Fishing for jaws in early vertebrate evolution: A new hypothesis of mandibular confinement

Abstract

The evolutionary origin of the vertebrate jaw persists as a deeply puzzling mystery. More than 99% of living vertebrates have jaws, but the evolutionary sequence that ultimately gave rise to this highly successful innovation remains controversial. A synthesis of recent fossil and embryological findings offers a novel solution to this enduring puzzle. The Mandibular Confinement Hypothesis proposes that the jaw evolved via spatial confinement of the mandibular arch (the most anterior pharyngeal arch within which the jaw arose). Fossil and anatomical evidence reveals: (i) the mandibular region was initially extensive and distinct among the pharyngeal arches; and (ii) with spatial confinement, the mandibular arch acquired a common pharyngeal pattern only at the origin of the jaw. The confinement occurred via a shift of a domain boundary that restricted the space the mesenchymal cells of the mandibular arch could occupy. As the surrounding domains replaced mandibular structures at the periphery, this shift allowed neural crest cells and mesodermal mesenchyme of the mandibular arch to acquire patterning programs that operate in the more posterior arches. The mesenchymal population within the mandibular arch was therefore no longer required to differentiate into specialized feeding and ventilation structures, and was remodelled into a jaw. Embryological evidence corroborates that the mandibular arch must be spatially confined for a jaw to develop. This new interpretation suggests neural crest as a key facilitator in correlating elements of the classically recognized vertebrate head 'segmentation'.
© 2015 Cambridge Philosophical Society.

Full text

BIOLOGICAL REVIEWS
Cambridge Philosophical Society
VOL. 91, NO. 3 AUGUST 2016
CONTENTS
ILIANA MEDINA & NAOMI E. LANGMORE. The evolution of acceptance and tolerance in hosts
of avian brood parasites 569
BERNARD W. T. COETZEE & STEVEN L. CHOWN. A meta-analysis of human disturbance
impacts on Antarctic wildlife 578
CATARINA FERREIRA, GUILLAUME BASTILLE-ROUSSEAU, AMANDA M. BENNETT,
E. HANCE ELLINGTON, CHRISTINE TERWISSEN, CAYLA AUSTIN, ADRIAN BORLESTEAN,
MELANIE R. BOUDREAU, KEVIN CHAN, ADRIAN FORSYTHE, THOMAS J. HOSSIE,
KRISTEN LANDOLT, JESSICA LONGHI, JOSÉE-ANNE OTIS, MICHAEL J. L. PEERS, JASON RAE,
JACOB SEGUIN, CRISTEN WATT, MORGAN WEHTJE & DENNIS L. MURRAY. The evolution
of peer review as a basis for scientific publication: directional selection towards
a robust discipline? 597
TETSUTO MIYASHITA. Fishing for jaws in early vertebrate evolution: a new hypothesis
of mandibular confinement 611
JULIAN C. EVANS, STEPHEN C. VOTIER & SASHA R. X. DALL. Information use in colonial
living 658
ANDREW N. OSTROVSKY, SCOTT LIDGARD, DENNIS P. GORDON, THOMAS SCHWAHA, GRIGORY
GENIKHOVICH & ALEXANDER V. ERESKOVSKY. Matrotrophy and placentation in invertebrates:
a new paradigm 673
JONATHAN BIRCH. Natural selection and the maximization of fitness 712
NOLAN N. BETT & SCOTT G. HINCH. Olfactory navigation during spawning migrations:
a review and introduction of the Hierarchical Navigation Hypothesis 728
DOMINIK THOM & RUPERT SEIDL. Natural disturbance impacts on ecosystem services
and biodiversity in temperate and boreal forests 760
ROMAN D. FURRER & GILBERTO PASINELLI. Empirical evidence for source–sink populations:
a review on occurrence, assessments and implications 782
ANDREW I. FURNESS. The evolution of an annual life cycle in killifish: adaptation to
ephemeral aquatic environments through embryonic diapause 796
WILFREDO A. MATAMOROS, CHRISTOPHER W. HOAGSTROM, JACOB F. SCHAEFER &
BRIAN R. KREISER. Fish faunal provinces of the conterminous United States
of America reflect historical geography and familial composition 813
PAUL-ANTOINE LIBOUREL & ANTHONY HERREL. Sleep in amphibians and reptiles: a review
and a preliminary analysis of evolutionary patterns 833
SCOTT C. BURGESS, MARISSA L. BASKETT, RICHARD K. GROSBERG, STEVEN G. MORGAN &
RICHARD R. STRATHMANN. When is dispersal for dispersal? Unifying marine and
terrestrial perspectives 867
Cover Illustration: A composite of lithographs by Jujiro Nomura (a research assistant of Edward P. Allis from 1887
to at least 1923) reveals complex head anatomy of the frilled shark Chlamydoselachus anguineus. See article by
Tetsuto Miyashita.
VOL. 91 NO. 3 BIOLOGICAL REVIEWS August 2016 (pp. 569–882)
Total-pages: Spine-width: Paper-type: UPM Classic 80
previous Spine-width: 12.1mm Paper-type: UPM Classic 80
Total-pages: 314
Current
Cambridge Philosophical Society
VOLUME 91 NUMBER 3 AUGUST 2016
BIOLO GICAL
REVIEWS
brv_91-3Cover.indd 1brv_91-3Cover.indd 1 29/06/16 4:45 PM29/06/16 4:45 PM

Biol. Rev. (2016), 91,pp.611–657. 611
doi: 10.1111/brv.12187
Fishing for jaws in early vertebrate evolution:
anewhypothesisofmandibularconfinement
Tetsuto Miyashita∗
Department of Biological Sciences, University of Alberta, Edmonton, Alberta T6G 2E9, Canada
ABSTRACT
The evolutionary origin of the vertebrate jaw persists as a deeply puzzling mystery. More than 99% of living vertebrates
have jaws, but the evolutionary sequence that ultimately gave rise to this highly successful innovation remains
controversial. A synthesis of recent fossil and embryological findings offers a novel solution to this enduring puzzle.
The Mandibular Confinement Hypothesis proposes that the jaw evolved via spatial confinement of the mandibular
arch (the most anterior pharyngeal arch within which the jaw arose). Fossil and anatomical evidence reveals: (i)the
mandibular region was initially extensive and distinct among the pharyngeal arches; and (ii) with spatial confinement,
the mandibular arch acquired a common pharyngeal pattern only at the origin of the jaw. The confinement occurred
via a shift of a domain boundary that restricted the space the mesenchymal cells of the mandibular arch could occupy.
As the surrounding domains replaced mandibular structures at the periphery, this shift allowed neural crest cells and
mesodermal mesenchyme of the mandibular arch to acquire patterning programs that operate in the more posterior
arches. The mesenchymal population within the mandibular arch was therefore no longer required to differentiate into
specialized feeding and ventilation structures, and was remodelled into a jaw. Embryological evidence corroborates that
the mandibular arch must be spatially confined for a jaw to develop. This new interpretation suggests neural crest as a
key facilitator in correlating elements of the classically recognized vertebrate head ‘segmentation’. Copyright ©2015
John Wiley & Sons, Ltd.
Key words:gnathostomes, cyclostomes, agnathans, pharyngeal arch, mandibular arch, premandibular, hyoid,
hypobranchial, neural crest, joint.
CONTENTS
I. Introduction: ‘branchiomery’ in evolution and development ............................................. 612
(1) Head segmentation and the origin of the vertebrate jaw ............................................. 612
(2) Basic scheme of pharyngeal patterning ............................................................... 613
(3) Evolution of mandibular arch derivatives ............................................................ 615
(a)Mandibularpatterninginthelivingvertebrates ................................................... 615
(b) Fossil evidence for distinct mandibular patterning ................................................ 615
(c) Embryological evidence for distinct mandibular patterning ....................................... 617
(d)Evaluationofevidencefortheancestrallyserialmandibularstructures ........................... 619
(e) Summary of evidence ............................................................................. 620
II. Review of jaw origin hypotheses .......................................................................... 621
(1) Gill Arch Hypothesis ................................................................................. 622
(2) Velum Hypothesis .................................................................................... 623
(3) Ventilation Hypothesis ............................................................................... 623
(4) Lateromedial Shift Hypothesis ....................................................................... 626
(5) Hox Hypothesis ...................................................................................... 626
(6) Heterotopy Hypothesis ............................................................................... 626
(7) Co-option Hypothesis ................................................................................ 627
(8) Summary of testing previous hypotheses ............................................................. 627
III. Mandibular Confinement Hypothesis .................................................................... 627
(1) Conditions for mandibular confinement .............................................................. 628
(2) Interfaces that confine mandibular arch derivatives .................................................. 628
*Address for correspondence (Tel: +17804923633;E-mail:tetsuto@ualberta.ca).
Biological Reviews 91 (2016) 611– 657 ©2015 Cambridge Philosophical Society

612 Tetsuto Miyashita
(a)Theinterfacewiththepremandibulardomaininlivingvertebrates .............................. 628
(b) Fossil inferences for the premandibular interface ................................................. 628
(c)Theinterfacewiththehyoidarchderivativesinlivingvertebrates ................................ 629
(d) Fossil inferences for the hyoid interface ........................................................... 629
(e)Theinterfacewiththehypobranchialmusculatureinlivingvertebrates .......................... 632
(f) Fossil inferences for the hypobranchial interface .................................................. 636
(3) The origin of jaws via mandibular confinement ...................................................... 636
(a) Mandibular Confinement Stage I ................................................................. 637
(b)MandibularConfinementstageII ................................................................ 637
(c) Mandibular Confinement Stage III ............................................................... 637
(d) Mandibular Confinement Stage IV ............................................................... 639
(e) Integration of previous jaw origin hypotheses ..................................................... 639
IV. Test of the Mandibular Confinement Hypothesis ........................................................ 640
(1) Testable predictions in the fossil record .............................................................. 641
(2) Testable predictions in embryological experiments ................................................... 641
(a) Embryological inferences for the premandibular interface ........................................ 641
(b) Embryological inferences for the hyoid interface ................................................. 642
(c) Embryological inferences for the hypobranchial interface ........................................ 642
(d)Predictionsthatremaintobetested ............................................................... 642
(3) Fossil evidence for functional correlates of jaws ...................................................... 643
(a) Feeding and ventilation ........................................................................... 643
(b)Dentition ......................................................................................... 644
(c) Perichondral association of early jaw elements .................................................... 646
(d)Thefirstjawjointandsynovialdiarthrosis ........................................................ 646
(e)Evolutionofthehyoidandbranchialarches ...................................................... 648
(4) Neural crest as a facilitator of ‘segmentation’ ......................................................... 649
V. Conclusions .............................................................................................. 650
VI. Acknowledgements ....................................................................................... 650
VII. References ................................................................................................ 651
‘‘You are old,’’ said the youth, ‘‘and your jaws are too weak
For anything tougher than suet;
Yet you finished the goose, with the bones and the beak —
Pray, how did you manage to do it?’’
‘‘In my youth,’’ said his father, ‘‘I took to the law
And argued each case with my wife;
And the muscular strength, which it gave to my jaw,
Has lasted the rest of my life.’’
Lewis Carroll (1865) Alice’s Adventures in Wonderland
I. INTRODUCTION: ‘BRANCHIOMERY’ IN
EVOLUTION AND DEVELOPMENT
(1) Head segmentation and the origin of the
vertebrate jaw
More than 99% of living vertebrate species have biting
jaws as a primary feeding apparatus (Janvier, 1996). Yet
the evolutionary origin of this highly successful structure
has plagued zoologists for generations. Classical hypotheses
invoked some form of segmented head, culminating in
an idealized model of the metameric vertebrate head
(Gegenbaur, 1859; Balfour, 1878; Goodrich, 1930; De Beer,
1937; Onai, Irie & Kuratani, 2014). This archetypical
vertebrate head is structured by a single scheme of
segmentation that includes the brain, peripheral nerves,
mesodermal mesenchyme (MMCs), neural crest cells (NCCs),
and pharyngeal endoderm from anterior to posterior,
as exists in trunk somites (Fig. 1). Therefore, any major
head structure in vertebrates – including the jaw – should
be explained as specialization in one of the metameres.
Nevertheless, many difficulties arise when attempting to
derive existing patterns from this archetype. A.S. Romer
lamented the complexity of the problem by likening it to
astudyoftheApocalypse:‘Thatwayleadstomadness’
(Thomson, 1993, p. 36).
This classical model no longer holds today. Developmen-
tal genetics has revealed that the head mesoderm is distinct
from the somatic mesoderm in vertebrates. For example,
head and trunk mesoderm express different regulatory
genes: Pitx2 is specific to the head and Pax3 specific to the
trunk (Mootoosamy & Dietrich, 2002; Sambasivan et al.,
2009; Adachi et al., 2012). The boundary between these
distinct regions is maintained by fibroblast growth factor
(FGF), retinoic acid, and bone morphogenetic protein (BMP)
signalling pathways (Tzahor et al.,2003;Botheet al.,2011).
In addition, the post-otic stream of NCCs clearly marks the
Biological Reviews 91 (2016) 611– 657 ©2015 Cambridge Philosophical Society

Fishing for jaws 613
Fig. 1. The classical scheme of the vertebrate head as
a single metameric series (modified from Goodrich, 1930):
red =mesoderm; yellow =neural crest; dull green =sensory
capsules; brown =nerves (dark =cranial nerves; light =somatic
spinal nerves); blue =endoderm. The broken line indicates
the boundary between the head and trunk. Abbreviations:
V, trigeminal nerve (note the two major trunks of this one
nerve); VII, facial nerve; IX, glossopharyngeal nerve; X, vagus
nerve; NCCs, neural crest cells; pot, post-trematic branch; prt,
pre-trematic branch.
head–trunk boundary (Kuratani, 1997; Lours-Calet et al.,
2014). Therefore, no single segmentation scheme can fully
describe the vertebrate head. Is it still reasonable to view the
origin of the jaw as specialization of a pharyngeal metamere?
This untested question poses formidable challenges to
existing hypotheses about the origin of the vertebrate jaw.
(2) Basic scheme of pharyngeal patterning
The embryonic vertebrate head has a pharyngeal region
where pharyngeal arches consisting of NCCs and MMCs
are set between the endodermal pharyngeal pouches (Fig. 2).
The NCCs primarily differentiate into cartilages, and the
MMCs mainly become muscles within these arches (Noden,
1983; Noden & Francis-West, 2006; Hall, 2009; Frisdal
& Trainor, 2014). These pharyngeal arches each have a
specific name, starting with the mandibular arch (developing
the jaw in jawed vertebrates), followed by the hyoid arch
(suspending the jaw), and then by a series of branchial arches
that develop skeletal branchial bars (5 on one side in most
crown gnathostomes, 7 in lampreys, variably more than 5
in hagfish, and variably between 5 and 45 in jawless stem
gnathostomes; Janvier, 2004). Anterior to the mandibular
arch lies the premandibular domain, which consists of NCCs
anterior to the notochord without a mesodermal component
(Couly, Coltey & Le Douarin, 1993; Kimmel & Eberhart,
2008; Wada, Nohno & Kuratani, 2011).
These anatomical terms have been used historically with
different assumptions. Therefore I use them under specific
definitions (Table 1). Pharyngeal arches are embryonic,
transient structures that confer topographic identity to
differentiated tissues. Once NCCs and MMCs differentiate
into cartilages and muscles within the pharyngeal arches,
Fig. 2. A schematic illustration of vertebrate pharyngeal pat-
terning with pharyngeal regions shaded in distinct colours by
innervations. Although omitted to avoid confusion, one sensory
sub-branch of the maxillary branch of the trigeminal nerve
(V2) – the palatine branch – extends into the premandibular
domain in many gnathostomes (Higashiyama & Kuratani,
2014). Colour codes are modified from Fig. 1 to highlight
the identity of each pharyngeal region: green =premandibular
domain; red =mandibular arch; greenish blue =hyoid arch;
light cyan =branchial arch innervated by the glossopharyn-
geal nerve (IX); purple =all non-glossopharyngeal branchial
arches [innervated by the vagus nerve (X)]. Abbreviations: V1,
ophthalmic branch of trigeminal nerve; V2+3, maxillomandibu-
lar branch of trigeminal nerve; Ant., anterior; NCC, neural
crest cell; Post., posterior. Other abbreviations are as in Fig. 1.
respectively, these tissues represent arch derivatives but
not arches themselves. Therefore, the serial organization
of pharyngeal arches does not necessarily mean serial
organization of the derivatives — this assumption is
tested throughout this review. The term visceral arches is
synonymous with pharyngeal arches and therefore is avoided,
whereas I refer to branchial arches (often used as a synonym
of pharyngeal arches in developmental biology) specifically as
asubsetofpharyngealarchesthatdevelopsfullgilllamellae.
Branchial bar is a preferred term over branchial (=gill) arch
for a skeletal element so not to confuse with the embryonic
structure, but the term gill arch is retained for the name of
the hypothesis (Gill Arch Hypothesis) for historical context.
Premandibular domain is not a pharyngeal arch, because
this region lacks key traits such as mesoderm, aortic arch,
pharyngeal pouch, and motor innervation (Fig. 2).
Confusion over phylogenetic inferences has also hindered
clarity in the discussion of jaw origins. In the current con-
sensus used herein, cyclostomes (hagfish and lampreys) form
a clade; stem gnathostomes include both jawless and jawed
forms; osteostracans represent the immediate outgroup of
jawed vertebrates, and galeaspids and pituriaspids are nested
outside that inclusive clade; placoderms represent jawed stem
Biological Reviews 91 (2016) 611– 657 ©2015 Cambridge Philosophical Society

614 Tetsuto Miyashita
Table 1. Definitions for terminology used in this review. To avoid confusion, synonymous terms in the literature (in brackets with an
equals sign) are not used. Branchial arch refers only to the pharyngeal arches that plesiomorphically support a full set of gill lamellae.
For the same reason, hypobranchial muscles and trabecula apply only to gnathostomes (although their probable homologues exist
in different forms; see main text). Likewise, an upper lip (anlage: posthypophyseal process) is specific to cyclostomes, and a maxillary
process specific to gnathostomes herein. Cranial nerves: V, trigeminal nerve; VII, facial nerve; IX, glossopharyngeal nerve; X, vagus
nerve
Terms Definitions
Anatomical terms
Branchiomeric nerves Cranial nerves VII, IX, X, and conventionally V. Each nerve innervates a specific pharyngeal arch
with motor and sensory neurons.
Hypobranchial muscles A group of somatic muscles along the floor of the pharynx in crown gnathostomes that includes jaw
depressors and tongue muscles.
Hypophyseal canal Houses an endocrine organ (adenohypophysis/pituitary gland); develops from nasohypophyseal
placode (cyclostomes) or Rathke’s pouch (gnathostomes).
Lingual apparatus A feeding structure along the floor of the pharynx in cyclostomes. Protractors and retractors slide
keratinous tooth plates over basal cartilages.
Maxillary process Secondary anterior extension from a dorsal part of the mandibular arch with NCCs and MMCs in
gnathostome embryos.
Neural crest cells (NCCs) Multipotent migratory cells that delaminate from neural plate border and form populations of
ectomesenchyme in vertebrate embryos.
Trigeminal NCCs A stream of NCCs that arises in association with trigeminal placodes, with three distinct migratory
routes: preoptic (premandibular), postoptic (premandibular), and mandibular; the latter stream
populates the mandibular arch.
Neurogenic placode An ectodermal thickening that gives rise to a sensory nervous system. Those for sensory capsules
include olfactory, adenohypophyseal (Rathke’s pouch), lens, and otic.
Epibranchial placodes Neurogenic placodes for ganglia of branchiomeric nerves associated with pharyngeal cleft: geniculate
(VII), petrosal (IX), and nodose (X).
Trigeminal placodes Neurogenic placodes for basal ganglia of trigeminal nerve; ophthalmic (for the first branch) and
maxillomandibular (for the second and third).
Pharyngeal arches (=visceral
arches)
AseriesofcolumnarstructuresofNCCsandMMCsseparatedbypharyngealpouchesinvertebrate
embryos.
Mandibular arch The most anterior pharyngeal arch, innervated by the trigeminal nerve. Its NCCs and MMCs
develop into a jaw apparatus in jawed vertebrates.
Hyoid arch The second pharyngeal arch, innervated by the facial nerve. Jaw suspension and middle ear
component in jawed vertebrates.
Branchial arches (=gill
arches)
The pharyngeal arches that follow the hyoid arch; innervated by the glossopharyngeal or vagus
nerve; branchial bars and lamellae.
Branchial bars (=gill
arches)
Skeletal elements (cartilages or bones) that support gills. A skeletal branchial bar develops from NCCs
within each branchial arch.
Pharyngeal pouches: A series of lateral protrusions of pharyngeal epithelium in vertebrate embryos. They externally open
as pharyngeal slits.
Hyomandibular pouch A pharyngeal pouch between the mandibular and hyoid arches; opens externally as a spiracle in some
vertebrates; Eustachian tube in tetrapods.
Posthypophyseal process A primordial upper lip in cyclostomes populated by postoptic trigeminal NCCs and mandibular
MMCs.
Premandibular domain A head region anterior to the mandibular arch and notochord; an embryonic snout. The domain
consists of NCCs and lacks mesoderm.
Trabecula Bilateral cartilaginous rods of the vertebrate chondrocranium that form around a hypophysis via
chondrification of premandibular NCCs.
Trigeminal nerve Cranial nerve V; three major branches: first (ophthalmic), second (maxillary), and third (mandibular).
The terms maxillary and mandibular only apply to jawed vertebrates.
Upper lip A structure that develops from the posthypophyseal process in cyclostomes, with the
postoptic-NCC-derived cartilages and the mandibular-MMC-derived muscles.
Velum A cartilaginous and muscular ventilation structure in cyclostomes, which extends into the pharynx
from the mandibular arch.
Taxonomic terms
Crown gnathostomes The last common ancestor of chondrichthyans and osteichthyans and all its descendants; a least
inclusive clade of all living jawed vertebrates.
Cyclostomata A least inclusive clade or grade that contains hagfish and lampreys but excludes gnathostomes; the
only living jawless vertebrates.
Placodermi A paraphyletic grade of jawed stem gnathostomes characterized by dermal plates over the head and
trunk.
Stem gnathostomes All vertebrates closer to the crown-group Gnathostomata than to cyclostomes; jawless forms include
galeaspids and osteostracans, whereas jawed forms include placoderms and possibly acanthodians
(generally considered as stem chondrichthyans).
MMCs, mesodermal mesenchymal cells; NCCs, neural crest cells.
Biological Reviews 91 (2016) 611– 657 ©2015 Cambridge Philosophical Society

Fishing for jaws 615
gnathostomes, paraphyletic with respect to the crown-group
Gnathostomata (chondrichthyans +osteichthyans); and
acanthodians are paraphyletic with respect to crown
chondrichthyans and perhaps to crown osteichthyans (see
Figs 11 and 17 for reference). Within this phylogenetic
framework, no single fossil or living terminal taxon
can surrogate for a common ancestor. Traditionally,
elasmobranch chondrichthyans were used as a model for
pan-vertebrate or pan-gnathostome features of the head (e.g.
Balfour, 1878; Goodrich, 1930; Mallatt, 1996). However,
it requires congruence with a tree to use a trait observed
in one terminal taxon as a primitive or derived state. For
example, a character observed in one species of lamprey is
likely plesiomorphic: (i) among lampreys if other species of
lamprey exhibit the same condition; (ii) within cyclostomes
if hagfish share the trait; and (iii) within vertebrates if
jawless stem gnathostomes have evidence for this character.
Similarly, for a character observed in chondrichthyans to be
extended to the node of crown gnathostomes, this character
must be shared with osteichthyans or stem gnathostomes.
(3) Evolution of mandibular arch derivatives
In the founding works of morphology, the vertebrate skull
was viewed as modified vertebrae (Goethe, 1790; Oken,
1807; Owen, 1848). Rejecting this idea – but still following
a similar philosophical vein – the classical model interpreted
the jaw and other mandibular arch derivatives as modified
conditions of the ancestral branchial arches (Fig. 1; Rathke,
1827; Huxley, 1864; Goodrich, 1930; Jollie, 1971): the upper
and lower jaws (palatoquadrate and Meckel’s cartilage) as
modified middle upper and middle lower branchial bar
elements (epibranchial and ceratobranchial); jaw muscles as
modified branchial constrictors; spiracular pseudobranch as
a vestigial gill; trigeminal nerve as a modified branchiomeric
nerve; and either maxillary or efferent pseudobranchial
artery as a modified aortic arch I (Fig. 3).
If the jaw originated as specialization of a metamere,
the evolutionary history of the mandibular arch derivatives
should initially parallel that of other pharyngeal arch
derivatives. That is, serial similarity of pharyngeal structures
in the vertebrate head must represent a primitive – but not
a derived – state within gnathostomes. This prediction is
tested via systematic comparison.
(a)Mandibular patterning in the living vertebrates
In crown gnathostomes (chondrichthyans +
osteichthyans), development of the mandibular arch
mirrors that of all other pharyngeal arches in patterns of
NCC migration and gene expression domains including
the Dlx code (Fig. 3; Depew, Lufkin & Rubenstein, 2001;
Kuratani, 2004, 2012; Minoux et al.,2009;Depewet al.,
2005; Minoux & Rijli, 2010; Compagnucci et al.,2013;Gillis,
Modrell & Baker, 2013). The adult morphology follows this
overall pattern in dorsoventral organization and muscle con-
figuration (Gegenbaur, 1859; Goodrich, 1930; Edgeworth,
1935; De Beer, 1937; Schilling & Kimmel, 1997; Frisdal
&Trainor,2014).Themandibularskeletonandmuscles
occupy a limited space that reflects organization of the origi-
nal embryonic arch, just like derivatives of other pharyngeal
arches that are delimited by the pharyngeal pouches. In doing
so, the mandibular arch derivatives are bounded anteriorly
and medially by the premandibular domain derivatives,
posteriorly by the hyomandibular pouch, and ventrally by
the somatic hypobranchial musculature (Fig. 4C–E).
In cyclostomes (hagfish +lampreys), however, the
mandibular skeleton and muscles (red and pink; Fig. 4A,
B) do not follow the gnathostome pattern, but show
unique anatomical differences. Anteriorly, the mandibular
arch derivatives in cyclostomes overlap the premandibular
domain from the lateral side to form an upper
lip – embryologically in a post-hypophyseal position (see
Fig. 10A; Kuratani et al.,2001;Oisiet al.,2013b). Posteriorly,
the mandibular derivatives extend within the pharynx to
form a velum. Ventrally, the derivatives sit below the floor
of the pharynx as a lingual apparatus. These positions
are otherwise occupied by non-mandibular structures in
crown gnathostomes. Therefore, the mandibular elements
are extensively distributed in these jawless lineages.
These feeding and ventilation structures in cyclostomes are
clearly identified as mandibular arch derivatives based on the
motor innervation by the trigeminal nerve and embryological
observations (Holmgren, 1946; Johnels, 1948; Lindstr¨
om,
1949; Kuratani et al.,1997,2001,2004;Kuratani,Horigome
&Hirano,1999;Kuratani,2012;Miyashita,2012;Oisiet al.,
2013a,b; Ziermann, Miyashita & Diogo, 2014). Trigeminal
motor innervation is specific to muscles derived from the
mandibular MMCs and is thus a useful marker (Song &
Boord, 1993; Higashiyama & Kuratani, 2014). Posterior to
the mandibular arch, the hyoid and branchial arches develop
a basket of cartilages that support pharyngeal structures such
as the gills, velum, lingual apparatus, and heart (Marinelli &
Strenger, 1954, 1956). Therefore, none of the structures that
develop from the mandibular arch of cyclostomes resembles
either those derived from other pharyngeal arches of the same
animals or the mandibular structures of crown gnathostomes
(Fig. 4C–E).
Crown gnathostomes (overall serially patterned arch
derivatives) and cyclostomes (distinctly patterned arch
derivatives) represent two different states of mandibular
development in vertebrates. If the jaw arose via modification
of a metamere as per the classical model: (i)stem
gnathostomes should share the crown-gnathostome-like
pattern; and (ii)serialsimilaritybetweenderivativesofall
pharyngeal arches must be due to the underlying metameric
pattern during development rather than to secondary
assimilation. However, various lines of evidence do not meet
these predictions.
(b)Fossil evidence for distinct mandibular patterning
Fossil evidence indicates that the serially patterned
derivatives of the mandibular arch in crown gnathostomes
represent a derived condition. This is because jawless
stem gnathostomes generally exhibit characters correlated
Biological Reviews 91 (2016) 611– 657 ©2015 Cambridge Philosophical Society

616 Tetsuto Miyashita
Fig. 3. Legend on Next page.
Biological Reviews 91 (2016) 611– 657 ©2015 Cambridge Philosophical Society

Fishing for jaws 617
with cyclostome-like mandibular patterning. Casts of the
trigeminal nerve indicate that a cyclostome-like upper
lip deriving from a mandibular arch (pink) laterally
overlapped the premandibular domain (green) at least in
osteostracans and galeaspids (Fig. 5; W¨
angsj¨
o, 1952; Janvier,
1981, 1985a,b,1996;Gaiet al., 2011). The presence of
muscle scars near the cast of the second branch of the
trigeminal nerve in osteostracans reinforces the idea of a
cyclostome-like muscular upper lip, because cyclostomes
have motor innervation by that branch in the equivalent
‘premandibular’ region (Song & Boord, 1993; Janvier, 2007;
Higashiyama & Kuratani, 2014).
None of these fishes has evidence for a branchial bar
in the position of the mandibular arch as predicted by the
classical model. Instead, a ventilation structure equivalent to
the cyclostome velum is inferred to have existed in galeaspids
and osteostracans (Janvier, 1981, 1985a,b,1996;Gaiet al.,
2011). As for correlates of a lingual apparatus, euphaneropids
have a mid-ventral skeletal rod likely consisting of calcified
cartilage where the piston cartilage or the longitudinal
ventral cartilages would sit in lampreys and hagfishes,
respectively (Janvier & Arsenault, 2007). Other controversial
candidates for topographical or functional similarity include
the mid-ventral pharyngeal groove in heterostracans and
the feeding apparatus of conodonts. The range of motions
in the latter implies anteroposterior rotation of S and M
elements in a way similar to the cyclostome lingual apparatus
(Goudemand et al., 2011). With the exception of the enigmatic
pharyngeal morphology of conodonts, these differentiated
‘mandibular’ structures are set in clear contrast with a series
of the branchial structures that follow. The full fossil evidence
is treated in more detail during presentation of the new
hypothesis (Section III.2).
(c)Embryological evidence for distinct mandibular patterning
Gene expression patterns in both cyclostomes and gnathos-
tomes suggest that the patterning of the mandibular arch
is distinct among all pharyngeal arches. Among numerous
examples, Hox genes are not expressed in the mandibular
arch in both cyclostomes and gnathostomes, even though
each of the other pharyngeal arches is specified by a unique
combination of collinearly expressed Hox genes (Fig. 6; Hunt
et al.,1991a,b; Trainor & Krumlauf, 2001; Takio et al., 2007).
Indeed, jaws fail to develop in gnathostome embryos with
ectopic Hox expression in the mandibular arch (Alexandre
et al., 1996; Couly et al., 1998; Pasqualetti et al.,2000;
Creuzet et al., 2002). A series of reverse experiments revealed
that ectopic jaw-like cartilages appear in non-mandibular
pharyngeal arches of Hoxa-deficient gnathostome embryos
(Rijli et al., 1993; Baltzinger et al., 2005; Minoux et al.,2009).
These experiments with the gnathostome pharyngeal Hox
code are generally interpreted as: (i) proper jaw development
requires absence of Hox expression; and (ii)allpharyngeal
arches share a mandibular-arch-like ‘default’ pattern in
the absence of Hox expression – seemingly conflicting
conclusions because a literal interpretation could imply an
improbable ancestor with jaws in every pharyngeal arch.
Contrary to that intuition, these revealing experiments
support different patterning requirements between the
mandibular and non-mandibular pharyngeal arches. The
ectopic jaw-like skeletons in Hox-deficient mutants indicate
that, at least in gnathostomes, non-mandibular pharyngeal
arches require a Hox-dependent patterning mechanism to
develop branchial bars properly (Minoux & Rijli, 2010).
Therefore, the gnathostome branchial bars arose evolution-
arily via aHox-dependent patterning and do not represent
a ‘default’ pattern of pharyngeal arches as previously
thought. In no known vertebrate does the mandibular arch
require this Hox-dependent program. Therefore, similarities
between the mandibular and non-mandibular pharyngeal
structures in gnathostomes require a patterning mechanism
that acts independent of (and was acquired later in evolution
than) the pharyngeal Hox code, such as the pan-pharyngeal
Dlx code (Fig. 3C).
Cyclostomes also exhibit such mandibular-arch-
specific gene-expression patterns and functions. In lamprey
development, morpholino knockdown of SoxE1 does not
affect the mandibular skeleton, even though the skeleton
in all the other pharyngeal arches is deleteriously affected
(McCauley & Bronner-Fraser, 2006). In addition, dorsoven-
trally patterned Dlx cognate expression differs between the
Fig. 3. Serial similarity of the mandibular, hyoid, and other pharyngeal structures in crown gnathostomes. (A) Sox10:GFP transgenic
zebrafish (Danio rerio) embryo (stage 22) in left ventrolateral view. Among other tissues, the chondrifying neural crest cells (NCCs) are
in fluorescent green, showing the primordial pharyngeal cartilages. (B) Chondrocranium of adult small spotted catshark (Scyliorhinus
canicula) in left lateral view, showing the serial pharyngeal cartilages (modified after Parker, 1878). (C) Schematic drawing of Dlx
expression domains in the pharyngeal region of the small spotted catshark embryo at the pharyngula stage in left lateral view (modified
after Gillis et al.,2013).(D)Schematicdrawingofseriallypatternedpharyngealcartilagesandmusclesinthezebrafishat96hpost
hatching (hph) (modified after Schilling & Kimmel, 1997). (E) Pseudobranch in a horizontal section of the actinopterygian Amia
calva in ventral view (modified after Allis, 1897). At the mandibular–hyoid boundary, a small, medially positioned, non-respiratory
pseudobranch is innervated by the palatine branch of the facial nerve but also sits in proximity of the glossopharyngeal nerve (which
sometimes innervates the pseudobranch). (F) Head circulation with an emphasis on pseudobranchial arteries in the frilled shark
Chlamydoselachus anguineus in left lateral view (modified after Allis, 1923). A sketch of the major cranial arteries was superimposed
on a slightly transparent gross anatomical drawing of the head in lateral view and adjusted in order to approximate correct
overall anatomical position. Colour codes for cartilages in (B–D): pink =dorsal (pharyngobranchial); red =intermediate dorsal
(epibranchial); black =joint; purple =intermediate ventral (ceratobranchial); blue =ventral (basibranchial). Colour codes for muscles
in (D): brown =dorsal; green =intermediate; yellow =ventral; orange =putative dorsal. Not to scale.
Biological Reviews 91 (2016) 611– 657 ©2015 Cambridge Philosophical Society

618 Tetsuto Miyashita
Fig. 4. The highly specialized and structurally complex mandibular musculoskeletal system in cyclostomes contrasted against
the spatially confined and serially patterned counterpart in gnathostomes. The differentiated skeleton and muscles are not serial
with those in the other pharyngeal arches in cyclostomes. (A) Hagfish (Eptatretus stoutii) in left lateral view, with colour-labelled
premandibular, mandibular, and hyoid skeletons and muscles (modified after Miyashita, 2012). (B) Lamprey (Lampetra fluviatilis)in
left lateral view, with colour-labelled premandibular, mandibular, and hyoid skeletons and muscles (modified after Johnels, 1948;
Miyashita, 2012). For (A) and (B), the cartilages developing from the mandibular neural crest cells (NCCs) are in pink, whereas
the muscles innervated by the second and third branches of the trigeminal nerve are in red. Muscular mandibular structures form
the upper lip, the velum, and the lingual apparatus. (C) Spiny dogfish (Squalus acanthias) in left lateral view, with colour-labelled
premandibular, mandibular, and hyoid skeletons and muscles (modified after Mallatt, 1997). (D) Chondrocranium of a 5-day-old
zebrafish (D. rerio) larva in left lateral view and ventral view, showing cartilages derived from the premandibular domain, mandibular
arch, and hyoid arch (modified after Hern´
andez, Barresi & Devoto, 2002). (E) The skull and cranial musculature of an adult
zebrafish in lateral view, showing the premandibular cranial elements and the mandibular and hyoid muscles (modified after
Diogo, Hinits & Hughes, 2008). Hypobranchial muscles in (E) are obscured by the skull and superficial muscles. For (C–E),
mandibular arch derivatives (pink and red) are spatially confined by premandibular, hyoid, and hypobranchial structures. Colour
codes: green =premandibular elements; greenish blue =hyoid elements; yellow green =hypobranchial muscles; red=mandibular
muscles [cartilages for (D, E)]; pink =mandibular cartilages; blue =endodermally derived tissue (spiracle, branchial slits). Not to
scale.
Biological Reviews 91 (2016) 611– 657 ©2015 Cambridge Philosophical Society

Fishing for jaws 619
Fig. 5. An example of general similarities between cyclostomes
and jawless stem gnathostomes. The osteostracan Scolenaspis
signata in ventral view, with the anatomical right side showing
reconstructions based on osteological correlates illustrated on the
anatomical left side (modified after Janvier, 1996). In general, the
mandibular region shows a cyclostome-like pattern in jawless
stem gnathostomes. The premandibular (green), mandibular
(red), hyoid (greenish blue), and following branchial (light cyan
and purple) regions are shaded over the reconstruction. Each
pharyngeal region was identified on the basis of osteological
correlates of soft tissues including cranial nerves.
mandibular arch and other pharyngeal arches in lampreys
(see Fig. 10B; Cerny et al.,2010).DlxA has dorsal and ventral
expression domains in the mandibular arch, whereas it
is expressed dorsally in other pharyngeal arches. DlxC/D
are expressed throughout the mandibular arch, but its
expression domain is restricted dorsally in others.
Cranial nerves also suggest distinction of the mandibular
arch. In both cyclostomes and gnathostomes, the trigeminal
nerve does not develop like branchiomeric nerves (cranial
nerves VII, IX and X; Figs 1 and 6). The trigeminal
ganglia are induced from an aggregate of placodes
(ophthalmic +maxillomandibular) adjacent to the mid-
and hindbrain, not from a single placode associated with
the pharyngeal cleft like branchiomeric nerves (Baker,
2005; Schlosser, 2005; Xu, Dude & Baker, 2008; Pieper
et al., 2011; Modrell et al., 2014). The trigeminal placodes
appear to sit intermediate between the anterior (olfactory,
adenohypophyseal, lens) and posterior (otic, epibranchial)
placodal regions, both spatially and in gene expression
profiles (Fig. 6; Saint-Jeannet & Moody, 2014). Otx2
expression is necessary for the anterior and trigeminal
placodes, as opposed to the mutually antagonistic Gbx2
and the posteriorly expressed Pax2 specifying the posterior
placodes (Bhat & Riley, 2011; Steventon, Mayor & Streit,
2012). On the other hand, Irx1 –3 expression domain includes
the posterior and trigeminal placodes, but not the anterior
(Schlosser & Ahrens, 2004). The trigeminal placodes are
unique among the cranial neurogenic placodes in expressing
Pax3 (Baker et al.,1999;O’Neill,McCole&Baker,2007;
Dude et al., 2009; Modrell et al., 2014). Upstream to this, the
platelet-derived growth factor (PDGF) signalling uniquely
specifies the ophthalmic trigeminal placode (McCabe &
Bronner-Fraser, 2008). The developed trigeminal nerve
does not have clear counterparts of pre- and post-trematic
branches – the former consisting only of sensory neurons and
the latter consisting of both sensory and motor neurons – as
do branchiomeric nerves (Song & Boord, 1993).
(d)Evaluation of evidence for the ancestrally serial mandibular
structures
Although the mandibular skeleton and muscles appears
to be patterned in series with the rest of the pharynx in
crown gnathostomes, the similarity needs not represent an
underlying metamerism. Historically, pharyngeal structures
that appear serial were used to reinforce the classical model,
but positing a serial pattern on the basis of superficial
similarities assumes the model a priori. These similarities
must be carefully evaluated. Even in the skeletons, no known
vertebrate has distinct mandibular elements that correspond
to pharyngobranchials or basibranchials in the branchial
arches. If the mandibular cartilages used to be patterned
exactly like other pharyngeal cartilages, that would require
multiple events of split, fusion, addition, and/or loss for parts
of the palatoquadrate and Meckel’s cartilage, without fossil
or developmental evidence.
Perhaps the best example to illustrate such imperfect cor-
respondence is the pseudobranch – an epithelial structure
that sits in the spiracle (elasmobranch chondrichthyans) or in
the subocular cavity (osteichthyans) for chemosensory, secre-
tory, and/or thermoregulatory functions (Figs 3E, F and 7B;
Laurent & Dunel-Erb, 1984). Its hemibranch-like appear-
ance, its position behind the mandibular arch, and its blood
irrigation via cavernous bodies in elasmobranchs all appears
consistent with the historical interpretation of it as a vestigial
gill (Mallatt, 1996). That is, the mandibular arch ancestrally
developed a gill like other pharyngeal arches.
However, the pseudobranch is likely not even an element
of the mandibular arch (Allis, 1916). The structure receives
innervation from the facial nerve (or sometimes the
glossopharyngeal nerve) and irrigation from an offshoot
of the oxygenated efferent hyoidean artery (afferent
pseudobranchial artery; Fig. 3F; Herrick, 1899; Laurent &
Dunel-Erb, 1984). Cavernous bodies are not diagnostic to
gills; they are a pressure-control circulatory apparatus present
in many organs. In humans, these include tear ducts, nasal
turbinates, and clitoris/penis (Ebbehøj & Wagner, 1979;
Baskin et al.,1999;Paulsenet al.,2000;Wexler&Davidson,
2004). These morphological features are consistent, however,
if the hyomandibular (spiracular) epithelium (i) secondarily
acquired folded structure for non-respiratory functions and
(ii) drew innervation and irrigation from the nearby posterior
arches.
Essentially, the interpretation of pseudobranch as a vesti-
gial gill circularly serves itself because the classical model of
perfect pharyngeal serial homology in an ancestral vertebrate
necessitates a gill in that position. This practice is tangible
in a proposed homology with a hyoidean ciliated groove
(pseudobranchial groove) in larval lampreys (Fig. 7; Mallatt,
1979, 1996). No anatomical or functional correspondences
Biological Reviews 91 (2016) 611– 657 ©2015 Cambridge Philosophical Society

620 Tetsuto Miyashita
Fig. 6. Developmental/anatomical scheme of the crown gnathostome head with an emphasis on constraints on neural crest cell
(NCC) migration and differentiation, modelled after a dogfish embryo (modified after Northcutt, 2008; Kuratani, 2012). Selected
placodal gene expression patterns follow Saint-Jeannet & Moody (2014). Colour codes for illustrated tissues follow Fig. 2, except for:
orange =mesoderm; yellow =placodes (lateral line placodes are omitted). Roman numerals =cranial nerves; black dots =ganglia.
NCCs are shaded with transparent colours specific to pharyngeal domains [green =premandibular (trigeminal NCCs, preoptic and
postoptic streams); red =mandibular (trigeminal NCCs, mandibular stream); greenish blue =hyoid; light cyan =glossopharyngeal;
purple =vagus], and dark thick lines of the same colour scheme indicate visceral cranial nerves. Abbreviations: II, optic nerve;
III, oculomotor nerve; IV, trochlear nerve; VI, abducens nerve; VIII, vestibulocochlear nerve; hb, hypobranchial muscles; hmp,
hyomandibular pouch; hy, hyoid arch; lep, lens placode; man, mandibular arch; oc, otic capsule; olp, olfactory placode; prm,
premandibular domain; r, rhombomere; rap, Rathke’s pouch (adenohypophyseal placode). See Figs 1 and 2 for other abbreviations.
exist between them. The groove is a row of ciliated epithelial
cells on the hyoid arch in larval lampreys, whereas the
pseudobranch is a highly irrigated epithelial structure in
the hyomandibular position in crown gnathostomes. An
interpretation that requires the fewest assumptions is that
these two structures evolved independently. Therefore, the
origin of the pseudobranch cannot be extended beyond
jawed vertebrates, let alone to a hypothetical mandibular gill.
(e)Summary of evidence
Evidence clearly contradicts the classical assumption that
the mandibular arch derivatives were patterned like other
pharyngeal arches in the earliest vertebrates. In order to
postulate a metameric pharynx in an ancestor, it must assume
implausibly more hypothetical character transformations
than to postulate non-serial mandibular patterning. At the
initial stage of development, the pharyngeal arches do appear
similar to each other: a body wall broken into segments by a
series of endodermal pouches regulated via Tbx1,Wnt11 and
Fgf3/8 (Piotrowski & N¨usslein-Volhard, 2000; Crump et al.,
2004; Graham, Okabe & Quinlan, 2005; Choe & Crump,
2014; Shone & Graham, 2014). However, the pharyngeal
arch derivatives are not exactly identical to one another
once the neurogenic placodes form, and once the NCCs
and MMCs differentiate into the musculoskeletal tissues.
Within crown gnathostomes, the mandibular and all other
pharyngeal arches develop structures that are similar to each
Biological Reviews 91 (2016) 611– 657 ©2015 Cambridge Philosophical Society

Fishing for jaws 621
Fig. 7. Position of the neural crest cell (NCC)-derived pharyngeal skeleton with respect to pharyngeal muscles and gill chambers in
a cyclostome and a gnathostome. The main skeletal support occurs on different sides of the branchial chambers. (A) Dorsal half of a
horizontally sectioned lamprey (Petromyzon marinus) ammocoete larva in ventral view, showing velum and branchial bars (dull yellow)
lateral to pharyngeal muscles (red) and branchial chambers (blue) (modified from Gaskell, 1908). (B) Dorsal half of a horizontally
sectioned dogfish (Sq. acanthias)inventralview,showingjawandinternalbranchialbars(dullyellow)medialtopharyngealmuscles
(red) and branchial chambers (blue) (modified from Mallatt, 1984a). Colour codes: blue =endodermal epithelium derived from
pharyngeal pouches; dull yellow =NCC-derived pharyngeal cartilages; red =pharyngeal muscles (excluding somatic muscles). Not
to scale.
other at the levels of musculoskeletal patterning. This state
represents a derived condition. As seen in cyclostomes and
jawless stem gnathostomes, the mandibular arch derivatives
plesiomorphically develop into a pattern distinct from those
in other pharyngeal arches. A successful jaw origin hypothesis
would be expected to accommodate this character polarity.
II. REVIEW OF JAW ORIGIN HYPOTHESES
A number of hypotheses have been offered to explain
the origin of the vertebrate jaw. Some of these explicitly
build on the metameric pharynx (Gill Arch and Ventilation
hypotheses). Some others focally explain one character
transition linked with the jaw origin (Velum, Lateromedial
Shift, and Heterotopy hypotheses). Yet others (Hox and
Co-option hypotheses) use expression patterns of a single
specific gene. I evaluate these hypotheses on the merit of
individual evidence (Table 2) and with two simple tests. Test
1:whethercharactertransitionspredictedbythehypothesis
depart markedly from those predicted parsimoniously from
the current consensus tree (see Figs 11 and 17). This test
discriminates against phenotypes that have never been
observed in terminal taxa or only observed in one lineage
distant from the node of all jawed vertebrates. Test 2:whether
the hypothesis explains character transitions that can be
constrained to a portion of the stem containing the node
of all jawed vertebrates. This test finds character changes
resolved outside that node and removes them from direct
causal relationships.
Biological Reviews 91 (2016) 611– 657 ©2015 Cambridge Philosophical Society

622 Tetsuto Miyashita
Table 2. Fit of different evidence and predictions to alternative hypotheses on the origin of the vertebrate jaw. This table lists only
those that discriminate multiple hypotheses by falsifying at least one. See main text (Section II) for details and justification
Types of evidence and predictions Alternative hypotheses on vertebrate and jaw origins
GA VL VN LS HX HT CO MC
(a)Fossil evidence
Premandibular domain did not constitute a
pharyngeal arch
–XXXXXXX
Mandibular arch derivative did not function as a
branchial bar
–X–/ – X / X
Gill on hyoid arch is a posterior hemibranch across
vertebrates
–X–X– / / X
Jawless stem gnathostomes had main pharyngeal
skeleton on lateral side of the mesoderm and gill
chambers
–/–X/ / /X
Jawed stem gnathostomes (placoderms) had jaw
cartilages lateral to adductors
–X–– / / / X
Velum-like ventilation structure existed in
hyomandibular boundary in jawless stem
gnathostomes
–X–/ – / / X
Paired nasal capsules and orally opening hypophyseal
canal appeared independent of jaw origins
///// – / X
Hyomandibular (adorbital) opening appeared
independent of jaw origins
/––// / /X
(b)Developmental evidence
Upper lips of jawless vertebrates are developmentally
incomparable to maxillary process of jawed
vertebrates
–/––/XXX
Velum develops as mandibular NCCs and MMCs
invaginate into a hyomandibular pouch
/––// / /X
Mandibular arch does not express Hox gene in
cyclostomes and gnathostomes
–/–/–X/X
Dorsoventrally patterned expression patterns of Dlx,
Hand, and Edn cognates in both cyclostomes and
gnathostomes
X/X–X – XX
Dlx genes evolved different repertoires in various
lineages, variably incorporated in patterning of the
mandibular arch
////–X–X
(c)Testable predictions to discriminate competing hypotheses
Extension of hypobranchial muscles to mid-ventral
pharyngeal floor did not precede a shift from lateral to
medial branchial bars
–/–X– / /X
Perturbation to the premandibular interface is linked
to jaw defects
–/–––X–X
Perturbation to the hyoid interface is linked to jaw
defects
–X–– / – – X
Perturbation to the hypobranchial interface is linked
to jaw defects
––/–– –XX
Codes for compatibility with evidence or prediction: X =evidence (or prediction) is compatible with a hypothesis; /=a hypothesis provides
no predicted outcome to explain the evidence; – =evidence (or prediction) is incompatible with a hypothesis.
Codes for alternative hypotheses: CO, Co-option Hypothesis; GA, Gill Arch Hypothesis; HT, Heterotopy Hypothesis; HX, Hox Hypothesis;
LS, Lateromedial Shift Hypothesis; MC, Mandibular Confinement Hypothesis; VL, Velum Hypothesis; VN, Ventilation Hypothesis.
(1) Gill Arch Hypothesis
The Gill Arch Hypothesis (Fig. 8A; Table 2; Gegenbaur,
1859; De Beer, 1937) explains the origin of the jaw
using the classical model of the ancestrally metameric
pharynx. In this hypothetical ancestor, the mandibular
and hyoid arches both developed gill structures. The jaw
later evolved when the mandibular skeleton was freed
from supporting the gills. This hypothesis also postulates
that the premandibular domain was once a fully formed,
most anterior pharyngeal arch, either later lost (Bjerring,
1977; Jarvik, 1980) or incorporated into the neurocranium
(De Beer, 1937). Such an archetypical pattern reduces the
observed vertebrate pharynx to a repetition of one phenotype
(branchial bars). Therefore, any other phenotype (e.g. jaw)
would be interpreted as modification of the ancestral state.
Biological Reviews 91 (2016) 611– 657 ©2015 Cambridge Philosophical Society

Fishing for jaws 623
Fig. 8. Graphic summary of three jaw origin hypotheses based on morphological observations. Each column represents a
hypothetical scenario for jaw origins from a general ancestral state (top) to a general gnathostome state (bottom). (A) Gill Arch
Hypothesis (modified from De Beer, 1937; Kuratani, 2012). (B) Velum Hypothesis (modified from Janvier, 1996). (C) Ventilation
Hypothesis (modified from Mallatt, 1996). Colour codes are enhanced from Figs 1 and 2 to highlight differentiated tissues within a
pharyngeal region: green =premandibular domain (light =skeleton; dark =V1); red =mandibular arch derivatives (light =skeleton;
dark =V2+3); blue =hyoid arch derivatives (greenish blue =skeleton; medium dark =VII), IX (pure cyan) and endoderm (pure
blue =pharyngeal epithelium; dark =pharyngeal slits; light cyan =pharyngeal cavity); purple=X; dull green =sensory capsules;
brown =CNS (medium) and notochord (dark). Abbreviations: CNS, central nervous system; hy arch, hyoid arch; man arch,
mandibular arch; prm arch, premandibular arch (hypothetical). Other abbreviations follow Figs 1 and 2.
However, an untested alternative is that jaw and branchial
bars assimilated each other by using a common patterning
mechanism.
The Gill Arch Hypothesis does not satisfy Test 1.
No vertebrate — fossil or living — is known to bear
respiratory gills on the mandibular arch or possess a fully
formed arch in the premandibular position (Janvier, 1993,
1996; Conway Morris & Caron, 2014). The premandibular
domain lacks multiple traits expected of a pharyngeal arch
(Fig. 2). To explain the loss of these traits, the Gill Arch
Hypothesis requires an unlikely scenario in which the
trabecula was incorporated into the braincase from the
lip (Fig. 8A; De Beer, 1937; Kuratani, 2012). The Gill Arch
Hypothesis does not meet the condition of Test 2 either. The
hypothesis requires homology of the pharyngeal skeleton
across the jaw origin. However, the skeleton develops on
different sides of the pharyngeal arches: lateral to the
gill, mesodermal core, nerve trunks, and aortic arches in
cyclostomes and jawless stem gnathostomes (euphaneropids,
galeaspids, heterostracans, and osteostracans; Janvier, 2004);
medial in crown gnathostomes and possibly in Metaspriggina
(Fig. 7; Schaeffer & Thomson, 1980; Janvier, 2007; Conway
Morris & Caron, 2014). The Gill Arch Hypothesis does not
explain this difference.
(2) Velum Hypothesis
The Velum Hypothesis (Fig. 8B; Table 2; Janvier, 1993;
Forey, 1995) postulates a velum – a ventilation structure
that develops from the mandibular arch in cyclostomes – as
a precursor to jaws. The velum is a mobile apparatus
with muscles and cartilages, and is innervated by the
third branch of the trigeminal nerve that is compared
to the mandibular branch of the same nerve in crown
gnathostomes (Song & Boord, 1993). However, the velar
cartilages and muscles have no exact counterparts in
gnathostomes, which contradicts simple co-option of the
velum into a jaw. Unlike the jaws having an upper and
lower component supporting the mouth, the cyclostome
velum develops into the pharynx as NCCs and MMCs
form a pocket along the anterior wall of the hyomandibular
pouch (Section III.2c; see Fig. 14G). The hypothesis requires
some mechanism of transposition for this cell population
with respect to the mouth. Otherwise, Test 2 is not
satisfied.
(3) Ventilation Hypothesis
The Ventilation Hypothesis (Fig. 8C; Table 2;
Mallatt, 1996, 2008) holds that the jaw appeared first
Biological Reviews 91 (2016) 611– 657 ©2015 Cambridge Philosophical Society

624 Tetsuto Miyashita
Fig. 9. Graphic summary of two jaw origin hypotheses based on comparative development. (A) Lateromedial Shift Hypothesis.
Left lateral view of a stage-25 lamprey P. marinus from McCauley & Bronner-Fraser (2003); horizontal section of the pharyngeal
region of a 16-day old P. marinus from Kimmel, Miller & Keynes (2001); left lateral view and horizontal sections of the pharyngeal
region of stage-34/35 axolotl (Ambystoma mexicanus) embryos from Cerny et al. (2004b). Neural crest cells (NCCs) wrap around the
pharyngeal arches in gnathostomes. Line across side view indicates plane of sectioning shown in the panels below. Horizontal
sections of axolotl show the left mandibular and hyoid NCCs labelled with either Snail riboprobe or anti-fibronectin/DAPI. Arrows
indicate NCCs. (B) Horizontal sections of stage-26/27 (left) and stage 29 (right) axolotl embryos stained with Snail riboprobe across
the same sectioning plane as in the axolotl embryo in (A), showing the inward migration of NCCs (from Cerny et al., 2004b). (C) Hox
Hypothesis [amphioxus (Branchiostoma sp.) and lamprey from Cohn (2002) with new labels; zebrafish (D. rerio) from Thisse & Thisse
(2005); amphioxus and zebrafish are horizontally inverted for consistency, showing left side]. Hox6 is not expressed in the pharyngeal
region in gnathostomes. Arrows indicate expression zones. Arabic numerals indicate numbers of pharyngeal arches. The asterisk
marks the cerebral vesicle. Not to scale. Abbreviations: a-fibr., anti-fibronectin; e, eye; llp, lower lip; mes, mesoderm; nc, notochord;
pe, pharyngeal endoderm; ulp, upper lip.
as a pumping structure modified from an internal branchial
bar. Then it was co-opted for feeding, as the mouth
migrated posteriorly. This hypothesis premises on the Gill
Arch Hypothesis with a number of proposed oropharyngeal
homologues such as musculoskeletal elements between the
cyclostome upper lip and the gnathostome upper jaw.
However, it does not consider the premandibular region
as a pharyngeal arch. The branchial bar homology is
reconciled by postulating an ancestor with both lateral
and medial branchial bars, and elasmobranch chon-
drichthyans are posited as retaining the ancestral lateral
branchial bars in the form of extrabranchial cartilages
(Mallatt, 1984a,1996).
The comparative basis of the Ventilation Hypothe-
sis – gross morphological similarities in the oropharyngeal
region across vertebrates – is difficult to reconcile with
developmental and fossil evidence. Although both the
cyclostome upper lip and the gnathostome upper jaw
appear to occupy similar positions, the former develops as a
posthypohyseal process during the primary NCC migration,
and the latter as a secondary extension of mandibular
elements into the premandibular region (Section II.6; see
Fig. 10A). In another example, the cyclostome velum and
the gnathostome oral valve are compared. However, the
oral valve is a medial extension of the oral epithelium, not
the anterior wall of the hyomandibular pouch as in the
cyclostome velum. For fossil evidence, the laterally posi-
tioned branchial bars represent a plesiomorphic condition
with respect to jawed vertebrates (see Fig. 11). There-
fore, the proposed alterations to the medially positioned
Biological Reviews 91 (2016) 611– 657 ©2015 Cambridge Philosophical Society

Fishing for jaws 625
Fig. 10. Graphical summary of two jaw origin hypotheses based on comparative development. (A) Heterotopy Hypothesis (all
schematic drawings are modified after Kuratani, 2012). From top to bottom, lamprey (Lethenteron japonicum)anddogfish(Scyliorhinus
torazame) embryos: in left lateral view, showing the premandibular and mandibular expression domains of oropharyngeal patterning
genes; in left lateral view, showing migration of trigeminal neural crest cells (NCCs); and in anteroventral view, showing migration
of trigeminal NCCs, early (left) to late (right) stage. Separation of the nasohypophyseal placode into the paired olfactory and
adenohypophyseal placodes resulted in formation of the trabecula via proliferation of preoptic and postoptic trigeminal NCCs
(premandibular domain) without contribution of mandibular mesodermal mesenchyme (MMCs). This change correlated with
heterotopy of the Dlx expression domain. A primary anterior migration of the postoptic trigeminal NCCs and the mandibular
MMCs forms the posthypophyseal process (php) in cyclostomes, whereas secondary forward extension of the mandibular arch
becomes the maxillary process (mx) in gnathostomes. Colour code for the middle and bottom panels of (A) follow Fig. 6. (B) Co-option
Hypothesis. Left lateral view of the lamprey (P. marinus) and zebrafish (D. rerio) embryos showing the pharyngeal skeleton, modified
after Medeiros & Crump (2012); ventral view of the embryos showing expression of Nkx3.2 from Cerny et al. (2010). The red asterisk
marks the presumptive position of the jaw joint, shown in the right panel in ventral view. The pharyngeal skeleton is patterned
dorsoventrally by the Dlx family and its pathway genes across vertebrates. Not to scale. Abbreviations: CNS, central nervous system;
man, mandibular arch; md, mandible; mx, maxillary process; nc, notochord; otp, otic placode; php, posthypophyseal process; prm,
premandibular region; tri, trigeminal. Other abbreviations follow Fig. 1.
mandibular branchial bar alone cannot explain the origin of
the jaw.
As such, this hypothesis fails to satisfy Tests 1 and
2. Homologies are proposed primarily on the basis of a
three-taxon comparison: a cephalochordate, larval lampreys,
and chondrichthyans, each assumed to approximate
‘protochordate’, ‘pre-gnathostome’, and crown gnathostome
states, respectively (Mallatt, 1996). A predicted series
of transformations dismisses all morphological patterns
in hagfish and stem gnathostomes inconsistent with the
scenario as independent specializations. Consequently, the
hypothesis is contradicted by synapomorphies predicted
for relevant nodes of the tree (see Fig. 11). For example,
the chondrichthyan-based hypothetical ancestor is seriously
challenged by the osteichthyan-like branchial skeleton of
the stem chondrichthyan Ozarcus (Pradel et al., 2014) or by
paraphyly of placoderms and acanthodians with respect to
crown gnathostome lineages (see Fig. 17; Brazeau, 2009;
Biological Reviews 91 (2016) 611– 657 ©2015 Cambridge Philosophical Society

626 Tetsuto Miyashita
Davis, Finarelli & Coates, 2012). This tree topology suggests
that placoderms, acanthodians, and stem osteichthyans
retain more symplesiomorphic conditions than living
chondrichthyans (Zhu et al., 2013; Brazeau & Friedman,
2014; Giles, Friedman & Brazeau, 2015).
(4) Lateromedial Shift Hypothesis
Although the Lateromedial Shift Hypothesis (Fig. 9A, B;
Table 2; Kimmel et al.,2001,2003;Cernyet al.,2004b)is
often presented as the ‘Outside-in’ Hypothesis, an alternative
name is used here to avoid confusion with the ‘Outside-in’
and ‘Inside-out’ hypotheses of tooth origins (Fraser et al.,
2010). According to this hypothesis, cyclostomes represent
an ancestral condition in which NCCs remain lateral to
the MMCs within the pharyngeal arches and chondrify
into the branchial bars at that position. At the origin of
the jaw, chondrogenic NCCs migrated to the medial side
by surrounding the mesodermal bar within the pharyngeal
arches (Fig. 9B).
The hypothesis is contradicted by the observation that
NCCs do form a cylinder around the MMCs within the pha-
ryngeal arches in lampreys (Meulemans & Bronner-Fraser,
2002; McCauley & Bronner-Fraser, 2003). Thus a modified
version of the hypothesis proposes that, instead of the location
of the NCCs, the endoderm–ectomesenchyme induction site
for chondrification may have switched to the medial side of
the gill (Cerny et al.,2004b). This suggestion conforms to Dlx
expression in the NCCs on the lateral side of the pharyngeal
arches in lampreys (Neidert et al.,2001).
However, this hypothesis assumes that internalization of
the skeleton occurred across the pharynx at once. This is
incompatible with fossil evidence (Test 2). In placoderms,
the jaw cartilages generally lay lateral with respect to the jaw
adductors (see Fig. 23A; Section IV.3c;Young,1984;Zhu
et al., 2013), even though the branchial bars did not form an
external pharyngeal basket (Stensi¨
o, 1969; Denison, 1978;
Forey & Gardiner, 1986). Therefore, the internalization
did not correlate between the mandibular and branchial
skeletons.
(5) Hox Hypothesis
The Hox Hypothesis (Fig. 9C; Table 2; Cohn, 2002) is
based on the expression of Hox6 cognate in the mandibular
arch of the lamprey Petromyzon marinus. In gnathostomes,
the mandibular arch does not express Hox, although other
pharyngeal arches each express a unique combination of Hox
genes (Hunt et al.,1991a,b; Trainor & Krumlauf, 2001). If the
mandibular Hox expression in P. marinus is a primitive state,
then its loss in a gnathostome ancestor may have allowed the
mandibular arch to acquire a novel developmental fate – a
jaw. However, other species of lamprey (Lampetra fluviatilis;
Lethenteron japonicum =Le. camtschaticum; Renaud, 2011) lack
Hox expression in the mandibular arch (Takio et al.,2004,
2007). This distribution suggests that the lack of mandibular
Hox expression is a primitive condition for lampreys, and
Hox6 likely has no role in the origin of the jaw. Hox6 is
neither collinearly expressed in P. marinus (Cohn, 2002) nor
part of the pharyngeal Hox code in crown gnathostomes
(Trainor & Krumlauf, 2001). Therefore, this hypothesis does
not satisfy Test 2.
(6) Heterotopy Hypothesis
The Heterotopy Hypothesis (Fig. 10A; Table 2; Kuratani
et al.,2001,2013;Shigetaniet al.,2002;Kuratani,2004,2005,
2012; Shigetani, Sugahara & Kuratani, 2005) maintains that
the jaw evolved with a posterior shift of Dlx-expressing NCCs.
It postulates that the Dlx expression was originally uniform
and broad across the trigeminal NCCs in vertebrates,
as in one species of lamprey (Le. japonicum; Myojin et al.,
2001; Kuraku et al.,2010).Withposteriorrestrictionof
the Dlx-expressing NCCs, the mandibular arch acquired
dorsoventrally cascading expression of Dlx genes to pattern
the upper and lower jaws with a joint (Depew et al.,2001,
2005). This shift spatially correlates with the more posterior
expression zone of the epithelial BMP4-induced Msx1 at both
ends of the FGF8-markered oral epithelium in gnathostomes
relative to the condition in lampreys (Shigetani et al.,2002).
The hypothesis predicts that the posterior shift followed
diplorhiny (paired nasal passages). The nasal cavity and
hypophyseal canal are a single structure that develops from
a single nasohypophyseal placode in cyclostomes (Janvier,
1996). In crown gnathostomes, however, the olfactory
placodes are paired and the single adenohypophyseal placode
sits in Rathke’s pouch within the oral cavity (Kuratani
et al., 2001). These tripartite nasohypophyseal placodes allow
proliferation of the non-Dlx-expressing trigeminal NCCs
in the periphery, leading to the formation of trabecula,
ethmoidal plate, and nasal septum. This proliferation
uncouples the Dlx-expressing trigeminal NCCs from the
premandibular posthypophyseal process (php; Fig. 10A; Oisi
et al.,2013b). Instead, a secondary anterior extension of
the mandibular arch forms a maxillary process in crown
gnathostomes (mx; Fig. 10A).
Phenotypes linked to the Dlx heterotopy – such as
the trabecula derivatives and oral connection of the
hypophyseal canal – occur in galeaspids (Gai et al.,2011).
However, these conditions in a jawless stem gnathostome
decouple heterotopy from the jaw origin. Parsimoniously, the
conditions observed in galeaspids arose either (i)inthelast
common ancestor of galeaspids and jawed vertebrates or (ii)
independently in the two lineages. Thus this transition cannot
be coupled to the jaw origin. Therefore, this hypothesis does
not satisfy Test 2. Once predicted correlates of the heterotopy
occur in a jawless form, no direct causal relationship can be
assumed. In addition, at least one species of lamprey (P.
marinus) has dorsoventrally patterned expression of Dlx genes
and a ventral expression domain of their upstream regulators
(Edn1 and Hand2)withinthemandibulararch(Cernyet al.,
2010). So the nested expression pattern of Dlx pathway genes
does not always require heterotopic conditions.
These incompatibilities partly arise from ambiguity in the
predictions of the Heterotopy Hypothesis. It is difficult to
distinguish two heterotopic phenomena: (i) posterior shift of
Biological Reviews 91 (2016) 611– 657 ©2015 Cambridge Philosophical Society

Fishing for jaws 627
Dlx-expressing NCCs and the oral epithelium or (ii)lossof
Dlx expression in the premandibular region with a shift of
the oral epithelium. The trigeminal NCCs migrate in three
distinct streams: preoptic, postoptic, and mandibular. The
first two populate the premandibular region anterior to the
mandibular arch (Fig. 10A). Dlx cognates are expressed in
all of these populations in cyclostomes (Cerny et al.,2010;
Kuraku et al., 2010; Fujimoto et al., 2013; Kuratani et al.,
2013), whereas the Dlx expression specifically marks the
mandibular stream in crown gnathostomes (Depew et al.,
2001, 2005). Meanwhile, the cyclostome posthypophyseal
process clearly contains the mandibular MMCs to form
the upper lip (Horigome et al.,1999;Kurataniet al.,2004,
2013; Kuratani, 2012). Nevertheless, the narrative of the
Heterotopy Hypothesis does not clearly indicate whether
the NCCs are recognized based on Dlx expression or
association with mandibular MMCs. Related to this, it
remains ambiguous about how BMP4/FGF8-markered oral
epithelium shifted.
(7) Co-option Hypothesis
The Co-option Hypothesis (Fig. 10B; Table 2; Cerny et al.,
2010; Medeiros & Crump, 2012) proposes that dorsoventrally
patterned Dlx expression already existed in the common
ancestor of cyclostomes and gnathostomes, and that the
jaw arose with co-option of the jaw-joint-specific Nkx3.2 (also
known as Bapx1) expression downstream of endotheline (Edn)
signalling (Miller et al.,2003).Dlx expression in the lamprey
Le. japonicum appears uniform at Tahara’s stages 22 –26 (Myo-
jin et al., 2001; Kuraku et al., 2010). However, this hypothesis
is based on the observation of dorsoventrally patterned Dlx
expression in another species of lamprey (P. marinus) at slightly
later developmental stage (Tahara’s stages 25–26.5).
Bosentan treatment provides circumstantial evidence that
Le.japonicum also dorsoventrally patterns the larval skull in
this region via Edn signalling (Yao et al., 2011). Furthermore,
dorsoventral expression patterns of Dlx and Edn in hagfish
(Fujimoto et al., 2013) suggest that dorsoventral patterning
of the mandibular arch arose before the origin of the jaw.
Metaspiriggina, a putative stem vertebrate, has seven pairs of
bipartite skeletal elements on the medial side of the gills
(Conway Morris & Caron, 2014). Although exact identity of
the elements remains to be resolved, it is consistent with the
prediction about dorsoventral pharyngeal skeletal patterning
before the jaw origin.
However, this hypothesis does not readily satisfy Test
2. Unclear orthology and paralogy of the Dlx cognates in
cyclostomes (Kuraku, 2013; Takechi et al.,2013)makesit
difficult to compare cyclostome Dlx expression patterns and
the gnathostome Dlx code. That said, Dlx-expressing NCCs
do not distribute in the same manner across vertebrates
(Fig. 10A). Between the cyclostome posthypophyseal process
and the gnathostome maxillary process, the trigeminal NCCs
come from different parts at different timings (Cerny et al.,
2004a). Therefore, the available data only suggest that the Dlx
pathway has been variably deployed in dorsoventral oropha-
ryngeal patterning across vertebrates. Additionally, Nkx3.2
expression – predicted to specify a jaw joint – appears to
exist in one species of lamprey (Le. japonicum) without a jaw
joint (see Fig. 22J, K; Kuraku et al.,2010).
(8) Summary of testing previous hypotheses
None of the previous jaw origin hypotheses is fully compatible
with current evidence. The hypotheses that did not satisfy
Test 1 (Gill Arch and Ventilation hypotheses) require a
set of hypothetical ancestral phenotypes that have never
been observed. As a result, they predict non-parsimonious
evolutionary transitions. The hypotheses that did not satisfy
Test 2 either make conflicting predictions with fossil evidence
or use character transitions phylogenetically decoupled from
the origin of the jaw.
One common thread for most of these hypotheses is an
ancestrally serially patterned pharynx. This assumption is
explicit in the Gill Arch and Ventilation hypotheses. The
Lateromedial Shift, Hox, and Co-option hypotheses are
clearly conceived to derive a jaw from a serially patterned
pharynx. It is difficult to reconcile these hypotheses with the
already outlined evidence for distinctly patterned mandibular
derivatives as a plesiomorphic condition. The Velum and
Heterotopy hypotheses do not consider the non-mandibular
part of the pharynx. These latter hypotheses still need to
account for character transitions constrained to the origin of
the jaw.
To make matters more challenging, many of these
hypotheses are not mutually exclusive, although it is unlikely
that many or all of the proposed key innovations for the jaw
appeared concurrently. Therefore, any new hypothesis must
predict a sequence of evolutionary events, each of which may
be explained by some current hypotheses. Several guidelines
would help building an inclusive hypothesis with as few ad
hoc explanations as possible. (i) A hypothetical transformation
series should concentrate on a segment of the main stem of the
vertebrate tree between the node of jawed vertebrates (pla-
coderms, acanthodians, chondrichthyans, and osteichthyans)
and the nodes with successive likely outgroups (osteostracans
and galeaspids). (ii) The initial and last stages of the transfor-
mation series should be informed by a set of likely synapo-
morphies predicted for the closest node to the stage along
the main stem. These traits are predicted by comparison
between (a) hagfish and lampreys for the node of cyclostomes,
(b) cyclostomes and jawless stem gnathostomes for the last
common ancestor of crown gnathostomes, osteostracans,
and galeaspids, (c) crown gnathostomes and jawed stem
gnathostomes for the node of jawed vertebrates, and (d)
chondrichthyans and osteichthyans for the node of crown
gnathostomes. (iii) Gene expression patterns should be linked
to anatomically observable phenotypes as far as possible.
III. MANDIBULAR CONFINEMENT HYPOTHESIS
A review of the evidence suggests that the jaw arose when
the mandibular arch acquired a pattern serial with the rest
Biological Reviews 91 (2016) 611– 657 ©2015 Cambridge Philosophical Society

628 Tetsuto Miyashita
of the pharynx, although none of the existing hypotheses
predicts such a transition. If this scenario is correct, the
mandibular derivatives should be patterned distinctly from
the hyoid and branchial derivatives among jawless stem
gnathostomes. Then the acquisition of serial mandibular
patterning should coincide with the origin of the jaw.
What sets these two states apart is spatial confinement of
the mandibular derivatives in jawed vertebrates relative to
jawless forms.
(1) Conditions for mandibular confinement
Many tissues are known to constrain NCC and MMC
migration. They include epithelium, neurogenic pla-
codes, and BMP4 signalling centres among others
(Graham et al.,1994;Hall,2009;Steventon,Mayor&
Streit, 2014). These markers indicate positions beyond
which NCCs and MMCs of the mandibular arch do not
migrate at the pharyngula stage. This occurs because a
space is pre-occupied by an epithelium or by cells from the
surrounding domains, or because it lacks proper signalling.
Unambiguous observations corroborate the role of sig-
nalling: (i) skeletal patterning of NCCs fails in the absence of
pharyngeal pouches (Crump et al.,2004;Jandziket al.,2014);
(ii)placodalcellsattractandthenrepulseNCCsvia local
inhibition of adhesion by non-canonical planar cell polarity
and N-cadherin signalling (Theveneau et al.,2013);and
(iii)NCCsmigratingfromdifferentlevelsalonghindbrain
rhombomeres do not mix (Fig. 6; K ¨
ontges & Lumsden, 1996;
Dahmann, Oates & Brand, 2011). Even in exceptional
cases in which NCC migration markedly departs from
the conserved anteroposterior pattern, NCCs of different
streams remain intact and have distinct fates, as indicated
by little area of overlap in the skull of the African clawed
frog Xenopus laevis (Gross & Hanken, 2008).
In the mandibular arch, many factors facilitate prolifer-
ation and differentiation of NCCs and MMCs at interfaces
with the surrounding tissues and cell populations. Among
them are the foregut and pharyngeal epithelia that pattern
the NCC-derived cartilages through signalling mechanisms.
Most notably, hedgehog signalling involving Shh ensures
condensation, survival, and competence of Hox-negative
NCCs to respond specifically to cues around the epithelium
differently from Hox-positive NCCs (Couly et al.,2002;
Ruhin et al.,2003;Creuzet,Couly&LeDouarin,2005;
Eberhart et al.,2006;Benouaiche et al., 2008; Balczerski et al.,
2012). Some boundaries may be defined by the reciprocal
inhibitory interactions among endodermal Shh and ecto-
dermal FGF8 and BMP4 (Shigetani, Nobusada & Kuratani,
2000; Haworth et al., 2007). FGF signalling plays multiple
roles in patterning of the pharyngeal series as well. Both
endodermally and ectodermally expressed FGF3 and FGF8
are required for: (i) formation of the pharyngeal pouches
(thus separating NCCs for distinct pharyngeal arches from
each other); (ii) survival of the post-migratory NCCs; and (iii)
chondrification of the NCCs around the endodermal epithe-
lium in both cyclostomes and gnathostomes (Sarker et al.,
2001; Abu-Issa et al.,2002;Crumpet al.,2004;Edlundet al.,
2014; Jandzik et al., 2014). Other various factors can also
influence NCC distribution. For example, Ednr expression
ensures survival and proliferation of NCCs in the ventral
region of the mandibular arch (Abe, Ruest & Clouthier,
2007). Disruptions to NCC patterning also affect the tightly
linked MMC patterning (Rinon et al.,2007;Grenieret al.,
2009; Sambasivan, Kuratani & Tajbakhsh, 2011).
(2) Interfaces that confine mandibular arch
derivatives
These mechanisms that delimit NCCs and MMCs imply
conditions for the mandibular arch to pattern a jaw. For
the mandibular arch derivatives, such spatial confinement
occurs at interfaces with three regions within the embryonic
head: premandibular domain, hyoid arch/hyomandibular
pouch, and hypobranchial musculature.
(a)The interface with the premandibular domain in living vertebrates
The interface with the premandibular domain forms
around the hypophyseal canal. The reciprocally inhibitory
Shh and Fgf8 expression zones around Rathke’s pouch
(hypophyseal canal) are necessary for proper development
of the mandibular derivatives and pituitary gland (Haworth
et al., 2007). This is exactly where the premandibular and
mandibular trigeminal NCC derivatives are set apart from
each other (Fig. 6; see Fig. 19A, B; Brito, Teillet & Le
Douarin, 2006; Khonsari et al.,2013).
This interface is associated with different sets of morpho-
logical structures between cyclostomes and gnathostomes.
In cyclostomes, the hypophyseal canal is associated with a
single nasal passage, as per a single nasohypophyseal placode
(Fig. 10A). The trigeminal NCCs of the postoptic stream
migrate with the mandibular MMCs to form a posthypophy-
seal process, which develops into an upper lip (Horigome
et al., 1999; Oisi et al.,2013b). In crown gnathostomes, how-
ever, the separation of the paired olfactory and single
adenohypophyseal placodes allows skeletogenic prolifera-
tion of the pre-/postoptic trigeminal NCCs in the periphery
without contribution of the mandibular MMCs (Fig. 10A).
These NCCs chondrify into a trabecula and an ethmoidal
plate (Kuratani et al.,2001,2004;Kuratani,2012).Unlike
cyclostomes, there is no posthypophyseal process. Instead, the
trigeminal NCCs of the mandibular stream secondarily form
a maxillary process with contribution of the MMCs. Thus,
these respective cyclostome and gnathostome structures are
mutually exclusive phenotypes. The latter traits qualify as
indicators for a gnathostome-like premandibular interface
because they confine the trigeminal NCCs of the mandibular
stream and the mandibular MMCs in that region.
(b)Fossil inferences for the premandibular interface
Jawless stem gnathostomes generally exhibit cyclostome-like
conditions for the premandibular interface (Figs 11 and 12).
These traits are a single nasal aperture, a single nasal capsule
and a hypophyseal canal associated with the nasal passage.
Biological Reviews 91 (2016) 611– 657 ©2015 Cambridge Philosophical Society

Fishing for jaws 629
They also generally lack the ethmoidal plate and nasal septum
(Fig. 12A –E). In addition, osteostracans show clear evidence
of the cyclostome-like upper lip in having the area of muscular
attachment beside the foramen for the second branch of the
trigeminal nerve [Fig. 5; character (ch.) 1 in Fig. 11].
However, relatively crownward taxa (galeaspids, heteros-
tracans, thelodonts, and possibly pituriaspids) had paired
nasal capsules (Janvier, 1993, 1996, 2007). Nasal capsule
morphology is unknown in arandaspids, but they had
paired nasal apertures with a septum (Fig. 13; Gagnier,
1993). Myllokunmingiids provide possibly the earliest evi-
dence for paired nasal structures (Shu et al.,2003;Conway
Morris & Caron, 2014), although it remains unresolved
whether the structures represent true nasal capsules. Galea-
spids had skeletal derivatives of trabecula (ethmoidal pro-
cess) and a hypophyseal canal that opened into the oral
cavity (Fig. 12F–H; Gai et al., 2011). These jawless verte-
brates exhibit a tendency to form a crown-gnathostome-like
premandibular –mandibular boundary. This transition could
have occurred well before the origin of the jaw. Attributes
of the crown gnathostome premandibular interface (nasal
septum, paired nasal capsules, trabecula-derived skeletal ele-
ments, orally opened hypophyseal canal) are identified across
placoderms (Stensi¨
o, 1963, 1969; Young, 1979, 1984; Gou-
jet, 1984; Dupret et al., 2014). Therefore, they were likely
present in the earliest jawed gnathostome.
These characters did not evolve all at once. Based on
the consensus tree (Fig. 11), paired nasal capsules (ch. 2) are
likely a plesiomorphic condition for jawed vertebrates, but
paired nasal apertures (ch. 3) in arandaspids is probably a
homoplasy. The trabecula derivative and the orally opening
hypophyseal canal in galeaspids (ch. 4, 5) either have a
common origin with those in jawed vertebrates (Gai et al.,
2011; Dupret et al., 2014) or they represent convergence.
Given the broad distribution of paired nasal capsules,
plesiomorphic states in osteostracans may represent a
reversal. An alternative possibility is that the reconstructed
‘telencephalon’ in osteostracans represents paired nasal
capsules (W¨
angsj¨
o, 1952). Whichever view is correct, the
crown-gnathostome-like premandibular interface probably
formed gradually and was fully established with a complete
loss of the muscular upper lip in part of the tree closer
towards jawed vertebrates than towards osteostracans.
(c)The interface with the hyoid arch derivatives in living vertebrates
Ahyomandibularpouchmarkstheinterfacebetween
the mandibular and hyoid arch derivatives. The
pouch separates the mandibular and hyoid streams
of NCCs and confers the mandibular identity to the
Hox-deficient NCCs anterior to it (Couly et al.,1998,2002;
Crump et al.,2004).
No clear boundary exists between the mandibular and
hyoid arch derivatives in cyclostomes at a gross anatomical
level. The velum – a ventilation structure – develops as the
NCCs and MMCs of the mandibular arch protrude into the
hyomandibular pouch posteriorly (Fig. 14F, G; Holmgren,
1946; Oisi et al.,2013a,b). Thus, the velar membrane,
along with peripheral endoderm epithelium, represents
an invaginated hyomandibular pouch. Once differentiated,
the velum extends within the pharynx posteriorly beyond
the hyoid region, and the proximal end of the velar
skeleton contacts the hyoid arch skeleton (Figs 4A, B and
7A). Furthermore, the hyoid muscles broadly overlap the
mandibular muscles in cyclostomes (Marinelli & Strenger,
1954, 1956). Such anatomical organization is consistent
with the observation that some NCCs of the hyoid stream
migrate to the mandibular arch in lampreys (McCauley &
Bronner-Fraser, 2003).
In crown gnathostomes, a number of hyoid structures
mark the interface along the hyomandibular pouch. The
hyomandibular pouch persists in adults with a spiracular
organ (non-tetrapod gnathostomes) or a paratympanic organ
(tetrapods) – a mechanosensory structure associated with the
geniculate ganglion of the facial nerve (Fig. 14A, B; Norris
& Hughes, 1920; Barry, Hall & Bennett, 1988; O’Neill et al.,
2012). The externally open spiracle (e.g. elasmobranchs) is a
ventilation outlet of the pouch. A pseudobranch develops
where the endodermal hyomandibular pouch meets or
approaches the ectodermal cleft (Wright, 1885; Laurent
& Dunel-Erb, 1984). In addition, the hyoid arch skeleton
suspends the jaw (Fig. 3A, B). Cyclostomes have none of
these structures.
These differences between cyclostomes and crown
gnathostomes reflect ventilation mechanisms. Non-
tetrapod vertebrates regulate respiratory flow by buccal and
parabranchial pumping (Hughes, 1960b). In crown gnathos-
tomes, the branchial bars medial to the gills are essential to
such dual-pump ventilation (Fig. 14C) because they set apart
the suction and pressure pumps (Hughes, 1960a; Brainerd
& Ferry-Graham, 2005). In addition, the medially located
branchial bars close in the pressure-reversal stage of respira-
tion (Fig. 14D; Summers & Ferry-Graham, 2002).
By contrast, cyclostomes maintain a dual-pump system
using the velum. The velum (i)pumpsrespiratoryflow
(hagfish and larval lampreys) and (ii) functions as a valve
to prevent backflow due to pressure reversal (cyclostomes
at all life stages) (Fig. 14E; Strahan, 1958; Mallatt, 1981).
Once water is pumped in from the oral cavity, constriction
of the basket-like pharyngeal skeleton lateral to the gills
pushes water out in lampreys (Fig. 4B; Mallatt, 1984a). In
hagfish, the muscular gill pouches and sphincters provide
supplementary pumping (Strahan, 1958). These respiratory
mechanisms in cyclostomes explain the lateral position of the
pharyngeal skeleton (Fig. 4A, B). In summary, the cyclostome
and crown gnathostome hyomandibular structures are
mutually exclusive phenotypes because they are functionally
redundant and topographically overlap.
(d)Fossil inferences for the hyoid interface
A velum-like water-pumping structure has been inferred
to sit in the prebranchial cavity at the mandibular–hyoid
boundary in osteostracans and tentatively in galeaspids
(Fig. 5 Janvier, 1981, 1985a,b;ch.6,Figs11Gaiet al.,
2011). Based on internal casts, the facial nerve innervated
Biological Reviews 91 (2016) 611– 657 ©2015 Cambridge Philosophical Society

630 Tetsuto Miyashita
Fig. 11. Phylogenetic tree of jawless and jawed vertebrates and distribution of characters used to determine degree of confinement
of the mandibular musculoskeletal system. Partial confinement occurred in independent lineages of jawless vertebrates, but only
jawed vertebrates have the full state of mandibular confinement for each interface. The phylogenetic tree is a consensus of the
current literature (Donoghue, Forey & Aldridge, 2000; Gess, Coates & Rubidge, 2006; Heimberg et al.,2010;Sansomet al., 2010;
Turner et al., 2010; Conway Morris & Caron, 2014) and as presented by Janvier (2007, 2008, 2010). For each character, the
first state in parentheses is plesiomorphic (represented by a red box) and the second state is apomorphic (represented by a blue
or green box). The most parsimonious character transformations were mapped onto the tree with a colour code matching the
character state in the branch. These changes are either unambiguous or optimized under ACCTRAN (accelerated transformation)
and DELTRAN (delayed transformation) models. To predict these changes, branches of uncertain affinities were collapsed into
polytomy at the least inclusive node that could nest that branch within (Pituriaspida in polytomy with Galeaspida and the clade
Osteostraci +Gnathostomata; Conodonta and Euphaneropidae in polytomy with Cyclotomata and the total group Gnathostomata).
Myllokunmingiids were treated as a lineage of stem vertebrates.
Biological Reviews 91 (2016) 611– 657 ©2015 Cambridge Philosophical Society

Fishing for jaws 631
Fig. 12. Features contributing to the premandibular –mandibular interface in various jawless vertebrates. (A) Nasal aperture
(nasohypophyseal opening) in adult lamprey (La. fluviatilis) in dorsal view. (B) The nasohypophyseal valve (boundary between the
nasal passage and the hypophyseal canal) in adult lamprey in dorsal view. (C) Sagittal section of the nasohypophyseal system of
an adult lamprey in lateral view, with nasal passage and hypophyseal canal differentially shaded. (D) Nasohypophyseal region of
the osteostracan Scolenaspis in dorsal view. (E) Sagittal section of the skull of Scolenaspis with nasal passage and hypophyseal canal
differentially shaded. (F) Schematic drawing of sagittal section of the skull of the galeaspid, Shuyu zheijiangensis.(G)HeadofShuyu with
soft tissue reconstruction in dorsal view. (H) Nasohypophyseal region of Shuyu reconstructed based on synchrotron microtomography
in dorsal view. (A–E) Modified after Janvier (2007); (F –H) modified after Gai et al. (2011). (A – F) Not to scale. (G, H) Scale
bar =2mm.
the anterior half of the first branchial cavity and extended
anteriorly up to the prebranchial ridge (Figs 5 and 12G, H;
W¨
angsj¨
o, 1952; Janvier, 1996; Gai et al., 2011). Therefore,
the posterior part of the prebranchial cavity marks the
mandibular–hyoid boundary.
In galeaspids and osteostracans, the tissue filling the
prebranchial cavity extended within the pharynx on the
medial side of the hyoid region, just like the velum in
cyclostomes. Some ostesotracans such as Dartmuthia have
preserved the dorsally facing attachment area for an
internal skeleton in this cavity, as expected if a velum was
present (Fig. 15C). This area of the prebranchial cavity was
drained by the marginal vein, which has topographical
correspondence to the anterior cardinal vein (Fig. 16D;
Janvier, 1981, 1985a). Galeaspids also have a set of correlates
for the velum: (i) endoskeletal attachment; (ii)avenoussinus
(lateral head vein) to the medial side; and (iii) multiple
depressions in between (Fig. 15D; Janvier, 1984; Gai et al.,
2011). These impressions are each consistent with cyclostome
counterparts: (i)avelarskeleton;(ii) a vein that drains from
the velar sinus to the anterior cardinal vein; and (iii)a
velar sinus and attachments for velar flexor muscles (Cole,
1926; Marinelli & Strenger, 1954, 1956; Tsuneki & Koshida,
1993; Miyashita, 2012). In both galeaspids and osteostracans,
the prebranchial cavity received innervation from the third
branch of the trigeminal nerve (Figs 5 and 12G, H). In
cyclostomes, this nerve innervates the velum (Lindstr¨
om,
1949).
The overall skeletal morphology is also consistent with the
presence of the velum. In those jawless stem gnathostomes
with rigid head shields and dermal scales, a simple
constriction of the branchial basket – like that in the
soft-bodied adult lamprey – alone would not have been
able to generate sufficient flow or control backflow caused by
pressure reversal. Although it requires additional data to test
whether or not the prebranchial structure was really a velum
proper, the current evidence supports a ventilation structure
derived from the mandibular arch within this cavity.
Jawless stem gnathostomes have other correlates of
cyclostome-like ventilation. As present in cyclostomes and
Biological Reviews 91 (2016) 611– 657 ©2015 Cambridge Philosophical Society

632 Tetsuto Miyashita
Fig. 13. Paired nasal apertures in the jawless arandaspid,
Sacabambaspis janvieri. (A) Reconstruction of Sacabambaspis in
anterior view, with the anatomical right side showing the
life reconstruction and with the left side showing the fossil
morphology (modified from Janvier, 1996). (B) Head region of
a cast of Sacabambaspis [MNHN (Mus´
eum National d’Histoire
Naturelle, Paris) 1005] in dorsal view. The median nasal plate
is lightly shaded with white and outlined with a dark line.
unambiguously preserved in euphaneropids (Fig. 15A, B;
Janvier et al., 2006), the branchial bars forming a basket
lateral to the gills are a symplesiomorphic condition for
jawless stem gnathostomes (ch. 7, Fig. 11). Where branchial
bar attachments are preserved, the individual branchial bars
formed lateral to the gill chambers (Figs 5 and 15C, D;
Janvier, 1993, 1996, 2004) with the possible exception of
Metaspiriggina (Conway Morris & Caron, 2014). It remains
ambiguous whether or not the branchial bars extended
dorsally in galeaspids, heterostracans, and osteostracans, but
they would have extended at least to the lateroventral side
(Stensi¨
o, 1964; Janvier, 1985a,2004).
Nevertheless, some stem gnathostomes have
crown-gnathostome-like conditions at this interface.
An adorbital opening occurs at a hyomandibular position
(behind the orbit but anterolateral to the otic capsule;
anterior to gill impressions) in some heterostracans and
pituriaspids (Fig. 15E; Halstead, 1971; Young, 1991). A
slit-like small ‘spiracle’ opens in a hyomandibular position
among some antiarch placoderms – the most stemward
jawed vertebrates (Young & Zhang, 1992; Arsenault et al.,
2004). Such an opening may well have existed in other pla-
coderms, but it would be difficult to distinguish it from a gap
between dermal elements. These openings likely represent
an outlet of the underlying structure, but they would spatially
conflict if mandibular elements were not restricted anterior to
this position.
More attributes of the hyoid interface are present in jawed
stem gnathostomes. From placoderms to crown gnathos-
tomes, configurations of the trigeminal and facial nerves,
major vessels, and the neurocranium–hyomandibular
contact remain consistent with respect to each other (Young,
1979, 1986; Goujet, 1984; Brazeau & Friedman, 2014;
Dupret et al.,2014;Gileset al., 2015). In the absence of
unambiguous osteological correlates, the presence of a
spiracular organ and pseudobranch cannot be precisely
evaluated. However, a groove for the ‘efferent pseudo-
branchial artery’ identified in jawed stem gnathostomes
and stem osteichthyans shows a configuration largely
consistent with a hypothetical pseudobranch (ch. 9,
Fig. 11; Fig. 3E, F; Goodrich, 1930; Gardiner, 1984;
Basden & Young, 2001). The vessel passed between
the palatoquadrate and hyomandibula (or between the
two palatoquadrate–neurocranium contacts close to this
boundary) from the suborbital shelf medially toward
the internal carotid foramen (Young, 1979, 1980, 1986;
Brazeau, 2009; Davis et al., 2012). In comparison, jawless
stem gnathostomes do not share such morphology. The
efferent artery of the mandibular arch branched from the
single dorsal aorta anterior to the prebranchial ridge in
osteostracans (Janvier, 1981, 1985a). In galeaspids, the
most anterior artery identified for the pharyngeal pouch
derivative is in the hyoid region, whereas the internal
carotid gives rise to the ophthalmic artery anterior to the
hypophyseal region (Fig. 12G, H; Gai et al.,2011).
Therefore, full attributes of the crown gnathostome
mandibular–hyoid interface cannot be extended beyond
the node of jawed vertebrates. Correlates for a ventilation
structure in the prebranchial region of galeaspids
and osteostracans indicate a cyclostome-like mandibular
patterning in this area (ch. 6, Fig. 11). Parsimoniously, the
adorbital opening in some heterostracans and pituriaspids
is likely independent of the spiracle in crown gnathostomes,
regardless of whether or not the position is hyomandibular
(ch. 8, Fig. 11).
(e)The interface with the hypobranchial musculature in living
vertebrates
In crown gnathostomes, the hypobranchial muscles
(including a tongue and jaw depressors) connect the lower
jaw and the pectoral girdle. Their progenitors originate in
anterior myomeres above the pharyngeal series, migrate
behind the pharyngeal arches, and extend anteriorly below
the pharynx at a level deeper than the branchial constrictors
(Lours-Calet et al.,2014),excludingmandibularelements
from this region. Thus, the interface with the hypobranchial
musculature forms mid-ventrally below the pharyngeal
floor. This region lies ventro-/postero-medial to the ventral
expression zone of the Edn signalling, which is upstream of
the nested Dlx expressions (Fig. 10B; Medeiros & Crump,
2012).
In cyclostomes, a lingual apparatus occupies this space as
amid-ventralextensionofthemandibularmusculoskeletal
system (Fig. 4A, B). This feeding apparatus is housed inside
Biological Reviews 91 (2016) 611– 657 ©2015 Cambridge Philosophical Society

Fishing for jaws 633
Fig. 14. The mandibular –hyoid interface and ventilation in vertebrates. (A) Longitudinal section of a small spotted catshark embryo
(Sc. canicula)atstage31,showingthespiracularorganatthedorsalpartofthehyomandibularpouch(fromO’Neillet al., 2012).
Scale bar =100 µm. (B) Schematic drawing of a cross section of the catshark embryo at the level of the spiracular organ, showing
its position within the pouch (modified from O’Neill et al., 2012). (C) Schematic drawing of ventilation in fish based on Hughes’s
(1960a,b) dual-pump model (modified from Brainerd & Ferry-Graham, 2005). The buccal and parabranchial (opercular) cavities
are separated by the branchial bars and lamellae, and each acts as a pump. (D) Example of dual-pump ventilation in the dogfish
(modified from Summers & Ferry-Graham, 2002; Brainerd & Ferry-Graham, 2005). For (C) and (D), colour indicates pressure level.
(E) Hagfish (Myxine glutinosa)asanexampleofcyclostomeventilationusingthevelum(pumpingandvalvebetweenthebuccaland
parabranchial cavities) and constriction of the branchial bars (modified from Strahan, 1958). In hagfish, the velum (shaded in red)
acts both as a buccal pump and a valve in a single stroke. (F) Pre-hatching hagfish embryo (Ep. stoutii)inrightlateralview,showing
the plane of section for (G) (from Dean, 1899). (G) Cross section (indicated in F) of a pre-hatching hagfish embryo (Ep. stoutii;
Dean-Conel collections, specimen number 2343; Museum of Comparative Zoology, Harvard University). This particular section
shows that the anlage of the velum forms with the mandibular neural crest cells (NCCs) extending into the hyomandibular pouch
(where a spiracular organ would develop in a gnathostome embryo) and eventually into the pharynx. The histological section also
shows the mesenchymal condensation for the cartilage of the velum and an anlage for the cardinal heart, a venous sinus between
the velar and facial cartilage. Anatomical terms with a prefix (a-) indicate anlagen; tissues have not differentiated yet. Abbreviations:
hyomand., hyomandibular; nc, notochord.
the pharyngeal skeleton, and its muscles protract and retract
keratinous tooth plates (Yalden, 1985). The apparatus is
anchored to the posterior end of the pharyngeal skeleton.
Cyclostomes do have the somatic muscles conventionally
called the hypobranchials, innervated by the somatic spinal
nerves as in the gnathostome hypobranchials (Kusakabe
& Kuratani, 2005, 2007). These muscles in cyclostomes,
however, neither delimit the mandibular structures nor
compare readily to the gnathostome counterpart. The
cyclostome hypobranchials differ in (i) drawing their
progenitors broadly from myomeres including those well
posterior to the pharyngeal series, (ii) differentiating external
to the pharyngeal skeleton, and (iii) wrapping around the
ventral side of the head as the most superficial muscle layer
Biological Reviews 91 (2016) 611– 657 ©2015 Cambridge Philosophical Society

634 Tetsuto Miyashita
Fig. 15. Features contributing to the mandibular – hyoid interface in various jawless stem gnathostomes. (A) The euphaneropid
Euphanerops logaevus in left lateral view, showing the lamprey-like branchial basket (MHNM 01-02). (B) Reconstruction of Euphanerops
in lateral view. (A, B) From Janvier & Arsenault (2007). (C) Right side of the head of the osteostracan Dartmuthia gemmifera in
ventral view [AMNH (American Museum of Natural History) Patten Collections 38.71.8750; modified from Janvier, 1985b] with
sketches of sections at positions indicated by the thick lines, showing the attachment site with a suture for the velar skeleton in the
hyomandibular position. (D) Right half of the head shield of the galeaspid Duyunaspis paoyangensis in ventral view (modified from
Janvier, 1984), showing the attachment site for endoskeletal elements in the hyomandibular position and associated depressions
that presumably represent muscular attachment and a sinus connected with the lateral head vein. (E) Ventilation aperture at the
hyomandibular position in the amphiaspid heterostracan Gabreyaspis tarda in dorsal view (modified from Halstead, 1971), with the
left half showing life reconstruction and the right half showing the fossil morphology.
(Kusakabe, Kuraku & Kuratani, 2011; Miyashita, 2012; Oisi
et al.,2015).
Despite different developmental origins, the mandibular
lingual apparatus in cyclostomes and the somatic hypo-
branchial muscles in gnathostomes are topographically and
functionally similar and cannot coexist in the pharyngeal
floor. In addition, three conditions are required to allow the
somatic muscles in this region. (i) The body wall musculature
must be differentiated into the epaxial and hypaxial groups, as
the gnathostome hypobranchials belong to the latter. (ii)The
pectoral girdle must be detached from the skull, because the
hypobranchials originate from the ventral pectoral girdle. A
free pectoral girdle also assumes the presence of muscles that
support the neck, the cucullaris and levator capitis (Fig. 16E;
Kuratani, 2008; Ericsson, Knight & Johanson, 2013). (iii)The
branchial bars must form on the medial side of the gills. The
laterally positioned pharyngeal skeleton precludes somatic
muscles from the pharyngeal floor at levels deeper than the
body wall. If the main skeletal support is to the medial side,
however, the pericardial entrance opens for somatic longi-
tudinal muscles to migrate from posterior (Matsuoka et al.,
2005). This latter configuration would limit a mandibular
Biological Reviews 91 (2016) 611– 657 ©2015 Cambridge Philosophical Society

Fishing for jaws 635
Fig. 16. Characters for the mandibular – hypobranchial interface in jawless and jawed stem gnathostomes. (A) Infilling of the
mid-ventral pharyngeal groove extending from the mandibular region in a cast of the dorsal surface of the ventral shield of a
cyathaspidiform heterostracan. The anatomical right side (left side of the illustration) is shaded according to pharyngeal arch
identities predicted based on the pattern common to all known vertebrates in which the most anterior gill lamellae represents a hyoid
hemibranch. The groove originates between the predicted left and right mandibular domains. (B) Reconstruction of the heterostracan
Protopteraspis gosseleti in ventral view, showing oral plates in position. (C, D) Postbranchial wall and associated osteological correlates
of soft tissues in the osteostracan Norselaspis in sagittal section, the right half in medial view (C) and in horizontal section, the
dorsal half in ventral view (D). The somatic musculature did not extend anteriorly beyond the postbranchial wall in osteostracans.
Therefore, gnathostome-like hypobranchial muscles did not exist in this immediate outgroup to gnathostomes. (E) Reconstructed
jaw mechanics in the arthrodire placoderm Coccosteus with two pairs of antagonising muscles and with reconstructed endoskeleton in
pink. (F) The arthrodire placoderm Compagopiscis croucheri in left lateral view (original image horizontally flipped for consistency) with
reconstruction of the hypobranchial muscles and a box indicating portions illustrated in (G–I). (G) Pectoral girdle of Compagopiscis
(position indicated in F) reconstructed from synchrotron microtomography, with sections showing internal microstructures in right
lateral view. (H) Reconstruction of Sharpey’s fibres, the connective tissue at the attachment site for the hypobranchial muscles, in the
position indicated in (G) and based on synchrotron microtomography. (I) Reconstructed hypobranchial muscles (blue arrows) from
the Sharpey’s fibres. (A, B) Modified from Janvier (1996). (C, D) Modified from Janvier (1981). (E) Modified from Miles (1969) with
input from Trinajstic et al. (2013). (F) Modified from Trinajstic & Hazelton (2007). (G –I) From Sanchez et al. (2013). Colour codes
for (A), (C) and (D) follow Fig. 6, except for: grey =mid-ventral pharyngeal groove; dark blue =spinal nerves; pink =circulatory
system. (E) Not to scale.
structure from extending mid-ventrally. These characters,
therefore, correlate with the hypobranchial interface.
As for the epaxial–hypaxial differentiation, paired fins
are a useful indicator. In gnathostomes, paired appendages
with mesodermal contributions (muscles and endoskeleton)
require (i) primaxial–abaxial differentiations bordered by
the lateral somitic frontier and (ii) dorsoventral compartmen-
talization within the abaxial domain and in the ectoderm
(Nowicki, Takimoto & Burke, 2003; Johanson, 2010; Oni-
maru et al.,2011;Tulenkoet al.,2013;delaRosa,M¨uller &
Biological Reviews 91 (2016) 611– 657 ©2015 Cambridge Philosophical Society

636 Tetsuto Miyashita
Metscher, 2014). These polarities in embryonic tissues largely
correspond to the adult epaxial–hypaxial distinction (epax-
ial: mainly derived primaxially; hypaxial: mainly derived
abaxially). These requirements for paired appendages
appear broadly conserved across crown gnathostomes.
(f)Fossil inferences for the hypobranchial interface
Neither the lingual apparatus nor the hypobranchial mus-
culature has hard parts amenable for fossilization, but their
correlates are occasionally preserved. A rare exception is
euphaneropids that have a mineralized mid-ventral skeletal
element extending into the branchial basket (ch. 10, Fig. 11;
Fig. 15A, B). This elongate element likely consists of calcified
cartilage and corresponds to the cyclostome lingual cartilage
in morphology and position (Janvier & Arsenault, 2007). The
element is associated with the ‘annular’ cartilage anterior to
the first full branchial bar, indicating its mandibular identity.
As for correlates of the lingual apparatus, osteostracans and
possibly pituriaspids have the postbranchial wall to close the
oropharyngeal region posteriorly with only several perfo-
rations required for the oesophagus and circulatory system
(Fig. 16C, D; Janvier, 1981, 1985a; Young, 1991). Not
only would the postbranchial wall have precluded somatic
muscles from the pharyngeal region, the first set of spinal
nerves extended to the brachial plexus without any evidence
of hypobranchial innervation (Janvier, 1981, 1985a). There-
fore, the ventral pharyngeal structures in osteostracans – in
whatever form required for feeding and ventilation (examples
in Stensi¨
o, 1964; Janvier, 1985a)–couldonlydevelopfrom
one or more of the pharyngeal arches as in cyclostomes.
Outside osteostracans, heterostracans have a longitudinal
groove from the oral region posteriorly on the ventral
shield (Fig. 16A), typically interpreted as an impression of
an endostyle (Halstead Tarlo & Whitting, 1965; Halstead,
1973). Because the groove originates near the mouth anterior
to the branchial region, this unknown tissue likely extended
from the mandibular region like a lingual apparatus, rather
than housing an endostyle that is more posteriorly positioned
in larval lampreys and cephalochordates. In parallel with
this, similarities in nasohypophyseal morphology between
heterostracans and hagfish led to reconstruction of a lingual
apparatus in the former taxa (Stensi¨
o, 1964; Janvier,
1974). Assuming little or no soft tissue cover, the lack
of significant wear at the tip of oral scales raises doubts
about whether such a feeding apparatus in heterostracans
permitted macrophagy (Fig. 16B; Purnell, 2002). Even if
heterostracans were microphagous, however, the oral scales
may not have been held rigid (Elliott, 2013). Beyond het-
erostracans, the predicted motion of the S and M conodont
elements implies skeletal base and mid-ventral protractors
and retractors in a way functionally similar to the cyclostome
lingual apparatus, despite the lack of correspondence in
tissue types (Purnell & Donoghue, 1997; Goudemand et al.,
2011). These correlates imply a broadly defined mid-ventral
mandibular structure across jawless stem gnathostomes.
Prerequisite conditions for the development of the
hypobranchial musculature occur variably among jawless
stem gnathostomes. The presence of paired fins in many stem
gnathostome lineages (ch. 11, Fig. 11) likely reflects some
degree of epaxial–hypaxial differentiation. Generally, these
fins develop above or behind the branchial series and/or
along the mid-ventral margin in anaspids, euphaneropids,
and thelodonts (Coates & Cohn, 1998; Coates, 2003; Wilson,
Hanke & M¨
arss, 2007; Sansom, Gabbott & Purnell, 2013).
The definitive pectoral fins with endoskeletal attachment
occur in many osteostracans and likely in pituriaspids
(Janvier, Arsenault & Desbiens, 2004). In none of these
fishes are the pectoral fins or the girdles detached from the
head. However, the epicercal tail in osteostracans (Janvier,
2007) suggests that the hypaxials are better differentiated
in this lineage than in other jawless stem gnathostomes to
support the caudal fin lobe along the ventral margin of
the tail.
The earliest definitive evidence for crown-
gnathostome-like hypobranchial morphology occurs
in placoderms. Arthrodire placoderms have osteological
correlates for well-developed hypobranchial musculature as
well as neck and fin muscles (ch. 12, 13, Fig. 11; Fig. 16E –I;
Miles, 1969; Johanson, 2003; Sanchez et al.,2013;Trinajstic
et al., 2013). In addition, the pectoral girdle is decoupled
from the head in all placoderms. This decoupling would
have allowed the neck muscles, including the levator
capitis and cucullaris, to elevate and depress the head
(Miles & Westoll, 1968). Although the epaxial–hypaxial
differentiation predates gnathostomes, the hypobranchial
muscles likely assumed their mid-ventral position below the
pharynx right around the origin of the jaw.
(3) The origin of jaws via mandibular confinement
A review of the fossil evidence indicates that a full set of
attributes of a spatially confined mandibular arch only
occurs in jawed vertebrates, whereas various jawless stem
gnathostomes developed some of the conditions (Fig. 11).
Based on this character distribution, the Mandibular Con-
finement (MC) Hypothesis proposes that premandibular,
hyoid, and hypobranchial structures spatially confined the
mandibular derivatives at respective interfaces. As a result,
the mandibular musculoskeletal system transitioned from
the highly expanded and differentiated state – like that seen
in cyclostomes – to its more restricted position in crown
gnathostomes. Mandibular confinement is necessary for a
jaw to evolve because progenitors in the mandibular arch
previously committed to pre-existing feeding and ventilation
structures could more easily acquire novel developmental
fates. Four stages are proposed (Fig. 17). Stage I: the earliest
common ancestor among crown gnathostomes, osteostra-
cans, and galeaspids but with exclusion of thelodonts and
heterostracans. Stage II: the earliest common ancestor of
all jawed vertebrates with exclusion of osteostracans. Stage
III: the earliest common ancestor of all jawed vertebrates in
exclusion of Stage II. Stage IV: t7he last common ancestor of
all jawed vertebrates. Stages I, II, and IV are reconstructed
from character states inferred for that or the closest nodes
(Fig. 11). Stages I–III are not set on a particular node.
Biological Reviews 91 (2016) 611– 657 ©2015 Cambridge Philosophical Society

Fishing for jaws 637
Fig. 17. A phylogenetic scheme for the Mandibular
Confinement Hypothesis, with each stage of mandibular
confinement (MC) between jawless stem gnathostomes
(galeaspids, pituriaspids, and osteostracans) and jawed stem
gnathostomes (placoderms) (Roman numerals, see main text
for details). Acanthodians are likely paraphyletic to crown
chondrichthyans, and also possibly to crown osteichthyans.
Relationships are based on consensus from recent literature
(Brazeau, 2009; Davis et al.,2012;Zhuet al.,2013;Brazeau
& Friedman, 2014; Dupret et al., 2014; Long et al., 2014;
Giles et al., 2015). MC Stage II may have preceded the
last common ancestor between osteostracans and crown
gnathostomes, if the nasohypophyseal split arose in the last
common ancestor between galeaspids and crown gnathostomes
(Gai et al., 2011; Dupret et al., 2014). However, osteostracans
have been consistently placed as the immediate outgroup
to jawed vertebrates (see Fig. 11). Under that topology, two
equally parsimonious interpretations exist: either osteostracans
independently had a reversal to the cyclostome-like condition or
galeaspids independently acquired the crown-gnathostome-like
condition.
For Stages I and II, this is because two different scenarios
take the same number of steps to explain the origin of
the crown-gnathostome-like premandibular phenotypes
in galeaspids (either at the last common ancestor with
crown gnathostomes or independently). Stage III represents
characters constrained in that segment of the stem that are
not explained by any other stages.
(a)Mandibular Confinement Stage I
Based on successive outgroup lineages to jawed ver-
tebrates, a likely ancestor hypothesized for Stage I is
an osteostracan/galeaspid-like jawless stem gnathosome
(Figs 17 and 18A). This stage exhibits the following condi-
tions: (i) an upper lip supporting a single nasohypophyseal
system (nasal capsules may have been paired); (ii)
velum-like hyomandibular ventilation structure; and (iii)
lingual-apparatus-like mid-ventral feeding structure. As in
galeaspids, heterostracans, and osteostracans, the branchial
bars lay lateroventral, and possibly laterodorsal, to the gills,
and were joined at the medial ends for structural support
(W¨
angsj¨
o, 1952; Stensi¨
o, 1964; Janvier, 1985a,2004;Gai
et al.,2011).
(b)Mandibular Confinement stage II
In Stage II, the mandibular structures were delimited
along the premandibular interface (Figs 17 and 18B).
This stage predicts an ancestor similar to galeaspids. A
tripartite split of the nasohypophyseal placode allowed the
adenohypophyseal portion to migrate posteriorly at the floor
of the forebrain away from the olfactory portions during
cephalic flexure, as in crown gnathostomes (Kawamura
et al.,2002;Herzoget al., 2003, 2004; Uchida et al.,2003).
The adenohypophyseal placode now sat within the oral
cavity (Rathke’s pouch). The preoptic and postoptic streams
of the trigeminal NCCs formed trabecular and ethmoidal
derivatives around these placode-derived structures. Shh
signalling from Rathke’s pouch ensured this proliferation
(Creuzet et al.,2004;Britoet al.,2006;Khonsariet al.,2013).
These streams of NCCs were no longer accompanied by
the MMCs, as the nasohypophyseal skeletal rods do not
assume muscular movement. This skeletal differentiation
freed the trigeminal NCCs in the premandibular region from
patterning muscular lip structures. Instead, a population
of mandibular trigeminal NCCs and MMCs secondarily
extended anteriorly to form a maxillary process. Restriction
of Dlx expression to the mandibular stream – perhaps via
acquisition of mesenchymal Bmp2 and Bmp4 expressions in
the distal domain of the maxillary process (Lee et al.,2001,
2004; He et al., 2014) –reinforced this separation of the
MMC- and non-MMC-associated trigeminal NCCs.
The differentiated premandibular skeletal elements
(trabecula-derived elements and oral hypophyseal opening)
mark this interface. Increased sensitivity of olfaction via
increased surface area may have driven this transition, as
the nasal septum bilaterally organizes nasal capsules. The
septum may also have reinforced the palatal skeleton by
acting as a strut between the palate and skull roof.
(c)Mandibular Confinement Stage III
The principal advances in Stage III were acquisition
of a joint within the mandibular arch – an alternative
pumping mechanism that rendered the hyomandibular
ventilation structure redundant – and establishment of the
crown-gnathostome-like hyoid interface (Figs 17 and 18C).
The dorsoventrally patterned mandibular skeleton deployed
Edn1-mediated focal expression of Nkx3.2 and Trps1 to
specify a joint within the cartilaginous endoskeleton of
the mandibular arch, as in crown gnathostomes (Fig. 10B;
Miller et al.,2003;Talbot,Johnson&Kimmel,2010).
This event was likely coupled with restriction of Barx1
Biological Reviews 91 (2016) 611– 657 ©2015 Cambridge Philosophical Society

638 Tetsuto Miyashita
Fig. 18. Legend on next page.
Biological Reviews 91 (2016) 611– 657 ©2015 Cambridge Philosophical Society

Fishing for jaws 639
expression to a more ventral region within the mandibular
arch. In crown gnathostomes, Barx1 (downstream of
Hand2) inhibits joint formation in the ventral part of the
mandibular arch (Nichols et al., 2013). In the lamprey
P. marinus,however,Barx1 is more broadly expressed in
intermediate and ventral zones of the mandibular arch
(Cerny et al.,2010).
With the advent of this presumptive jaw joint, the
cartilaginous mandibular skeleton became highly mobile.
Via perichondral ossification, the cartilaginous endoskeleton
became associated with the dermal elements around the
mouth. This enabled buccal pumping and led to reduction of
the velum-like hyomandibular structure. With the reduction,
NCCs on the medial side of the branchial arches gained Dlx
expression, and the main site of chondrogenesis shifted to
that side. The cartilages were dorsoventrally patterned by
the Dlx code with an epibranchial–ceratobranchial hinge to
prevent flow reversal, giving rise to gnathostome ventilation
driven by buccal–parabranchial pumping. The reduction of
the hyomandibular ventilation structure also allowed: (i)a
spiracular organ and pseudobranch to develop at the wall of
the hyomandibular pouch; and (ii) the hyoid arch cartilage
to attach to the neurocranium dorsally and the mandibular
skeleton ventrally (not illustrated).
(d)Mandibular Confinement Stage IV
In Stage IV, the mid-ventral interface with the hypobranchial
musculature was established (Figs 17 and 18D). With the
medial shift of the branchial bars, the mid-ventral mandibu-
lar structure could no longer be housed and anchored within
a basket of cartilages and became reduced. This shift allowed
the hypobranchial myoblasts arising in the lower part of ante-
rior somites to migrate in this region. The somatic myoblast
migration into the pharyngeal floor completed spatial
confinement of mandibular structures. The hypobranchial
muscles extended from the pectoral girdle to the lower jaw.
The pectoral girdle was independent of the head, which
allowed the neck musculature to evolve in tandem with elabo-
ration of the hypobranchial system (not illustrated in Fig. 18).
This newly evolved jaw mechanics increased suction volume
and food size (Miles, 1969). At later stages, acquisition(s) of
dentition permitted macrophagous carnivory.
With all of the boundaries for the mandibular elements
now established, the jaw appeared as a highly integrated
feeding and ventilation apparatus. The NCCs and
MMCs that would have differentiated into cyclostome-like
mandibular structures became remodelled as progenitors
for the jaw apparatus. Therefore, the MC Hypothesis does
not propose any specific evolutionary precursor of the jaw
in the cyclostome skull.
(e)Integration of previous jaw origin hypotheses
The MC Hypotheis draws on elements from earlier
hypotheses, but it reconciles previously incompatible
predictions. (i) It incorporates a central premise of the
Heterotopy Hypothesis – altered distribution of NCCs that
migrate with mandibular MMCs in the premandibular
region – without assuming its direct causal link to the
jaw. (ii) It follows the Co-option Hypothesis in explaining
the origin of the jaw by acquisition of a joint, but it
meets the challenge of incorporating fossil phenotypes and
providing functional context. (iii) It incorporates the Velum
Hypothesis in reconstructing a velum in a jawless ancestor,
but avoids the difficulty of treating the velum as a direct
precursor of the jaw. (iv) It draws from the Lateromedial
Shift Hypothesis to explain the position of the branchial
Fig. 18. The Mandibular Confinement Hypothesis illustrated, showing successive confinement of the mandibular arch over four
stages. Transition from one stage to another describes a change in the development of the pharyngeal region in an embryo (left,
lateral view and ventral view) and a fully differentiated state in an adult for a functional explanation (middle, dorsal view; right,
lateral view). Embryos in ventral view depict the early migratory (left, anatomical right side) and the late post-migratory (right,
anatomical left side) phase of neural crest cells (NCCs). Both embryos and adults are diagrams of character states predicted for
this transition. Anatomical terms inside apostrophes reflect functional and topographical similarities and are used for convenience.
‘Velum’ is used for a predicted hyomandibular ventilation structure similar to the cyclostome velum, whereas ‘lingual apparatus’
is for a predicted mid-ventral pharyngeal structure extending from the mandibular arch like the cyclostome lingual apparatus.
Facial cartilage is a general term for chondral endoskeleton that occurs anterior and ventral to the parachordal cartilage (skeletal
rod between the polar cartilage/acrochordal commissure and the otic capsule) in cyclostomes and gnathostomes. (A) Mandibular
confinement (MC) Stage I. The NCCs and mesodermal mesenchyme (MMCs) of the mandibular arch are not confined. (B) MC
Stage II with the crown-gnathostome-like premandibular interface. The split of the nasohypophyseal placode led to the loss of
cyclostome-like upper lip and the gain of gnathostome-like trabecula. (C) MC Stage III with the crown-gnathostome-like hyoid
interface. The loss of hyomandibular extension of the NCCs and MMCs of the mandibular arch for a ventilation structure (velum)
led to: (i) hyomandibular sensory structures (spiracular organ, pseudobranch); (ii) a shift of the skeletal support of the gills from
the lateral to medial side; and (iii) a new mechanism for the buccal pump using the cranial endoskeleton. (D) MC Stage IV with
the crown-gnathostome-like hypobranchial interface, showing a fully developed jaw apparatus. The medial shift of the branchial
bars allowed the hypobranchial muscles to extend along the pharynx towards the jaw. By this point, the mandibular NCCs
were spatially confined into the position seen among gnathostome embryos, and all the mandibular cell populations that would
otherwise have differentiated into specialized feeding and ventilation structures were released from such fates and remodelled into
the jaw apparatus. Abbreviations: a-velum, anlage of velum-like hyomandibular ventilation structure; CNS, central nervous system;
hyomand., hyomandibular; lep, lens placode; man, mandibular stream of trigeminal NCCs +mandibular MMCs; nc, notochord;
poo, postoptic stream of trigeminal NCCs; pro, preoptic stream of trigeminal NCCs; tri, trigeminal.
Biological Reviews 91 (2016) 611– 657 ©2015 Cambridge Philosophical Society

640 Tetsuto Miyashita
Fig. 19. Jaw defects as a result of disruption of the premandibular interface. (A) Fate map of post-migratory trigeminal neural crest
cells (NCCs) (coloured for skeletogenic fates in B) superimposed onto a sketch of wildtype zebrafish embryo [24 h post fertilization
(hpf)]. The NCCs are segregated spatially for their skeletogenic fates, and a hypophyseal canal sits at the boundary between the
premandibular and mandibular subpopulations. (B) Zebrafish chondrocranium coloured according to fates of the trigeminal NCCs
that correspond to (A). (C, D) Alcian-blue-stained nocm579 mutant zebrafish [5 days post fertilisation (dpf)] in lateral and ventral
views. The ethmoidal plate and trabecula (*) are absent, and the Meckel’s cartilage is rotated dorsally to overlap the snout in a
manner similar to the cyclostome upper lip. The normal skeletal phenotype is shown in Fig. 4D. (E) Chondrogenic condensations
of trigeminal NCCs in a sox9a RNA-stained zebrafish (48 hpf) in dorsal view. (F) Lack of chondrogenic NCC condensations in the
anterior neurocranium of a sox9a RNA-stained smo-zebrafish (48 hpf). Hedgehog signalling is inhibited in this mutant, and the NCCs
in the premandibular domain fail to proliferate. Same scale as (E). (G) Alcian-blue-stained shh-zebrafish in dorsal view, showing
a fusion between the left and right palatoquadrates at the pterygoid processes in the absence of the premandibular cartilages.
This phenotype is similar to the cyclostome chondrocranium in which the palatal cartilages fuse to each other at the midline. (H)
Alcian-blue-stained smo-zebrafish in dorsal view, showing a failure of the NCCs to chondrify following the inhibition of hedgehog
signalling. (I) Nine-day-old [Hamburger–Hamilton (HH) stage 35] chick embryo stained with alcian blue and alizarin red and
its interpretive drawing, showing normal skull morphology. (J) Nine-day old chick embryo stained with alcian blue and alizarin
red after surgical ablation of endodermal zone 1 and its interpretive drawing, showing the reduced ethmoidal elements and the
abnormal Meckel’s cartilage. Premandibular cartilages are shaded in the interpretive drawing. (A, B) Modified from Eberhart et al.
(2006). (C, D) Provided by G¨
okhan Unlu and Ela W. Knapik (Vanderbilt University). (E–H) From Eberhart et al. (2006). (I, J) From
Benouaiche et al. (2008). Abbreviations: CNS, central nervous system; ect, ectethmoid (red); ep, ethmoidal plate; is, infraorbital
septum (green); mes, mesethmoid (blue); nc, notochord; pc, polar cartilage (derived at the anterior end of paraxial mesoderm);
ppq, pterygoid process of palatoquadrate; tr, trabecula. Colour codes for (I) and (J): blue =mesethmoid; green =interobital septum;
red =ectethmoid. (A–D, G –J) Not to scale.
bars in the context of ventilation, but it does not require
simultaneous internalization of the jaw and branchial bars.
(v) It also explains why some jawless stem gnathostomes
have crown-gnathostome-like conditions without developing
jaws: all of the premandibular, hyoid, and hypobranchial
interfaces had to be present to constrain the mandibular
arch derivatives before a jaw could emerge.
IV. TEST OF THE MANDIBULAR CONFINEMENT
HYPOTHESIS
The MC Hypothesis is constructed as a tree-based scenario
on considerable fossil and embryological evidence. For this, it
was designed to satisfy both Tests 1 and 2 (Section III). Some
of the characters used by the MC Hypothesis cannot be
Biological Reviews 91 (2016) 611– 657 ©2015 Cambridge Philosophical Society

Fishing for jaws 641
readily constrained to a single node, and exact states of these
characters remain rather vague (e.g. mid-ventral pharyngeal
structure). However, these characters were identified via
skeletal correlates in more than one lineage of jawless stem
gnathostomes (Fig. 11). Similarly, gene expression patterns
that can only be documented in living taxa were linked
to phenotypes that can be observed in fossils. Therefore,
these uncertainties did not increase the number of character
transitions from what is parsimoniously required on the
consensus tree. In addition to these standard tests, the MC
Hypothesis makes a number of falsifiable predictions.
(1) Testable predictions in the fossil record
The MC Hypothesis premises on the ancestrally distinct
mandibular patterning. Therefore, it makes two predictions
testable through the fossil record. (i) Distinct mandibular
patterning is a symplesiomorphic condition along the stem
of jawless gnathostomes. (ii) A spatially confined, serially
patterned mandibular arch is a necessary condition for a
jaw to arise. Fossil evidence that suggests otherwise would
falsify the hypothesis. Ideally, such evidence could come in
two forms nested immediately outside placoderms: (a) jawless
stem gnathostomes that have all the attributes of mandibular
confinement; or (b) jawed stem gnathostomes that have a
cyclostome-like, unconfined state of mandibular patterning.
No known taxon qualifies either condition to my knowledge.
Although the myllokunmingiid Metaspriggina appears to lack
acyclostome-likeupperlipandhaveabipartitepharyngeal
skeleton (Conway Morris & Caron, 2014), no correlates of
pharyngeal arch identities can be discerned, and its inferred
phylogenetic position as a stem vertebrate poses difficulty
in using this one taxon to reverse the character polarity
supported by multiple stem gnathostome lineages.
(2) Testable predictions in embryological
experiments
A survey of fossil and embryological evidence (Fig. 11)
indicates that a full set of phenotypic conditions for
the interfaces of the mandibular arch only occurs in
jawed gnathostomes. Although this empirical approach
corroborates a proposal of the MC Hypothesis that
ajawrequiresaspatiallyconfined,seriallypatterned
mandibular arch, the hypothesis can be tested further
through experimental embryology. Disruption of each of
the interfaces of the mandibular arch in crown gnathostome
embryos should have a negative effect on jaw formation.
In reverse, knockdown/knockout/mutant phenotypes that
result in jaw defects may be linked to one of the interfaces.
Such phenotypes may even partly resemble conditions in
cyclostomes. Challenges to such a test are the specificity
of manipulation and the dynamic nature of development.
Conspicuous defects would involve a suite of structures not
restricted to jaws. Conversely, development can compensate
for experimental manipulations. These factors may explain
why the jaw still forms, albeit mildly disrupted, when the
olfactory placode and anterior neural fold ectoderm are
either ablated or suppressed in chick embryos (Szabo-Rogers
et al.,2009;Gittonet al., 2011). Nevertheless, numerous
examples exist where altered properties of one or more
interfaces specifically led to jaw defects.
(a)Embryological inferences for the premandibular interface
Multiple phenotypes support the suggestion that failure to
form key structures properly at the premandibular interface
(hypophyseal canal; trabecula and ethmoidal-plate deriva-
tives) results in jaw defects. In nocyranom579 (noc) mutant
zebrafish, the trabecula and ethmoidal plate do not form, and
the Meckel’s cartilage shows delayed development and dorsal
rotation towards the presumptive ethmoidal plate (Fig. 19C,
D; Neuhauss et al., 1996). The rotated Meckel’s cartilage
superficially resembles cyclostomes in that a lower labial car-
tilage is upturned anteriorly onto a premandibular position
(Fig. 4A, B). This parallel does not necessarily mean that these
cartilages are evolutionarily related. Instead, the mandibular
trigeminal NCCs in noc mutants shifted over the premandibu-
lar region in the absence of pre- and postoptic NCC prolifer-
ation, precisely in a manner similar to the cyclostome upper
lip. This parallel meets the prediction of cyclostome-like
phenotypes in the absence of mandibular confinement. Like-
wise, Otx2 heterozygote mice show jaw defects ranging from
deformation to loss along with severe defects of the trabecula
and ethmoidal elements (Matsuo et al.,1995).
Experiments with chicks and zebrafish indicate a crucial
role of the hypophyseal canal primordium in the patterning
of the jaw. In this region, epithelial expression of Bmp4 and
Fgf8 specify the oral epithelium and delimit the mandibular
arch (Fig. 10A; Richman et al., 1997; Shigetani et al., 2000).
In this antagonistic signalling environment, the reciprocally
inhibitory Shh in the foregut endoderm maintains the
trigeminal NCCs of the mandibular stream. In chicks,
inhibition of Shh or grafting of the foregut therefore results
in jaw defects, and also impairs development of the nasal
and ethmoidal cartilages formed by the trigeminal NCCs
of the preoptic and postoptic streams (Fig. 19I, J; Creuzet
et al.,2004;Britoet al., 2006; Brito, Teillet & Le Douarin,
2008; Benouaiche et al., 2008; Balczerski et al.,2012).In
shh-zebrafish, the trabecula and ethmoidal plate do not
form, and the deformed left and right palatoquadrates fuse
(Fig. 19E, G; Eberhart et al., 2006). The trigeminal NCCs
do not even form skeletogenic condensation in mutant
zebrafish for the hedgehog co-receptor smo (Fig. 19F, H;
Eberhart et al.,2006).Likenoc-zebrafish, fusion of the
palatoquadrates in shh-zebrafish parallels cyclostomes in
which the left and right palatal cartilages meet each other
anterior to the adenohypophysis and medial to the upper
lip (Miyashita, 2012; Oisi et al.,2013a). This parallel again
indicates that a loss of mandibular confinement at the domain
boundary leads to a cyclostome-like phenotype. In both
shh and smo phenotypes, the mandibular NCCs could not
have chondrified medially if the premandibular NCCs had
proliferated around the hypophyseal canal.
Biological Reviews 91 (2016) 611– 657 ©2015 Cambridge Philosophical Society

642 Tetsuto Miyashita
(b)Embryological inferences for the hyoid interface
Numerous genes affect cartilages of the pharyngeal arches,
but I excluded phenotypes in which the defects cannot
be specifically linked to the pharyngeal pouches and/or
the mandibular and hyoid arches. They could result from
mutations that broadly affect NCC migration, survival, or
chondrogenesis (e.g. Piotrowski et al., 1996; Schilling et al.,
1996).
Multiple experiments demonstrate the role of the
hyomandibular pouch in proper jaw development. In
zebrafish, fgf8 mutants and morphants show reduction
of the pharyngeal pouches and, as a result, develop a
variety of defects in jaw, hyoid, and branchial cartilages
(Fig. 20A–C; Crump et al.,2004).Thesezebrafishalsofail
to form the hyomandibula–neurocranium contact. In fras1
mutant zebrafish, the hyomandibular pouch is reduced and
the mandibular and hyoid cartilages are defective and show
dorsoventral fusion without a joint (Fig. 20D; Talbot et al.,
2012).
An intact hyoid arch is crucial to jaw development.
In myoDfh261 mutant zebrafish, cranial and hypobranchial
muscles are missing except those in the hyoid arch (Fig. 20E;
Hinits et al., 2011). In these mutants, dlx2 is still normally
expressed, and the jaw and branchial bars do form, with
some defects (including the jaw being lowered and open).
However, the hyoid cartilage and muscles are relatively
unaffected, and the presence of the trabecula indicates an
intact premandibular interface. Therefore, the mandibular
NCCs are still able to form a jaw, albeit lacking muscular
components, if the premandibular and hyoid interfaces
maintain integrity. In addition, several zebrafish mutants
show specific reduction of the skeleton within the mandibular
and hyoid arches, and yet the branchial bars and the
premandibular skeleton are relatively unaffected [sucker
(suc), schmerle,sturgeon,andhoover ;Piotrowskiet al. (1996);
csnk1a1,dirty south,no soul,sec61a,andword of mouth;
Nissen, Amsterdam & Hopkins (2006)]. These phenotypes
reinforce the link between hyoid patterning and jaw
development.
(c)Embryological inferences for the hypobranchial interface
As predicted, disruption of the interface with the
hypobranchial musculature results in failure to form a
crown gnathostome pattern in the mandibular arch. In
each of Dlx5/6-/-,Hand2cko and EdnRA-/- mice, both the
mandibular skeleton and hypobranchial muscles are severely
affected (Clouthier et al.,1998;Ruest&Clouthier,2009;
Heude et al.,2010;Barronet al.,2011).EdnRA, a receptor
for Edn1, induces Dlx5/6 in the intermediate region of the
mandibular trigeminal NCCs and Hand2 in the ventral region
(Fig. 10B; Ruest et al.,2004;Medeiros&Crump,2012).
In Hand2 and EdnRA mutants, the mid-ventral extension
of the mandibular elements (ectopic maxilla or incisor)
replaces the posteriorly reduced tongue (Fig. 21A–D).
That is, the mandibular elements extend mid-ventrally
in the absence of the hypobranchial musculature, as in
cyclostomes that develop the lingual apparatus in this
region. Therefore, proper development of the hypobranchial
musculature (thereby delimiting the mandibular arch) is
necessary to pattern a jaw (Ruest et al., 2004; Ruest &
Clouthier, 2009). In the absence of this confinement, the
jaw defect is coupled with a cyclostome-like phenotype.
Asimilarhypobranchial–mandibularlinkofdefective
phenotypes also exists in Fuzzy-/- mice (Zhang et al.,
2011).
In zebrafish, the suc-(zebrafish edn-1) mutants have jaw
and hypobranchial defects (Fig. 21E– H; Miller et al.,2000).
Like in myoD mutants (Hinits et al.,2011;SectionIV.2b), the
suc mutants fail to form a jaw joint, and the hypobranchial
muscles are highly reduced. Here again, the link between the
hypobranchial and jaw defects indicates a critical role of the
hypobranchial interface in jaw patterning.
(d)Predictions that remain to be tested
Further to test predictions of the MC Hypothesis, there is
an urgent need to resolve discrepancies between reports of
key gene expression patterns and functions in cyclostomes.
In some cases, reported gene expression patterns differ
markedly among species (Dlx in the trigeminal NCCs:
uniform and broad in Le. japonicum; dorsoventral patterns
in P. marinus and E. burgeri;HoxL6 in the mandibular NCCs:
positive in P. marinus and negative in La. fluviatilis and Le.
japonicum;Nkx3.2 in the mandibular NCCs: positive in Le.
japonicum and negative in P. marinus; Fig. 22I –M; Cohn,
2002; Takio et al., 2004; Cerny et al., 2010; Kuraku et al.,
2010; Fujimoto et al., 2013). These discrepancies may come
from species-specific patterns, different developmental stages,
or false positive/negative results.
AbetterunderstandingoftheEdnsignallingpathwayand
its function in cyclostomes can also test the MC Hypothesis
further. Edn1 expression domains in the ectodermal
epithelium of the pharyngeal arches differ markedly between
cyclostomes and gnathostomes (Fig. 22A–H). Edn1 has
epithelial expression in the ventral regions of all pharyngeal
arches in gnathostomes (Clouthier et al.,1998;Milleret al.,
2000), but its lamprey cognates have epithelial expression
in the dorsal part of the non-mandibular pharyngeal arches
in Le.japonicum (Fig. 22A–F; Kuraku et al., 2010). If this
pattern is general among cyclostomes, then confinement
of the mandibular NCCs correlates with the gain of Edn1
expression in the ventral region. Furthermore, the absence
of lateral branchial bars may correlate with the lack of
dorsal epithelial expression of Edn1.Ednr and Dlx2 are
more strongly expressed in NCCs on the lateral side in
lampreys (Kimmel et al.,2001;Cernyet al.,2010;Kuraku
et al., 2010), whereas the expressions appear uniform across
NCCs in gnathostomes (Nair et al.,2007).Therefore,Edn1
may have ancestrally mediated Dlx2 expression downstream
of FGF signalling in the pharyngeal skeletogenic pathway,
and may have been incorporated later into the dorsoventrally
patterned Dlx expressions. Ongoing work aims to test these
predictions.
Biological Reviews 91 (2016) 611– 657 ©2015 Cambridge Philosophical Society

Fishing for jaws 643
Fig. 20. Jaw defects as a result of disruption to the hyoid interface. (A –C) Pharyngeal pouch formation and jaw and hyoid cartilage
morphology in fgf8 wildtype, mutant, and morphants (from Crump et al., 2004). (A) Wildtype zebrafish, showing pharyngeal arches
labelled with fli1:GFP (Arabic numerals) and pharyngeal pouches (p1, p2 ...) at 28 h post fertilization (hpf) (left) and jaw and hyoid
cartilage stained with alcian blue at 4 dpf (right). (B) fgf8-mutant zebrafish, showing the pharyngeal arches labelled with fli1:GFP
and pharyngeal pouches labelled with zn8 probes at 28 hpf (left) and jaw and hyoid cartilage stained with alcian blue at 4 days post
fertilization (dpf) (right). Pharyngeal pouches, including the hyomandibular pouch (white arrowhead), are reduced, and an ectopic
pouch (white arrow) is present. Same scale as (A). (C) Jaw and hyoid cartilages of various zebrafish morphants treated with fgf8
morpholino, stained with alcian blue at 4 dpf. Black arrowhead points to a defective hyoid cartilage associated with the ectopic
pouch (white arrow in B). (D) Comparison of the mandibular and hyoid skeletal phenotypes between the wildtype and frasb1048
mutants at 7 dpf. Drawings are based on Sox9a:EGFP zebrafish from Talbot et al. (2012). Arrow indicates fusion of the cartilages
at the presumptive jaw joint. Colour codes follow Fig. 4: pink=mandibular arch cartilage; blue green =hyoid arch cartilage. (E)
Comparison of cranial phenotypes between myoDfh261 mutant and wildtype siblings (sib) stained for muscles with MyHC in ventral
and lateral (GFP fluorescent image) views (96 hpf), and stained for skeleton with alcian blue and alizarin red in ventral and lateral
views (5 dpf) (from Hinits et al., 2011). Dark arrowheads indicate some of the muscles in the mandibular arch that are missing in
myoDfh261 mutants. Muscles in the hyoid arch (ao, lo) and the hypobranchial muscle inserted to the hyoid arch (sh) are still present
in myoDfh261 mutants, although other cranial muscles fail to develop. The loss of muscles leads to skeletal defects, but the jaw
(mk) still develops. Abbreviations: ah, adductor hyomandibulae; ao, adductor operculi; ch, ceratohyal; cl, cleithrum; do, dilatator
operculi; eth, ethimoidal plate; h, heart; l–5, levator 5; lo, levator operculi; mk, Meckel’s cartilage; nc, notochord; op, operculum;
pp, protractor pectoralis; ps, parasphenoid; s, somitic muscles; sh, sternohyoideus.
(3) Fossil evidence for functional correlates of jaws
(a)Feeding and ventilation
The MC Hypotheses posits an osteostracan/galeaspid-like
stem gnathostome (Fig. 18A), thereby implying
microphagous deposit feeding as its possible feeding
mode. This scenario differs from the Ventilation Hypothesis
that postulated microphagous suspension feeding at the
initial stages of jaw evolution (Mallatt, 1984b). Galeaspids,
osteostracans, and thelodonts were likely deposit feeders on
the basis of a generally depressiform profile, ventrally located
mouth, and well-developed branchial series (Denison,
1961; Janvier, 1996). Thelodont body fossils are sometimes
associated with mud behind the branchial series, which
in furcacaudiform thelodonts is identified as a stomach
(Turner, 1982; Wilson & Caldwell, 1993). Indirect inferences
for detrital grazing exist in (i) euphaneropids for gut trace
consisting of organic-rich sediment (Stensi¨
o, 1939) and
(ii) the anaspid Birkenia and/or the thelodont Loganellia
for possible association with fine-grained spiral coprolites
(Gilmore, 1992).
Biological Reviews 91 (2016) 611– 657 ©2015 Cambridge Philosophical Society

644 Tetsuto Miyashita
Fig. 21. Jaw defects as a result of disruption to the hypobranchial interface. (A) Ednrafl/fl mouse embryo (E18.5) in lateral view,
showing normal conditions. (B) Sagittal section of Ednrafl/fl mouse embryo (E18.5) near the midline of the head stained with eosin
and hematoxylin, showing normal skeletal phenotype with tongue (t) in mid-ventral position. (C) Ednra-/- mouse embryo (E18.5)
in lateral view, showing the reduced lower jaw. (D) Sagittal section of Ednra-l- mouse embryo (E18.5) near the midline of the head
stained with eosin and hematoxylin, showing the reduced tongue (arrow) and the ectopic maxilla (mx*) in mid-ventral position.
(E) Wildtype (WT) zebrafish [5 days post fertilization (dpf)] stained with 1025 for muscles in ventral view. (F) suc-mutant zebrafish
(5 dpf) stained with 1025 for muscles in ventral view, showing the loss of cranial muscles. (G) Cartilages of the mandibular and
hyoid arches of a wildtype zebrafish (5 dpf) stained with alcian blue in left lateral view. Same scale as (H). (H) Mandibular and
hyoid cartilages of a suc-mutant zebrafish (5 dpf) stained with alcian blue in left lateral view, showing skeletal defects including
reduction and fusion between the elements. (A–D) From Ruest & Clouthier (2009). (E –H) From Miller et al. (2000). Abbreviations:
am, adductor mandibulae; hh, hyohyoideus; i, incisor; ih, interhyoideus; ima, intermandibularis anterior; imp, intermandibularis
posterior; md, mandible; mi, upper incisior; mx*, ectopic maxilla; t, tongue. For other abbreviations, see Fig. 20.
Deposit feeding was likely the primary feeding mode
in primitive placoderms as well. The antiarch Bothri-
olepis has a sediment-filled intestine (Denison, 1941). The
hypobranchial-assisted suction implicated for the phyllolepid
arthrodire Cowralepis (Carr, Johanson & Ritchie, 2009) is
also consistent with deposit feeding, but whether exhibiting
microphagy or macrophagy remains unclear for this and
other depressiform placoderms. Some placoderms closer
to crown gnathostomes were unmistakably macrophagous.
Stomach contents indicate that the eubrachythoracid
arthrodire Coccosteus fed on smaller fish (Miles & Westoll,
1968). The chimaeroid-like bodyplan with tooth plates
of pleromic dentine indicates durophagy in ptyctodonts
(Ørvig, 1980).
However, discussion of feeding modes risks overgener-
alization because the phylogenetic pattern is not linear.
Suspension feeding in the water column is inferred for
arthrodires with elongate and edentulous gnathal plates,
such as the large eubrachythoracid Titanichthys (Denison,
1978). Evidence for occlusion in P elements of conodonts
implies a particle size large enough to require mechani-
cal processing (Purnell, 1995; Jones et al.,2012;Purnell&
Jones, 2012). Still, precision of these inferences is a problem.
Heterostracans, for example, have been reconstructed as
macrophagous predators, deposit feeders, or microphagous
suspension feeders on various indirect reasoning (Denison,
1961; Halstead, 1973; Janvier, 1974; Soehn & Wilson, 1990;
Janvier & Blieck, 1993; Purnell, 2002). Overall, all suggested
feeding modes require powerful ventilation, particularly for
most stem gnathostomes for which ram feeding is unlikely.
A strong buccal pump would be favoured in any feeding
styles that utilize suction. Therefore, ventilation was likely
an essential functional driver through the initial stages of jaw
evolution.
(b)Dentition
If jaw evolution was initially driven by ventilation, early teeth
and tooth-like structures could have evolved independent
of the presence/absence of the jaw. This prediction is
compatible with current evidence (Smith & Hall, 1990,
1993; Smith & Coates, 2000, 2001; Donoghue & R¨ucklin,
2014). Although sometimes touted as the precursor of
teeth, enamelized euconodont elements likely evolved
independently from crown gnathostome teeth (Murdock
et al.,2013).Dermaldenticlesoccuracrossjawlessstem
gnathostomes, with conspicuous oral plates in heterostracans
and osteostracans (Fig. 16B; Janvier, 1996). Thelodonts
have both external oral and internal pharyngeal scales,
and the latter in Loganellia have polarized development
(R¨ucklin et al., 2011). Among placoderms, some arthrodires
have tooth cusps of dentine and bone with a pulp cavity,
distinguishable from dermal tubercles (ch. 15, Fig. 11; Smith
&Johanson,2003;R¨ucklin et al., 2012). On the other hand,
acanthothoracids, antiarchs, ptyctodonts, and rhenanids lack
dental lamina, but variably have oral tubercles or beak-like
tooth plates (ptyctodonts) (Johanson & Smith, 2005). Dermal
Biological Reviews 91 (2016) 611– 657 ©2015 Cambridge Philosophical Society

Fishing for jaws 645
Fig. 22. Different gene expression patterns generate testable predictions based on the MC Hypothesis. (A–F) Edn cognate
(EdnA/B/C) expression in the lamprey Le. japonicum, showing a lack of expression in the ectodermal epithelium of the mandibular
arch (yellow arrowhead) and the presence of expression in the ectodermal epithelium of other pharyngeal arches (red arrowhead). (A)
In situ hybridization of Edn cognates in Le. japonicum embryo at Tahara’s stage 24 in left lateral view. (B) Same embryo in ventral view,
enlarged. (C) In situ hybridization of Edn cognates in Le. japonicum embryo at Tahara’s stage 25 in left lateral view, showing planes
of section for (D–F). (D –F) Cross sections of Le. japonicum embryo, each showing a lack of epithelial expression in the mandibular
arch (D, E), expression in the mandibular neural crest cells (NCCs) (D, E), and ectodermal epithelial expression in the dorsal part
of the non-mandibular pharyngeal arches (F). (G, H) edn1 expression in zebrafish D. rerio [36 h post fertilization (hpf)], showing the
presence of epithelial expression in the ventral region of all pharyngeal arches. (G) Close-up of the pharyngeal region in left lateral
view, showing the epithelial expression of edn1 (red arrowheads) in the ventral region of the pharyngeal arches (ectodermal) and in
pharyngeal pouches (endodermal; labelled with Arabic numerals; strong expression zone in pharyngeal pouch 2 is delineated with
a broken line). Black dots indicate a core of a pharyngeal arch. (H) Close-up of the pharyngeal region in ventral view, showing
edn1 expression in NCCs and mesodermal mesenchyme (MMCs) in the ventral region of the pharyngeal arches (black arrowheads).
Asterisks indicate non-expressing NCCs that surround the mesenchymal expression zone. (I–K) Nkx3.2 expression in the lamprey
Le. japonicum.(I)Le. japonicum embryo at Tahara’s stage 24 in left lateral view, showing expression at the trigeminal ganglia (white
arrowhead). (J, K) Le. japonicum embryo at Tahara’s stage 25 in left lateral view (J) and in cross section (K) showing expressions in
the mandibular NCCs between the lower lip and velum (black arrowheads) and in trigeminal ganglia (white arrowheads). (L, M)
Nkx3.2 expression in lamprey P. marinus at Tahara’s stage 26.5 in left lateral view (L) and in longitudinal section stained with eosin
(M). By contrast to Le. japonicum,Nkx3.2 is not expressed in the mandibular NCCs, but in the pharyngeal ectodermal epithelium (red
arrowheads). Black arrowhead =expression in NCCs; red arrowhead =expression in ectodermal or endodermal epithelium; white
arrowhead =expression in nerve ganglia; yellow arrowhead =lack of expression. Photographs for (A –F) and (I– K) from Kuraku
et al. (2010); (G) and (H) from Miller et al. (2000); (L) and (M) from Cerny et al. (2010). Abbreviations: aa1, aortic arch 1; llp, lower lip;
m, mouth; n, notochord; nt, neural tube; PA, pharyngeal arch; ph, pharynx; se, surface ectoderm; ulp, upper lip; vel, velum; Vg,
trigeminal ganglia.
Biological Reviews 91 (2016) 611– 657 ©2015 Cambridge Philosophical Society

646 Tetsuto Miyashita
gnathal elements appear to be absent in petalichthyids
(Johanson & Smith, 2005). Acanthodians collectively exhibit
a wide variety of tooth forms: teeth on either endochondral or
dermal jaw elements, isolated teeth, tooth whorls, crushing
plates, pharyngeal denticles/cones/rakers, and cheek and
lip scales that migrated into the mouth (Gross, 1971;
Ørvig, 1973; Blais, MacKenzie & Wilson, 2011). Given
this distribution of teeth and tooth-like structures, and given
such wide variation in developmental modes among and
within species of stem and crown osteichthyans (Botella et al.,
2007; Soukup et al., 2008; Huysseune, Sire & Witten, 2009;
Qu et al., 2013), dentitions likely evolved, perhaps multiple
times, independent of the origin of the jaw.
(c)Perichondral association of early jaw elements
The MC Hypothesis predicts that the endoskeleton
became integrated with dermal elements via perichon-
dral ossification. Perichondral ossification occurs in
osteostracans (Donoghue, Sansom & Downs, 2006),
whereas perichondral calcification exists in galea-
spids (Zhu & Janvier, 1996; Wang et al., 2005). The
branchial bars were calcified in euphaneropids (Janvier &
Arsenault, 2002, 2007), and pyritized internal elements in
arandaspids may represent a remnant of calcifying cartilages
(Gagnier, 1993). Therefore, perichondrial mineralization
preceded the jaw, and perichondral ossification likely arose
in the last common ancestor of osteostracans and jawed
gnathostomes (ch. 16, 17, Fig. 11).
The proposed perichondral association (Fig. 18C) is
supported by the jaw morphology of placoderms. In
this group, the dermal gnathal plates attach to the
sometimes perichondrally ossified palatoquadrate and
Meckel’s cartilage (Fig. 23A, B; Goujet, 1984; Young,
1984). The chronological trend of dermal dominance in
arthrodire jaws positively correlates with both ontogenetic
and phylogenetic trends (Fig. 23C –F; Young, 2010; R¨ucklin
et al., 2012). In many osteichthyans, jaw elements dermally
ossify around the palatoquadrate and Meckel’s cartilage
to form upper and lower jaws, whereas chondrichthyans
secondarily evolved calcification of the cartilaginous jaw
elements. The Silurian placoderm Entelognathus suggests
conservation of dermal elements between osteichthyans
and placoderms (Fig. 23G–J; Zhu et al., 2013). Thus, the
osteichthyan pattern appears more plesiomorphic than the
chondrichthyan pattern.
Entelognathus also suggests that the commissural lamina
shifted to the medial side of the adductor mandibulae in the
stem of the crown-group gnathostomes, thus documenting
a final stage of transition of the pharyngeal skeleton from
the lateral to the medial side (Fig. 23F). This series provides
morphological evidence for a shift of chondrogenetic site in
the mandibular arch, but not in conjunction with that in the
branchial arches. The medial shift of the jaw was later than
that of the branchial bars, as indicated by medial adductor
fossa of the jaw elements and the internalized branchial bars
in placoderms (Fig. 23A, B).
(d)The first jaw joint and synovial diarthrosis
The MC Hypothesis postulates co-option of jaw-joint-specific
gene expression at Stage III (Fig. 18C; Cerny et al.,2010;
Medeiros & Crump, 2012). This first jaw joint likely arose via
modification of a pre-existing structure, because a synovial
diarthrosis is a complex association of mineralized skeletal
elements, fibrous tissues (ligaments and synovial membrane),
and synovial fluid.
The closest model in cyclostomes to a synovial diarthrosis
is a venous sinus associated with the velum. In hagfish,
the hyaline-like cartilage of the velum approaches the
facial skeleton, and a velar sinus (cardinal heart) sits
between the two cartilages and drains into the anterior
cardinal vein (Fig. 24D; Cole, 1926; Miyashita, 2012).
The ligamentous wall of the cardinal heart links the two
cartilages via the perichondrium and is devoid of myofibrils.
The flexion of the velum pressurizes the cardinal heart
and aids in circulation (Forster, 1997). A similar velar
sinus occurs in lampreys (Tsuneki & Koshida, 1993). The
cyclostome intercartilaginous velar sinus closely resembles a
synovial joint both developmentally and morphologically
(Fig. 24A–C), with ligaments as in a joint capsule and
with blood instead of synovial fluid. To close the gap
further, the synovial fluid derives from blood plasma in
gnathostomes, and the ligamentous tissue for the joint
capsule is highly vascularized. Vascular endothelial growth
factor (VEGF) that promotes angiogenesis is also linked to
synovial fluid accumulation (Fava et al., 1994). Furthermore,
the anlage of the cardinal heart in hagfishes forms between
the two chondrifying mesenchymal populations (velar and
facial cartilages) (Fig. 14G), much like cavitation in the
development of a synovial joint (Fig. 24A).
These similarities suggest that the first jaw joint was
a co-opted intercartilaginous blood sinus like that in
cyclostomes. Spatially, the hyomandibular position of the
velar sinus in cyclostomes approximately agrees with the
position of a jaw joint in gnathostomes. Anatomically, the
third branch of the trigeminal nerve innervates the velum in
cyclostomes and the mandible in gnathostomes (Lindstr¨
om,
1949; Song & Boord, 1993). Functionally, the velar sinus is
the closest structure to diarthrosis among jawless vertebrates.
Consistent with this comparison, Nkx3.2 that has focal expres-
sion in the jaw joint is expressed along the boundary between
the lower lip and velum in at least one species of lamprey
(Fig. 22J, K; Le. japonicum; Kuraku et al.,2010).Correlates
of a velar sinus occur in the successive outgroups of jawed
vertebrates, galeaspids, and osteostracans (Section III.2d).
Further evidence exists that an intercartilaginous blood
sinus was a primitive form of a joint capsule in general.
In osteostracans, the joint surface for the endoskeleton of
the pectoral fin does not form a glenoid fossa. Unfinished
bone surface (likely capped by cartilage) represents the joint
surface within the fossa irrigated and drained by a pair of
large vascular foramina. This area is further surrounded
by a fossa for muscular attachment anteriorly and several
more foramina from the marginal venous sinus, all within
the pectoral fenestra (Fig. 24E; W¨
angsj¨
o, 1952; Janvier,
Biological Reviews 91 (2016) 611– 657 ©2015 Cambridge Philosophical Society

Fishing for jaws 647
Fig. 23. Early evolution of the jaw via incorporation of endochondral and dermal elements. The stem gnathostome jaws are a
composite of dermal and chondral elements. (A, B) Jaw of the antiarch placoderm Bothriolepis. (A) Reconstruction of dermal and
chondral jaw elements in dorsal and ventral view. (B) Upper and lower jaws preserved in situ in ventral view [FMNH (Field Museum
of Natural History) PF 3812]. (C–E) Variation in gnathal elements among arthrodire placoderms indicates chronological and
phylogenetic trends: the mobile endochondral elements deployed the dermal elements via perichondral ossification as a functional
biting element. (C) Lower jaw of a buchanosteid placoderm in which the dermal infragnathal is attached to the main jaw element, a
perichondrally ossified Meckel’s cartilage. (D) Lower jaw of a coccosteomorph placoderm in which the dermal infragnathal makes
a greater contribution to the jaw. The perichondrally ossified Meckel’s cartilage still acts as a main lever. (E) Lower jaw of the
placoderm Dunkleosteus terrelli in which the perichondrally ossified Meckel’s cartilage is reduced in size but still composes the jaw
articulation. (F) Early evolution of the jaw skeleton, with cross sections of the head for each early gnathostome lineage. In the initial
stage, the commissural lamina of the endochondral palatoquadrate extended to the medial side of the adductor muscles, eventually
completing the lateromedial shift of the pharyngeal endoskeletal elements. Meanwhile, the perichondrally ossified skeletal elements
deployed dermal elements, which became the main functional levers in osteichthyans. The dermal components were reduced in
chondrichthyans and acanthodians, in which the endochondral elements function as the main levers. (G–J) Comparison of the
dermal skull elements between osteichthyan and placoderm representatives reveals conservation of many dermal skull elements, and
indicates that dermal marginal jaw elements (perichondrally associated with the palatoquadrate and Meckel’s cartilage) appeared
early in gnathostome evolution. Colours indicate dermal elements conserved between osteichthyans and placoderms (Zhu et al.,
2013). (G) Bowfin (Amia calva) in lateral view as an example of the actinopterygian skull. The pectoral girdle is omitted. (H)
Eusthenopteron foordi in lateral view as an example of the sarcopterygian skull and pectoral girdle. (I) Skull and pectoral girdle of
the placoderm Entelognathus primordialis in lateral view. (J) The arthrodiran Dicksonosteus arcticus in lateral view as an example of the
placoderm skull and pectoral girdle. Many elements are obscured in lateral view, and some elements are independently lost or
reduced within each lineage. See Zhu et al. (2013) for a complete list of dermal elements. (A) Modified after Young (1984); (C –E)
from Young (2010); (F, H–J) modified after Zhu et al. (2013); (G) modified after Grande & Bemis (1998). (C –J) Are not to scale.
Biological Reviews 91 (2016) 611– 657 ©2015 Cambridge Philosophical Society

648 Tetsuto Miyashita
1981, 1984, 1985a). A similar morphology occurs in the
pectoral fenestra of antiarch placoderms. Again, the pectoral
fenestra does not have a glenoid fossa. The dorsal and ventral
articular facets sit within a deeply embayed fossa, which was
irrigated by the vascular axillary foramen (Fig. 24F; Goujet,
1984; Young & Zhang, 1992; Janvier, 1984, 1995). This
pattern is regarded as homologous to the pectoral attachment
in osteostracans (Johanson, 2002). These morphological
features are more consistent with an intercartilaginous blood
sinus than a typical synovial diarthrosis. In arthrodires, the
pectoral fin attachment still does not form a glenoid fossa but
is along the longitudinal ridge within the pectoral fenestra,
which is perforated by a number of neurovascular foramina
(Stensi¨
o, 1969). The rhenanid placoderm Gemuendina has
asemicircularscapularprocessthatfitstoasinglebasal
cartilage (Gross, 1963).
This ‘blood sinus’ hypothesis conflicts with the previously
proposed ‘mucocartilage’ hypothesis, which holds that the
jaw joint evolved from soft mucocartilage-like connective
tissue (Cattell et al., 2011). This hypothesis compares the
mucocartilage of larval lampreys to the jaw joint as
SoxE-positive and Runx/Barx-negative mesenchymal tissues.
However, the mucocartilage is a transient fibrous larval
tissue. Neither mucocartilage nor similar tissues have been
observed in other vertebrates. As for expression patterns,
many skeletal and connective tissues express SoxE in
both lampreys and crown gnathostomes, and the reported
expression domains of Runx and Barx are spatially and
functionally not identical between them (Cattell et al.,2011;
Lakiza et al., 2011). Highly general expression patterns
alone are therefore insufficient to postulate mucocartilage
as present in a gnathostome ancestor. As opposed to
the mucocartilage, a synovial diarthrosis is not simply
mesenchymal, but a complex of well-differentiated tissues.
Neither does the mucocartilage hypothesis explain the
origins of synovial fluid, joint capsule, cavitation, or
articular cartilages. Thus, mucocartilage likely represents
an autapomorphic condition of lampreys.
At any rate, jaw joint morphology in the earliest
gnathostomes remains elusive. In the antiarch Bothriolepis,
the quadrate remains unossified posterior to the adductor
fossa, and the Meckel’s cartilage is cartilaginous around
the joint (Fig. 23A, B). In other placoderms preserved
with a palatoquadrate, the quadrate has the unossified
‘condyle’ and the sometimes absent detent process, and
the glenoid fossa of the articular is a large, round, flat area
with a narrow rim below the dorsal margin (e.g. Young,
1986; Gardiner & Miles, 1990; Young, Leli`
evre & Goujet,
2001). In ptyctodonts, the glenoid fossa receives the lateral
quadrate condyle, whereas the mesial ventral margin of the
quadrate fits to the articular (e.g. Trinajstic et al.,2012).In
acanthodians, and likely in Entelognathus,thequadrateclasps
the preglenoid process of the articular between the articular
condyle and the prearticular process (Miles, 1968; Zhu et al.,
2013). None of these cases presents a definitive correlate of an
intercartilaginous blood sinus, but the cartilaginous jaw joint
in antiarchs is reminiscent of the proximal velar contact in
cyclostomes. Otherwise, the double joint and the attachment
surface for the mandibulohyoid ligament (Johanson, 2003)
resemble a typical synovial joint of crown gnathostomes.
(e)Evolution of the hyoid and branchial arches
The MC Hypothesis postulates a hyoid arch as supporting
ahemibranchinstemgnathostomes,asexhibitedby
both cyclostomes and crown gnathostomes. By contrast,
the ‘aphetohyoidean’ hypothesis (Gregory, 1904; Mallatt,
1996) posits a hyoid arch of the earliest gnathostome as
a full branchial arch. For that latter hypothesis to be
parsimoniously probable, such aphetohyoidean condition
should be shared among stem gnathostomes. Indeed, the
first full branchial arch has sometimes been interpreted as
a hyoid arch in heterostracans and osteostracans (Mallatt,
1996; Brazeau & Friedman, 2014).
This interpretation of the aphetohyoidean condition in
stem gnathostomes does not agree with fossil evidence.
In heterostracans, the most anterior gill impression has
been identified as hyomandibular on the basis of its
position posterior to the orbit (Mallatt, 1996). However,
heterostracans variably have one to three such impressions
posterior to the orbit and anterior to the otic capsule
(Fig. 16A; Stensi¨
o, 1964; Halstead, 1973). In the absence
of other correlates that can be used for identification of
pharyngeal arches (e.g. cranial nerves), such a vaguely defined
position alone is insufficient to identify the gill impression
as a non-conventionally holobranchial hyoid arch – a
pattern never observed otherwise. Similarly, osteostracans
have been misinterpreted. In this group, the first branchial
pouch was posited as hyoidean because the cast of the
facial nerve is exposed ventrally in that position (Mallatt,
1996). However, the nerve extends further anteriorly within
adeepgrooveproximaltothehyomandibular–palatinesplit
(Fig. 5; Stensi ¨
o, 1927, 1932, 1964; W¨
angsj¨
o, 1952). This
observation clearly indicates that the main site of innervation
lay below and anterior to the distal part of the nerve cast, not
immediately below the most proximal part of the exposed
nerve cast (Janvier, 1981, 1984, 1985a,b). This peculiar
morphology is better explained by an anterior shift of the
oropharyngeal structures relative to the neurocranium.
In acanthodians and placoderms, the hyoid arch skeleton
is suspended from the braincase and closely associated with
the jaw elements (Miles, 1968; Stensi¨
o, 1969; Forey &
Gardiner, 1986; Young, 1986; Trinajstic et al.,2012),and
the position of the hyoid–braincase contact constitutes a
gnathostome synapomorphy (Brazeau & Friedman, 2014).
Although the hyoid skeleton in the stem chondrichthyan
Ozarcus resembles a branchial bar (Pradel et al., 2014), the
most parsimonious explanation for this exceptional case is
an autapomorphy. Taken together, these observations reject
the aphetohyoidean hypothesis. Namely, the hyoid arch
supported a hemibranch in stem gnathostomes, and the
endoskeletal attachment for the ‘velum’ was hyoidean as in
cyclostomes (Figs 4A, B and 15C).
Rejection of the aphetohyoidean hypothesis leads to an
intriguing possibility: a hyoid arch may also be considered
Biological Reviews 91 (2016) 611– 657 ©2015 Cambridge Philosophical Society

Fishing for jaws 649
Fig. 24. Possible origin of the jaw joint via an intercartilaginous blood sinus (MC Stage III). (A) Schematic drawing for
development of a synovial joint in gnathostomes (based on Archer, Dowthwaite & Francis-West, 2003; Pitsillides & Beier,
2011). (B) Three-dimensional reconstruction of a jaw joint of the gecko Eublepharis macularius using X-ray microcomputed
tomography (from Payne, Holliday & Vickaryous, 2011), showing the gross anatomy of a typical gnathostome synovial
joint. (C) Transverse histological section of the jaw joint of Eu. macularius (from Payne et al.,2011),showingthefine-scale
anatomy of a typical gnathostome synovial joint. Hyaline cartilage caps the element as an articular cartilage. The synovial
cavity is encapsulated within the ligamentous tissue, and the cavity filled with synovial fluid acts as a lubricant and shock-
absorbing agent. (D) Transverse histological section of the head of an adult hagfish (Ep. stoutii) at the level of the cardinal
heart. The cardinal heart is an intercartilaginous venous sinus between the velar and facial cartilages, and is functionally
and anatomically similar to a synovial joint. The cardinal heart lacks myofibrils, and its wall consists of ligament. The
movement of the velum aids in circulation, while the blood acts as a lubricant. (E) Pectoral fenestra of the osteostracan
Norselaspis glacialis in left lateral view (modified from Janvier, 1984), showing the attachment area of the pectoral fin. The
pectoral fenestra is shaded green, and the external surface of the head shield light brown. The attachment of the pectoral fin
is a pad of perichondrally ossified cartilage (blue) within a fossa irrigated and drained by the several major vascular
foramina (red). (F) The area of pectoral fin attachment in the antiarch Bothriolepis mcpharsoni in left lateral view (modified
from Janvier, 1995). The pectoral articulation is via perichondrally ossified cartilaginous pads divided dorsoventrally by the
funnel pit, which may have hosted a blood sinus irrigated and drained by the axillary foramen. Colour codes are as in (E).
as a distinctly patterned region from branchial arches.
The hyoid arch appears intermediate between mandibular
and branchial arches for supporting a hemibranch, sitting
between two pharyngeal pouches, and expressing the Hox
code (Fig. 2). It largely remains unexplored how such a
combination of features evolutionarily arose in the hyoid
arch. This perspective is likely to provide a critical insight
into the origin and early evolution of the pharyngeal
apparatus.
(4) Neural crest as a facilitator of ‘segmentation’
Evidence that contributed to constructing the MC
Hypothesis reframes the question of jaw origins as the
Biological Reviews 91 (2016) 611– 657 ©2015 Cambridge Philosophical Society

650 Tetsuto Miyashita
making of a segment – remodelling a distinct region into
aseriallypatternedcompartmentofthemusculoskeletal
system. The classical model coupled the idealized visceral
branchiomery with the extrapolated paraxial somitomery
and the ectodermal rhombomery in the hindbrain. Because
the paraxial somitomery does not extend into the head
in vertebrates (Fig. 6), truly segmental head structures in
vertebrates appear to be the hindbrain rhombomeres and
pharyngeal pouches. Although the origin of Hox-regulated
rhombomeres (Parker, Bronner & Krumlauf, 2014) remains
poorly understood beyond the node of crown vertebrates,
Tbx1-dependent specification of pharyngeal pouches in
vertebrates likely shares a root with a similar mechanism
behind the pharyngeal slits of cephalochordates (Koop et al.,
2014). Therefore, these repeating patterns in the ectoderm
and endoderm probably evolved independently (Graham
et al.,2014).Whatpossiblemechanismcouldexplainthese
remarkable anteroposterior and dorsoventral correlations
between the serial embryonic structures (Fig. 6)?
One alternative to the unified metamerism is reinforce-
ment of robusticity in tissue interactions during development.
When distinctly fated and spatially patterned cell aggregates
interact to form multi-component structures – such as the
elements of the pharyngeal apparatus (Frisdal & Trainor,
2014) – selection for robusticity to developmental pertur-
bations would reinforce spatial and temporal correlations
between the anlagen.
A key facilitator underlying this vertebrate head
‘segmentation’ is neural crest. Cranial NCCs interact with
all of the anlagen: the migratory paths are constrained to
the rhombomeric borders; interactions with the placodes
form basal ganglia and sensory capsules; and interactions
with MMCs and the endodermal epithelium pattern
the pharyngeal musculoskeletal system (Fig. 6). This view
emphasizes the abilities of NCCs to migrate and interact
with different tissues as explanations for the highly integrated
vertebrate head (Hall, 2009; Minoux & Rijli, 2010; Trainor,
2014).
V. CONCLUSIONS
(1) Previous hypotheses about jaw origins either required
phenotypes that are never observed in terminal taxa or
characters allegedly causally linked but phylogenetically
decoupled from the origin of the jaw. The classical model of
a metameric pharynx in a jawless ancestor is abandoned, as
exceptions are far more common than the supposed rule.
(2) Considerable anatomical, embryological, and fossil
evidence indicates that the seemingly serial mandibular
musculoskeletal system was initially distinct from that of
all other pharyngeal arches before the origin of the jaw.
(3) The Mandibular Confinement Hypothesis postulates
a set of developmental rules successively introduced in stem
gnathostomes, which allow mandibular arch derivatives
to serially assimilate musculoskeletal patterning of other
pharyngeal arches. The initially distinct and structurally
complex mandibular elements were spatially confined
as surrounding domains took over functions previously
performed by the mandibular structures. The NCCs and
MMCs of the mandibular arch freed from cyclostome-like
differentiation patterns could then be remodelled into a jaw
apparatus.
(4) Three interfaces confined the NCCs and MMCs of
the mandibular arch in gnathostomes. The premandibular
interface resulted from proliferation and differentiation of
trigeminal NCCs of the premandibular region around the
split nasohypophyseal placodes, thereby decoupling the
mandibular contribution. The hyoid interface resulted from
loss of the hyomandibular extension of the mandibular
elements. The hypobranchial interface resulted from
occupation of the mid-ventral pharyngeal space by the
hypobranchial musculature.
(5) The MC Hypothesis also proposes the fol-
lowing functional adaptations in the deposit-feeding
ancestors: (i) medial shift of the branchial bars
for buccal–parabranchial ventilation; (ii) chondral–
dermal association via perichondral ossification in the earliest
jaw; and (iii) co-option of an intercartilaginous blood sinus
for a synovial diarthrosis.
(6) The MC Hypothesis reconciles and complements
previous hypotheses. This is accomplished because the MC
Hypothesis: (i)justifiestheuseofcyclostomesasasurrogate
for ancestral states through tests with fossil inference; (ii)
provides functional explanations for differentiated tissues
when contrasting comparable developmental stages between
cyclostomes and crown gnathostomes; and (iii) views the
origin of the jaw as the result of mandibular patterning
assimilating the pharyngeal patterning, rather than as
specialization of a pharyngeal metamere.
(7) The MC Hypothesis suggests that multi-tissue
interactions underlie the vertebrate head ‘segmentation’.
Neural crest was a key facilitator of correlations between
seemingly corresponding serial embryonic structures. In the
vertebrate head, only the hindbrain rhombomeres and the
pharyngeal pouches appear to be truly segmented, but they
likely have independent evolutionary origins.
VI. ACKNOWLEDGEMENTS
The ideas presented in this paper developed over years
of discussion with my supervisors A. R. Palmer and
P. J. Currie and numerous other people, particularly
P. Ahlberg, T. Allison, M. Coates, P. Donoghue, B.
Hall, P. Janvier, Z. Johanson, S. Kuratani, and K. Ota.
I thank K. Miyashita, E. Koppelhus, A. Oatway, H.
McDermid, Bamfield Marine Sciences Centre, and the
Palmer lab for technical assistance. G. Unlu and E. Knapik
provided photographs for Fig. 19. My collections visits
were assisted by: P. Janvier and G. Cl´
ement (Mus´
eum
national d’Histoire naturelle); H. Maddin, J. Hanken,
K. Hartel, and J. Cundiff (Museum of Comparative
Zoology, Harvard University); J. Maisey and A. Gishlick
Biological Reviews 91 (2016) 611– 657 ©2015 Cambridge Philosophical Society

Fishing for jaws 651
(American Museum of Natural History); W. Simpson
(Field Museum of Natural History). This paper would not
have materialized without fellow students, instructors, and
assistants from Embryology 2013 at the Marine Biological
Laboratory and Larval Biology 2010 at Friday Harbor
Laboratories. We shared days and nights of experiments,
discussions, and adventures (MBL: A. Accorsi, C. Baker,
R. Behringer, M. Colasanto, A. Collazo, A. Edgar, A.
Gillis, J. Henry, K. Hosoda, N. Kenny, B. Kumar, E.
Kunttas, L. Linden, L. Maya, K. McClelland, S. Miller,
J. Miranda, E. Mis, D. ¨
Ozpolat, J. Park, N. Patel, T.
Piotrowski, P. Ray, M. Riddle, A. S´
anchez-Alvarado,
E. Schock, A. Shyer, G. Stooke-Vaughan, B. Swalla,
S. Tintori, W. Wang, and W. Weßels; FHL: S. Bayer,
K. Chan, M. Contins, D. Gr¨unbaum, L. Kapsenberg,
E. Kosman, R. Neves, E. Norton, E. Renborg, M. Rock,
L. Sam, and R. Strathmann). I appreciate W. Foster and
four anonymous referees for providing useful critiques over
a wide range of topics and A. Cooper for carefully editing
the final draft. This work was supported by scholarships
from Alberta Innovates and Vanier Canada to me and by
NSERC Discovery Grant (A7245) to A.R. Palmer.
VII. REFERENCES
Abe,M.,Ruest, L.-B. & Clouthier,D.B.(2007). Fate of cranial neural crest cells
during craniofacial development in endothelin-A receptor-deficient mice. International
Journal of Developmental Biology 51,97–105.
Abu-Issa,R.,Smyth,G.,Smoak,I.,Yamamura,K.&Meyers,E.N.(
2002).
Fgf8 is required for pharyngeal arch and cardiovascular development in the mouse.
Development 129,4613–4625.
Adachi,N.,Takechi,M.,Hirai,T.&Kuratani, S. (2012). Development of the
head and trunk mesoderm in the dogfish, Scyliorhinus torazame:II.Comparisonof
gene expression between the head mesoderm and somites with reference to the
origin of the vertebrate head. Evolution & Development 14,257–276.
Alexandre,D.,Clarke,J.D.W.,Oxtoby,E.,Yan, Y.-L., Jowett,T.&Holder,
N. (1996). Ectopic expression of Hoxa-1 in the zebrafish alters the fate of the
mandibular arch neural crest and phonocopies a retinoic acid-induced phenotype.
Development 122,735–746.
Allis,E.P.(
1897). The cranial muscles and cranial and first spinal nerves in Amia
calva.Journal of Morphology 12,487–808.
Allis,E.P.(
1916). The so-called mandibular artery and the persisting remnant
of the mandibular aortic arch in the adult selachian. Journal of Morphology 27,
99– 118.
Allis,E.P.(
1923). The cranial anatomy of Chlamydoselachus anguineus.Acta Zoologica 4,
123– 221.
Archer,C.W.,Dowthwaite,G.P.&Francis-West,P.(
2003). Development of
synovial joints. Birth Defects Research Part C 69,144–155.
Arsenault,M.,Desbiens, S., Janvier,P.&Kerr,J.(
2004). New data on
the soft tissues and external morphology of the antiarch Bothriolepis canadensis
(Whiteaves, 1880), from the Upper Devonian of Miguasha, Quebec. In
Recent Advances in the Origin and Early Radiation of Vertebrates (eds G. Arratia,
M. V. H. Wilson and R. Cloutier), pp. 439– 454. Verlag Dr. Friedrich Pfeil:
Munich.
Baker,C.V.H.(
2005). Neural crest and cranial ectodermal placodes. In Developmental
Neurobiology, Fourth Edition (eds M. S. Rao and M. Jacobson), pp. 67– 127. Kluwer
Academic Publishers, New York.
Baker,C.V.,Stark,M.R.,Marcelle,C.&Bronner-Fraser,M.(
1999).
Competence, specification and induction of Pax-3 in the trigeminal placode.
Development 126,147–156.
Balczerski,B.,Matsutani,M.,Castillo,P.,Osborne,N.,Stainier,D.Y.
R. & Crump,J.G.(
2012). Analysis of Sphingosine-1-phosphate signaling mutants
reveals endodermal requirements for the growth but not dorsoventral patterning of
jaw skeletal precursors. Developmental Biology 362,230–241.
Balfour,F.M.(
1878). The development of the elasmobranchial fishes. Journal of
Anatomy and Physiology 11,405–706.
Baltzinger,M.,Ori,M.,Pasqualetti,M.,Nardi,I.&Rijli,F.M.(
2005).
Hoxa2 knockdown in Xenopus results in hyoid to mandibular homeosis. Developmental
Dynamics 234,856–867.
Barron,F.,Woods,C.,Kuhn,K.,Bishop,J.,Howard,M.J.&Clouthier,D.
E. (2011). Downregulation of Dlx5 and Dlx6 expression by Hand2 is essential for
initiation of tongue morphogenesis. Development 138,2249–2259.
Barry,M.A.,Hall,D.H.&Bennett, M. V. L. (1988). The elasmobranch
spiracular organ. I. Morphological studies. Journal of Comparative Physiology A 163,
85– 92.
Basden,A.M.&Young,G.C.(
2001). A primitive actinopterygian neurocranium
from the Early Devonian of southeastern Australia. Journal of Vertebrate Paleontology
21,754–766.
Baskin, L. S., Erol,A.,Li,Y.W.,Liu,W.H.,Kurzrock,E.&Cunha,G.
R. (1999). Anatomical studies of the human clitoris. The Journal of Urology 162,
1015– 1020.
Benouaiche, L., Gitton,Y.,Vincent,C.,Couly,G.&Levi,G.(
2008). Sonic
hedgehog signalling from foregut endoderm patterns the avian nasal capsule.
Development 135,2221–2225.
Bhat,N.&Riley,B.B.(
2011). Integrin-α5coordinatesassemblyofposteriorcranial
placodes in zebrafish and enhances Fgf-dependent regulation of otic/epibranchial
cells. PLoS ONE 6,e27778.
Bjerring,H.C.(
1977). A contribution to structural analysis of the head of craniate
animals. Zoologica Scripta 6,127–183.
Blais, S. A., MacKenzie, L. A. & Wilson,M.V.H.(
2011). Tooth-like scales in
Early Devonian eugnathostomes and the ‘outside-in’ hypothesis for the origins of
teeth in vertebrates. Journal of Vertebrate Paleontology 31,1189–1199.
Botella,H.,Blom,H.,Dorka,M.,Ahlberg,P.E.&Janvier,P.(
2007). Jaws
and teeth of the earliest bony fishes. Nature 448,583–586.
Bothe,I.,Tenin,G.,Oseni,A.&Dietrich, S. (2011). Dynamic control of head
mesoderm patterning. Development 138,2807–2821.
Brainerd,E.L.&Ferry-Graham, L. A. (2005). Mechanics of respiratory pumps.
Fish Physiology 23,1–28.
Brazeau,M.D.(
2009). The braincase and jaws of a Devonian ‘acanthodian’ and
modern gnathostome origins. Nature 457,305–308.
Brazeau,M.D.&Friedman,M.(
2014). The characters of Paleozoic jawed
vertebrates. Zoological Journal of the Linnean Society 170,779–821.
Brito,J.,Teillet,M.-A.&Le Douarin,N.M.(
2006). An early role for Sonic
hedgehog from foregut endoderm in jaw development: ensuring neural crest cell
survival. Proceedings of the National Academy of Sciences of the United States of America 103,
11607– 11612.
Brito,J.,Teillet,M.-A.&Le Douarin,N.M.(
2008). Induction of
mirror-image supernumerary jaws in chicken mandibular mesenchyme by Sonic
Hedgehog-producing cells. Development 135,2311–2319.
Carr,R.K.,Johanson, Z. & Ritchie,A.(
2009). The phyllolepid placoderm
Cowralepis mclachlani: insights into the evolution of feeding mechanisms in jawed
vertebrates. Journal of Morphology 270,775–804.
Cattell,M.,Lai, S., Cerny,R.&Medeiros,D.M.(
2011). A new mechanistic
scenario for the origin and evolution of vertebrate cartilage. PLoS ONE 6,e22474.
Cerny,R.,Cattell,M.,Sauka-Spengler,T.,Bronner-Fraser,M.,Yu,F.&
Meulemans Medeiros,D.(
2010). Evidence for the prepattern/cooption model of
vertebrate jaw evolution. Proceedings of the National Academy of Sciences of the United States
of America 107,17262–17267.
Cerny,R.,Lwigale,P.,Ericsson,R.,Meulemans,D.,Epperlein,H.-H.&
Bronner-Fraser,M.(
2004a). Developmental origins and evolution of jaws: new
interpretation of ‘‘maxillary’’ and ‘‘mandibular’’. Developmental Biology 276,225–236.
Cerny,R.,Meulemans,D.,Berger,J.,Wilsch-Bra ¨
uninger,M.,Kurth,T.,
Bronner-Fraser,M.&Epperlein,H.-H.(
2004b). Combined intrinsic and
extrinsic influences pattern cranial neural crest migration and pharyngeal arch
morphogenesis in axolotl. Developmental Biology 266,252–269.
Choe,C.P.&Crump,J.G.(
2014). Tbx1 controls the morphogenesis of
pharyngeal pouch epithelia through mesodermal Wnt11r and Fgf8a. Development
141,3583–3593.
Clouthier,D.E.,Hosoda,K.,Richardson,J.A.,Williams,S. C., Yanagisawa,
H., Kuwaki,T.,Kumada,M.,Hammer,R.E.&Yanagisawa,M.(
1998). Cranial
and cardiac neural crest defects in endothelin-A-deficient mice. Development 125,
813– 824.
Coates,M.I.(
2003). The evolution of paired fins. Theory in Biosciences 122,266–287.
Coates,M.I.&Cohn,M.J.(
1998). Fins, limbs, and tails: outgrowths and axial
patterning in vertebrate evolution. BioEssays 20,371–381.
Cohn,M.J.(
2002). Lamprey Hox genes and the origin of jaws. Nature 416,386–387.
Cole,F.J.(
1926). A Monograph on the general morphology of the myxinoid fishes,
based on a study of Myxine. Part VI. The morphology of the vascular system.
Transactions of the Royal Society of Edinburgh 54,309–342.
Compagnucci,C.,Debiais-Thibaud,M.,Coolen,M.,Fish,J.,Griffin,J.N.,
Bertocchini,F.,Minoux,M.,Rijli,F.M.,Borday-Birraux,V.,Casane,
D., Mazan, S. & Depew,M.J.(
2013). Pattern and polarity in the development
and evolution of the gnathostome jaw: both conservation and heterotopy in the
branchial arches of the shark, Scyliorhinus canicula.Developmental Biology 377,428–448.
Biological Reviews 91 (2016) 611– 657 ©2015 Cambridge Philosophical Society

652 Tetsuto Miyashita
Conway Morris,S.&Caron,J.-B.(2014). A primitive fish from the Cambrian of
North America. Nature 512,419–422.
Couly,G.F.,Coltey P. M. & Le Douarin,N.M.(
1993). The triple origin of skull
in higher vertebrates: a study in quail-chick chimeras. Development 117,409–429.
Couly,G.,Creuzet, S., Bannaceur, S., Vincent,C.&Le Douarin,N.M.
(2002). Interactions between Hox-negative cephalic neural crest cells and the foregut
endoderm in patterning facial skeleton in the vertebrate head. Development 125,
3445– 3459.
Couly,G.,Grapin-Botton,A.,Coltey,P.,Ruhin,B.&Le Douarin,N.
(1998). Determination of the identity of the derivatives of the cephalic neural
crest: incompatibility between Hox gene expression and lower jaw development.
Development 125,3445–3459.
Creuzet, S., Couly,G.&Le Douarin,N.M.(
2005). Patterning the neural crest
derivatives during development of the vertebrate head: insights from avian studies.
Journal of Anatomy 207,447–459.
Creuzet, S., Couly,G.,Vincent,C.&Le Douarin,N.M.(
2002). Negative
effect of Hox gene expression on the development of the neural crest-derived facial
skeleton. Development 129,4301–4313.
Creuzet, S., Schuler,B.,Couly,G.&Le Douarin,N.M.(
2004). Reciprocal
relationships between Fgf8 and neural crest cells in facial and forebrain development.
Proceedings of the National Academy of Sciences of the United States of America 101,4843–4847.
Crump,J.G.,Maves, L., Lawson,N.D.,Weinstein,B.M.&Kimmel,C.B.
(2004). An essential role for Fgfs in endodermal pouch formation influences later
craniofacial skeletal patterning. Development 131,5703–5716.
Dahmann,C.,Oates,A.C.&Brand,M.(
2011). Boundary formation and
maintenance in tissue development. Nature Reviews Genetics 12,43–55.
Davis, S. P., Finarelli,J.A.&Coates,M.I.(
2012). Acanthodes and shark-like
conditions in the common ancestor of modern gnathostomes. Nature 486,247–250.
Dean,B.(
1899). On the embryology of Bdellostoma stouti.Ageneralaccountof
myxinoid development from the egg and segmentation to hatching, Festschrift zum
70ten Geburststag Carl von Kupffer,Jena:GustavFischer,pp.220–276.
De Beer,G.(
1937). The Development of the Vertebrate Skull,OxfordUniversityPress,
London.
De La Rosa, L. N., M¨
uller,G.B.&Metscher,B.D.(
2014). The lateral
mesodermal divide: an epigenetic model of the origin of paired fins. Evolution &
Development 16,38–48.
Denison,R.H.(
1941). The soft anatomy of Bothriolepis.Journal of Paleontology 15,
553– 561.
Denison,R.H.(
1961). Feeding mechanisms of Agnatha and early gnathostomes.
American Zoologist 1,177–181.
Denison,R.H.(
1978). Placodermi. In Handbook of Paleoichthyology (Volume II,ed.
H.-P. Schultze). 128p. Gustav Fischer, Stuttgart.
Depew,M.J.,Lufkin,T.&Rubenstein, J. L. R. (2001). Specification of jaw
subdividions by Dlx genes. Science 298,381–385.
Depew,M.J.,Simpson,C.A.,Morasso,M.&Rubenstein, J. L. R. (2005).
Reassessing the Dlx code: the genetic regulation of branchial arch skeletal pattern
and development. Journal of Anatomy 207,501–561.
Diogo,R.,Hinits,Y.&Hughes, S. M. (2008). Development of mandibular, hyoid,
and hypobranchial muscles in the zebrafish: homologies and evolution of these
muscles within bony fishes and tetrapods. BMC Developmental Biology 8,24.
Donoghue,P.C.J.,Forey, P. L. & Aldridge,R.J.(
2000). Conodont affinity and
chordate phylogeny. Biological Reviews 75,191–251.
Donoghue,P.C.J.,Sansom,I.J.&Downs,J.P.(
2006). Early evolution of
vertebrate skeletal tissues and cellular interactions, and the canalization of skeletal
development. Journal of Experimental Zoology Part B 306,278–294.
Donoghue,P.C.J.&R¨
ucklin,M.(
2014). The ins and outs of the evolutionary
origin of teeth. Evolution & Development (doi: 10.1111/ede.12099).
Dude,C.M.,Kuan,C.Y.K.,Bradshaw,J.R.,Greene,N.D.,Relaix,F.,Stark,
M. R. & Baker,C.V.H.(
2009). Activation of Pax3 target genes is necessary but
not sufficient for neurogenesis in the ophthalmic trigeminal placode. Developmental
Biology 326,314–326.
Dupret,V.,Sanchez, S., Goujet,D.,Tafforeau,P.&Ahlberg,P.(
2014). A
primitive placoderm sheds light on the origin of the jawed vertebrate face. Nature
507,500–503.
Ebbehøj,J.,&Wagner,G.(
1979). Insufficient penile erection due to abnormal
drainage of cavernous bodies. Urology 13,507–510.
Eberhart,J.K.,Swartz,M.E.,Crump,J.G.&Kimmel,C.B.(
2006). Early
Hedgehog signaling from neural to oral epithelium organizes anterior craniofacial
development. Development 133,1069–1077.
Edgeworth,F.H.(
1935). The Cranial Muscles of Vertebrates. Cambridge
University Press.
Edlund,R.K.,Ohyama,T.,Kantarci,H.,Riley,B.B.&Groves,A.K.(
2014).
Foxi transcription factors promote pharyngeal arch development by regulating
formation of FGF signaling centers. Developmental Biology 390,1–13.
Elliott,D.K.(
2013). A new cyathaspid (Agnatha, Heterostraci) with an articulated
oral cover from the Late Silurian of the Canadian Arctic. Journal of Vertebrate
Paleontology 33,29–34.
Ericsson,R.,Knight,R.&Johanson, Z. (2013). Evolution and development of
the vertebrate neck. Journal of Anatomy 222,67–78.
Fava,R.A.,Olsen,N.J.,Spencer-Green,G.,Yeo, K.-T., Yeo, T.-K., Berse,B.,
Jackman,R.W.,Senger,D.R.,Dvorak,H.F.&Brown, L. F. (1994). Vascular
permeability factor/endothelial growth factor (VPF/VEGF): accumulation and
expression in human synovial fluids and rheumatoid synovial tissue. Journal of
Experimental Medicine 180,341–346.
Forey, P. L. (1995). Agnathans recent and fossil, and the origin of jawed vertebrates.
Reviews in Fish Biology and Fisheries 5,267–303.
Forey,P.L.&Gardiner,B.G.(
1986). Observations on Ctenurella (Ptyctodontida)
and the classification of placoderm fishes. Zoological Journal of the Linnean Society 86,
43– 74.
Forster,M.E.(
1997). The blood sinus system of hagfish: its significance in a
low-pressure circulation. Comparative Biochemistry and Physiology Part A, Physiology 116,
239– 244.
Fraser,G.J.,Cerny,R.,Soukup,V.,Bronner-Fraser,M.&Streelman,J.T.
(2010). The odontode explosion: the origin of tooth-like structures in vertebrates.
BioEssays 32,808–817.
Frisdal,A.&Trainor,P.A.(
2014). Development and evolution of the pharyngeal
apparatus. WIREs Developmental Biology 3,403–418.
Fujimoto, S., Oisi,Y.,Kuraku, S., Ota,K.G.&Kuratani, S. (2013).
Non-parsimonious evolution of hagfish Dlx genes. BMC Evolutionary Biology 13,
15.
Gagnier,P.-Y.(
1993). Sacabambaspis janvieri,Vert
´
ebr´
e ordovicien de Bolivie. 1. Analyse
morphologique. Annales de Pal´eontologie (Vert´ebr´es) 79,19–69.
Gai, Z., Donoghue,P.C.J.,Zhu,M.,Janvier,P.&Stampanoni,M.(
2011).
Fossil jawless fish from China foreshadows early jawed vertebrate anatomy. Nature
476,324–327.
Gardiner,B.G.(
1984). The relationships of placoderms. Journal of Vertebrate Paleontology
4,379–395.
Gardiner,B.G.&Miles, R. S. (1990). A new genus of eubrachythoracid arthrodire
from Gogo, Western Australia. Zoological Journal of the Linnean Society 99,159–204.
Gaskell,W.H.(
1908). On the Origin of Vertebrates. Longmans, Green: London and
New York.
Gegenbaur,C.(
1859). Grundz¨uge der vergleichenden Anatomie, Wilhelm
Engelmann: Leipzig.
Gess,R.W.,Coates,M.I.&Rubidge, B. S. (2006). A lamprey from the Devonian
period of South Africa. Nature 443,981–984.
Giles, S., Friedman,M.&Brazeau,M.D.(
2015). Osteichthyan-like
cranial conditions in an Early Devonian stem gnathostome. Nature (doi:
10.1038/nature14065).
Gillis,J.A.,Modrell, M. S. & Baker,C.V.H.(
2013). Developmental evidence
for serial homology of the vertebrate jaw and gill arch skeleton. Nature Communications
4,1436.
Gilmore,B.(
1992). Scroll coprolites from the Silurian of Ireland and the feeding of
early vertebrates. Palaeontology 35,319–333.
Gitton,Y.,Benouaiche, L., Vincent,C.,Heude,E.,Soulika,M.,Bouhali,
K., Couly,G.&Levi,G.(
2011). Dlx5 and Dlx6 expression in the anterior neural
fold is essential for patterning the dorsal nasal capsule. Development 138,897–903.
Goethe,J.W.(
1790). Das Sch¨
adelgr¨ut aus sechs Wirbelknochen aufgebaut. Zur
Naturwissenschaft ¨uberhaupt, besonders zur Morphologie (Vol. II). J. G. Gotta, Stuttgart.
Goodrich, E. S. (1930). Studies on the Structure and Development of Vertebrates.The
Macmillan Company, London.
Goudemand,N.,Orchard,M.J.,Urdy, S., Bucher,H.&Tafforeau,P.
(2011). Synchrotron-aided reconstruction of the conodont feeding apparatus and
implications for the mouth of the first vertebrates. Proceedings of the National Academy of
Sciences of the United States of America 108,8720–8724.
Goujet,D.(
1984). Les poisons Placodermes du Spitsberg. Arthrodires Dolichothoraci de la
Formation de Wood Bay (D´evonien inferieur),Cahiers de Pal´eontologie, Centre national de la
Recherche scientifique, Paris.
Graham,A.,Butts,T.,Lumsden,A.&Kiecker,C.(
2014). What can vertebrates
tell us about segmentation? EvoDevo 5,24.
Graham,A.,Francis-West,P.,Brickell,P.&Lumsden,A.(
1994). The signalling
molecule BMP4 mediates apoptosis in the rhombencephalic neural crest. Nature 372,
684– 686.
Graham,A.,Okabe,M.&Quinlan,R.(
2005). The role of the endoderm in
the development and evolution of the pharyngeal arches. Journal of Anatomy 207,
479– 487.
Grande, L. & Bemis,W.E.(
1998). A comprehensive phylogenetic study of amiid
fishes (Amiidae) based on comparative skeletal anatomy. An empirical search for
interconnected patterns of natural history. Memoir of the Society of Vertebrate Paleontology
4,1–690.
Gregory,W.K.(
1904). The relations of the anterior visceral arches to the
chondrocranium. The Biological Bulletin 7,55–69.
Grenier,J.,Teillet, M.-A., Grifone,R.,Kelly,R.G.&Duprez,D.(
2009).
Relationship between neural crest cells and cranial mesoderm during head muscle
development. PLoS ONE 4,e4381.
Biological Reviews 91 (2016) 611– 657 ©2015 Cambridge Philosophical Society

Fishing for jaws 653
Gross,W.(1963). Gemuendina stuertzi Traquair. Neuuntersuchung. Notizblatt des
Hessischen Landesamtes f¨ur Bodenforschung zu Wiesbaden 91,36–73.
Gross,W.(
1971). Downtonische und dittonische acanthodier-reste des ostseegebietes.
Palaeontographica Abteilung A 136,1–82.
Gross,J.B.&Hanken,J.(
2008). Segmentation of the vertebrate skull: neural-crest
derivation of adult cartilages in the clawed frog, Xenopus laevis.Integrative and Comparative
Biology 48,681–696.
Hall,B.K.(
2009). The Neural Crest and Neural Crest Cells in Vertebrate Development and
Evolution. Springer, New York.
Halstead, L. B. (1971). The presence of a spiracle in the Heterostraci (Agnatha).
Zoological Journal of the Linnean Society 50,195–197.
Halstead, L. B. (1973). The heterostracan fishes. Biological Reviews 48,279–332.
Halstead Tarlo,L.B.&Whitting,H.P.(
1965). A new interpretation of the
internal anatomy of the Heterostraci (Agnatha). Nature 206,148–150.
Haworth,K.E.,Wilson,J.M.,Grevellec,A.,Cobourne,M.T.,Healy,C.,
Helms,J.A.,Sharpe,P.T.&Tucker, A. S. (2007). Sonic hedgehog in the
pharyngeal endoderm controls arch pattern via regulation of Fgf8 in head ectoderm.
Developmental Biology 303,244–258.
He,F.,Hu,X.,Xiong,W.,Li, L., Lin, L., Shen,B.,Yang, L., Gu, S., Zhang,Y.&
Chen,Y.(
2014). Directed Bmp4 expression in neural crest cells generates a genetic
model for the rare human bony syngnathia birth defect. Developmental Biology 391,
170– 181.
Heimberg,A.M.,Cowper-Sal-Lari,R.,S´
emon,M.,Donoghue,P.C.J.
&Peterson,K.J.(
2010). MicroRNAs reveal the interrelationships of hagfish,
lampreys, and gnathostomes and the nature of the ancestral vertebrate. Proceedings of
the National Academy of Sciences of the United States of America 107,19379–19383.
Hern´
andez, L. P., Barresi,M.J.F.&Devoto,S. H. (2002). Functional morphology
and developmental biology of zebrafish: reciprocal illumination from an unlikely
couple. Integrative and Comparative Biology 42,222–231.
Herrick,C.J.(
1899). The cranial and first spinal nerves of menidia: a contribution
upon the nerve components of the bony fishes. Journal of Comparative Neurology 9,
153– 455.
Herzog,W.,Sonntag,C.,Von Der Hardt, S., Roehl,H.H.,Varga, Z. M.
&Hammerschmidt,M.(
2004). Fgf3 signaling from the ventral diencephalon
is required for early specification and subsequent survival of the zebrafish
adenohypophysis. Development 131,3681–3692.
Herzog,W.,Zeng,X.,Lele, Z., Sonntag,C.,Ting, J.-W., Chang,C.-Y.&
Hammerschmidt,M.(
2003). Adenohypophysis formation in the zebrafish and its
dependence on Sonic hedgehog. Developmental Biology 254,36–49.
Heude,´
E., Bouhali,K.,Kurihara,Y.,Kurihara,H.,Couly,G.,Janvier,P.
&Levi,G.(
2010). Jaw muscularization requires Dlx expression by cranial neural
crest cells. Proceedings of the National Academy of Sciences of the United States of America 107,
11441– 11446.
Higashiyama,H.&Kuratani, S. (2014). On the maxillary nerve. Journal of
Morphology 275,17–38.
Hinits,Y.,Williams,V.C.,Sweetman,D.,Donn,T.D.,Ma,T.P.,Moens,C.B.
&Hughes, S. M. (2011). Defective cranial skeletal development, larval lethality and
haploinsufficiency in Myod mutant zebrafish. Developmental Biology 358,102–112.
Holmgren,N.(
1946). On two embryos of Myxine glutinosa.Acta Zoologica 27,1-90.
Horigome,N.,Myojin,M.,Ueki,T.,Hirano, S., Aizawa, S. & Kuratani, S.
(1999). Development of cephalic neural crest cells in embryos of Lampetra japonica,
with special reference to the evolution of the jaw. Developmental Biology 207,287–308.
Hughes,G.M.(
1960a). A comparative study of gill ventilation in marine teleosts.
Journal of Experimental Biology 37,28–45.
Hughes,G.M.(
1960b). The mechanism of gill ventilation in the dogfish and skate.
Journal of Experimental Biology 37,11–27.
Hunt,P.,Gulisano,M.,Cook,M.,Sham,M.H.,Faiella,A.,Wilkinson,D.,
Boncinelli,E.&Krumlauf,R.(
1991a). A distinct Hox code for the branchial
region of the vertebrate head. Nature 353,861–864.
Hunt,P.,Whiting,J.,Muchamore,I.,Marshall,H.&Krumlauf,R.(
1991b).
Homeobox genes and models for patterning the hindbrain and branchial arches.
Development 1(Suppl.), 187– 196.
Huxley,T.H.(
1864). Lectures on the Elements of Comparative Anatomy on the Classification of
Animals and the Vertebrate Skull, John Churchill & Sons, London.
Huysseune,A.,Sire,J.Y.&Witten,P.E.(
2009). Evolutionary and developmental
origins of the vertebrate dentition. Journal of Anatomy 214,465–476.
Jandzik,D.,Hawkins,M.B.,Cattell,M.V.,Cerny,R.,Square,T.A.&
Medeiros,D.M.(
2014). Roles for FGF in lamprey pharyngeal pouch formation
and skeletogenesis highlight ancestral functions in the vertebrate head. Development
141,629–638.
Janvier,P.(
1974). The structure of the naso-hypophyseal complex, and the mouth in
fossil and extant cyclostomes, with remarks on amphiaspiforms. Zoologica Scripta 3,
193– 200.
Janvier,P.(
1981). Norselaspis glacialis n.g., n.sp. et les relations phylog´
en´
etiques entre
les Kiaeraspidiens (Osteostraci) du D´
evonien inf´
eerieur du Spitsberg. Palaeovertebrata
11,19–131.
Janvier,P.(
1984). The relationships of the Osteostraci and Galeaspida. Journal of
Vertebrate Paleontology 4,344–358.
Janvier,P.(
1985a). Les C´ephalaspides du Spitsberg. Anatomie, phylog´enie et syst´ematique
des Ost´eostrac´es siluro-d´evoniens. R´evision des Ost´eostrac´es de la Formation de Wood Bay
(D´evonien inf´erieur du Spitsberg),Cahiers de Pal´eontologie, Centre national de la Recherche
scientifique, Paris.
Janvier,P.(
1985b). Les Thyestidiens (Ost´eostraci) du Silurien de Saaremaa (Estonie). Premi`
ere
partie: morphologie et anatomie. Annales de Pal´eontologie 71,83–147
Janvier,P.(
1993). Patterns of diversity in the skull of jawless fishes. In The
Skull: Patterns of Structural and Systematic Diversity (Volume II,eds J. Hanken and
B. K. Hall), pp. 131– 188. The University of Chicago Press: Chicago.
Janvier,P.(
1995). The brachial articulation and pectoral fin in antiarchs (Placodermi).
Bulletin du Mus´eum National d’Histoire Naturelle 4e s´erie 17,143–162.
Janvier,P.(
1996). Early Vertebrates.ClarendonPress:Oxford.
Janvier,P.(
2004). Early specializations in the branchial apparatus of jawless
vertebrates: a consideration of gill number and size. In Recent Advances in the Origin and
Early Radiation of Vertebrates (eds G. Arratia,M.V.H.Wilson and R. Cloutier),
pp. 29– 52. Verlag Dr. Friedrich Pfeil: Munich.
Janvier,P.(
2007). Homologies and evolutionary transitions in early vertebrate history.
In Major Transitions in Vertebrate Evolution (eds J. S. Anderson and H. D. Sues), pp.
57– 121. Indiana University Press: Bloomington.
Janvier,P.(
2008). Early jawless vertebrates and cyclostome origins. Zoological Science
25,1045–1056.
Janvier,P.(
2010). microRNAs revive old views about jawless vertebrate divergence
and evolution. Proceedings of the National Academy of Sciences of the United States of America
107,19137–19138.
Janvier,P.&Arsenault,M.(
2002). Calcification of early vertebrate cartilage. Nature
417,609.
Janvier,P.&Arsenault,M.(
2007). The anatomy of Euphanerops longaevus
Woodward, 1900, an anaspid-like jawless vertebrate from the Upper Devonian
of Miguasha, Quebec, Canada. Geodiversitas 29,143–216.
Janvier,P.,Arsenault,M.&Desbiens, S. (2004). Calcified cartilage in the paired
fins of the osteostracan Escuminaspis laticeps (Traquair 1880), from the Late Devonian
of Miguasha (Qu´
ebec, Canada), with a consideration of the early evolution of the
pectoral fin endoskeleton in vertebrates. Journal of Vertebrate Paleontology 24,773–779.
Janvier,P.&Blieck,A.(
1993). L. B. Halstead and the heterostracan controversy.
Modern Geology 18,89–105.
Janvier,P.,Desbiens, S., Willett,J.A.&Arsenault,M.(
2006). Lamprey-like
gills in a gnathostome-related Devonian jawless vertebrate. Nature 440,1183–1185.
Jarvik,E.(
1980). Basic Structure and Evolution of Vertebrates (II Volumes). Academic Press:
London.
Johanson, Z. (2002). Vascularization of the osteostracan and antiarch (Placodermi)
pectoral fin: similarities, and implications for placoderm relationships. Lethaia 35,
169– 186.
Johanson, Z. (2003). Placoderm branchial and hypobranchial muscles and the origins
in jawed vertebrates. Journal of Vertebrate Paleontology 23,735–749.
Johanson, Z. (2010). Evolution of paired fins and the lateral somitic frontier. Journal
of Experimental Zoology B 314,347–352.
Johanson, Z. & Smith,M.M.(
2005). Origin and evolution of gnathostome dentitions:
aquestionofteethandpharyngealdenticlesinplacoderms.Biological Reviews 80,
1–43.
Johnels,A.G.(
1948). On the development and morphology of the skeleton of the
head of Petromyzon.Acta Zoologica 29,139–279.
Jollie,M.(
1971). A theory concerning the early evolution of the visceral arches. Acta
Zoologica 52,85–96.
Jones,D.,Evans,A.R.,Siu,K.K.,Rayfield,E.J.&Donoghue,P.C.J.(
2012).
The sharpest tools in the box? Quantitative analysis of conodont element functional
morphology Proceedings of the Royal Society of London Series B doi: 10.1098/rspb20120147.
Kawamura,K.,Kouki,T.,Kawahara,G.&Kikuyama, S. (2002). Hypophyseal
development in vertebrates from amphibians to mammals. General and Comparative
Endocrinology 126,130–135.
Khonsari,R.H.,Seppala,M.,Pradel,A.,Dutel,H.,Clement,G.,
Lebedev,A.,Ghafoor, S., Rothova,M.,Tucker,A.,Maisey,J.G.,Fan,
C.-M., Kawasaki,M.,Ohazama,A.,Tafforeau,P.,Franco,B.,Helms,J.,
Haycraft,C.J.,David,A.,Janvier,P.,Cobourne,M.T.&Sharpe,P.T.
(2013). The buccohypophyseal canal is an ancestral vertebrate trait maintained by
modulation of sonic hedgehog signaling. BMC Biology 11,27.
Kimmel,C.B.&Eberhart,J.K.(
2008). The midline, oral ectoderm, and the arch-0
problem. Integrative and Comparative Biology 48,668–680.
Kimmel,C.B.,Miller,C.T.&Keynes,R.J.(
2001). Neural crest patterning and
the evolution of the jaw. Journal of Anatomy 199,105–119.
Kimmel,C.B.,Ullmann,B.,Walker,M.,Miller,V.T.&Crump,J.G.(
2003).
Endothelin 1-mediated regulation of pharyngeal bone development in zebrafish.
Development 130,1339–1351.
K¨
ontges,G.&Lumsden,A.(
1996). Rhombencephalic neural crest segmentation is
preserved throughout craniofacial ontogeny. Development 122,3229–3242.
Koop,D.,Chen,J.,Theodosiou,M.,Carvalho,J.E.,Alvarez, S., De Lera,A.
R., Holland, L. Z. & Schubert,M.(
2014). Roles of retinoic acid and Tbx1/10 in
pharyngeal segmentation: amphioxus and the ancestral chordate condition. EvoDevo
5,36.
Biological Reviews 91 (2016) 611– 657 ©2015 Cambridge Philosophical Society

654 Tetsuto Miyashita
Kuraku, S. (2013). Impact of asymmetric gene repertoire between cyclostomes and
gnathostomes. Seminars in Cell & Developmental Biology 24,119–127.
Kuraku, S., Takio,Y.,Sugahara,F.,Takechi,M.&Kuratani, S. (2010).
Evolution of oropharyngeal patterning mechanisms involving Dlx and endothelins in
vertebrates. Developmental Biology 341,315–323.
Kuratani, S. (1997). Spatial distribution of postotic crest cells defines the head/trunk
interface of the vertebrate body: embryological interpretation of peripheral nerve
morphology and evolution of the vertebrate head. Anatomia Embryologia 195,1–13.
Kuratani, S. (2004). Evolution of the vertebrate jaw: comparative embryology and
molecular developmental biology reveal the factors behind evolutionary novelty.
Journal of Anatomy 205,335–347.
Kuratani, S. (2005). Developmental studies of the lamprey and hierarchical
evolutionary steps towards the acquisition of the jaw. Journal of Anatomy 207,
489– 499.
Kuratani, S. (2008). Evolutionary developmental studies of cyclostomes and the
origin of the vertebrate neck. Development, Growth & Differentiation 50, S189– S194.
Kuratani, S. (2012). Evolution of the vertebrate jaw from developmental perspectives.
Evolution & Development 14,76–92.
Kuratani, S., Adachi,N,Wada,N.,Oisi,Y.&Sugahara,F.(
2013).
Developmental and evolutionary significance of the mandibular arch and
prechordal/premandibular cranium in vertebrates: revising the heterotopy scenario
of gnathostome jaw evolution. Journal of Anatomy 222,41–55.
Kuratani, S., Horigome,N.&Hirano, S. (1999). Developmental morphology of
the head mesoderm and reevaluation of segmental theories of the vertebrate head:
evidence from embryos of an agnathan vertebrate, Lampetra japonica.Developmental
Biology 210,381–400.
Kuratani, S., Nobusada,Y.,Horigome,N.&Shigetani,Y.(
2001). Embryology
of the lamprey and evolution of the vertebrate jaw: insights from molecular and
developmental perspectives. Philosophical Transactions of the Royal Society B 356,15–32.
Kuratani, S., Murakami,Y.,Nobusada,Y.,Kusakabe,R.&Hirano, S.
(2004). Developmental fate of the mandibular mesoderm in the lamprey, Lethenteron
japnicum:comparativemorphologyanddevelopmentofthegnathostomejawwith
special reference to the nature of the trabecular cranii. Journal of Experimental Zoology
B302,458–468.
Kuratani, S., Ueki,T.,Aizawa, S. & Hirano, S. (1997). Peripheral development
of cranial nerves in a cyclostome, Lampetra japonica: morphological distribution of
nerve branches and the vertebrate body plan. Journal of Comparative Neurology 384,
483– 500.
Kusakabe,R.,Kuraku, S. & Kuratani, S. (2011). Expression and interaction of
muscle-related genes in the lamprey imply the evolutionary scenario for vertebrate
skeletal muscle, in association with the acquisition of the neck and fins. Developmental
Biology 350,217–227.
Kusakabe,R.&Kuratani,S. ( 2005). Evolution and developmental patterning of the
vertebrate skeletal muscles: perspectives from the lamprey. Developmental Dynamics,
234,824–834.
Kusakabe,R.&Kuratani, S. (2007). Evolutionary perspectives from development
of mesodermal components in the lamprey. Developmental Dynamics 236,2410–2420.
Lakiza,O.,Miller, S., Bunce,A.,Lee,E.M.J.&McCauley,D.W.(
2011). SoxE
gene duplication and development of the lamprey branchial skeleton: Insights into
development and evolution of the neural crest. Developmental Biology 359,149–161.
Laurent,P.&Dunel-Erb, S. (1984). The pseudobranch: morphology and function.
Fish Physiology 10B,285–323.
Lee, S.-H., B´
edard,O.,Buchtov´
a,M.,Fu,K.&Richman,J.(
2004). A new origin
for the maxillary jaw. Developmental Biology 276,207–224.
Lee, S.-H., Fu,K.K.,Hui,J.N.&Richman,J.M.(
2001). Noggin and retinoic acid
transform the identity of avian facial prominences. Nature 414,909–912.
Lindstr¨
om,T.(
1949). On the cranial nerves of the cyclostomes with special reference
to n. trigeminus. Acta Zoologica 30,315–458.
Long,J.A.,Mark-Kurik,E.,Johanson, Z., Lee, M. S. Y., Young,G.C.,Zhu,
M., Ahlberg,P.E.,Newman,M.,Jones,R.,Den Blaauwen,J.,Choo,B.
&Trinajstic,K.(
2014). Copulation in antiarch placoderms and the origin of
gnathostome internal fertilization. Nature 517,196–199.
Lours-Calet,C.,Alvares, L. E., El-Hanfy, A. S., Gandesha, S., Walters,
E. H., Sobreira,D.R.,Wotton,K.R.,Jorge,E.C.,Lawson,J.A.,
Lewis,A.K.,Tada,M.,Sharpe,C.,Kardon,G.&Dietrich, S. (2014).
Evolutionarily conserved morphogenetic movements at the vertebrate head– trunk
interface coordinate the transport and assembly of hypopharyngeal structures.
Developmental Biology 390,231–246.
Mallatt,J.(
1979). Surface morphology and functions of pharyngeal structures in the
larval lamprey Petromyzon marinus. Journal of Morphology 162,249–273.
Mallatt,J.(
1981). The suspension feeding mechanism of the larval lamprey Petromyzon
marinus.Journal of Zoology 194,103–142.
Mallatt,J.(
1984a). Early vertebrate evolution: pharyngeal structure and the origin
of gnathostomes. Journal of Zoology 204,169–183.
Mallatt,J.(
1984b). Feeding ecology of the earliest vertebrates. Zoological Journal of the
Linnean Society 82,261–272.
Mallatt,J.(
1996). Ventilation and the origin of jawed vertebrates: a new mouth.
Zoological Journal of the Linnean Society 117,329–404.
Mallatt,J.(
1997). Shark pharyngeal muscles and early vertebrate evolution. Acta
Zoologica 78,279–294.
Mallatt,J.(
2008). The origin of the vertebrate jaw: neoclassical ideas versus newer,
development-based ideas. Zoological Science 25,990–998.
Marinelli,W.&Strenger,A.(
1954). Vergleichende Anatomie und Morphologie der
Wirbeltiere. I Lieferung. Petromyzon marinus (L),1–80.
Marinelli,W.&Strenger,A.(
1956). Vergleichende Anatomie und Morphologie der
Wirbeltiere. II Lieferung. Myxine glutinosa (L).81–172.
Matsuo,I.,Kuratani, S., Kimura,C.,Takeda,N.&Aizawa, S. (1995). Mouse
Otx2 functions in the formation and patterning of rostral head. Genes & Development
9,2646–2658.
Matsuoka,T.,Ahlberg,P.E.,Kessaris,N.,Iannarelli,P.,Dennehy,U.,
Richardson,W.D.,McMahon,A.P.&Koentges,G.(
2005). Neural crest
origins of the neck and shoulder. Nature 436,347–355.
McCabe,K.L.&Bronner-Fraser,M.(
2008). Essential role for PDGF signaling in
ophthalmic trigeminal placode induction. Development 135,1863–1874.
McCauley,D.W.&Bronner-Fraser,M.(
2003). Neural crest contributions to the
lamprey head. Development 130,2317–2327.
McCauley,D.W.&Bronner-Fraser,M.(
2006). Importance of SoxE in neural
crest development and the evolution of the pharynx. Nature 441,750–752.
Medeiros,D.M.&Crump,J.G.(
2012). New perspectives on pharyngeal dorsoventral
patterning in development and evolution of the vertebrate jaw. Developmental Biology
371,121–135.
Meulemans,D.&Bronner-Fraser,M.(
2002). Amphioxus and lamprey AP-2
genes: implications for neural crest evolution and migration patterns. Development
129,4953–4962.
Miles, R. S. (1968). Jaw articulation and suspension in Acanthodes and their significance.
In Current Problems of Lower Vertebrate Phylogeny, Nobel Symposium 4 (ed. T. Ørvig),
pp. 109– 127. Almqvist & Wiksell, Stockholm.
Miles, R. S. (1969). Features of placoderm diversification and the evolution of
the arthrodire feeding mechanism. Transactions of the Royal Society of Edinburgh 68,
123– 170.
Miles, R. S. & Westoll, T. S. (1968). The placoderm fish Coccosteus cuspidatus Miller
ex Agassiz from the Middle Old Red Sandstone of Scotland. Part I. Descriptive
morphology. Transactions of the Royal Society of Edinburgh 67,373–464.
Miller,C.T.,Schilling,T.F.,Lee,K.-H.,Parker,J.&Kimmel,C.B.
(2000). sucker encodes a zebrafish Endothelin-1 required for ventral pharyngeal arch
development. Development 127,3815–3828.
Miller,C.T.,Yellon,D.,Stainier,D.Y.R.&Kimmel,C.B.(
2003). Two
endothelin 1 effectors, hand2 and bapx1,patternventralpharyngealcartilageandthe
jaw joint. Development 130,1353–1365.
Minoux,M.,Antonarakis, G. S., Kmita,M.,Duboule,D.&Rijli,F.M.(
2009).
Rostral and caudal pharyngeal arches share a common neural crest ground pattern.
Development 136,637–645.
Minoux,M.&Rijli,F.M.(
2010). Molecular mechanisms of cranial neural crest cell
migration and patterning in craniofacial development. Development 137,2605–2621.
Miyashita,T.(
2012). Comparative analysis of the anatomy of the myxinoidea and the ancestry
of early vertebrate lineages.UnpublishedMScThesis:UniversityofAlberta,Edmonton.
Modrell, M. S., Hockman,D.,Uy,B.,Buckley,D.,Saula-Spengler,T.,
Bronner,M.E.&Baker,C.V.H.(
2014). A fate-map for cranial sensory ganglia
in the sea lamprey. Developmental Biology 385,405–416.
Mootoosamy,R.C.&Dietrich, S. (2002). Distinct regulatory cascades for head
and trunk myogenesis. Development 129,573–583.
Murdock,D.J.E.,Dong,X.-P.,Repetski,J.E.,Marone,F.,Stampanoni,M.&
Donoghue,P.C.J.(
2013). The origin of conodonts and of vertebrate mineralized
skeletons. Nature 502,546–549.
Myojin,M.,Ueki,T.,Sugahara,F.,Murakami,Y.,Shigetani,Y.,Aizawa, S.,
Hirano, S. & Kuratani, S. (2001). Isolation of Dlx and Emx gene cognates in an
agnathan species, Lampetra japonica,andtheirexpressionpatternsduringembryonic
and larval development: conserved and diversified regulatory patterns of homeobox
genes in vertebrate head evolution. Journal of Experimental Zoology B 291,68–84.
Nair, S., Li,W.,Cornell,R.&Schilling,T.F.(
2007). Requirements for
Endothelin type-A receptors and Endothelin-1 signaling in the facial ectoderm
for the patterning of skeletogenic neural crest cells in zebrafish. Development 134,
335– 345.
Neidert,A.H.,Virupannavar,V.,Hooker,G.W.&Langeland,J.A.(
2001).
Lamprey Dlx genes and early vertebrate evolution. Proceedings of the National Academy
of Sciences of the United States of America 98,1665–1670.
Neuhauss, S. C. F., Solnica-Krezel, L., Schier,A.F.,Zwartkruis,F.,Stemple,
D. L., Malicki,J.,Abdelilah, S., Stainier,D.Y.R.&Driever,W.(
1996).
Mutations affecting craniofacial development in zebrafish. Development 123,357–367.
Nichols,J.T.,Pan, L., Moens,C.B.&Kimmel,C.B.(
2013). barx1 represses joints
and promotes cartilage in the craniofacial skeleton. Development 140,2765–2775.
Nissen,R.M.,Amsterdam,A.&Hopkins,N.(
2006). A zebrafish screen for
craniofacial mutants identifies wdr68 as a highly conserved gene required for
endothelin-1 expression. BMC Developmental Biology 6,28.
Noden,D.M.(
1983). The role of the neural crest in patterning cranial skeletal,
connective, and muscle tissues. Developmental Biology 96,144–165.
Biological Reviews 91 (2016) 611– 657 ©2015 Cambridge Philosophical Society

Fishing for jaws 655
Noden,D.M.&Francis-West,P.(2006). The differentiation and morphogenesis
of craniofacial muscles. Developmental Dynamics 235,1194–1218.
Norris,H.W.&Hughes, S. P. (1920). The spiracular sense-organ in elasmobranchs,
ganoids and dipnoans. The Anatomical Record 18,205–209.
Northcutt,R.G.(
2008). Historical hypotheses regarding segmentation of the
vertebrate head. Integrative and Comparative Biology 48,611–619.
Nowicki, J. L., Takimoto,R.&Burke,A.C.(
2003). The lateral somitic
frontier: dorso-ventral aspects of anterio-posterior regionalization in avian embryos.
Mechanisms of Development 120,227–240.
Oisi,Y.,Fujimoto, S., Ota,K.G.&Kuratani, S. (2015). On the peculiar
morphology and development of the hypoglossal, glossopharyngeal and vagus
nerves and hypobranchial muscles in the hagfish. Zoological Letters 1,6.
Oisi,Y.,Ota,K.G.,Fujimoto, S. & Kuratani, S. (2013a). Development of
the chondrocranium in hagfishes, with special reference to the early evolution of
vertebrates. Zoological Science 30,944–961.
Oisi,Y.,Ota,K.G.,Kuraku, S., Fujimoto, S. & Kuratani, S. (2013b). Craniofacial
development of hagfishes and the evolution of vertebrates. Nature 493,175–180.
Oken, L. (1807). ¨
Uber die Bedeutung der Sch¨adelknochen.G
¨
obhardt Bamberg, Jena.
Onai,T.,Irie,N.&Kuratani, S. (2014). The evolutionary origin of the vertebrate
body plan: the problem of head segmentation. Annual Review of Genomics and Human
Genetics 15,443–459.
O’Neill,P.,Mak, S.-S., Fritzsch,B.,Ladher,R.K.&Baker,C.V.H.(
2012).
The amniote paratympanic organ develops from a previously undiscovered sensory
placode. Nature Communications 3,1041.
O’Neill,P.,McCole,R.B.&Baker,C.V.H.(
2007). A molecular analysis
of neurogenic placode and cranial sensory ganglion development in the shark,
Scyliorhinus canicula.Developmental Biology 304,156–181.
Onimaru,K.,Shoguchi,E.,Kuratani, S. & Tanaka,M.(
2011). Development
and evolution of the lateral plate mesoderm: comparative analysis of amphioxus and
lamprey with implications for the acquisition of paired fins. Developmental Biology 359,
124– 136.
Ørvig,T.(
1973). Acanthodian dentition and its bearing on the relationships of the
group. Palaeontographica Abteilung A 143,119–150.
Ørvig,T.(
1980). Histological studies of ostracoderms, placoderms and fossil
elasmobranchs. 4. Ptyctodontid tooth plates and their bearing on holocephalian
ancestry: the condition of Ctenurella and Ptyctodus.Zoologica Scripta 9,219–239.
Owen,R.(
1848). On the Archetype and Homologies of the Vertebrates Skeleton.Johnvan
Voorst, London.
Parker,W.K.(
1878). On the structure and development of the skull in sharks and
skates. Philosophical Transactions of the Zoological Society of London 10,189–234.
Parker,H.J.,Bronner,M.E.&Krumlauf,R.(
2014). A Hox regulatory network of
hindbrain segmentation is conserved to the base of vertebrates. Nature 514,490–493.
Pasqualetti,M.,Ori,M.,Nardi,I.&Rijli,F.M.(
2000). Ectopic Hoxa2 induction
after neural crest migration results in homeosis of jaw elements in Xenopus.Development
127,5367–5378.
Paulsen,F.P.,Thale,A.B.,Hallmann,U.J.,Schaudig,U.&Tillmann,B.
N. (2000). The cavernous body of the human efferent tear ducts: function in tear
outflow mechanism. Investigative Ophthalmology & Visual Science 41,965–970.
Payne, S. L., Holliday,C.M.&Vickaryous,M.K.(
2011). An osteological
and histological investigation of cranial joints in geckos. The Anatomical Record 294,
399– 405.
Pieper,M.,Eagleson,G.W.,Wosniok,W.&Schlosser,G.(
2011). Origin and
segregation of cranial placodes in Xenopus laevis.Developmental Biology 360,257–275.
Piotrowski,T.&N¨
usslein-Volhard,C.(
2000). The endoderm plays an important
role in patterning the segmented pharyngeal region in zebrafish (Danio rerio).
Developmental Biology 225,339–356.
Piotrowski,T.,Schilling,T.F.,Brand,M.,Jiang,Y.-J.,Heisenberg,C.-P.,
Beuchle,D.,Grande,H.,Van Eeden,J.M.,Furutani-Seiki,M.,Granato,
M., Haffter,P.,Hammerschmidt,M.,Kane,D.A.,Kelsh,R.N.,Mullins,
M. C., Odenthal,J.,Warga,R.M.&N¨
usslein-Volhard,C.(
1996). Jaw and
branchial arch mutants in zebrafish II: anterior arches and cartilage differentiation.
Development 123,345–356.
Pitsillides,A.A.&Beier,F.(
2011). Cartilage biology in osteoarthritis – lessons
from developmental biology. Nature Reviews Rheumatology 7,654–663.
Pradel,A.,Maisey,J.G.,Tafforeau,P.,Mapes,R.H.&Mallatt,J.(
2014). A
Palaeozoic shark with osteichthyan-like branchial arches. Nature 509,608–611.
Purnell,M.A.(
1995). Microwear on conodont elements and macrophagy in the first
vertebrates. Nature 374,798–800.
Purnell,M.A.(
2002). Feeding in extinct jawless heterostracan fishes and testing
scenarios of early vertebrate evolution. Proceedings of the Royal Society of London B 269,
83– 88.
Purnell,M.A.&Donoghue,P.C.J.(
1997). Architecture and functional
morphology of the skeletal apparatus of ozarkodinid conodonts. Philosophical
Transactions of the Royal Society of London, Series B 352,1545–1564.
Purnell,M.A.&Jones,D.(
2012). Quantitative analysis of conodont tooth wear and
damage as a test of ecological and functional hypotheses. Paleobiology 38,605–626.
Qu,Q.,Sanchez, S., Blom,H.,Tafforeau,P.&Ahlberg,P.E.(
2013). Scales
and tooth whorls of ancient fishes challenge distinction between external and oral
‘teeth’. PLoS ONE 8,e71890.
Rathke,H.(
1827). Bemerkungen ¨uber den innern Bau des Qerders (Ammocoetes branchialis)
und des kleinen Neunauges ( Petromyzon planeri)(Bd II), Neueste Schriften der Naturf.
Gesellsch. Danzog.
Renaud,C.B.(
2011). Lampreys of the world. An annotated and illustrated catalogue
of lamprey species known to date. FAO Species Catalogue for Fisheries Purposes 5,1–109.
Richman,J.M.,Herbert,M.,Matovinovic,E.&Walin,J.(
1997). Effect of
fibroblast growth factors on outgrowth of facial mesenchyme. Developmental Biology
189,135–147.
Rijli,F.M.,Mark,M.,Lakkaraju, S., Dietrich,A.,Doll ´
eP&Chambon,P.
(1993). A homeotic transformation is generated in the rostral branchial region of the
head by disruption of Hoxa-2, which acts as a selector gene. Cell 75,1333–1349.
Rinon,A.,Lazar, S., Marshall,H.,B¨
uchmann-Møller, S., Neufeld,A.,
Elhanany-Tamir,H.,Taketo,M.M.,Sommer, L., Krumlauf,R.&
Tzahor,E.(
2007). Cranial neural crest cells regulate head muscle patterning
and differentiation during vertebrate embryogenesis. Development 134,3065–3075.
R¨
ucklin,M.,Donoghue,P.C.J.,Johanson, Z., Trinajstic,K.,Marone,F.
&Stampanoni,M.(
2012). Development of teeth and jaws in the earliest jawed
vertebrates. Nature 491,748–751.
R¨
ucklin,M.,Giles, S., Janvier,P.&Donoghue,P.C.J.(
2011). Teeth before
jaws? Comparative analysis of the structure and development of the external and
internal scales in the extinct jawless vertebrate Loganella scotica.Evolution & Development
13,523–532.
Ruest, L.-B. & Clouthier,D.E.(
2009). Elucidating timing and function of
endothelin-A receptor signaling during craniofacial development using neural crest
cell-specific gene depletion and receptor antagonism. Developmental Biology 328,
94– 108.
Ruest, L.-B., Xiang,X.,Lim, K.-C., Levi,G.&Clouthier,D.E.(
2004).
Endothelin-A receptor-dependent and -independent pathways in establishing
mandibular identity. Development 131,4413–4423.
Ruhin,B.,Creuzet, S., Vincent,C.,Beunouaiche, L., Le Douarin,N.M.&
Couly,G.(
2003). Patterning of the hyoid cartilage depends upon signals arising
from the ventral foregut endoderm. Developmental Dynamics 228,239–246.
Saint-Jeannet,J.-P.&Moody, S. A. (2014). Establishing the pre-placodal region
and breaking it into placodes with distinct identities. Developmental Biology 389,13–27.
Sambasivan,R.,Gayraud-Morel,B.,Dumas,G.,Cimper,C.,Paisant,S.,
Kelly,R.&Tajbakhsh,S. (2009). Distinct regulatory cascades govern extraocular
and pharyngeal arch muscle progenitor cell fates. Developmental Cell 16,810–821.
Sambasivan,R.,Kuratani, S. & Tajbakhsh, S. (2011). An eye on the head: the
development and evolution of craniofacial muscles. Development 138,2401-2415.
Sanchez, S., Dupret,V.,Tafforeau,P.,Trinajstic,K.M.,Ryll,B.,
Gouttenoire,P.-J.,Wretman, L., Zylberberg, L., Peyrin,F.&Ahlberg,
P. E. (2013). 3D microstructural architecture of muscle attachemnts in extant and
fossil vertebrates revealed by synchrotron microtomography. PLoS ONE 8,e56992.
Sansom, R. S., Freedman,K.,Gabbott, S., Aldridge,R.J.&Purnell,M.
A. (2010). Taphonomy and affinity of an enigmatic Silurian vertebrate, Jamoytius
kerwoodi White. Palaeontology 53,1393–1409.
Sansom, R. S., Gabbott, S. E. & Purnell,M.A.(
2013). Unusual anal fin in a
Devonian jawless vertebrate reveals complex origins of paired appendages. Biology
Letters 9,20130002(doi:10.1098/rsbl.2013.0002).
Sarker, S., Petiot,A.,Copp,A.,Farretti,P.&Thorogood,P.(
2001). FGF2
promotes skeletogenic differentiation of cranial neural crest cells. Development 128,
2143– 2152.
Schaeffer,B.&Thomson, K. S. (1980). Reflections on agnathan –gnathostome
relationships. In Aspects of Vertebrate History (ed. L. L. Jacobs), pp. 19– 33. Museum
of Northern Arizona Press: Flagstaff.
Schilling,T.F.&Kimmel,C.B.(
1997). Musculoskeletal patterning in the
pharyngeal segments of the zebrafish embryo. Development 124,2945–2960.
Schilling,T.F.,Piotrowski,T.,Grandel,H.,Brand,M.,Heisenberg,C.-P.,
Jiang, Y.-J., Beuchie,D.,Hammerschmidt,M.,Kane,D.A.,Mullins,M.C.,
Van Eeden,J.M.,Keish,R.N.,Furutani-Seiki,M.,Granato,M.,Haffter,
P., Odenthal,J.,Warga,R.M.,Trowe,T.&N¨
usslein-Volhard,C.(
1996).
Jaw and branchial arch mutants in zebrafish I: branchial arches. Development 123,
329– 344.
Schlosser,G.(
2005). Evolutionary origins of vertebrate placodes: insights from
developmental studies and from comparisons with other deuterostomes. Journal of
Experimental Zoology Part B 304,347–399.
Schlosser,G.&Ahrens,K.(
2004). Molecular anatomy of placode development in
Xenopus laevis.Developmental Biology 271,439–466.
Shigetani,Y.,Nobusada,Y.&Kuratani, S. (2000). Ectodermally derived
FGF8 defines the maxillomandibular region in the early chick embryo:
epithelial-mesenchymal interactions in the specification of the craniofacial
ectomesenchyme. Developmental Biology 228,73–85.
Shigetani,Y.,Sugahara,F.,Kawakami,Y.,Murakami,Y.,Hirano, S. &
Kuratani, S. (2002). Heterotopic shift of epithelial-mesenchymal interactions in
vertebrate jaw evolution. Science 296,1316–1319.
Biological Reviews 91 (2016) 611– 657 ©2015 Cambridge Philosophical Society

656 Tetsuto Miyashita
Shigetani,Y.,Sugahara,F.&Kuratani, S. (2005). A new evolutionary scenario
for the vertebrate jaw. BioEssays 27,331–338.
Shone,V.&Graham,A.(
2014). Endodermal/ectodermal interfaces during
pharyngeal segmentation in vertebrates. Journal of Anatomy 225,479–491.
Shu, D.-G., Conway Morris, S., Han,J.,Zhang, Z.-F., Yasui,K.,Janvier,P.,
Chen, L., Zhang, X.-L., Li,Y.&Liu,H.-Q.(
2003). Head and backbone of the
Early Cambrian vertebrate Haikouichthys. Nature 421,526–529.
Smith,M.M.&Coates,M.I.(
2000). Evolutionary origins of teeth and jaws:
developmental models and phylogenetic pattern. In Development, Function and Evolution
of Teeth (eds M. F. Teaford,M.M.Smith and W. J. Ferguson), pp. 133 –151.
Cambridge University Press, Cambridge.
Smith,M.M.&Coates,M.I.(
2001). The evolution of vertebrate dentitions:
phylogenetic pattern and developmental methods. In Major Events in Early Vertebrate
Evolution, Systematics Association Special Volume Series, Volume 61 (ed. P. E. Ahlberg),
pp. 223– 240. Taylor & Francis, London.
Smith,M.M.&Hall,B.K.(
1990). Development and evolutionary origins of
vertebrate skeletogenic and odontogenic tissues. Biological Reviews 65,277–373.
Smith,M.M.&Hall,B.K.(
1993). A developmental model for evolution of
the vertebrate exoskeleton and teeth: the role of cranial and trunk neural crest.
Evolutionary Biology 27,387–447.
Smith,M.M.&Johanson, Z. (2003). Separate evolutionary origins of teeth from
evidence in fossil jawed vertebrates. Science 299,1235– 1236.
Soehn, K. L., & Wilson,M.V.H.(
1990). A complete, articulated
heterostracan from Wenlockian (Silurian) beds of the Delorme Group, Mackenzie
Mountains, Northwest Territories, Canada. Journal of Vertebrate Paleontology 10,
405– 419.
Song,J.&Boord, R. L. (1993). Motor components of the trigeminal nerve and
organization of the mandibular arch muscles in vertebrates: phylogenetically
conservative patterns and their ontogenetic basis. Acta Anatomica 148,139–149.
Soukup,V.,Epperlein,H.H.,Hor ´
acek,I.&Cerny,R.(
2008). Dual epithelial
origin of vertebrate oral teeth. Nature 455,795–798.
Stensi¨
o,E.A.(
1927). The Downtonian and Devonian vertebrates of Spitsbergen.
Part I. Family Cephalaspidae. Skrifter 12,1–391.
Stensi¨
o,E.A.(
1932). The Cephalaspids of Great Britain, British Museum (Natural
History), London.
Stensi¨
o,E.A.(
1939). A new anaspid from the Upper Devonian of Scaumenac Bay
in Canada, with remarks on the other anaspids. Kungliga Svenska Vetenskapsakademiens
Handlingar 18,1–25.
Stensi¨
o,E.A.(
1963). Anatomical studies on the arthrodiran head. Part I. Preface,
geological and geographical distribution, the organization of the head in the
Dolichothoraci, Coccosteomorphi and Pach osteomorphi. Taxonomic appendix.
Kungliga Svenska Ventenskapsakademiens Handlingar 9(2), 1 –419.
Stensi¨
o,E.A.(
1964). Agnathes. In Trait´e de Pal´eontologie (Volume IV (1), ed. J.
Piveteau), pp. 93– 382. Masson, Paris.
Stensi¨
o,E.A.(
1969). Elasmobranchiomorphi Placodermata Arthrodires. In Trait´ede
Pal´eontologie (Volume IV (2), ed. J. Piveteau), pp. 71– 692. Masson, Paris.
Steventon,B.,Mayor,R.&Streit,A.(
2012). Mutual repression between Gbx2
and Otx2 in sensory placodes reveals a general mechanism for ectodermal patterning.
Developmental Biology 367,55–65.
Steventon,B.,Mayor,R.&Streit,A.(
2014). Neural crest and placode interaction
during the development of the cranial sensory system. Developmental Biology 389,
28– 38.
Strahan,R.(
1958). The velum and the respiratory current of Myxine.Acta Zoologica
39,227–240.
Summers,A.P.&Ferry-Graham, L. A. (2002). Respiration in elasmobranchs: new
models of aquatic ventilation. In Vertebrate Biomechanics and Evolution (eds V. L. Bels,
J.P. Gasc and A. Casinos), pp. 87– 100. Bios Scientific Publishers Ltd.: Oxford.
Szabo-Rogers, H. L., Geetha-Loganathan,P.,Whiting,C.J.,Nimmagadda,
S., Fu,K.&Richman,J.M.(
2009). Novel skeletogenic patterning roles for the
olfactory pit. Development 136,219–229.
Takechi,M.,Adachi,N.,Hirai,T.,Kuratani, S. & Kuraku, S. (2013). The Dlx
genes as clues to vertebrate genomics and craniofacial evolution. Seminars in Cell &
Developmental Biology 24,110–118.
Takio,Y.,Kuraku, S., Murakami,Y.,Pasqualetti,M.,Rijli,F.M.,Narita,
Y., Kuratani, S. & Kusakabe,R.(
2007). Hox gene expression patterns in
Lethenteron japonicum embryos – insights into the evolution of the vertebrate Hox
code. Developmental Biology 308,606–620.
Takio,Y.,Pasqualetti,M.,Kuraku, S., Hirano, S., Rijli,F.M.&Kuratani,
S. (2004). Lamprey Hox genes and the evolution of jaws. Nature 429 (electronic
supplement) doi: 10.1038/nature02616).
Talbot,J.C.,Johnson, S. L. & Kimmel,C.B.(
2010). hand2 and Dlx genes specify
dorsal, intermediate, and ventral domains within zebrafish pharyngeal arches.
Development 137,2507–2517.
Talbot,J.C.,Walber,M.B.,Carney,T.J.,Huycke,T.R.,Yan,Y.-L.,Bremiler,
R. A., Gai, L., Delaurier,A.,Postlethwait,J.H.,Hammerschmidt,M.
&Kimmel,C.B.(
2012). fras1 shapes endodermal pouch I and stabilizes zebrafish
pharyngeal skeletal development. Development 139,2804–2813.
Theveneau,E.,Steventon,B.,Scarpa,E.,Garcia, S., Trepat,X.,Streit,A.
&Mayor,R.(
2013). Chase-and-run between adjacent cell populations promotes
directional collective migration. Nature Cell Biology 15,763–772.
Thisse,C.&Thisse,B.(
2005). High throughput expression analysis of ZF-models
consortium clones. ZFIN Direct Data Submission ZDB-PUB-051025-1. Electronic
file available at http://zfin.org. Accessed 30.12.2013.
Thomson, K. S. (1993). Segmentation, the adult skull, and the problem of homology.
In The Skull: Patterns of Structural and Systematic Diversity (Volume II,edsJ.Hanken
and B. K. Hall), pp. 36– 68. The University of Chicago Press: Chicago.
Trainor, P. A. (ed.) (2014). Neural Crest Cells: Evolution, Development and Disease,Academic
Press, London.
Trainor,P.A.&Krumlauf,R.(
2001). Hox genes, neural crest cells and branchial
arch patterning. Current Opinion in Cell Biology 13,698–705.
Trinajstic,K.&Hazelton,M.(
2007). Ontogeny, phenotypic variation and
phylogenetic implications of arthrodires from the Gogo Formation, Western
Australia. Journal of Vertebrate Paleontology 27,571–583.
Trinajstic,K.,Long,J.A.,Johanson, Z., Young,G.C.&Senden,T.(
2012). New
morphological information on the ptyctodontid fishes (Placodermi, Ptyctodontida)
from Western Australia. Journal of Vertebrate Paleontology 32,757–780.
Trinajstic,K.,Sanchez, S., Dupret,V.,Tafforeau,P.,Long,J.,Young,G.,
Senden,T.,Boisvert,C.,Power,N.&Ahlberg,P.E.(
2013). Fossil musculature
of the most primitive jawed vertebrates. Science 341,160–164.
Tsuneki,K.&Koshida,Y.(
1993). Structural organization of the blood-sinus systems
in lampreys and hagfish: functional and evolutionary interpretations. Acta Zoologica
74,227–238.
Tulenko,F.J.,McCauley,D.W.,Mackenzie, E. L., Mazan, S., Kuratani, S.,
Sugahara,F.,Kusakabe,R.&Burke,A.C.(
2013). Body wall development in
lamprey and a new perspective on the origin of vertebrate paired fins. Proceedings of
the National Academy of Sciences of the United States of America 110,11899–11904.
Turner, S. (1982). A new articulated thelodont (Agnatha) from the Early Devonian
of Britain. Palaeontology 25,879–889.
Turner, S., Burrow,C.J.,Schultze,H.-P.,Blieck,A.,Reif, W.-E., Rexroad,
C. B., Bultynck,P.&Nowlan, G. S. (2010). False teeth: conodont-vertebrate
phylogenetic relationships revisited. Geodiversitas 32,545–594.
Tzahor,E.,Kempf,H.,Mootoosamy,R.C.,Poon,A.C.,Abzhanov,A.,
Tabin,C.J.,Dietrich, S. & Lessar,A.B.(
2003). Antagonists of Wnt and BMP
signaling promote the formation of vertebrate head muscle. Genes and Development 17,
3087– 3099.
Uchida,K.,Murakami,Y.,Kuraku, S., Hirano, S. & Kuratani, S. (2003).
Development of the adenohypophysis in the lamprey: evolution of epigenetic
patterning programs in organogenesis. Journal of Experimental Zoology B 300,32–47.
Wada,N.,Nohno,T.&Kuratani, S. (2011). Dual origins of the prechordal cranium
in the chicken embryo. Developmental Biology 356,529–540.
Wang, N.-Z., Donoghue,P.C.J.,Smith,M.M.&Sansom,I.J.(
2005). Histology
of the galeaspid dermoskeleton and endoskeleton, and the origin and early evolution
of the vertebrate cranial endoskeleton. Journal of Vertebrate Paleontology 25,745–756.
W¨
angsj¨
o,G.(
1952). The Downtonian and Devonian vertebrates of Spitsbergen. IX.
Morphologic and systematic studies of the Spitsbergen cephalaspids. Skrifter 97,
1–611.
Wexler,D.B.&Davidson,T.M.(
2004). The nasal valve: a review of the anatomy,
imaging, and physiology. American Journal of Rhinology 18,143–150.
Wilson,M.V.H.&Caldwell,M.W.(
1993). New Silurian and Devonian
fork-tailed ‘thelodonts’ are jawless vertebrates with stomachs and deep bodies. Nature
361,442–444.
Wilson,M.V.H.,Hanke,G.F.&M¨
arss,T.(
2007). Paired fins of jawless
vertebrates and their homologies across the ‘‘agnathan’’-gnathostome transition. In
Major Transitions in Vertebrate Evolution (eds J. S. Anderson and H.-D. Sues), pp.
122– 149. Indiana University Press, Bloomington.
Wright,R.R.(
1885). On the hyomandibular clefts and pseudobranchs of Lepidosteus
and Amia.Journal of Anatomy and Physiology 19,476–499.
Xu,H.,Dude,C.M.&Baker,C.V.H.(
2008). Fine-grained fate maps for
the ophthalmic and maxillomandibular trigeminal placodes in the chick embryo.
Developmental Biology 317,174–186.
Yalden,D.W.(
1985). Feeding mechanisms as evidence for cyclostome monophyly.
Zoological Journal of the Linnean Society 84,291–300.
Yao,T.,Ohtani,K.,Kuratani, S. & Wada,H.(
2011). Development of lamprey
mucocartilage and its dorsal-ventral patterning by endothelin signaling, with insight
into vertebrate jaw evolution. Journal of Experimental Zoology B 316,339–346.
Young,G.C.(
1979). New information on the structure and relationships of Buchanosteus
(Placodermi: Euarthrodira) from the Early Devonian of New South Wales. Zoological
Journal of the Linnean Society 66,309–352.
Young,G.C.(
1980). A new Early Devonian placoderm from New South Wales,
Australia, with a discussion of placoderm phylogeny. Palaeontographica Abteilung A 167,
10– 76.
Young,G.C.(
1984). Reconstruction of the jaws and braincase in the Devonian
placoderm fish Bothriolepis.Palaeontology 27,625–661.
Young,G.C.(
1986). The relationships of placoderm fishes. Zoological Journal of the
Linnean Society 88,1–57.
Biological Reviews 91 (2016) 611– 657 ©2015 Cambridge Philosophical Society

Fishing for jaws 657
Young,G.C.(1991). The first armoured agnathan vertebrates from the Devonian
of Australia. In Early Vertebrates and Related Problems of Evolutionary Biology (eds
M. M. Chan,Y.H.Liu and G. R. Zhang), pp. 67– 85. Science Press,
Beijing.
Young,G.C.(
2010). Placoderms (armored fish): dominant vertebrates of the Devonian
Period. Annual Review of Earth and Planetary Sciences 38,523–550.
Young,G.C.,Leli `
evre,H.&Goujet,D.(
2001). Primitive jaw structure in
an articulated brachythoracid arthrodire (placoderm fish; Early Devonian) from
southeastern Australia. Journal of Vertebrate Paleontology 21,670–678.
Young,G.C.&Zhang,G.-R.(
1992). Structure and function of the pectoral
joint and operculum in antiarches, Devonian placoderm fishes. Palaeontology 35,
443– 464.
Zhang, Z., Wlodarcsyk,B.J.,Niederreither,K.,Venugopalan, S., Florez,
S., Finnell,R.H.&Amendt,B.A.(
2011). Fuz regulates craniofacial development
through tissue specific responses to signaling factors. PLoS ONE 6,e24608.
Zhu,M.&Janvier,P.(
1996). The histological structures of the endoskeleton in
galeaspids (Galeaspida, Vertebrata). Journal of Vertebrate Paleontology 18,650–654.
Zhu,M.,Wu,X.,Ahlberg,P.E.,Choo,B.,Lu,J.,Qiao,T.,Zhao,W.,Jia, L.,
Blom,H.&Zhu,Y.(
2013). A Silurian placoderm with osteichthyan-like marginal
jaw bones. Nature 302,188–193.
Ziermann,J.M.,Miyashita,T.&Diogo,R.(
2014). Cephalic muscles of
Cyclostomes (hagfishes and lampreys) and Chondrichthyes (sharks, rays and
holocephalans): comparative anatomy and early evolution of the vertebrate head
muscles. Zoological Journal of the Linnean Society 172,771–802.
(Received 19 February 2014; revised 18 March 2015; accepted 19 March 2015; published online 21 April 2015)
Biological Reviews 91 (2016) 611– 657 ©2015 Cambridge Philosophical Society