J Mammal Evol DOI 10.1007/s10914-015-9299-4 ORIGINAL PAPER Functional Morphology of the Forelimb of the Nine-Banded Armadillo (Dasypus novemcinctus): Comparative Perspectives on the Myology of Dasypodidae R. A. Olson 1,2 & M. D. Womble 1 & D. R. Thomas 1 & Z. D. Glenn 1 & M. T. Butcher 1 # Springer Science+Business Media New York 2015 Abstract The nine-banded armadillo, Dasypus novemcinctus, is a member of the family Dasypodidae, which contains all extant species of armadillos and represents the most diverse group of xenarthran mammals by their speciation, form, and range of scratch-digging ability. This study aims to identify muscle traits that reflect specialization for fossorial habit by observing forelimb structure in D. novemcinctus and comparing it among armadillos using available myological data. A number of informative traits were observed in D. novemcinctus and among Dasypodidae, including the absence of m. rhomboideus profundus, the variable presence of a m. articularis humeri and m. coracobrachialis, two heads of m. triceps brachii with scapular origin, and a lack of muscle mass devoted to antebrachial supination. Muscle mass and myosin heavy chain (MHC) isoform content were also quantified from our forelimb dissections. New osteological indices are additionally calculated and reported for D. novemcinctus. Collectively, the findings emphasize muscle mass and power output for limb retraction and specialization of the distal limb for sustained purchase of soil by strong pronation and carpal/digital flexion. Moreover, the myological traits assessed here provide a valuable resource for Electronic supplementary material The online version of this article (doi:10.1007/s10914-015-9299-4) contains supplementary material, which is available to authorized users. * M. T. Butcher mtbutcher@ysu.edu 1 Department of Biological Sciences, Youngstown State University, 4013 Ward Beecher Science Hall, Youngstown, OH, USA 2 Department of Biological Sciences, Ohio University, Athens, OH, USA interpretation of muscle architecture specializations among digging mammals and future reassessment of armadillo phylogeny. Keywords Dasypus novemcinctus . Forelimb . Myology . Myosin . Osteology . Scratch-digging . Semi-fossorial Introduction The nine-banded armadillo, Dasypus novemcinctus, is a member of the order Cingulata and superorder Xenarthra, which contains armadillos (Cingulata), sloths (Phyllophaga), and anteaters (Vermilingua). Cingulata represents a major placental mammalian clade from South America, and although their phyletic nature is not fully resolved (Gaudin and Wible 2006), armadillos are currently recognized as a monophyletic group (Wilson and Reeder 2005) of armored New World mammals that comprise the sole family Dasypodidae (Delsuc and Douzery 2008; Loughry and McDonough 2013). Cingulata is estimated to have emerged 41 mya, while the genus Dasypus arose approximately 11 mya during the middle Miocene (Delsuc et al. 2012), and then further diversified with migration to both Central and North America from South America (McBee and Baker 1982; Loughry and McDonough 2013). Dasypus novemcinctus is found throughout Central and South America, and is the only armadillo to inhabit the southern regions of North America. In addition, it has the highest population density of any extant armadillo species (Aguiar and Fonseca 2008). The ability of D. novemcinctus to occupy niches in such a broad range of microclimates is related to both their flexible diet and thermoregulatory strategy. Adults primarily feed on insects and other invertebrates, have relatively low basal J Mammal Evol metabolic rates (for placentals), and commonly walk between feeding sites to conserve energy (Loughry and McDonough 2013). Moreover, foraging activity peaks near sunset, while the rest of their daily time budget is spent dormant inside one of their burrows (McDonough and Loughry 1997). Ninebanded armadillos will excavate burrows or scavenge an existing burrow, but do not appear to dig as an essential part of their foraging strategy (Nowak 1999; Vizcaíno and Milne 2002). Because of its observed habits, D. novemcinctus is behaviorally classified as semi-fossorial (Talmage and Buchanan 1954; Hildebrand 1985; Vizcaíno et al. 1999; Kley and Kearney 2007). Moreover, it has been proposed that fossorial ability and morphological adaptations for digging were strongly selected for in the basal stock of xenarthrans (Nyakatura and Fischer 2011), and this likely constrained ecomorphology for scratch-digging in all extant armadillos despite their range of fossorial habits. Typical of mammals that employ scratch-digging, armadillo forelimbs display a number of modifications to the limb skeletal elements, including large areas for muscle attachment, a prominent olecranon process that increases the mechanical advantage of the elbow extensor muscles, and large claws for piercing soil (Hildebrand 1985; Vizcaíno et al. 1999). Other notable digging modifications found in armadillos include a greatly enlarged teres process of the scapula, an enlarged deltoid tuberosity on the humerus, and a short, highly fused radius and ulna (Miles 1941; Vizcaíno and Milne 2002). Despite these previous observations of skeletal adaptation for digging in armadillos, modifications to the muscular anatomy of their forelimbs has largely been ignored. Descriptions of limb muscle form do exist for six-banded, Euphractus sexcinctus (Cuvier and Laurillard 1849; Galton 1869; Windle and Parsons 1899), large hairy, Chaetophractus villosus (Windle and Parsons 1899), and pink fairy armadillos, Chlamyphorus truncatus (Macalister 1875; Windle and Parsons 1899), but are limited to the shoulder and brachial musculature of D. novemcinctus (Miles 1941), although the myology of a species synonymous with D. novemcinctus (Tatusia peba of Macalister 1875) is available. No quantitative data on muscle architecture or myosin heavy chain (MHC) isoform expression have been reported for armadillos. In this study, we describe the myology of the forelimb of D. novemcinctus, and evaluate the extent to which osteology, muscle mass, and MHC isoform content reflect the evolution of fossorial ability in D. novemcinctus. In addition, available myology and morphological characters from other species of dasypodids will be evaluated to help remedy gaps in our knowledge of comparative muscle structure and digging function in armadillos. This work will also build on our previous studies of muscle architectural properties (Moore et al. 2013; Rose et al. 2013; Rupert et al. 2015) to improve understanding of muscle traits that indicate specialization for fossorial habit as observed across diverse orders of mammals. Materials and Methods A total of eight forelimbs from four nine-banded armadillos were dissected for this study. Armadillos were originally collected and used in a study of in vivo femoral strains (Copploe et al. 2015) after which, the animals were euthanized, frozen, and stored at −20 °C until observation. Specimens were allowed to thaw for 36–48 h at 4 °C prior to dissection and measurement. Both limbs were used for myological observations, and the best preserved limb was used for making muscle architectural measurements and harvesting of muscle tissue blocks for MHC isoform analysis. This work was conducted at Youngstown State University (YSU) in 2014. All procedures complied with protocols approved by the Institutional Animal Care and Use Committees (IACUC) and adhered to the legal requirements of the United States. Muscle origin and insertion for D. novemcinctus dissection followed those of Macalister (1875), Windle and Parsons (1899), and Miles (1941). Muscle nomenclature closely followed the Nomina Anatomica Veterinaria (International Committee on Veterinary Gross Anatomical Nomenclature 2012) with a few exceptions. The term m. triceps brachii angular head (caput angulare: Ercoli et al. 2015) was adopted to designate a second head of the m. triceps brachii that originates on the scapula, and subdivisions of the m. flexor digitorum profundus humeral head (caput humerale) are named after those of Fisher et al. (2009). Synonyms of several muscles (e.g., digital extensors) are provided in the Results section to facilitate comparisons with previous works on armadillo myology. Muscle names, abbreviations, actions, and fiber architecture are presented in Table 1. Briefly, forelimbs were skinned and muscles (excluding those of the manus) were identified, systematically dissected, and their attachments and fiber architecture recorded. All dissections were documented by photographs taken with an α-nex 5 digital camera (Sony, Japan) and these images were used to create limb muscle maps and illustrations of the limb muscle topography. Muscles were periodically moistened with phosphate buffered saline (PBS) to prevent desiccation during dissection. Following the removal of each muscle and all associated free tendons, wet muscle mass (to the nearest 0.001 g) was recorded using an analytical balance (Model: AB104-S/Fact; Mettler-Toledo, USA). Means ± SD (standard deviation) of muscle masses are also presented in Table 1. In addition, muscle architectural measurements (e.g., muscle belly length, fascicle length, and pennation angle) were taken using methods previously described in detail (Moore et al. 2013; Rose et al. 2013). These data, along with calculations of physiological crosssectional area (PCSA), maximum isometric force (Fmax), joint torque, and muscle power are reported in Table 5. Ten bone length and width measurements (to the nearest 0.01 mm) were taken using digital calipers (Model: CD-8 CSX; Mitutoyo, Japan) as previously described and illustrated J Mammal Evol Table 1 Forelimb muscles per limb region, their masses, actions, and fiber architecture for D. novemcinctus Muscle Abbreviation N Muscle mass (g) Action Fiber architecture Parallel Extrinsic Trapezius pars cervicalis TC 3 6.83±2.0 Scapula elevation, stabilization, protraction Trapezius pars thoracica TT 3 7.58±1.2 Scapula elevation, stabilization, retraction Parallel Rhomboideus cervicis et capitis RC 4 5.30±0.7 Scapula elevation, stabilization, protraction Parallel Parallel Rhomboideus thoracis RT 4 2.68±0.5 Scapula elevation, stabilization, retraction Serratus ventralis SV 3 9.82±1.4 Scapula stabilization, limb protraction/retraction Multipennate Latissimus dorsi LAT 4 11.9±5.1a Limb retraction Parallel Pectoralis superficialis PS 3 14.9±1.4 Limb retraction, adduction Unipennate Subclavius SC 4 2.59±0.3 Scapula stabilization, rotation Unipennate Deltoideus pars acromialis DA 4 1.64±0.1 Shoulder flexion, limb retraction Parallel Deltoideus pars scapularis DS 4 5.72±0.6 Shoulder flexion, limb retraction Parallel Deltoideus pars clavicularis DC 4 0.43±0.1 Shoulder flexion, limb retraction Parallel Supraspinatus SSP 4 5.57±0.8 Shoulder extension, limb stabilization Bipennate Shoulder Infraspinatus ISP 4 3.24±0.4 Shoulder flexion, limb retraction Unipennate Teres major TMJ 4 2.67±0.3 Shoulder flexion, limb retraction Unipennate Subscapularis SUB 4 4.05±0.1 Shoulder extension; limb adduction Multipennate Unipennate Brachium Coracobrachialis CCB 4 0.47±0.03 Limb protraction, adduction Biceps brachii BB 4 1.65±0.2 Elbow flexion; shoulder extension Unipennate Brachialis BCH 4 1.57±0.1 Elbow flexion Parallel Elbow extension; shoulder flexion, humeral retraction Unipennate Triceps brachii – angular TBA 4 11.64±1.1 Triceps brachii – long TBLO 4 3.87±0.2 Elbow extension; shoulder flexion, humeral retraction Unipennate Triceps brachii – medial TBM 2 3.44±0.4 Elbow extension Unipennate Triceps brachii – lateral TBL 4 3.31±0.4 Elbow extension Parallel Tensor fasciae antebrachii TFA 4 7.53±1.6 Antebrachial fascia tension; elbow extension Parallel Anconeus ANC 4 0.40±0.1 Elbow extension Parallel Pronator teres PT 4 0.63±0.1 Antebrachial pronation Unipennate Flexor carpi ulnaris FCU 4 0.56±0.2 Carpal flexion Unipennate Flexor carpi radialis FCR 4 0.69±0.1 Carpal flexion Bipennate Flexor digitorum superficialis FDS 4 1.32±0.1 Flexion digits II–IV Bipennate Flexor digitorum profundus – humeral medial FDPHM 4 6.06±0.2b Flexion digits III–V Flexor digitorum profundus – ulnar FDPU 4 Multipennate Flexor digitorum profundus – radial FDPR 4 0.86±0.1 Unipennate Antebrachium Flexor digitorum profundus – humeral profundus FDPHP 4 0.36±0.1 Extensor carpi radialis 4 2.12±0.03 ECR Multipennate Parallel Carpal extension Bipennate Extensor carpi ulnaris ECU 4 0.48±0.1 Carpal extension Unipennate Extensor digiti IIc ED2 4 0.84±0.3 Extension digit III Unipennate Extensor digitorum communis EDC 4 0.55±0.3 Extension digits IV, V Unipennate Abductor digiti I longusd AD1L 4 0.67±0.1 Abduction digit II Bipennate Abductor digiti V longuse AD5L 4 0.55±0.3 Abduction digit V Unipennate N = Number of limbs Muscle masses are mean ± SD (standard deviation) a In one specimen, the m. latissimus dorsi was incompletely separated from the m. tensor fasciae antebrachii, resulting in a lower mass for m. latissimus dorsi and a higher mass for m. tensor fasciae antebrachii. This explains the high SD for this muscle b Mass of both muscle heads combined; muscle bellies fused near origin c Digit I is absent and the attachments of digits II and III are modified. This muscle only extended digit III d Previously named m. extensor pollicus longus or m. extensor ossis metacarpi pollicis. This muscle abducted digit II e Previously named m. extensor digiti minimi. This muscle is a separate belly and abducted digit V, thus it is renamed m. adbuctor digiti V longus J Mammal Evol (Rose et al. 2014). A summary of the measurements, abbreviations, and recorded values are reported in Table 6. From the raw osteological measurements, six functional indices that indicate either area available for muscle attachment or mechanical advantage of the forelimb were calculated: Scapular index (SI=SW/SL); Epicondyle index (EI=HEW/HL); Humeral robusticity index (HRI=HW/HL); Triceps metacarpal out-force index (TMOI=OL/(UL+MCL III – OL)); Ulnar robusticity index (URI=UW/UL); and Manus proportions index (MANUS=PPL III/MCL III). The specific mechanical significance of each osteological index is indicated by Rose et al. (2014) and Samuels et al. (2013), and the references therein. Small blocks of muscle tissue were harvested from selected forelimb muscles from each specimen (N=4). Muscle tissue was prepared for SDS-PAGE by flash freezing in liquid nitrogen, grinding to powder, homogenizing 50 mg of muscle powder in 800 ml of Laemmli buffer (Laemmli 1970; Toniolo et al. 2008), and centrifugation of the homogenates at 13 k rpm for 10 min (Rupert et al. 2014). Samples for gel loading were diluted in gel sample buffer (Mizunoya et al. 2008) to a final protein concentration of ~0.125 μg/ml. MHC isoforms were identified on SDS-PAGE gels using established methods (Talmadge and Roy 1993) performed with slight modifications (Mizunoya et al. 2008) as previously described (Hazimihalis et al. 2013; Rupert et al. 2014). Gels were loaded with a total of ~1 μg of protein per lane, stained with silver (Bio-Rad, Hercules, CA), and imaged using Pharos FX Plus (Quantity One software: Bio-Rad). MHC isoform content was quantified by densitometry in ImageJ (v.1.43: NIH) using a brightness area product method (Rupert et al. 2014). Band intensity values in each gel lane were summed across two to four independent gel runs per muscle (and individual) and used to calculate a percentage for each MHC isoform expressed in each muscle analyzed. Percentages of the MHC isoforms for each muscle were then averaged across individuals to provide a final mean percentage content of both the slow and fast MHC isoforms. trapezius pars cervicalis originates from the dorsal aponeurosis, which is anchored along the midline to the cervical and upper thoracic regions of the vertebral column. It inserts onto the distal one third of the clavicle, the acromion, and cranial spine of the scapula (Fig. 1). The m. trapezius pars thoracica is almost rectangular in shape and noticeably lighter in color than the pars cervicalis. It arises from the spines of all thoracic vertebrae via the dorsal aponeurosis and inserts upon the caudal one third of the scapular spine (Fig. 1). The origin from the upper thoracic vertebrae is fleshy and is located deep to the origin of the m. trapezius pars cervicalis. The origin from the middle and lower thoracic vertebrae is increasingly aponeurotic. The m. trapezius pars cervicalis was previously described as the upper portion (Galton 1869) or superficial portion (Miles 1941), while the m. trapezius pars thoracica was described as the lower (Galton 1869) or deep portion (Miles 1941). The presence of a clavicular insertion was noted by Cuvier and Laurillard (1849) in E. sexcinctus and by Macalister (1875) in C. truncatus, but was not reported by either Galton (1869) or Miles (1941). M. Rhomboideus The m. rhomboideus has two parts: m. rhomboideus cervicis et capitis and m. rhomboideus thoracis. M. rhomboideus cervicis et capitis is found in the dorsal region of the neck. It originates via tendinous fibers from the lambdoidal crest of the cranium, just caudal to the mastoid process, and from the spinous processes of the cervical vertebrae. It inserts onto the lateral side of the cranial (coracovertebral) angle of the scapula (Fig. 1). Miles (1941) referred to this muscle as the m. rhomboideus pars capitis and it is also known as the m. occipitoscapularis (Macalister 1875). The m. rhomboideus thoracis originates from the dorsal aponeurosis and middle thoracic vertebrae, and has a fan-shaped appearance. It inserts along the dorsal border of the scapula and onto the deep aspect of the teres process of the scapula (Fig. 2). M. Serratus Ventralis Results Muscle mass, action, and fiber architecture are reported for all muscles in Table 1. Extrinsic muscles are not illustrated. These illustrations are available in Miles (1941: Figs. 3, 4, and 5). Extrinsic Muscles of the Forelimb M. Trapezius The m. trapezius is a thin muscle located deep to the m. platysma, m. panniculus, and a variable layer of fat. It is composed of the pars cervicalis and pars thoracica. The m. The m. serratus ventralis shows a fairly extensive origin from the angle of the first rib and the lateral, mid-shaft area of ribs 2–7. It passes deep to the scapula to insert along the deep side of the entire medial aspect of the dorsal border of the scapula, from the cranial angle to the distal end of the teres process (Fig. 2). Miles (1941) referred to this muscle as the m. serratus anterior with a broad origin from the transverse processes of the last 2–3 cervical vertebrae and extending to rib 6. It was documented to have two muscle bellies, the cervical and posterior, in D. novemcinctus. This muscle is also referred to as the m. serratus magnus, and in E. sexcinctus, it displays an upper portion (from ribs 1–2) and lower portion (from ribs 3– 6) (Galton 1869), while the lower portion extends to rib 7 in J Mammal Evol Fig. 1 Lateral muscle map diagram of the forelimb skeleton indicating the locations of muscle origins and insertions for D. novemcinctus. Muscle nomenclature abbreviations are given in Table 1. Limb outline adapted from Miles (1941) RC SSP TT TC DA ISP ISP SSP DC TBL BCH DS TMJ DA TBA DS RT AD5L ECR AD1L ECU ECR EDC TBM EDC TBA ED2 TFA C. truncatus (Macalister 1875). In our dissections, the muscle could not be clearly separated into two parts and had no discernable attachment to the cervical vertebrae. M. Latissimus Dorsi The m. latissimus dorsi originates from the middle and lower thoracic vertebrae via the dorsal aponeurosis, FDP ED2 TBL ECU AD5L ECU muscular slips from ribs 5–9, and from a small area on the medial and caudal aspects of the teres process of the scapula (Fig. 2). Its medial margin is extensively fused with the m. pectoralis superficialis. The muscle displays external tendinous fibers on its superficial side as it nears its insertion point. It inserts via a flat tendon onto the medial lip of the bicipital groove of the humerus (Fig. 2). At its insertion, the muscle shows a Fig. 2 Medial muscle map diagram of the forelimb skeleton indicating the locations of muscle origins and insertions for D. novemcinctus. Muscle nomenclature abbreviations are given in Table 1. Limb outlines adapted from Miles (1941) SV TC SC BB CCB SUB SUB LAT TMJ PS TBLO RT DC TBM LAT FCR CCB FDPR PT FDPHP FDPHM ANC PT TBLO FDS ANC FCU FDPU AD1L FDS FDP FCR FDP FCU BCH BB J Mammal Evol one-half twist, with the most dorsal fibers inserting more distally along the proximal, medial humerus. M. Pectoralis Superficialis The m. pectoralis superficialis is a large muscle that originates near the midline, along the entire length of the lateral sternum. It forms a tendinous insertion onto the medial humerus, along the pectoral crest extending proximally from the distal tip of the deltoid tuberosity (Fig. 2). This insertion shows a one-half twist, with the cranial-most fibers, which originate from the cranial end of the sternum, inserting more distally. In one individual, the m. pectoralis superficialis could be divided into two muscle bellies. The second belly (or abdominal head) had no bony origin and instead arose from the fascia of the abdominal aponeurosis. Its insertion onto the medial humerus was continuous with the muscle as a whole, extending diagonally proximally and caudally to just below the caudal side of the humeral head. Miles (1941) described this abdominal belly as a separate muscle known as m. pectoralis abdominis. M. Pectoralis Profundus This muscle is absent in the nine-banded armadillo. M. Subclavius The m. subclavius arises from the lateral surface of the first rib. Distally, it forms a wide tendon that passes under the coracoacromial ligament to insert onto the coracoid process, the upper margin of the glenoid fossa, and the medial side of the acromion of the scapula (Fig. 2). This muscle was previously misidentified as the m. pectoralis minor by Miles (1941). Intrinsic Muscles of the Forelimb –Shoulder M. Deltoideus The m. deltoideus is composed of the pars acromialis, pars scapularis, and pars clavicularis. The m. deltoideus pars acromialis originates from the entire length of the acromion process of the scapula and inserts near the middle of the lateral aspect of the deltoid tuberosity of the humerus (Fig. 1). This head has superficial tendinous fibers at its origin and a deep central tendon running the length of the muscle to its insertion. The m. deltoideus pars scapularis originates from the deep surface of the proximal acromion, and via an extensive aponeurosis along the length of the scapular spine and dorsal border of the scapula, from the base of the scapular spine to the teres process (Fig. 1). It becomes tendinous on its deep side and inserts onto the distal and caudal aspects of the deltoid tuberosity, just caudal to the insertion of the m. deltoideus pars acromialis. The m. deltoideus pars clavicularis (or m. cleidobrachialis by some authors; see Discussion) originates from the distal one-third of the clavicle and inserts onto the humerus along the medial and distal aspects of the deltoid tuberosity on the humerus (Figs. 1 and 2). Both the pars clavicularis and pars acromialis are highly fused along the length of their muscle bellies. M. Supraspinatus The m. supraspinatus originates from the supraspinous fossa of the scapula. Its tendon passes deep to the acromion and over the head of the humerus to insert onto the superior-cranial aspect of the greater tubercle of the humerus (Figs. 1 and 3). M. Infraspinatus The m. infraspinatus originates from the infraspinous fossa of the scapula, with additional tendinous fibers arising from the dorsal border of the scapula, inferior to the scapular spine. It has a broad, flat tendon that extends the length of the muscle belly. The muscle passes deep to the acromion and inserts via both its tendon and some fleshy fibers onto the greater tubercle of the humerus, immediately caudal to the insertion of m. supraspinatus. (Figs. 1 and 3). M. Teres Minor The m. teres minor was not observed as a separate muscle in our dissections. However, it was reported by Macalister (1875), and was previously noted by Miles (1941) to be highly fused with m. infraspinatus in D. novemcinctus. M. Teres Major The m. teres minor originates from the fossa on the lateral aspect of the teres process of the scapula (Figs. 1 and 4). It inserts via a tendon to the medial lip of the bicipital groove of the humerus, immediately distal to the insertion of the m. latissimus dorsi (Fig. 2). It has superficial tendinous fibers near its insertion. M. Subscapularis The m. subscapularis has an extensive origin from the subscapularis fossa of the scapula. The muscle belly is partitioned into three distinct portions by internal tendons. It inserts via a broad tendon onto the lesser tubercle of the proximal, medial humerus (Figs. 2 and 4). J Mammal Evol Fig. 3 Lateral view of the muscles of the brachium in D. novemcinctus –Brachium M. Articularis Humeri M. Coracobrachialis This muscle is absent in the nine-banded armadillo. The m. coracobrachialis originates near the tip of the coracoid process of the scapula (Fig. 2). Its tendon of origin and proximal muscle belly are fused with the short head of m. biceps brachii. The m. coracobrachialis then passes superficial to the m. subscapularis and inserts via a fleshy attachment onto the medial aspect of the humeral mid-shaft, opposite the deltoid tuberosity (Figs. 2 and 4). M. Biceps Brachii Fig. 4 Medial view of the muscles of the brachium in D. novemcinctus The m. biceps brachii has long and short heads of origin. The long head (caput longum) originates as a tendon from the superior lip of the glenoid fossa of the scapula (Fig. 2). This tendon passes over the head of the humerus and then runs distally within the bicipital groove of the humerus, where it J Mammal Evol is held in place by a strong ligament spanning between the greater and lesser tubercles. The diminutive short head (caput breve) arises from the coracoid process of the scapula, where it is fused extensively with the origin of the m. coracobrachialis and only a few thin fibers of this muscle head are visible. The two muscle heads join to form a single muscle belly, which then inserts onto the shaft of the proximal ulna, just distal to the trochlear notch (Figs. 2, 3 and 4). The muscle displays tendinous fibers on its deep surface near its insertion. A short head of the m. biceps brachii was not reported by Macalister (1875). M. Brachialis The m. brachialis originates from an area on the lateral aspect of the proximal humerus, caudal to the greater tubercle. It inserts via a short tendon onto the ulna, immediately distal to the insertion of m. biceps brachii (Figs. 1, 2, 3 and 4). M. Tensor Fasciae Antebrachii The m. tensor fasciae antebrachii (or m. dorsoepitrochlearis: Miles 1941) has an extensive fleshy origin from the surface fascia of the m. latissimus dorsi, inferior to the level of the teres process, near the line at which the m. latissimus dorsi becomes less fleshy and more tendinous (not illustrated). Near the shoulder joint, both muscles show a variable amount of fusion as m. tensor fascia antebrachii passes over the muscle belly of m. latissimus dorsi. The m. tensor fascia antebrachii shows fleshy insertions onto the m. triceps brachii long head near the olecranon process, and into the medial antebrachial fascia, where it overlies the posterior area of origin for the ulnar head of m. flexor digitorum profundus. The m. tensor fascia antebrachii also shows a small insertion onto the tip of the olecranon process of the ulna (Fig. 1). M. Triceps Brachii M. Anconeus The m. triceps brachii has four distinct heads of origin: angular, long, medial, and lateral. The angular head (caput angulare) arises from an extended line along the lateral side of the caudal border of the scapula and inserts onto the lateral tip of the olecranon process of the ulna (Figs. 1, 3 and 4). The long head (caput longum) originates from a small area on the medial side of the caudal border of the scapula, near the glenoid fossa. It inserts medially near the tip of the olecranon (Figs. 2 and 4). Miles (1941) described a single head of origin on the scapula (referred to as the scapular head) that is split by a deep fissure into anteromedial and posterolateral portions, which correspond to the long and angular heads, respectively. However, Windle and Parsons (1899) clearly described two scapular heads for the m. triceps brachii and suggested that this condition is present in all species of armadillos, and may be even further modified (i.e., additional scapular heads) in closely related anteaters. The medial and lateral heads of m. triceps brachii arise from the humerus. The medial head (caput mediale) shows a fleshy origin from an extensive area of the caudal and medial humerus and inserts onto much of the cranial border of the olecranon process (Figs. 1, 2 and 4). The distal one-half of the medial and long heads are fused. The lateral head (caput laterale) has a small area of origin on the proximal, lateral humerus, caudal to the origin of m. brachialis. It inserts onto the inferior portion of the olecranon, just distal to the insertion of the long head (Figs. 1 and 3). It shows external tendinous fibers superficially near its insertion. The m. anconeus is a small muscle that originates from the caudal aspect of the medial epicondyle of the humerus and inserts onto the medial border of the olecranon process, at its tip (Figs. 2 and 4). –Antebrachium M. Pronator Teres The m. pronator teres originates as part of the common flexor tendon from the medial epicondyle of the humerus (Fig. 2). It forms a flattened tendon that inserts onto the medial (Fig. 2) and distal portions of the radial ridge of the radius (not illustrated). M. Flexor Carpi Ulnaris The m. flexor carpi ulnaris has an extended line of origin from the inferior-medial surface of the mid- and distal ulna, where the muscle displays tendinous fibers on its deep surface (Fig. 2). In addition, the muscle also has a long, thin tendon that arises from the medial tip of the olecranon process of the ulna. This tendon runs along the medial side of the ulnar head of m. flexor digitorum profundus before joining into the distal belly of m. flexor carpi ulnaris (Fig. 5a and b). At its insertion, the muscle shows a medial fleshy attachment to the pisiform (accessory carpal bone), and laterally a tendinous portion that continues distally to insert onto the palmar side of the shaft of metacarpal V. J Mammal Evol Fig. 5 Illustrations of the antebrachial muscles in D. novemcinctus. a medial view; b lateral view M. Flexor Carpi Radialis Palmaris Longus The m. flexor carpi radialis originates as part of the common flexor tendon off the medial epicondyle of the humerus, where it is fused with the humeral head of the m. flexor digitorum profundus. The m. flexor carpi radialis inserts via a tendon to the palmar surface at the proximal base of metacarpal I (Figs. 2 and 6a). Just prior to this insertion, a ligament spanning between the styloid process of the distal radius and the scapholunate carpal bone loops around the inserting tendon. The m. flexor carpi radialis also has external tendinous fibers on the proximal one half of its medial side. This muscle is absent in the nine-banded armadillo. M. Flexor Digitorum Superficialis The m. flexor digitorum superficialis originates from the medial and inferior aspects of the medial epicondyle of the humerus as part of the common flexor tendon (Fig. 2). The more superficial muscle fibers have a fleshy origin, while the deeper fibers are more tendinous. These two sets of fibers quickly join together to form the muscle belly. Distally, two tendons emerge from the medial and lateral sides of the muscle belly. J Mammal Evol Fig. 6 Illustrations of the antebrachial muscles in D. novemcinctus. a caudal view; b cranial view J Mammal Evol The medial tendon emerges from the deep fibers and is the thicker, major muscle tendon (Fig. 5a). It passes primarily to digit III, with a small tendinous slip going to the distal phalanx of digit II. The lateral muscle tendon is formed by the superficial fibers and then passes to digit IV. This muscle does not provide a tendon to digit V. Within digits III and IV, the tendons bifurcate to insert along much of the medial and lateral margins on the palmar side of the middle phalanx. Just prior to this insertion, the split in each tendon forms a broad, flat area that overlies the deeper tendons of m. flexor digitorum profundus, thus holding the profundus tendons against the proximal phalanx bones. m. flexor digitorum profundus tendon runs deep to the tendons of m. flexor digitorum superficialis. M. Pronator Quadratus This muscle is absent in the nine-banded armadillo. M. Brachioradialis This muscle is absent in the nine-banded armadillo. M. Extensor Carpi Radialis M. Flexor Digitorum Profundus The m. flexor digitorum profundus is a large and complex muscle that is composed of two humeral (medial and profundus), one ulnar, and one radial head. The humeral medial head (caput humerale mediale) has a tendinous origin from the medial epicondyle of the humerus as part of the common flexor tendon (Figs. 2, 5a and 6a). The ulnar head (caput ulnare) has an extensive origin off of the medial ulna and olecranon process, with a tendinous attachment coming from the tip of olecranon (Figs. 2 and 5a). The proximal portion of this origin is deep to the m. tensor fascia antebrachii. The humeral medial head quickly joins with the large ulnar head and the two are highly fused along their muscle belly lengths (Fig. 6a). These two muscle heads are the largest of the m. flexor digitorum profundus and both display a heavy investment of internal tendon. The radial head (caput radiale) originates from the superior radial ridge, near the middle onethird of the radius (Figs. 2 and 6a). The small humeral profundus head (caput humerale profundus) lies between the humeral medial and radial heads (not illustrated). It arises from the cranial portion of the medial epicondyle of the humerus (Fig. 2). Variable degrees of fusion of these heads were also observed between limbs on some specimens. In one individual, the radial head showed extensive fusion with the humeral medial head. The humeral profundus head also showed a high degree of fusion with the humeral medial head in another specimen. All four muscle heads merge together to form a single muscle belly distally that gives rise to a broad common tendon that contains a large sesamoid bone. The humeral profundus and radial heads join the common tendon on the medial side of the sesamoid bone, with the radial head joining more superficially (Fig. 6a). Beyond the sesamoid bone, the common tendon then divides into three large tendons, which continue through the manus to insert onto the distal phalanx bones of digits III–V. No tendon is provided to digit II. As they enter digits III and IV, the The primary area of origin for m. extensor carpi radialis is from the lateral supracondylar ridge of the humerus, where it arises via tendinous fibers located on its deep side (Fig. 1). In addition, fibers also join the main belly from the lateral side of the proximal radius, just distal to the elbow joint (Figs. 1 and 5b). These two components join to form a muscle belly that is somewhat tendinous on its deep surface. Distally, at the point where the muscle passes deep to m. abductor digiti I longus (Fig. 5a), two tendons emerge from the belly. The medial tendon splits and inserts onto the dorsal aspect of the shaft of metacarpal III. The lateral tendon inserts onto the medial side of the shaft of metacarpal IV. M. Extensor Carpi Ulnaris The m. extensor carpi ulnaris originates from the lateral-most edge of the mid- and proximal ulna (Fig. 1), with additional fibers coming from the fascia separating it from the m. abductor digiti V longus (see below). A small slip also originates from the common extensor tendon off of the lateral epicondyle of the humerus. The muscle belly runs distally along the lateral side of the ulnar ridge and styloid process of the distal ulna (Fig. 6b). An internal tendon runs through the mid-belly and superficial tendinous fibers are present near its origin. It inserts onto the proximolateral end of metacarpal V. Near its insertion, the tendon contains a small sesamoid bone. M. Extensor Digiti II The m. extensor digiti II (or m. extensor indicis: Macalister 1875) originates from a small area on the lateral aspect of the olecranon process, near the ulnar ridge (Fig. 1). The muscle shows varying degrees of fusion with m. extensor digitorum communis. The tendon passes deep to the extensor retinaculum to enter the manus (Fig. 5b) and inserts on the dorsal surface of the distal phalanx of digit III (Fig. 6b); the muscle insertion is modified in D. novemcinctus, which have lost digit I. J Mammal Evol M. Extensor Digitorum Communis The m. extensor digitorum communis arises from the lateral epicondyle of the humerus via the common extensor tendon, where is it is fused with the origin of m. extensor carpi ulnaris. The tendon passes deep to the extensor retinaculum along with the tendon of m. extensor digiti II and inserts primarily on the distal phalanx of digit IV, with an additional small tendon splitting off to attach to the distal phalanx of digit V (Fig. 6b). M. Abductor Digiti I Longus The m. abductor digiti I longus shows an extensive origin along the length of the lateral ulna and interosseous membrane, and a small area of fleshy origin from the distal radial ridge of the radius (Fig. 1). These muscle components merge and then form a thin tendon that passes deep to the extensor retinaculum to insert onto the lateral mid-shaft of metacarpal II (Figs. 5a and 6b); the muscle insertion is modified in D. novemcinctus, which have lost digit I. The muscle displays external tendinous fibers on its superficial aspect near the origin, and a central tendon that runs the length of the belly to the insertion. As it passes over the distal radial ridge, it is the most superficial muscle of the cranial antebrachium (Figs. 5b and 6b). This muscle was previously misidentified as either the m. extensor ossis metacarpi pollicis (Macalister 1875) or the m. extensor pollicis longus (Windle and Parsons 1899). M. Abductor Digiti V Longus The m. abductor digiti V longus (or m. extensor minimi digiti: Macalister 1875) has an origin via the common extensor tendon from the lateral epicondyle of the humerus and laterally from the intermuscular septum separating it from the m. extensor digitorum communis (Fig. 1). This muscle forms a wide tendon that passes deep to the extensor retinaculum at the carpal joint to then insert onto the mid-lateral aspect of metacarpal V (Figs. 5b and 6b). It displays external tendinous fibers on the deep side of its muscle belly at the origin and on the superficial side near the insertion, with these tendinous inscriptions continuing almost the entire length of the muscle belly. This muscle is also synonymous with the m. extensor digitorum lateralis. Fig. 7. Mean total forelimb muscle mass for the armadillo is 131±8.9 g, which accounts for 3.58±0.2 % of total body mass. Mean total intrinsic muscle mass is 61.7±4.4 g, which accounts for 1.70±0.2 % of total body mass. The limb retractors are the most massive group, accounting for 51.2±0.1 % of total forelimb muscle mass. The m. pectoralis superficialis and m. latissimus dorsi, along with the m. triceps brachii angular head, are the major mass contributors to this group. The second most massive group is the elbow extensors, which account for 22.7±1.9 % of total forelimb muscle mass (Fig. 7). This functional group contains the four relatively massive heads of the m. triceps brachii, along with the thin m. tensor fasciae antebrachii and the small m. anconeus. The scapular elevator/stabilizers (e.g., m. trapezius and m. rhomboideus) account for 24.0±1.8 % of total forelimb muscle mass, and all of the muscles in this functional group are extrinsic muscles of the forelimb. The complex digital flexors account for a moderate amount of total forelimb muscle mass at 6.55± 0.3 %, and are more massive than all other functional muscle groups of the antebrachium combined (Fig. 7). Functional Osteological Indices Means (± SD) of each osteological index are calculated and presented in Table 2. The large teres process results in a high scapular index (SI) for the robust scapula of D. novemcinctus, indicating a large area for muscle attachment. The index that indicates the relative area for muscle attachment about the distal humerus epicondyles (EI) is also high and has a mean value of 0.33±0.02 (Table 2). Average values of the bone robusticity indices are 0.10 (HRI) and 0.12 (URI) for the humerus and ulna, respectively, and reflect high bone bending strength and overall large areas for muscle attachment in the brachium and antebrachium. The manus proportions index (MANUS) relates to the out-lever length of the manus (Samuels and Van Valkenburgh 2008). The MANUS mean value of 0.48±0.02 is the first reported for any species of armadillo (Table 2). Lastly, the index that relates to the estimated outforce applied to the substrate per unit muscle in-force (TMOI) is substantial. The large olecranon process and stout manus (i.e., shortened metacarpal III and proximal phalanx III lengths) results in a high TMOI, indicating the ability to apply high out-force to the substrate during scratch-digging. M. Supinator MHC Isoform Composition This muscle is absent in the nine-banded armadillo. Limb Muscle Mass Distribution The forelimb of D. novemcinctus contains eight extrinsic and 30 intrinsic muscles (excluding intrinsic muscles of the manus) that were studied. Muscles are organized into functional groups based on their main action (Table 1) and shown in Means (± SD) of percent MHC content of each muscle are calculated and presented in Table 3. All except three forelimb muscles studied (TT, FCU, and FDS) show expression of four MHC isoform bands: MHC-1, 2A, 2X, and 2B. However, MHC-2B content is generally low, with the SD often exceeding the mean because of inter-individual variability in expression of this fast isoform. In only three muscles (TMJ, BB, and J Mammal Evol 60 Percentage Forelimb Muscle Mass (%) Fig. 7 Distribution of functional group muscle mass to total forelimb muscle mass in D. novemcinctus. Total forelimb muscle mass was calculated as the summed mass of all individual muscles studied. Proximal-todistal muscle group mass is expressed as a percentage, with bars representing means for each functional group. Error bars represent the SD (standard deviation). Muscles with synergistic functions are combined in one functional group. Biarticular muscles are also included in more than one functional group 50 40 30 20 10 0 Scapula Limb Limb Limb Elevator/ Retractors Adductors Protractors Stabilizers PT) is the fast 2B isoform expressed in all individuals. The percent content of slow MHC-1 is the next lowest and its expression also shows a range of variability across individuals (Table 3). Overall, fast MHC-2X is the predominant isoform expressed in all muscles studied, particularly in the extrinsic and intrinsic muscles of the shoulder joint. Beyond the shoulder joint, there is a proximal-to-distal decrease in the mean percentage of the fast MHC-2X isoform accompanied by a shift to relatively greater expression of slow MHC-1 and fast 2A isoforms (Fig. 8). Overall, the limb retractor muscles collectively contain 16.4 % MHC-1, 30.1 % MHC-2A, 46.9 % MHC-2X and 6.6 % MHC-2B; the elbow extensor muscles contain 23.6 % MHC-1, 31.2 % MHC-2A, 37.1 % MHC-2X, and 8.1 % MHC-2B; and the carpal/digital flexor muscles contain 25.3 % MHC-1, 42.5 % MHC-2A, 30.5 % MHC2X, and 1.7 % MHC-2B isoforms (Fig. 8). Discussion The forelimb musculature, osteological indices, and MHC isoform content reported here for the nineTable 2 Means ± SD for each functional osteological indices of D. novemcinctus Elbow Flexors Elbow Extensors Carpal Flexors Carpal Extensors Digital Flexors Digital Pronators Extensors banded armadillo all provide strong support for its behavioral classification as a semi-fossorial mammal. Similar to other terrestrial scratch-diggers studied (e.g., Gambaryan 1974; Moore et al. 2013; Warburton et al. 2013; Rupert et al. 2015), forelimb muscle mass in D. novemcinctus is concentrated in the limb retractor, elbow extensor, and digital flexor muscle groups. These functional groups combined accounted for over 80 % of total forelimb muscle mass. Correspondingly, the scapula and humerus displayed enlarged surface areas for muscle origins, resulting in high values for SI and EI. In addition to area for muscle attachment, brachial and antebrachial bone bending strength are also enhanced, as indicated by high values for HRI and URI. Functionally, high values for the MANUS and TMOI indices further indicate that the forelimb has an enhanced mechanical advantage for the application of force to the substrate during scratch-digging. Analysis of MHC isoforms showed predominant expression of the fast MHC-2X isoform along with a progressive proximal-to-distal shift to slower isoforms among the three major functional muscle groups. Fast-contracting limb retractor and SI EI HRI URI MANUS TMOI 0.90±0.02 0.33±0.02 0.10±0.01 0.12±0.004 0.48±0.02 0.50±0.04 Means were computed from N=4 specimens analyzed SI Scapular index, EI Epicondyle index, HRI Humeral robusticity index, URI Ulnar robusticity index, MANUS Manus proportions index, TMOI Triceps metacarpal out-force index Definitions for each index are found in text J Mammal Evol Table 3 Mean % MHC isoform content in selected forelimb muscles from D. novemcinctus Muscle N MHC isoform content 1 2A 2X 2B TT 3 21.2±7.9 6.5±7.9 72.2±14.8 0.0 LAT PS DA DS ISP TMJ BB TBA TBM TBL PT FCU FDS FDPHM 4 3 4 3 3 3 4 4 3 4 4 4 4 4 13.9±7.3 16.4±21.7 15.7±11.5 6.21±4.0 16.7±6.5 16.9±4.7 11.5±2.8 24.5±14.8 38.9±18.4 7.4±10.2 15.3±2.5 26.9±9.1 33.0±9.5 16.0±13.4 36.0±4.2 36.5±8.3 30.2±4.6 26.5±14.2 33.5±5.8 33.2±2.4 21.3±4.3 38.6±10.4 31.8±1.0 23.2±17.6 35.8±4.7 35.0±7.9 43.0±1.9 49.4±5.2 46.8±12.1 40.7±24.7 52.6±16.9 55.4±6.9 45.8±8.3 35.0±3.1 46.6±12.0 26.7±13.2 23.4±9.9 61.1±25.3 42.5±6.7 38.1±14.9 23.9±7.5 29.5±9.0 3.3±4.4 6.5±7.7 1.59±3.2 12.0±15.1 4.0±7.0 14.9±3.6 11.8±7.6 10.2±11.9 5.9±7.8 8.3±6.0 6.4±1.6 0.0 0.0 5.1±5.9 N = number of muscles All data are mean ± SD Means for each muscle were computed from 2 to 4 independent gel experiments per muscle/individual TT Trapezius pars thoracica, LAT Latissimus dorsi, PS Pectoralis superficialis, DA Deltoideus pars acromialis, DS Deltoideus pars scapularis, ISP Infraspinatus, TMJ Teres major, BB Biceps brachii, TBA Triceps brachii angular head, TBM Triceps brachii medial head, TBL Triceps brachii lateral head, PT Pronator teres, FCU Flexor carpi ulnaris, FDS Flexor digitorum superficialis, FDPHM Flexor digitorum profundus humeral medial head Fig. 8 MHC isoform content for the major functional muscle groups associated with scratchdigging in D. novemcinctus: limb retractors, elbow extensors, and carpal/digital flexors elbow extensor muscles that have enhanced mass and mechanical advantage indicate a limb system capable of excavating earth with high force and power. This conclusion will need to be verified with future evaluations of similar muscle traits in additional species of armadillos. At present, few comparative data for muscle mass distribution are available for armadillos. The fossorial pink fairy armadillo (C. truncatus), which burrows extensive fodder tunnels, has a large forelimb-to-hind limb muscle mass ratio of 2.4 (Gambaryan et al. 2009). Although some armadillos display more highly fossorial lifestyles, they still have relatively well-developed hind limbs for terrestrial locomotion, as compared to subterranean species such as golden moles whose ratios may surpass 4.0 (Gambaryan et al. 2009). The large hairy armadillo (C. villosus) has a ratio of 0.8 for forelimb-to-hindlimb muscle mass, which is more similar to a ratio of 0.4 reported for D. novemcinctus (Gambaryan et al. 2009). However, C. villosus relies on digging as a large part of its foraging strategy in addition to digging burrows for shelter (Smith 2008). Thus, the comparatively low ratio for D. novemcinctus supports its semi-fossorial classification and reflects its more generalized habits. The low forelimb-to-hind limb muscle mass ratio of D. novemcinctus is also suggestive of hind limbs better suited for terrestrial locomotion (Vizcaíno and Milne 2002; Copploe et al. 2015); however, nine-banded armadillos rarely use their fast running walk gait, instead preferring to walk slowly while searching for food, with short bursts of speed reserved for evasion of predators. Nine- MHC-1 MHC-2A MHC-2X MHC-2B Limb Retractors Elbow Extensors Carpal/Digital Flexors 0 10 20 30 50 60 40 Percent Distribution (%) 70 80 90 100 J Mammal Evol banded armadillos also display vertical jumping as a defensive and evasive behavior (Loughry and McDonough 2013). Despite the observation of robust hind limbs potentially for rapid evasive behaviors, the forelimb skeletal anatomy and muscle mass distribution among the limb retractors, elbow extensors, and carpal/digital flexors reported here for D. novemcinctus indicate numerous modifications for scratch-digging behavior. The power stroke for scratch-digging involves retraction of the limb at the shoulder joint, along with extension of the elbow and flexion of the carpus and digits (Hildebrand 1985; Moore et al. 2013). Forelimb retraction is emphasized in D. novemcinctus by the extensive muscle mass invested in this action. Limb retractors act on both the scapula and humerus and have the ability to apply a substantial flexor moment to the shoulder joint. The majority of these muscles have long, parallel fiber architecture indicating their ability to shorten at high velocity. Correspondingly, these are powerful muscles that have a relatively fast MHC isoform composition, with well over double the percentage of fast MHC-2X to slow MHC-1. MHC-2X fibers have intermediate glycolytic and oxidative properties, which are appropriate for moderate bouts of sustained high force and power output. This method of burrow excavation seems to match descriptions of the observed digging habits of D. novemcinctus (Taber 1945; Nowak 1999). Numerous modifications to the forelimb skeleton also indicate the importance of limb retraction in this species. For example, the elongated scapular spine and acromion provide an extended area for the insertion of m. trapezius pars thoracica and the origin of m. deltoideus, while the greatly enlarged teres process of the scapula not only provides a large area of origin for the unipennate m. teres major, but also increases its muscle moment arm allowing for the application of appreciable flexor torque at the shoulder joint. In addition, the teres process and extensive posterior scapular border give a large area of insertion for m. rhomboideus thoracis, a muscle that is advantageous for scratch-digging, as evident in other species that also require powerful retraction of the forelimb (Ercoli et al. 2015). The elbow extensors account for a significant fraction of the total forelimb muscle mass in D. novemcinctus, and by in large, occupy the majority of muscle mass in the brachium. In mustelids, the large mass of elbow extensors is directly correlated to the degree of fossorial ability (Ercoli et al. 2015). However, scratch-digging caniforms may have greater elbow flexor muscle mass (Ercoli et al. 2015) than is seen in D. novemcinctus, where m. biceps brachii and m. brachialis contribute little mass to the limb system, indicating their potentially lower functional significance during scratch-digging in the nine- banded armadillo. That said, other functional implications for the elbow flexors (e.g., prey manipulation) cannot be discarded without further study, because the diversity of habits of armadillos are not well known. The primary muscles of the elbow extensor group are the multiple heads of m. triceps brachii (angular, long, medial, and lateral) and m. tensor fasciae antebrachii. Together, these muscles provide the large torque needed for extension of the elbow during the power stroke. The biarticular m. triceps brachii caput angulare also contributes to limb retraction by the application of high flexor torque at the shoulder joint (Table 5). This is evident by its unipennate fascicles, large PCSA, and long muscle moment arm at the shoulder joint. A similar observation and functional interpretation for both the caput angulare and caput longum were made for the m. triceps brachii of the semi-fossorial American badger, Taxidea taxus (Moore et al. 2013). The elbow extensors also have an overall slower MHC isoform content than that of the limb retractors, with less fast MHC-2X and more slow MHC-1, thus providing greater fatigue resistance. Therefore, the contractile physiology of the elbow extensors is consistent with the high mechanical advantage (Vizcaíno and Milne 2002) and out-force potential (TMOI; Table 2) of the functional group as provided by the long olecranon process and shortened radius. The modifications to the forelimb skeleton in D. novemcinctus enhance muscle torque at the elbow joint without requiring all of the elbow extensors to be composed of primarily fast-contracting MHC-2X and 2B fibers. In the antebrachium, the carpal and digital flexor muscles together accounted for 7.6 % of total forelimb muscle mass. Moreover, the digital flexors of D. novemcinctus are 13.8 % of intrinsic forelimb muscle mass (see Supplemental Data Figure), a value that is intermediate to that of T. taxus [17.7 % (Moore et al. 2013)] and the groundhog, Marmota monax [10.9 % (Rupert et al. 2015)]. During scratch-digging, these muscles are expected to initially flex the digits and carpus, causing the robust claws to pierce and dislodge the substrate, and then maintain this flexed position for substrate displacement during the power stroke (Stalheim-Smith 1984; Thewissen and Badoux 1986; Moore et al. 2013). The high EI value in D. novemcinctus reflects the enlarged size of the humeral medial epicondyle, which provides attachment for the large antebrachial flexor muscles. MHC isoform content for the carpal/digital flexors also continues the trend of increasingly slower-contracting fibers in the distal forelimb. The composition of slow MHC isoforms contributes to the fatigue resistance of these muscles, allowing them to sustain force of contraction to maintain the carpal, metacarpal, and interphalangeal joints in flexion throughout the digging power stroke. The relatively higher J Mammal Evol expression of the MHC-2A isoform in the digital flexor muscles seems most appropriate for this functional role in scratch-digging. Future studies of additional species of armadillos are needed verify the significance of MHC-2A expression in the distal limbs; however, these findings are consistent with recent observations in the carpal/ digital flexors of burrowing M. monax (Rupert et al. 2015). The largest muscle of the antebrachium is the complex m. flexor digitorum profundus. This muscle is used for unified movements of the digits, with all four heads initially joining into a single common tendon that incorporates a large sesamoid bone before dividing into individual digital tendons. The sesamoid bone increases its muscle moment arm about the carpus and thus its mechanical advantage, allowing the m. flexor digitorum profundus to augment the total force applied to the substrate. In contrast, the much smaller m. flexor digitorum superficialis appears to allow for greater fine control of the digits, with separate tendons for digits III and IV emerging from the muscle belly. Both digital flexor muscles have relatively slow MHC isoform content, with MHC-2A being predominant. Except for the m. pronator teres, where fast MHC-2B was universally expressed, the 2B isoform was generally not found in distal forelimbs (two individuals contained a very small percentage of MHC-2B in only the m. flexor digitorum profundus). Slower contractile velocity and sustained force is needed for scratch-digging where the power stroke is performed repeatedly. Therefore, digging requires high metabolic energy expenditure and this is difficult for armadillos to sustain with their heavy, robust limbs and low basal metabolic rates (Vizcaíno and Milne 2002; Loughry and McDonough 2013). Perhaps it is for this reason, that the limbs of armadillos have retia mirabilia as an energy saving adaptation (Vaughn et al. 2013). MHC-2A fibers and the ability to supply the carpal/digital flexors with continuous oxygenated blood supports the postural function of these muscles to maintain flexion of the distal limb joints during scratchdigging. Lastly, the m. flexor carpi radialis of D. novemcinctus was previously determined to have a large distribution of fast, glycolytic fibers (McConathy et al. 1983), which infers the MHC-2B isoform fiber type. This finding should be interpreted with caution because expression of the fast 2B isoform in a mammal as large as D. novemcinctus is somewhat surprising, and thus it is not known if MHC-2B expression represents an adaptive function or phylogenetic retention. Fast MHC-2B isoform fibers in some species may be attributed to burst escape behaviors such as jumping and rapid excavation (Nowak 1999), in addition to possibly being retained for thermoregulatory purposes. Further study of internal muscle properties across armadillo species is needed to confirm these hypotheses; however, we speculate that expression of the 2B isoform is adaptive for armadillos. For example, despite having the highest mechanical advantage of all armadillo forelimbs (Vizcaíno and Milne 2002), C. truncatus is capable of high-speed excavation in loose soils (Rood 1970; Nowak 1999). To overcome the extreme force advantage of its forelimbs, we predict that muscles with long, fast MHC-2B fibers would be required to achieve the velocity of joint rotation needed to dig rapidly. The relatively lower mechanical advantage of the shoulder musculature of D. novemcinctus compared with that of other species of armadillos (Vizcaíno and Milne 2002) may also help explain the faster MHC isoform composition of the limb retractor group and the progressive proximal-to-distal shift to slower contractile properties. Comparative Myology and Phylogenetic Perspectives Understanding the functional morphology of D. novemcinctus furthers our understanding of the family Dasypodidae, as well as specialization in other scratchdigging mammals. Table 4 compares myological data obtained from D. novemcinctus with that of other dasypodids for which similar data exists (Galton 1869; Macalister 1875; Windle and Parsons 1899; Miles 1941). This information is useful for relating species both functionally and phylogenetically, and it is critical for our future phylogenetic reconstructions of the Cingulata based on analyses including numerous postcranial morphological characters. In dasypodids observed thus far, there are no records of m. omotransversarius and m. cleidocephalicus (Cuvier and Laurillard 1849; Macalister 1875). Moreover, there are no accounts of the m. rhomboideus profundus (Table 4). Each of these muscles contributes to protraction of the limb (Ercoli et al. 2015), and thus their absence contributes to the disparity between limb protractor and retractor mass in D. novemcinctus. In caniforms, the absence of m. rhomboideus profundus is considered an ancestral feature (Fisher et al. 2009), and this also appears to be true for xenarthrans, which represent one of the most basal placental mammal clades. In some species of scratch-digging sciurids, the m. deltoideus pars clavicularis and m. cleidobrachialis are separate and distinct, with the m. cleidobrachialis commonly joining with the m. brachialis to insert onto the ulna (Thorington et al. 1997; Bezuidenhout and Evans 2005). However, because the m. deltoideus pars clavicularis is often used interchangeably with m. cleidobrachialis (Ercoli et al. 2015), this makes it difficult to directly compare with some historical myological descriptions of the m. deltoideus. We consider the m. cleidobrachialis to be absent in D. novemcinctus and choose to maintain nomenclature separate from m. J Mammal Evol Table 4 Comparative forelimb muscle traits in the nine-banded armadillo and other armadillos Trait Dasypus novemcinctusa,c Euphractus sexcinctusb,c,d Chaetophractus villosusd Chlamyphorus truncatusc,d Origin of trapezius pars thoracica Rhombiodeus profundus Origin of serratus ventralis (number of muscle bellies) Articularis humeri Coracobrachialis Teres minor Biceps brachii short head Brachioradialis Triceps brachii angular (number of scapular heads) Pronator quadratus Palmaris longus Digits served by flexor digitorum superficialis Insertion of flexor carpi radialis Supinator Digit served by abductor digiti I longus Digits served by abductor digiti V longusi Number of digits on forefeet All thoracic vertebrae Absent Ribs 2–7 (variable) C4–T12 Absent Ribs 1–7 (no data) C4–T12 Absent Ribs 1–8, 9 (2) – Absent Ribs 1–7 (7–8) Variable Present Variablee Variable Absent Present (2) Present Present Present Present Absent Present (2) Present Present – Present Absent Present (2) Variable Variable Present Absent Absent Present (variable) Absent or reduced Absentf II–IV Absent Absent or fibers fused with FDS II, III Absent Fibers fused with FDS III Absent or reduced – – Metacarpal II Absent IIh To digit V 4 Trapezium Variableg I To digits IV, V 5 Trapezium Variable I To digits IV, V 3–5 Metacarpal II Absent I To digits IV, V 5 Data sources: a Miles (1941) and this study; b Galton (1869); c Macalister (1875); d Windle and Parsons (1899) e often highly fused with m. infraspinatus; f may be fused with FDS (see text for discussion); g highly reduced when present; h digit I is absent, and digits II and Vare reduced: muscle is modified to abduct digit II; i synonymous with m. extensor digitorum lateralis in species with five digits: muscle belly is separate and has an abductor action deltoideus until our future observations can confirm its presence (or absence) in dasypodids. The m. articularis humeri is also referred to as m. coracobrachialis brevis, and it is observed in both E. sexcinctus and C. villosus (Windle and Parsons 1899), with a variable presence in C. truncatus and D. novemcinctus (Macalister 1875). Lastly, the presence of m. coracobrachialis (or m. coracobrachialis longus) is a primitive feature in caniforms (Fisher et al. 2009), and its presence in D. novemcinctus, E. sexcinctus, and C. villosus (Galton 1869; Macalister 1875; Windle and Parsons 1899), along with its attachment to the humerus, may also be suggestive of ancestral traits in armadillos (Table 4). The insertion of m. biceps brachii exclusively onto the ulna in D. novemcinctus is somewhat unique compared with other scratch-digging taxa, where a radial insertion is most common (Bezuidenhout and Evans 2005; Moore et al. 2013). This modified insertion is shared with C. truncatus (Macalister 1875). In E. sexcinctus, the m. biceps brachii is described as inserting onto both the radius and ulna (Galton 1869), and this condition is unknown in other armadillo species (Windle and Parsons 1899). An insertion on the ulna alone prevents the m. biceps brachii from acting on the radius for antebrachial supination. The absence of m. brachioradialis (or m. supinator longus) and m. supinator (or m. supinator brevis) in D. novemcinctus further indicates a lack of supination ability. The m. brachioradialis is reported as absent in all species of armadillo for which data are available (Table 4), but a rudimentary m. supinator may be present in both E. sexcinctus (Galton 1869) and C. villosus (Windle and Parsons 1899), potentially allowing for some degree of antebrachial supination in these species. In other animals, the presence of each muscle is often associated with manipulation of objects or subduing of prey, as is commonly observed in feliforms with a high capacity for distal limb supination (Julik et al. 2012). These abilities are also related to an un-fused radius and ulna; therefore object manipulation and prey capture are well developed in carnivorans that are less cursorial or noncursorial (Iwaniuk et al. 2000). Dasypus novemcinctus is not known to demonstrate either of these behaviors (McBee and Baker 1982; Loughry and McDonough 2013), and the radius and ulna are strongly fused in this species, further constraining its foot posture to plantigrade. Other species of armadillos may also show high degree of fusion between the radius and ulna and J Mammal Evol consequently, retain little-to-no muscle mass for pronation-supination (Table 4), and these features may be associated with specialization for scratch-digging in the Cingulata as opposed to cursorial habit. Another anatomical constraint for use of the distal forelimb for scratch-digging in armadillos is reflected in the myology of the elbow extensors. Complexity involving the number of heads of the m. triceps brachii reflects the degree of joint position control that can be achieved in elbow flexion-extension (Julik et al. 2012; Ercoli et al. 2015). Accessory heads of the m. triceps brachii have not been observed in armadillos, suggesting a limb design for gross movement rather than fine motor control. However, two muscle heads of the m. triceps brachii having an origin on the scapula are present in all armadillos studied (Table 4), along with anteaters (Windle and Parsons 1899), and mustelids and mephitids (Ercoli et al. 2015), which shows good correspondence with fossorial habits (see Ercoli et al. 2015 for an alternative functional explanation relative to the acquisition of this muscle belly in carnivorans). Moreover, the m. tensor fasciae antebrachii is present in all xenarthrans (Windle and Parsons 1899). It has been suggested that when present in caniforms, the m. triceps brachii caput angulare may have its evolutionary origin as a specialization of the m. tensor fasciae antebrachii (Ercoli et al. 2015). A similarly strong selection for scratch-digging may have resulted in convergent evolution of a distinct m. triceps brachii caput angulare in armadillos. Finally, the m. palmaris longus was not observed in any of our specimens of D. novemcinctus. However, the superficial portion of m. flexor digitorum superficialis may serve the function of, or contribute fibers to, a rudimentary m. palmaris longus, as observed in E. sexcinctus (Galton 1869) and C. villosus (Windle and Parsons 1899) (Table 4). The muscle belly arrangement of m. flexor digitorum superficialis in D. novemcinctus is notable, with each of two sets of fibers (superficial and deep) providing a tendon to separate digits, but lacking a tendon to digit V. In E. sexcinctus, the m. flexor digitorum superficialis serves digits II and III along with tendons from m. palmaris longus (Galton 1869). Precise control of individual digits is not as necessary for scratch-digging behaviors and therefore, the apparent lower complexity of armadillo digital flexors may be a derived feature. Accounts of myology for additional armadillo species are needed to support this hypothesis. In regards to the digital extensors, historically the m. abductor digiti I longus and m. abductor digiti V longus, which are both present in all armadillos previously studied, have been incorrectly referred to as m. extensor pollicis and m. extensor digiti minimi, respectively (Windle and Parsons 1899). Discrepancies in nomenclature for armadillos are related to the number of digits on the forefeet (Table 4) and the use of human anatomical references. A clear action of abduction for each muscle was observed in our study of D. novemcinctus and from the descriptions available in the literature, we suspect this is also the condition in other species. However, the loss of digit I in D. novemcinctus most likely resulted in a reorganization of m. abductor digiti I longus to insert onto digit II and a modification of m. extensor indicis to insert on digit III. Because other species of armadillo with five digits on their forefeet possess an m. extensor indicis (Macalister 1875; Windle and Parsons 1899), we chose nomenclature that maintained consistency with function rather than a reassignment of digit number (and potential misidentification of muscle action). For reference, some members of the Carnivora also have abductors of digits I and V, and these muscles are considered to be a primitive feature in the order (Fisher et al. 2009). Therefore, the condition of armadillos with fewer than five digits on their forefeet is likely a derived feature for which the functional significance requires further investigation. Concluding Remarks Postcranial morphological data are severely lacking for basal placental mammalian clades. In particular, the general lack of limb muscle characters from armadillos (Dasypodidae) and the other xenarthrans limits our interpretations of their phylogeny and ability to statistically evaluate interspecies variability. Myological studies involving quantification of muscle architectural properties provide additional information regarding soft tissue structure, which are traits often assumed in previous studies. A more complete account of muscle (and bone) characters from both the cranium (Gaudin and Wible 2006) and postcranium of armadillos would be valuable to help clarify uncertainty surrounding monophyletic or paraphyletic origins of the family Dasypodidae. Data from this study and our future studies will be used to answer questions about armadillo ecomorphology and phylogeny. Re-evaluation of the phylogeny of the Cingulata will not only provide insight into structure and function in armadillos and other scratch-digging mammals, but also may help bring a greater understanding to the evolution of eutherian mammals. Acknowledgments We sincerely thank W. Loughry (Valdosta State University) and J. Copploe (YSU) for armadillo collection. A very special thanks to Theresanne DeMartino for the anatomical illustrations. We also thank M. Thornhill for preparation of muscle protein stocks and H. Suzuki for help with gel photography. The YSU Department of Biological Sciences and Ohio University Department of Biomedical Sciences are also gratefully acknowledged. J Mammal Evol Appendix 1 Table 5 Muscle architectural property measurements for the forelimb muscles of D. novemcinctus Muscle ML (cm) lF (cm) θ (°) TC TT RC RT LAT PS SC DA DS DC SSP ISP TMJ SUB CCB BB 6.83±0.4 7.54±1.3 5.42±0.9 3.96±1.0 13.0±1.9 12.9a 5.6a 2.92±0.6 8.47±0.8 2.26±0.3 6.80±0.2 6.89±0.4 6.43±0.9 5.76±0/5 4.13±0.6 3.99±0.3 5.00±1.0 5.70±1.6 5.11±1.1 3.29±0.9 9.96±3.2 7.67±4.6 3.23±0.4 2.45±0.6 6.27±1.5 1.94±0.4 2.21±0.4 1.50±0.5 1.99±1.0 1.89±0.8 0.79±0.1 1.31±0.6 0 0 0 0 0 24±4 21±2 0 0 0 21±5 19±5 18±5 34±7 18±5 17±5 BCH TBA 4.74±0.5 6.05±1.4 3.67±0.5 3.80±1.6 0 24±5 TBLO 5.37±0.9 2.32±0.8 18±5 TBM TBL TFA ANC PT FCU FCR FDS -deep fibers -superficial FDPHM/FDPU FDPR FDPHP ECR ECU ED2 EDC AD1L AD5L 4.58±1.1 5.73±0.6 10.6±2.3 2.36±0.2 2.94±0.4 3.79±0.7 3.49±0.3 2.32±0.8 4.38±1.2 7.17±3.0 1.55±0.3 1.15±0.4 0.97±0.6 0.95±0.3 21±3 0 0 0 21±4 19±2 19±3 3.74±0.4 4.06±0.6 7.34±0.4 3.46±0.5 3.14±0.3 5.21±0.2 4.54±1.3 5.66±1.2 4.48±1.0 4.28±1.1 4.32±0.5 0.71±0.2 0.89±0.4 1.49±0.5 1.22±0.3 2.25±0.3 1.84±0.6 0.91±0.3 0.99±0.3 1.29±0.8 0.81±0.3 0.93±0.3 19±5 19±4 21±5 17±3 0 14±4 18±3 18±4 19±3 26±6 16±3 rm (cm) 2.83±0.8 0.77±0.1 1.92±0.6 1.09±0.3 0.90±0.5 1.65±0.6 0.62±0.1 0.42±0.1 0.90±0.3s 0.64±0.2e 0.52±0.1 3.42±0.2 s 2.71±0.4e 1.38±0.2s 2.51±0.1e 1.30±0.02 1.41±0.6 0.67±0.3 0.54±0.2 0.21±0.1 0.95±0.2 1.10±0.1 0.76±0.1 0.50±0.1 0.61±0.3 0.73±0.2 0.46±0.3 Volume (cm3) PCSA (cm2) Fmax (N) Power (W) 6.45 7.15 5.00 2.52 13.3 14.1 2.44 1.54 5.40 0.41 5.25 3.05 2.52 3.82 0.45 1.55 1.29 1.26 0.10 0.77 1.33 1.67 0.71 0.63 0.86 0.21 2.22 1.93 1.20 1.68 0.53 1.13 38.7 37.7 29.4 23.0 40.1 50.2 21.2 18.9 25.8 6.27 66.7 57.8 35.9 50.4 16.0 33.9 1.55 1.72 1.20 0.61 3.20 3.09 0.55 0.37 1.30 0.10 1.18 0.70 0.58 0.76 0.10 0.36 1.48 11.0 0.40 2.65 12.1 79.5 0.36 2.42 3.65 1.49 44.8 0.84 3.24 3.12 4.82 0.37 0.60 0.53 0.65 1.20 0.71 0.67 0.24 0.49 0.52 0.65 39.0 21.4 20.2 7.21 14.6 15.5 19.4 0.73 0.75 1.16 0.09 0.13 0.12 0.15 0.48 0.88 5.72 0.81 0.34 2.00 0.45 0.79 0.52 0.63 0.43 0.64 0.93 3.57 0.63 0.15 1.05 0.47 0.77 0.38 0.70 0.43 19.3 28.0 107 19.0 4.47 31.5 14.2 23.0 11.4 21.1 13.5 0.11 0.20 1.28 0.19 0.08 0.47 0.10 0.18 0.12 0.14 0.10 Torque (N.cm) 113 0.15 49.6 6.83 58.3 59.7 35.4 6.70 31.5s 21.6e 6.12 272s 215e 61.7s 112e 50.8 30.2 4.81 8.36 3.98 18.4 30.9 81.5 9.41 2.74 22.9 6.54 All measurements were made following the procedures of Moore et al. (2013) and Rose et al. (2013) on the left forelimbs of four individuals. Muscle moment arm (rm) was measured in situ with digital calipers; muscle belly length (ML) was measured with digital calipers following removal of any free tendons; muscle belly mass (MM) was recorded using an electronic balance; fascicle length (lF ) was measured from 5 to 10 random fascicles inside the muscle belly using digital calipers; and pennation angle (θ: to the nearest degree) was measured at 5–10 random locations inside the muscle belly using a goniometer Muscle volume was calculated by dividing MM by a muscle density of 1.06 g cm−3 . PCSA was calculated as (muscle volume/mean lF )×cos θ. Isometric force (Fmax) was estimated by multiplying PCSA by a maximum isometric stress of 30 N cm−2 . Joint torque was calculated as Fmax ×rm. Muscle power (W) was estimated to be one tenth the product of Fmax and Vmax, where Vmax is maximum fiber shortening velocity (in FL s−1 ). A value of 8.03 FL s−1 was calculated as Vmax for fast MHC-2A fibers at physiologic temperature for armadillos (see Moore et al. 2013 for details of these calculations) a single measurement s measured at the shoulder joint e measured at the elbow joint rm Muscle moment arm J Mammal Evol Appendix 2 Table 6 Raw osteological measurements from the left forelimb bones of D. novemcinctus Animal Body mass (kg) SL (cm) SW (cm) HL (cm) HW (cm) HEW (cm) OL (cm) UL (cm) UW (cm) MCL III (cm) PPL III (cm) A1 A2 A3 A4 5.40 5.91 5.75 5.55 5.66±0.2 2.10 2.01 2.04 1.98 2.03±0.05 1.65 1.91 1.90 1.73 1.80±0.1 3.8 3.8 3.5 3.6 3.7±0.2 6.12 6.60 6.17 6.27 6.29±0.2 5.84 6.42 6.22 6.15 6.16±0.2 0.58 0.65 0.63 0.68 0.63±0.04 2.84 2.95 2.71 2.89 2.85±0.1 6.57 6.99 6.77 6.56 6.72±0.2 0.81 0.83 0.82 0.75 0.80±0.04 0.77 0.91 0.88 0.87 0.86±0.06 Scapula length (SL), taken as the maximum length from the posterior border to the tip of the supraglenoid tubercle; scapula width (SW), taken as the maximum dorsoventral width between the posterior end of the vertebral border and the teres process; humerus length (HL), functional length from humeral head to distal end of the medial epicondyle; humerus width (HW), mediolateral width at mid-shaft; humerus epicondylar width (HEW), maximum width between the distal epicondyles; olecranon length (OL), length from proximal end of olecranon to center of the trochlear notch; ulna length (UL), functional length from olecranon process to the styloid process; ulna width (UW), mediolateral width at mid-shaft; metacarpal III length (MCL III), functional length from the proximal to the distal end; and proximal phalanx III length (PPL III), functional length from the proximal to the distal end In bold are mean ± SD References Aguiar JM, Fonseca GAB (2008) Conservation status of the Xenarthra. In: Vizcaíno SF, Loughry WJ (ed) The Biology of the Xenarthra. University Press of Florida, Gainesville, pp 215–231 Bezuidenhout AJ, Evans HE (2005) Anatomy of the woodchuck (Marmota monax). Am Soc Mammal, Stillwater Copploe JV II, Blob RW, Parrish JHA, Butcher MT (2015) In vivo strains in the femur of the nine-banded armadillo (Dasypus novemcinctus). J Morphol doi: 10.1002/jmor.20387 Cuvier G, Laurillard M (1849) Recueil de Planches de Myologie. Dusacq, Paris Delsuc F, Douzery EJP (2008) Recent advances and future prospects in xenarthran molecular phylogenetics. In: Vizcaíno SF, Loughry WJ (ed) The Biology of the Xenarthra. University Press of Florida, Gainesville, pp 11–23 Delsuc F, Superina M, Tilak M-K, Douzer EJP, Hassanin A (2012) Molecular phylogenetics unveils the ancient evolutionary origins of the enigmatic fairy armadillos. Mol Phylogenet Evol 62:673–680 Ercoli MD, Álvarez A, Stefanini MI, Busker F, Morales MM (2015) Muscular anatomy of the forelimbs of the lesser grison (Galictis cuja), and a functional and phylogenetic overview of the Mustelidae and other Caniformia. J Mammal Evol 22:57–91 Fisher RE, Adrian B, Barton M, Holmgren J, Tang ST (2009) The phylogeny of the red panda (Ailurus fulgens): evidence from the forelimb. J Anat 215:611–635 Galton JC (1869) The muscles of fore and hind limbs in Dasypus sexcinctus. Trans Linn Soc 26:523–565 Gambaryan PP (1974). How Mammals Run. John Wiley & Sons, New York Gambaryan PP, Zherebtsova OV, Perepelova AA, Platonov VV (2009) Pes muscles and their action in giant anteater Myrmecophaga tridactyla (Myrmecophagidae, Pilosa) compared with other plantigrade mammals. Russian J Theriol 8:1–15 Gaudin TJ, Wible JR (2006) The phylogeny of living and extinct armadillos (Mammalia, Xenarthra: Cingulata): a craniodental analysis. In: Carrano MT, Gaudin TJ, Blob RW, Wible JR (ed) Amniote Paleobiology: Perspectives on the Evolution of Mammals, Birds, and Reptiles. University of Chicago Press, Chicago, pp 153–198 Hazimihalis PJ, Gorvet MA, Butcher MT (2013) Myosin isoform fiber type and fiber size in the tail of the Virginia opossum (Didelphis virginiana). Anat Rec 296:96–107 Hildebrand M (1985) Digging of quadrupeds. In: Hildebrand M, Bramble DM, Liem KF, Wake DB (ed) Functional Vertebrate Morphology. The Belknap Press of Harvard University Press, Cambridge, pp 89–109 International Committee on Veterinary Gross Anatomical Nomenclature (2012) Nomina Anatomica Veterinaria, 5th edn. Hannover, Germany; Columbia, MO; Ghent, Belgium; Sapporo, Japan: Editorial Committee Iwaniuk AN, Pellis SM, Whishaw IQ (2000) The relative importance of body size, phylogeny, locomotion, and diet in the evolution of forelimb dexterity in fissiped carnivores (Carnivora). Can J Zool 78: 1110–1125 Julik E, Zack S, Adrian B, Maredia A, Parsa A, Poole M, Starbuck A, Fisher RE (2012) Functional anatomy of the forelimb muscles of the ocelot (Leopardus pardalis). J Mammal Evol 19:277–304 Kley NJ, Kearney M (2007) Adaptations for digging and burrowing. In: Hall B (ed) Fins into Limbs. University of Chicago Press, Chicago, pp 284–309 Laemmli UK (1970) Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 227:680–685 Loughry WJ, McDonough CM (2013) The Nine-Banded Armadillo: A Natural History (Animal Natural History Series). University of Oklahoma Press, Norman Macalister A (1875) A monograph of the anatomy of Chlamydophorus truncatus (Harlan), with notes on the structure of other species of Edentata. Trans R Irish Acad 25:219–278 McBee K, Baker RJ (1982) Dasypus novemcinctus. Mammal Species 162:1–9 McConathy D, Giddings CJ, Gonyea WJ (1983) Structure-function relationship of the flexor carpi radialis muscle among four species of mammals. J Morphol 175:279–292 McDonough CM, Loughry WJ (1997) Influences on activity patterns in a population of nine-banded armadillos. J Mammal 78:932–94. J Mammal Evol Miles SS (1941) The shoulder anatomy of the armadillo. J Mammal 22: 157–169 Mizunoya W, Wakamatsu J, Tatsumi R, Ikeuchi Y (2008) Protocol for high-resolution of rodent myosin heavy chain isoforms in a mini gelelectrophoresis system. Anal Biochem 377:111–113 Moore AL, Budny JE, Russell AP, Butcher MT (2013) Architectural specialization of the intrinsic forelimb musculature of the American badger (Taxidea taxus). J Morphol 274:35–48 Nowak RM (1999) Walker’s Mammals of the World, 6th ed. Johns Hopkins University Press, Baltimore Nyakatura JA, Fischer MS (2011) Functional morphology of the muscular sling at the pectoral girdle in tree sloths: convergent morphological solutions to new functional demands? J Anat 219:360–374 Rood JP (1970) Notes on the behavior of the pygmy armadillo. J Mammal 51:179 Rose JA, Sandefur A, Huskey S, Demler JL, Butcher MT (2013) Muscle architecture and out-force potential of the thoracic limb in the eastern mole (Scalopus aquaticus). J Morphol 274:1277–1287 Rose JA, Moore AL, Russell AP, Butcher MT (2014) Functional osteology of the forelimb digging apparatus in badgers. J Mammal 95:543–558 Rupert JE, Cordero Schmidt E, Moriera-Soto A, Rodriguez Herrera B, VandeBerg JL, Butcher MT (2014) Myosin heavy chain (MHC) isoform expression in the prehensile tails of didelphid marsupials: functional differences between arboreal and terrestrial opossums. Anat Rec 297:1364–1376 Rupert JE, Rose JA, Organ JM, Butcher MT (2015) Forelimb muscle architecture and myosin isoform composition in the groundhog (Marmota monax). J Exp Biol 218:194–205 Samuels JX, Van Valkenburgh B (2008) Skeletal indicators of locomotor adaptations in living and extinct rodents. J Morphol 269:1387–1411 Samuels JX, Meachen JA, Sakai SA (2013) Postcranial morphology and locomotor habits of living and extinct carnivorans. J Morphol 274: 121–146 Smith P (2008) Greater hairy armadillo Chaetophractus villosus. Mammals of Paraguay 11:1–15 Stalheim-Smith A (1984) Comparative study of the forelimbs of the semifossorial prairie dog, Cynomys gunnisoni, and the scansorial fox squirrel, Sciurus niger. J Morphol 180:55–68 Taber FW (1945) Contribution to the life history and ecology of the ninebanded armadillo. J Mammal 26:211–226 Talmadge RJ, Roy RR (1993) Electrophoretic separation of rat skeletal muscle myosin heavy-chain isoforms. J Appl Physiol 75: 2337–2340 Talmage RV, Buchanan GD (1954) The Armadillo (Dasypus novemcinctus): A Review of Its Natural History, Ecology, Anatomy, and Reproductive Physiology. Rice Institute, Houston Thewissen JG, Badoux DM (1986) The descriptive and functional myology of the fore-limb of the aardvark (Orycteropus afer, Pallas 1766). Anat Anz 162:109–123 Thorington RW Jr, Darrow K, Betts ADK (1997) Comparative myology of the forelimb of squirrels. J Morphol 234:155–182 Toniolo L, Cancellara P, Maccatrozzo L, Patruno M, Mascarello F, Reggiani C (2008) Masticatory myosin unveiled: first determination of contractile parameters of muscle fibers from carnivore jaw muscles. Am J Physiol Cell Physiol 295:1535–1542 Vaughn TA, Ryan JM, Czaplewski NJ (2013) Cingulata, Pilosa, and Pholidota. In: Mammalogy, 5th ed. Jones and Bartlett Publishers, Sudbury, Massachusetts, pp 147–160 Vizcaíno SF, Milne N (2002) Structure and function in armadillo limbs (Mammalia: Xenarthra: Dasypodidae). J Zool 257:117–127 Vizcaíno SF, Fariña RA, Mazzetta GV (1999) Ulnar dimensions and fossoriality in armadillos. Acta Theriol 44:309–320 Warburton NM, Grégoire L, Jacques S, Flandrin C (2013) Adaptations for digging in the forelimb muscle anatomy of the southern brown bandicoot (Isoodon obesulus) and bilby (Macrotis lagotis). Aust J Zool 61:402–419 Wilson DE, Reeder DM (2005) Mammal Species of the World. A Taxonomic and Geographic Reference, 3rd ed. Johns Hopkins University Press, Baltimore Windle BCA, Parsons FG (1899) On the myology of the Edentata. Proc Zool Soc London 1899:314–339