essner, 2002). While not observed within the context of the study, it was observed frequently as an escape response during recapture. The circumstances under which parachuting is initiated in the wild are currently unclear. The data presented in the current study lead to the prediction that parachuting has relevance in the wild as a predator avoidance strategy and not as a locomotor behavior.

Gliding comprises a considerable amount of the behavioral repertoire of flying squirrels, occurring in 6.8% of locomotor bouts. This is comparable to data presented for

Petaurista (7%; Stafford et al., 2003). The frequency with which gliding is initiated is relatively impressive considering that glides are single events that cannot be reinitiated until the animal climbs back up a tree. In contrast, locomotor behaviors such as climbing are frequently reinitiated following a pause or a change in support type. Also noteworthy are the higher frequencies of climbing up observed in flying squirrels (28.4% versus

17.0% in chipmunks and 19.1% in red squirrels) and the lower frequencies of climbing

107 down (14.6% in flying squirrels versus 17.1% in chipmunks and red squirrels). While not statistically significant, they suggest a relationship between gliding and climbing that may bear on models of energetics that have tested the efficiency of gliding versus climbing (Scholey, 1986; Scheibe and Robins, 1998).

Biomechanical Predictions

Table 3.6 lists biomechanical predictions associated with other studies of squirrel morphology, behavior, and ecology. The prediction that small squirrels should leap more than large squirrels due to their relatively longer hindlimbs was supported (Table 3.6).

While chipmunks are classified as small squirrels, they do not represent a valid test of the hypothesis since they do not possess the elongated hindlimbs of small arboreal squirrels.

Flying squirrels have elongated these elements (Fig. 3.3) and do leap slightly more frequently than the larger red squirrel (9.3% versus 8.7%). However, the difference between flying and red squirrels is even clearer if gliding, which also involves hindlimb propulsion (Essner, 2002), is included (16.5% versus 8.7%).

The prediction that small squirrels use smaller supports was not supported (Table

3.6), consistent with other studies of squirrel positional behavior (Youlatos, 1999;

Stafford et al., 2003). In the present study the two smallest species, flying squirrels and chipmunks, used large diameter supports more frequently than the largest species despite their absolutely smaller arm spans. Claws make it possible for squirrels to climb infinitely wide supports, unlike most primates that rely on friction grips for climbing

(Cartmill, 1974; 1985). Nevertheless, Thorington and Thorington (1989) predicted that squirrels might have problems climbing relatively large trunks if they possessed smooth bark that made embedding claws difficult. No attempt was made in the present study to

108 include a diverse array of textures. Therefore, it is not possible to evaluate the extent to which bark texture had an effect on the results. Stafford et al. (2003) found that

Petaurista, a large squirrel, used smaller supports than the smaller tree squirrels studied by Youlatos (1999) and suggested that ecology may be of greater importance in determining support use than morphology. This would seem to be a reasonable assertion, but given the patterns observed in studies thus far, allometry should not be disregarded.

Squirrels with elongated forearms were predicted to climb more frequently (Table

3.6). Muscles from the trunk insert almost to the midshaft of the humerus of tree squirrels, limiting mobility during climbing (Thorington and Thorington, 1989).

Elongated forearms were predicted to be an adaptation for overcoming the lack of mobility during climbing by increasing forelimb protraction (Thorington and Thorington,

1989; Youlatos, 1999). This prediction was supported (Table 3.6). Flying squirrels, with elongated forearms, climbed up trees more frequently than the other two species

(although they climbed down less frequently, presumably as a result of gliding).

Elongation of the forelimbs is a pattern frequently observed in mammalian gliders that results in increased patagial surface areas (Thorington and Heaney, 1981; Runestad and

Ruff, 1995; Jackson, 1999). However, the degree to which the forearm is lengthened relative to the upper arm among gliders may be somewhat dependent on the site of patagial attachment (e.g. wrist/digit versus elbow attachment sites; Runestad and Ruff,

1995). An examination of climbing frequencies among gliding forms that vary in forelimb proportions would be an intriguing test of the relationship between forearm elongation and climbing.

109

The prediction that gliding forms would use vertically oriented supports more frequently was also supported (Table 3.6C). Petaurista leucogenys apparently does not fit this pattern and uses horizontally oriented supports more frequently. This led Stafford et al., (2003) to suggest that models of gliding energetics that incorporate the cost of vertical climbing are overemphasizing its importance (Scholey, 1986; Scheibe and

Robins, 1998). The present study indicates that such models may be suitable for

Glaucomys volans, which uses vertical supports during the majority of bouts (51%).

Petaurista leucogenys was found to use small supports primarily and often engaged in scrambling behavior (moving on terminal branches) during foraging for the leaves and buds which comprise 36% of its diet (Kawamichi, 1997; Goldingay, 2000). Glaucomys volans is much more of a dietary generalist than Petaurista leucogenys and does not engage in folivory to a great extent (Goldingay, 2000) so variation in preference for support type might be expected as a result of foraging differences. Perhaps of greater import are the observations by Ando and Shiraishi (1991) that Petaurista leucogenys lacks the ability to descend trunks head first. Small flying squirrels such as Glaucomys do not appear to have such problems which may explain the greater use of vertical trunks in the present study (e.g. climbing down accounts for 14.6% of locomotor bouts in G. volans). This reiterates the potentially important variation among gliding forms in locomotor behavior and habitat use and the need for further study.

Additional Predictions

Shock Absorption

Gliding is a locomotor behavior unique among sciurids to flying squirrels. The suite of morphological changes associated with gliding has resulted in their distinct

110 position in morphospace (Fig. 3.3). Flying squirrels have elongated the proximal limb elements (except for the thigh) and shortened distal limb elements (fore and hind feet) relative to the ancestral condition. Elongating the fore and hindlimbs increases patagial surface area, directly improving gliding performance by increasing the amount of lift or drag that can be generated (depending upon angle of attack; Norberg, 1990). However, the impressive morphological changes associated with gliding raise additional biomechanical questions that require further study.

While flying squirrels in this study generally landed on the ground, they prefer vertical trunks as landing sites in the wild (e.g. Scholey, 1986; Ando and Shiraishi, 1993;

Vernes, 2001). In either case their preferred landing sites are relatively noncompliant.

The elongated forelimbs and hindlimbs of flying squirrels lead to the biomechanical prediction that they play an important role in shock absorption during landings by increasing deceleration time, resulting in lower peak forces (Preuschoft et al., 1996).

Additional evidence of a potential role in shock absorption is provided by the landing kinematics presented by Essner (Chapter 2). A key finding of that study was that fore and hindlimbs reached peak extension as they contacted the landing substrate suggesting an attempt to maximize deceleration distance. The forces associated with landings have never been measured in gliding mammals and the degree to which limb elongation affects impulse (area under force time curve) at landing requires additional investigation.

Reducing Distal Mass and Gliding

Another finding was that flying squirrels have reduced the size of their most distal elements. This was also noted by Scheibe et al. (1990) in a comparative investigation of morphological variation across gliding, arboreal, and terrestrial mammals. Given their

111 greater reliance upon leaping and gliding, flying squirrels might have been expected to have longer hind feet than chipmunks, in order to maximize takeoff velocity (Emerson,

1985). This leads to the question of the role of the distal elements during gliding. One possibility is that shortening both fore and hindfeet results in lower mass on the distal part of the limb. Essner (Chapter 2) describes rapid, coordinated rotations of the fore and hindlimbs (fluttering) immediately following takeoff. The presence of a large mass on the distal limb elements would hinder such rapid limb movements (Preuschoft, 1990) and have dramatic implications for a number of gliding maneuvers, including fluttering, turning, and landing. Experimentally increasing the mass of the distal elements (e.g. attaching weights to hands and feet) may be an effective way of determining effects on gliding performance.

Multilevel Comparisons

The prediction that chipmunks and flying squirrels would occupy extreme regions of multivariate space compared with red squirrels was generally supported. All three species differed significantly at the morphological level, with flying squirrels exhibiting the greatest degree of divergence in limb morphology (Fig. 3.8). Flying squirrels have elongated the upper arm, forearm, and shank and have shortened the fore and hindfeet.

Chipmunks have made the opposite changes placing them at the other extreme.

However, they have diverged less from the tree squirrel condition (e.g. red squirrels) thought to be close to the ancestral morphology for the sciurid clade (Emry and

Thorington, 1984). Flying squirrels and chipmunks also diverged at the behavioral level, occupying opposite regions of ethospace. Flying squirrels occupied a region defined by aerial locomotion while chipmunks were located in a region defined by ground running

112 locomotion. Thus, red squirrels were intermediate in both morphospace and ethospace.

Ecologically, red squirrels occupied a distinct region of multivariate space but exhibited overlap with flying squirrels in their use of high supports (> 4-meters). Chipmunks diverged in the opposite direction using ground habitat more frequently. Interestingly, chipmunks and flying squirrels differed from red squirrels by their greater use of large diameter supports (Fig. 3.8). However, the use of large diameter supports may not be comparable since flying squirrels also had a greater tendency to use vertical supports

(51.0% in flying squirrels versus 35.2% in chipmunks) while chipmunks made greater use of horizontal supports (53.0% in chipmunks versus 35.7% in flying squirrels), indicating differing locomotor behaviors are involved. This may indicate that chipmunks do not represent a valid test of the biomechanical relationship between body size and support size discussed earlier.

Ecologically, the three species are distinct, having partitioned the habitat according to height and support diameter (and possibly support orientation, see above).

Chipmunks make greater use of large supports near the ground, red squirrels make greater use of medium supports high in the canopy, and flying squirrels make greater use of large supports high in the canopy. Lines connecting taxa across ecological, behavioral, and morphological levels (Fig. 3.8) show that ecological separation is consistent with variation at the morphological and behavioral levels. Furthermore, each taxon occupies a distinct region of multivariate space at each of these levels. The evolutionary transitions to gliding and semiarboreality involved changing limb morphology in opposing directions. Morphologically, the elongated fore and hindlimbs of flying squirrels are consistent with their use in gliding (aerial) behavior (increased

113 patagial surface area) as well as their use of large diameter supports during climbing (see above). The shortened fore and hindlimbs of chipmunks are consistent with their relatively generalized terrestrial lifestyle, while their use of large diameter supports is also an indicator of terrestriality since it actually differs from that of flying squirrels due to their propensity for using horizontally oriented downed logs rather than vertically oriented trunks. The intermediate morphology of red squirrels is evident in their intermediate locomotor behavior and use of medium diameter supports.

Examining interspecific variation across multiple levels of analysis has been used as an effective tool for uncovering the mechanistic bases for behavioral evolution in other taxa (Reilly and Lauder, 1992; Lauder and Reilly, 1996). It allows for the determination of the sequence of character changes leading to evolutionary novelty (e.g. Morphological change preceding behavioral change on the phylogeny). The small number of taxa included in this study precludes a thorough investigation of the evolutionary trajectories that led to gliding and semiarboreality. Furthermore, the absence of a complete sciurid phylogeny makes it difficult at present to identify and include potentially informative taxa. This may be especially problematic in squirrels, since they are a relatively ancient group (e.g. Douglassciurus dates back to 35 myp) and intermediate forms are unlikely to have persisted to the present day. Nevertheless, future research should attempt to include additional taxa and examine them at additional levels (e.g. performance, motor patterns).

This approach should provide further insight into the evolution of sciurid locomotor novelty.

114

References

Ando, M. and Shiraishi, S. (1991). Arboreal quadrupedalism and gliding adaptations in

the Japanese Giant Flying Squirrel, Petaurista leucogenys. Honyurui Kagaku

(Mammalian Science). 33, 167-181.

Ando, M. and Shiraishi, S. (1993). Gliding flight in the Japanese Giant Flying Squirrel

Petaurista leucogenys. J. Mamm. Soc. Japan 18, 19-32.

Bryant, M. D. (1945). Phylogeny of the Nearctic Sciuridae. Am. Midl. Nat. 33, 257-

390.

Cartmill, M. (1974). Pads and claws in arboreal locomotion. In: Primate Locomotion.

pp 249-272. Jenkins, F. A. (ed). New York: Academic Press.

Cartmill, M. (1985). Climbing. In: Functional vertebrate morphology. 73-88.

Hildebrand, M., Bramble, D. M., Liem, K. F., and Wake, D. B. (Eds). Cambridge:

Belknap Press.

Chickering, J. G. (1996). Scaling of skeletal adaptation, locomotor performance and

muscular morphology in sciurids. Ph.D. Dissertation.

Dagosto, M. and Gebo, D. L. (1998). Methodological issues in studying positional

behavior: Meeting Ripley’s challenge. In: Primate Locomotion: Recent

Advances. pp. 5-29. Strasser, E., Fleagle, J. G., Rosenberger, A. L., and McHenry,

H. M. (Eds). New York: Plenum Press.

Elliott, L. (1978). Social behavior and foraging ecology of the eastern chipmunk

(Tamias striatus) in the Adirondack Mountains. Smithson. Contrib. Zool. 265, 1-

107.

115

Emerson, S. B. (1985). Jumping and leaping. In: Functional Vertebrate Morphology. pp

58-72. Hildebrand, D. M. Bramble, K. F. Liem, and Wake, D. B. (Eds),

Cambridge: Belknap Press.

Emerson, S. B. and Koehl, M. A. R. (1990). The interaction of behavioral and

morphological change in the evolution of a novel locomotor type: “Flying” frogs.

Evolution. 44, 1932-1946.

Emerson, S. B., Travis, J., and Koehl, M. A. R. (1990). Functional complexes and

additivity in performance: A test case with “flying” frogs. Evolution. 44, 2153-

2157.

Emry, R. J. and Korth, W. W. (1996). The Chadronian squirrel “Sciurus” jeffersoni

Douglass, 1901: A new generic name, new material, and its bearing on the early

evolution of Sciuridae (Rodentia). J. Vert. Paleo. 16, 775-780.

Emry, R. J. and Korth, W. W. (2001). Douglassciurus, new name for Douglassia

Emry and Korth, 1996, not Douglassia Bartsch, 1934. J. Vert. Paleo. 21, 400.

Emry, R. J. and Thorington, Jr., R. W. (1982). Descriptive and comparative osteology

of the oldest fossil squirrel, Protosciurus (Rodentia: Sciuridae). Smith. Contrib.

Paleo. 47, 1-35.

Emry, R. J. and Thorington, Jr., R. W. (1984). The tree squirrel Sciurus (Sciuridae,

Rodentia) as a living fossil. In: Living fossils: 23-31. Eldridge, N. and Stanley, S.

M. (Eds). New York: Springer Verlag.

Essner, Jr., R. L. (2002). Three-dimensional launch kinematics in leaping, parachuting,

and gliding squirrels. J. Exp. Biol. 205, 2469-2477.

116

Fleagle, J. G. (1976). Locomotor behavior and skeletal anatomy of sympatric Malaysian

leaf-monkeys (Presbytis obscura and Presbytis melalophos). Year. Phys.

Anthropol. 20, 440-453.

Fleagle, J. G. and Mittermeier, R. A. (1980). Locomotor behavior, body size, and

comparative ecology of seven Surinam monkeys. Am. J. Phys. Anthrop. 52, 301-

314.

Gebo, D. L., and Chapman, C. A. (1995). Habitat, annual and seasonal effects on

positional behavior in red colobus monkeys. Am. J. Phys. Anthrop. 96, 73-82.

Goldingay, R. L. (2000). Gliding mammals of the world: Diversity and ecological

requirements. In: Biology of Gliding Mammals. 9-44. Goldingay, R. L. and

Scheibe, J. S. (Eds). Fürth: Filander Verlag.

Greenacre, M. J. (1984). Theory and applications of correspondence analysis. London:

Academic Press.

Hight, M. E., Goodman, M., and Prychodko, W. (1974). Immunological studies of the

sciuridae. Syst. Zool. 23, 12-25.

Jackson, S. M. (1999). Glide angle in the genus Petaurus and a review of gliding in

mammals. Mammal Rev. 30, 9-30.

Kawamichi, T. (1997). Seasonal changes in the diet of Japanese giant flying squirrels in

relation to reproduction. J. Mamm. 78, 204-212.

Lauder, G. V. (1991). Biomechanics and evolution: Integrating physical and historical

biology in the study of complex systems. In Biomechanics in evolution: 1-19.

Rayner, J. M. V. and Wootton, R. J. (Eds). Cambridge: Cambridge University

Press.

117

Lauder, G. V. and Reilly, S. M. (1996). The mechanistic bases of behavioral evolution:

A multivariate analysis of musculoskeletal function. In: Phylogenies and the

Comparative Method in Animal Behavior: pp. 105-137. Martins, E. P. (ed). New

York: Oxford University Press.

Losos, J. B. (1990). The evolution of form and function: Morphology and locomotor

performance in West Indian Anolis lizards. Evolution. 44, 1189-1203.

Moermond, T. C. (1979). Habitat constraints on the behavior, morphology, and

community structure of Anolis lizards. Ecology. 60, 152-164.

Mosimann, J. E. and James, F. C. (1979). New statistical methods for allometry with

application to Florida red-winged blackbirds. Evolution. 33, 444-459.

Norberg, U. M. (1990). Vertebrate flight. Berlin: Springer-Verlag, 291 pp.

Norberg, U. M. (1994). Wing design, flight performance, and habitat use in bats. In:

Ecological morphology: Integrative organismal biology. 205-239. Wainwright, P.

C. and Reilly, S. M. (Eds). Chicago: University of Chicago Press.

Oliver, J. A. (1951). “Gliding” in amphibians and reptiles, with a remark on an arboreal

adaptation in the lizard, Anolis carolinensis carolinensis Voigt. Am. Nat. 85, 171-

176.

Oshida, T., Masuda, R., and Yoshida, M. (1996). Phylogenetic relationships among

Japanese species of the family Sciuridae (Mammalia, Rodentia), inferred from

nucleotide sequences of mitochondrial 12s ribosomal RNA genes. Zool. Soc. Jap.

13, 615-620.

Pounds, J. A. (1988). Ecomorphology, locomotion, and microhabitat structure: Patterns

in a tropical mainland Anolis community. Ecol. Mono. 58, 299-320.

118

Pounds, J. A. (1991). Habitat structure and morphological patterns in arboreal

vertebrates. In: Habitat Structure and Morphological Patterns in Arboreal

Vertebrates. pp. 109-119. Bell, S. S., McCoy, E. D., and Mushinsky, H. R. (Eds).

London: Chapman and Hall.

Preuschoft, H. (1990). Gravity in primates and its relation to body shape and

locomotion. Human Evol. 5, 559-578.

Preuschoft, H., Witte, H., Christian, A., and Fischer, M. (1996). Size influences on

primate locomotion and body shape, with special emphasis on the locomotion of

“small mammals.” Folia Primatol. 66, 93-112.

Rayner, J. M. V. (1981). Flight adaptations in vertebrates. Symp. Zool. Soc. Lond. 48,

137-172.

Reilly, S. M. and Lauder, G. V. (1992). Morphology, behavior, and evolution:

Comparative kinematics of aquatic feeding in salamanders. Brain Behav. Evol.

40, 182-196.

Reilly, S. M. and Wainwright, P. C. (1994). Conclusion: Ecological morphology and

the power of integration. In: Ecological morphology: Integrative organismal

biology. 339-354. Wainwright, P. C. and Reilly, S. M. (Eds). Chicago: University

of Chicago Press.

Rosenberger, A. L. and Stafford, B. J. (1994). Locomotion in captive Leontopithecus

and Callimico: A multimedia study. Am. J. Phys. Anthrop. 94, 379-394.

Roth, V. L. (1996). Cranial integration in the Sciuridae. Am. Zool. 36, 14-23.

119

Runestad, J. A. and Ruff, C. B. (1995). Structural adaptations for gliding in mammals

with implications for locomotor behavior in Paromomyids. Am. J. Phys. Anthrop.

98, 101-119.

Savile, D. (1962). Gliding and flight in the vertebrates. Am. Zool. 2, 161-166.

Scheibe, J. S., Figgs, D., and Heiland, J. (1990). Morphological attributes of gliding

rodents: A preliminary analysis. Trans. Missouri Acad. Sci. 24, 49-55.

Scheibe, J. S., and Robins, J. H. (1998). Morphological and behavioral attributes of

gliding mammals. In Ecology and Evolutionary Biology of Tree Squirrel. pp. 131-

144. Steele, M. A., Merritt, J. F. and Zegers, D. A. (Eds). Martinsville, VA:

Virginia Museum of Natural History.

Scholey, K. D. (1986). The climbing and gliding locomotion of the Giant Red Flying

Squirrel, Petaurista petaurista (Sciuridae). BIONA-report 5, 187-204.

Sokal, R. R. and Rohlf, F. J. (1995). Biometry. San Francisco: Freeman Press.

Stafford, B. J., Thorington, R. W. Jr., and Kawamichi, T. (2003). Positional behavior

of Japanese Giant Flying Squirrels (Petaurista leucogenys). J. Mamm. 84, 263-

271.

Thorington, R. W., Jr. and Heaney, L. R. (1981). Body proportions and gliding

adaptations of flying squirrels (Petauristinae). J. Mamm. 62, 101-114.

Thorington, Jr., R. W. and Thorington, E. M. (1989). Postcranial proportions of

Microsciurus and Sciurillus, the American pygmy tree squirrels. Adv. Neotrop.

Mamm. 125-136.

Thorington, Jr., R. W., Darrow, K., and Betts, A. D. K. (1997). Comparative myology

of the forelimb of squirrels (Sciuridae). J. Morph. 234, 155-182.

120

Thorington, Jr., R. W., Miller, A. M. L., and Anderson, C. G. (1998). Arboreality in

tree squirrels (Sciuridae). In: Ecology and evolutionary biology of tree squirrels.

119-130. Steele, M. A., Merritt, J. F., and Zegers, D. A. (Eds). Martinsville, VA:

Virginia Museum of Natural History.

Vanhooydonck, B., Van Damme, R., and Aerts, P. (2002). Variation in speed, gait

characteristics and microhabitat use in lacertid lizards. J. Exp. Biol. 205, 1037-

1046.

Vernes, K. (2001). Gliding performance of the northern flying squirrel (Glaucomys

sabrinus) in mature mixed forest of eastern Canada. J. Mamm. 82: 1026-1033.

Wainwright, P. C. (1987). Biomechanical limits to ecological performance: mollusk-

crushing by the Caribbean hogfish, Lachnolaimus maximus (Labridae). J. Zool.

Lond. 213, 283-297.

Wainwright, P. C. (1994). Functional morphology as a tool in ecological research. In:

Ecological morphology: Integrative organismal biology. 42-59. Wainwright, P.

C. and Reilly, S. M. (Eds). Chicago: University of Chicago Press.

Ward, S. C., and Sussman, R. W. (1979). Correlates between locomotor anatomy and

behavior in two sympatric species of Lemur. Am. J. Phys. Anthrop. 50, 575-590.

Warren, R. D. and Crompton, R. H. (1997). Locomotor ecology of Lepilemur

edwardsi and Avahi occidentalis. Am. J. Phys. Anthrop. 104, 471-486.

Wells-Gosling, N. (1985). Flying squirrels: Night gliders in the dark. Washington D.C.:

Smithsonian Institution Press.

121

Youlatos, D. (1999). Locomotor and postural behavior of Sciurus igniventris and

Microsciurus flaviventer (Rodentia, Sciuridae) in eastern Ecuador. Mammalia. 63,

405-416.

122

Table 3.1. Limb proportions in extant and fossil squirrels (Ra=radius, Hu=humerus, Ti=tibia, Fe=femur; Modified from Thorington and Heaney, 1981 and Emry and Thorington, 1982).

Species N Ra/Hu Ti/Fe Hu/Fe Ra/Ti Ra + Hu/ Ti + Fe Sciurus niger 19 0.94 ± 0.02 1.12 ± 0.02 0.79 ± 0.02 0.66 ± 0.02 0.72 ± 0.01

Tamiasciurus 34 0.92 ± 0.02 1.13 ± 0.02 0.81 ± 0.01 0.66 ± 0.01 0.73 ± 0.01 hudsonicus Glaucomys volans 26 1.13 ± 0.02 1.15 ± 0.02 0.85 ± 0.01 0.84 ± 0.01 0.85 ± 0.01 Douglassciurus jeffersoni 1 0.96 1.07 0.75 0.67 0.71 (USNM 243981)

123

Table 3.2. Variation in body size and six limb morphological variables in chipmunks, red squirrels, and flying squirrels. Sample sizes are in parentheses.

Chipmunk Red Squirrel Flying Squirrel Variable (n=13) (n=15) (n=20) Mass (g) 94.3 ± 6.5 166.8 ± 10.0 83.5 ± 5.3 Forefoot (mm) 19.0 ± 0.4 27.4 ± 0.6 18.5 ± 0.4 Forearm (mm) 24.4 ± 0.5 36.5 ± 0.3 34.7 ± 0.6 Upper arm (mm) 25.3 ± 0.6 36.1 ± 0.5 30.4 ± 0.6 Hindfoot (mm) 31.8 ± 0.4 43.2 ± 0.5 28.3 ± 0.4 Shank (mm) 33.6 ± 0.7 47.6 ± 0.8 40.1 ± 0.6 Thigh (mm) 29.3 ± 0.9 41.5 ± 0.6 33.0 ± 0.5

Values are mean ± S.E.M. (N=48)

124

Table 3.3. Factor loadings for PCA of six limb morphological variables in chipmunks, red squirrels and flying squirrels.

Variables Factor 1 Factor 2 Factor 3 Factor 4 Factor 5 Forefoot length 0.897191 0.341981 -0.010320 0.275702 0.044494 Forearm length -0.945175 0.159472 -0.216890 -0.079562 0.166856 Upper arm length -0.680170 -0.394284 0.530712 0.232416 0.215027 Hindfoot length 0.943182 -0.133537 -0.016495 -0.286489 0.101132 Shank length -0.647568 0.354062 0.368502 -0.287666 -0.486571 Thigh length -0.044640 -0.631710 -0.589832 0.294561 -0.405318 Variance Explained 71.8% 9.79% 7.56% 6.80% 4.02%

125

Table 3.4. Factor loadings for correspondence analysis of six locomotor behavior variables in chipmunks, red squirrels, and flying squirrels.

Variables Factor 1 Factor 2 Factor 3 Factor 4 Factor 5 Climb up 0.096 0.021 0.090 0.218 0.358 Climb down 0.000 0.244 0.000 0.247 0.347 Run branch 0.003 0.022 0.608 0.075 0.054 Run ground 0.678 0.134 0.013 0.000 0.002 Run incline 0.003 0.181 0.287 0.412 0.002 Aerial 0.220 0.398 0.001 0.047 0.237 Variance explained 65.67% 15.54% 10.11% 6.89% 1.79%

126

Table 3.5. Factor loadings for correspondence analysis of nine habitat use variables in chipmunks, red squirrels, and flying squirrels.

Variables Factor 1 Factor 2 Factor 3 Factor 4 Factor 5 Factor 6 Factor 7 Factor 8 Ground 0.511 0.008 0.023 0.074 0.001 0.208 0.113 0.000 Low 0.007 0.001 0.269 0.041 0.044 0.080 0.374 0.000 High 0.211 0.015 0.388 0.006 0.110 0.002 0.179 0.000 Small 0.014 0.094 0.048 0.005 0.472 0.023 0.058 0.172 Medium 0.009 0.301 0.114 0.013 0.320 0.032 0.036 0.105 Large 0.001 0.419 0.001 0.019 0.046 0.069 0.075 0.221 Horizontal 0.132 0.026 0.082 0.098 0.001 0.207 0.074 0.229 Vertical 0.098 0.085 0.039 0.021 0.000 0.337 0.070 0.210 Oblique 0.017 0.050 0.036 0.723 0.005 0.042 0.021 0.064 Variance 59.81% 18.28% 11.69% 5.42% 3.62% 1.18% 0.00% 0.00% explained

127

Table 3.6. Biomechanical predictions from Thorington and Thorington (1989), Youlatos (1999), and Stafford et al. (2003).

Biomechanical Prediction Support Small squirrels leap more frequently Yes Small squirrels use smaller supports more frequently No Squirrels with elongated forearms climb more frequently Yes Gliding forms use vertical supports more frequently Yes

128

Figure 3.1. Phylogenetic relationships of the sciurid taxa included in this study (Hight et al., 1974; Oshida et al., 1996; Roth, 1996) and the ancestral tree squirrel represented by fossil taxa such as Douglassciurus. Modern tree squirrels (e.g. red squirrel) are thought to retain ancestral features of morphology, behavior, and ecology (Emry and Thorington, 1984).

129

Figure 3.2. Enclosure constructed to study squirrel behavior and habitat use. Trees and branches were erected inside the enclosure and the substrate was partitioned according to height, diameter, and orientation.

130

Figure 3.3. Size removed principal components analysis (PCA) of six limb morphological variables in chipmunks (▲), red squirrels (■) and flying squirrels (●). Factor one explained the majority of the variance (72%) and discriminated among the species based upon a contrast between proximal and distal limb elements. Flying squirrels possess elongated upper arms, forearms, and shanks with shortened fore and hindfeet. Chipmunks have made opposing changes. Red squirrels are intermediately located. MANOVA indicates all three species differ significantly in morphology (Wilks’ λ=0.056, P<0.00001). Significant factors indicated by *.

3.0

FS RS 1.0 (10%)

Factor 2 -1.0 CH

-3.0 -2.0 -0.8 0.3 1.5 Long upper/forearm, shank Factor 1 (72%)* Short upper/forearm, shank Short fore/hindfeet Long fore/hindfeet

131

Figure 3.4. Proportions for the seven locomotor behaviors examined in chipmunks, red squirrels, and flying squirrels. Chipmunks used ground running to a greater degree than flying squirrels and red squirrels. Flying squirrels had a greater propensity to climb up and leap. Flying squirrels were unique in their use of parachuting/gliding locomotion (although red squirrels use it as an escape behavior).

Chipmunk 29.9 11.5 20.1 17.1 17.0 3.9

Red squirrel 15.2 11.9 27.9 17.1 19.1 8.7

Flying squirrel 6.8 11.2 22.5 14.6 28.4 9.3 7.2

0% 20% 40% 60% 80% 100%

run ground run incline run branch climb down climb up leap parachute/glide

132

Figure 3.5. Correspondence analysis (CA) of six locomotor behaviors in chipmunks (▲), red squirrels (■) and flying squirrels (●). Empty symbols indicate species means. Factor 1 explains the majority of the variance (66%) and reveals a strong association between chipmunks and run ground. In contrast, flying squirrels are associated with aerial locomotion. Red squirrels are intermediately located. MANOVA indicates all three species differ significantly in locomotor behavior (Wilks’ λ=0.155, P<0.0001). Significant factors indicated by *.

0.4

0.3

0.2 RS CH 0.1 FS 0.0

-0.1 Factor 2 (16%) Factor 2 (16%) -0.2

-0.3

-0.4 -0.8 -0.4 0.0 0.1 0.2 0.3 0.4 0.5 -1.0 -0.8 -0.6 -0.4-0.8 -0.2 0.0 0.2 0.4 0.6 Run ground Factor 1 (66%)* Aerial

133

Figure 3.6. Proportions for the nine habitat use variables examined in chipmunks, red squirrels, and flying squirrels. Chipmunks utilized ground based locomotion more frequently than the other two species. In contrast, flying squirrels and red squirrels moved high in the trees more frequently (A). Flying squirrels and chipmunks made greater use of large diameter supports than red squirrels. Red squirrels made greater use of medium-sized substrates (B). Flying squirrels used vertically oriented supports most frequently. Chipmunks and red squirrels used horizontal substrates more often than flying squirrels (C).

134

Figure 3.7. Correspondence analysis (CA) of nine habitat use variables examined in chipmunks (▲), red squirrels (■) and flying squirrels (●). Empty symbols indicate species means. Factor 1 explained the majority of the variance (60%) and discriminated between chipmunks (ground) and red squirrels and flying squirrels (high) based upon support height. Factor two (18%) discriminated between red squirrels (medium) and flying squirrels and chipmunks (large) based upon support diameter. MANOVA indicates all three species differ significantly in habitat use (Wilks’ λ=0.241, P<0.0005). Significant factors indicated by *.

0.3

RS 0.2

0.1 CH

0.0

-0.1

Factor 2 (18%)* FS -0.2

-0.3

-0.4 Large Medium Medium Large -0.4 -0.2 0.0 0.2 0.4 0.6 0.8

High Factor 1 (60%)* Ground

135

Figure 3.8. Multilevel comparisons of sciurid morphospace, ethospace, and ecospace. Lines connect taxa across multiple levels. The three species exhibited significant variation at each level. Morphologically, flying squirrels and chipmunks have changed limb morphology in opposing directions while red squirrels have retained the intermediate ancestral morphology. Morphological differences are consistent with differences at the behavioral and ecological levels.