American J Phys Anthropol - 2001 - Jernvall - Genotype Phenotype and Developmental Biology of Molar Tooth Characters

YEARBOOK OF PHYSICAL ANTHROPOLOGY 43:171–190 (2000)
Genotype, Phenotype, and Developmental Biology of Molar

Tooth Characters
JUKKA JERNVALL1 AND HAN-SUNG JUNG2,3
1
Viikki Biocenter, Institute of Biotechnology, University of Helsinki,
Helsinki FIN-00014, Finland
2
Division of Histology, Department of Oral Biology, College of Dentistry,
Yonsei University, Seoul 120-752, Korea
3
Cutaneous Biology Research Center, Massachusetts General Hospital-
East, Harvard Medical School, Charlestown, Massachusetts 02129
KEY WORDS dentition; teeth; morphogenesis; evolvability; dental

traits
ABSTRACT Primate molar shapes reflect developmental and ecological

processes. Development may constrain as well as facilitate evolution of new
tooth shapes, affecting how reliable dental characters are in phylogenetic
studies. Much of the genetic machinery of development uses the same genes
among different organs, including teeth, limbs, and feathers. Furthermore,
within a tooth, the development of individual cusps repeatedly uses the same
set of developmental genes, forming a “developmental module.” The repeated
activation of the developmental module can explain the cumulative variation
in later-developing cusps. Therefore short, later-developing cusps may be
evolvable but also more homoplastic. This patterning cascade mode of cusp
development can be used to explain the variational properties of dental
characters and character states related to cusp initiation. The developmental
basis and variational properties of crown termination, cusp shape, and cusp
configuration characters are currently less well understood. It is unlikely
that there is a simple “gene to phenotype” map for dental characters. Rather,
the whole cusp pattern is a product of a dynamic developmental program
manifested in the activation of the developmental modules. Yrbk Phys An-
thropol 43:171–190, 2000. © 2000 Wiley-Liss, Inc.
TABLE OF CONTENTS
Introduction ......................................................................................................................... 172

Return of Developmental Biology Into the Modern Synthesis ........................................ 172
Tooth Shape Development and Evolution ........................................................................ 174
Primate teeth ................................................................................................................... 174
Development of molar cusps ........................................................................................... 175
Formation of a signaling center before the cusps ..................................................... 176
Embryonic role of the enamel knot ............................................................................ 177
Cusps are enamel knot replays .................................................................................. 178
Cusps as feathers ........................................................................................................ 179
Evolvability of tooth cusps ............................................................................................. 181
Is there a homeobox code for teeth? .............................................................................. 184
Defining Dental Characters ............................................................................................... 185
Grouping characters based on development ................................................................. 185
Initiation characters .................................................................................................... 186
Termination characters ............................................................................................... 186
Cusp shape characters ................................................................................................ 187
© 2000 WILEY-LISS, INC.

10968644, 2000, S31, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/1096-8644(2000)43:31+<171::AID-AJPA6>3.0.CO;2-3 by Cochrane Mexico, Wiley Online Library on [30/05/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
172 YEARBOOK OF PHYSICAL ANTHROPOLOGY [Vol. 43, 2000
Configuration characters ............................................................................................. 187

Conclusions: Should We Worry About Development? ...................................................... 187
Acknowledgments ............................................................................................................... 188
Literature Cited .................................................................................................................. 188
INTRODUCTION tire human genome may transform the field

of developmental biology even further. Cur-
Were it not for teeth, anthropology would
rently, most attempts to understand devel-
be a different subject. Teeth, living and
opment stem from research on individual
dead, have much to contribute in the study
gene pathways, but this approach is much
of ecology, paleontology, functional mor-
like trying to describe the ecology of a whole
phology, and systematics. Teeth seem to be
forest from studying a single primate spe-
a good, and durable, source of characters for
cies. The advent of genome-wide data will
phylogenetic analysis. These characters in-
allow population ecology-like approaches on
clude the surface features of tooth crowns,
developmental biology, as large groups of
most notably cusps and crests, structural genes can be analyzed using new techniques
details of enamel, and measures of size and like microarrays.
shape. Moreover, these dental characters of- Acknowledging the role of development in
ten appear discrete and heritable, and they morphological evolution can quickly become
have different degrees of expression, i.e., a practical problem if we want to estimate
different character states. The two most how much and what kinds of effects devel-
prevalent uses of teeth are studies of phy- opment has on frequently used dental char-
logeny and ecological adaptation. Tooth acters. Do dental characters really possess
characters have a role in phylogenetic re- the properties of being discrete and herita-
constructions of primates in general and ble, having biologically meaningful charac-
hominid interrelations in particular. More- ter states, and can they really change char-
over, detailed dental characters have a wide acter states independently from each other?
use in comparisons of human populations. All these properties are considered impor-
Different aspects of teeth can also be asso- tant for a reliable character in reconstruct-
ciated with diet, allowing the estimation of ing phylogenies (Lieberman, 1999). More-
dietary specializations in extinct species. over, once we have a clear idea about tooth
Thus, both dental micro- and macrostruc- development, we can begin to ask questions,
ture are correlated experimentally or ana- e.g., what kinds of developmental criteria
lytically to different foods. However, like can be used to define reliable characters? In
any biological structure, dental characters other words, this review explores whether
reflect not only functional demands and di- one should worry about development when
etary adaptations in the context of phylog- considering dietary causality for evolution
eny, but also developmental processes con- of new tooth cusps or when choosing dental
trolling morphogenesis. This latter is more characters for phylogenetic studies.
apparent in teeth than in other skeletal el-
ements because after eruption crown shape RETURN OF DEVELOPMENTAL
changes only by wear, and not through re- BIOLOGY INTO THE MODERN
modeling as in bones. SYNTHESIS
Since the previous Yearbook articles ad- Several researchers have pointed out the
dressing tooth development (Weiss, 1990; absence of developmental biology from the
Keene, 1991), there has been a renaissance modern synthesis of evolutionary biology
of developmental biology, mostly propelled (e.g., Gilbert et al., 1996; Hamburger, 1980).
by new discoveries from developmental ge- Gilbert et al. (1996) state that classical em-
netics. Indeed, enough new investigators bryology, so central to Darwin’s arguments,
have begun to examine development in the was later seen as old-fashioned, as genetics
context of evolution that a new field of evo- was found to be a powerful paradigm for
lutionary developmental biology (evo-devo) explaining the evolution of characters. In-
is in the making. The sequencing of the en- deed, comparative tooth development was
J. Jernvall and H.-S. Jung] DEVELOPMENTAL BIOLOGY OF MOLAR CHARACTERS 173
we can indeed infer phenotype from geno-

type, and genetic determinism is warranted
(Fig. 1). This would mean that natural se-
lection could act directly on genes. However,
explaining macroevolution as merely the ex-
tension of (gradual) microevolution, and ul-
timately as the extension of changing geno-
type, is problematic (Bateson, 1894, 1913;
Gilbert, 1994; Gilbert et al., 1996; Mayr,
1959). Large morphological changes, when
developmental processes underlying their
variation properties were ignored, seemed
to be beyond the explanatory power of grad-
ual changes in the genotype. Eventually,
Gould (1977) triggered the incorporation of
development, in terms of heterochrony and
allometry, in the study of morphological
Fig. 1. Changing practices in evolutionary studies.
The modern synthesis (top) has largely concerned link-
evolution.
ing genetic level (as population genetics) to microevolu- Interestingly, the reductionist approach
tion. Macroevolution is an extension of microevolution; of genetic determinism continues to mani-
development is not included. Evo-devo studies (middle)
link genetic level (as developmental genetics) to devel- fest itself in much research on development,
opment. Changes in development explain macroevolu- but in a slightly transformed way. For a
tion; microevolution is not necessarily explained. Com- long time, the study of development was the
bined approaches (bottom) use genetic level (as
developmental and population genetics) to link develop- study of cell differentiation, which was di-
ment to microevolution and macroevolution. rected by differential gene expression. This
focus on cell differentiation was contrasted
in the 1980s by the discovery that homeobox
intensively studied in diverse groups of genes were found to direct the differentia-
mammalian species until the 1930s (e.g., tion of whole body parts. Moreover, as ho-
Ahrens, 1913; Bolk, 1920, 1921, 1922; Hei- meobox genes were found to be shared
nick, 1908; Leche, 1915; Ohlin, 1896; von among practically every type of animal, de-
Korff, 1932). After that, research on teeth velopmental biology underwent a paradigm
began to focus on quantifying traits relevant shift. Gilbert et al. (1996) pointed out that
to the genetics of microevolution. One such the discovery of shared homeobox genes
research focus was the analysis of correla- marked a rediscovery of the concept of ho-
tion fields in which tooth measurements mology among developmental biologists
were correlated with each other along the (i.e., the concept that many features are
tooth row and among the jaws (e.g., Kurtén, shared among organisms due to common
1953). As population genetics became the ancestry, and that this could be used as a
central model used to explore evolution, tool in research). As a great deal of the re-
development and even phenotypes were search on tooth development had focused on
mostly considered the ephemeral extension enamel and dentine differentiation, now
of the genotype (discussed in Dawkins, “developmental regulatory” genes entered
1982; Gilbert et al., 1996; Mayr, 1997). Yet, the spotlight. These were genes that were
this paradigm relies on the assumption that found to be expressed in other organs such
there is an almost one-to-one relationship as limbs and, moreover, they were usually
between genotype and phenotype. Wagner genes that had homologues across all stud-
and Altenberg (1996) called this a “repre- ied animals. And in the 1990s, as in popu-
sentation problem,” or “gene to phenotype lation genetics in the 1930s, molecular ge-
mapping” (also see Minelli, 1998), and netics could be used to explain evolution.
Weiss and Fullerton (2000) referred to a The most simplistic view of evolution im-
“phenogenetic” relationship. Essentially, if plied that changes in the homologous genes
the phenogenetic relationship is very strong, and their regulation resulted in the evolu-
tion of new morphologies. A central point to 2000), defined as how much genotype can
remember is that the “model” organisms evolve without a corresponding change in
that were behind the discovery of homeobox phenotype. Phenogenetic drift is a crucial
genes included species as disparate as the issue because if phenogenetic drift is com-
mouse and the fruit fly. Therefore, new evo- mon, simple genetic reductionism is unwar-
lutionary developmental studies were di- ranted and some higher-level developmen-
rectly involved with macroevolution, be- tal principles must be discovered in order to
cause the conserved developmental genes understand the developmental basis of mor-
were homologous across classes and phyla. phological evolution. In this case, epigeno-
These evo-devo studies, addressing the ori- type (the whole developmental complex),
gin of new organs and tissue-types across rather than genotype, would form the basis
higher taxa, are quite different from the for the developmental control of phenotype.
studies of the “old” modern synthesis involv- Phenogenetic drift will also affect how we
ing genetics and microevolution. In the most approach homologies. Indeed, it is likely
simplified terms, changes in the conserved that the discovery of homologous develop-
developmental regulatory genes cause mental genes, interacting in operational
changes in development, in turn causing groups or modules (see Raff, 1996), already
changes in morphology (Fig. 1). Whereas demands the reappraisal of character ho-
the modern synthesis lacked developmental mologies. For example, as recognition of
biology, the evo-devo studies are often lack- characters depends on their being inherited
ing microevolution. In practical terms, units (see Lieberman, 1999), the problem
rather than extrapolating microevolution- becomes, what exactly is the developmental
ary changes to macroevolution, expression information underlying the characters if
domains of homeobox genes account for their development is controlled by modules
large changes in morphology directly (e.g., that are themselves shared among different
Cohn and Tickle, 1999; Burke et al., 1995). organs? A reasonable solution is perhaps to
Without doubt, recent advances in devel- rely on the homology concept of Van Valen
opmental biology have enabled us to add (1982, 1994), in which continuity of develop-
new information on the early evolution of mental information, without strict assump-
higher taxa previously beyond the resolu- tions on the physical nature of that informa-
tion, or reach, of the fossil record. Fields like tion, causes homology.
biological anthropology are in the fortunate
position of being able to integrate develop- TOOTH SHAPE DEVELOPMENT
ment into evolutionary studies. Detailed or- AND EVOLUTION
ganismal-level knowledge of tooth evolution
Primate teeth
in ecological contexts can be used to link
new information on development to micro- In this paper we concentrate on molar
and macroevolution (Fig. 1). This may pro- tooth shapes in order to illustrate a more
duce better appraisals of the developmental focused picture of the links between devel-
basis for characters used in phylogenetic opment and evolution. Another reason to
studies (Holland, 1999). The risk involved in limit the discussion to molar crowns is the
applying molecular-level models of charac- fact that the standard mammal used in de-
ters directly to macroevolution is that it is velopmental studies, the mouse, has only
difficult to test whether character codings incisor and molar teeth, and thus practi-
are meaningful when considering natural cally nothing is known about the molecular
selection and microevolution. In other aspects of development in adjacent tooth
words, it is uncertain how variational prop- families. The evolution of dentition as a
erties of developmental systems can be in- whole, and its resemblance to the evolution
ferred by extrapolating from molecules to of other segmented structures such as the
fossils. Furthermore, examining develop- vertebral column, are discussed extensively
ment in the context of micro- and macroevo- in Weiss (1990).
lution can be used to estimate the extent of First, we will examine the diversity of
phenogenetic drift (Weiss and Fullerton, primate molar teeth themselves. Unlike
common taxonomic diversity measures,

morphological diversity measures that cut
across phylogenetic boundaries are useful in
delineating the minimum range of morpho-
logical diversity that is possible to con-
struct. By grouping primate upper molar
shapes into discrete crown types, we can
first examine just diversity, or the number
of morphological types. Crown type scheme
is a simple classification system of tooth
crown topography, where no phylogenetic
relations (i.e., homologies) are implied
(Jernvall et al., 1996, 2000). Five variables
(characters) are used to define each crown Fig. 2. Complexity of primate upper molars. Fre-
type: cusp shape, number of buccal cusps, quency distribution of crown types with different num-
bers of cusps and lophs on second upper molars. Note
number of lingual cusps, number of longitu- that complexity (as the summed-up number of cusps
dinal lophs, and number of transverse lophs and lophs) is lower in primates than in artiodactyls or
perissodactyls, but similar to archaic ungulates, the
(Jernvall et al., 1996, 2000). Lophs are well- condylarths. The average disparity (morphological dif-
developed shearing crests, and both lophs ference among crown types) is 2.7, 2.7, 3.3, and 3.6 in
and cusps have to be two thirds of the crown primates, condylarths, artiodactyls, and perissodactyls,
respectively.
height to be tabulated, making the crown
type scheme conservative but robust.
Extant primates have seven crown types, phylogenetic studies (Wagner, 2000), as the
which is more diverse than in the ungulate chance for convergent evolution of similar
orders Artiodactyla (three) and Perissodac- dental morphologies can be hypothesized to
tyla (four). Therefore, primates seem to be higher among primates than, for exam-
have high molar shape diversity compared ple, among ungulates. This stresses the
to ungulates, whose dentition is tradition- need to understand the developmental basis
ally considered diverse, reflecting different of dental characters.
dietary specializations to herbivory (Janis
Development of molar cusps
and Fortelius, 1988). However, if we include
the fossil record, the number of primate Of the main characters of tooth shape, the
crown types (order Primates only) rises to development of cusps is perhaps currently
nine, whereas Artiodactyla and Perissodac- the best understood. Mammalian tooth
tyla have 13 and 17 crown types, respec- shape is a product of the folding and growth
tively (Jernvall et al., 2000). Moreover, of the interface between epithelium and
when estimating dental complexity in terms neural crest-derived mesenchyme. As fold-
of the number of cusps and lophs, primate ing proceeds, the differentiating inner
molars are simpler than ungulate molars enamel epithelium facing the mesenchyme
(Fig. 2). Indeed, the crown type complexity gives rise to enamel-forming ameloblasts,
of primate upper molars is closer to that of while the mesenchyme below gives rise to
Condylarthra, archaic ungulates that have the dentine-forming odontoblasts (Butler,
been implicated in the ancestry of perisso- 1956; Jernvall and Thesleff, 2000). The de-
dactyls and artiodactyls in the Eocene epoch velopmental stage when individual teeth
(Fig. 2). Macroevolutionary diversity of pri- become visible begins by the formation of a
mate molar shapes can be thus described as tooth bud that can first be seen as a slight
having a relatively high diversity of basic outgrowth of the dental lamina (Fig. 3A,B).
morphotypes with only a moderate range of These earliest stages of tooth development
anatomical design, in that the number of resemble morphologically other epithelial
cusps and lophs is generally low and in that appendages, such as hairs and glands (Jern-
the topographies of crown types are quite vall and Thesleff, 2000). Also, in all these
similar (Fig. 2). This limited range of ana- organs, development is regulated by inter-
tomical design may be in itself a concern in actions between the epithelial and underly-
Fig. 3. Schematic representations of tooth development. A: Beginning in the cap stage, the tooth
crown forms by unequal growth of the inner enamel epithelium (shaded). B: Molecular signals considered
to be important for particular developmental stages are indicated above the morphology. Note how the
same signaling pathways are used repeatedly during advancing tooth development. BMP, bone morpho-
genetic protein; FGF, fibroblast growth factor; SHH, sonic hedgehog; WNT, wingless-integrated.
ing mesenchymal tissues. The development Nieminen et al., 1998) manifests the pri-
of molar shape begins by the folding of the mary enamel knot as an embryonic signal-
bud tip that results in a cap-resembling ing center. Signaling centers interact with
structure surrounding the mesenchymal surrounding tissues by secreting molecular
dental papilla (Fig. 3A). Unlike the iterative signals that are required for the progression
branching morphogenesis of lungs and of morphogenesis and they are understood,
glands, the tooth bud only forms one set of in developmental biology terms, to “regu-
lateral branches along the anterio-posterior late” or “control” different aspects of mor-
axis of the bud. phogenesis. These views stem from classic
Formation of a signaling center before grafting experiments showing that certain
the cusps. The formation of the cap stage regions of the embryo (i.e., signaling cen-
marks the future tooth crown base. This ters) are able to reorganize embryogenesis
developmental stage begins with the forma- (Tickle et al., 1975; Spemann and Mangold,
tion of the primary enamel knot at the tip of 1924). Other comparable embryonic signal-
the tooth bud. The primary enamel knot is ing centers include the zone of polarizing
an epithelial structure that is histologically activity (ZPA) and the apical ectodermal
visible as a cluster of densely packed cells. ridge (AER), which together participate in
These cells are nondividing and express sev- the patterning and growth of limbs
eral signaling molecules (Vaahtokari et al., (Niswander and Martin, 1993; Tickle et al.,
1996a,1996b). The coexpression of over 10 1975). The notochord is another signaling
signals (http://honeybee.helsinki.fi/toothexp; center in the patterning of the neural tube
and somites (Echelard et al., 1993, Fan and only expressed in the enamel knot cells, and
Tessier-Lavigne, 1994, Liem et al., 1995), Fgf3 is expressed in the enamel knot and
and the tips of the lung buds have been the underlying mesenchyme. Fgf10, the
proposed to function as signaling centers in fourth Fgf transcribed in teeth, is expressed
lung branching (Hogan, 1999; Metzger and during the cap stage only in the mesen-
Krasnow, 1999). While the developmental chyme (Kettunen and Thesleff, 1998). More-
context and outcome differ among signaling over, the different FGF proteins (we follow
centers, the genes used in molecular signal- the practice of marking genes in italicized
ing are usually the same. Indeed, discover- letters and protein products in capital let-
ies where signaling centers could be recip- ters) do not stimulate all dental tissues
rocally replaced among organs and equally. FGF4 and FGF9 have been shown
organisms had already pointed toward a to stimulate cell proliferation of both dental
common basis for embryonic signaling epithelium and mesenchyme, whereas
(Hornbruch and Wolpert, 1986). For exam- FGF10 stimulates cell division only in den-
ple, the Sonic hedgehog (Shh) gene is ex- tal epithelium (Kettunen et al., 1998, 2000).
pressed in the developing limb in the ZPA, Thus it appears that the primary enamel
in the lung bud, notochord, and enamel knot knot promotes tooth germ growth around
(Echelard et al., 1993; Riddle et al., 1993; itself, which may play a role in the folding of
Roelink et al., 1994). The anterior-posterior dental epithelia during the formation of the
limb patterning requiring ZPA can be mim- tooth cap and crown base (Jernvall et al.,
icked by implanting a bead releasing SHH- 1994).
protein (Riddle et al., 1993) or a tooth germ While the embryological role of the
(Koyama et al., 1996) into limbs. These dis- enamel knot may be related to epithelial
coveries make it obvious that one gene does folding, the molecular-level process of the
not produce one structure. enamel knot expressing growth-stimulating
Fgfs while not growing itself has been a
Embryonic role of the enamel knot. puzzle. Partly, the nonproliferation of
The formation of the tooth cap is basically enamel knot cells can be linked to the
all the epithelial folding that is needed in a dearth of FGF receptors in the enamel knot
unicusped tooth, e.g., in a canine. The tip of area (Kettunen et al., 1998). Thus, the
the tooth cap becomes the tip of the cusp, enamel knot appears to be “deaf” to the
and the sides of the cap continue to grow stimulatory signals of FGFs. Additionally,
down, forming the cervical loop and eventu- an early marker for the enamel knot is cy-
ally the root (Fig. 3B). The formation of clin-dependent kinase inhibitor p21 expres-
tooth cap and enamel knot is accompanied sion in the enamel knot. p21 expression is
by unequal growth in different regions of associated with cells stopping to divide and
the inner enamel epithelium. Only epithe- beginning terminal differentiation. There-
lium has been found to have a distinct pat- fore, as in differentiating muscle cells
tern in cell proliferation during early crown (Parker et al., 1995), p21 may be involved in
morphogenesis, suggesting that the forma- the differentiation and withdrawal of
tion of the crown base is essentially a fold- enamel knot cells from the cell cycle (Bloch-
ing of a two-dimensional sheet. The role of Zupan et al., 1998; Jernvall et al., 1998). In
the cells of the enamel knot as the first area the case of the enamel knot, the differenti-
of mitotically quiescent cells has been in- ated stage involves the expression of
tensely studied in the folding and unequal growth-stimulatory genes. p21 is also ex-
growth of the tooth cap. Indeed, it is curious pressed in the limb AER that also expresses
that the nondividing cells of the primary FGFs while remaining nonproliferative
enamel knot express fibroblast growth fac- (Parker et al., 1995). It is worthwhile to
tor (Fgf) genes, which are known mitogens. point out that experimental knockout mice
At least four Fgfs are expressed in a de- lacking functional p21 had no reported den-
veloping tooth, of which all are found to tal (or limb) defects (Deng et al., 1995). This
promote cell division in cultured dental tis- illustrates the possibility that p21 may have
sues. Two of the Fgfs, Fgf4 and Fgf9, are functional redundancy with other inhibitors
of cell proliferation as a part of a develop-

mental “insurance policy” against loss of
normal cell cycle control. The biology of re-
dundancy, however, is not fully understood.
The primary enamel knot begins to disap-
pear soon after the bud has branched out to
form the cap (Fig. 3). The disappearance of
signaling centers is in itself a subject of
research. It is now known that the cells of
the enamel knot disappear via apoptosis, or
programmed cell death. This apoptosis is
associated with the expression of another
growth factor, bone morphogenetic protein 4
(Bmp4), in the enamel knot (Jernvall et al.,
Fig. 4. Occlusal view (three-dimensional reconstruc-
1998). While Bmps have multiple roles in tions of inner enamel epithelium viewed from the above)
development, they appear also to be gener- of developing mouse and vole lower molar. Note how
Fgf4 expression marks the adult cusp patterns.
ally involved in apoptosis in other organs,
such as rhombomeres and digits (Graham et
al., 1994; Pizette and Niswander, 1999). Bi-
Unfortunately, experimental inactiva-
ologically, apoptosis has been suggested to
tions of genes affecting tooth development
be a mechanism involved in the control of
result in the total lack of some of or all of the
duration of molecular signaling of the
teeth (Chen et al., 1996; Ferguson et al.,
enamel knot (Vaahtokari et al., 1996b; Jern- 1998; van Genderen et al., 1994; Hardcastle
vall et al., 1998), and this has been sug- et al., 1998; Kratochwil et al., 1996; Peters
gested to be the case in the apoptotic re- et al., 1998; Thomas et al., 1997), an obvious
moval of the AER in the limb bud as well limitation if one wants to study changes in
(Pizette and Niswander, 1999). cusp patterns. Thus knockout mice are not
Cusps are enamel knot replays. In necessarily informative for considering cusp
development per se. However, a natural
teeth that have many cusps, i.e., usually
mouse mutant, Tabby, has defects in molar
molar teeth, new enamel knots appear soon
cusps that are informative about develop-
after the primary enamel knot disappears.
ment. The Tabby gene has a corresponding
The new enamel knots, called secondary
human gene, the anhidrotic ectodermal dys-
enamel knots (Jernvall et al., 1994), appear
plasia (EDA) gene (Ferguson et al., 1997;
at the places of future cusp tips. As for the Srivastava et al., 1997), and both EDA pa-
primary enamel knot, secondary enamel tients and Tabby mice have abnormal devel-
knots are also nonproliferative, express opment of epithelial appendages such as
Fgf4, and are removed apoptotically (Vaah- teeth, hair, and sweat glands. In Tabby
tokari et al., 1996b; Coin et al., 1999). mice, molar cusps are compressed, as the
The secondary enamel knots are the first tips of buccal and lingual cusps are close to
signs of cusps (Fig. 4), and using Fgf4 as an each other or completely united. In addition,
enamel knot marker, the onset of species- the last developing cusps are often missing.
specific cusp patterns was recently analyzed This altered cusp pattern is already visible
in developing mouse and vole lower molars at the level of secondary enamel knots,
(Keränen et al., 1998). Cusp patterning was which are often fused (Pispa et al., 1999).
estimated to begin as early as a day after The altered cusp pattern in Tabby molars
the formation of the tooth cap, thus before appears to stem from an affected primary
cusps are visible morphologically. This re- enamel knot that is small, while still ex-
sult confirms the notion of Butler (1982) pressing all the known enamel knot genes.
that tooth germ presumptive domains exist Therefore, a general retardation of growth
in the early tooth germ from which cusps may cause the delayed formation of cusps
develop. and the lack of the later developing molars
in Tabby mice. Indeed, cusp development in so that the smallest cusps are the last to
Tabby molars is partially corrected when form (Butler, 1956). Thus, for example, the
Tabby tooth germs are cultured in vitro phylogenetically recent but large metacone
with FGF4 and -10 present in culture me- forms prior to the phylogenetically older
dium (Pispa et al., 1999). This effect sug- paracone in opossums (Berkowitz, 1967).
gests that Fgfs may indeed have a role in This kind of development, where there is no
cusp development in vivo. single gene for a single cusp, requires us to
Considering the development of individ- explain the evolution of cusps using dy-
ual cusps, no cusp seems to differ from an- namic patterning mechanisms (Jernvall,
other in terms of secondary enamel knots. 2000; Weiss et al., 1998; Zhao et al., 2000a).
Every cusp studied expresses genes of the
families Fgf, Shh, Wnt, and Bmp. Indeed, Cusps as feathers. In teeth, an accurate
there are currently no genes that would, control of secondary enamel knot spacing is
individually or in a combination, mark the likely to exist during morphogenesis, as this
identity of, for example, the metaconid. process results in correct cusp position and
However, while several signals appear to be size. Both these characters (position and
active in the initiation of cusp development, size) are critical for a functioning occlusion
only Fgf4 expression remains strictly re- and feeding. The iterative activation of the
stricted to the cusp tips. For example, p21 secondary enamel knots resembles that of
expression begins to spread around the feather primordia development. Feather
enamel knots, and this spreading is associ- primordia patterning is currently perhaps
ated with cessation of cell proliferation and the best characterized system of iterative
subsequently, together with other signals, patterning. Feather primordia are arranged
including Fgf9 and Shh, with the differen- in specific tracts over the avian body. Feath-
tiation of enamel-forming ameloblasts. The ers, like teeth, are formed by interactions
primary enamel knot, however, seems to between the epithelium and mesenchyme.
possess slightly different gene activity from In feathers, the spacing between feather
the secondary enamel knots, as it also ex- primordia can be altered by affecting molec-
presses Bmp2, the transcription of which ular signaling in vitro. Using small beads
quickly becomes undetectable as the first releasing BMP4 protein, Jung et al. (1998)
secondary enamel knots are activated inhibited feather formation at the distance
(Åberg at al., 1997). Also, as the primary of the protein-releasing bead. This regional
knot disappears apoptotically, its gene ac- inhibition effect, or field, strictly depended
tivity is silenced. on the concentration of BMP protein. In con-
The iterative activation of the secondary trast, FGF4 and SHH protein stimulated
enamel knots allows us to rephrase the the formation of feathers nearby, by produc-
problem of tooth development in slightly ing ectopic feather primordia around the
new terms. The development of tooth shape protein-releasing bead. Jung et al. (1998)
may require the spatial and temporal con- concluded that Fgf and Shh may be the ac-
trol of the secondary enamel knots that are tivators, while Bmp is the inhibitor involved
no different from each other genetically. in the spacing of feathers. Furthermore,
The spatial pattern of the enamel knot ac- both FGF4 and SHH protein induced the
tivation influences the relative position of expression of Bmp4, while BMP4 protein
cusps to each other, while the temporal inhibited the expression of the two former
spacing of the knots influences cusp size. genes (Jung et al., 1998). These experiments
While phylogenetically ancient cusps have are compatible with a reaction-diffusion
been documented to begin their develop- model of morphogenesis (Jung et al., 1998).
ment earlier than phylogenetically recent In this model, an initial point (i.e., signaling
cusps, the main determinant of initiation of center) expressing an activator (e.g., Fgf4)
cusp development is the relative size of each will cause local growth and also the expres-
cusp (Butler, 1956). Tooth development sion of an inhibitor (e.g., Bmp4) to prevent
starts from the formation of the tallest the formation of new signaling centers
cusps, and the growth and folding proceed nearby. However, as the inhibitor will also
cesses may be involved in the production of

patterns.
The reaction-diffusion model might not be
limited to feathers and teeth. Localized ex-
pression of Fgfs, Bmps, and Shh can be
found in other parts of the skin, such as in
the developing mouse tongue papilla (Jung
et al., 1999b), and even the branching mor-
phogenesis of mammal lungs and insect tra-
chea shows similarities with avian feathers
(Hogan, 1999; Metzger and Krasnow, 1999).
Indeed, secondary enamel knots do not
strike us as markedly different from feather
primordia. In teeth, the control of both the
timing and location of secondary enamel
knots within the crown base has to be accu-
rate (Fig. 5). As the same signaling mole-
Fig. 5. Feather patterning and cusp patterning have
many similarities. Feathers and cusps form as itera- cules are used in teeth, it could be that the
tions of similar sets of signaling molecules. For exam- process of patterning is quite similar to that
ple, as a model to regulate patterning, FGF4 functions
as an activator promoting cusp/feather bud initiation
of feather primodia. However, some differ-
and growth. Inhibitors such as BMP4, which have a ences do exist. The exact locations of gene
more diffuse expression domain, control the minimal expression are slightly different. For exam-
distance between adjacent secondary enamel knots/
feathers. The total number of cusps on a tooth would be ple, Bmp4 is expressed in a diffusive, inhi-
limited by the size of the tooth crown base. bition-field manner in the dental mesen-
chyme, and not in the epithelium. Of course,
other Bmps (at least Bmp9) are known to be
inhibit the production of the activator that expressed in the dental epithelium (Åberg
is needed for the production of the inhibitor et al., 1997). Indeed, one of the main differ-
itself, production and spreading of the in- ences among epithelial organs appears to be
hibitor are eventually terminated (Fig. 5). in the details of domains that express the
The result is a spatial pattern of differenti- shared molecular signals. For example,
ated domains that can be altered by any enamel knots express Fgf4 in the epithelia,
changes in activator-inhibitor production, but lung buds express Fgf10 in the mesen-
decay, diffusion, or interaction. As Weiss et chyme next to the tips. However, the com-
al. (1998) pointed out, reaction-diffusion mon property for signaling centers (partly
models describe self-organized systems (or by definition) is their superimposed pattern
Bateson-Turing processes), which are sys- of gene expression.
tems that do not need external information One aspect that seems to set teeth apart
to execute a developmental program once it from feathers is their robustness in defying
is initiated. It is noteworthy that many pro- external perturbations. Experimentally,
cesses, and models, can produce similar ef- teeth cultured in vitro are remarkably resis-
fects, and even cell density is known to af- tant to perturbation by added proteins. It is
fect feather patterning (Jung et al., 1999a). known that, as in feathers and limbs, FGF
Furthermore, in reaction-diffusion models, proteins do, while BMP proteins do not,
the inhibitors would have to diffuse faster stimulate growth in separated tooth epithe-
than the activators in order to create spac- lium and mesenchyme. But only in rare
ing among buds. Interestingly, Hammer cases will the addition of FGF protein in-
(1998) showed that analytically, the effects crease the number of cusps, as in the rescue
of signaling can also be realized by cell-cell- of mutant Tabby teeth (Pispa et al., 1999).
mediated propagation of the signal and not Indeed, teeth appear to be exceedingly self-
by long-distance diffusion. Thus, while reac- organized units. A single tissue containing
tion-diffusion models are compatible with primodia for a mouse first molar can give
current data, several molecular-level pro- rise to all the molars in vitro. We suspect
that the robustness of tooth development is moderate evolutionary diversity in the num-
related to the fact that teeth can change ber of cusps and conules which may have
shape after eruption only by wear, not by properties of evolvability. Indeed, the itera-
remodeling. Thus, evolutionarily speaking, tive activation of secondary enamel knots
there has been perhaps a good reason to may in itself be a mechanism of evolvability,
make tooth development as robust as possi- as the enamel knots can be understood as
ble. Interestingly, Coin et al. (2000) showed developmental “modules.” Modularity of de-
experimentally that dividing cap-staged velopment has been suggested to be one of
mouse molar germs to anterior and poste- the mechanisms of evolvability (Gerhart
rior halves caused disorganization of the and Kirschner, 1997) because it facilitates
primary enamel knot that subsequently be- local, independent changes in morphology.
came restored in the halves. When cultured, That is, pleiotropic effects of genes may be
the distal halves produced almost normal limited if development uses modules that
sets of cusps, suggesting that portions of have only partly overlapping pleiotropic ef-
early tooth germs have a high self-organiz- fects (Minelli, 1998; Wagner and Altenberg,
ing capability. 1996). For example, while gene knockout
experiments with dental defects usually
Evolvability of tooth cusps
have effects on a wide range of organs, in
Secondary enamel knots are helpful in some cases there is redundancy in gene
explaining cusp development, but to use function, and several genes have to be deac-
them to examine molar evolution requires tivated in order to have a specific effect. One
us to interpret them in their micro- and such case is with Dlx homeobox genes,
macroevolutionary contexts. Evolution of where both Dlx1 and Dlx2 have to be deac-
molar shapes is manifested by the conver- tivated in order to have a dental effect,
gent acquisition of new cusps. Most notably, which is the lack of upper molars (Thomas
the distolingual cusp in the upper molars, et al., 1997). The other mouse teeth are un-
the hypocone, has evolved at least 20 times affected, most likely due to the expression of
in mammals (Hunter and Jernvall, 1995). other Dlx genes in their regions (Zhao et al.,
Even among primates, the hypocone has 2000b). However, no cusp-specific differ-
evolved multiple times (Fleagle, 1998). In ences in homeobox gene expression have
many cases, the details (e.g., which part of been reported (Zhao et al., 2000b), and the
the crown gives rise to the hypocone) differ, secondary enamel knot “cusp-making mod-
and subsequently different names are often ules” appear to be alike among all the cusps
used. However, the derived morphologies at the level of molecular signaling. This
can often be such that it is difficult would mean that pleiotropy among second-
(Sánchez-Villagra and Kay, 1996) or impos- ary enamel knots, and cusps in general,
sible to infer the mode of hypocone evolution should be high. Obviously, cis-regulatory el-
without a good fossil record. The evolution ements can be assumed to be involved in the
of the hypocone has been suggested to man- activation of individual secondary enamel
ifest the capacity of tooth development to knots. Indeed, regulatory regions of genes
promote the evolution of new morphologies evolve faster than actual coding regions
(Gerhart and Kirschner, 1997). The capacity (Huynene and Bork, 1998). Could this rate
to evolve, or evolvability, has been tradition- difference, for example, implicate that new
ally inferred in population genetics studies regulatory elements evolve with every cusp?
(in response to selection), but evolvability is This scenario would be a case of extreme
also considered to be a property of biological genetic determinism that does not necessar-
systems that can be studied at various lev- ily make it much easier to evolve new cusps.
els, e.g., by studying development (Capo- But how then has molar tooth diversity
rale, 1999; Gerhart and Kirschner, 1997; evolved and, moreover, can teeth really be
Kirschner and Gerhart, 1998; Wagner and evolvable?
Altenberg, 1996; West-Eberhard, 1998). A possible solution is a patterning cas-
Primate molars, while lacking some more cade mode of cusp spacing that may pro-
extreme morphologies (Fig. 2), still exhibit mote the evolution of new cusps (Jernvall,
growth of the inner enamel epithelium, de-

lay in cusp initiation results in a shorter
cusp (Fig. 6). And as the secondary enamel
knot program is repeated for every cusp,
any small difference in cusp spacing will
have a cumulative effect on later-developing
cusps (Jernvall, 2000) (Fig. 6). The pattern-
ing cascade mode of cusp development pre-
dicts that variation in small cusps should be
cumulative and that random variation
should not necessarily increase in short, lat-
er-developing cusps. This seems to be the
case at least in a population sample of seal
dentition and where the height of the three
tallest cusps can be effectively used to pre-
dict the height and number of short cusps in
a tooth (Jernvall, 2000). However, seal den-
tition is essentially two-dimensional, and it
remains to be seen whether the patterning
cascade explains cusp size variation in more
complex teeth, such as in primate molars.
Interestingly, Polly (1998) studied variation
in cusp position, apart from height, in viver-
ravid carnivores and found that the degree
of growth between cusps may contribute ad-
Fig. 6. Patterning cascade mode of cusp develop- ditional variation to cusp positions. How-
ment. A: Small initial differences in cusp spacing have
cumulative effects on later developing cusps. B: Only ever, it is not known if this is also the case
small changes in the spacing of cusps can increase or with cusp height.
decrease cusp number and the size of small cusps. The A possibly useful character in human mo-
crown height affects the realized cusp numbers globally,
while the patterning cascade affects the potential cusp lars that could be linked to the patterning
pattern. More complex teeth are likely to have partly cascade development is Carabelli’s cusp, an
independent patterning cascades at different parts of
the crown (e.g., paracone-protocone-Carabelli cascade;
extra mesiolingual cusp emerging from the
paracone-metacone cascade). protocone (Hillson, 1996). If a simple seal-
like patterning cascade is applicable to hu-
man molars, then a larger Carabelli’s cusp
2000). As in the models explaining feather would be expected to be present when the
bud patterning, the position of a new sec- height of the protocone, relative to the para-
ondary enamel knot may depend on the pre- cone, is large. However, as the differences
vious enamel knot, and the evolution of new are cumulative, only a very small increase
cusps would not require the evolution of any in size of the protocone would be needed to
new developmental information. Rather, a have a large difference in the size of Cara-
parameter change in the enamel knot pro- belli’s cusp (Fig. 6). While it remains to be
gram affecting the reaction-diffusion pro- documented, this kind of variation could ex-
cess would alter cusp patterns. This also plain why the number and size of small
implies that no external positional informa- cusps and cingula are variable in popula-
tion is needed for any particular cusp, but tions, as natural selection limiting variation
that the whole cusp pattern is a product of in large cusps would not be as effective on
the dynamic program itself (Jernvall, 2000; small cusps. Inversely, selection to increase
Zhao et al., 2000a). For example, a broader the size of small cusps would have relatively
inhibition field around an enamel knot will small effects on large cusps. Thus, the evo-
displace the formation of new secondary lution of new cusps may be a good example
knots further away. Furthermore, as crown of evolvability. Obviously variation in the
shape results from the folding and down- parameters affecting the patterning cascade
mode of cusp formation has to be inherited affecting, for example, the binding property
in order to respond to natural selection of a cis-regulatory element. Flickering can,
(thus, to be evolvable in terms of population however, be even more likely if we consider
genetics). Indirect evidence can be found in the reaction-diffusion models for signaling
the case of seals where a hybrid cusp be- center activation discussed above. For ex-
tween two species (Halchoerus grypus and ample, the evolution, or loss, of the hypo-
Phoca hispida) had an intermediate dental cone does not need to involve the same gene
morphology, including the configuration or molecular-level change in every case. A
and number of cusps (Lönnberg, 1929). similar developmental effect (e.g., delay in
Thus the patterning cascade parameters secondary knot activation) can be produced
are inheritable, but to what extent remains in some evolutionary lineages by altering
to be determined. It is worthwhile to point production/diffusion of activators/inhibi-
out that the patterning cascade mode of tors, and in other lineages by altering the
cusp development is quite compatible with growth rate or differentiation rate. This sit-
the “morphogenetic triangle” that Keene uation, while currently hypothetical, would
(1991) used to approach heterochrony in be an extreme case of phenogenetic drift
tooth development. For example, the itera- (Weiss and Fullerton, 2000), as the same
tions that the patterning cascade is allowed change in phenotype could be produced by
to go through appear to be controlled inde- different changes in genotype. It is worth-
pendently by crown height (Jernvall, 2000). while to bear in mind that while several
Therefore, the potential cusp number and mouse knockout experiments, including the
the actually developmentally realized cusp inactivation of homeobox genes, have pro-
number result from the patterning cascade duced morphologies that superficially re-
and crown height, respectively. semble evolutionary reversals, these changes
The repeated activation of molecularly fail in closer scrutiny (Smith and Schneider,
similar secondary enamel knots as a basis 1998). Thus the lack of “true” evolutionary
for cusp evolvability may have influenced reversals in knockout experiments suggests
the likelihood of cusp evolution in general. that at least the simplest genotype to phe-
Not only may the evolution of new cusps be notype map is unwarranted.
a relatively small change developmentally, Considering cusps as products of the re-
but the loss of cusps may also be equally peated activation of the same molecular ma-
simple. Nevertheless, loss of cusps seems to chinery raises the question, how much indi-
be rare in mammalian radiations (e.g., vidual variation is really allowed for a
mainly in seals and whales). Among pri- single cusp? Evolutionary versatility (Ver-
mates, the loss of the hypocone in marmo- meij, 1973) is likely to play a role in the
sets and tamarins (Callitrichinae) is per- development of cusps in different parts of
haps the only documented case (Kay, 1994: the crown. Thus, trigonid and talonid por-
Fleagle, 1998). tions of the crown may have partly indepen-
Marshall et al. (1994) suggested that re- dent patterning cascades (Jernvall, 2000).
activation of genes can be quite common Moreover, Schwartz (2000) demonstrated a
after up to 6 million years of evolution. This highly diverse patterning of enamel thick-
reactivation can greatly increase the likeli- ness on hominoid cusps. This diversity
hood of repeated disappearance-appear- shows that at least enamel formation can be
ance, or “flickering,” of character states, as controlled in a highly specific manner. In-
they might be lost and gained repeatedly deed, structural enamel genes may be a case
after divergence of lineages. Whatever the of “structural” genetic determinism. Mathur
actual time limits really are, the likelihood and Polly (2000) showed that the diversity
for flickering can be predicted to be substan- of amelogenin splicing variants correlates
tially greater in a situation such as the re- positively with the diversity of enamel mi-
peated activation of secondary enamel crostructure among mammalian lineages.
knots. The increase and decrease in cusp Interestingly, the formation of enamel
number do not need to depend on complete seems to be decoupled with, or at least more
activation of single gene but in a mutation sensitive than, crown growth, as repeated
enamel hypoplasia has been found not to be 2000; Zhao et al., 2000a). For example,
correlated with crown height (Guatelli- Barx1 is found to be expressed only in the
Steinberg and Skinner, 2000). future molar region, while Msx1 and Msx2
are expressed in the preincisor region. Six
Is there a homeobox code for teeth?
known Dlx genes have a complicated combi-
Visible tooth crown development com- natorial code-like expression pattern where
mences with the formation of the primary the region of the upper incisors lacks Dlx
enamel knot. The repeated use of signaling activity altogether, the region for the upper
centers in cusp development is compatible molars expresses two (Dlx1 and Dlx2), the
with dynamic models of development (such lower molars express four (Dlx1, -2, -5, and
as the reaction-diffusion models). The simi- -6), and the lower incisors express three
larities in molecular signaling in cusp, (Dlx2, -5, and -6) genes (Zhao et al., 2000b).
feather, and other periodic patternings sug- The only informative Dlx knockout experi-
gest that the principles for adjusting cusp ment involves the deactivation of Dlx1 and
spacing may be quite “commonly” used dur- Dlx2. These mice lack upper molars (Qiu et
ing development. Another aspect of develop- al., 1997; Thomas et al., 1997). Therefore,
ment inferred to play a role in morphologi- Dlx genes appear not to be an indication by
cal evolution is a combinatorial code of themselves of a dental homeobox code, as
homeobox transcription factors that select the Dlx1 and Dlx2 knockout upper molars
different differentiation fates. In particular, should transform into upper incisors (Weiss
Hox genes are sequentially arranged in the et al., 1998). Weiss et al. (1998) suggested
genome, and they are expressed in a cumu- that the Dlx5 and Dlx6 double knockout
lative sequence along the body axis. Differ- would be an effective test for the Dlx code,
ent combinations of Hox genes are required as these mutants can be hypothesized to
for the correct identity of body segments, transform their lower molars into upper mo-
and altering this code can cause homeotic lars.
changes, i.e., changes in the morphology of Interestingly, Tucker et al. (1998) showed
one part into that of another (Krumlauf, how inhibition of Bmp4 signaling in the pre-
1994). The homeobox code is thus an exam- incisor region of the jaw can alter incisors to
ple of genetic determinism where struc- become molars. As cusp number is in-
ture’s identity can be derived from its indi- creased and root development seems appar-
vidual pattern of gene expression. However, ent in the transformed incisors, this effect
the identity of structures (e.g., thoracic or can indeed be interpreted as an identity
cervical vertebrae; Burke et al., 1995) does change, and not just addition of a new cusp.
not equate with the morphology of individ- Moreover, Barx1, a homeobox gene ex-
ual structures. By analogy, a dental ho- pressed in the premolar region, is expressed
meobox code could potentially be involved in in the preincisor region as a result of Bmp4
the developmental separation of incisor, ca- inhibition, suggesting that homeobox genes
nine, premolar, and molar tooth families are involved in the identity change of incisor
rather than in the morphology of an individ- to molar. It is noteworthy that the tooth
ual molar per se. family-specific expression of genes, includ-
Of the various families of homeobox ing the Dlx genes, is limited to early
genes, Hox genes cannot be used to explain branchial arch development. In later devel-
tooth patterning because Hox gene expres- opmental stages, the genes are reactivated
sion does not extend to the first branchial in all the teeth. The signaling molecule
arch, which gives rise to jaws and teeth Bmp4 appears to be used repeatedly in var-
(Hunt and Krumlauf, 1992; Krumlauf, ious stages of development, including per-
1994). Thus, other homeobox genes would haps even in the induction of the primary
have to be involved in tooth development enamel knot as a mesenchymally expressed
(Weiss, 1993; Weiss et al., 1998). At least signal (Jernvall et al., 1998).
Dlx, Msx, and Barx homeobox-containing Evolutionarily, an apparent role for a ho-
genes have been recognized as having po- meobox tooth code could be a role in the
tential for tooth identity-coding (Sharpe, molarization of the premolars. Currently
there is no comparative evidence to test this opmental processes generating a character

hypothesis, as all studied mammals have can we make reliable inferences on inheri-
only incisors and molars. There are even tance and character independence. One ex-
fewer data to infer how a homeobox code ample is character redundancy. The corre-
could direct the morphology of individual lated variation of cusp heights as a function
teeth (e.g., evolution of the hypocone, evolu- of patterning cascade may cause phenotypic
tion of carnivore carnassials) or tooth size. correlations that reduce character indepen-
Presumably this would require a highly ac- dence. However, simply ignoring short
curate “cocktail” of homeobox genes regulat- cusps in a phylogenetic study is perhaps too
ing all aspects of tooth shape (see below) at a simplistic “cure,” as short cusps are inher-
stages which would precede the formation of ently more variable in a population, making
the primary enamel knot by 2 days in the them useful for lower taxonomic distinc-
mouse. In practice, the homeobox code tions. Studies mapping the extent of mor-
might influence the initial parameters in phological variation in small cusps can be
the enamel knot program, which in turn especially valuable in estimating the varia-
would influence the patterning cascade tional properties of primate teeth (Cuozzo
mode of cusp development. In the end, a and Sauther, 1999).
dental homeobox code is a question of the Below we present groupings of some com-
degree and nature of genetic determinism in monly used dental characters. We stress
the control of morphology. that the groupings of characters are based
on inferred mode of tooth development only,
DEFINING DENTAL CHARACTERS allowing us to make evolutionary predic-
tions when the variational properties of
Grouping characters based
traits are known. This also implies that our
on development
discussion on the likelihood of homoplasy is
The concern that morphologic characters strictly meant as the developmental poten-
may be unreliable in phylogenetic studies tial for character state reversals. Character
stems partly from inferred frequency of con- grouping schemes incorporating additional
vergent evolution (e.g., Collard and Wood, knowledge of fitness, functional biology, and
2000; Hunter and Jernvall, 1995). Indeed, natural selection in connection to traits
detailed studies of morphological charac- have also been suggested (Lovejoy et al.,
ters, stimulated by molecular hypotheses, 1999). While this discerning use of informa-
have illustrated the subtle differences and tion is proposed to allow “formally state the
similarities in dental character evolution presumed morphogenetic basis” of traits, it
(e.g., Fleagle and McGraw, 1999). The evolv- remains to be seen how the role of develop-
ability of tooth cusps may add to this con- ment can be discerned from the other fac-
cern, as from the developmental point of tors involved in character grouping. An-
view, there is very little to prevent conver- other general problem with groupings of
gent evolution of new cusps. Some morphol- character is that our knowledge of varia-
ogies may even be rare due to the patterning tional properties of characters in connection
cascade mode of cusp development (Jern- to development is quite limited. Defining
vall, 2000) increasing the likelihood for con- tooth characters based on molecular-level
vergent evolution. Another concern of char- models directly (e.g., enamel knots, ho-
acters in phylogenetic studies is even more meobox code) is of limited use without a
fundamental, i.e., what is a good, or real, population-level model of development. For
character anyway? Ideally, each character example, inferring cusps to be indepen-
provides reliable information about ances- dently variable because they have their own
try and descent (discussed in Lieberman, signaling centers (secondary enamel knots)
1999; Wiesmüller et al., 2000). Taking a de- or inferring cusps to be highly redundant
velopmental biology point of view, a reliable characters because their secondary enamel
dental character should be discretely iden- knots express the same genes are both ex-
tifiable so that its developmental basis can ceedingly simplistic conclusions. This is par-
be understood. Only by knowing the devel- ticularly apparent when the population-
Fig. 7. Basic tooth crown characters grouped into four developmental categories. Note that degree of
homoplasy is here meant as developmental potential for homoplasy, not actual realized homoplasy.
level mode of variation is known (Jernvall, cumulative variation. The initiation charac-
2000) (Fig. 5). Furthermore, taking the op- ters may not always be independent from
posite strategy by measuring character cor- each other, but this would only be a problem
relations without any consideration of de- when, for example, a large hypocone is
velopment may miss developmentally found to have phenotypic correlation with a
relevant correlations altogether (Jernvall, large metacone. However, the patterning
2000). Thus, classification of characters cascade mode of development may increase
without any mechanistic models of develop- the likelihood of homoplasy and flickering
ment in microevolution could be little more (repeated character state reversals). This is
than geometrical exercises on character cor- particularly the case with short cusps and
relations. These reservations in mind, we dental features. In general, it will be inter-
grouped dental characters below into four esting to test whether evolvability of mor-
categories: initiation, termination, cusp phology increases the actually realized ho-
shape, and configuration characters. moplasy.
Initiation characters. In this paper we Termination characters. These charac-
have mostly addressed initiation characters ters fall into two groups: global and local
(Fig. 7) because their variational properties (Fig. 7). Global termination of tooth crown
are somewhat known (Fig. 6). These charac- formation will affect tooth crown size and
ters are likely to stem from the patterning height and indirectly the presence of short
cascade mode of development that makes cusps. Thus, variation in tooth crown height
short cusps evolvable by increasing their is an additional factor increasing the likeli-
hood of homoplasy and flickering in short ple, an increase in the transverse orienta-
cusps. It is worthwhile to bear in mind that tion of lophs in upper and lower molars may
tooth size characters may be indirect results be developmentally the same process. Fur-
of global termination of tooth crown growth, thermore, the configuration of cusps is
and thus size measures may actually corre- likely to limit configuration and number of
late with the presence of, for example, a roots (Butler, 1956).
small hypocone. Also, the global termina-
CONCLUSIONS: SHOULD WE WORRY
tion of crown development may actually
ABOUT DEVELOPMENT?
have local components (e.g., lingual/buccal).
The local termination of growth affecting Perhaps the greatest value of teeth in evo-
intercuspal areas may influence presence of lutionary biology is their usefulness in in-
lophs. The varitional properties of the ter- ferring diets. The extensive convergence in
mination of intercuspal growth remain to be tooth morphology allows reliable inferences
determined. of dietary correlates and morphology (e.g.,
Fleagle and McGraw, 1999; Jablonski,
Cusp shape characters. The shape of 1994). The evolvability of tooth cusps may
cusps, apart from enamel thickness, is even promote this usefulness, as detailed
largely due to the relative growth of the dental morphologies (e.g., Kay, 1984; Ya-
inner enamel epithelium and the underly- mashita, 1998) may be hypothesized to
ing mesenchyme (Fig. 7). Cusp shape char- evolve quickly to reflect dietary demands.
acters are likely to be global (affecting the However, most evolutionary studies rely on
whole crown), but this remains to be prop- phylogenetic hypotheses that often use den-
erly tested. As these characters do not re- tal characters. And while the evolvability of
sult from simple initiation or termination new cusps may be an exciting phenomenon
events, but instead from changes in growth from the evolutionary biology point of view,
rate, it will be interesting to see if cusp it may also create complications when den-
shapes may be less prone to homoplasy and tal characters are used in constructing phy-
flickering than the initiation and termina- logenies.
tion characters. A developmental basis for characters may
be particularly important to know in cases
Configuration characters. The lateral where tooth shape (and biological shapes in
cusp topography is one of the main aspects general) turns out to be coded by dynamic
of tooth shapes as they ensure, in addition processes that cannot be reduced into a sim-
to cusp heights, proper occlusion of opposing ple genetic code. Already the repeated acti-
teeth. There is preliminary evidence that vation of the developmental modules (i.e.,
changes in lateral topography result from the enamel knots) during tooth develop-
coordinated changes in molecular prepat- ment suggests that homologous cusps and
terns. This prepatterning happens within crests are not coded as such into the ge-
the primary enamel knot preceding the ap- nome, but that the whole cusp pattern is a
pearance of individual cusps (unpublished product of a dynamic program (Jernvall,
results). The early prepatterning of lateral 2000; Zhao et al., 2000a). Also, the pattern-
topography may promote the likelihood of ing cascade mode of cusp development and
homoplasy, because the positions of cusps its morphological manifestation, the initia-
can be altered once during development. tion characters, point toward dynamic pro-
This could be developmentally a simpler cesses in the control of tooth morphogenesis.
and a more robust mechanism than altering It is important to note that cusps can still be
cusp positions by allometric growth during historically homologous. Homology is not
the whole period of crown formation. The just a static genetic code readable deep in-
degree of character independence within a side the genome, but rather, it is a readout
crown remains to be determined. As is the of the information stored in the dynamic
case with initiation characters, these char- cusp-making program. Comparisons to
acters are likely to be strongly linked other developmental systems, e.g., to the
between occluding teeth and thus, for exam- patterning of avian feather primodia,
should aid us in discerning the molecular- Butler PM. 1956. The ontogeny of molar pattern. Biol
Rev. 31:30 –70.
lever components of dynamic processes (e.g., Butler PM. 1982. Some problems of the ontogeny of
reaction-diffusion systems). Another aspect tooth patterns. In: Kurtén B, edtor. Teeth: form, func-
where comparative work is needed is to tion, and evolution. New York: Columbia University
Press. p 44 –51.
understand the role of homeobox codes in Caporale LH. 1999. Molecular strategies in biological
dental patterning. Particularly within pri- evolution. New York: New York Academy of Sciences.
mates, comparative work has great poten- Chen Y, Bei M, Woo I, Satokata I, Maas R. 1996. Msx1
controls inductive signaling in mammalian tooth mor-
tial to help to quantify variational proper- phogenesis. Development 122:3035–3044.
ties of developmental systems. Indeed, in Cohn MJ, Tickle C. 1999. Developmental basis of limb-
lessness and axial patterning in snakes. Nature 399:
contrast to the initiation characters, the 474 – 479.
variational properties of other character Coin R, Lesot H, Vonesch JL, Haikel Y, Ruch JV. 1999.
types are less understood from the develop- Aspects of cell proliferation kinetics of the inner den-
tal epithelium during mouse molar and incisor mor-
mental biology point of view. We predict phogenesis: a reappraisal of the role of the enamel
that examining dental variation by incorpo- knot area. Int J Dev Biol 43:261–269.
rating developmental information in pri- Coin R, Schmitt R, Lesot H, Vonesch JL, Ruch JV. 2000.
Regeneration of halved embryonic lower first mouse
mate populations will offer new ways to test molars: correlation with the distribution pattern of non-
models of development. In this respect, bio- dividing IDE cells, the putative organizers of morpho-
genetic units, the cusps. Int J Dev Biol 44:289–295.
logical anthropology may prove to be a test Collard M, Wood B. 2000. How reliable are human
bench (and hopefully not only the source) for phylogenetic hypotheses? Proc Natl Acad Sci USA
molecular-level “just-so stories” about de- 97:5003–5006.
Cuozzo FP, Sauther ML. 1999. The paleotaxonomic im-
velopment and evolution. And, almost as a plications of dental variation in a living, wild popula-
side product, we may finally be able to dis- tion of Lemur catta. J Vert Pal 19:40A.
cern reliable dental characters for recon- Dawkins R. 1982. The extended phenotype. Oxford: Ox-
ford University Press.
structing phylogenies. Maybe then we Deng C, Zhang P, Harper JW, Elledge SJ, Leder PJ
would not have to worry about development 1995. Mice lacking p21CIP1/WAF1 undergo normal de-
velopment, but are defective in G1 checkpoint control.
but instead use it as a tool in evolutionary Cell 82:675– 684.
studies. Echelard Y, Epstein DJ, St-Jacques B, Shen L, Mohler
J, McMahon JA, McMahon AP. 1993. Sonic hedgehog,
ACKNOWLEDGMENTS a member of a family of putative signaling molecules,
is implicated in the regulation of CNS polarity. Cell
We thank Mikael Fortelius, John Hunter, 75:1417–1430.
Fan CM, Tessier-Lavigne M. 1994. Patterning of mam-
Soile Keränen, Irma Thesleff, and Ken malian somites by surface ectoderm and notochord:
Weiss for discussions or collaborations on evidence for sclerotome induction by a hedgehog ho-
empirical work. We are grateful to Dan molog. Cell 79:1175–1186.
Ferguson BM, Brockdorff N, Formstone E, Ngyuen T,
Lieberman, Chris Ruff, Pat Wright, and an Kronmiller J, Zonana J. 1997. Cloning of Tabby, the
anonymous reviewer for excellent com- murine homolog of the human EDA gene: evidence for
a membrane-associated protein with a short collage-
ments. nous domain. Hum Mol Genet 6:1589 –1594.
Ferguson CA, Tucker AS, Christensen L, Lau AL, Mat-
LITERATURE CITED zuk MM, Sharpe PT. 1998. Activin is an essential
early mesenchymal signal in tooth development that
Åberg T, Wozney J, Thesleff I. 1997. Expression pat- is required for patterning of the murine dentition.
terns of bone morphogenetic proteins (Bmps) in the Genes Dev 12:2636 –2649.
developing mouse tooth suggests roles in morphogen- Fleagle JG. 1998. Primate adaptation and evolution.
esis and differentiation. Dev Dyn 210:383–396. San Diego: Academic Press.
Ahrens K. 1913. Die Entwicklung der menschlichen Fleagle JG, McGraw WS. 1999. Skeletal and dental
Zahne. Arb Anat Inst Wiesbaden 48:169 –266. morphology supports diphyletic origin of baboons and
Bateson W. 1894. Materials for the study of variation, mandrills. Proc Natl Acad Sci USA 96:1157–1161.
treated with special regard to discontinuity in the Gerhart J, Kirschner M. 1997. Cell, embryos and evo-
origin of species. London: Macmillan. lution. Oxford: Blackwell Scientific.
Bateson W. 1913. Problems of genetics. New Haven: Gilbert SF. 1994. Dobzhansky, Waddington, and
Yale University Press. Schmalhausen: embryology and the modern synthe-
Berkovitz BKB. 1967. The dentition of a 25-day pouch- sis. In: Adams M, editor. The evolution of Theodosius
young specimen of Didelphis virginiana (Didelphidae: Dobzhansky. Princeton: Princeton University Press.
Marsupialia). Arch Oral Biol 12:1211–1212. p 143–154.
Bloch-Zupan A, Leveillard T, Gorry P, Fausser JL, Ruch Gilbert SF, Opitz JM, Raff RA. 1996. Resynthesizing
JV. 1998. Expression of p21(WAF1/CIP1) during evolutionary and developmental biology. Dev Biol
mouse odontogenesis. Eur J Oral Sci 1:104 –111. 173:357–372.
Bolk L. 1920. Odontological essays. J Anat 55:138 –186. Gould SJ 1977. Ontogeny and phylogeny. Cambridge,
Burke AC, Nelson CE, Morgan BA, Tabin C. 1995. Hox MA: Harvard University Press.
genes and the evolution of vertebrate axial morphol- Graham A, Francis-West P, Brickell P, Lumsden A.
ogy. Development 121:333–346. 1994. The signalling molecule BMP4 mediates apo-
ptosis in the rhombencephalic neural crest. Nature Jung HS, Wolpert L, Davidson D. 1999a. The formation
372:684 – 686. of the feather pattern in chick skin after a proportion
Guatelli-Steinberg D, Skinner M. 2000. Prevalence and of cells have been killed by X-irradiation. Int J Dev
etiology of linear enamel hypoplasia in monkeys and Biol 43:117–123.
apes from Asia and Africa. Folia Primatol (Basel) Jung HS, Oropeza V, Thesleff I. 1999b. Shh, Bmp-2,
71:115–132. Bmp-4 and Fgf-8 are associated with initiation and
Hamburger V. 1980. Embryology and the modern syn- patterning of mouse tongue papillae. Mech Dev 81:
thesis in evolutionary theory. In: Mayr E, Provine W, 179 –182.
editors. The evolutionary synthesis: perspectives on Kay RF. 1984. On the use of anatomical features to infer
the unification of biology. New York: Cambridge Uni- foraging behavior in extinct primates. In: Rodman
versity Press. p 97–112. PS, Cant JGH, editors. Adaptations for foraging in
Hammer Ø. 1998. Diffusion and direct signaling models nonhuman primates. New York: Columbia University
are numerically equivalent. J Theor Biol 192:129–130. Press. p 21–53.
Hardcastle Z, Mo R, Hui CC, Sharpe PT. 1998. The Shh Kay RF. 1994. “Giant” tamarin from the Miocene of
signalling pathway in tooth development: defects in Colombia. Am. J Phys Anthropol 95:333–353.
Gli2 and Gli3 mutants. Development 125:2803–2811. Keene HJ. 1991. On hetorochrony in heterodonty: a
Heinick P. 1908. Über die Entwicklung des Zahnsys- review of some problems in tooth morphogenesis and
tems von Castor fiber L. Zool Jahrb Abt Anat Ont evolution. Yrbk Phys Anthropol 34:251–282.
Tiere 26:355– 402. Keränen S, Åberg T, Kettunen P, Thesleff I, Jernvall J.
Hillson S. 1996. Dental anthropology. Cambridge: Cam- 1998. Association of developmental regulatory genes
bridge University Press. with the development of different molar tooth shapes
Hogan BL. 1999. Morphogenesis. Cell 96:225–233. in two species of rodents. Dev Genes Evol 208:477–
Holland PWH. 1999. The future of evolutionary devel- 486.
opmental biology. Nature 402:41– 44. Kettunen P, Thesleff I. 1998. Expression and function of
Hornbruch A, Wolpert LJ. 1986. Positional signaling by FGFs-4, -8, and -9 suggests functional redundancy
Hensen’s node when grafted to the chick limb bud. and repetitive use as epithelial signals during tooth
Embryol Exp Morphol 94:257–265. morphogenesis. Dev Dyn 211:256 –268.
Hunt P, Krumlauf R. 1992. Hox codes and positional Kettunen P, Karavanova I, Thesleff I. 1998. Respon-
specifications in vertebrate embryonic axes. Ann Rev siveness of developing dental tissues to fibroblast
Cell Biol 8:227–256. growth factors: expression of splicing alternatives of
Hunter JP, Jernvall J. 1995. The hypocone as a key FGFR1, -2, -3, and of FGFR4; and stimulation of cell
innovation in mammalian evolution. Proc Natl Acad proliferation by FGF-2, -4, -8, and -9. Dev Genet 22:
Sci USA 92:10718 –10722. 374 –385.
Huynene MA, Bork P. 1998. Measuring genome evolu-
Kettunen P, Laurikkala J, Itäranta P, Vainio S, Itoh N,
tion. Proc Natl Acad Sci USA 95:5849 –5856.
Thesleff I. 2000. Associations of FGF-3 and FGF-10
Jablonski NG. 1994. Convergent evolution in the denti-
with signalling networks regulating tooth morpho-
tion of grazing Macropodine marsupials and the
genesis. Dev Dyn 219:322–332.
grass-eating Cercopithecine primate Theropithecus
Kirschner M, Gerhart J. 1998. Evolvability. Proc Natl
gelada. J R Soc West Aust 77:37– 43.
Janis CM, Fortelius M. 1988. On the means whereby Acad Sci USA 95:8420 – 8427.
mammals achieve increased functional durability of Koyama E, Yamaai T, Iseki S, Ohuchi H, Nohno T,
their dentitions, with special reference to limiting Yoshioka H. 1996. Polarizing activity, Sonic hedge-
factors. Biol Rev 63:197–230. hog, and tooth development in embryonic and postna-
Jernvall J. 2000. Linking development with generation tal mouse. Dev Dyn 206:59 –72.
of novelty in mammalian teeth. Proc Natl Acad Sci Kratochwil K, Dull M, Farinas I, Galceran J, Gross-
USA 97:2641–2645. chedl R. 1996. Lef1 expression is activated by BMP-4
Jernvall J, Thesleff I. 2000. Reiterative signaling and and regulates inductive tissue interactions in tooth
patterning during mammalian tooth morphogenesis. and hair development. Genes Dev 10:1382–1394.
Mech Dev 92:19 –29. Krumlauf R. 1994. Hox genes in vertebrate develop-
Jernvall J, Kettunen P, Karavanova I, Martin LB, ment. Cell 78:191–201.
Thesleff I. 1994. Evidence for the role of the enamel Kurtén B. 1953. On the variation and population dy-
knot as a control center in mammalian tooth cusp namics of fossil and recent mammal populations. Acta
formation: non-dividing cells express growth stimu- Zool Fennica 76:1–122.
lating Fgf-4 gene. Int J Dev Biol 38:463– 469. Leche W. 1915. Zur Frage nach der stammesgeschich-
Jernvall J, Hunter JP, Fortelius M. 1996. Molar tooth tlichen Bedeutung des Milchgebisses bei den Säuge-
diversity, disparity, and ecology in Cenozoic ungulate tieren. Zool Jahrb [Abt I] 38:275–370.
radiations. Science 274:1489 –1492. Lieberman DE. 1999. Homology and hominid phylog-
Jernvall J, Åberg T, Kettunen P, Keränen S, Thesleff I. eny: problems and potential solutions. Evol Anthropol
1998. The life history of an embryonic signaling cen- 7:142–151.
ter: BMP-4 induces p21 and is associated with apo- Liem KF Jr, Tremml G, Roelink H, Jessell TM. 1995.
ptosis in the mouse tooth enamel knot. Development Dorsal differentiation of neural plate cells induced by
125:161–169. BMP-mediated signals from epidermal ectoderm. Cell
Jernvall J, Hunter JP, Fortelius M. 2000. Trends in the 82:969 –979.
evolution of molar crown types in ungulate mammals: Lönnberg E. 1929. A hybrid between grey seal, Halicho-
evidence from the Northern Hemisphere. In: Teaford erus grypus and Baltic ringed seal, Phoca hispida
M, Smith MM, Ferguson M, editors. Development, annellata. Ark Zool 21:1– 8.
function and evolution of teeth. Cambridge: Cam- Lovejoy CO, Cohn MJ, White TD. 1999. Morphological
bridge University Press. p 269 –281. analysis of the mammalian postcranium: A develop-
Jung HS, Francis-West PH, Widelitz RB, Jiang T-X, mental perspective. Proc Natl Acad Sci USA
Ting-Berreth S, Tickle C, Wolpert L, Chuong C-M. 96:13247–13252.
1998. Local inhibitory action of BMPs and their rela- Marshall CR, Raff EC, Raff RA. 1994. Dollo’s law and
tionships with activators in feather formation: impli- the death and resurrection of genes. Proc Natl Acad
cations for periodic patterning. Dev Biol 196:11–23. Sci USA 91:12283–12287.
Mathur AK, Polly PD. 2000. The evolution of enamel Tabby phenotype is caused by mutation in a mouse
microstructure: how important is amelogenin? J homologue of the EDA gene that reveals novel mouse
Mamm Evol 7:23– 42. and human exons and encodes a protein (ectodyspla-
Mayr E. 1959. Where are we? Cold Spring Harbor Symp sin-A) with collagenous domains. Proc Natl Acad Sci
Quant Biol 24:1–9. USA 94:13069 –13074.
Mayr E. 1997. The objects of selection. Proc Natl Acad Spemann H, Mangold H. 1924. Uber Induktion von
Sci USA 94:2091–2094. Embryonanlagen durch Implantation artfremder Or-
Metzger RJ, Krasnow MA. 1999. Genetic control of ganisatoren. Roux Arch Entw Mech 100:599 – 638.
branching morphogenesis. Science 284:1635–1639. Thomas BL, Tucker AS, Qiu M, Ferguson CA, Hard-
Minelli A. 1998. Molecules, developmental modules, castle Z, Rubenstein JL, Sharpe PT. 1997. Role of
and phenotypes: a combinatorial approach to homol- Dlx-1 and Dlx-2 genes in patterning of the murine
ogy. Mol Phyl Evol 9:340 –347. dentition. Development 124:4811– 4818.
Nieminen P, Pekkanen M, Aberg T, Thesleff I. 1998. A Tickle C, Summerhill D, Wolpert L. 1975. Positional
graphical WWW-database on gene expression in signalling and specification of digits in chick limb
tooth. Eur J Oral Sci 1:7–11. morphogenesis. Nature 254:199 –202.
Niswander L, Martin GR. 1993. FGF-4 and BMP-2 have Tucker AS, Matthews KL, Sharpe PT. 1998. Transfor-
opposite effects on limb growth. Nature 361:68 –71. mation of tooth type induced by inhibition of BMP
Ohlin A. 1896. Om tandutvecklingen hos Hyperoodon. signaling. Science 82:1136 –1138.
Bihang. Kung Svensk Vet Akad Handl 22:1–33. Vaahtokari A, Åberg T, Jernvall J, Keränen S, Thesleff
Parker SB, Eichele G, Zhang P, Rawls A, Sands AT, I. 1996a. The enamel knot as a signaling center in the
Bradley A, Olson EN, Harper JW, Elledge SJ 1995. developing mouse tooth. Mech Dev 54:39 – 43.
p53-independent expression of p21Cip1 in muscle and Vaahtokari A, Åberg T, Thesleff I. 1996b. Apoptosis in
other terminally differentiating cells. Science 267: the developing tooth—association with an embryonic
1024 –1027. signaling center and suppression by egf and fgf-4.
Peters H, Neubuser A, Balling R. 1998. Pax genes and Development 122:121–129.
organogenesis: Pax9 meets tooth development. Eur van Genderen C, Okamura RM, Farinas I, Quo RG,
J Oral Sci 1:38 – 43. Parslow TG, Bruhn L. 1994. Development of several
Pispa J, Jung HS, Jernvall J, Kettunen P, Mustonen T, organs that require inductive epithelial-mesenchy-
Tabata MJ, Kere J, Thesleff I. 1999. Cusp patterning mal interactions is impaired in LEF-1-deficient mice.
defect in Tabby mouse teeth and its partial rescue by Genes Dev 8:2691–2703.
FGF. Dev Biol 216:521–534. Van Valen L. 1982. Homology and causes. J Morphol
Pizette S, Niswander L. 1999. BMPs negatively regu- 173:305–312.
late structure and function of the limb apical ectoder- Van Valen L. 1994. Serial homology: the crests and
mal ridge. Development 26:883– 894. cusps of mammalian teeth. Acta Palaeontol Pol 38:
Polly PD. 1998. Variability, selection, and constraints: 145–158.
development and evolution in viverravid (Carnivora, Vermeij GJ. 1973. Adaptation, versatility, and evolu-
Mammalia) molar morphology. Paleobiology 24:409 – tion. Syst Zool 22:466 – 477.
429. von Korff K. 1932. Zusammensetzung und Formbildung
Qui M, Bufone A, Ghattas I, Menses JJ, Sharpe PT, der kompliziert Gebauten Zähne. Jahrb Morphol
Presley R, Pedersen RA, Rubenstein JLR. 1997. Role Mikrosk Anat [Abt 2] 29:561–591.
of the Dlx homeobox genes in proximodistal pattern- Wagner GP, Altenberg L. 1996. Complex adaptations
ing of the bronchial arches: mutations of Dlx-1, Dlx2, and the evolution of evolvability. Evolution 50:967–
and Dlx-1 and -2 after morphogenesis of proximal 976.
skeletal and soft tissue structures derived from the Wagner PJ. 2000. Exhaustion of morphologic character
first and second arches. Dev Biol 185:165–184. states among fossil taxa. Evolution 54:365–386.
Raff RA. 1996. The shape of life. Genes, development Weiss KM. 1990. Dublication with variation: metameric
and the evolution of animal form. Chicago: Chicago logic in evolution from genes to morphology. Yrbk
University Press. Phys Anthropol 33:1–23.
Riddle RD, Johnson RL, Laufer E, Tabin C. 1993. Sonic Weiss KM. 1993. A tooth, a toe, and a vertebra: the
hedgehog mediates the polarizing activity of the ZPA. genetic dimensions of complex morphological traits.
Cell 75:1401–1416. Evol Anthropol 2:121–134.
Roelink H, Augsburger A, Heemskerk J, Krozh V, Nor- Weiss KM, Fullerton SM. 2000. Phenogenetic drift and
lin S, Ruiz i Altaba A, Tanabe Y, Placzek M, Edlund the evolution of genotype-phenotype relationships.
T, Jessell TM, Dodd J. 1994. Floor plate and motor Theor Popul Biol 57:187–195.
neuron induction by vhh-1, a vertebrate homolog of Weiss KM, Stock D, Zhao Z. 1998. Dynamic interactions
hedgehog expressed by the notochord. Cell 76:761–775. and the evolutionary genetics of dental patterning.
Sánchez-Villagra MR, Kay RF. 1996. Do Phalangeri- Crit Rev Oral Biol Med 9:369 –398.
forms (Marsupialia: Diprotodontia) have a “hypo- West-Eberhard MJ. 1998. Evolution in the light of de-
cone”? Aust J Zool 44:461– 467. velopmental and cell biology, and vice versa. Proc
Schwartz GT. 2000. Taxonomic and functional aspects Natl Acad Sci USA 95:8417– 8419.
of the patterning of enamel thickness distribution in Wiesemüller B, Rothe H, Henke W, Lieberman DE.
extant large-bodied hominoids. Am J Phys Anthropol 2000. Good characters and homology. Evol Anthropol
111:221–244. 9:151–152.
Sharpe PT. 2000. Homeobox genes in initiation and Yamashita N. 1998. Functional dental correlates of food
shape of teeth during development in mammalian properties in five Malagasy lemur species. Am J Phys
embryos. In: Teaford M, Smith MM, Ferguson M, Anthropol 106:169 –188.
editors. Development, function and evolution of teeth. Zhao Z, Weiss KM, Stock DW. 2000a. Development and
Cambridge: Cambridge University Press. p 3–12. evolution of dentition patterns and their genetic ba-
Smith KK, Schneider RA. 1998. Have gene knockouts sis. In: Teaford M, Smith MM, Ferguson M, editors.
caused evolutionary reversals in the mammalian first Development, function and evolution of teeth. Cam-
arch? Bioessays 20:245–255. bridge: Cambridge University Press. p 152–172.
Srivastava AK, Pispa J, Hartung AJ, Du Y, Ezer S, Zhao Z, Stock DW, Buchanan AV, Weiss KM. 2000b.
Jenks T, Shimada T, Pekkanen M, Mikkola ML, Ko Expression of Dlx genes during the development of
MS, Thesleff I, Kere J, Schlessinger D. 1997. The the murine dentition. Dev Genes Evol 210:270 –275.

American J Phys Anthropol - 2001 - Jernvall - Genotype Phenotype and Developmental Biology of Molar Tooth Characters

Uploaded by

Copyright:

Available Formats

You might also like

American J Phys Anthropol - 2001 - Jernvall - Genotype Phenotype and Developmental Biology of Molar Tooth Characters

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

American J Phys Anthropol - 2001 - Jernvall - Genotype Phenotype and Developmental Biology of Molar Tooth Characters

Uploaded by

Copyright:

Available Formats

YEARBOOK OF PHYSICAL ANTHROPOLOGY 43:171–190 (2000)

Genotype, Phenotype, and Developmental Biology of Molar

KEY WORDS dentition; teeth; morphogenesis; evolvability; dental

ABSTRACT Primate molar shapes reflect developmental and ecological

Introduction ......................................................................................................................... 172

© 2000 WILEY-LISS, INC.

Configuration characters ............................................................................................. 187

INTRODUCTION tire human genome may transform the field

we can indeed infer phenotype from geno-

common taxonomic diversity measures,

of cell proliferation as a part of a develop-

cesses may be involved in the production of

growth of the inner enamel epithelium, de-

there is no comparative evidence to test this opmental processes generating a character

You might also like