• We are sorry, but NCBI web applications do not support your browser and may not function properly. More information
Logo of janatLink to Publisher's site
J Anat. Nov 2004; 205(5): 335–347.
PMCID: PMC1571356

Evolution of the vertebrate jaw: comparative embryology and molecular developmental biology reveal the factors behind evolutionary novelty


It is generally believed that the jaw arose through the simple transformation of an ancestral rostral gill arch. The gnathostome jaw differentiates from Hox-free crest cells in the mandibular arch, and this is also apparent in the lamprey. The basic Hox code, including the Hox-free default state in the mandibular arch, may have been present in the common ancestor, and jaw patterning appears to have been secondarily constructed in the gnathostomes. The distribution of the cephalic neural crest cells is similar in the early pharyngula of gnathostomes and lampreys, but different cell subsets form the oral apparatus in each group through epithelial–mesenchymal interactions: and this heterotopy is likely to have been an important evolutionary change that permitted jaw differentiation. This theory implies that the premandibular crest cells differentiate into the upper lip, or the dorsal subdivision of the oral apparatus in the lamprey, whereas the equivalent cell population forms the trabecula of the skull base in gnathostomes. Because the gnathostome oral apparatus is derived exclusively from the mandibular arch, the concepts ‘oral’ and ‘mandibular’ must be dissociated. The ‘lamprey trabecula’ develops from mandibular mesoderm, and is not homologous with the gnathostome trabecula, which develops from premandibular crest cells. Thus the jaw evolved as an evolutionary novelty through tissue rearrangements and topographical changes in tissue interactions.

Keywords: evolutionary novelty, Hox genes, jaw, lamprey, neural crest, pharynx

Introduction: the mandibular arch and the jaw

The head of the vertebrate embryo is characterized by the possession of neural crest-derived ectomesenchyme and the pharyngeal arches (PAs), which are primarily equivalent to the gill arches. The skeletal elements in the PAs are derived exclusively from the ectomesenchyme, not from the mesoderm (Le Lièvre, 1974, 1978; reviewed by Le Douarin, 1982; Gans & Northcutt, 1983; Hall & Hörstadius, 1988; Noden, 1988; Hall, 1999; Le Douarin & Kalcheim, 1999; Kimmel et al. 2001; Morriss-Kay, 2001; Trainor et al. 2003). The jaw in gnathostomes (jawed vertebrates) is one of the earliest innovations in the evolution of vertebrates and is derived from the mandibular arch (MA). Evolution of the jaw therefore can be viewed as the establishment of a developmental programme for the ectomesenchyme of the MA to form a dorsoventrally articulated pattern, consisting of upper and lower jaws. However, the evolutionary scenario of the jaw, or the history of changes in the developmental programmes to create the jaws, remains largely unknown. The lamprey, a jawless vertebrate, is thought to represent the outgroup to the jawed vertebrates and may suggest the ancestral developmental programmes shared by the common ancestor, as well as the changes introduced to form the jaw in gnathostome lineages.

According to the classic morphological concept, the jaw in gnathostomes is assumed to have arisen by transforming one of the rostral gill arches of the ancestral vertebrate (reviewed by Sewertzoff, 1911, 1928; Goodrich, 1930; Gregory, 1933; de Beer, 1937; Romer, 1966; Moy-Thomas & Miles, 1971; Romer & Parsons, 1977; Jarvik, 1980; Mallatt, 1984, 1996; Carroll, 1988; Janvier, 1996; Kuratani et al. 2001). However, the fossil record has not revealed any ancestral animals with an undifferentiated series of gill arches in their pharynx. Moreover, in all the gnathostome embryos observed so far, PA1 and PA2 can be recognized as modified from the rest of the arches (branchial arches), and specifically called the mandibular (MA) and hyoid arches (HA), respectively (Fig. 1; and see Gregory, 1933; Edgeworth, 1935; de Beer, 1937; Romer, 1966; Jarvik, 1980). This is also true for the lamprey, a modern agnathan (jawless) vertebrate (Kuratani et al. 2001). In this animal, the MA differentiates into the velum, the pumping apparatus that lets water into the pharynx, as well as the lower lip, which resembles the gnathostome lower jaw (Fig. 2; Mallatt, 1996; Kuratani et al. 2001; Shigetani et al. 2002; see below). An ancestral animal with simple gill arches with no mandibular or hyoid identities is purely hypothetical. Recent embryological and molecular developmental analyses of lampreys, the living agnathans, have suggested instead a more complicated scenario for the evolution of the gnathostome jaw.

Fig. 1
Visceral skeletal systems in various gnathostomes. Visceral skeletons of Chimaera monstrosa (A: Holocephali), Callorhynchus (B: Holocephali), Chlamidoselachus (C: Elasmobranchii), Acipenser sturio (D: Chondrostei), Gasterosteus aculeatus (E: Teleostei) ...
Fig. 2
Pharyngeal anatomy of the ammocoete larva of the lamprey. The pharynx of the lamprey larva has been cut horizontally and its dorsal half is illustrated from the ventral view. Mandibular, hyoid (HA), and two branchial arches (Ba1–2) are shown. ...

Mandibular arch and the Hox-free default state

The idea that the jaw is a transformed PA fits the developmental sequence of the gnathostome embryo better than the actual fossil record. A specific class of homeobox-containing genes, called Hox genes, are expressed sequentially along the anteroposterior axis of the embryonic pharynx, thereby constituting a nested pattern of gene expression, or the ‘Hox code’ in the ectomesenchyme (Fig. 3A; Hunt et al. 1991a,b). Hox genes in amniotes are arranged tandemly in four clusters, each of which is found on a different chromosome (reviewed by McGinnis & Krumlauf, 1992). There is a tendency called ‘spatial collinearity’ in that the genes located in the 3′ direction of a cluster are more likely to be up-regulated in the anterior part of the embryo, whereas the more 5′ genes are transcribed towards the posterior part of the embryo. Thus, each one of the PAs carries a different and specific subset of Hox transcripts that determines its specific developmental pathway (Fig. 3; Hunt et al. 1991a, b). Hox genes, encoding transcription factors, play developmental roles as the ‘homeotic selector’ genes by providing positional cues to the ectomesenchyme filling each PA. It is important to note that there are no Hox genes expressed in the MA (Fig. 3A).

Fig. 3
Evolution of the gene expression patterns and the origin of the jaw. (A) Simplified Hox codes in gnathostomes (top) and the lamprey larva (bottom) are summarized. Gnathostome-specific regulatory gene expression domains are shown in bold, and crest cells ...

Part of the jaw-patterning programme in gnathostomes appears to be regulated by the ‘Hox-free default’ state of the MA, as has been shown by a number of experiments. First, gain- and loss-of-function experiments on Hoxa-2, a Hox gene expressed in the HA and posterior to it, result in the transformation of the MA into the HA, and that of the HA into the MA, respectively (Gendron-Maguire et al. 1993; Rijli et al. 1993; Grammatopoulos et al. 2000; Pasqualetti et al. 2000). This is consistent with the ‘law of posterior prevalence’ (Lufkin et al. 1992), in which loss of function of a Hox gene leads to anteriorization, whereas gain-of-function leads to the posteriorization of morphological identities. Second, transplantation of the Hox-free neural crest into the rhombomere 4 level of the hindbrain, which is destined to become the HA, leads to the duplication of MA skeletal elements in the HA domain (Noden, 1983; Couly et al. 1998). The growth factor FGF8 (fibroblast growth factor 8), released from the embryonic mid-hindbrain boundary, possibly inhibits expression of Hox genes in the rostral hindbrain crest (Trainor et al. 2002). Thus, in the gnathostome developmental process, the differentiation of the jaw from the rostral crest cells is permitted by the absence of the Hox transcripts from the crest cells in the MA. The question then arises as to whether the Hox-free state of the MA was established at the outset of gnathostome evolution, or if it was already present in agnathans, or even in cephalochordates (e.g. amphioxus).

In the above context, Cohn (2002) reported in a preliminary study on a lamprey species, Lampetra fluviatilis, that one of the Hox genes, HoxL6, was developmentally up-regulated throughout the PAs, implying that the presence of Hox transcripts in the agnathan MA inhibited the differentiation of the jaw in this animal group. There is no doubt that an MA can be identified morphologically in these animals, even if they lack jaws (de Beer, 1937; reviewed by Kuratani et al. 2001). By contrast, recent analyses by Takio et al. (2004) did not confirm this scenario: 11 Hox genes were isolated from a species, Lethenteron japonicum, including the orthologue of HoxL6, but none of the genes was expressed in the embryonic MA. This is consistent with the finding that LjFgf8/17 (the lamprey cognate for Fgf8) is expressed in the mid-hindbrain boundary, as in gnathostome embryos (Fig. 3; Murakami et al. 2001; Shigetani et al. 2002), and with the generally accepted notion for Hox code evolution in chordates (Schilling & Knight, 2001, and references therein). Moreover, Hox2 was clearly expressed in the crest cells of the HA and in the more posterior PAs, and similarly, Hox3 was expressed in the crest cells of the PA3 and more posterior regions, reminiscent of the nested, collinear pattern of the gnathostome Hox code (Fig. 3). Although the difference of Hox6 expression between Lampetra fluviatilis and Lethenteron japonicum maybe due to a species- or genus-specific difference in the regulatory mechanism for Hox6, at the very least it is conceivable that some agnathans and gnathostomes share the same basic Hox code (Hox-free default state of PA1; PG2 and PG3 genes expressed in PA1 and PA2, respectively). Importantly, almost all the vertebrate species possess more or less differentiated MAs (jaws in gnathostomes; velum and lower lip in the lamprey) and HAs, followed by respiratory branchial arches that are similar across species (Figs 1 and and2).2). It seems most likely that this type of ‘primitive’ Hox code was already established in the common ancestor of the lamprey and gnathostomes – with differentiated PA1 and PA2 – with distinctive identities as opposed to the morphologically identical, more posterior PAs (Fig. 3). In this connection, it is interesting to note that the Cambrian fossil animal Haikouella appeared to have possessed an oral apparatus that resembled that of the lamprey ammocoete larva (Mallatt & Chen, 2003; Mallatt et al. 2003; but also see Shu et al. 2003): well-differentiated oral apparatus, which would have been at least in part differentiated from the MA in this animal, is consistent with the deep origin of the Hox-free default MA.

This scenario further implies that evolution of the developmental programme that forms the jaw involved changes in molecular mechanisms downstream of the shared Hox code of ancestral vertebrates, again consistent with the idea that the MA is morphologically equivalent as a developmental unit between the lamprey and gnathostomes by having the same topographical relationships with other embryonic structures, as well as with the same positional value defined by the absence of Hox transcripts (Figs 3 and and5;5; see also Kuratani et al. 2001, for morphological value of the vertebrate MA).

Fig. 5
Comparisons of oral patterning. Distribution patterns of cephalic crest (A), expression of Fgf8/17 and Bmp2/4 cognates in the ectoderm (B), ventral view of embryonic oral regions (C), and the medial sagittal sections (D) are schematically represented. ...

Dlx and Otx genes

The shared developmental features identified between the lamprey and gnathostomes are most likely to represent the ancestral programme possessed by the common ancestor (Fig. 3B; also see Trainor et al. 2003, for a similar method of speculation), whereas there can also be programmes that have arisen or lost only in one of the lineages. It is mportant to realize that molecular phylogenetic analyses of regulatory gene cognates often indicate the polarity of changes along the phylogenetic tree, making the evolutionary scenario a more plausible one (reviewed by Kuratani et al. 2002). For a possible example of the gnathostome-specific derived feature (gnathostome synapomorphy), it is worth noting that gnathostome MA is also dorsoventrally patterned through the nested expression of Dlx genes along the dorsoventral axis of the arch (Depew et al. 2002). Thus, simultaneous disruption of Dlx5 and Dlx6 expressed in the ventral half of the MA leads to the mirror-image duplication of the upper jaw segment in the domain of the lower jaw. Preliminary data on the lamprey Dlx genes and their expression patterns, and the dorsoventrally symmetrical morphology of the lamprey branchial cartilages, indicate that the Dlx code for dorsoventral polarity is not a pan-vertebrate developmental feature: expression patterns of Dlx cognates are not localized in the MA, as seen in mouse embryos (Myojin et al. 2001; Neidert et al. 2001; also see Kuratani et al. 2002; Fig. 3).

Like the Hox code in vertebrates, expression of Otx cognates is highly conserved between gnathostomes and the lamprey, both in the neural tube and in the MA (Ueki et al. 1998; Tomsa & Langeland, 1999; Murakami et al. 2001). In the mouse embryo, Otx2-expression and Hox-free default states appear to pattern the distal and proximal portions of the MA in a complementary fashion (Rijli et al. 1993; Gendron-Maguire et al. 1993; Matsuo et al. 1995; also see Kuratani et al. 1997, and 2001, 2002, for reviews). Such a division appears to be partly due to the migration and distribution patterns of the crest cells in MA, the rostral part of which preferentially receives cells from the Otx2-positive midbrain and segregates from the Hox-free crest originated from the rostral hindbrain (Osumi-Yamashita et al. 1994; see also Köntges & Lumsden, 1996, for chick development). It will be intriguing to determine if a similar complementary pattern exists in the lamprey MA during development. In this context, based on vital-dye labelling studies, the origin and migration patterns of crest cells in the MA are not identical between the lamprey and amniotes, as discussed below (Shigetani et al. 2002; McCauley & Bronner-Fraser, 2003).

The gnathostome jaw now seems to have arisen as one of several variations in differentiation programmes at the radiation of the ancestral vertebrates, based on the shared ground plan of craniofacial patterning with the shared basic expression patterns of some homeobox genes. However, the variation in gnathostome patterning also seems to involve a change in global interactions between mesenchyme and epithelium, as seen in the various types of craniofacial designs found in the Palaeozoic fossils (Janvier, 1996).

Oral apparatus and the mandibular arch: heterotopy and loss of homology

According to Wagner & Müller (2002), an ‘evolutionary novelty’ can be defined as a new structure that arises by overriding ancestral developmental constraints, so that morphological homology is lost between the novel and ancestral structures. Thus, the gnathostome jaw could be counted as an evolutionary novelty, as discussed below. In the following discussion, we have to bear in mind that the term ‘mandibular arch’ universally refers to an identical developmental unit among vertebrates (morphologically homologous throughout vertebrates), whereas the ‘oral apparatus’ or ‘oral region’ may differentiate from different regions of the embryonic head in each animal group.

Although the classical transformation theory of the jaw predicts the initially identical, undifferentiated pharyngeal arches, the cephalic crest cells (ectomesenchyme) never simply form single divided cell streams each filling a single PA. Instead, in all the vertebrate embryos observed, there are three distinct crest cell populations, and the most rostral one not only populates the MA, but also expands rostrally to the entire pharynx (Noden, 1988; Osumi-Yamashita et al. 1994; Kuratani, 1997; Graham, 2001; Kuratani et al. 2001; Graham et al. 2004; reviewed by Horigome et al. 1999, and Kuratani et al. 2001; Figs 4 and and55 top). At the early pharyngula stage, when cephalic crest cells cease emigration, the distribution pattern of the crest cells is very similar between gnathostome and the lamprey – the cephalic crest cells form three major streams of migration, and the most rostral cell population (trigeminal crest cells) is distributed in the MA as well as in the premandibular (PM) region (Fig. 5A; Horigome et al. 1999; Shigetani et al. 2000). Here the term ‘premandibular’ does not imply the presence of a ‘premandibular arch’ once assumed by comparative morphologists, but simply indicates the position rostral to the MA. [In this review, the premandibular crest cells are defined as those crest cells that surround the premandibular mesoderm, as opposed to the MA crest cells that surround the mandibular mesoderm. Owing to the absence of the pharyngeal pouch that normally limits the MA rostrally, only the mesodermal component can be used as a landmark. This terminology inevitably relates to the classical concept of head segmentation in vertebrates, which used to be based primarily on the segments in the head mesoderm (see de Beer, 1937; Jarvik, 1980; Jefferies, 1986). However, as has been discussed previously (Kuratani et al. 1999), neither lamprey nor gnathostome embryos show any sign of segmentation in the cephalic mesoderm, except for the late-forming premandibular mesoderm that is more or less separated from the rest of the cell population, and the lateral part of the head mesoderm, which is partially and secondarily divided by the growth of pharyngeal pouches. Thus, primary mesodermal segmentation is not assumed in the vertebrate head in this review. For more discussion on head segmentation and metamerism, see Kuratani (2003). The term ‘trigeminal crest cells’ stems from the fact that the distribution of these cells corresponds to the peripheral distribution patterns of the trigeminal nerve in the later embryo (see Kuratani, 1997; Kuratani et al. 2001; Fig. 6). The crest cells in the PM region are further subdivided into preoptic and postoptic cell populations (Figs 4, ,5A5A and and66).

Fig. 4
Comparison of mesenchymal developmental patterning. In the lamprey and gnathostome embryonic heads, shared patterns of mesodermal (yellow) and ectomesenchymal (green) distribution can be detected. The endodermal derivatives are coloured pink. The crest ...
Fig. 6
Comparisons of equivalent ectomesenchymal regions between the lamprey and gnathostomes. Terminology for the crest cell populations and subpopulations is based on the topographical distribution of the crest cells with respect to the other embryonic structures ...

Importantly, at an early stage of pharyngular development, there is no clear boundary between the PM and MA regions (Kuratani et al. 1999; Shigetani et al. 2000, 2002), and the oral apparatus is formed in different ways between gnathostomes and the lamprey (it is composed of upper and lower jaws in the gnathostomes; of upper and lower lips in the ammocoete larva of the lamprey). Each domain of the trigeminal ectomesenchyme can be identified by its topographical relationships with the mesodermal components (reviewed by Kuratani et al. 2001; Figs 46). The MA crest cells surround the mandibular mesoderm and the postoptic part of the PM crest cells surrounds the premandibular mesoderm (primordia of the extrinsic eye muscles in gnathostomes; Koltzoff, 1901; see Boorman & Shimeld, 2002, for a case of specific gene expression in the premandibular mesoderm of lamprey embryos; and see Kuratani et al. 1999, for comparative embryology of this mesoderm). Interestingly, both the upper and the lower jaws develop from the MA crest cells, whereas in the lamprey the MA differentiates only into the velum and the lower lip, and the upper lip is derived from PM crest cells (Kuratani et al. 2001; Shigetani et al. 2002). Based on the basic architecture of the embryo therefore the term ‘mandibular arch’ should be reserved for the specific PA found in the same relative position in the embryo, as a developmental unit that is specifically and topographically associated with the mandibular mesoderm, irrespective of its fate in later developmental stages, which can differ in each animal lineage. Thus, in gnathostomes, the oral apparatus called the ‘jaw’ develops from the MA, whereas the lamprey oral region also requires the PM region as an embryonic material (Figs 46). The term ‘oral apparatus’, on the other hand, implies only a functional resemblance of structures, which can be derived from varied sets of embryonic tissues. It is important to note that a gene expression pattern is not always associated with a homologous set of cell populations, as is discussed below.

In the amniote MA, BMP4 in the distal ectoderm induces the expression of the target gene, Msx1, in the underlying ectomesenchyme, and the proximally distributed FGF8 induces Dlx1 expression in the proximal ectomesenchyme (Barlow & Francis-West, 1997; Neubüser et al. 1997; Tucker et al. 1998; Shigetani et al. 2000). Shigetani et al. (2002) have also found that molecules involved in the proximodistal patterning of the gnathostome jaw are apparently used in similar patterning of the upper and lower lips of the lamprey larva, as if lips and jaws were homologous to each other (Fig. 5B). Namely, LjFgf8/17 (the lamprey cognate of Fgf8) and LjDlx1/6 (the lamprey cognate of Dlx1) are expressed widely in the oral ectoderm and mesenchyme, respectively. Moreover, LjBmp2/4a (the lamprey cognate of Bmp4) and LjMsxA (the lamprey cognate of Msx1) are expressed in the tips of lips, respectively (Fig. 5B; also see Shigetani et al. 2002). As the functions of ectomesenchymally expressed homeobox genes are most prominent in the MA-derivatives in gnathostomes (Satokata & Maas, 1994; Martin et al. 1995; Qiu et al. 1995, 1997; Yamada et al. 1995; reviewed by Hall, 1998), and apparently masked by the Hox gene expression in HA and posterior PAs (Mallo & Gridley, 1996), the Hox-free default state of the lamprey MA is again consistent with the apparently similar functions of these gene cognates in the lamprey (proximodistal specification in oral patterning). In the above comparison, homologous sets of genes are not expressed in homologous embryonic materials, but rather are associated with functionally similar structures, namely the oral apparatus (see below for a similar discussion on adenohypophysial differentiation). To resume the agnathan mode of Dlx1 gene expression, which denotes the oral and pharyngeal region in the ectomesenchyme, the domain of the upstream factor, FGF8, has to be expanded in the gnathostome embryo. Actually, implantation of an FGF8-soaked bead into the PM region mimics the lamprey-type Dlx1 expression in the chick embryo (Shigetani et al. 2000, 2002).

These experimental results imply that epithelial–mesenchymal interactions have been topographically shifted in the transition from the lamprey-like agnathan to the gnathostome states, based on a shared pattern of embryonic tissues. As a result, morphological homology was apparently lost between the lamprey and gnathostome oral apparatus; expression patterns of orthologous genes are not associated with morphologically equivalent cell populations. This evolutionary implication is curious: even if these genes always functioned in defining the oral apparatus, their regulation does not seem to be restricted to the same (homologous) embryonic component during this transition (see Manzanares & Nieto, 2003, for a similar discussion on gene usage). Such a situation simultaneously implies that the invention of the jaw deserves to be understood in the context of evolutionary novelty as defined by Wagner & Müller (2002): because the newly acquired pattern is not homologous with the ancestral pattern, the former was brought about by overriding ancestral developmental constraints, not simply modifying it for adaptation.

Alternatively, it is also possible that lips and jaws are homologous, derived from homologous cell populations with homologous gene expressions. In this case, however, morphological identities of crest cell populations cannot rely on the mesodermal components that are shared in vertebrates (Kuratani et al. 1999). Furthermore, the developmental nature of the lamprey trabecula, the premandibular cartilage, does present a conundrum and cannot be explained using this consideration, as will be discussed below.

Trabecula cranii and hypophysis

The above heterotopic scenario of jaw evolution leads us to question the homology of the so-called ‘trabecular cartilage’ reported in various vertebrate embryos and larvae. In many, the trabecular cartilage has been illustrated to arise as a pair of rod-like primordia, rostral to the MA domain, and for this mode of development, this cartilage has often been equated with the pharyngeal arch skeleton as a remnant of the premandibular arch (reviewed by de Beer, 1931, 1937; Kuratani et al. 1997, 2001). However, there is no clear embryological or palaeontological evidence to support the presence of the premandibular arch as a basic component of the vertebrate head. Regardless, in gnathostome development, this cartilage appears most likely to develop from the PM ectomesenchyme (Le Lièvre & Le Douarin, 1975; Couly et al. 1993), and is then incorporated into the rostral part of the cranial base. As opposed to the more caudal part of the neurocranium, which is derived from the mesoderm and requires the presence of the notochord to chondrify, the trabecula derivatives and some associated cartilages similarly derived from the crest are called the ‘prechordal cranium’ (Couly et al. 1993). If this cartilage is derived from the PM crest cells in gnathostomes, can it also develop in the lamprey, which uses the PM crest cells to differentiate into the upper lip? A pair of rod-like cartilages actually develops beneath the brain in the chondrocranial base of the lamprey, and they are termed ‘trabeculae’ (Johnels, 1948; see McBurney & Wright, 1996, for its histogenesis).

By injecting vital dyes into the mandibular mesoderm in the young lamprey embryo, it has been shown that both the trigeminal nerve-innervated muscles and the trabecular cartilage primordium were labelled (Kuratani et al. 2004). The latter was seen as a strand of mesenchymal condensation lateral to the notochord. As noted above this is a site that is more suitable for the mesodermally derived neurocranium, which does require the notochord (Couly et al. 1993). Moreover, at the initial stage of its development, the trabecula has been found at the level of the MA (dorsal to the first aortic arch) by Johnels (1948). Therefore, it seems that the so-called trabecula in the lamprey might represent a rostrally elongated parachordal cartilage, not a crest-derived prechordal skeletal element (see Fig. 6), although it cannot be ruled out that the rostral part of the lamprey trabecula may receive contributions from the neural crest, or there would be a cryptic boundary in the rostral part of the lamprey trabecula, delineating the mesodermally derived and crest-derived parts as seen in the gnathostome neurocranium. Moreover, several studies have so far alluded to the contribution of the neural crest to the lamprey trabecula (Newth, 1956; Langille & Hall, 1988; but also see Newth, 1951). Importantly, although the detailed mapping of the lamprey cranium is still incomplete, the mesodermal contribution to the lamprey trabecula is consistent with the heterotopy theory of jaw evolution.

The lamprey-specific use of PM crest cells is made possible by a unique patterning of the nasohypophysial placode. In the gnathostomes, nasal placodes are patterned as a pair of structures beneath the forebrain, rostral to the Rathke's pouch anlage (which differentiates as a part of the oral ectoderm). Such placodal morphology allows premandibular crest cells of gnathostomes to invade rostrally in the cranial base to form the prechordal cranium (Fig. 5C, right). In the lamprey, by contrast, nasal and hypophysial placodes initially form a single ectodermal plate rostral to the oral ectoderm (Fig. 5C; for embryology see Gorbman & Tamarin, 1985; Kuratani et al. 2001; Uchida et al. 2003). Thus the premandibular crest cells in the lamprey cannot grow rostrally to form a median septum in the cranial base as seen in the gnathostomes; instead the upper lip primordia arise behind this hypophysial plate and grow beneath the plate to form the floor of the nostril, or the nasohypophysial duct (Fig. 5A,C; see Kuratani et al. 2001). The hagfish, which follows a similar developmental pattern of placodes, has an equivalent duct that opens into the pharynx, unlike the lamprey (Gorbman, 1983; Gorbman & Tamarin, 1986; Janvier, 1996). It is important to realize that the hypophysis arises through interaction between the ectoderm and the ventral diencephalon, or the hypothalamic anlage. Again, the topographical relationships of tissues form a central problem. With respect to the mouth openings (oral ectoderm) and nasal placodes, the hypophysial primordia arise in non-equivalent topographies between the lamprey and gnathostomes – again the morphological homology is lost in a strict sense. Still, the primordia have to come into contact with the same inducer, or the hypothalamic anlage, to differentiate as the hypophysis in both animal groups. A close relationship between the oral ectoderm and hypothalamic anlage is established early in amniote development (Couly & Le Douarin, 1985; reviewed by Uchida et al. 2003). This relationship does not greatly change through later development. In the lamprey, by contrast, the hypophysial placode has to grow towards the hypothalamus secondarily in the late embryonic period. No similar pattern of tissue growth appears in gnathostome development.

Uchida et al. (2003) examined regulatory gene expression patterns and speculated on the developmental patterning of the hypophysis in the lamprey. For example, Pitx genes are known to specify the rostral ectoderm during early gnathostome embryogenesis, and play essential roles in development of the hypophysis (Szeto et al. 1999). In the lamprey, the Pitx and Pax6 cognates are also expressed in rostral ectoderm, and the expression domain becomes divided anteroposteriorly into two parts, the nasohypophysial plate and the oral ectoderm, by the secondary growth of the upper lip primordia. Similarly, TTF-1, a marker gene for the gnathostome hypothalamus, is also expressed in an equivalent portion of the brain anlage in the lamprey (Fig. 5D; Murakami et al. 2001; Ogasawara et al. 2001; Uchida et al. 2003). Of the genes that have been examined thus far, the expression of transcription factors, which act in a cell-autonomous manner, is associated with the equivalent cell type or specific structure in both lamprey and gnathostomes. In contrast, the expression patterns of genes encoding non cell-autonomously functioning signalling molecules, such as growth factors, are not comparable between the lamprey and gnathostomes (Uchida et al. 2003). Unlike the mouse, neither the Fgf8/17 or the Bmp2/4 cognates are expressed in the lamprey hypothalamus, whereas the Fgf8/17 cognate is expressed in the hypophysial placode of the lamprey.

As already seen in the relatively expanded expression domain of Fgf8 (LjFgf8/17) in the lamprey, changes in the distribution patterns of signalling molecules in the two animal groups may explain the different topography and behaviour of embryonic tissues: it is conceivable that these growth factor-encoding genes have to be regulated differently in those embryos with non-comparable topography to realize identical tissue interactions. Morphological homology may be lost during evolutionary changes of oral patterning, but the cell-autonomous developmental roles of regulatory genes tend to be conserved and are thus always attached to specific cell types or structures, such as Msx cognates always expressed in the tips of oral fringes (lips and jaws). It may rather be the changes in regulation of signalling molecule-encoding genes that form the base for heterotopy.

In conclusion, comparative embryology and molecular developmental biology of the lamprey embryo have allowed us to distinguish between the common features in development shared by lampreys and gnathostomes, and unique developmental programmes possessed by each of these animal lineages. As seen in the basic Hox code in PAs and the Hox-free default state of the MA, these shared developmental programmes are most likely to have been established in their common ancestor in the Cambrian sea, whereas the unique features in gnathostome embryos are likely to indicate the evolutionary changes in developmental programmes behind the acquisition of the jaw, unless the features were secondarily lost in the lineage of the lamprey. As has been discussed elsewhere (Kuratani et al. 2001), the craniofacial pattern with a pair of nostrils (diplorhiny) in gnathostomes appears to be apomorphic with respect to that with single nostril (monorhiny) in many of agnathans (also see Janvier, 1996). In this transition of developmental programmes, there is a tendency that cell-autonomously functioning genes, mostly transcription factor-encoding genes, are always associated with the functionally equivalent structures or cell types, whereas the non-cell-autonomously functioning genes, such as growth factor-encoding, genes tend to shift their regulation topographically, possibly as the molecular basis for heterotopy. In this way, a specific function is not always associated with the same cell lineage or homologous embryonic cell population, and morphological homology is often lost during this type of evolution. The gnathostome jaw therefore is apparently an ‘evolutionary innovation’ by the definition of Wagner & Müller (2002), being made possible by a heterotopic shift of gene regulation. For the same reason, the morphological concepts ‘oral’ and ‘mandibular’ must be dissociated in the discussion of vertebrate history.

As far as the ‘jaw’ is defined as a derivative from the ‘mandibular arch’, the jaw homologue cannot be found in the lamprey, no matter how well the larval lips resemble jaws. As suggested from some fossil records (Janvier, 1996), agnathans would have already enjoyed a dorsoventrally movable oral apparatus patterned through identical molecular cascades (FGF8–BMP4 signalling cascades onto Hox-negative ectomesenchyme) that now pattern the vertebrate jaw. In other words, shape and function were already there, but the place to create them was not fixed. This is reminiscent of the hypothesis by Janvier (1996) that the morphological pattern of the gnathostome head, including the patterning of the mouth, nose and hypophysis, would have been merely one of the various possible evolutionary experiments invented in the Palaeozoic era. With the advance of genomic sciences, it may become possible in the near future to compare the regulation of genes in various animals at the genomic level in an evolutionary developmental context, and we will be able to relate such changes directly to the heterotopic changes in embryonic patterning programmes at the morphological level.


I thank Filippo M. Rijli for his critical reading of the manuscript and valuable discussion. This work was supported by Grants-in-Aid from the Ministry of Education, Science and Culture of Japan (Specially Promoted Research).


  • Barlow AJ, Francis-West PH. Ectopic application of recombinant BMP-2 and BMP-4 can change patterning of developing chick facial primordia. Development. 1997;124:391–398. [PubMed]
  • de Beer GR. On the nature of the trabecula cranii. Q. J. Microsc. Sci. 1931;74:701–731.
  • de Beer GR. The Development of the Vertebrate Skull. London: Oxford University Press; 1937.
  • Boorman CJ, Shimeld SM. Cloning and expression of a Pitx homeobox gene from the lamprey, a jawless vertebrate. Dev. Genes Evol. 2002;212:349–353. [PubMed]
  • Carroll RL. Vertebrate Paleontology and Evolution. New York: W.H. Freeman; 1988.
  • Cohn MJ. Evolutionary biology: lamprey Hox genes and the origin of jaws. Nature. 2002;416:386–387. [PubMed]
  • Couly GF, Le Douarin NM. Mapping of the early neural primordium in quail–chick chimeras. I. Developmental relationships between placodes, facial ectoderm, and prosencephalon. Dev. Biol. 1985;110:422–439. [PubMed]
  • Couly GF, Coltey PM, Le Douarin NM. The triple origin of skull in higher vertebrates: a study in quail-chick chimeras. Development. 1993;117:409–429. [PubMed]
  • Couly G, Grapin-Botton A, Coltey P, Ruhin B, Le Douarin NM. Determination of the identity of the derivatives of the cephalic neural crest: incompatibility between Hox gene expression and lower jaw development. Development. 1998;125:3445–3459. [PubMed]
  • Depew MJ, Lufkin T, Rubenstein JL. Specification of jaw subdivisions by Dlx genes. Science. 2002;298:371–373. [PubMed]
  • Edgeworth FH. The Cranial Muscles of Vertebrates. Cambridge: Cambridge University Press; 1935.
  • Gans C, Northcutt RG. Neural crest and the origin of vertebrates: a new head. Science. 1983;220:268–274. [PubMed]
  • Gaskell WH. On the Origin of Vertebrates. London and New York: Longmans, Green; 1908.
  • Gendron-Maguire M, Mallo M, Zhang M, Gridley T. Hoxa-2 mutant mice exhibit homeotic transformation of skeletal elements derived from cranial neural crest. Cell. 1993;75:1317–1331. [PubMed]
  • Goodrich ES. Studies on the Structure and Development of Vertebrates. London: McMillan; 1930.
  • Gorbman A. Early development of the hagfish pituitary gland: evidence for the endodermal origin of the adenohypophysis. Am. Zool. 1983;23:639–654.
  • Gorbman A, Tamarin A. Early development of oral, olfactory and adenohypophyseal structures of agnathans and its evolutionary implications. In: Foreman RE, Gorbman A, Dodd JM, Olsson R, editors. Evolutionary Biology of Primitive Fishes. New York: Plenum; 1985. pp. 165–185.
  • Gorbman A, Tamarin A. Pituitary development in cyclostomes compared to higher vertebrates. In: Yoshimura F, Gorbman A, editors. Pars Distalis of the Pituitary Gland–Structure, Function and Regulation. Amsterdam: Elsevier; 1986. pp. 3–14.
  • Graham A. The development and evolution of the pharyngeal arches. J. Anat. 2001;199:133–141. [PMC free article] [PubMed]
  • Graham A, Begbie J, McGonnell I. Significance of the cranial neural crest. Dev. Dyn. 2004;229:5–13. [PubMed]
  • Grammatopoulos GA, Bell E, Toole L, Lumsden A, Tucker AS. Homeotic transformation of branchial arch identity after Hoxa2 overexpression. Development. 2000;127:5355–5365. [PubMed]
  • Gregory WK. Fish Skulls: a Study of the Evolution of Natural Mechanisms. Malabar, FL: American Philosophical Society, reprinted by Krieger Publishers Co.; 1933.
  • Hall BK, Hörstadius S. The Neural Crest. New York: Oxford University Press; 1988.
  • Hall BK. Evolutionary Developmental Biology. 2. London: Chapman & Hall; 1998.
  • Hall BK. The Neural Crest in Development and Evolution. New York: Springer Verlag; 1999.
  • Horigome N, Myojin M, Hirano S, Ueki T, Aizawa S, Kuratani S. Development of cephalic neural crest cells in embryos of Lampetra japonica, with special reference to the evolution of the jaw. Dev. Biol. 1999;207:287–308. [PubMed]
  • Hunt P, Whiting J, Muchamore I, Marshall H, Krumlauf R. Homeobox genes and models for patterning the hindbrain and branchial arches. Development Suppl. 1991a;1:187–196. [PubMed]
  • Hunt P, Wilkinson D, Krumlauf R. Patterning the vertebrate head: Murine hox 2 genes mark distinct subpopulations of premigratory and migrating cranial neural crest. Development. 1991b;112:43–50. [PubMed]
  • Janvier P. Early Vertebrates. New York: Oxford Scientific Publications; 1996.
  • Jarvik E. Basic Structure and Evolution of Vertebrates. Vol. 2. New York: Academic Press; 1980.
  • Jefferies RPS. The Ancestry of the Vertebrates. London: British Museum (Natural History); 1986.
  • Johnels AG. On the development and morphology of the skeleton of the head of Petromyzon. Acta Zool. 1948;29:139–279.
  • Kimmel CB, Miller CT, Keynes RJ. Neural crest patterning and the evolution of the jaw. J. Anat. 2001;199:105–120. [PMC free article] [PubMed]
  • Koltzoff NK. Entwicklungsgeschichte des Kopfes von Petromyzon planeri. Bull. Soc. Nat. Moscou. 1901;15:259–289.
  • Köntges G, Lumsden A. Phombencephalic neural crest segmentation is preserved throughout craniofacial ontogeny. Development. 1996;122:3229–3242. [PubMed]
  • Kuratani S. Distribution of postotic crest cells in the chick embryo defines the trunk/head interface: embryological interpretation of crest cell distribution and evolution of the vertebrate head. Anat. Embryol. 1997;195:1–13. [PubMed]
  • Kuratani S, Matsuo I, Aizawa S. Developmental patterning and evolution of the mammalian viscerocranium: genetic insights into comparative morphology. Dev. Dyn. 1997;209:139–155. [PubMed]
  • Kuratani S, Horigome N, Hirano S. Developmental morphology of the cephalic mesoderm and re-evaluation of segmental theories of the vertebrate head: evidence from embryos of an agnathan vertebrate, Lampetra japonica. Dev. Biol. 1999;210:381–400. [PubMed]
  • Kuratani S, Horigome N. Development of peripheral nerves in a cat shark, Scyliorhinus torazame, with special reference to rhombomeres, cephalic mesoderm, and distribution patterns of crest cells. Zool. Sci. 2000;17:893–909.
  • Kuratani S, Nobusada Y, Horigome N, Shigetani Y. Embryology of the lamprey and evolution of the vertebrate jaw: insights from molecular and developmental perspectives. Phil. Trans. R. Soc. Lond. B Biol. Sci. 2001;356:15–32. [PMC free article] [PubMed]
  • Kuratani S, Kuraku S, Murakami Y. Lamprey as an Evo-Devo model: lessons from comparative embryology and molecular phylogenetics. Genesis. 2002;34:175–195. [PubMed]
  • Kuratani S. Evolutionary developmental biology and vertebrate head segmentation: a perspective from developmental constraint. Theory Biosci. 2003;122:230–251.
  • Kuratani S, Murakami Y, Nobusada Y, Kusakabe R, Hirano S. Developmental fate of the mandibular mesoderm in the lamprey, Lethenteron japonicum: comparative morphology and development of the gnathostome jaw with special reference to the nature of trabecula cranii. J. Exp. Zool. 2004;302B:458–468. [PubMed]
  • Langille RM, Hall BK. Role of the neural crest in development of the trabeculae and branchial arches in embryonic sea lamprey, Petromyzon marinus (L.) Development. 1988;102:301–310.
  • Le Douarin NM. The Neural Crest. Cambridge: Cambridge University Press; 1982.
  • Le Douarin NM, Kalcheim C. The Neural Crest. 2. Cambridge: Cambridge University Press; 1999. Developmental and Cell Biology Series.
  • Le Lièvre CS. Rôle des cellules mesectodermiques issues des crêtes neurales céphaliques dans la formation des arcs branchiaux et du skelette viscéral. J. Embryol. Exp. Morph. 1974;31:453–577. [PubMed]
  • Le Lièvre CS, Le Douarin NM. Mesenchymal derivatives of the neural crest: analysis of chimeric quail and chick embryos. J. Embryol. Exp. Morph. 1975;34:125–154. [PubMed]
  • Le Lièvre CS. Participation of neural crest-derived cells in the genesis of the skull in birds. J. Embryol. Exp. Morph. 1978;47:17–37. [PubMed]
  • Lufkin T, Mark M, Hart CP, Dollé P, Le Meur M, Chambon P. Homeotic transformation of the occipital bones of the skull by ectopic expression of a homeobox gene. Nature. 1992;359:835–841. [PubMed]
  • Mallatt J. Early vertebrate evolution: pharyngeal structure and the origin of gnathostomes. J. Zool. 1984;204:169–183.
  • Mallatt J. Ventilation and the origin of jawed vertebrates: a new mouth. Zool. J. Linn. Soc. 1996;117:329–404.
  • Mallatt J, Chen JY. Fossil sister group of craniates: predicted and found. J. Morph. 2003;258:1–31. [PubMed]
  • Mallatt J, Chen J, Holland ND. Comment on ‘A new species of yunnanozoan with implications for deuterostome evolution’ Science. 2003;300:1372. [PubMed]
  • Mallo M, Gridley T. Development of the mammalian ear: coordinate regulation of formation of the tympanic ring and the external acoustic meatus. Development. 1996;122:173–179. [PubMed]
  • Manzanares M, Nieto MA. A celebration of the new head and an evaluation of the new mouth. Neuron. 2003;37:895–898. [PubMed]
  • Martin J, Bradley A, Olson E. The paired-like homeobox gene MHox is required for early events of skeletogenesis in multiple lineages. Genes Dev. 1995;9:1237–1249. [PubMed]
  • Matsuo I, Kuratani S, Kimura C, Takeda N, Aizawa S. Mouse Otx2 functions in the formation and patterning of rostral head. Genes Dev. 1995;9:2646–2658. [PubMed]
  • McBurney KM, Wright GM. Chondrogenesis of a non-collagen-based cartilage in the sea lamprey, Petromyzon marinus. Can. J. Zool. 1996;74:2118–2130.
  • McCauley DW, Bronner-Fraser M. Neural crest contributions to the lamprey head. Development. 2003;130:2317–2327. [PubMed]
  • McGinnis W, Krumlauf R. Homeobox genes and axial patterning. Cell. 1992;68:283–302. [PubMed]
  • Morriss-Kay GM. Derivation of the mammalian skull vault. J. Anat. 2001;199:143–151. [PMC free article] [PubMed]
  • Moy-Thomas JA, Miles RS. Paleozoic Fishes. London: Chapman & Hall; 1971.
  • Murakami Y, Ogasawara M, Sugahara F, Hirano S, Satoh N, Kuratani S. Identification and expression of the lamprey Pax-6 gene: Evolutionary origin of segmented brain of vertebrates. Development. 2001;128:3521–3531. [PubMed]
  • Murakami Y, Pasqualetti M, Takio Y, Hirano S, Rijli F, Kuratani S. Segmental development of reticulospinal and branchiomotor neurons in the lamprey: insights into evolution of the vertebrate hindbrain. Development. 2004;131:983–995. [PubMed]
  • Myojin M, Ueki T, Sugahara F, et al. Isolation of Dlx and Emx gene cognates in an agnathan species, Lampetra japonica, and their expression patterns during embryonic and larval development: Conserved and diversified regulatory patterns of homeobox genes in vertebrate head evolution. J. Exp. Zool. (Mol. Dev. Evol.) 2001;291:68–84. [PubMed]
  • Neidert AH, Virupannavar V, Hooker GW, Langeland JA. Lamprey Dlx genes and early vertebrate evolution. Proc. Natl. Acad. Sci. USA. 2001;98:1665–1670. [PMC free article] [PubMed]
  • Neubüser A, Peters H, Balling R, Martin GR. Antagonistic interactions between FGF and BMP signalling pathways: a mechanism for positioning the site of tooth formation. Cell. 1997;90:247–255. [PubMed]
  • Newth DR. Experiments on the neural crest of the lamprey embryo. J. Exp. Biol. 1951;28:247–260.
  • Newth DR. On the neural crest of the lamprey embryo. J. Embryol. Exp. Morph. 1956;4:358–375.
  • Noden DM. The role of the neural crest in patterning of avian cranial skeletal, connective, and muscle tissues. Dev. Biol. 1983;96:144–165. [PubMed]
  • Noden DM. Interactions and fates of avian craniofacial mesenchyme. Development. 1988;103(Suppl.):121–140. [PubMed]
  • Ogasawara M, Shigetani Y, Suzuki S, Kuratani S, Sato N. Expression of thyroid transcription factor-1 (TTF-1) gene in the ventral forebrain and endostyle of the agnathan vertebrate, Lampetra japonica. Genesis. 2001;30:51–58. [PubMed]
  • Osumi-Yamashita N, Ninomiya Y, Doi H, Eto K. The contribution of both forebrain and midbrain crest cells to the mesenchyme in the frontonasal mass of mouse embryos. Dev. Biol. 1994;164:409–419. [PubMed]
  • Pasqualetti M, Ori M, Nardi I, Rijli FM. Ectopic Hoxa2 induction after neural crest migration results in homeosis of jaw elements in Xenopus. Development. 2000;127:5367–5378. [PubMed]
  • Qiu M, Bulfone A, Martines S, et al. Null mutation of Dlx-2 results in abnormal morphogenesis of proximal first and second branchial arch derivatives and abnormal differentiation in the forebrain. Genes Dev. 1995;9:2523–2538. [PubMed]
  • Qiu M, Bulfone A, Ghattas I, et al. Role of the Dlx homeobox genes in proximodistal patterning of the branchial arches: mutations of Dlx-1, Dlx-2, and Dlx-1 and–2 alter morphogenesis of proximal skeletal and soft tissue structures derived from the first and second arches. Dev. Biol. 1997;185:165–184. [PubMed]
  • Rijli FM, Mark M, Lakkaraju S, Dierich A, Dollé P, Chambon P. Homeotic transformation is generated in the rostral branchial region of the head by disruption of Hoxa-2, which acts as a selector gene. Cell. 1993;75:1333–1349. [PubMed]
  • Romer AS. Vertebrate Paleontology. Chicago: Chicago University Press; 1966.
  • Romer AS, Parsons TS. The Vertebrate Body. 5. Philadelphia: Saunders; 1977.
  • Satokata I, Maas R. Msx1 deficient mice exhibit cleft palate and abnormalities of craniofacial and tooth development. Nat. Genet. 1994;6:348–356. [PubMed]
  • Schilling TF, Knight RD. Origins of anteroposterior patterning and Hox gene regulation during chordate evolution. Philos. Trans. R. Soc. Lond. B Biol. Sci. 2001;356:1599–1613. [PMC free article] [PubMed]
  • Sewertzoff AN. Die Kiemenbogennerven der Fische. Anat. Anz. 1911;38:487–495.
  • Sewertzoff AN. The head skeleton and muscles of Acipenser ruthenus. Acta Zool. 1928;9:1–127.
  • Shigetani Y, Nobusada Y, Kuratani S. Ectodermally-derived FGF8 defines the maxillomandibular region in the early chick embryo: epithelial–mesenchymal interactions in the specification of the craniofacial ectomesenchyme. Dev. Biol. 2000;228:73–85. [PubMed]
  • Shigetani Y, Sugahara F, Kawakami Y, Murakami Y, Hirano S, Kuratani S. Heterotopic shift of epithelial–mesenchymal interactions for vertebrate jaw evolution. Science. 2002;296:1319–1321. [PubMed]
  • Shu D, Morris SC, Zhang ZF, et al. A new species of yunnanozoan with implications for deuterostome evolution. Science. 2003;299:1380–1384. [PubMed]
  • Szeto DP, Rodriguez-Esteban C, Ryan AK, et al. Role of the Bicoid-related homeodomain factor Pitx1 in specifying hindlimb morphogenesis and pituitary development. Genes Dev. 1999;15:484–494. [PMC free article] [PubMed]
  • Takio Y, Pasqualetti M, Kuraku S, Hirano S, Rijli FM, Kuratani S. Lamprey Hox genes and the evolution of jaws. Nature. 2004. following 262, Brief Communication Arising On Line. [PubMed]
  • Tomsa JM, Langeland JA. Otx expression during lamprey embryogenesis provides insights into the evolution of the vertebrate head and jaw. Dev. Biol. 1999;207:26–37. [PubMed]
  • Trainor PA, Ariza-McNaughton L, Krumlauf R. Role of the isthmus and FGFs in resolving the paradox of neural crest plasticity and prepatterning. Science. 2002;295:1288–1291. [PubMed]
  • Trainor PA, Melton KR, Manzanares M. Origins and plasticity of neural crest cells and their roles in jaw and craniofacial evolution. Int. J. Dev. Biol. 2003;47:541–553. [PubMed]
  • Tucker AS, Matthews KL, Sharpe PT. Transformation of tooth type induced by inhibition of BMP signaling. Science. 1998;282:1136–1138. [PubMed]
  • Uchida K, Murakami Y, Kuraku S, Hirano S, Kuratani S. Development of the adenohypophysis in the lamprey: evolution of the epigenetic patterning programs in organogenesis. J. Exp. Zool. (Mol. Dev. Evol.) 2003;300B:32–47. [PubMed]
  • Ueki T, Kuratani S, Hirano S, Aizawa S. otd/Otx cognates in a lamprey, Lampetra japonica. Dev. Genes Evol. 1998;208:223–228. [PubMed]
  • Wagner GP, Müller GB. Evolutionary innovations overcome ancestral constraints: a re-examination of character evolution in male sepsid flies (Diptera: Sepsidae) Evol. Dev. 2002;4:1–6. [PubMed]
  • Yamada G, Mansouri M, Terres M, et al. Targeted mutation of the mouse goosecoid gene results in craniofacial defects and neonatal death. Development. 1995;121:2917–2922. [PubMed]

Articles from Journal of Anatomy are provided here courtesy of Anatomical Society of Great Britain and Ireland
PubReader format: click here to try


Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...


Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...