The hidden teeth of sloths: Evolutionary vestiges and the development of a simplified dentition

Abstract

Xenarthrans are unique among mammals in retaining simplified teeth that are rootless and homodont, which makes it difficult to determine dental homologies. We apply computerized tomography to prenatal developmental series of extant sloths, Bradypus and Choloepus, to further elucidate the patterns of morphological variation in their dentition. We also propose new criteria based on sequences of dental mineralization, and the presence of vestigial teeth, to distinguish between caniniforms and postcaniniforms. We report for the first time the presence of vestigial incisors in Bradypus. We also show the presence of a vestigial tooth in front of the lower caniniform in both extant sloth genera and the existence of two generations for the upper caniniform in Choloepus. The study of their sequence of mineralization indicates that the lower and upper caniniform teeth are not homologous in sloths, and suggests that upper caniniforms are not homologous between the two extant sloth genera. Our results show that assessing the developmental processes and functional constraints remains crucial to understand the dental variations observed in sloths, and more generally, tooth class homology issues in mammals. Applied to the tooth row of all extinct sloths, these developmental data illuminate a potentially ancestral dental formula for sloths.

Full text

1
Scientific RepoRts | 6:27763 | DOI: 10.1038/srep27763
www.nature.com/scientificreports
The hidden teeth of sloths:
evolutionary vestiges and the
development of a simplied
dentition
Lionel Hautier1, Helder Gomes Rodrigues2,3, Guillaume Billet2 & Robert J. Asher4
Xenarthrans are unique among mammals in retaining simplied teeth that are rootless and homodont,
which makes it dicult to determine dental homologies. We apply computerized tomography to
prenatal developmental series of extant sloths, Bradypus and Choloepus, to further elucidate the
patterns of morphological variation in their dentition. We also propose new criteria based on sequences
of dental mineralization, and the presence of vestigial teeth, to distinguish between caniniforms and
postcaniniforms. We report for the rst time the presence of vestigial incisors in Bradypus. We also
show the presence of a vestigial tooth in front of the lower caniniform in both extant sloth genera and
the existence of two generations for the upper caniniform in Choloepus. The study of their sequence
of mineralization indicates that the lower and upper caniniform teeth are not homologous in sloths,
and suggests that upper caniniforms are not homologous between the two extant sloth genera. Our
results show that assessing the developmental processes and functional constraints remains crucial to
understand the dental variations observed in sloths, and more generally, tooth class homology issues
in mammals. Applied to the tooth row of all extinct sloths, these developmental data illuminate a
potentially ancestral dental formula for sloths.
Like other xenarthrans (sloths, armadillos, and anteaters), living and extinct sloths (Folivora) depart from the rest
of mammals by the simplied nature of their dentition. Teeth present in most xenarthran adults lack enamel and
are usually homodont, ever-growing, tubular and primarily composed of orthodentine and vasodentine1, which
makes it dicult to identify homologies with the teeth and cusps of other mammals. Both extant sloth genera are
functionally monophyodont, and their dentition is generally considered to constitute a single set of permanent
teeth2–6. e sloth dentition contrasts with the complete lack of teeth in anteaters, and the supernumerary teeth
of armadillos. It mainly diers from that of other xenarthrans in showing a morphological distinction between
caniniforms and molariforms, a dierence based on the general morphology, occlusion, and position of their
teeth.
Recent morphological and molecular phylogenetic analyses7–10 suggested that the two modern genera are only
distantly related, with a divergence time that could be as long as 30 million years ago11. Despite this independent
evolutionary history, both two-toed and three-toed sloths display identical dental formulae with ve upper and
four lower teeth, as do the majority of extinct sloth genera1,10. is apparent stability in number associated with
the dierentiation of the tooth row observed in extant forms masks a complex evolution of the dentition in foli-
vorans (i.e., modern sloths end extinct gravigrade sloths). Bradypus shows a closely t toothrow, lacking diastema
with each tooth showing a peg-like morphology. Choloepus displays an enlarged, chisel-shaped caniniform at the
front of the dentition and isolated from the molariforms by a diastema1,12.
While the intriguing nature of the xenarthran teeth has attracted a lot of attention, few studies have focused
on the development of the whole dentition13, especially in sloths. Early workers have only described isolated
1Institut des Sciences de l’Evolution de Montpellier, Université Montpellier, CNRS, IRD, EPHE, Cc 064; place Eugène
Bataillon, 34095 Montpellier Cedex 5, France. 2Sorbonne Universités, CR2P, UMR CNRS 7207, Univ Paris 06, Muséum
national d’Histoire naturelle, 8 rue Buon, 75005 Paris, France. 3Mécanismes adaptatifs et évolution (MECADEV),
UMR 7179, CNRS, Funevol team, Muséum national d’Histoire naturelle, 55 rue Buon, Bat. Anatomie Comparée,
CP 55, 75005 Paris, France. 4Department of Zoology, University of Cambridge, Downing St., Cambridge CB2 3EJ,
UK. Correspondence and requests for materials should be addressed to L.H. (email: lionel.hautier@univ-montp2.fr)
received: 14 February 2016
Accepted: 24 May 2016
Published: 14 June 2016
OPEN

www.nature.com/scientificreports/
2
Scientific RepoRts | 6:27763 | DOI: 10.1038/srep27763
foetuses of sloths or focused on the developmental sequence of their skeleton14,15. However, these do not detail
the development of the teeth or provide a comparative basis upon which to analyse possible homologies with the
dentition of other mammals. Using a large dataset of scanned foetuses of sloths, we provide data on xenarthran
prenatal dental ontogeny, identify some developmental criteria with which to recognize homologies with other
mammalian teeth, and propose a new hypothesis for the development of heterodonty in sloths.
Results
A terminology specic to sloths has been used to avoid confusion and in order to draw reliable comparisons (S1):
pmx stands for premaxilla; d stands for deciduous teeth; cf and mf stand for lower caniniforms and molariforms
respectively, while Cf and Mf stand for upper teeth; lower loci 1–3 and upper loci 1–4 involve functional molar-
iform teeth; v and V stand for vestigial lower and upper teeth respectively (i.e., these loci are absent in adults).
Prenatal dental development in three-toed sloths. The sequence of prenatal dental eruption in
Bradypus is well resolved with 18 specimens that represent a variety of developmental stages. Eleven of 18 foe-
tuses display a number of developing teeth dierent from those observed in adults, which (as noted previously)
are characterized by ve upper and four lower teeth (i.e., 5/4). All the alveoli of the adult teeth are present early
during dental development, but lack teeth. is implies that dental buds are developing but not yet mineralizing;
these buds cannot be directly observed because so tissues are dicult to detect using X-ray microtomography
Figure 1. Lateral view of three-dimensional reconstruction of CT-scans of skull in the three-toed sloth
Bradypus. (A) Bradypus variegatus (ZMB 33812), SL = 23 mm; (B) Bradypus variegatus (ZMB 41122),
SL = 26 mm; (C) Bradypus variegatus (MNHN-ZM-MO-1995-326A), SL = 26 mm; (D) Bradypus variegatus
(ZMB 41120), SL = 42 mm; (E) Bradypus tridactylus (BMNH 52-1173), SL = 42 mm; (F) Bradypus sp. (MNHN-
ZM-MO-1995-327), SL = 38 mm. Upper teeth are in violet; lower teeth are in green; premaxillary bone is in red.

www.nature.com/scientificreports/
3
Scientific RepoRts | 6:27763 | DOI: 10.1038/srep27763
without so-tissue staining. In the youngest specimen (ZMB 33812, SL = 23.7 mm, Figs1A and 2A), only the
mesialmost pairs of teeth are mineralizing on the maxillary and the dentary (dCf and dv), meaning that dentine
formation has started. In contrast to other teeth, the rst pair of uppers (dCf) to appear are not centered in their
alveolus, but sit o-centre in the anterolateral corner of the alveolus. e dCf is the rst locus to mineralize, but
its size does not change drastically during the rst stages. Other teeth mineralize aer dCf, but grow more quickly
(Figs1A–D and 2). e second youngest specimen (ZMB 41122, SL = 25.9 mm, Figs1B and 2B) shows mineral-
ized dCf, dMf2, dMf3, dv, dmf1, and dmf2, with empty alveoli (i.e., teeth not yet mineralizing) at the dMf1, dMf4,
dcf, and dmf3 loci. All the alveoli include mineralized teeth in subsequent stages.
More importantly, four of 18 specimens have six upper and ve lower teeth; they show an extra pair of teeth
on the premaxilla (dVpmx, Fig.1C,D), which correspond to rudimentary incisors, absent in the adults. ese
incisors can be retained until relatively late in development (e.g., ZMB 41120, SL = 41.84 mm, Fig.1D), but are
resorbed before birth. Five specimens have a dental formula composed of ve upper and ve lower teeth. e rst
pair of lower teeth (dv), just mesial to the lower caniniform (dcf), is also resorbed during development, likely aer
the small incisors, and is absent in later stages. Both extra upper and lower teeth (dVpmx and dv) are apparent on
both right and le sides, they do not have visible extension of the root, and are never associated with alveoli. We
observed no major dental dierences between B. variegatus and B. tridactylus, both of which exhibited similar
morphology at comparable stages.
Prenatal dental development in two-toed sloths. e skull length (see S2 for measurements) and
number of discrete ossication centres of the cranium indicate that most of the specimens of Choloepus corre-
spond to relatively late stages compared to Bradypus. However, as for Bradypus, the number of teeth varies greatly
among our specimens and diers from the morphology observed in adult Choloepus. While the adult dental
formula is 5/4, the foetuses showed either ve upper and ve lower teeth (60% of the cases) or six upper and ve
lower teeth (40% of the cases). All specimens display an extra tooth on the mandible in front of the functional
adult tooth row (dv), i.e., in front of the lower caniniforms (dcf). Two younger specimens also show an extra
tooth in the maxilla (dCf; Fig.3A) in front of the upper caniniforms (Cf); these extra teeth are located in the same
alveoli as Cf and are oriented mesio-buccally (Fig.3A,B). As observed for Bradypus, all extra teeth are present
on both right and le sides; however, in contrast to Bradypus, no vestigial incisor (dVpmx) was observed in the
premaxilla of Choloepus, although we cannot rule out its presence in specimens younger than those in our sample.
No extra teeth were detected at the level of the diastema that separates the mesialmost functional tooth from the
Figure 2. Palatal view of three-dimensional reconstruction of the maxillary bones in early developmental
stages of the three-toed sloth Bradypus. (A) Bradypus variegatus (ZMB 33812), SL = 23 mm; (B)
Bradypus variegatus (ZMB 41122), SL = 26 mm; (C) Bradypus variegatus (MNHN-ZM-MO-1995-326A),
SL = 26 mm; (D) Bradypus variegatus (MNHN-ZM-MO-1995-326B), SL = 30 mm; (E) Bradypus sp. (MNHN-
ZM-MO-1902-325), SL = 30 mm; (F) Bradypus sp. (MNHN-ZM-MO-1995-327), SL = 38 mm. Upper teeth are
in violet; premaxillary bone is in red. Dashed lines represent dental alveoli. Abbreviations: b.c., bony crypt.

www.nature.com/scientificreports/
4
Scientific RepoRts | 6:27763 | DOI: 10.1038/srep27763
molariforms in any of the specimens studied. As observed in Bradypus, the lower teeth (dv) do not have visible
extension of the root and are never associated with alveoli, in contrast to the teeth present in adults. We observed
no major dental dierences between the two species of Choloepus.
Discussion
A history of prenatal dental development in extant sloths. For Böker16, “sloths will always remain
the poor sibling of comparative anatomy, because it is neither possible to obtain complete anatomical series, nor
does the possibility currently exist to gain good insights from paleontology. Only ontogeny promises good prospects
to enlighten the history of sloths’ modications. Yet this does not promise an easy path either, because not only are
important developmental stages very dicult to nd, but even when embryos are available, they show that devel-
opment of key anatomical structures seems to occur at very early stages” (in German in the text). Comparisons of
tooth development in sloths suggests that, for at least some anatomical regions, Böker is correct that ontogeny
is a key source of information about homology. However, he is too pessimistic that sloths will remain poorly
understood compared to other animals. Extant and extinct sloths display quite homodont teeth, which makes it
dicult to determine dental homologies. Prenatal dental development in sloths shows teeth that are cone-shaped
and monocuspid (see also17,18, and demonstrates that robust hypotheses of homologies cannot be drawn based
on occlusal patterns alone (e.g.19,20), which simply result from rapid wear (i.e., cusp-like pattern). However, our
data show that Bradypus and Choloepus display several pairs of supernumary teeth in the mesial part of their
dentition during prenatal ontogeny. is is consistent with previous, anecdotal accounts of dental “anomalies”
in sloths2,6,17,21–23. Brandts21 (cited in Röse22) reported the rst record of vestigial lower teeth in Bradypus and
believed them to be canines. Parker2 observed similar vestiges in Choloepus embryos, which made him recognize
the lower caniniform as a premolar locus. Gervais17 seemed unaware of Brandts’ reference when he described the
presence of vestigial teeth in the mandible of a foetus of Bradypus, which he considered to be incisors based on
their closeness to the symphysis as well as their position compared to the caniniforms. Simon23 reached similar
conclusions with two foetuses of Bradypus (CRL = 23.5 and 24.2 cm); he also proposed, probably for the rst time,
the non-homology between the lower and upper caniniforms based on dental eruption sequences (the upper
caniniform erupting much later than the upper molariform and the lower caniniform).
Figure 3. A comparison of the tooth rows in Choloepus and Bradypus. (A) lateral view of the skull of
Choloepus. (B,C) palatal views of the tooth rows in C. didactylus ((B) MNHN-ZM-MO-1882-625, SL = 45 mm)
and B. variegatus ((C) MNHN-ZM-MO-1995-326B, SL = 30 mm). Note the similar (o-centre) position of the
dCf in the mesialmost alveolus.

www.nature.com/scientificreports/
5
Scientific RepoRts | 6:27763 | DOI: 10.1038/srep27763
Occurrence of vestigial teeth and dental anomalies in sloths. Our data show that Bradypus and
Choloepus display several pairs of supernumary teeth in the mesial part of their dentition during prenatal ontog-
eny. is is consistent with previous, anecdotal accounts of dental “anomalies” in sloths. High-resolution X-ray
computed tomography on a large number of foetuses demonstrated that these extra teeth cannot simply be
explained by individual variation that occurs exceptionally within a population. Such methods also revealed that
these teeth lack typical roots and do not develop in clearly individualized alveoli, unlike other teeth. More impor-
tantly, these vestigial teeth occur on both right and le sides, which is rarely the case with sloth dental anoma-
lies12. Sloths display fewer dental anomalies compared to other mammals (i.e., 2.4% of adult specimens observed
by McAfee12 exhibited any sort of anomalies), and increases in tooth number (i.e., hyperdontia) occur at a much
lower rate than reductions12. Bradypus is more prone to lose teeth and all cases of tooth loss involve the dCf12,
which could be logically expected since it is the most reduced tooth of the dentition24. Interestingly, nearly all of
the anomalies observed in adults (unpaired hyperdontia and anodontia) aected the upper dentition in sloths12
while vestigial teeth were most consistently observed on the mandible. In fact, the only paired hyperdontia anom-
aly ever reported on the mandible (Fig.1C)12 probably corresponds to a specimen that failed to resorb the mesial
vestigial teeth (dcf) observed in the foetuses. is reinforces the idea that the mineralization and resorption of the
vestigial teeth is an integral part of prenatal dental development in sloths. All of these teeth (dVpmx in Bradypus,
dCf in Choloepus, and dv in both) correspond to the denition of vestigial structures given by Peterkova et al.25;
they occur transiently during development in all members of a population, and on occasion persist into maturity.
While both extant genera have similar development of the lower teeth, including the mineralization of paired
mesial vestigial teeth (dv), dierences are evident on the upper jaw, with Bradypus displaying premaxillary vestig-
ial teeth and Choloepus maxillary but no premaxillary vestigial teeth.
First evidence of tooth replacement in sloths. Vestigial teeth were observed in the maxillae of
Choloepus (dCf). ese vestigial teeth are very close to the caniniform teeth (Cf). ey are located in the same
alveolus, and appear apically with respect to the Cf (Fig.3A,B), as expected in vertical dental replacement
(e.g.26,27). Determining the epithelial connections of teeth during early developmental stages provides the best
criterion for dening the deciduous or permanent homologies of individual teeth28,29, but unstained CT data do
not convey this information on the dierentiation of the dental lamina. However, based on positional data we
interpret the vestigial upper tooth (dCf) and the caniniform (Cf) in Choloepus as deciduous and permanent teeth
of the same locus30. Such an occurrence of non-functional vestigial deciduous teeth, rapidly replaced by perma-
nent teeth, has already been reported in other mammalian groups such as marsupials (e.g. Perameles29), soricids
(e.g., Sorex, Suncus27,31), and mustelids (e.g., Mephitis, Enhydra32,33). e presence of two dental generations in a
folivoran is here reported for the rst time. Our results hint at diphyodonty for at least one locus in sloths (i.e.,
the caniniform in Choloepus), and are consistent with the expectation that ancestral xenarthrans possessed tooth
replacement as typical for mammals. It can then be stated that the diphyodonty is a symplesiomorphy of the
Xenarthra, as it is shared by the extant sloth Choloepus and the armadillo Dasypus13,34.
Dental homologies between sloths. When considering the dental mineralization sequence in Choloepus,
homologies are not obvious between the upper and lower tooth rows. is is partly due to the limited resolution
of our ontogenetic sequence and the lack of data on early developmental stages for this genus. e youngest spec-
imen already shows advanced stage of mineralization for the whole dentition. is might explain why vestigial
incisors (dVpmx), rapidly resorbed in Bradypus, are not observed in Choloepus.
In contrast, the dierent ontogenetic stages of Bradypus enable precise hypotheses of dental homologies in
extant sloths. Upper caniniforms (dCf) and lower vestigial (dv) teeth are the rst teeth to start their mineral-
ization in Bradypus (Fig.1A–C) and probably belong to the same locus since they do so simultaneously35. e
similar development of the dCf in Choloepus and in Bradypus (Fig.3B,C), both in terms of size and position in
the alveolus, and the similar early stages of development between dCf and dv in Bradypus, allow us to hypothesise
that upper and lower vestigial mineralized buds of Choloepus (dCf and dv) are homologous. Such an explana-
tion would imply that the upper caniniforms are not homologous in the two extant genera of sloths, with adults
Choloepus showing a permanent caniniform (Cf) for that locus while adults Bradypus retain a deciduous canini-
form (dCf). Following this hypothesis, the deciduous upper teeth present at a vestigial state in Choloepus would
be functional in Bradypus in concert with an absence of a permanent generation for that locus. e large bony
crypt long observed during the mineralization of the dCf of Bradypus would then represent an embryological
holdover when the permanent tooth primordia (Cf) was still activated for that locus. Such an assumption is
supported by a case of bilateral anomaly in Bradypus12 (Fig.3A) involving both occurrences of mesial dCf and a
distally large Cf, which corresponds to the conguration observed in the foetal series of Choloepus.
Alternatively, rather than a retained deciduous caniniform in adult Bradypus, it could be proposed that succes-
sion at this locus is not represented in our ontogenetic series for that genus. Following this alternative hypothesis,
the upper functional caniniforms of both genera are homologous and correspond to permanent teeth. en,
dCf observed during the ontogeny of Bradypus would correspond to a vestigial deciduous canine that would
eventually be replaced later on by a permanent tooth (Cf). is would imply that we are missing several early
developmental events in Bradypus, between putative mineralization of Cf and reabsorption of dCf, which appears
unlikely considering that we were able to trace the evolution in shape and size of the outline of the mesialmost
alveolus (Fig.2) and that Bradypus is the best-sampled genus in terms of the number of dierently sized stages.
e rst hypothesis is thus preferred here.
Simplied dentition vs the mammalian tooth row. Upper caniniforms (dCf) and lower vestigial (dv)
teeth are the rst teeth to start their mineralization in Bradypus (Figs1A and 2A). Following the hypothesis that
dCf and dv represent homologous deciduous teeth in Choloepus and Bradypus, these teeth are deciduous canines

www.nature.com/scientificreports/
6
Scientific RepoRts | 6:27763 | DOI: 10.1038/srep27763
since the dC is one of the earliest tooth germs to dierentiate in eutherian mammals29,35,36. e extreme mesial
position of dCf on the maxillary bone (Figs1 and 2), near the suture with the premaxilla, also provides further
support in favour of this attribution.
e mineralization of dCf and dv is followed by the second and third upper molariforms (dMf2-3), which
mineralize simultaneously with the rst and second lower molariforms (dmf1-2). Such development is remi-
niscent of the initiation of distal milk premolars in mammals (e.g., dP3-428,29) although the proposed sequence
is only based on two foetal Bradypus specimens. e vestigial premaxillary teeth (dVpmx) of Bradypus likely
correspond to vestigial incisors that are resorbed during development and are never observed in adult specimens,
not even as anomalies12. e relative timing of mineralization of the dVpmx, rst upper molariform (dMf1),
and lower caniniform (dcf) cannot be precisely determined based on our dataset. However, their mineralization
likely occurs shortly before that of the distalmost upper and lower molariform teeth (dMf4 and dmf3) since the
latter only show an incipient mineralization in the youngest specimen with evidence bearing on that locus (spec-
imen MNHN CG 1995-326A, Figs1C and 2C). If the general trends observed when studying the mammalian
dental mineralization sequence (e.g.28,29,37) are valid for sloths, the last molariforms (dMf4 and dmf3) should be
considered as rst molars, preceded by three deciduous premolars (dMf1-3; dcf-dmf2) and a deciduous canine
(dCf; dv). Notably, the early diverging living cingulate (Dasypus) has also been interpreted as exhibiting a single,
unreplaced M1 locus in each jaw quadrant, preceded by replaced premolariforms and possibly a canine locus34.
However, in contrast to Dasypus, no known folivoran shows replacement of functional milk teeth, and there is
no empirical evidence to support this hypothesis since the possibility of a violation of the “normal” mammalian
sequence cannot be entirely ruled out. e sloth dental formula might include supernumary teeth, as present
for instance among the premolars of cingulates13,34 or in Mesozoic groups like docodonts or morganucodonts38.
In any case, these results support the assumption that the upper caniniforms present in adult Bradypus likely
represent canines and that the upper and lower caniniforms (dCf, dcf) are not homologous since they mineral-
ize at very dierent times during ontogeny35. e dcf can then be considered as a premolar locus, which might
be homologous to dMf1 when compared their timing of mineralization. Such a hypothesis of non-homology
between dCf and dcf was proposed early on and stemmed mainly from the fact that the upper caniniforms in
sloths occlude with the mesial surface of the lower caniniforms, while upper canines occlude with the distal edge
of the lower canines in other placentals3,24. Simon23 also noted that upper caniniforms of Bradypus erupt well aer
the lower caniniforms, although some studies on both extinct and extant mammals prefer developmental prenatal
dental data over eruption sequences to establish dental homologies29,35.
Reconstructing the ancestral dental formula of sloths. Most fossil sloths show a very similar number
of teeth compared to adult specimens of extant species with ve upper and four lower teeth. Except for dental
anomalies12 and for the dubious fossil sloth Entelops39,40, this number is never exceeded in extinct folivorans (S3).
Some mylodontids and nothrotheriids show a reduction in tooth number with a loss of upper caniniforms10
(S3). It is therefore parsimonious to propose an ancestral dental formula of ve upper and four lower teeth for
Folivora10 and match our reconstruction of the sloth ancestral dental formula (S3). Our developmental data
oer new evidence for the loss of teeth during the evolutionary history of sloths in showing that some loci have
been retained in foetal stages of extant forms. e vestigial upper incisors found in Bradypus embryos was never
reported in any other sloth, but may have been present in the earliest folivorans. e lower vestigial tooth dv was
reported in both Bradypus and Choloepus. Given the phylogenetic distance between the two genera7–10, it is likely
that such a vestige was also present in early ontogenetic stages of the most recent common ancestor of Folivora.
is idea is corroborated by the rare occurrence of teeth at a similar position in some fossil sloths41–43.
We showed that the functional upper caniniforms are very dierent in the two genera: Choloepus shows an
ephemeral and tiny mineralized bud of dCf associated with a massive caniniform Cf, whereas Bradypus shows a
moderately-sized, peg-like tooth dCf. Intermediate stages between these extreme patterns may well have occurred
in fossil sloths and would have shown the succession of a Cf to a well-mineralized dCf. However, evidence for such a
succession (for instance an erupting Cf in juvenile or subadult stages) remains unknown in the fossil record of sloths.
is absence might lie in the scarcity of well-documented ontogenetic series for fossil sloths (e.g.18), although rela-
tively few subadult stages are needed to document a succession of dental generations (e.g.44). So far, no succession
of dental generations at the Cf locus has been observed in megatherioids, one of the potential allies to Bradypus9,45.
Another explanation for the absence of dental replacement at the Cf locus in the fossil record of sloths may actually
lie in the potential autapomorphic condition of the pattern observed in Bradypus. is genus is thought to represent
a paedomorphic lineage when compared its skull morphology to other folivorans46,47. e retention of the dCf and
absence of functional Cf in adults supports the concept of a paedomorphic Bradypus and could constitute another
retained juvenile feature. If the retention and subsequent growth of dCf are unique to Bradypus, it is not surprising
that no intermediate stage was found in the fossil record as close fossil relatives of three-toed sloths remain virtually
unknown45. A highly autapomorphic condition in Bradypus can also account for many morphological discrepancies
and could have erased or modied several inherited folivoran synapomorphies in this genus; this could explain why
it is retrieved as fully basal10 while it might instead be more apically nested within the folivoran clade9,12.
In Bradypus, the retention of a short rostrum associated with the development of a large dMf1 may have inhib-
ited the development and mineralization of a large permanent caniniform (Cf, Fig.2D). is ontogenetic pattern
gives room for a complete mineralization of the deciduous caniniform (dCf), which remains reduced (Fig.2E).
e growth of dCf seems to be “reactivated” only when the mineralization of all other teeth is well underway
(Fig.2E); only then does it quickly acquire its adult size. Interestingly, such a development of vestigial teeth that
recover functionality has been proposed in a few mammalian species that show a reduced and simplied den-
tition, explained by minor developmental modications (e.g., frequent recovering of dP4 in the murine rodent,
Rhynchomys48). is lends further credence to the ndings of Simon23 who noticed the late eruption of the upper
caniniforms compared to the lower caniniforms and all molariforms in Bradypus.

www.nature.com/scientificreports/
7
Scientific RepoRts | 6:27763 | DOI: 10.1038/srep27763
Following the hypothesis of an autapomorphic condition in Bradypus, the small dCf mineralized bud observed
during the ontogeny of both extant genera might represent an ancestral feature for Folivora. Most of the diversity in
shape of the “caniniforms” observed during the evolutionary history of sloths (i.e., caniniform, incisiform, peg-like,
entirely absent; see S3) could then originate from a permanent Cf, as in Choloepus. A large caniniform is present
in earliest fossil sloths like Octodontotherium49 or Pseudoglyptodon50, and contrasts with our reconstruction of the
sloth ancestral dental formula (S3) that is ultimately inuenced by the basal rooting position of Bradypus on sloth
phylogeny. Our results illuminate a potentially dierent ancestral dental formula for sloths that challenges the tra-
ditional assumption that the bradypodid tooth row is primitive and megalonychid dental features derived10. Such a
hypothesis also mitigates the potential weight of dental features in phylogenetical and systematic studies, especially
those related to the size and shape of the caniniforms. As a matter of fact, Gaudin10 (p. 275) commented that “the
family [Megalonychidae] is united largely by features associated with the caniniform rst upper and lower teeth”.
In sloths, the diversication in shape of the mesialmost teeth is frequently associated with a variation in ros-
tral length and the presence of a pre and/or post diastema (S3). Such a diastema, which is oen considered as a
toothless gap, could challenge the homology of the teeth between taxa. However, the intercalation of additional
teeth in the diastema, as observed in armadillos13, seems unlikely because of the relative stability of the dental
formula in the sloth fossil record (S3). Our observations are consistent with McAfee’s view on the development
of the diastema in Choloepus12, which he proposed could result from an increase of skull length and migration of
mesial teeth rather than a loss of teeth between the caniniforms (dCf/Cf-dcf) and molariforms (dMf1-dmf1). e
lesser development of the diastema in the youngest stages of Choloepus and the complete absence of vestigial teeth
at its level are also in line with this assertion.
In conclusion, we showed that vestigial teeth are informative in understanding dental homologies, especially
in assessing the deciduous or successional nature of individual teeth. Our developmental data for extant sloths
bear directly on the claim that their lower caniniform teeth are not homologous to canines of other mammals and
that upper caniniforms are not homologous between the two-toed and the three-toed sloths. ese results under-
line that dening dental homologies in extant and extinct sloths is complex and that, where possible, characters
based on dental features should be augmented with developmental data to ensure proper homology assessment.
Development of discrete shapes and functional domains in the tooth row is governed by developmental processes
that are still poorly known in mammals and for which further investigations on non-model mammals, such as
sloths, are timely and topical.
Methods
We sampled material from collections of the Museum für Naturkunde Berlin (ZMB), the Natural History
Museum of London (BMNH), the Muséum National d’Histoire Naturelle in Paris (MNHN), and the Institut
Royal des Sciences Naturelles de Belgique in Brussels (IRSNB). A total of 25 unsexed sloth foetuses were exam-
ined, representing four species of both extant genera: Bradypus tridactylus, Bradypus variegatus, Choloepus didac-
tylus, and Choloepus homanni51. Species identication was based on collection data (especially geographical
origin) and cranial anatomy51,52 and was possible for 17 of our 25 specimens (S2). ey range in size from 70 to
200 mm crown rump length (CRL), measured from the vertex of the skull to the base of the tail. Collections of
such non-model organisms oen include specimens collected decades ago and invariably lack data on individual
age. Assignment to a relative developmental stage was based on the Skull Length (SL) and the number of discrete
ossication centres throughout the skeleton15,53,54.
3-D data acquisition. Skulls were imaged using high-resolution microtomography (μCT) at the Helmholtz
Zentrum (Berlin, Germany), at the Natural History Museum (London, UK), at the AST-RX platform MNHN
(Paris, France), and at VISCOM SARL (Saint Ouen l’Aumône, France). is method allows 3D renderings of ossi-
ed tissues, as well as non-invasive virtual extractions of dental elements. Due to the scan resolution, we could not
test for the putative presence of a small enamel cap at the tips of the forming teeth. ese reconstruction and vis-
ualization were performed using stacks of digital CT images with the AVIZO 7.1 (Visualization Sciences Group)
soware. 3D reconstruction of the specimens were deposited in MorphoMuseum (http://www.morphomuseum.
com/; M3#109 to M3#115) and Morph-D-base (https://www.morphdbase.de/).
Reconstruction of the ancestral dental morphotype (S3). e datamatrix of Gaudin10 was downloaded
from Morphobank and the following characters of interest were selected for study: character n°2: dental formula;
n°6: diastema; n°13: size of Cf ; n°14: size of cf; n°19: morphology of Cf/cf; n°21: position of Cf relative to the ante-
rior edge of the maxilla. e cladogram corresponds to the topology of the strict consensus obtained by Gaudin10:
Fig.1 when all characters were weighted equally. For our analysis of character optimizations, this cladogram was
pruned in order to contain only sloth taxa (Folivora) (i.e., all non-folivoran successive outgroups originally included
in Gaudin’s analysis10 were excluded). Parsimonious reconstruction of the hypothetical ancestral morphotype (S3)
for the selected characters was undertaken on this reduced cladogram using the soware Mesquite 2.7555.
References
1. Vizcaíno, S. F. e teeth of the ‘“ toothless ”’: novelties and ey innovations in the evolution of xenarthrans (Mammalia, Xenarthra).
Paleobiology 35, 343–366 (2009).
2. Parer, W. . On the structure and development of the sull in Mammalia Part II. Edentata. Philos. Trans. . Soc. London 176, 1–119
(1885).
3. McDonald, H. G. In Morphol. Stud. Foss. extat Xenarthra (Fariña, . A., Vizcaíno, S. F. & Storch, G.) 5–17 (Sencenbergiana
biologica, 2003).
4. Naples, V. L. & McAfee, . . Chewing through the Miocene: an examination of the feeding musculature in the ground sloth
Hapalops from South America (Mammalia: Pilosa). F1000esearch 86, 1–22 (2014).
5. Tomes, C. S. A Manual of Dental Anatomy, Human and Comparative. (Presley Blaiston, 1882).
6. Leche, W. Studien über die Entwiclung des Zahnsystems bei den Säugethieren. Morphol. Jahrb. 19, 502–547 (1892).

www.nature.com/scientificreports/
8
Scientific RepoRts | 6:27763 | DOI: 10.1038/srep27763
7. Delsuc, F., Vizcaíno, S. F. & Douzery, E. J. P. Inuence of Tertiary paleoenvironmental changes on the diversication of South
American mammals : a relaxed molecular cloc study within xenarthrans. BMC Evol. Biol. 13, 1–13 (2004).
8. Greenwood, A. D., Castresana, J., Feldmaier-Fuchs, G. & Pääbo, S. A Molecular Phylogeny of Two Extinct Sloths. Mol. Phylogenet.
Evol. 18, 94–103 (2001).
9. Poinar, H., uch, M., Mcdonald, G., Martin, P. & Pääbo, S. Nuclear gene sequences from a Late Pleistocene sloth coprolite. Curr. Biol.
13, 1150–1152 (2003).
10. Gaudin, T. J. Phylogenetic relationships among sloths (Mammalia, Xenarthra, Tardigrada): the craniodental evidence. Zool. J. Linn.
Soc. 140, 255–305 (2004).
11. Delsuc, F., Superina, M., Tila, M., Douzery, E. J. P. & Hassanin, A. Molecular Phylogenetics and Evolution Molecular phylogenetics
unveils the ancient evolutionary origins of the enigmatic fairy armadillos. Mol. Phylogenet. Evol. 62, 673–680 (2012).
12. McAfee, . . Dental anomalies within extant members of the mammalian Order Pilosa. Acta Zool. 96, 301–311 (2014).
13. Martin, B. E. Tooth development in Dasypus novemcinctus. J. Morphol. 27, 647–691 (1916).
14. Hautier, L., Weisbecer, V., Sánchez-Villagra, M. ., Goswami, A. & Asher, . J. Seletal development in sloths and the evolution of
mammalian vertebral patterning. Proc. Natl. Acad. Sci. USA 107, 18903–18908 (2010).
15. Hautier, L. et al. Seletal ossication and sequence heterochrony in xenarthran evolution. Evol. Dev. 13, 460–476 (2011).
16. Böer, H. Beobachtungen und Untersuchungen an Säugetieren (einschließlich südamerianischer Edentaten) während einer
biologisch-anatomischen Forschungsreise nach Brasilien. Morphol Jahrb 70, 1–66 (1932).
17. Gervais, P. emarque au sujet du système dentaire de l’aï. J. Zool. 437–437 (1873).
18. Cartelle, C. & De Iuliis, G. Eremotherium Laurillardi (Lund) (Xenarthra, Megatheriidae), the Panamerican giant ground sloth:
Taxonomic aspects of the ontogeny of sull and dentition. J. Syst. Palaeontol. 4, 199–209 (2006).
19. Bargo, M. S., Vizcaíno, S. F. & ay, . F. Predominance or orthal masticatory movements in the early miocene Eucholaeops
(Mammalia, Xenarthra, Tardigrada, Megalonychidae) and other Megatherioid sloths. J. Vertebr. Paleontol. 29, 870–880 (2009).
20. Pujos, F., Iuliis, G. De. & Quispe, B. M. Hisatherium saintandrei, gen. et sp. nov.: an unusual sloth from the Santacrucian of
Quebrada Honda (Bolivia) and an overview of middle Miocene, small megatherioids. J. Vertebr. Paleontol. 31, 1131–1149 (2011).
21. Brandts, P. Dissertation Inaugural de Tradigradis (Lugduni Batav, 1828).
22. öse, C. Beiträge zur Zahnentwiclung der Edentaten. Anatomischer Anzeiger. 7, 495–512 (1892).
23. Simon, L. Beiträge Anatomie und Entwiclung der Bradypodiden. 68, 239–260 (1902).
24. Naples, V. L. Cranial osteology and function in the tree sloth, Bradypus and Choloepus. Am. Museum Novit. 2739, 1–41 (1982).
25. Peterova, ., Lesot, H. & Petera, M. Phylogenetic memory of developing mammalian dentition. J. Exp. Zool. B. Mol. Dev. Evol.
306, 234–250 (2006).
26. Järvinen, E., Tummers, M. & esle, I. e role of the dental lamina in mammalian tooth replacement. J Exp Zool 312, 281–291
(2009).
27. Järvinen, E., Välimäi, ., Pummila, M., esle, I. & Jernvall, J. e taming of the shrew mil teeth. Evol D ev 10, 476–485 (2008).
28. Lucett, W. P. Ontogenetic staging of the mammalian dentition, and its value for assessment of homology and heterochrony. J.
Mamm. Evol. 1, 269–282 (1993).
29. Lucett, W. P. In Mammal phylogeny Mesozoic Dier. multituberculates, monotremes, early therians, marsupials. (Szalay, F. S.,
Novace, M. J. & Mcenna, M. C.) 182–204 (Springer, 1993).
30. Moss-Salentijn, L. In Dev. Funct. Evol. teeth (Butler, P. M. & Joysey, . A.) 13–29 (Academic Press, 1978).
31. Yamanaa, A., Yasui, ., Sonomura, T., Iwai, H. & Uemura, M. Development of deciduous and permanent dentitions in the upper
jaw of the house shrew (Suncus murinus). Arch Oral Bi ol 55, 279–287 (2010).
32. Verts, B. J. e biology of the striped sun. (University of Illinois Press, 1967).
33. enyon, . e sea otter in the eastern Pacic Ocean. N Am Fauna 68, 1–352 (1969).
34. Ciancio, M. ., Castro, M. C. & Asher, . J. Evolutionary implications of dental eruption in Dasypus (Xenarthra). J. Mamm. Evol. 19,
1–8 (2012).
35. Lucett, W. P. & Maier, W. Development of Deciduous and Permanent Dentition in Tarsius and Its Phylogenetic Signicance. Folia
Primatol. 37, 1–36 (1982).
36. Butler, P. In Dent. Anthropol. (Brothwell, D.) 1–13 (Pergamon Press, 1963).
37. van Nievelt, A. F. H. & Smith, . . To replace or not to replace: the signicance of reduced functional tooth replacement in
marsupial and placental mammals. Paleobiology 31, 324–346 (2005).
38. Luo, Z.-X., ielan-Jaworowsa, Z. & Cifelli, . Evolution of dental replacement in mammals. Carnegie Mus Nat Hist Bull 36,
159–175 (2004).
39. Pascual, . Una nueva superfamilia ‘Entelopsoidea’ descripcion de la nueva especie ‘Entelops parodii’. Acta Geol. Lilloana 3, 127–146
(1960).
40. Hoffstetter, . Les Edentés Xénarthres, un groupe singulier de la faune néotropicale (origine, affinités, radiation adaptative,
migrations et extinctions). In Proc. First Int. Meet. “Palaeontology, Essent. Hist. Geol. 385–443 (1982).
41. Gervais, P. Zoologie et Paléontologie générales: Nouvelles echerches sur les Animaux vertébrés vivants et fossiles Vol. 1 (Libraire de la
Société de Géographie, 1867).
42. Burmeister, H. Atlas de la description physique de la république Argentine. Mammifères (Coni, 1881).
43. Lydeer, . Contribution to a nowledge of the fossil Vertebrates of Argentina 2: e extinct Edentates of Argentina. An. del Mus.
La Plata, Paleontol. Argentina 32, 1–118 (1895).
44. Cifelli, . L. et al. Fossil evidence for the origin of the marsupial pattern of tooth replacement. Nature 379, 715–718 (1996).
45. Patterson, B. & Pascual, . e fossil mammal fauna of South America. Q. ev. Biol. 43, 409–451 (1968).
46. Patterson, B., Seagall, W., Turnbull, W. D. & Gaudin, T. J. e ear region in Xenarthrans (=Edentata, Mammalia). Part II. Pilosa
(slogs, anteaters), palaeanodonts, and a miscellany. Fieldiana, Geol. 24, 1–79 (1992).
47. Gaudin, T. J. e ear region of edentate and the phylogeny of the tardigrada (Mammalia, Xenarthra). J. Vertebr. Paleontol. 15,
672–705 (1995).
48. Charles, C., Solé, F., Gomas odrigues, H. & Viriot, L. Under Pressure? Dental Adaptations To Termitophagy and Vermivory Among
Mammals. Evolution (N. Y.). 67, 1792–1804 (2013).
49. Shocey, B. J. & Anaya, F. Grazing in a New Late Oligocene Mylodontid Sloth and a Mylodontid adiation as a Component of the
Eocene-Oligocene Faunal Turnover and the Early Spread of Grasslands/Savannas in South America. J. Mamm. Evol. 18, 101–115
(2011).
50. Mcenna, M. C., Wyss, A. . & Flynn, J. J. Paleogene Pseudoglyptodont Xenarthrans from Central Chile and Argentine Patagonia.
Am. Museum Novit. 3536, 1–18 (2006).
51. Wetzel, . M. In Evol. Ecol. armadillos, sloths vermilinguas (Montgomery, G.) 5–21 (Smithsonian Institution Press, 1985).
52. Hautier, L., Billet, G., Eastwood, B. & Lane, J. Patterns of Morphological Variation of Extant Sloth Sulls and their Implication for
Future Conservation Eorts. Anat. ec. 297, 979–1008 (2014).
53. Hautier, L. et al. Patterns of ossication in southern versus northern placental mammals. Evolution (N. Y.). 67, 1994–2010 (2013).
54. Hautier, L., Stanseld, F. J., Allen, W. . T. & Asher, . J. Seletal development in the African elephant and ossication timing in
placental mammals. Proc. . Soc. B Biol. Sci. 279, 2188–2195 (2012).
55. Maddison, W. P. & Maddison, D. . (2011). Mesquite: A Modular System for Evolutionary Analysis, Version 2.75. University of
British Columbia, Vancouver, Canada. UL http://mesquiteproject.org.

www.nature.com/scientificreports/
9
Scientific RepoRts | 6:27763 | DOI: 10.1038/srep27763
Acknowledgements
We are grateful to M. Herbin, C. Bens, G. Véron, A. Verguin, F. Renoult, C. Denys and J. Cuisin (Museum National
d’Histoire Naturelle, Paris), Peter Giere and Frieder Mayer (Museum für Naturkunde, Berlin) Paula Jenkins and
Roberto Portela Miguez (Natural History Museum, London), and their colleagues for access to comparative
material. N. Karjilov (Helmholtz Zentrum Berlin), R. Abel (Natural History Museum), M. García-Sánz (AST-RX
platform, Muséum national d’Histoire naturelle, Paris, France), F. Landru, C. Morlier, G. Guillemain and all the
sta from Viscom SARL (St Ouen l’Aumône, France) provided generous help and advice with CT acquisition.
We thank Dennyss Lelaurin and Mélanie Canas-Grosso for their help in the data acquisition. We acknowledge
financial support from the Grant F/09 364/I from the Leverhulme Trust. This work has benefited from an
“Investissements d’Avenir” grant managed by Agence Nationale de la Recherche, France (CEBA, ref. ANR-10-
LABX-25-01). is publication is contribution No. ISEM 2016-081 of the Institut des Sciences de l’Evolution de
Montpellier (UMR 5554 – UM2 + CNRS + IRD).
Author Contributions
L.H. initiated the project; L.H. designed the research plan; L.H. and G.B. collected CT data; L.H. reconstructed
the 3D data; L.H., H.G.R. and G.B. analyzed the data; L.H., H.G.R., G.B. and R.J.A. discussed the results and wrote
the manuscript.
Additional Information
Supplementary information accompanies this paper at http://www.nature.com/srep
Competing nancial interests: e authors declare no competing nancial interests.
How to cite this article: Hautier, L. et al. e hidden teeth of sloths: evolutionary vestiges and the development
of a simplied dentition. Sci. Rep. 6, 27763; doi: 10.1038/srep27763 (2016).
is work is licensed under a Creative Commons Attribution 4.0 International License. e images
or other third party material in this article are included in the article’s Creative Commons license,
unless indicated otherwise in the credit line; if the material is not included under the Creative Commons license,
users will need to obtain permission from the license holder to reproduce the material. To view a copy of this
license, visit http://creativecommons.org/licenses/by/4.0/