The ancestral form of the vertebrate trunk is a long tube enclosing, along much of its length, the coelomic cavity (Gegenbaur et al.,1878). The lateral plate mesoderm and the musculature that invest it form the walls and floor of this cavity, which are generally referred to as the body-wall and abdominal musculature, respectively (Burke and Nowicki,2003). Despite the structural importance of the body-wall and abdominal muscles and connective tissue, they have been curiously neglected in anatomical studies. As layered, laminar structures that differentiate from primordial embryonic mesodermal folds (Maurer,1898), they are best described with detailed attention to the topology of these layers. This level of detail is rarely achieved outside of human anatomy (e.g., Gray et al.,1995). In fact, it often seems in works focusing on the myology or gross anatomy of the entire body that descriptions of the deep dorsal axial musculature are relatively detailed but that the topological relations of body-wall muscles to abdominal muscles are not explicitly described (e.g., Sanders,1870,1872,1874; Osawa,1898).
In extant non-tetrapod vertebrates, the thick myotomal segmental musculature occupies most of the dorso–ventral extent of the body (Gegenbaur et al.,1878; Burke and Nowicki,2003). Separate abdominal musculature is not differentiated, though there is often a ventral section of muscle fibers with relatively horizontal orientations that is rolled inward toward the ventral midline in a transverse plane. Within tetrapods, a highly conservative arrangement generally exists (Gegenbaur et al.,1878; Francis,1934; Carrier,1989,1990,1993; Gray et al.,1995). The outermost layer of the body wall deep to the skin is the bilaminar superficial fascia, followed by the m. obliquus externus and its connective tissue, including an extensive aponeurosis, then by either the m. obliquus internus or, more anteriorly, the ribs and intercostal musculature. Deeper layers include the m. transversus abdominis, the transversalis fascia, and the peritoneum.
Along the ventral midline is the segmented m. rectus abdominis. The m. obliquus externus aponeurosis is either entirely superficial to the m. rectus abdominis or splits around it. The m. obliquus internus variably runs deep or superficial to the m. rectus abdominis, or splits around it. Together, the ventral body wall connective tissue forms the rectus sheath (Chouke,1934; Walmsley,1937; Anson et al.,1938; McVay and Anson,1940; Rizk,1980; Carr and Altig,1992). Some work suggests that the body-wall muscles, instead of being paired as they initially appear, are actually large, sheetlike digastric muscles whose aponeuroses decussate in complex ways at the linea alba in the ventral midline (Rizk,1980).
An unusual configuration of the m. rectus abdominis has long been reported in certain lizards, specifically those of the clade Autarchoglossa, which is consistently recovered by gross-anatomy-based phylogenies (Camp,1923; Moody,1983; Estes et al.,1988; Conrad,2008) but is paraphyletic with respect to Iguania in DNA-sequence-based phylogenies (Townsend et al.,2004). In autarchoglossan squamates, which include the skinks, teiids, anguimorphs, and possibly snakes, there is an unsegmented m. rectus (abdominis) lateralis, which “may usually be recognized by its position alongside the Rectus medianus and profundus and between the Obliquus externus superficialis and the median portion of the Obliquus externus profundus” (Camp,1923, p. 378). The m. rectus abdominis lateralis is not related to the “m. rectus abdominis superficialis” of Camp (1923) and other authors, which is part of the m. rectus abdominis proper as discussed at length by Moody (1983).
The m. rectus abdominis lateralis has come to be perceived as a “key” synapomorphy of Autarchoglossa (Estes et al.,1988). However, since the work of Camp (1923), only one major work (Moody,1983) has elaborated upon the anatomy and distribution of the m. rectus abdominis lateralis. In this work, the muscle was claimed to be absent in Iguania, and its absence was confirmed in Gekkota as also asserted by Camp (1923) and Kluge (1976). Nevertheless, a precise topological description of the relations and range of variation of the m. rectus abdominis lateralis was lacking.
In this context, and spurred by early investigations which revealed a muscle fitting the description of the m. rectus abdominis lateralis in Iguania, I untertook a study of the abdominal musculature and its precise topological relations to the external body-wall layers in autarchoglossan and non-autarchoglossan squamates (Bhullar,2007). My initial hypothesis was that the m. rectus abdominis lateralis, if defined by subtle differences in topological relations of muscle and connective-tissue layers, may be more widely distributed than previously realized.
MATERIALS AND METHODS
Dissections were performed under an Olympus stereo microscope fitted with a camera lucida, which was used to produce the non-schematic illustrations. A summary of taxa dissected appears below, with specimen number in parentheses in the case of catalogued specimens. Anolis carolinensis were obtained from Ward's Biological. The dissected Leiolepis belliana was THNC 56654 (Texas Natural History Collections, The University of Texas at Austin, Herpetology).
Minimally destructive dissections were performed on the TNHC specimens by the “windowing” of body-wall layers at various anteroposterior levels. Body-wall layer topology in A. carolinensis was confirmed by transverse sectioning, with the sections observed wet and after air-drying, when the connective tissue had shrunken and separation of layers was more evident.
A. carolinensis and Crotaphytus sp.
In A. carolinensis, the inner and outer layers of superficial fascia lie immediately under the skin (Fig. 1b). The outer layer is loose and contains fat deposits. The inner layer is tighter and more membranous, transmitting cutaneous branches of the segmental spinal nerves. Deep to the inner layer of superficial fascia, the m. obliquus externus is composed of two layers of fibers trending posteroventrally, but a split between the layers is not evident anywhere along the body, and they appear to be surrounded by a single membranous sheath. Along the mid-venter posterior to the m. pectoralis, the m. rectus abdominis is covered by a thick, gray–blue, fibrous sheath that bears an array of collagen fibers, mostly parallel on each side with those of the descending m. obliquus externus aponeurosis, which in turn continue the fiber direction of the more dorsal fleshy part of that muscle. The fibers thus cross each other at approximately right angles. The m. rectus abdominis is divided by the inscriptional ribs (Fig. 2).
Flanking the m. rectus abdominis on each side and partially covering its lateral edges is a narrow, unsegmented slip of muscle, surrounded by a distinct, tough connective tissue sheath and extending from the posterior end of the m. pectoralis to the pubic membrane (Fig. 2). This muscle is here interpreted as the m. rectus abdominis lateralis, contrary to previous reports of its absence in Iguania (summarized by Moody,1983). At the posteriormost inscriptional rib, it makes an abrupt medial turn and begins to narrow perpendicular to its long axis; at this turn there is sometimes also a posteriorly pointing chevron of melanocytes. Its anterior fibers are continuous with those of the posterolateral portion of the m. pectoralis in two of eight specimens examined and are separate from the m. pectoralis, inserting onto a connective tissue sheath and part of the first inscriptional rib, in the other four. The m. rectus abdominis lateralis is the most superficial muscle in the trunk, superficial even to the m. obliquus externus. Its connective tissue sheath is outlined by a scattering of melanocytes. This outline also corresponds to a weak attachment to the dermis, sandwiching the superficial fascia between the muscle fascia and the skin. A light line of connective tissue is visible paralleling the lateral edge about one quarter of the way to the medial edge. On the deep surface of the m. rectus abdominis lateralis (Fig. 3), this band represents the ventral terminus of the muscle fibers of the m. obliquus externus, though the aponeurosis of that muscle continues toward the ventral midline. Also, upon the deep surface, a short pigmented strip of perichondral membrane attaches each inscriptional rib to the m. rectus abdominis lateralis, carrying branches of the segmental spinal nerves to supply the muscle (as reported for the m. rectus abdominis lateralis by Maurer,1896). This attachment is about one quarter of the way to the medial edge of the muscle. The medial edge of the m. rectus abdominis lateralis is formed when the m. obliquus externus aponeurosis and possibly part of the costal/m. obliquus internus aponeurosis plunges through the muscle from its deep surface, cutting it off and emerging as the superficial contribution to the sheath of the m. rectus abdominis proper (Fig. 1b).
The crotaphytine iguanian Crotaphytus shows a similar morphology to A. carolinensis, but with less prominent melanocyte concentrations in the periphery of the m. rectus abdominis lateralis. Otherwise, the topological relations and extents of the various muscles and connective tissue sheets are comparable between these taxa.
In the corytophanine iguanian Laemanctus longipes, the m. obliquus externus is clearly divided into superficialis and profundus portions. Both extend deep to the thick, well-developed the m. rectus abdominis lateralis, where the m. obliquus externus becomes membranous and then separates the m. rectus abdominis lateralis from the m. rectus abdominis at the common contact point with the m. obliquus internus aponeurosis, forming the superficial part of the rectus sheath proper. As in A. carolinensis (Fig. 2), but to a greater extent, a slight medial flap of the m.rectus abdominis lateralis is superficial to the true rectus (Fig. 1b). The m. rectus abdominis lateralis is considerably thicker than the m. rectus abdominis proper.
In the phrynosomatine iguanian Holbrookia texana, the m. obliquus externus extends even farther medially as a muscular body deep to the m. rectus abdominis lateralis than it does in A. carolinensis. It also extends deep to the m. pectoralis posteriorly. However, the connective tissue sheath overlying the m. rectus abdominis lateralis is more continuous with the tough rectus sheath proper than in A. carolinensis, where it is strictly delineated. In Sceloporus sp. and Cophosaurus sp., many of the fibers of the m. obliquus externus seem to blend into the m. rectus abdominis lateralis. One deep slip of the m. obliquus externus plunges deep to it.
In the acrodontan iguanian L. belliana, the m. pectoralis cutaneous is present as figured by Moody (1983, Fig. 4). Moody (1983) reported that this muscle, identified by Camp (1923) as the m. rectus abdominis lateralis, is part of the m. pectoralis and therefore non-homologous to the former. He found it in many agamid lizards. However, the m. pectoralis cutaneous shares with the m. rectus abdominis lateralis its superficial position and, with some instances of the m. rectus abdominis lateralis, a continuity with the m. pectoralis. Additionally, the m. pectoralis cutaneous, like the remainder of the posterolateral portion of the m. pectoralis and like the m. rectus abdominis lateralis, is innervated by segmental spinal nerves (Moody,1983; Gray et al.,1995). As discussed later, the development of the m. pectoralis is closely linked with that of the m. rectus abdominis and the body wall muscles, and distinctions among parts of these muscles can at times be hard to make. Thus, the m. pectoralis cutaneous may represent a reduced m. rectus abdominis lateralis depending on the optimization of character states upon the squamate tree, with the latter name taking priority in the absence of developmental data, as discussed further below.
In the primitive gecko Eublepharis macularius there is no well-developed the m. rectus abdominis lateralis. The m. obliquus externus aponeurosis does not extend deep to any part of the m.rectusabdominis, though the costal/m. obliquus internus membrane still does (Fig. 1a). Rather, the m. obliquus externus aponeurosis continues directly into the sheath of the m. rectus abdominis proper, the muscle fibers abutting but not extending deep or superficial to that muscle.
A. sexlineatus, an autarchoglossan, has an extensive the m. rectus abdominis lateralis in which the two sides meet at the midline and thereby entirely obscure the m. rectus abdominis proper (Fig. 1c). In the thoracic region, the muscle fibers of the m. obliquus externus abut the lateral edge of the m. rectus abdominis lateralis while the m. obliquus externus aponeurosis extends deep to it, but more posteriorly, the fibers plunge deep to the m. rectus abdominis lateralis before the m. obliquus externus becomes membranous. Further, the m. rectus abdominis lateralis is strongly attached to the dermis at the imbricae between the rows of large, platelike ventral scales (Fig. 4). If one were to remove the skin without carefully separating the muscle from the skin, much of it would be removed with the skin (J.A. Gauthier, pers. comm.). Even when detached, many fibers remain on the skin, and the muscle itself is impressed into transverse steps by the scale rows. Certain of the cutaneous muscles of snakes may also be homologous to the m. rectus abdominis lateralis and therefore show the “autarchoglossan” condition of extensive close association with the dermis (Tsuihiji,2007 and personal communication).
The dissections described earlier demonstrate the widespread presence in Iguanidae (pleurodontan Iguania) of a muscle topologically identical to the autarchoglossan the m. rectus abdominis lateralis. The m. rectus abdominis lateralis has never before been reported in Iguania (Sanders,1872,1874; Moody,1983). I suspect that this deficit owes to the subtle nature of the relatively thin muscle in that clade and the absence of extensive contact with the skin.
The absence of the m. rectus abdominis lateralis was confirmed in the gekkotan Eublepharis macularius, and the presence of a well-developed the m. rectus abdominis lateralis with strong attachments to the dermis was confirmed in the autarchoglossan A. sexlineatus. This extensive development and association with the skin is widely reported in autarchoglossans (e.g., Camp,1923 for various lizards, though noting that those with derived cycloid scales lose the dermal association; Surahya,1989 for Varanus komodoensis).
Unique to Lizards?
An initial question that arises given the apparently broad distribution of the m. rectus abdominis lateralis in Squamata is whether a similar muscle has been reported outside of that clade. Within Amphibia, Caudata have the least modified abdominal musculature. Some descriptions indicate that the extensive the m. rectus abdominis of some salamanders is actually superficial to the m. obliquus externus (Simons and Brainerd,1999; Brainerd and Simons,2000, in a range of salamanders bracketing Caudata; Walthall and Ashley-Ross,2006, in Taricha) or at least that the profundus portion of the m. obliquus externus is deep to the m. rectus abdominis (Carrier,1993, in Dicamptodon). Others report an amniote-like condition in which the m. obliquus externus and is largely superficial to the m. rectus abdominis (Fig. 1a; Francis,1934, in Salamandra; Naylor,1978, in a range of salamanders bracketing Caudata). However, in none of these taxa does the m. obliquus externus aponeurosis then emerge to separate a m. rectus abdominis superficialis from the m. rectus abdominis proper. Although Naylor (1978) called a lateral strip of muscle “m. rectus lateralis” in some taxa, this slip is not formed by the connective tissue of the m. obliquus externus, and it is deeper, not more superficial, than the major portion of the m. rectus abdominis. The m. obliquus internus generally meets or splits around (in its aponeurotic part) the m. rectus abdominis; descriptions are sometimes unclear as to which occurs.
In frogs, based on the accounts of Rana esculenta (Ecker and Haslam,1889) and Ascaphus truei (Ritland,1955), together bracketing Anura, an amniote-like condition exists in which the m. obliquus externus is superficial to the m. rectus abdominis and the m. obliquus internus meets or splits around the m. rectus abdominis. In Gymnophiona, the same topology has been reported by Naylor and Nussbaum (1980), who also identifies a “m. rectus lateralis.” However, that muscle in caecilians is dorsal on the body wall, and deep to the m. obliquus externus, sometimes having a close association with the m. obliquus internus.
Among mammals, in general, the condition described for humans (Fig. 1a), with a few variations, occurs. Monotremes and primitive marsupials have a mediolaterally narrow the m. rectus abdominis deep to the m. obliquus externus (Reilly and White,2003). Eutheria and Metatheria differ slightly in whether the aponeurosis of the m. obliquus internus splits around the m. rectus abdominis or contributes only to the deep layer of the rectus sheath, and these relations change at the arcuate line of the rectus sheath (Rizk,1980; Lancaster and Henson,1995).
Finally, among non-squamate reptiles, the myology of Sphenodon punctatus, sister taxon to Squamata, has been described several times (Maurer,1896; Osawa,1898; Byerly,1925). However, the relations of the body wall and abdominal muscles have never been adequately described or figured. The existing descriptions appear to indicate that there is no m. rectus abdominis lateralis and that the m. obliquus externus is superficial to the m. rectus abdominis proper, in which are embedded the gastralia. Turtles have a highly modified body wall and abdominal muscle complex, but no structure that particularly resembles a m. rectus abdominis lateralis (Bojanus,1819; Gaunt and Gans,1969). However, works on turtle myology, especially of the trunk, are scarce. Birds are also modified and have reduced abdominal musculature related to their extremely shortened and stiffened trunks (George and Berger,1966). Finally, in crocodylians, the major part of the m. rectus abdominis is covered superficially by the aponeurotic part of the m. obliquus externus (Romer,1923). However, at the posterior end of the trunk lies a superficial slip of muscle called the m. trunco-caudalis by Maurer (1896) but also suggested to be the m. rectus abdominis lateralis (Gadow,1887; Romer,1923). This muscle is superficial to the m. obliquus externus and extends posteriorly to attach to the posterior edge of the pubis and the m. ischio-caudalis. It is unclear whether the m. trunco-caudalis is part of the body-wall or the appendicular muscle system.
A topological correspondence among the muscles I have called the m. rectus abdominis lateralis is evident from the results of the dissections. However, it is not entirely clear whether these muscles have a special relationship to the m. rectus abdominis as the name would imply, that is, whether they differentiated from a single primordium the other part of which became only the m. rectus abdominis, since muscle primordia generally follow a pattern of progressive division (Maurer,1898; Rowe,1986). The m. rectus abdominis lateralis in squamates is sometimes continuous, in part, with the posterolateral portion of the m. pectoralis (Fig. 2). The possible homology of the “m. pectoralis cutaneous” in agamids with the m. rectus abdominis lateralis was mentioned earlier (Camp,1923; Moody,1983). In both amphibians and reptiles, a m. pectoralis abdominis is occasionally described that extends posteriorly for a considerable distance (Camp,1923; Manzano and Perotti,1999), and the posterior part of the m. pectoralis is supplied by segmental spinal nerves as is the m. rectus abdominis lateralis where present (Ashley,1952; Gray et al.,1995). Additionally, in salamanders (Naylor,1978), the m. rectus abdominis proper is sometimes continuous with the m. obliquus layers and in at least some frogs it is continuous with part of the abdominal portion of the m. pectoralis (Ecker and Haslam,1889). The dilineations among the m. rectus abdominis, the m. rectus abdominis lateralis, the m. pectoralis, and even some of the body wall musculature therefore are not as strict across tetrapods as they are in humans.
Developmentally, recent and groundbreaking work (Burke and Nowicki,2003; Nowicki et al.,2003) has shown that the ribs and most of the intercostal musculature (as well as the deep dorsal musculature) belong to a “primaxial” domain in which both musculature and connective tissue are derived from the somites. The body wall muscles, the abdominal muscles, and most of the appendicular muscles including the m. pectoralis are derived instead from an “abaxial” domain in which myocytes still originate from somites but connective tissue is contributed by the lateral plate mesoderm. It is in this lateral-plate-derived connective tissue that partitions among muscles are formed as the myocytes migrate down from the dermomyotomal parts of the somites. This migration occurs gradually and bilaterally, and the m. rectus abdominis as well as the m. rectus abdominis lateralis where present arise from masses at the ventral edges of the descending masses and later differentiate further (Maurer,1898; Bardeen and Lewis,1901; Straus and Rawles,1953; Parry,1968; Christ et al.,1983; Lynch,1984; Manzano and Perotti,1999). The downgrowth is associated in part with expression of Pax3 and Pax7 in the descending lateral lip of the dermomyotomes according to studies in the chicken Gallus gallus (Tremblay et al.,1998; Otto et al.,2006). According to Maurer (1898), the m. rectus abdominis lateralis in Lacerta differentiates from a common primordium with the m. rectus abdominis, but he did not clearly show the time-sequence of muscle differentiation within the body wall, nor the early formation of the more anterior m. pectoralis.
Given the intimate developmental association of all abaxial muscles and their late differentiation from each other as laminae of muscle and connective tissue, it is difficult to determine whether the m. rectus abdominis lateralis is better conceived as associated with the m. rectus abdominis, the m. pectoralis, the body wall muscles, some combination thereof, or none of these—rather being a neomorphic layer of the abaxial system. Further and more detailed comparative histological and developmental study is required to dissect the ontogeny of the complex and highly integrated abaxial system.
Because further study is required to determine whether the m. rectus abdominis lateralis is closely associated with any of the ancestral amniote abaxial muscles, I suggest retaining this name for the time being for reasons of priority and historical continuity. The data presented here are evidence for its broad homology and distribution within Squamata. The precise hypothesis of homology differs according to the phylogenetic hypothesis used in the analysis as discussed in the next section. However, structural identity is undoubted in that the m. rectus abdominis lateralis, regardless of potential name changes, is situated superficial to the m. obliquus externus and either flanking or superficially covering the m. rectus abdominis proper.
Phylogeny and Function
Within Squamata, the presence of the m. rectus abdominis lateralis optimizes as ancestral for Iguanidae (in the sense of pleurodont Iguania), no matter which of the varied hypotheses of the relationships within that clade is used (Etheridge and de Queiroz,1988; Frost et al.,2001; Conrad,2008). Within Squamata, optimization differs depending on whether the given phylogenetic hypothesis is a more traditional gross-morphology-based topology with Iguania as sister to the remaining squamates (Fig. 5a, e.g., Estes et al.,1988; Conrad,2008) or the topology suggested by recent molecular studies, in which Iguania is sister to Anguimorpha and Gekkota is sister to the remaining squamates (Fig. 5b; Townsend et al., 2004). If the m. pectoralis cutaneous in Acrodonta is not taken as a form of the m. rectus abdominis lateralis, then, under the morphological phylogeny, the m. rectus abdominis lateralis has appeared twice, once along the stem of Iguanidae and again along the stem of Autarchoglossa (Fig. 5a). Under the molecular phylogeny, it optimizes as having appeared only once, along the internode between Gekkota and the remaining squamates, and subsequently lost along the stem of Acrodonta (Fig. 5b). If acrodontans do show remnants of the m. rectus abdominis lateralis, then even under the morphological hypothesis, the optimization becomes ambiguous. Given a single origin, the morphological hypothesis requires elaboration and intimate association with the dermis to have appeared along the stem of Autarchoglossa, whereas the molecular hypothesis requires reduction and dissociation of the muscle from the dermis along the stem of Iguania (Fig. 5b). The presence of the m. rectus abdominis lateralis in the common ancestor of all squamates save Gekkota would represent rare gross-anatomical support for the molecular-structure-based hypothesis of squamate relationships, in which Gekkota is sister to the remaining taxa (Townsend et al., 2008).
The entire optimization would change if Sphenodon were shown to have a the m. rectus abdominis lateralis or if the m. trunco-caudalis in Crocodylia were to be scored as the m. rectus abdominis lateralis. Some topologies would then suggest that the muscle is a synapomorphy appearing along the stem of Lepidosauria or the stem of Diapsida.
The considerable mass and high degree of differentiation of the squamate m. rectus abdominis lateralis suggests a functional role. In particular, the extensive development and close association with the large belly scales in autarchoglossan squamates (also observed by Camp,1923) suggests that the m. rectus abdominis lateralis and the rows of large belly scales are a developmental and/or functional module. As with many, perhaps most, aspects of the body wall and abdominal muscle complex in vertebrates, the function of the m. rectus abdominis lateralis, now shown to be widely distributed in lizards, remains unexplored.
My primary sounding board for new projects is, and likely always will be, my first advisor, Jacques Gauthier. He encouraged my work and generously shared his observations on the m. rectus abdominis lateralis in Autarchoglossa and the way in which it tends to come off with the skin. Tim Rowe was an additional source of inspiration and, in particular, instruction on basic techniques of dissection. Takanobu Tsuihiji, perhaps the most gifted living comparative myologist, provided much-needed expert advice. He was also so kind as to accidentally leave behind a jar of Anoliscarolinensis in the back of a cabinet at our old Yale lab (Jacques' lab), which I duly found when rummaging around on a return visit. This jar initiated my attempts to dissect Anolis. Annie Burke, Shigeru Kuratani, Anthony Herrel, Farish Jenkins, Arkhat Abzhanov, and Karel Liem indulged me with further discussion regarding the minutiae of muscle anatomy and development. Krister Smith, Gabe Bever, and Chris Bell provided useful commentary on the presentation of the data herein. Anthony Herrel and one anonymous reviewer made perceptive and entirely helpful suggestions that allowed considerable refinement of the article. Travis LaDuc and David Cannatella (Catfish) allowed me essentially free reign in picking out specimens to dissect from the Texas Natural History Collections, an important and progressive philosophy regarding collections use for anatomical purposes for which I am exceedingly thankful. This work was supported by a Donald D. Harrington Graduate Fellowship from The University of Texas at Austin, a National Science Foundation Graduate Research Fellowship, and a James Mills Peirce Fellowship from Harvard University. The initial presentation of these data took place amidst the wonders of Paris, memories of which have inspired me ever since.