ArticlePDF Available

Extensive Diversity and Disparity of the Early Miocene Platanistoids (Cetacea, Odontoceti) in the Southeastern Pacific (Chilcatay Formation, Peru)

Authors:

Abstract and Figures

Several aspects of the fascinating evolutionary history of toothed and baleen whales (Cetacea) are still to be clarified due to the fragmentation and discontinuity (in space and time) of the fossil record. Here we open a window on the past, describing a part of the extraordinary cetacean fossil assemblage deposited in a restricted interval of time (19–18 Ma) in the Chilcatay Formation (Peru). All the fossils here examined belong to the Platanistoidea clade as here redefined, a toothed whale group nowadays represented only by the Asian river dolphin Platanista gangetica. Two new genera and species, the hyper-longirostrine Ensidelphis riveroi and the squalodelphinid Furcacetus flexirostrum, are described together with new material referred to the squalodelphinid Notocetus vanbenedeni and fragmentary remains showing affinities with the platanistid Araeodelphis. Our cladistic analysis defines the new clade Platanidelphidi, sister-group to Allodelphinidae and including E. riveroi and the clade Squalodelphinidae + Platanistidae. The fossils here examined further confirm the high diversity and disparity of platanistoids during the early Miocene. Finally, morphofunctional considerations on the entire platanistoid assemblage of the Chilcatay Formation suggest a high trophic partitioning of this peculiar cetacean paleocommunity.
Content may be subject to copyright.
Life 2020, 10, 27; doi:10.3390/life10030027 www.mdpi.com/journal/life
Article
Extensive Diversity and Disparity of the Early
Miocene Platanistoids (Cetacea, Odontoceti) in the
Southeastern Pacific (Chilcatay Formation, Peru)
Giovanni Bianucci
1,
*, Christian de Muizon
2
, Mario Urbina
3
and Olivier Lambert
4
1
Dipartimento di Scienze della Terra, Università di Pisa, Pisa 56126, Italy
2
CR2P (CNRS, MNHN, SU), Muséum National d’Histoire Naturelle, Département Origines et Évolution,
Paris 75005, France; muizon@mnhn.fr
3
Departamento de Paleontología de Vertebrados, Museo de Historia Natural de la Universidad Nacional
Mayor de San Marcos, Lima 15072, Peru; mariourbina01@hotmail.com
4
Institut Royal des Sciences Naturelles de Belgique, D.O. Terre et Histoire de la Vie, 1000 Brussels, Belgium;
olivier.lambert@naturalsciences.be
* Correspondence: giovanni.bianucci@unipi.it
Received: 14 February 2020; Accepted: 16 March 2020; Published: 18 March 2020
Abstract: Several aspects of the fascinating evolutionary history of toothed and baleen whales
(Cetacea) are still to be clarified due to the fragmentation and discontinuity (in space and time) of
the fossil record. Here we open a window on the past, describing a part of the extraordinary cetacean
fossil assemblage deposited in a restricted interval of time (19–18 Ma) in the Chilcatay Formation
(Peru). All the fossils here examined belong to the Platanistoidea clade as here redefined, a toothed
whale group nowadays represented only by the Asian river dolphin Platanista gangetica. Two new
genera and species, the hyper-longirostrine Ensidelphis riveroi and the squalodelphinid Furcacetus
flexirostrum, are described together with new material referred to the squalodelphinid Notocetus
vanbenedeni and fragmentary remains showing affinities with the platanistid Araeodelphis. Our
cladistic analysis defines the new clade Platanidelphidi, sister-group to Allodelphinidae and
including E. riveroi and the clade Squalodelphinidae + Platanistidae. The fossils here examined
further confirm the high diversity and disparity of platanistoids during the early Miocene. Finally,
morphofunctional considerations on the entire platanistoid assemblage of the Chilcatay Formation
suggest a high trophic partitioning of this peculiar cetacean paleocommunity.
Keywords: Odontoceti; Squalodelphinidae; Platanistidae; early Miocene; Peru; phylogeny;
paleoecology
1. Introduction
The evolutionary history of cetaceans is overall increasingly well documented by a
representative fossil record scattered in various parts of the world [1–3]. This record describes in
detail: (1) The progressive adaptation of ancient cetaceans, named archaeocetes, to life in the sea [4–
6]; (2) the origin of mysticetes and their later evolution characterized by the replacement of teeth with
baleen [7–10] and the tendency for extreme gigantism [11,12]; and (3) the great radiation of
odontocetes that over time explored a large number of feeding strategies and ecological niches,
thanks to their ability to echolocate [13–15] and to their marked cranial plasticity [16,17]. In spite of
this general picture, our knowledge on this highly successful clade of marine mammals is still far
from exhaustive. In recent years, new taxa have been continuously described, highlighting the fact
that we do not yet possess a solid dataset of past cetacean diversity. Furthermore, two weak points
Life 2020, 10, 27 2 of 62
in the framework of the evolutionary history of cetaceans are (1) the discontinuity of the fossil record,
from both a temporal and a geographical point of view; and (2) the low geochronological resolution
featuring many fossils or fossil assemblages. These critical issues should be taken into consideration
when attempting to reconstruct in detail the ecological structure of ancient cetacean
paleocommunities, to analyze with a statistically significant approach some evolutionary trends, and
to tentatively correlate these to the main abiotic and biotic changes observed at a global scale [7,18,19].
In this context, this paper focuses on a part of the fossil cetacean assemblage coming from the
extraordinary Cenozoic marine vertebrate Lagerstätte of the East Pisco Basin (Peru). The fossils here
examined were collected in the lower Miocene layers of the Chilcatay Formation (Chilcatay Fm)
exposed in the vertebrate-bearing fossil localities of Ullujaya and Zamaca, western Ica Valley, Ica
Region. More than 180 partial skeletons of cetaceans, together with remains of other vertebrates, have
been discovered in these two localities and most of these fossils are marked on two published
geological maps [20–22]. Furthermore, all the fossils reported in the maps have been included in
detailed stratigraphic columns accompanied by a precisely defined geochronological and
biostratigraphic framework (ca 19–18 Ma) for the deposition of the entire sequence of fossil-bearing
marine sediments (Figure 1).
More specifically, this paper focuses on the fossils of the Ullujaya-Zamaca assemblage belonging
to the superfamily Platanistoidea, an odontocete clade that underwent a major radiation during the
early Miocene and which today is only represented by the South Asian river dolphin (Platanista
gangetica) confined to the freshwaters of the Indus and Ganges river systems [23]. Following other
works already published by us on the platanistoids from the Chilcatay Formation [24–26], the fossils
here described further support the great diversity and morphological disparity of this clade.
2. Materials and Methods
2.1. Institutional Abbreviations
GAS, Georgian Academy of Sciences, Tbilisi, Georgia; GMNH, Gunma Museum of Natural
History, Tomioka, Japan; IRSNB, Institut Royal des Sciences Naturelles de Belgique, Brussels,
Belgium; LACM, Natural History Museum of Los Angeles County, Los Angeles, U.S.A.; LDUCZ,
Grant Museum of Zoology, University College London, United Kingdom; MGP-PD, Museo di
Geologia e Paleontologia, Università di Padova; MNHN, Muséum National d’Histoire Naturelle,
Paris, France; MLP, Museo de Ciencias Naturales de La Plata, Buenos Aires, Argentina; MSNUP,
Museo di Storia Naturale, Università di Pisa, Italy; MUSM, Museo de Historia Natural, Universidad
Nacional Mayor de San Marco, Lima, Peru; MZUF, Museo di Storia Naturale, Zoological collection,
Università degli Studi di Firenze, Italy; NMG, National Museum of Georgia, Tbilisi, Georgia; NMV,
Museum Victoria Palaeontology Collections, Melbourne, Australia; OU, Geological Museum,
University of Otago, Dunedin, New Zealand; SBCM, San Bernardino County Museum, Redlands,
U.S.A.; UCMP, University of California Museum of Paleontology, Berkeley, U.S.A.; USNM, National
Museum of Natural History, Smithsonian Institution, Washington, D.C., U.S.A.
2.2. Anatomical Abbreviation
BZW, bizygomatic width of the skull; CBL, condylobasal length of the skull; TBL, total body
length.
2.3. Collection and Preparation
The platanistoid specimens described here where discovered during several field expeditions
from 2010 to 2019 that involved all the authors of this paper. The fossils were excavated by one of the
authors (M.U.) and by W. Aguirre and subsequently transported to the MUSM for preparation and
storage. The preparation and consolidation of these fossils was made by W. Aguirre using mechanical
tools and standard fossil vertebrate preparation techniques.
Life 2020, 10, 27 3 of 62
2.4. Anatomical Terminology
The anatomical terminology follows Mead and Fordyce [27] for the skull and mostly Evans and
de Lahunta [28] for the postcranial skeleton.
2.5. Cladistic Analysis
The phylogenetic analysis was performed using a modified version of the matrix published by
Bianucci et al. [26] (Table B1 in the Appendix B). The new genera Ensidelphis and Furcacetus and seven
new characters (characters 42–48 in the Appendix A) were added, whereas the fragmentary USNM
475496 specimen and one controversial character (character 42 in [26]) were removed from the matrix.
Some character states were coded differently from a previous version of the matrix due to the
discovery of new material (e.g., for Notocetus) or a better preparation of previously described fossils
(e.g., MUSM 603). The final matrix includes 24 taxa coded for 48 morphological characters.
The parsimony analysis was executed with the software PAUP (version 4.0b10; [29]),
considering all characters unordered and unweighted, and using the tree bisection and reconnection
(TBR) algorithm with ACCTRAN optimization.
3. Geological, Stratigraphical, and Paleontological Setting
The marine sediments constituting the Chilcatay Fm as exposed in the vertebrate-bearing fossil
localities of Ullujaya and Zamaca were deposited in the onshore portion of the East Pisco Basin, which
extends over some 30 km across strike between the present-day Coastal Cordillera to the west and
the Western Cordillera to the east, and about 200 km from Pisco to Nazca (Figure 1) [30–32]. The
complete Eocene–Pliocene succession filling this basin overlies a pre-Cenozoic crystalline basement
consisting of a variety of Precambrian metamorphic rocks, known as the Arequipa Massif [33,34],
intruded by a complex assemblage of lower Paleozoic gabbroic to granitoid rocks forming the San
Nicolás Batholith [35] that, in turn, is unconformably overlain by Jurassic volcano-sedimentary rocks
[32]. From its base upward the Cenozoic fill of the basin comprises the Caballas, Paracas, Otuma,
Chilcatay, and Pisco formations [21,22,31,36–39].
Life 2020, 10, 27 4 of 62
Figure 1. Geographical position of Ullujaya and Zamaca (Pisco Basin, southern coast of Peru) (a, b)
and related composite stratigraphic sections (c, d) showing the distribution of fossil platanistoids in
the Chilcatay Fm, including the specimens with known stratigraphical position described in this
paper. Red silhouette indicates holotype. Absolute datings (
40
Ar/
39
Ar on ash layers) constraining the
age of the fossil platanistoids are also reported along the sections.
On the whole the Eocene–Pliocene sedimentary succession of the East Pisco Basin represents one
of the most significant marine vertebrate Lagerstätte of the Cenozoic Era due to the exceptional
preservation and the elevated concentration of fossils [21,22,40–46] referred to cetaceans
(archaeocetes [47–49]; odontocetes [24,25,50–62]; mysticetes [10,63–68]), pinnipeds [69,70], marine
birds [71–74], marine turtles [75], marine sloths [76–81], and sharks and rays [82–88].
Life 2020, 10, 27 5 of 62
By using an allostratigraphic approach, Di Celma et al. [21,22] subdivided the Chilcatay strata
exposed in the vertebrate-bearing fossil localities of Ullujaya and Zamaca into two distinctive
sediment wedges, informally designated Ct1 and Ct2 in ascending stratigraphic order, separated by
a major intraformational unconformity (Figure 1). In the Zamaca area the base of Ct1 rests with an
angular unconformity on the Otuma Formation. This basal unconformity does not occur in the
Ullujaya area, where the lowermost portion of the Chilcatay Fm is not exposed. The Ct1 allomember
comprises three facies associations indicative of shoreface (Ct1c), offshore (Ct1a), and subaqueous
delta (Ct1b) depositional settings. The Ct1c association, only exposed at the Zamaca locality, is 10.5
m-thick and consists of massive or weakly bedded sandstones with scattered boulders alternating
with pebble- to boulder-sized conglomerate beds with abundant shelly calcarenite matrix. The Ct1a
association is 35 m and 31 m-thick at Ullujaya and Zamaca, respectively, and consists of silty to sandy
mudstones interbedded with occasional very fine- to fine-grained sandstone beds, as well as a few
volcanic ash layers. The Ct1b association is comprised of mixed siliciclastic-carbonate oblique and
sigmoidal clinoforms downlapping onto the underlying sediments of Ct1a; Ct1b thickness decreases
basinward from about 20 m at Ullujaya to a zero-edge in the central part of the Zamaca area. The Ct2
allomember comprises two facies associations recording shoreface (Ct2a) and offshore (Ct2b) marine
depositional settings. The Ct2a association is about 3.5 m-thick at both Ullujaya and Zamaca and is
composed of medium- to very coarse-grained sandstones containing sub-rounded to sub-angular
pebbles. The Ct2b association is 7 m and more than 15 m-thick at Ullujaya and Zamaca, respectively,
and consists of a heterolithic succession of weakly bioturbated, thinly-bedded silty mudstone
intercalated with minor, laterally persistent, very fine-grained sandstone interbeds and occasional
volcanic ash layers [21,22].
The entire stratigraphical succession of the Chilcatay Fm exposed at the Zamaca and Ullujaya
localities has been roughly constricted through radiometric dating of ash layers to an interval
between 19.25 and 18.02 Ma (late early Miocene, Burdigalian), considering that a volcanic ash layer
sampled at Zamaca, 4 m above the contact between the Chilcatay Fm and the underlying Otuma
Formation, gave an
40
Ar/
39
Ar age of 19.25 ± 0.05 Ma, and that a volcanic ash layer sampled at Ullujaya,
just 1 m below the contact between the Chilcatay and overlying Pisco Formation, provide an age of
18.02 ± 0.07 Ma [89]. Moreover, four samples for
87
Sr/
86
Sr stratigraphy were collected along the Ct1a
sequence at Zamaca and Ullujaya and gave ages for the whole stratigraphical sequence comprised
between 18.85 and 18.00 Ma [90]. These
40
Ar/
39
Ar and
87
Sr/
86
Sr geochronological ages are consistent with
biostratigraphic results obtained with silicoflagellates and diatoms, both further constraining the
deposition of the Chilcatay Fm in the Ullujaya-Zamaca area between 19 and 18 Ma [21,60]. Another
volcanic ash sample, collected from the basal portion of the Ct1a facies association exposed at Ullujaya,
gave a
40
Ar/
39
Ar ages of 19.00 ± 0.28; consequently, the age of the underlying Ct1c facies association
exposed at Zamaca can be further constricted between 19.25 ± 0.08 Ma and 19.00 ± 0.28 Ma.
The vertebrate fossil assemblage of the Chilcatay Fm exposed at Ullujaya and Zamaca is
dominated by cetaceans (mostly odontocetes) and elasmobranches (mostly lamniformes and
carcharhiniformes); large bony fishes and sea turtles were also recorded [20–22,86]. Besides the
platanistoid remains here described, the odontocete assemblage includes already published material
belonging to the squalodelphinids Huaridelphis raimondii [24] and Notocetus vanbenedeni [25], the
longirostrine homodont Chilcacetus cavirhinus [57], the heterodont Inticetus vertzi [60], and
undescribed eurhinodelphinids, kentriodontids, and physeteroids [20–22].
4. Systematic Paleontology
Cetacea Brisson, 1762
Neoceti Fordyce and Muizon, 2001
Odontoceti Flower, 1867
Platanistoidea Gray, 1863
Life 2020, 10, 27 6 of 62
Remarks on the superfamily Platanistoidea and its content. In its first definition proposed by
Simpson [91] the superfamily Platanistoidea included Platanista, all other extant "river dolphins"
(Lipotes, Inia, and Pontoporia, the latter being listed there as Stenodelphis), and their closest fossil
relatives. Currently, the only extant genus recognized as belonging to this clade is Platanista, whereas
the other "river dolphins" are placed within the Delphinida clade [92–97]. However, in the past
decades a number of fossil taxa have been included in the Platanistoidea, radically changing the
concept of this superfamily. Besides the Platanistidae, Muizon [92] first included in the Platanistoidea
the extinct families Squalodelphinidae and Squalodontidae, and, a few years later [98], also the
Dalpiazinidae and Prosqualodon. Later, Fordyce [99] added the Waipatiidae and Barnes [100] the
Allodelphinidae. This broad concept of the Platanistoidea has been questioned in several recent
phylogenies. For example, the heterodont Squalodontidae and Prosqualodon were recovered in a more
basal position in several analyses [26,95,101,102], whereas in other analyses the position of these taxa
inside or outside Platanistoidea depends on the settings of the phylogenetic analyses (e.g.,
homoplastic characters being down weighted or not [103–107]). The family Waipatidae was also
removed from Platanistoidea in some phylogenies (e.g., [26,95,101,102]), whereas the poorly known
Dalpiazinidae were never included in a software-assisted phylogenetic analysis. By contrast, the
families Allodelphinidae and Squalodelphinidae appear as two distinct clades closely related to the
Platanistidae in part of the recent phylogenies (e.g., [24,26,101,108]), with allodelphinids being
recovered in the basalmost position, sister group of the clade formed by the platanistids and
squalodelphinids. In several other recent papers (e.g., [103–107]) allodelphinids were not included in
the phylogenetic analyses and squalodelphinids were paraphyletic.
Since the phylogenetic analysis presented below confirms again the close relationships between
the Platanistidae, Squalodelphinidae, and Allodelphinidae, a new definition of the Platanistoidea
sensu stricto is proposed below, only including the above mentioned three families and thus
excluding the Dalpiazinidae, Squalodontidae, Waipatidae, and Prosqualodon.
Proposing such a less inclusive, but more stable definition of this superfamily, we do not a priori
exclude that in future analyses the Platanistoidea clade falls near to one or more of the above excluded
families. In fact, the aim of this restrictive choice is essentially to put order in the controversial
systematics of the large odontocete group including extinct platanistoid-like taxa.
Finally, the New Clade Name (NCN) Platanidelphidi is below defined following the rules
reported in the International Code of Phylogenetic Nomenclature (PhyloCode [109]). The
Platanidelphidi clade includes the new genus Ensidelphis described below and the platanistoid
MUSM 603 previously referred to aff. Huaridelphis raimondii [24], together with the clade formed by
Platanistidae + Squalodelphinidae, the latter being repeatedly recovered in morphological
phylogenies since the first analyses by Muizon [92,98].
Emended diagnosis of Platanistoidea. The members of the Platanistoidea are nearly homodont
odontocetes having single-rooted teeth and sharing the following characters: (1) Vertex distinctly
shifted to the left compared to the sagittal plane of the skull (absent in Allodelphis and Ninjadelphis);
(2) long hamular fossa of the pterygoid sinus extending anteriorly on the palatal surface of the
rostrum (also present in Ziphiidae); (3) presence of an articular rim on the lateral surface of the
periotic; (4) elongated anterior spine on the tympanic bulla, associated with a marked anterolateral
convexity; (5) great reduction of the coracoid process of the scapula and the acromion located on the
anterior edge of the scapula (also present in Squalodon and Prosqualodon); (6) neurocranium distinctly
shorter than wide, with ratio between neurocranium length (longitudinal, from occipital condyles to
level of antorbital notches) and postorbital width < 0.90 (also present in Eurhinodelphinidae); and (7)
anterior portion of the zygomatic process of the squamosal in contact with the postorbital process of
the frontal (also present in Eurhinodelphinidae).
Platanidelphidi (NCN)
Definition. The branch-based Platanidelphidi consists of the extant Platanista and all species that
share a more recent common ancestor with Platanista than with Allodelphis.
Life 2020, 10, 27 7 of 62
Etymology. From Platanista, the type genus of the Platanistidae; and from delphinus, dolphin in
Latin. The name is also a combination of Platanistidae and Squalodelphinidae, the two families
currently included in the Platanidelphidi.
Diagnosis. The members of the Platanidelphidi clade are platanistoids sharing the following
synapomorphies: (1) Asymmetry of the premaxillae on the rostrum at some distance anterior to the
premaxillary foramina, with the right premaxilla being distinctly narrower than the left in dorsal
view; (2) posterior dorsal infraorbital foramen(ina) along the vertex more medial than the lateralmost
margin of the premaxilla on the cranium; (3) deep fossa in the frontal on orbit roof, at the level of the
frontal groove (presumably for orbital lobe of pterygoid sinus); (4) palatines not contacting each other
on the sagittal plane and displaced dorsolaterally; (5) thick zygomatic process of the squamosal (ratio
between the maximum distance from the anteroventral margin of the zygomatic process to the
posterodorsal margin, in lateral view, and the vertical distance from the lower margin of the occipital
condyles to the vertex of the skull > 0.35); (6) dorsal outline of the zygomatic process of the squamosal
in lateral view being dorsally convex (also present in Squalodon and in some specimens of the
eurhinodelphinid Xiphiacetus); and (7) ventral edge of the zygomatic process of squamosal in lateral
view being almost straight or convex.
Platanidelphidi incertae sedis
Ensidelphis, gen. nov.
LSID: zoobank.org:act: C0D8D0CE-1769-4BA1-85D9-054889976B18
Type and only known species. Ensidelphis riveroi, sp. nov.
Diagnosis. As for the type and only referred species.
Etymology. From ensis’, Latin name of a Roman sword similar to the gladius but longer and
narrower; for the very elongated and narrow rostrum; and from ‘delphis’, dolphin in Latin. Gender
masculine.
Ensidelphis riveroi, sp. nov.
Figures 2–9, Tables 1–2
LSID: zoobank.org:act: act:1DEC0DD9-FEFA-4CC8-A668-11934B4256AA
Holotype and only referred specimen. MUSM 3898 consists of an almost complete and well-
preserved cranium (only small portions of both antorbital processes and both the jugals are missing)
with articulated, complete mandibles. Both tympanic bullae are exposed on the ventral surface,
probably hiding the in situ periotics and accessory ossicles. Only six incomplete teeth are visible,
embedded in their alveoli on the right mandible. The atlas, axis, and two additional cervical vertebrae
from the same animal are also preserved.
Type locality. Zamaca locality, Western Ica Valley, Ica Region, Southern Peru (Figure 1a,b).
Geographic coordinates: 14°37'1.65" S, 75°37'31.25" W; 307 m above sea level. This specimen was
reported in the Zamaca fossil map of Di Celma et al. [22] with the field number ZM 97 and
provisionally referred to “aff. Platanistidae indet.”
Type horizon. The holotype of Ensidelphis riveroi MUSM 3898 was collected in the Chilcatay Fm
exposed at Zamaca locality, and more precisely at 10.7 m above the contact with the underlying
Otuma Formation, near the top of the Ct1c facies association of the Ct1 allomember [21,22] (Figure
1d). The age of the Ct1c facies association is constricted between 19.25 ± 0.08 Ma and 19.00 ± 0.28 Ma
(early Burdigalian) on the basis of two volcanic ash layer samples dated by
40
Ar/
39
Ar [89].
Diagnosis. Ensidelphis is a small odontocete having a narrow and very elongated rostrum (about
80% of the CBL) bearing about 64 small single-rooted teeth in each quadrant. It differs from all other
odontocetes in having a protuberance (‘temporal swelling’, new term) on the lateral surface of the
temporal fossa dorsal to the squamosal-parietal suture. It is referred to the Platanistoidea as defined
above by having: Elongated hamular fossa of the pterygoid sinus extending anteriorly on the palatal
surface of the rostrum; neurocranium shorter than wide (ratio < 0.90); and elongated anterior spine
Life 2020, 10, 27 8 of 62
on the tympanic bulla associated with a marked anterolateral convexity. It differs from all other
Platanistoidea in having a lesser posterior extension of both right and left ascending processes of the
premaxillae, ending at the contact with the anterolateral angles of the nasals, and in having an even
more elongated anterior spine of the tympanic bulla. It differs from all other Platanistoidea, with the
exception of Platanista, for the lesser minimal distance between the temporal crests (see quantification
below; character possibly related to the temporal swelling mentioned above). It differs from the
similarly hyper-longirostrine Allodelphinidae in having the dorsal opening of the mesorostral groove
anterior to the rostrum base narrower than the premaxilla, the lateral rostral suture between
premaxilla and maxilla not deeply grooved, proportionally wider premaxillae at rostrum base (>60%
of the width of the rostrum), and vertex not strongly transversely pinched. Ensidelphis belongs to the
Platanidelphidi clade in having: Asymmetry of the premaxillae on the rostrum at some distance
anterior to the premaxillary foramina, with the right premaxilla distinctly narrower than the left in
dorsal view; posterior dorsal infraorbital foramen located along the vertex, more medial than the
lateralmost margin of the premaxilla in the cranium; vertex distinctly shifted to the left compared to
the sagittal plane of the skull; thick zygomatic process of the squamosal (ratio between the maximum
distance from the anteroventral margin of the zygomatic process to the posterodorsal margin, in
lateral view, and the vertical distance from the lower margin of the occipital condyles to the vertex of
the skull > 0.35); and ventral edge of zygomatic process of squamosal in lateral view almost straight.
Within the Platanidelphidi, Ensidelphis differs from Platanista, Pomatodelphis, and Zarhachis in the
lateral rostral suture between premaxilla and maxilla being not deeply grooved; from Huaridelphis,
Notocetus, and Squalodelphis in the presence of a deep lateral groove on the mandible; from
Phocageneus, Notocetus, and Squalodelphis in the ventral groove of the tympanic not affecting the whole
length of the bone, including the anterior spine; from Araeodelphis, Dilophodelphis, Furcacetus,
Huaridelphis, Platanista, and Squalodelphis in the very elongated rostrum (>70% of the CBL); from
Furcacetus, Huaridelphis, Macrosqualodelphis, and Notocetus in the smaller size of teeth at rostrum mid-
length (diameter <2% of BZW); from Platanista, Pomatodelphis, and Zarhachis in the less elongated
mandibular symphysis (61% of the total mandibular length contra >65%) and a consequent smaller
angle between the mandibular rami (25° contra roughly 60°).
Etymology. riveroi, honoring Mariano Eduardo de Rivero y Ustariz (1798–1857), prominent
Peruvian geologist and archaeologist.
Description and Comparison
Ontogeny. We consider the holotype of Ensidelphis riveroi as an adult animal, having: Nasals and
frontals fused together at the vertex, fusion of the premaxillae and maxillae at the anterior end of the
rostrum, well individualized upper and lower alveoli, and all epiphyses of preserved post-atlas
cervical vertebrae completely fused.
Total body length estimate. The TBL of Ensidelphis was estimated to 1.95 m, using a BZW value
of 196 mm in the equation proposed by Pyenson and Sponberg [110] for the stem platanistoids.
However, considering the extreme elongation of its rostrum, we suspect that the TBL of Ensidelphis
was greater than the one calculated with this equation. Therefore, we tried another way to get a better
estimate using the extant Platanista gangetica and the Miocene Zarhachis flagellator for comparison. We
chose Zarhachis because it is a relative of Ensidelphis having the same hyper-longirostrine cranium
(see [111,112]; ratio between BZW and CBL equals 0.23 and about 0.22 for the holotype of Ensidelphis
riveroi and Z. flagellator USNM 10485, respectively), and Platanista because it is the closest extant
relative of Ensidelphis.
Life 2020, 10, 27 9 of 62
Table 1. Measurements on the skulls of Ensidelphis riveroi MUSM 3898 (holotype) and Platanidelphidi
indet. MUSM 3897 from the Chilcatay Fm (early Miocene, Peru). All measurements are in mm.
Dimension
Ensidelphis riveroi
MUSM 3898
Platanidelphidi
indet. MUSM 3897
Condylobasal length 865 +430
Length of rostrum
697
-
Width of rostrum its rostrum
e111
103
Width of premaxillae at base of rostrum
70
65
Bizygomatic width of skull
196
183
Width of maxillae at midlength of rostrum
33
-
Width of premaxillae at midlength of rostrum
22
-
Maximum width between temporal crests
120
118
Minimum posterior distance between temporal crests
90
98
Length of right orbit
31
-
Height of right temporal fossa
81
-
Length of right temporal fossa
72
-
Length of zygomatic process of squamosal from
posglenoid process to tip of zygomatic process
68 -
Maximum width of premaxillae on neurocranium
89
82
Width of right premaxillary sac fossa
37
31
Width of left premaxillary sac fossa
37
31
Width of bony nares
31
34
Minimum posterior distance between maxillae
e42
-
Distance from foramen magnum to nuchal crest
81
-
Width between lateral margins of occipital condyles
80
79
Height of right occipital condyle
41
44
Width of foramen magnum
-
28
Height of foramen magnum
-
22
Maximum width between the exoccipitals
169
149
Length of alveoli at midlength of rostrum
4.0
-
Transverse width of alveoli at midlength of rostrum
3.4
-
Number of teeth per upper tooth row e64 +21
Total length of mandibles
775
-
Length of symphyseal portion
476
-
Width of mandibles at posterior end of symphysis
46
-
Height of mandibles at posterior end of symphysis
27
-
+, incomplete; - missing data; e, estimate.
The length of the subcomplete thoracic portion (10 vertebrae) of Z. flagellator USNM 10485 is
about 64.05 cm (measurement after Kellogg [111]; note that the last thoracic of USNM 10485 lacks the
centrum, so we estimated its length as a mean value between the ninth thoracic and the first lumbar).
The vertebral column of three measured skeletons of Platanista gangetica (LDUCZ Z2282, MHNP
A7945, MSNUP M272) is between 4.12 and 4.76 times the length of its thoracic portion. Using the
same proportions, the length of the postcranial skeleton of Z. flagellator can be estimated between 264
and 305 cm and, adding the skull length (119.5 cm according to Kellogg [111]), we obtain a skeletal
length between 383 to 424 cm. Based on these estimations, the skeletal length of Z. flagellator is
between 3.21 to 3.55 times the length of its skull. Considering that the skull length of Ensidelphis is
865 cm and applying the same proportions as for Zarhachis, we obtain an estimate of the skeletal
length of Ensidelphis between 277 (= 865 × 3.21) and 307 cm (865 × 3.55). However, the actual body
length is slightly greater than the skeletal length due to soft tissues, including intervertebral disks.
Consequently a few more centimeters should be added. Therefore, the body length of Ensidelphis
could have been around 3 m, a length significantly larger that the estimation obtained with the
Pyenson and Sponberg [110] equation.
Cranium
Life 2020, 10, 27 10 of 62
General morphology. The most conspicuous character of the cranium of Ensidelphis is the
extreme elongation of its rostrum (81% of the CBL) (Figure 2a,b; Table 1). Among odontocetes, such
an elongated rostrum, a state defined as hyper-longirostry [17], has only been observed in all the
allodelphinids (Allodelphis, Goedertius, Ninjadelphis, and Zarhinocetus), pomatodelphinine platanistids
(Pomatodelphis and Zarhachis), eurhinodelphinids (i.e., Eurhinodelphis, Schizodelphis, Xiphiacetus, and
probably Ziphiodelphis), the eoplatanistid Eoplatanista, and the lipotid Parapontoporia. Compared with
the other hyper-longirostral platanistoids, the rostrum of Ensidelphis displays a dorsoventral
compression intermediate between the less compressed rostra of allodelphinids and the more
compressed rostra in pomatodelphinids. As in Pomatodelphis and Zarhachis, the dorsoventral
compression in Ensidelphis is more pronounced near the apex of the rostrum. The rostrum of
Ensidelphis clearly differs from the rostrum of the extant Platanista by not displaying a marked
transverse compression, a character also observed in a fragmentary early Miocene fossil rostrum
referred to Platanistinae [100]. Moreover, the rostrum of the E. riveroi holotype is markedly bent
towards the right, a possibly natural condition also observed in some adult females of extant
Platanista gangetica [113] and in some specimens of Pontoporia blanivillei [114] (see below for a possible
interpretation of this peculiar asymmetry).
As in all platanistoids and in eurhinodelphinids the neurocranium of Ensidelphis is
anteroposteriorly shortened, its length being about 90% of the postorbital width.
As in all Platanidelphidi and the allodelphinid Zarhinocetus, the vertex and the bony nares of
Ensidelphis are shifted on the left side and the main transverse axis of the nasals is obliquely oriented,
even if this feature is less marked than in some other platanistoids (e.g., Notocetus and Huaridelphis).
As a consequence of this shifting, the left frontal is markedly anteroposteriorly shorter than the right
on the vertex.
The vertex is low, flat, and weakly sloping anteroventrally; moreover, in lateral view it does not
form a pointed crest as observed in several other platanistoids. The transverse compression of the
vertex in Ensidelphis is lesser than observed in all Platanidelphidi with the exception of Huaridelphis.
In fact, in Ensidelphis and Huaridelphis the minimum transverse width of the vertex is only slightly
greater than the transverse width of bony nares, whereas in all other members of the Platanidelphidi
the minimum transverse width of the vertex is the same or slightly lower than the width of bony
nares. An exception is represented by Platanista, displaying a strongly transversely pinched vertex,
as observed in the allodelphinids Allodelphis, Goedertius, Ninjadelphis, and Zarhinocetus, but not in
Arktocara, a putative allodelphinid lacking transverse compression of the vertex [115].
The temporal fossa displays a remarkable height (ratio between vertical height of the fossa, in
lateral view, and vertical distance from the lower margin of the occipital condyles to the vertex of the
skull estimated to 0.70) (Figures 4 and 5). Among platanistoids, a similar height of the temporal fossa
(ratio > 0.60) is only observed in Furcacetus, Macrosqualodelphis, Notocetus, and Platanista. Nevertheless,
the temporal fossa of Ensidelphis differs from the temporal fossa of the aforementioned platanistoids
in being anteroposteriorly compressed (height > anteroposterior length).
Life 2020, 10, 27 11 of 62
Figure 2. Skull in dorsal view of the holotype (MUSM 3898) of Ensidelphis riveroi from the lower Miocene
Chilcatay Fm (Zamaca, Pisco Basin, Peru). (a,b), complete skull; (c,d), detail of the neurocranium. Linear
hatching indicates major breaks, dark shading areas covered by sediment or dental alveoli, and beige
shading reconstructed missing parts. In (b) and (d) the mandibles are shown in blue.
A clear autapomorphy of Ensidelphis is the presence of a peculiar protuberance, here named
temporal swelling, on the medial wall of the temporal fossa, just above the squamosal–parietal suture.
Clearly visible in lateral (Figures 4c,d, 5c,d) and posterior (Figure 6c,f) views of the cranium, this
swelling is interpreted here as an original character, neither due to a pathology, or a trauma, or even
post-mortem deformation, since it is present with the same shape and exactly in the same position in
both the right and the left fossae. Further supporting the non-artificial nature of this character is the
low minimum transverse distance between the temporal crests at the very same level of these
swellings (best observed in posterior view of the cranium). The ratio between this distance and BZW
is 0.55 for Ensidelphis; among other platanistoids, only Platanista has a lower value. It is possible that
the strong transverse posterior compression of the skull has been compensated (possibly for
biomechanical reasons, in relation to the origin of the temporal muscles) by the appearance of the
temporal swellings.
Life 2020, 10, 27 12 of 62
Figure 3. Skull in ventral view of the holotype (MUSM 3898) of Ensidelphis riveroi from the lower Miocene
Chilcatay Fm (Zamaca, Pisco Basin, Peru). (a,b), complete skull; (c,d), detail of the neurocranium. Linear
hatching indicates major breaks, dark shading areas covered by sediment or dental alveoli, and beige
shading reconstructed missing parts. In (b) and (d) the mandibles are shown in blue.
Premaxilla. The apex of the rostrum is formed by the premaxillae only, as clearly showed in
ventral view by the oblique maxillary/premaxillary suture that runs anterolaterally, reaching the
lateral margin of the rostrum about 60 mm from the end (Figure 3a,b). Among other platanistoids,
the anterior portion of the rostrum being formed by the premaxillae only is also observed in
Dilophodelphis, Furcacetus, Huaridelphis, Notocetus, and Squalodelphis, whereas in all the allodelphinids
and platanistids both premaxillae and maxillae are proposed to reach the apex of the rostrum
[108,116]. Outside platanistoids, eurhinodelphinids display an even longer anterior premaxillary part
of the rostrum (e.g., [117]).
The premaxillae are joined together dorsomedially, closing the mesorostral groove from the apex
of the rostrum to 150 mm anterior to the base of the rostrum (Figure 2a,b). However, the two
premaxillae remain distinct from one another by a narrow but clear medial sulcus. Roughly at the
level of the right antorbital notch, the opening of the mesorostral groove reaches its maximum
transverse width (11 mm). Among platanistoids a wider dorsal opening of the mesorostral groove
near the rostrum base is observed in Medocinia, Squalodelphis, and in all the allodelphinids.
In dorsal view each premaxilla is laterally fused to the maxilla for about the anterior 100 mm of
the rostrum, then the premaxilla–maxilla suture is marked by a thin sulcus becoming deeper towards
Life 2020, 10, 27 13 of 62
the posterior portion of the rostrum, but remaining narrow for all the anteroposterior extension of
the rostrum, without forming a deep and wide lateral groove as observed in the platanistids
Platanista, Pomatodelphis, and Zarhachis.
In the anterior portion of the rostrum the premaxillae are dorsoventrally flattened, then,
proceeding posteriorly, their cross section becomes hemicylindrical at rostrum mid-length, before
flattening again towards the posterior portion of the rostrum. At the base of the rostrum the dorsal
surface of the premaxillae is flat but markedly medioventrally sloping to form a prenarial depression,
also observed in most other platanistoids and some other archaic odontocetes (e.g., squalodontids).
About 150 mm anterior to the base of the rostrum, at the level where the premaxillae begin to
diverge, the right premaxilla is distinctly transversely narrower than the left premaxilla. This
character is observed in all platanistoids, with the exception of Araeodelphis. At the base of the rostrum
the premaxillae are moderately wide (transverse width of the premaxillae equals 63% of the width of
the rostrum), an intermediate condition between allodelphinids (<60%) on the one side, and Medocinia
and Squalodelphis (>75%) on the other side.
A single large premaxillary foramen is present on each premaxilla 35 mm anterior to the rostrum
base. The premaxillary foramen is also anterior to the antorbital notch in other platanistoids, with the
exception of Dilophodelphis, Macrosqualodelphis, Platanista, and the Notocetus skulls from the Chilcatay
Fm, all having the premaxillary foramen roughly level with the antorbital notch. The anteromedial
and posterolateral sulci are wide and clearly discernible, especially on the right side, whereas the
posteromedial sulcus is weakly excavated. The premaxillary sac fossae are moderately transversely
concave, they slope medioventrally, and they have the same transverse width. The right ascending
process of the premaxilla ends with a posterior point incised by a notch followed anteriorly by a
longitudinal wide groove, similar to the premaxillary cleft described in Waipatia [99], also observed
in most other platanistoids (e.g., [24,26,112]), and in several other archaic odontocetes (e.g., [105,118]).
The left ascending process of the premaxilla has a rounded posterior margin without incision or
groove. The posterior end of both the right and left ascending processes of the premaxillae contacts
the anterolateral angle of the corresponding nasal. Such a limited posterior extension of the premaxillae
distinguishes Ensidelphis from the other platanistoids, all having the posteromedial margin of the
ascending process of both premaxillae in contact with the lateral margin of the nasal. A partial exception
is observed in Furcacetus, whose right premaxilla only contacts the anterolateral angle of the right nasal.
Nevertheless, in Furcacetus the left premaxilla displays a significant posterior extension, as in all other
platanistoids except Ensidelphis. A greater posterior extension of the premaxillae is present in
allodelphinids, all having the premaxillae extending posteriorly beyond the nasals.
Life 2020, 10, 27 14 of 62
Figure 4. Skull in left lateral view of the holotype (MUSM 3898) of Ensidelphis riveroi from the lower
Miocene Chilcatay Fm (Zamaca, Pisco Basin, Peru). (a,b), complete skull; (c,d), detail of the
neurocranium. Linear hatching indicates major breaks, dark shading areas covered by the sediment
or dental alveoli, and beige shading reconstructed missing parts. In (b) and (d) the mandibles are
shown in blue and the tympanic bulla in brown.
Maxilla. In dorsal view (Figure 2) the maxilla remains transversely narrow along the entire
length of the rostrum, showing a flat dorsal surface roughly parallel to the horizontal plane, only
weakly ventrolaterally sloping in the mid portion of the rostrum.
Formed by the maxilla, the lateral margin of the posterior portion of the rostrum is
dorsoventrally thin and blade-like as in all platanistoids of the Platanidelphidi clade, with the
exception of Huaridelphis, Medocinia, and Notocetus, having a markedly thicker margin.
Both ascending processes of the maxillae are partly broken; consequently, their original dorsal
elevation cannot be assessed, although their preserved parts are already higher than the dorsal
margin of the rostrum base, a condition shared with all other platanistoids. Due to the incompleteness
of the antorbital processes, the shape of the two antorbital notches is unknown.
No dorsal infraorbital foramina are visible around the base of the rostrum, whereas a posterior
dorsal infraorbital foramen is located more medial than the lateralmost margin of the premaxilla. The
same position of the posterior infraorbital foramina is observed in all known skulls of Platanidelphidi
where these foramina are visible.
Life 2020, 10, 27 15 of 62
Figure 5. Skull in right lateral view of the holotype (MUSM 3898) of Ensidelphis riveroi from the lower
Miocene Chilcatay Fm (Zamaca, Pisco Basin, Peru). (a,b), complete skull; (c,d), detail of the
neurocranium. Linear hatching indicates major breaks, dark shading areas covered by the sediment
or dental alveoli, and beige shading reconstructed missing parts. In (b) and (d) the mandibles are
shown in blue, the tympanic bulla in brown, and the teeth in orange.
The posteromedial portions of the ascending processes of the maxillae lateral to the vertex are
weakly asymmetrical with the right more anteroposteriorly elongated than the left. Furthermore, the
right maxilla descends more abruptly ventrolaterally from the vertex than the left (a condition
opposite to that observed in squalodelphinids, all having the left maxilla sloping more abruptly
ventrolaterally), to form a deep fossa posterolateral to the right nasal.
The palatal surface of the maxillae is partly covered by the articulated mandibles (Figure 3);
however, the latter are partly shifted to the left, making the ventral surface of the right maxilla largely
visible along most of the rostrum. This surface is flat and horizontal; near the lateral margin of the
rostrum, it is pierced by well-defined alveoli.
Presphenoid and cribriform plate. The well-preserved nasal septum runs anteroposteriorly
along the sagittal plane from the base of the rostrum to the cribriform plate, separating the bony nares
(Figure 2c,d). The cribriform plate borders anteriorly the nasals, reaching dorsally the anterodorsal
margin of these bones.
Nasal. The vertex is a flat, trapezoidal area where the sutures between bones are not clearly
visible (Figure 2c,d). This could be due to some post-mortem abrasion of the skull or, more likely, it
could be a genuine anatomical feature. In fact, bone fusion at the vertex is observed in other
odontocetes, as for example in the beaked whale Tusciziphius [119]. If abrasion can be excluded, the
nasals of Ensidelphis have a flat dorsal surface. They are probably anteroposteriorly compressed,
being shorter than the frontals at the vertex, and, together in dorsal view, form a rectangle with an
angle of about 7° between its anterior side and a coronal plane of the skull. A similar oblique
Life 2020, 10, 27 16 of 62
orientation of the longitudinal axis of the nasals is present in most Platanidelphidi and also in the
archaic odontocete Waipatia. Although the nasal-frontal suture is not clearly discernible, being only
tentatively reconstructed in Figure 2b,d, it seems that the maximum transverse compression across
the vertex occurs at the level of the nasals, with the frontals progressively widening posteriorly. In
lateral view the dorsal margin of the nasals slopes anteroventrally, forming an angle of 12° with the
dorsal margin of the rostrum. An anterior slope of the nasals is also observed in Furcacetus and
Huaridelphis, being more pronounced in the latter (22–35°), whereas it is absent in other platanistoids
such as Macrosqualodelphis and Notocetus, both displaying an inflated and subhorizontal dorsal
surface of the nasals.
Frontal. The frontals at the vertex are trapezoidal, flat and slope anteroventrally with the same
inclination as the nasals (Figure 2c,d). The suture between the right and left frontals is not discernible.
Both preorbital processes of the frontals are broken and only a posteromedial portion is
preserved. In lateral view the best-preserved right process (Figure 5c,d) displays a thickening (ratio
between the height of this process measured in lateral view perpendicular to the maxilla-frontal
suture and the vertical distance from the lower margin of the occipital condyles to the vertex of the
skull = 0.10) that is lesser than in Dilophodelphis, Furcacetus, Medocinia, Pomatodelphis, and Zarhachis,
all having ratios >0.30. However, this value in Ensidelphis was measured along the broken lateral
surface and a thickening of the preorbital process in its missing lateral portion cannot be excluded.
Along the same longitudinal break of the supraorbital process, the section visible in lateral view
suggests that the orbit was short and that the postorbital process was robust and triangular.
The articulated mandibles cover most of the ventral surface of the two orbit roofs (Figure 3);
consequently, it is not possible to check for the presence of a deep fossa in the medial portion of the
orbit roof, as observed in all other Platanidelphidi.
Supraoccipital. In lateral view the nuchal crest is not prominent, even if it represents the highest
part of the skull (Figures 4 and 5). In anterior (Figure 6a,b) and posterior (Figure 6c,d) views of the
cranium this crest draws a straight horizontal line, whereas in dorsal view it displays a low posterior
concavity.
The lateral margin of the supraoccipital forms, together with the parietal, the temporal crest that
delimits posterodorsally the temporal fossa. In posterior view the temporal crest is obliquely oriented
due to the gradual transverse narrowing of the occipital shield (supraoccipital + exoccipitals)
ventrally, as mentioned above. A maximum transverse constriction of the occipital shield is similarly
located ventrally, although less marked, in Macrosqualodelphis. All the other platanistoids have lateral
margins of the occipital shield less laterally concave in posterior view; in several cases these margins
are almost straight (i.e., pomatodelphinines and allodelphinids). In all these cases the maximum
transverse constriction of the shield is not located ventrally as in Ensidelphis and Macrosqualodelphinus.
A peculiar condition is observed in Platanista, displaying a remarkable transverse narrowing of the
shield related to the transverse widening of the temporal fossae. However, in Platanista the narrowest
portion of the occipital is located more dorsally, suggesting a non-homologous origin of this feature
in Ensidelphis and in the extant South Asian river dolphin.
Life 2020, 10, 27 17 of 62
Figure 6. Skull of the holotype (MUSM 3898) of Ensidelphis riveroi from the lower Miocene Chilcatay
Fm (Zamaca, Pisco Basin, Peru). (a,b), anterodorsal view; (c,d), posterior view. Linear hatching
indicates major breaks, dark shading areas covered by the sediment, and beige shading reconstructed
missing parts. In (b) the mandibles are shown in blue.
The posterodorsal surface of the occipital shield is weakly transversely concave and exhibits two
wide, roughly circular depressions with a diameter of about 30 mm, one for each side of the
supraoccipital. These depressions might indicate the origin of neck muscles, such as m. semispinalis
capitis or m. rhomboideus capitis (see [120]). There is no external occipital crest (sensu [27]).
Palatine. In ventral view, palatines are not visible in the well exposed posteromedial portion of
the rostrum (Figure 3c,d), possibly because they are fully covered by the pterygoids. However, the
lateral portion of the neurocranium is still covered by sediment and by the two mandibular rami;
consequently, it is not possible to check for the presence of a lateral exposure of the palatine.
Pterygoid. The right and left pterygoids, joined together medially, form a narrow point that
extends 55 mm beyond the level of the right antorbital notch (Figure 3c,d). Partially covered by the
pterygoid plates, the pterygoid sinus fossae also reach beyond the base of the rostrum, as in all
platanistoids. The well-preserved left lateral lamina of the pterygoid is a rectilinear plate that contacts
posterolaterally the falciform process of the squamosal.
Jugal-Lacrimal. The lacrimal and the jugal are lost, on both sides of the skull, due to the breakage
of the antorbital processes (Figures 4 and 5).
Squamosal. In lateral view (Figures 4 and 5), the zygomatic process of the squamosal is short
and robust, with a maximum thickness making 35% of the vertical distance from the lower margin of
the occipital condyles to the vertex. A value similar or greater than in Ensidelphis was observed in all
Platanidelphidi. The zygomatic process of Ensidelphis also shares with other Platanidelphidi the
globose shape in lateral view, due to the dorsal margin being convex and the ventral margin being
not concave (in this case it is straight and obliquely oriented). In particular the dorsal margin of the
zygomatic process of Ensidelphis draws a regular arch, as in most Platanidelphidi with the exception
Life 2020, 10, 27 18 of 62
of Pomatodelphis and Zarhachis, having the posterior portion of the dorsal margin that slopes more
abruptly posteriorly. In Ensidelphis, the anterodorsal surface of the zygomatic process is tightly
appressed to the postorbital process of the frontal, a feature due to the anterodorsal development of
the zygomatic process and shared with all platanistoids and eurhinodelphinids. In lateral view the
sternocephalicus fossa extends on the posteroventral portion of the zygomatic process as a narrow
and elongate groove that forms an angle of about 65° with the horizontal plane. The postglenoid
process is small and anteroventrally direct as in other Platanidelphidi. The squamosal plate is visible
in lateral view, forming the ventral portion of the medial wall of the temporal fossa. This wall is
locally laterally inflated, forming the peculiar temporal swelling mentioned above, with the
maximum lateral expansion (better seen in posterior view) in correspondence to the squamosal-
parietal suture. In Platanista this suture line is only slightly swollen.
The ventral surface of the zygomatic processes is almost entirely covered by the articulated
mandibular condyles (Figure 3). The thin plate of the left falciform process can be observed,
contacting anteriorly the posterolaterally elongated, plate-like lateral lamina of the pterygoid.
Parietal. Visible on the medial wall of the temporal fossa the parietal forms most of the lateral
wall of the neurocranium (Figures 4 and 5). The aforementioned temporal swelling involves also the
ventral portion of the parietal exposed in the temporal fossa.
Exoccipital. The occipital condyles are large and posteriorly prominent: they are bordered
dorsolaterally by well-excavated dorsal condyloid fossae (visible in posterior view: Figure 6c,d) and
ventrally by ventral condyloid fossae (visible in ventral view: Figure 3). The foramen magnum is
circular and the jugular notches are deeply incised. The paroccipital process is robust; in lateral view
it is thicker and significantly more ventrally extended than the postglenoid process of the squamosal,
a condition observed in all Platanidelphidi, with the exception of Platanista, which has an atrophied
paroccipital process and a large, ventrally extended postglenoid process.
Basioccipital. The basioccipital basin is transversely narrow, delimited laterally by robust
basioccipital crests that are posterolaterally bent, drawing together a small angle (about 30°) (Figure 3).
Vomer. The vomer is visible in ventral view and delimits posteriorly and medially each choana
(Figure 3).
Tympanic bulla. Both tympanic bullae are preserved in situ, well exposed on the ventral surface
of the skull (Figures 3, 7a,b). They are also visible, partially covered by the mandibles, with the skull
in lateral view (Figures 4, 5, 7c,d). As in the platanistids Platanista, Pomatodelphis, and Zarhachis, the
median furrow is partially filled with spongy bone, and, unlike the squalodelphinids for which the
tympanic bulla is preserved (Notocetus, Phocageneus, and Squalodelphis), it does not extend anteriorly
on the anterior spine (Figure 7a,b). The outer posterior prominence is slightly shorter than the inner
posterior prominence, a condition shared with the allodelphinids Allodelphis, Ninjadelphis, and
Zarhinocetus, whereas in the squalodelphinids Notocetus, Phocageneus, and Squalodelphis the outer and
inner posterior prominences have approximately the same posterior extent, and in the platanistids
Platanista, Pomatodelphis, and Zarhachis the outer posterior prominence extends farther posteriorly
than the inner prominence.
Life 2020, 10, 27 19 of 62
Figure 7. Right tympanic bulla articulated to the skull of the holotype (MUSM 3898) of Ensidelphis
riveroi from the lower Miocene Chilcatay Fm (Zamaca, Pisco Basin, Peru). (a,b), ventral view; (c,d),
lateral view. In (b) and (d) the tympanic bulla is shown in brown and the mandible in blue.
Abbreviations: as, anterior spine; basiocc, basioccipital; cp, conical process; ipp, inner posterior
prominence; ls, lateral furrow; mf, median furrow; opp, outer posterior prominence; postgl proc,
postglenoid process; pp, posterior process; sp, sigmoid process.
The anterior spine is thin and very long. A more or less elongated anterior spine is present in all
platanistoids whose tympanic bulla is preserved, but we observed an extreme elongation as in
Ensidelphis (anterior spine ca 27% of the total length of the bulla) only in the tympanic bulla associated
with the skull MUSM 603 (Figure 7 in [24]). Interestingly enough, the tympanic bulla MUSM 603
shares other features with Ensidelphis (e.g., spongy bones in the median furrow, outer posterior
prominence posteriorly shorter than the inner posterior prominence), supporting a congeneric
referral of the two specimens (see below). The partly exposed lateral surface of the tympanic bulla of
Ensidelphis (Figures 4, 5, 7c,d) reveals a high and inflated outer lip, as in other platanistoids (e.g., [92]).
The conical process is moderately developed and almost in contact posteriorly with the elongated
sigmoid process, the latter having its distal portion posterodorsally bent. The lateral furrow is clearly
visible. A small portion of the posterior process of the tympanic bulla, articulated with the squamosal
and the exoccipital, is also visible.
Mandible
The mandibles are tightly articulated with the cranium, the right and the left mandibular
condyles being inside their respective mandibular fossae on the ventral surface of the zygomatic
processes of the squamosals (Figure 3). The two mandibles are fused together as in all other
Platanistoidea having mandibles preserved and in the Eurhinodelphinoidea (Eurhinodelphinidae +
Eoplatanistidae), but not in the longirostral homodont odontocete Chilcacetus [57]. In particular, along
the 150 mm-long anterior portion of the symphysis the mandibles are ankylosed, with the medial
suture being invisible in ventral view. The long mandibular symphysis represents 61% of the total
length of the mandibles. Posterior to the symphysis the two mandibular rami draw an angle of 25° in
ventral view. The same proportions of the symphysis and a similar angle between the rami are
observed in all allodelphinids, whereas the platanistids Araeodelphis, Platanista, Pomatodelphis, and
Zarhachis have a more elongated symphysis (>65%) and a consequently larger angle between the
mandibular rami (roughly 60°). Among other platanistoids the mandible is only well known in the
squalodelphinids Notocetus and Squalodelphis, both having a shorter symphysis (40% and 43%,
respectively) and an angle between the mandibular rami equal to 38°. These values are consistent
with a rostrum significantly shorter than in Ensidelphis.
The symphysis of Ensidelphis is dorsoventrally flattened and it is longitudinally furrowed by two
deep lateral grooves, one for each mandible, running from 65 mm from the anterior end of the
mandible to the posterior end of the symphysis (Figures 4a,b, 5a,b). Such grooves are present in all
platanistoids whose mandible is preserved, with the exception of the squalodelphinids Huaridelphis,
Notocetus, and Squalodelphis.
In lateral view the height of the postsymphyseal portion of the mandible increases gradually
posteriorly, with both the dorsal and ventral margins forming a low angle with the horizontal axis of
Life 2020, 10, 27 20 of 62
the mandible. Among other platanistoids with a mandible preserved, a similar shape of the
mandibular rami in lateral view is observed in the allodelphinid Goedertius and the squalodelphinid
Notocetus, whereas a more abrupt posterior elevation of the dorsal margin is present in Zarhinocetus
and, to an even greater extent, in Platanista and Zarhachis. The mandible of Squalodelphis is too
damaged to assess this feature.
Dentition
In ventral view on the right side of the rostrum, 54 alveoli are visible, the six anteriormost being
in the premaxilla (Figure 5a,b). However, this value does not represent the total tooth count of the
upper right quadrant, considering that the posterior portion of the upper right alveolar row is
covered by the mandible. The complete alveolar row is instead exposed on the left mandible (Figure
4a,b); here 62 alveoli are counted and two additional alveoli are estimated to have been originally
present in a 28 mm-long reconstructed mid-length portion of the rostrum, resulting in a total tooth
count for each mandible of about 64. Since the exposed alveoli on the rostrum have roughly the same
longitudinal length and the same spacing as on the mandible, approximately 64 teeth could also be
inferred for each upper quadrant. Therefore, the total tooth count of Ensidelphis is estimated at 256. A
tooth count > 200 characterizes all hyper-longirostrine platanistoids, with a remarkable count of about
315 teeth in Zarhachis [111].
The alveoli are small and roughly circular, only slightly longer than transversely wide. Their
transverse width varies between 2.4 and 4.4 mm, with an average value of 3.5 mm. The ratio between
the transverse width for alveoli at mid rostrum length and the BZW is 0.018. Similarly, low values
(<0.02) are observed in all other platanistoids, except in squalodelphinids (ratio > 0.03). The
interalveolar septa range between 2.3 and 8.0 mm in length, with an average value of 5.5 mm. On
each tooth row, the 6–8 anteriormost alveoli (for premaxillary teeth) are slightly smaller and less
spaced than more posterior alveoli. However, with the exception of a few cases, interalveolar septa
are longer than the adjacent alveoli, suggesting that when the mouth was closed interlocking teeth of
the upper and lower tooth rows did not systematically contact each other. It is significant to underline
that for example in the extant longirostrine dolphin Pontoporia blainvillei interalveolar spaces increase
with the age of the animal (C.M. personal observation); therefore, in Ensidelphis riveroi this character
could also be subject to ontogenetic variation.
Figure 8. In situ lower teeth of the holotype (MUSM 3898) of Ensidelphis riveroi from the lower Miocene
Chilcatay Fm (Zamaca, Pisco Basin, Peru). (a), posterior tooth; (b), two anterior teeth. Tympanic bulla
shown in brown and mandibles in blue.
Six teeth are preserved in their alveoli on the right mandible (Figures 5a,b, 8): two in the anterior
part of the symphyseal portion (sixth and seventh teeth from the apex) and four in the
postsymphyseal portion. All preserved teeth are single rooted, with a simple conical crown lacking
Life 2020, 10, 27 21 of 62
accessory denticles or cingula. However, it cannot be excluded that accessory denticles and/or cingula
were present in the lost posteriormost teeth, as in some other platanistoids (e.g., Notocetus and
Phocageneus). The crowns of the two preserved anterior teeth are broken and only their basal portion
is preserved; their diameter is 4.3 mm. The best-preserved posterior tooth is slender, having a
diameter at the base of the crown of 2.8 mm; the height of its almost complete crown is 4.6 mm.
Cervical Vertebrae
None of the atlas, axis, and two other cervicals were fused (Figure 9). They were kept attached
by sediment in their position as found in the field, which is not in anatomical connection but strictly
associated. We think it is plausible that the original anatomical sequence was maintained and that,
therefore, the two vertebrae posterior to the axis represent the third and fourth cervicals.
Life 2020, 10, 27 22 of 62
Figure 9. Cervical vertebrae of the holotype (MUSM 3898) of Ensidelphis riveroi from the lower
Miocene Chilcatay Fm (Zamaca, Pisco Basin, Peru). (a,b), overall view of the disarticulated cervicals
joined by sediment. (ch), atlas in anterior (c,d), lateral (e,f), and ventral (g,h) views. (in), axis in
anterior (i, j), dorsal (k,l), and ventral (m, n) views. (or), fourth cervical in anterior (o,p) and ventral
(q,r) views. Linear hatching indicates major breaks. The atlas is shown in blue, the axis in orange, the
third cervical in green, the fourth cervical in yellow, and a fragment of unidentified cervical in brown.
Atlas. The dorsal transverse processes of the atlas are slightly more elongated and robust than
the ventral transverse processes (Figure 9c–h). As far as the degree of reduction of the ventral
transverse processes is concerned, Ensidelphis is intermediate between squalodelphinids
(Huaridelphis, Macrosqualodelphis, Notocetus, and Phocageneus), all having an extreme reduction of this
process, and the other platanistoids, all having similarly elongated dorsal and ventral transverse
processes. As in Macrosqualodelphis and Notocetus the dorsal transverse processes are widely visible
in anterior view, unlike in other platanistoids whose atlas is preserved. This is due to the less
posteriorly projected condition of these processes. The neural canal is transversely compressed (ratio
between width and height = 0.81), as in Macrosqualodelphis (ratio 0.76), whereas it is circular or slightly
dorsoventrally compressed in other platanistoids with the atlas preserved. The neural arch is low,
pierced by large lateral vertebral foramina for the vertebrarterial canal, and with a transversally thin
and short neural spine. The ventral tubercle is robust with a posteroventrally directed pointed tip.
Measurements of the cervical vertebrae of MUSM 3898 are provided in Table 2.
Table 2. Measurements on the cervical vertebrae of Ensidelphis riveroi MUSM 3898 (holotype) from
the Chilcatay Fm (early Miocene, Peru). All measurements are in mm.
Dimension C1 C2 C3 C4
Width of vertebra 121 110 97 e100
Height of vertebra 78 110 71 78
Centrum length - 33 33 33
Centrum width - 43 e50 46
Centrum height - 41 39 40
Neural canal width 36 - - 33
Neural canal height 41 - - 20
- missing data; e, estimate.
Axis. Among Platanidelphidi the axis is only known in Araeodelphis (USNM 16569), Huaridelphis
(MUSM 1403, incomplete), Platanista, and USNM 206006, an undescribed platanistid from the Calvert
Formation (U.S.A.) showing affinities with Pomatodelphis and Zarhachis. The transverse processes of
Ensidelphis are similar to those of Huaridelphis, being short and robust, triangular in dorsal and ventral
view, and posteriorly projected (Figure 9i–n). The transverse processes of Araeodelphis, Platanista, and
USNM 206006 are slender, longer, and posteroventrally projected. The neural arch and the neural
spine of Ensidelphis are massive and anteroposteriorly thick in lateral view, whereas they are thin in
Araeodelphis and USNM 206006, and dorsoventrally short and overall triangular in Platanista.
Third and fourth cervicals. Among Platanidelphidi, the cervical vertebrae posterior to the axis
are only known in Araeodelphis (USNM 16569), Huaridelphis (MUSM 1403, incomplete), Phocageneus
(USNM 21036, only the third and fifth) Platanista, and Notocetus (only one in MUSM 1395). Better
preserved than the third, the fourth cervical of Ensidelphis shows the closest similarities with the third
cervical of Phocageneus, both having an almost circular centrum in anterior view, a wider than high
neural canal, and a low and transversely wide medial keel (Figure 9o–r). However, the ventral
transverse processes in C4 of Ensidelphis are more ventrally projected and the transverse foramina for
the vertebrarterial canal are smaller than in the C3 of Phocageneus. Cervicals of the allodelphinids
Allodelphis, Goedertius, and Ninjadelphis differ from the cervicals of Ensidelphis and of the other
Platanidelphidi by having significantly more anteroposteriorly elongated centra [121].
Life 2020, 10, 27 23 of 62
Platanidelphidi indet.
Figure 10, Table 1
Referred specimen, locality, and age. MUSM 3899 is an incomplete cranium with several
missing parts including the anterior portion of the rostrum and the right orbital region. Moreover,
several areas are damaged, such as the premaxillae on the rostrum, the vertex, and the supraoccipital,
and the posterior half of the neurocranium is shifted to the left in respect to the sagittal plane, as
clearly seen in ventral view. On the palatal surface of the rostrum 21 and 24 small alveoli for single-
rooted teeth are counted on the right and left side, respectively. Earbones and teeth are not preserved.
Zamaca locality, Western Ica Valley, Ica Region, Peru (Figure 1a,b). Geographic coordinates:
14°37'28.77" S, 75°38'22.92" W; 345 m above sea level. This specimen was reported in the Zamaca fossil
map of Di Celma et al. [22] with the field number ZM 128 and provisionally referred to
“Platanistoidea indet.” From the Chilcatay Fm, 38.1 m above the contact with the underlying Otuma
Formation, in the Ct1a facies association of the Ct1 allomember [21,22] (Figure 1d). The age of this
portion of the Ct1a facies association can be constricted between 19.00 ± 0.25 Ma and 18.08 ± 0.07 Ma
(early Burdigalian) on the basis of two volcanic ash layer samples dated by
40
Ar/
39
Ar [89].
Figure 10. Cranium (MUSM 3899) of Platanidelphidi indet. from the lower Miocene Chilcatay Fm
(Zamaca, Pisco Basin, Peru). (a,b), dorsal view. (c,d), ventral view. (e,f), left lateral view. Oblique
linear hatching indicates major breaks, horizontal dotted hatching eroded surface, and dark shading
areas covered by the sediment or dental alveoli.
Brief Description and Comparison
The long hamular fossa of the pterygoid sinus extending anteriorly on the palatal surface of the
rostrum, the cranium distinctly shorter than wide, and the anterior portion of the zygomatic process
of the squamosal tightly appressed to the postorbital process of the frontal are all derived characters
allowing us to refer MUSM 3899 to the Platanistoidea superfamily (Figure 10). In particular, MUSM
3899 belongs to the Platanidelphidi clade in having: (1) Asymmetry of the premaxillae on the rostrum
at some distance anterior to the premaxillary foramina, with the right premaxilla being distinctly
narrower than the left in dorsal view; (2) posterior dorsal infraorbital foramen along the vertex more
medial than the lateralmost margin of the premaxilla in the cranium; (3) deep fossa in the frontal on
the orbit roof, at the level of the frontal groove, presumably for an orbital lobe of the pterygoid sinus;
Life 2020, 10, 27 24 of 62
(4) vertex distinctly shifted to the left compared to the sagittal plane of the skull; (5) palatine not
exposed anterior to the pterygoid; (6) dorsoventrally thick zygomatic process of the squamosal; and
(7) straight ventral edge of the zygomatic process in lateral view.
Within the Platanidelphidi MUSM 3899 shares affinities with Ensidelphis in its small size (BZW
equals 183 cm in MUSM 3899 and 196 cm in the holotype of Ensidelphis riveroi; bicondylar width
equals 79 cm in MUSM 3899 and 80 cm in the holotype of E. riveroi), the narrow and possibly
elongated rostrum, the shape of the zygomatic process of the squamosal (with half-circle shaped
dorsal margin and straight anteroventral margin), the moderately transversely wide dorsal opening
of the mesorostral groove in the rostrum base area, the transversely wide and anteriorly located
premaxillary foramen, the limited posterior extension of the ascending processes of the premaxillae,
and the small maxillary alveoli (the average of the transverse diameter for the 21 posteriormost
alveoli is 3.8 m in both MUSM 3899 and Ensidelphis). Moreover MUSM 3899 shares with Ensidelphis
the absence of any of the derived characters distinguishing both the squalodelphinids and the
platanistids within the Platanidelphidi. Nevertheless, MUSM 3899 differs from Ensidelphis in the
dorsoventrally thinner supraorbital process of the frontal, the transversely narrower vertex (20% and
25% of the BZW in MUSM 3899 and Ensidelphis, respectively), the right and left posterior dorsal
infraorbital foramina being more laterally located, and the apparent absence of the peculiar temporal
swelling on the medial wall of the temporal fossa (but this area is badly preserved on both sides of
MUSM 3899's cranium). Moreover, MUSM 3899 differs from Ensidelphis in the narrower space
between the alveoli (the average length of interalveolar septa between the 21 posteriormost alveoli is
3.0 m in MUSM 3899 and 4.6 in the holotype of E. riveroi), although this character, as outlined above,
could be subject to intraspecific ontogenetic variation. Based on these observations and considering
the fragmentary state of the specimen, we assign MUSM 3899 to an indeterminate basal
Platanidelphidi, pending the discovery of more complete specimens.
Squalodelphinidae Dal Piaz, 1917
Emended diagnosis. The Squalodelphinidae have the following synapomorphies, absent in
other members of the Platanidelphidi clade: (1) Deep, V-shaped, left antorbital notch, related to an
anteriorly pointed left antorbital process; (2) left-side torsion of the rostrum with the longitudinal axis
of the neurocranium forming an angle of about 5° with the main axis of the rostrum in dorsal view,
generating asymmetry of the posterior portion of the rostrum; (3) pars cochlearis of the periotic
square-shaped in ventral view; (4) large and thin-edged aperture of the cochlear aqueduct of the
periotic; (5) median furrow of the tympanic affecting the whole length of the bone, including the
anterior spine; (6) apical extension of the manubrium of the malleus; (7) strong development of the
dorsal transverse process of the atlas and extreme reduction of its ventral process.
Type genus. Squalodelphis Dal Piaz, 1917
Other genera included. Furcacetus gen. nov., Huaridelphis, Macrosqualodelphis, Medocinia,
Notocetus, Phocageneus.
Furcacetus, gen. nov.
LSID: zoobank.org:act: 7B69F853-98CE-4824-88AC-3294D1B0580D
Type and only known species. Furcacetus flexirostrum, sp. nov.
Diagnosis. As for the type and only referred species.
Etymology. From ‘furca’, fork in Latin, and ‘cetus’, whale in Latin. For the procumbent
anterior upper teeth, which, together with the rostrum, look like a fork. Gender masculine.
Furcacetus flexirostrum, sp. nov.
Figures 11-13, Table 3
LSID: zoobank.org:act: B290D90E-D943-4768-956F-28DD1DF75988
Life 2020, 10, 27 25 of 62
Holotype and only referred specimen. MUSM 487 consists of a cranium damaged in some parts;
in particular the left lateral and the posteroventral portions of the neurocranium are missing. Eight
broken teeth are inside their alveoli on the rostrum (five on the maxilla and three on the premaxilla).
The incomplete right periotic is still articulated to the cranium.
Type locality. MUSM 487 was collected several years ago from layers of the Chilcatay Fm in the
Zamaca-Ullujaya area, western Ica Valley, Ica Region, southern Peru (Figure 1a,b). Approximate
geographic coordinates: 14°36’ S, 75°38’ W.
Type horizon. The exact horizon of the Chilcatay Fm where the holotype of Furcacetus
flexirostrum, MUSM 487, was discovered is unknown. Nevertheless, the entire stratigraphical
sequence of the Chilcatay Fm exposed at Zamaca and Ullujaya localities has been roughly constricted
through radiometric dating of ash layers to an interval between 19.25 ± 0.05 and 18.02 ± 0.07 Ma (early
Burdigalian) [89].
Diagnosis. Furcacetus is a small odontocete having an asymmetrical cranium with a narrow and
moderately elongated rostrum (about 68% of the CBL), bearing about 25 single-rooted teeth in each
upper quadrant. Its rostrum differs from that of all other Platanistoidea s.s., as defined above, in
having a sigmoid shape in lateral view and bearing procumbent anterior teeth. Furcacetus shares with
the other platanistoids the elongated hamular fossa of the pterygoid sinus extending anteriorly on
the palatal surface of the rostrum and the neurocranium being shorter than wide (ratio < 0.90).
Furcacetus belongs to the Platanidelphidi clade in having: asymmetry of the premaxillae on the
rostrum at some distance anterior to the premaxillary foramina (ca 18 cm in this case), with the right
premaxilla distinctly narrower than the left in dorsal view; posterior dorsal infraorbital foramen
along the vertex more medial than the lateralmost margin of the premaxilla on the cranium; deep
fossa in the frontal on orbit roof, at the level of the frontal groove; vertex distinctly shifted to the left
compared to the sagittal plane of the skull; and thick zygomatic process of the squamosal (ratio
between the maximum distance from the anteroventral margin of the zygomatic process to the
posterodorsal margin, in lateral view, and the vertical distance from the lower margin of the occipital
condyles to the vertex of the skull > 0.35). Furcacetus is referred to the Squalodelphinidae in having:
pars cochlearis of the periotic square-shaped in ventral view; and longitudinal axis of the
neurocranium forming an angle of about 5° with the main axis of the rostrum in dorsal view,
generating asymmetry in the posterior portion of the rostrum.
It differs from all other squalodelphinids in having a greater asymmetry of the ascending processes
of the premaxillae, the left process being significantly wider and more posteriorly extended than the
right; and in having about 25 teeth for upper quadrant, a tooth count that is greater than in Notocetus
(22–23) and Squalodelphis (15), and lower than in Dilophodelphis (ca 35) and Huaridelphis (28–30) (exact
tooth count unknown in Macrosqualodelphis, Medocinia, and Phocageneus). It shares with
Macrosqualodelphis and Notocetus the anteroposteriorly elongated temporal fossa (ratio between
horizontal width and vertical height = 1.42) and the correspondingly elongated zygomatic process of
the squamosal. It shares with Dilophodelphis and Huaridelphis a deep, V-shaped right antorbital notch
drawing an angle of about 60°. It shares with Notocetus the transverse widening of the premaxillae in
the anterior portion of the rostrum. It shares with Dilophodelphis and Medocinia a marked dorsoventral
thickening of the preorbital process of the frontal. It differs from Medocinia, and Squalodelphis in the
narrower transverse opening of the mesorostral groove near the rostrum base and in the lesser
transverse widening of the premaxilla at rostrum base.
Life 2020, 10, 27 26 of 62
Figure 11. Cranium in dorsal view of the holotype (MUSM 487) of Furcacetus flexirostrum from the
lower Miocene Chilcatay Fm (Pisco Basin, Peru). Linear hatching indicates major breaks. Teeth are
shown in orange.
Etymology. From ‘flexus’, bent in Latin, and ‘rostrum’. For the sinusoidal shape of the rostrum in
lateral view of the cranium.
Description and Comparison
Ontogeny. The closed sutures of the cranial bones, the medial fusion of the frontals on the vertex,
and the well-individualized alveoli on the rostrum suggest that the holotype and only referred
specimen of Furcacetus flexirostrum MUSM 487 was an adult animal.
Total body length estimate. Estimating the BZW of the holotype MUSM 487 (left zygomatic
process of the squamosal lost) at 240 mm, we used the equation proposed by Pyenson and Sponberg
[110] for stem Platanistoidea to obtain an approximate TBL of 234 cm for Furcacetus flexirostrum, a
value slightly lower than the TBL of Notocetus vanbenedeni (237–255 cm based on the BZW of MUSM
3896, and MUSM 3897).
Cranium
General morphology. By adding about 20 mm to the length for the missing part of the occipital
condyles, the cranium of Furcacetus could have reached a CBL of roughly 585 mm, 67% of which
being occupied by the rostrum (Table 3). These values are in the range of Notocetus (CBL = 580–634
mm; rostrum = 62%–68% of CBL). CBL of other squalodelphinids is either greater (Macrosqualodelphis:
> 770 mm; Squalodelphis: 640 mm) or lesser (Huaridelphis: 494 mm; Dilophodelphis: 440) than in
Furcacetus, but in all of them the rostrum is moderately elongated (63–70% of CBL), as in Furcacetus.
The rostrum of Furcacetus has a peculiar sinusoidal shape in lateral view (Figure 13a–d): from its
base, it curves upward until about 50 mm from the apex, where it curves downward until its anterior
end. The 50 mm downward-bent anterior portion of the rostrum is formed by the premaxillae only
and hosted procumbent incisors (Figure 13e,f). Among other squalodelphinids, we also observed a
Life 2020, 10, 27 27 of 62
curved (but not sinusoidal) rostrum in the cranium MUSM 1403 of Huaridelphis raimondii (but not in
the holotype) and in the crania MUSM 1395 and MUSM 3897 of Notocetus vanbenedeni (but not in the
holotype and the other referred specimens). Among extant odontocetes, although not associated with
procumbent incisors a sinusoidal shape of the rostrum was observed in a few crania of Pontoporia
blainvillei [114], whereas a rostrum raising anterodorsally, but lacking the anterior downward
counter-curvature and, again, the procumbent incisors, was noted in one cranium of the brevirostrine
pontoporiid Brachydelphis mazeasi [122] and in some crania of Platanista gangetica [113,123] and of few
delphinid species (e.g., Delphinus delphis MZUF 12484, G.B. personal observation; Delphinus capensis,
Figure 2 in [124] ).
The rostrum of Furcacetus is dorsoventrally compressed in its anterior and posterior portions,
whereas it becomes more laterally compressed towards mid-length. A deep transverse concavity,
involving both the maxillae and the premaxillae, occurs on the dorsal surface of the cranium around
the base of the rostrum (Figures 11,13c). This concavity is laterally overhung by the elevated
antorbital regions, which, as in all Platanidelphidi, are distinctly higher than the dorsal margin of the
rostrum base in lateral view.
The right antorbital notch is deep and V-shaped, drawing an angle of about 60°, as in
Dilophodelphis and Huaridelphis, whereas other squalodelphinids have a more open right antorbital
notch. The left antorbital notch is not preserved and, consequently, it is not possible to check if the
right and left antorbital notches were asymmetrical, as in most other platanistoids.
Table 3. Measurements on the crania of Furcacetus flexirostrum MUSM 487 (holotype) and Notocetus
vanbenedeni (MUSM 3896, 3897, 1395) from the Chilcatay Fm (early Miocene, Peru). The crania of N.
vanbedeni are compared with the holotype (MLP 5-5) and referred specimen (AMNH 9485) from the
Monte León Formation (early Miocene, Argentina). All measurements are in mm.
Dimension
Furcacetus
flexirostrum Notocetus vanbenedeni
MUSM 487
MUSM
3896
MUSM
3897
MUSM
1395
AMNH
9485
MLP 5-5
holotype
Condylobasal length
e585
590
+585
600
634*
580
Length of rostrum 392 380 +388 403 433* 360
Width of rostrum at its base
e97
117
126
136
142*
120
Width of premaxillae at base of rostrum 68 67 73 78 89* 81
Orbital width of skull - 208 238 227 252* 230
Bizygomatic width of skull
e240
243
263
-
-
+220
Width of maxillae at mid-length of
rostrum 39 50 49 49 52* 47
Width of premaxillae at mid-length of
rostrum 19 27 32 28 30* 25
Maximum width between temporal crests
118
125
162
145
142*
145
Minimum posterior distance between
temporal crests - 117 138 134 128* 140
Length of orbit - 53 56 55 - 58
Height of temporal fossa 71 83 72 66 74* 80
Length of temporal fossa
101
123
118
108
115*
118
Length of zygomatic process of
squamosal from posglenoid process to tip
of zygomatic process
- 90 90 91 - -
Maximum width of premaxillae on
neurocranium 101 110 e106 e105 - 108
Width of right premaxillary sac fossa 36 39 40 40 - 58
Width of left premaxillary sac fossa 33 40 40 41 - 48
Maximum distance between premaxillae
anterior to bony nares 7 - - e15 28* -
Width of bony nares 40 - e41 45 - 44
Anterior width of nasals - - - 43 48* 45
Length of medial suture of nasals - - - 14 - 19
Length of medial suture of frontals at
vertex - - - 28 - 21
Life 2020, 10, 27 28 of 62
Minimum posterior distance between
maxillae 34 - - 40 - 47
Distance between foramen magnum and
nuchal crest - 95 83 93 97* 91*
Width between lateral margins of
occipital condyles - 85 87 88 - 76
Height of right occipital condyle - 44 48 48 46* 45
Width of foramen magnum
-
35
41
e39
40*
34
Height of foramen magnum - 35 34 e29 40* 33*
Length of alveoli at mid-length of rostrum
7
-
10
9
9.5*
7*
Transverse width of alveoli at mid-length
of rostrum 7 - 10 9 8* 5.5*
Length of upper tooth row - 317 +285 +302 363* 315*
Number of teeth per upper tooth row e25 - 18 +18 21* 23
+, incomplete; - missing data; e, estimate; *, measurement from
[125]
.
As mentioned above, the original BZW of Furcacetus is estimated at 240 mm, suggesting that the
neurocranium was proportionally short (79% of BZW), a condition observed in all Platanistoidea.
Figure 12. Cranium in ventral view of the holotype (MUSM 487) of Furcacetus flexirostrum from the
lower Miocene Chilcatay Fm (Pisco Basin, Peru). Linear hatching indicates major breaks and dark
shading areas covered by the sediment. Right periotic shown in brown and teeth in orange.
The asymmetry of the cranium of Furcacetus is remarkable, as in most other Platanidelphidi. This
asymmetry concerns: (1) The vertex and the bony nares, being shifted to the left side; (2) the right
bony naris being transversely broader and more posteriorly located than the left; (3) the left
premaxilla being markedly more posteriorly extended than the left; (4) the left frontal at the vertex
being anteroposteriorly shorter than the right; and (5) the missing left nasal being probably originally
smaller than the right. Moreover, as in all other squalodelphinids the asymmetry of the cranium of
Furcacetus involves also the posterior portion of the rostrum due to the left lateral shift of the rostrum
Life 2020, 10, 27 29 of 62
(main axis of the rostrum in dorsal view forming an angle of about 5° with the longitudinal axis of
the neurocranium).
The temporal fossa is significantly elevated (ratio between the vertical height of the fossa, in
lateral view, and the vertical distance from the lower margin of the occipital condyles to the vertex of
the skull = 0.70), as in Ensidelphis, Macrosqualodelphis, Notocetus, and Platanista. The temporal fossa of
Furcacetus is also anteroposteriorly elongated (ratio between the horizontal length and vertical height
= 1.42), similarly to Macrosqualodelphis and Notocetus [26].
Figure 13. Holotype (MUSM 487) of Furcacetus flexirostrum from the lower Miocene Chilcatay Fm (Pisco
Basin, Peru). (a,b), cranium in right lateral view; (c) cranium in right anterolateral view; (d) cranium in
posterior view; (e,f), detail of the anterior portion of the rostrum in lateral view; (g,h), posterior tooth in
lateral (g) and anterior (h) views. Linear hatching indicates major breaks and dark shading areas covered
by the sediment or the dental alveoli. Right periotic shown in brown and teeth in orange.
Premaxilla. The anterior portion of the rostrum is formed by the premaxilla alone for 50 mm
(Figure 11), a condition also observed in all squalodelphinids having the apex of the rostrum
preserved (Dilophodelphis, Huaridelphis, Notocetus, and Squalodelphis), and in Ensidelphis. Partially
related to this feature, in dorsal view the lateral premaxilla-maxilla suture is laterally bent and the
premaxilla widens transversely towards the apex of the rostrum. A similar widening is observed in
Life 2020, 10, 27 30 of 62
Notocetus, as well as in squalodontids; in the latter it is similarly associated with procumbent
premaxillary teeth (e.g., [98]).
As outlined above, this anterior premaxillary portion of the rostrum is curved downward,
dorsoventrally compressed, and bears three alveoli on each side (Figure 13).
The dorsal surface of the anterior portion of the rostrum is poorly preserved, but a foramen that
pierces the left premaxilla is clearly visible 55 mm posteriorly to the apex. Foramina piercing the
premaxillae on the anterior portion of the rostrum are also observed in Araeodelphis, Dilophodelphis,
and Notocetus.
The dorsal opening of the mesorostral groove is very narrow (maximum transverse width = 2
mm) on the 160 mm-long anterior portion of the rostrum; the mesorostral groove is fully closed
dorsally, in the middle portion of the rostrum, for a tract 75 mm long, then the premaxillae gradually
diverge towards the base of the rostrum, although the opening remains narrow for the whole
posterior tract of the mesorostral groove, reaching a maximum transverse width of 8 mm. Among
other squalodelphinids, a similarly narrow opening of the mesorostral groove near the rostrum base
is present in Huaridelphis, Macrosqualodelphis, and Notocetus, whereas Medocinia and Squalodelphis
display a wider opening.
The premaxilla–maxilla suture is distinct along the whole rostral length, but it is not located in
a deep lateral groove as in the platanistids Platanista, Pomatodelphis, and Zarhachis.
The right and the left premaxillae on the rostrum maintain almost the same transverse width for
their whole anteroposterior extension, with the right premaxilla only slightly narrower than the left
near rostrum mid-length. The asymmetry of the premaxillae, a peculiar feature of the Platanidelphidi,
is more pronounced in other squalodelphinids, as for example Notocetus.
At the base of the rostrum the premaxilla is transversely wider than the maxilla as in most other
squalodelphinids, but not as much as in Medocinia and Squalodelphis. At this level each premaxilla
exhibits a marked medial slope, bounding a deep medial depression that extends posteriorly on the
neurocranium, with the premaxillary sac fossae also ventromedially sloping.
The premaxillary foramina (one on each premaxilla) are located roughly at the level of the right
antorbital notch, a condition observed, among other platanistoids, in Dilophodelphis,
Macrosqualodelphis, Platanista, and the Notocetus skulls from the Chilcatay Fm.
The weakly excavated premaxillary sac fossa is laterally delimited by a wide and deep
posterolateral sulcus that ends posteriorly where the premaxilla reaches it maximum transverse
width. The anteromedial sulcus is shallower than the posterolateral sulcus and, as in Huaridelphis,
Macrosqualodelphis, and Notocetus, is significantly elongated, extending about 120 mm anterior to the
premaxillary foramen. The posteromedial sulcus is not clearly discernible.
The anterior limit of the bony nares is defined by an angle of the medial margin of each
premaxilla, the angle of the right premaxilla being 20 mm posterior to the angle of the left, generating
the marked asymmetry of the bony nares mentioned above.
The right and left ascending processes of the premaxilla are strongly asymmetrical. The right
process ends with a posteromedial point that contacts the anterolateral angle of the right nasal. The
left process extends significantly more posteriorly, far beyond the nares; it was probably medially in
contact with the lost left nasal, and reaches posteriorly the frontal. Among other squalodelphinids,
an asymmetry of the ascending processes of the premaxillae is also observed in Dilophodelphis,
Huaridelphis, and Notocetus, but less marked than in Furcacetus, being limited to a more pointed end
of the right ascending process (left process not significantly longer than the right). Unlike in other
squalodelphinids and some platanistids (e.g., Zarhachis), there is no trace of a longitudinal groove on
the posterior portion of the ascending processes of the premaxillae of Furcacetus MUSM 487.
In ventral view (Figure 12), on the rostrum the premaxilla–maxilla suture runs obliquely from
the posterior margin of the third incisor to a medial point located 138 mm anterior to the right
antorbital notch. Consequently, the premaxillae display a narrow and elongated ventral exposure
between the palatal processes of the maxillae.
Maxilla. The rostral portions of the maxillae are not well preserved, with some missing parts,
especially for the left maxilla (Figure 11). However, it is clearly discernible that the transverse width
Life 2020, 10, 27 31 of 62
of the dorsal exposure of the maxilla remains narrow for the entire length of the rostrum. Moreover,
for most of its anteroposterior extension on the rostrum, the maxilla appears to slope laterally.
Approaching the antorbital notch, the maxilla first becomes flat and horizontal, then slopes medially,
forming, with the premaxilla, the deep dorsomedian depression at the base of the rostrum.
The only dorsal infraorbital foramen discernible in MUSM 487 is a small posterior foramen
piercing the ascending process of the right maxilla very close to the posterior end of the right
premaxilla and more medial than the lateralmost margin of the premaxilla. A similar position of the
posterior dorsal infraorbital foramen is observed in all the known skulls of Platanidelphidi for which
this area is well preserved.
Apparently, the maxilla displays a limited anterolateral extension above the preorbital process
of the frontal, although the maxilla–frontal suture it is not clearly discernible on the right side of the
cranium and the left side is incompletely preserved. Above the orbit, the right maxilla exhibits a
longitudinal bulge posterolateral to the antorbital notch (Figure 13c). A similar bulge is present in
other squalodelphinids, but less marked, except in Dilophodelphis (thickening significantly greater
than in Furcacetus) and Squalodelphis (thickening similar to Furcacetus). It is important to note that the
comparison was made using the right side of the neurocranium and that in other squalodelphinids
the right preorbital + orbital region is less elevated than the left. It is therefore probable that the
thickening on the partly preserved left side of MUSM 487 was even greater than that of the right side.
Be that as it may, the thickening observed in Furcacetus does not produce an individualized crest
forming an acute angle in cross section as observed in the platanistids Platanista, Pomatodelphis, and
Zarhachis. As in other squalodelphinids, a marked asymmetry characterizes the posteromedial portion
of the ascending processes of the maxillae lateral to the vertex, the right maxilla being significantly
anteroposteriorly longer than the left maxilla. Both posteromedial portions of the maxillae slope steeply
laterally from the vertex, with the right maxilla being almost vertical.
In ventral view (Figure 12), the palatal surface of the rostrum is mainly formed by the maxillae.
It is flat along its anterior two thirds, becoming weakly transversely convex towards the base of the
rostrum. Badly preserved and partly covered by sediment, the lateral portions of the palatal surface
of the maxillae display relatively large alveoli, some of them being filled by broken teeth, as
described in detail below.
Presphenoid and cribriform plate. The nasal septum separating the asymmetrical nares is well
ossified and elevated (Figure 11). Its posterodorsal margin (medial portion of the cribriform plate)
is in contact with the right nasal.
Nasal. Only the right nasal is preserved. In dorsal view (Figure 11), it is weakly inflated,
rectangular, with a main axis that is obliquely oriented in respect to the frontal plane of the skull.
Similar features are observed in Huaridelphis, Macrosqualodelphis, and Notocetus. In lateral view
(Figure 13a,b), the nasal of Furcacetus appears to slope anteriorly (at least in its anterior portion) as
in Huaridelphis, but not in Macrosqualodelphis and Notocetus, having the dorsal surface of the nasal
roughly horizontal. The lost left nasal was probably smaller than the right, judging by the size and
the shape of the depressed area between the right nasal and the medial margin of the posterior
portion of the left premaxilla. If this interpretation is correct, the asymmetry of the nasals in
Furcacetus is opposite to the asymmetry observed in the other squalodelphinids, all of them having
the left nasal slightly larger than the right. The nasal–frontal suture is sinusoidal, with a weak
anteromedial convexity, resembling the holotype of Huaridelphis raimondii MUSM 1396 and differing
from Macrosqualodelphis (suture roughly straight), Notocetus (suture distinctly anteromedially
pointed), and Medocinia (suture posteromedially pointed).
Frontal. At the vertex, the lateral margin of the right frontal is longitudinally more elongated
than the left (Figure 11), a condition shared with Dilophodelphis, Huaridelphis, Notocetus, and
Squalodelphis. The medial suture between the frontals at the vertex is not discernible. In lateral view
(Figure 13a,b), the dorsal surface of the frontals at the vertex appears horizontal, as in Notocetus, but
not in Huaridelphis and Macrosqualodelphis, both having frontals anteroventrally sloping.
On the anterolateral portion of the neurocranium, the frontal is widely exposed dorsally, since
the maxilla only partially covers the supraorbital and preorbital processes, a condition also observed
Life 2020, 10, 27 32 of 62
in Dilophodelphis, among other squalodelphinids. In lateral view both these processes appear
dorsoventrally thickened. The thickening of the preorbital process is greater than in all other
squalodelphinids except Dilophodelphis, Medocinia, and the possibly related skull USNM 475596
[24,126]. A more developed preorbital process is also seen in the platanistids Pomatodelphis and
Zarhachis. The postorbital process of the frontal of Furcacetus is robust and trapezoidal in lateral
view, similar to that of Notocetus. The medial portion of the ventral surface of the frontal of Furcacetus
is excavated in the orbit region by a deep and obliquely oriented fossa (Figure 12). A similar fossa,
probably corresponding to an extension of the pterygoid sinus in the orbit region, has been observed
in all other Platanidelphidi.
Supraoccipital. The supraoccipital is poorly preserved. The eroded nuchal crest is roughly
straight in dorsal view (Figure 11). The supraoccipital slopes posteriorly from the vertex with its
posterodorsal surface drawing, in lateral view, an angle of ca 50° with the horizontal plane (Figure
13a,b). A similar inclination of the supraoccipital is observed in Notocetus, whereas Huaridelphis and
Macrosqualodelphis display a lower inclination and Squalodelphis an almost vertical supraoccipital.
Palatine. The palatines are not discernible on the ventral surface of the skull (Figure 12). Their
anterior portions are probably fully covered by the pterygoids, as generally observed in other
Platanidelphidi. Lateral to the pterygoids, the ventral surface of the skull is damaged and partially
covered by sediment; it does not show any trace of the lateral exposure of the palatines.
Pterygoid. On the posterior palatal surface of the rostrum (Figure 12), the right and left
pterygoids are medially sutured and each is excavated by a pterygoid sinus fossa extending about
30 mm anterior to the right antorbital notch. An elongated hamular fossa of the pterygoid sinus,
extending anterior to the rostrum base, is a derived feature shared by all platanistoids. The lateral
lamina of the pterygoid runs posterolaterally, reaching the anterior margin of the falciform process
of the squamosal.
Jugal–Lacrimal. There is no trace of the jugals and lacrimals. Nevertheless, the ventral surface
of the well-preserved preorbital process of the right frontal is marked by a deep oblique groove that
represents the suture for the missing lacrimal (Figure 12). This suture indicates that the lacrimal was
anteroposteriorly narrow along the lateral wall of the antorbital notch, as in all other platanistoids.
Squamosal. In lateral view (Figure 13a,b), the zygomatic process of the squamosal is robust
and displays a convex dorsal margin, two features shared with all Platanidelphidi. More
specifically, the dorsal margin of the zygomatic process of Furcacetus is more similar to that of
Medocinia, contrasting with those, more regularly arched, of other squalodelphinids. The
anterodorsal surface of the zygomatic process of Furcacetus tightly contacts the postorbital process
of the frontal, a feature related to the anterodorsal extension of the zygomatic process and shared with
all platanistoids and eurhinodelphinids. The anteroventral margin of the zygomatic process it not
preserved in the holotype of Furcacetus flexirostrum. The anteroposterior elongation of the process is
significant, similar to that observed in Macrosqualodelphis and Notocetus, but not in Dilophodelphis,
Huaridelphis and Ensidelphis, all having a shorter zygomatic process. In ventral view, posteromedial to
the mandibular fossa a longitudinally elongated tympanosquamosal recess is visible. The falciform
process projects anteromedially and articulates with the lateral lamina of the pterygoid.
Parietal. In lateral view (Figure 13a,b), the parietal is widely exposed in the temporal fossa.
There is no trace of the peculiar temporal swelling observed in Ensidelphis.
Exoccipital. The whole left exoccipital is lost and the right paroccipital process is badly
preserved (Figure 13 c). A shallow dorsal condyloid fossa is visible dorsolateral to the broken right
occipital condyle.
Basioccipital. The basioccipital is damaged, but the right basioccipital crest is partially
preserved (Figure 12). The angle between right and left basioccipital crests in ventral view is
estimated to about 40°.
Vomer. The vomer is exposed in ventral view (Figure 12), posterior and medial to each choana,
and on the palatal surface of the rostrum, between the maxillae, for a tract extending from roughly
100 to 150 mm anterior to the right antorbital notch.
Life 2020, 10, 27 33 of 62
Periotic. The incomplete right periotic is preserved in articulation with the corresponding
squamosal (Figure 12). The posterior process is lost and the pars cochlearis is damaged. As in all
other squalodelphinids, the anterior process is elongate and not transversely thickened; it does not
display the peculiar marked anteromedial bending that characterizes platanistids, and its ventral
surface is excavated by an elongate and deep anterior bullar facet. The posterior portion of the pars
cochlearis is lost, but its well-preserved anterior wall is rectilinear, suggesting that originally this
part of the periotic was square-shaped in dorsal and ventral view, as in all other squalodelphinids.
Dentition
The estimated tooth count for each upper quadrant is 25 (Figure 12). Excluding the longirostrine
platanistoids, all having smaller and more numerous teeth, the tooth count of Furcacetus is slightly
higher than in Notocetus (18−23), significantly higher than in Squalodelphis (15), and lower than in
Araeodelphis (ca 50), Dilophodelphis (ca 35), and Huaridelphis (28−30).
The transverse diameter of the preserved alveoli ranges from 4.5 to 7.2 mm and the spacing
between the alveoli increases posteriorly; the premaxillary alveoli are almost in contact and the
alveoli at rostrum mid-length are about 15 mm from each other. The posterior alveoli are either not
preserved or covered by hard sediment. The ratio between the maximum transverse width for alveoli
at rostrum mid-length and the estimated BZW is 0.025, a value similar to Huaridelphis (0.026) and
smaller than in Macrosqualodelphis (0.042) and Notocetus (0.040). Judging by the marked oblique
orientation of the axis of the three roots embedded in the premaxillae, the incisors (premaxillary
teeth) were anterolaterally procumbent (Figure 13e,f), a condition absent in all other Platanistoidea
s.s. and more similar to Waipatia and squalodontids. The maximum diameter of the root of these incisors
is 7 mm. The posteriormost preserved tooth, located about 100 mm anterior to the right antorbital notch,
displays a small proximal portion of crown (Figure 13g,h). In particular, only the lateral surface of the
crown is well preserved, showing cusp-like cingular nodules and enamel ornamented with longitudinal
striations. The maximum diameter of the root and crown of this tooth is 6.0 and 6.5 mm, respectively.
About 70 mm anterior to this tooth, along the upper alveolar groove, another tooth has a root with a
diameter of 6.0 mm and a crown without accessory denticles (Figure 13a,b).
Notocetus Moreno, 1892
Type and only included species. Notocetus vanbenedeni Moreno, 1892.
Notocetus vanbenedeni Moreno, 1892
Figures 14–17; Tables 3–6
Holotype. MLP 5-5, skull including the mandible but without ear bones; early Miocene of
Chubut Province, Argentina [127–129].
Previously referred specimens. AMNH 9485, skull including the right tympanic bulla,
mandibles and some vertebrae and ribs; lower Miocene, Santa Cruz Province, Argentina [129];
AMNH 29026, fragmentary skull including the squamosal, part of the exoccipital, the right periotic,
tympanic bulla and malleus, several teeth, a scapula and fragments of vertebrae and ribs; lower
Miocene, Chubut Province, Argentina [92]. MUSM 1395, incomplete cranium with associated
periotics and without teeth, and one cervical vertebra of the same animal [25]; Ullujaya, Western Ica
Valley, Ica Region, Peru, from an unknow horizon of the Chilcatay Fm. The age of the Chilcatay Fm
exposed at Ullujaya can be constricted between 19.00 ± 0.25 Ma and 18.02 ± 0.07 Ma (early
Burdigalian) on the basis of two volcanic ash layer samples dated by
40
Ar/
39
Ar [89].
New referred specimens, localities, and ages. MUSM 3896 (Figures 14a–e, 15a,b), an almost
complete skull including the cranium (only the jugals are missing, the dorsal surface of the vertex is
covered by a concretion, and the dorsal surface of the rostrum at mid-length is damaged), fused
mandibles in articulation with the cranium, articulated ear bones, and presumably all upper and
lower teeth in their respective alveoli. Zamaca locality, western Ica Valley, Ica Region, Peru.
Geographic coordinates: 14°37'28.77" S, 75°38'22.92" W; 340 m above sea level. This specimen was
reported in the Zamaca fossil map of Di Celma et al. [22] with the field number ZM 47 and
Life 2020, 10, 27 34 of 62
provisionally referred to Notocetus sp.”. From the Chilcatay Fm, 7 m above the contact with the
underlying Otuma Formation, in the Ct1c facies association of the Ct1 allomember [21,22]. The age of
the Ct1c facies association is constricted between 19.25 ± 0.08 Ma and 19.00 ± 0.28 Ma (early
Burdigalian) on the basis of two volcanic ash layer samples dated by
40
Ar/
39
Ar [89]. More details for
this age are reported above in the horizon and age description of the holotype of Ensidelphis riveroi
found ca 3 m above MUSM 3896 in the same locality.
MUSM 3897 (Figures. 14f–k, 15c–f), a cranium with eroded vertex, lacking jugals and ear bones,
and without teeth in situ; incomplete mandibles fused, but not articulated to the cranium and with
two teeth inside alveoli; and 4 detached teeth, all belonging to the same animal. Zamaca locality,
Western Ica Valley, Ica Region, Peru. Geographic coordinates: 14°37'18.05" S, 75°38'34.55" W; 340 m
above sea level. This specimen was recently discovered and consequently it was not reported in the
previously published Zamaca fossil map [22]. From the Chilcatay Fm, 35 m above the contact with
the underlying Otuma Formation, in the upper portion of the Ct1a facies association of the Ct1
allomember [21,22]. The age of this portion of the Ct1a facies association can be constricted between
19.00 ± 0.25 Ma and 18.08 ± 0.07 Ma (early Burdigalian) on the basis of two volcanic ash layer samples
dated by
40
Ar/
39
Ar [89].
MUSM 1484 (Figures 16 and 17) consists of the left tympanic bulla, the incomplete fused
mandibles, one tooth, the atlas, the right and left humeri, the left radius and the left ulna of a single
individual. Other bones of the same animal, including some vertebrae and ribs are still in the field
(Figure 8h–i in [20]); the cranium is not preserved. Ullujaya locality, western Ica Valley, Ica Region,
Peru. Geographic coordinates: 14°34'50.8" S, 75°38'44.9" W; 325 m above sea level. This specimen was
reported in the Ullujaya fossil map [21] with the field number O4 and provisionally referred to
“Squalodelphinidae indet.” From the Chilcatay Fm, 27.9 m above the base of the exposed section at
Ullujaya, in the Ct1a facies association of the Ct1 allomember [21,22]. The age of the Chilcatay Fm
exposed at Ullujaya can be constricted between 19.00 ± 0.25 Ma and 18.02 ± 0.07 Ma (early
Burdigalian) on the basis of two volcanic ash layer samples dated by
40
Ar/
39
Ar [89].
Brief Description and Comparison of the Newly Referred Specimens
MUSM 3896 and MUSM 3897 fall in the size, shape, and measurements ranges of variation of
the skulls previously referred to Notocetus vanbenedeni (Table 3). In fact, both skulls show the
diagnostic characters of N. vanbenedeni as redefined by Bianucci et al. [25] for the description of
MUSM 1395. The only significant difference is observed in the tooth count for the upper quadrant of
the skull MUSM 3897, being smaller (18) and out of the range of variation (21–23) of the two
Argentinian specimens (tooth count unknown in MUSM 1395, 1484, and 3896). Associated with a
larger size of the alveoli, this difference could be due to intraspecific variation as observed in the
extant Platanista gangetica, for which the tooth count of the upper quadrant varies from 26 to 39 [23].
This new Zamaca material confirms that Notocetus was a squalodelphinid with a powerful
feeding apparatus characterized by robust mandibles, large, interlocking, and moderately heterodont
teeth, and anteroposteriorly elongated temporal fossa and zygomatic process of the squamosal. These
characters are particularly evident in MUSM 3896 (Figure 14a–e), the only skull of the species that
preserves the complete mandibles firmly articulated to the cranium (and probably all the teeth and
ear bones in place). For these features Notocetus is found to be intermediate between the slightly
smaller Furcacetus and the significantly larger Macrosqualodelphis.
Life 2020, 10, 27 35 of 62
Figure 14. Skulls of Notocetus vanbenedeni from the lower Miocene Chilcatay Fm (Zamaca, Pisco Basin,
Peru). (ae), cranium and articulated mandibles (MUSM 3896) in dorsal (a), posterior (b), ventral (c),
and lateral (d) views, and detail of the left tympanic bulla in ventral view (e); (fk), cranium and
mandibles (MUSM 3897) in dorsal (f, g), ventral (g,h), and left lateral (j,k) views. In (e) the linear
hatching indicates a major break.
On the ventral surface of the tympanic bulla of MUSM 3896, the wide and deep median furrow
extends clearly on the anterior spine (Figure 14e). Such an anterior extension of the median furrow is
present in N. vanbenedeni AMNH 29026 from Argentina and is considered a synapomorphy of the
family Squalodelphinidae. By contrast, the median furrow of the well-preserved tympanic bulla of
MUSM 1484 apparently does not extend anteriorly along the anterior spine (Figure 16a–d). However,
some longitudinal striations on the ventral surface of the anterior spine could represent spongy bone
filling the anterior portion of the median furrow, suggesting that this character could be subject to
intraspecific variation, possibly due to ontogenesis. Both tympanic bullae of MUSM 1484 and MUSM
3896 display an elongated anterior spine (although incomplete in MUSM 1484) associated with a
marked anterolateral convexity, a derived character observed in all platanistoids. Moreover, the inner
and outer posterior prominences of both tympanic bullae have roughly the same posterior extent, a
condition shared with the squalodelphinids Notocetus, Phocageneus, and Squalodelphis. In lateral view
the tympanic bulla of MUSM 1484 is almost identical to that of Notocetus vanbenedeni AMNH 29026,
figured by Muizon ([92], Figures. 4a,7a).
Life 2020, 10, 27 36 of 62
Table 4. Measurements on detached teeth of Notocetus vanbenedeni (MUSM 3897, 1484) from the
Chilcatay Fm (early Miocene, Peru). All measurements are in mm.
Dimension
MUSM 3897
MUSM 1484
a
b
c
d
Total length
31.5
+28.5
+25.8
25.0
28.6
Root length
27.2
+24.8
+19.9
+19.8
19.7
Crown length
9.4
+7.3
+2.0
5.5
9.2
Maximum transverse diameter of root
8.0
7.6
8.5
7.6
7.1
Maximum mesiodistal diameter of root
10.2
9.4
10.8
10.0
6.2
Transverse diameter at crown base
6.9
6.8
5.9
5.5
7.1
Mesiodistal diameter at crown base
6.5
6.8
5.6
6.8
6.2
+, incomplete.
Figure 15. Teeth of Notocetus vanbenedeni from the lower Miocene Chilcatay Fm (Zamaca, Pisco Basin,
Peru). (a,b), details of the skull MUSM 3896 in right (a) and left (b) lateral view showing the lower
and upper teeth in place; (cf), two detached teeth associated with the skull MUSM 3897 in lateral
(c,d) and posterior (e,f) views.
The preservation of a complete set of teeth implanted in their alveoli in MUSM 3896 (Figure
15a,b), together with the four detached teeth of MUSM 3897 (Figure 15c–f; Table 4), is relevant since,
to date, teeth of Notocetus were only known in the fragmentary specimen AMNH 29026 [92]. The
Peruvian teeth confirm the observations made by Muizon [92] on AMNH 29026: the two posterior
teeth embedded in the left mandible of MUSM 3896 have a low triangular crown with two-three small
posterior accessory denticles and cusp-like cingular nodules on the lateral surface; moving to the
anterior portion of the mandibles and rostrum of MUSM 3896, tooth crowns become gradually
higher, forming a conical point, slightly recurved posteriorly, and lacking the posterior denticles;
their enamel is ornamented only by thin longitudinal striations. The four detached teeth of MUSM
3897 have a fusiform root, proportionally large compared to the crown. The two best-preserved teeth
have a low triangular crown with weak anterior and posterior carinae and several cusp-like cingular
nodules and papillae forming a cingulum near the base of the crown. One of these two teeth displays
a wide wear surface on the posterolingual surface of the crown, probably due to the contact with the
opposite tooth (attritional wear facet). The same tooth has a peculiar swelling on the posterior surface
of the root. The only preserved tooth of MUSM 1484 (Figure 16h) fully overlaps in size and shape
Life 2020, 10, 27 37 of 62
with the described anterior teeth of N. vanbendeni [92]. This tooth is fusiform, weakly curved, and its
crown is conical, weakly mesiodistally compressed, and bearing a posterior carina. The enamel is
ornamented by thin anastomosed longitudinal grooves and ridges, more marked on the labial surface
of the crown, where an ectocingulum is also visible.
Figure 16. Fragmentary specimen (MUSM 1484) of Notocetus vanbenedeni from the lower Miocene
Chilcatay Fm (Ullujaya, Pisco Basin, Peru). (ad), left tympanic bulla in dorsal (a), ventral (b), medial
(c), and lateral (e) views; (eg), fragment of left mandible in lateral (e) and dorsal (f, g) views; (h),
tooth in medial view; (i,j), atlas in anterior (h) and posterior (i) views. Abbreviations: ipp, inner
posterior prominence; opp, outer posterior prominence. Linear hatching indicates major breaks.
As in the mandibles of the Argentinian specimens, in MUSM 3896 (Figure 14a,c,d), MUSM 3897
(Figure 14g,i,k), and MUSM 1484 (Figure 16e–g) the mandibular symphysis is fused, the symphyseal
portion is markedly dorsoventrally flattened, and, as in Huaridelphis and Squalodelphis, lacking the
pair of lateral grooves observed in members of the families Allodelphinidae and Platanistidae, and
in the basal Platanidelphidi Ensidelphis. The symphyseal portion of the complete mandibles of MUSM
3896 represents 40% of the total mandibular length, measured parallel to the sagittal plane, a value
close to Squalodelphis (42%), the only other squalodelphinid for which complete mandibles are known.
The angle between the two mandibles equals 38° in both MUSM 3896 and MUSM 3897, similar to
Squalodelphis but significantly smaller than in all platanistids (> 50°). Embrasure pits are observed
between and lateral to the alveoli on the mandibles of MUSM 3897 (Figure 14g) and MUSM 1484
(Figure 16f–g). In lateral view, posterior to the alveolar row, the robust ramus (preserved in MUSM
3896 and MUSM 3897) raises posterodorsally and the coronoid process is significantly elevated. The
angular process is prominent, separated by a wide notch from the mandibular condyle.
Life 2020, 10, 27 38 of 62
The almost complete atlas of MUSM 1484 (Figure 16i,j) is close in shape to the atlas of the
specimen of N. vanbenedeni AMNH 9485, described by True [125], in the extreme reduction of the
ventral transverse processes, the neural arch being proportionally lower, and the roughly circular
neural canal.
Table 5. Measurements on the forelimb bones of Notocetus vanbenedeni (MUSM 1487) compared with
the holotype of Macrosqualodelphis ukupachai (MUSM 2545), both from the Chilcatay Fm (early
Miocene, Peru). All measurements are in mm.
Dimension
Notocetus vanbenedeni
MUSM 1484
Macrosqualodelphis ukupachai
MUSM 2545
Humerus
Total proximodistal length
128
160
Anteroposterior width at distal end
51
79
Transverse width at distal end
31
52
Anteroposterior width at mid-length
48
71.
Transverse width at mid-length 26 54
Anteroposterior width of the head
47
64
Transverse width of the head
44
62
Radius
Total proximodistal length
83
129
Transverse width at mid-length
42
60
Ulna
Total proximodistal length
63
-
Transverse width at mid-length
31
-
+, incomplete; - missing data; e, estimate.
Description of the Forelimb Bones of MUSM 1484
The humerus, radius, and ulna, preserved in MUSM 1484 (Figure 17; Table 5), are here described
in detail because up to now these bones were unknown in Notocetus vanbenedeni. Among other
squalodelphinids the forelimb was described only in Macrosqualodelphis ukupachai, whereas among
other platanistoids these bones are known in the extant Platanista gangetica and in the allodelphinids
Allodelphis pratti, A. woodburnei, Goedertius oregonensis, and Zarhinocetus errabundus ([121], Figure 39).
Comparisons (Figure 18; Table 6) were made also with the few other archaic odontocetes having some
of these bones preserved (Awamokoa tokarahi, Kelloggia barbara, Otekaikea huata, Schizodelphis sp.,
Squalodon bellunensis, S. calvertensis, and with derived archaeocetes (e.g., Cynthiacetus peruvianus).
Table 6. Measurement ratios of the forelimb bones of Notocetus vanbenedeni (MUSM 1487) compared
with those of other platanistoids, some archaic odontocetes, and the archaeocete Cynthiacetus peruvianus.
Species Inv. Number B/A D/C F/E C/A E/A
Notocetus vanbenedeni MUSM 1484 0.41 0.51 0.47 0.69 0.51
Macrosqualodelphis
ukupachai
MUSM 2545 0.46 0.49 - 0.77 -
Platanista gangetica
MNHN-ZM-
2018-2918
0.40
0.68
0.87
0.56
0.55
Platanista gangetica IRSNB 1507 0.41 0.74 0.74 0.52 0.55
Platanista gangetica MSNUP M272 0.36 0.90 0.72 0.50 0.47
Allodelphis pratti UCMP 83791 0.29 - 0.29 - 0.71
Allodelphis woodburnei SBCM L3210-1 0.31 - - - -
Ninjadelphis ujiharai GMNH-PV-2570 0.28 0.24 0.25 0.90 0.87
Goedertius oregonensis LACM 123887 0.30 - - - -
Zarhinocetus errabundus LACM 21031 0.26 - - - -
Xiphiacetus bossi USNM 8842 0.45 - - - -
Awamokoa tokarahi OU 22125 - - 0.24 - -
Squalodon bellunensis MGP-PD 26114 - 0.38 0.34 - -
Squalodon calvertensis USNM 10484 - - 0.34 - -
Life 2020, 10, 27 39 of 62
Kelloggia barbara GAS IP S2/S6 0.42 0.40 0.42 0.69 0.60
Waipatiidae indet. NMV P48861 0.37 0.25 - 0.96 -
Otekaikea huata OU 22306 0.44 0.32 0.37 0.73 0.67
Cynthiacetus peruvianus MNHN.F.PRU10 0.26 0.19 0.21 0.60 0.57
Humerus. The humerus of MUSM 1484 (Figure 17) is similar in shape but smaller than in
Macrosqualodelphis ukupachai; among platanistoids it shares only with M. ukupachai the large,
hemispherical and posterolaterally protruding head, the lesser tubercle being higher than the head,
the salient and distally elongated deltopectoral crest, and the large and deep fossa for insertion of M.
infraspinatus. In particular, the large, hemispherical and posterolaterally protruding head may be a
diagnostic character of squalodelphinids since it was not observed in the humerus of any other
cetacean, whereas the lesser tubercle being higher than the head is also observed in Kelloggia barbara
[130]. Moreover, the humeri of MUSM 1484 and M. ukupachai are clearly anteroposteriorly wider than
allodelphinid humeri. In fact, the ratio between the anteroposterior width at mid-length and the
proximodistal length (B/A in Table 6) is 0.41 and 0.46 in MUSM 1848 and M. ukupachai, respectively,
whereas in allodelphinids it ranges from 0.26 in Zarhinocetus errabundus to 0.31 in Allodelphis woodburnei.
A humerus as robust as in MUSM 1848 and M. ukupachai is instead observed in Platanista gangetica (B/A
= 0.36–0.41), although the humerus of the only extant platanistid differs from all other platanistoids in
lacking the delctopectoral crest and in its pronounced distal anteroposterior widening.
Figure 17. Fragmentary specimen (MUSM 1484) of Notocetus vanbenedeni from the lower Miocene
Chilcatay Fm (Ullujaya, Pisco Basin, Peru). (a,b), right humerus in lateral (a) and anterior (b) views;
(c), left humerus, radius, and ulna in lateral view. Linear hatching indicates major breaks.
Life 2020, 10, 27 40 of 62
Figure 18. Comparison of the shape of the left humerus, radius, and ulna of Notocetus vanbenedeni with
the squalodelphinid Macrosqualodelphis ukupachai, the extant platanistid Platanista gangetica, the
allodelphinids Allodelphis pratti and Ninjadelphis ujiharai (redrawn from [121]), the waipatiid Otekaikea
huata (redrawn from [104]), an indeterminate waipatiid (redrawn from [131]), and the squalodontid
Kelloggia barbara (redrawn from [130]) in lateral view. Scale bars equal 5 cm.
Radius. The radius of MUSM 1484 (Figure 17c) is a mediolaterally flat bone that distally widens
anteroposteriorly, more so than in M. ukupachai but to a lesser degree than in P. gangetica. The
proximodistal length of the radius of MUSM 1484 is significantly smaller than the proximodistal length
of the humerus, with a ratio (C/A in Table 6) = 0.69, lower than in M. ukupachai (0.77), but higher than
in P. gangetica (0.50–0.56), whereas in the allodelphinid Ninjadelphis ujiharai the two bones almost have
the same length (C/A = 0.96).The ratio between the anteroposterior width at mid-length and the
proximodistal length (D/C) is relatively high (0.51), if compared to N. ujiharai (0.24) and some archaic
odontocetes, whereas it is similar to M. ukupachai (0.49) and lower than in P. gangetica (0.68–0.74).
The radius of MUSM 1484 is articulated with the ulna only for a small proximal portion of its
posterior margin; distal to this articulation, the interosseous space between the radius and ulna is
very broad, a condition similarly observed in M. ukupachai and P. gangetica. The articular surface for
the first carpals is distally convex and arched in lateral view, as in P. gangetica.
Ulna. The ulna of MUSM 1484 (Figure 17c), is proximodistally shorter than the radius and, as in
M. ukupachai and P. gangetica, at its mid-length it is anteroposteriorly narrower than the radius. Like
the radius, it is mediolaterally flattened and the ratio between its anteroposterior width at mid-length
and its proximodistal length (F/E in Table 6) has a value (0.51) higher than in allodelphinids and other
archaic odontocetes, but lower than in P. gangetica (0.72–0.87) (ratio not computable for the ulna of
M. ukupachai due to its incompleteness). The olecranon of the ulna of MUSM 1484 is even smaller
than in M. ukupachai, clearly less developed anteroposteriorly and proximally than in allodelphinids
(plesiomorphic condition), whereas the olecranon is completely lacking in Platanista.
cf. Notocetus sp.
Figure 19
Referred specimen, locality, and age. MUSM 1485 consists of a right isolated tympanic bulla.
Ullujaya locality, Western Ica Valley, Ica Region, Peru. Exact locality and stratigraphical horizon
unknown. Approximate geographic coordinates: 14°35’ S, 75°38’ W. The age of the Chilcatay Fm
exposed at Ullujaya can be constricted between 19.00 ± 0.25 Ma and 18.02 ± 0.07 Ma (early
Burdigalian) on the basis of two volcanic ash layer samples dated by
40
Ar/
39
Ar [89].
Life 2020, 10, 27 41 of 62
Figure 19. Right tympanic bulla (MUSM 1485) of cf. Notocetus sp. from the lower Miocene Chilcatay
Fm (Ullujaya, Pisco Basin, Peru) in dorsal (a), ventral (b), medial (c), and lateral (d) views.
Abbreviations: ipp, inner posterior prominence; opp, outer posterior prominence.
Brief Description and Comparison
Like MUSM 1484 described above this tympanic bulla is similar in size and shape to those
referred to Notocetus vanbenedeni. The median furrow (Figure 19b) is only moderately deep, but it
extends anteriorly on the preserved portion of the broken anterior spine, a feature that, as outlined
above, has been observed in all squalodelphinid tympanic bullae (Phocageneus, Notocetus, and
Squalodelphis). Here also, as in the other squalodelphinids, the inner and outer posterior prominences
have approximately the same posterior extent. Considering the incompleteness of this specimen, we
cautiously refer it to cf. Notocetus sp.
Platanistidae Gray, 1846
Emended diagnosis. The Platanistidae are characterized by the following synapomorphies,
absent in the other members of the Platanidelphidi clade: (1) extremely elongated mandibular
symphysis (>65% of the total length of the mandible) and wide angle (>50°) between the mandibular
rami; (2) distinct dorsal crest in the antorbital-supraorbital region (character absent in Araeodelphis);
(3) hook-like articular process on the lateral surface of the periotic (periotic unknown in Araeodelphis);
(4) outer posterior prominence of the tympanic bulla posteriorly longer than the inner posterior
prominence (tympanic bulla unknown in Araeodelphis).
Platanistidae indet. aff. Araeodelphis
Figure 20; Table 7
Referred specimen. MUSM 631, fragmentary skull consisting of rostrum and fused symphyseal
portion of mandibles.
Locality and horizon. Western Ica Valley, Ica Region, Zamaca, Peru, Chilcatay Fm. Exact locality
and stratigraphical horizon unknown. Approximate geographic coordinates: 14°37’ S, 75°38’ W. The
entire stratigraphical sequence of the Chilcatay Fm exposed at Zamaca can be roughly constricted
between 19 and 18 Ma (early Burdigalian), considering that a volcanic ash layer located 4 m above
the contact between the Chilcatay Fm and the underlying Otuma Formation in the Zamaca area gave
an Ar/Ar age of 19.25 ± 0.05 Ma, and a volcanic ash layer from the nearby locality of Ullujaya, located
below the contact between the Chilcatay and Pisco formations, gave an age of 18.02 ± 0.07 Ma [89].
Life 2020, 10, 27 42 of 62
Table 7. Measurements on the rostrum and associated mandibles of aff. Araeodelphis MUSM 631 from
the Chilcatay Fm (early Miocene, Peru). All measurements are in mm.
Length of the rostrum as preserved
500
Length of left upper tooth row
440
Number of teeth per upper tooth row
55
Length of the mandibles as preserved 455
Length of symphyseal portion of mandibles
415
Width of mandibles at posterior end of symphysis
57
Height of mandible at posterior end of symphysis
22
Brief Description and Comparison
MUSM 631 shows marked affinities with the rostrum and associated mandibles of the early
diverging platanistid Araeodelphis natator (holotype USNM 10478 and referred cranium USNM 526604
[108,132]), from the late early Miocene (Burdigalian) of Maryland (USA). Shared characters between
MUSM 631 and A. natator are the following:
1) Rostrum and mandibles elongated and narrow, with lateral margins parallel in dorsal view
(Figure 20a,b), and strongly dorsoventrally compressed in lateral view (Figure 20e,f),
especially in their anterior half;
2) premaxilla fused to the maxilla in the anteriormost portion of the rostrum;
3) premaxilla-maxilla suture along the rostrum outlined by a distinct sulcus (more excavated in
the anterior half of the rostrum), but without the deep lateral groove featuring more derived
platanistids;
4) mesorostral canal dorsally closed or very narrow for the whole rostrum length;
5) oblique sulci on the dorsal surface of the premaxillae; similar sulci are also present in the
longirostral odontocete Chilcacetus, but not in other platanistoids;
6) firmly ankylosed mandibular symphysis being extremely elongated (Figure 20g–j), a
distinctive character of the platanistids within the platanistoids; indeed, even if it is not
possible to calculate the ratio between the length of the symphysis and the total length of the
mandible, the extreme elongation of MUSM 631's symphysis is evidenced by the fact that the
posterior end of the symphysis almost reaches the pterygoid-maxilla suture;
7) ventral surface of the symphyseal portion of each mandible marked by a deep longitudinal
groove; this character is also present in all other platanistoids with the exception of
squalodelphinids;
8) symphyseal portion of the mandible bearing a similar number of teeth: 39–40 for MUSM 631
and 38 for USNM 10478;
9) similar tooth count for each upper quadrant (Figure 20g,h); in fact Godfrey et al. [108] estimated
approximately 50 teeth for each quadrant on the rostrum of USNM 526604, a number close to
that counted (55) for the almost complete rostrum of MUSM 631; USNM 10478 has 47 alveoli
for each upper quadrant, but the rostrum is not as complete as in MUSM 631;
10) presence of a medial trough between the maxillae on the palatal surface of the rostrum.
Life 2020, 10, 27 43 of 62
Figure 20. Rostrum and associated mandible (MUSM 631) of Platanistidae indet. aff. Araeodelphis from
lower Miocene of Chilcatay Fm (Zamaca, Pisco Basin, Peru). (a,b), rostrum in dorsal view; (c,d),
rostrum in ventral view; (e,f), rostrum in left lateral view; (g,h), mandibles in dorsal view; (i,j),
mandibles in ventral view; (k), mandibles in right lateral view. Linear hatching indicates major breaks
and beige shading reconstructed missing part.
Significant differences between MUSM 631 and Araeodelphis natator are:
1) the size, MUSM 631 being larger than USNM 10478 (the length of the mandibular
symphyseal portion is 415 mm in MUSM 631 contra 291 mm in USNM 10478; the transverse
Life 2020, 10, 27 44 of 62
width of the fused mandibles at the posterior end of the symphysis is 57 mm in MUSM 631
contra 50 mm in USNM 10478) (Table 7); A. natator USNM 526604 is even smaller than the
holotype (approximately 10%–15% smaller, according to Godfrey et al. [108]);
2) the mandibular alveoli being proportionally larger and not as many, compared to the upper
alveoli in MUSM 631, contra the same size and number of alveoli in lower and upper
quadrant in A. natator USNM 10478; in fact the transverse diameters of alveoli in MUSM 631
range between 4 and 7 mm in the mandible and 2 and 5.5 mm in the rostrum and there are
40 alveoli in the symphyseal portion of the mandible, for about 53 in the corresponding
rostral portion;
3) the angle formed by the mandibular rami is apparently more acute in MUSM 631 than in A.
natator USNM 10478; the wide angle (>50°) between the mandibles is an important feature
related to the extreme elongation of the symphysis shared by all platanistoids, including
Araeodelphis; however, the two mandibular rami of MUSM 631 are fragmentary and show
some post-mortem fractures that could have changed their original orientation; it is therefore
probable that the two mandibles originally formed a wide angle, as expected considering the
very long symphysis; and
4) on the dorsal surface of the rostrum of A. natator USNM 526604 the lateral margin of the
premaxilla is laterally convex near the antorbital notch, a feature not observed in MUSM 631;
however this convexity of the premaxilla could have been originally present in the missing
posterior portion of the rostrum of MUSM 631 (the same could be true for A. natator USNM
10478, apparently lacking a premaxillary convexity, but with a rostrum incomplete at its base).
Considering the affinities and differences above mentioned, it is possible that MUSM 631 either
belongs to: (1) A large specimen of A. natator, considering that the size of the mandible and of the
alveoli are subject to ontogenetic variation; (2) an undescribed, larger species of Araeodelphis; or (3)
an unknow genus and species of basal platanistid. Pending the discovery of a more complete
specimen, MUSM 631 is here assigned to Platanistidae indet. aff. Araeodelphis.
5. Phylogeny
The cladistic analysis produced 12 equally parsimonious trees, with tree length = 98, consistency
index (CI) = 0.60, and retention index (RI) = 0.83. The strict consensus tree, the bootstrap values, and
the Adams consensus tree are presented in Figure 21.
Beside the addition of new taxa, the topology here obtained does not differ significantly from
those published by Lambert et al. [24], Godfrey et al. [108], Kimura [133], and Bianucci et al. [26]. In
fact, in all these analyses the clade that was previously informally named ‘homodont platanistoids’
and that is here redefined as the superfamily Platanistoidea includes three monophyletic families: the
Allodelphinidae in the basalmost position, and the sister group-related Platanistidae +
Squalodelphinidae. Similarly, our analysis confirms the position of the Eurhinodelphinidae
branching between the more basal Squalodon + Waipatia and the Platanistoidea, supporting the
paraphyly of the Platanistoidea as defined by Muizon [92] and Fordyce [99]. The monophyly of the
Platanistoidea as redefined here is supported by a bootstrap value of 72 and by
the following
synapomorphies: (1) Vertex distinctly shifted to the left compared to the sagittal plane of the skull
(char. 14, state 1; reversal to state 0 in Allodelphis and Ninjadelphis); (2) long hamular fossa of the
pterygoid sinus extending anteriorly on the palatal surface of the rostrum (char. 17, state 1); (3)
presence of an articular rim on the lateral surface of the periotic (char. 20, states 1,2); (4) elongated
anterior spine on the tympanic bulla, associated with a marked anterolateral convexity (char. 27,
states 1,2); and (5) great reduction of the coracoid process of the scapula and the acromion being
located on the anterior edge of the scapula (char. 36, state 1; also present in Squalodon).
Within the Platanistoidea, the clade Platanistidae + Squalodelphinidae forms, together with the
new genus Ensidelphis and the possibly congeneric MUSM 603, the Platanidelphidi. This new clade is
supported by a bootstrap value of 100 and by the following synapomorphies: (1) Asymmetry of the
premaxillae on the rostrum at some distance anterior to the premaxillary foramina, with the right
Life 2020, 10, 27 45 of 62
premaxilla being distinctly narrower than the left in dorsal view (char. 4, state 1; reversal to state 0 in
Araeodelphis); (2) posterior dorsal infraorbital foramen(ina) along the vertex more medial than the
lateralmost margin of the premaxilla on the cranium (char. 12, state 1); (3) deep fossa in the frontal
on orbit roof, at the level of the frontal groove (presumably for orbital lobe of pterygoid sinus) (char.
13, state 1); (4) palatines not contacting each other on the sagittal plane and displaced dorsolaterally
(char. 16, states 1,2); (5) thick zygomatic process of the squamosal (ratio between the maximum
distance from the anteroventral margin of the zygomatic process to the posterodorsal margin, in
lateral view, and the vertical distance from the lower margin of the occipital condyles to the vertex of
the skull > 0.35) (char. 18, state 1); (6) dorsal outline of the zygomatic process of the squamosal in
lateral view dorsally convex (char. 19, states 1,2; also present in Squalodon and in some specimens of
Xiphiacetus); and (7) ventral edge of the zygomatic process of the squamosal almost straight or convex
in lateral view (char. 42, state 1).
Figure 21. Results of the main phylogenetic analysis showing the relationships of Ensidelphis and
Furacetus with the other Platanistoidea. (a), consensus tree of 12 equally parsimonious trees, with tree
length = 98, consistency index (CI) = 0.60 and retention index (RI) = 0.82; numbers associated with the
nodes are bootstrap values; (b), Adams consensus tree.
The new phylogeny here presented supports the referral of Furcacetus to the Squalodelphinidae,
although the relationships within this family remain poorly resolved, as in previous analyses
[24,26,108]. However, the Adams consensus tree shows a more satisfactory result, with Furcacetus in
a derived position among squalodelphinids, forming a clade together with Huaridelphis and Medocinia
+ Squalodelphis. Interestingly enough, in our new analysis Dilophodelphis is placed within the
Squalodelphinidae, instead of Platanistidae as proposed by Boersma et al. [101] and Bianucci et al. [26].
Also recovered by Kimura [133], this different familial attribution could be due here to some changes
of character states when re-examining the holotype of this genus and to the reformulation of the
character 9 dealing with the dorsal crest in the antorbital-supraorbital region. In fact, we consider the
prominent dorsal swelling in the antorbital-supraorbital region characterizing the cranium of
Life 2020, 10, 27 46 of 62
Dilophodelphis homologous to the similar, although less developed thickening observed in other
squalodelphinids (e.g., Squalodelphis). Such a thickening does not generate a true crest, forming an acute
angle in cross section, as observed instead in the platanistids Platanista, Pomatodelphis, and Zaharachis.
6. Shape, Function, and Ecology of the Platanistoids of the Chilcatay Fm
The new fossil specimens described here indicate that the diversity and disparity of the
platanistoids living along the Peruvian coast during the early Miocene was twice as high as
previously pointed out [26]. For a short time interval, well-defined chronostratigraphically between
ca 19 and 18 Ma, and in a restricted geographical area, six distinct monogeneric species recovered in the
families Squalodelphinidae and Platanistidae and more stemward among basal Platanidelphidi have
been described (Figure 22). Concerning morphological disparity, the features showing the most striking
variation are the body size, the rostrum length, and the number and size of the teeth (Figure 23). These
and other characters related to distinct feeding strategies are reported below for each platanistoid of
the Chilcatay Fm, also including the previously described squalodelphinids Marcrosqualodelphis
ukupachai and Huaridelphis raimondii [24,26].
6.1. Ensidelphis riveroi
The basal Platanidelphidi Ensidelphis riveroi was a medium size odontocete (estimated TBL = ca
3 m) characterized by an extremely elongated rostrum (RI = 0.81) bearing a large number (about 64
for each quadrant) of small teeth (transverse width of alveoli at rostrum mid-length = 1.18 % BZW).
It represents a new hyper-longirostrine dolphin, falling in the same category as a series of previously
described taxa with a similar cranial morphology (allodelphinids, eoplatanistids, eurhinodelphinids,
the lipotid Parapontoporia, and pomatodelphinines). Boessenecker et al. [16] and McCurry and
Pyenson [17] outlined that hyper-longirorstrine odontocetes are mainly restricted to the early
Miocene epoch. The description of E. riveroi further supports the radiation of this peculiar
morphotype during this short temporal range. The extremely elongated rostrum, being
dorsoventrally flattened in its anterior portion and combined with a mandible that is as long as the
rostrum are all features shared by E. riveroi with the pomatodelphinines Pomatodelphis inaequalis and
Zarhachis flagellator. McCurry and Pyenson [17] suggested that these two pomatodelphinines stunned
fish with the mouth closed, performing oscillatory rotation movements mainly along the horizontal
plane, like the swordfish Xiphias gladius [134], or used the jaws sweeping through the water to grasp
the prey as the Indian gharial (Gavialis gangeticus). A feeding strategy similar to that of the gharial
was also observed in the Amazon river dolphin Inia geoffrensis [135]. Nevertheless, if the extreme
elongation of the rostrum allows for great angular acceleration and greater speed during sweep
feeding [135], the bite force decreases significantly near the anterior end of elongated jaws [136,137],
probably preventing the capture of large and medium sized prey. Furthermore, considering the
relatively poor maneuverability of a body with such a long rostrum when swimming, we consider a
valid alternative hypothesis to the ones mentioned above that E. riveroi used its long and slender
snout to probe soft sediment for hidden prey, as proposed for the hyper-longirostrine
eurhinodelphinids [3,138].
Life 2020, 10, 27 47 of 62
Figure 22. Evolution of size and morphology of the cranium among Platanistoidea using the Adams
consensus tree of Figure 21b (the outgroup Zygorhiza is not reported). The rostral index is defined as
the ratio between the rostrum length and the skull length [17]. Chronostratigraphic scale follows
Cohen et al. [139].
Interestingly enough, E. riveroi exhibits a temporal fossa (the area of insertion for the temporalis
muscles, involved in jaw adduction) that is significantly smaller than in Platanista gangetica and I.
geoffrensis, suggesting a lower bite force compared to these extant longirostrine dolphins (see Lambert
[117] for a similar observation among eurhinodelphinids). On the other side, the peculiar
protuberance (temporal swelling) observed on the temporal fossa of E. riveroi may provide a stronger
attachment surface for part of the temporalis muscles subject to strong stress, due to the very
elongated mandible, during the closing of the mouth. Combined with relatively short tooth rows, the
significant length of the post-symphyseal portion of the mandible and the marked posterodorsal
elevation of its dorsal margin (mirrored by a gradual posterodorsal elevation of the lateral margin of
the rostrum) in the skull of E. riveroi results in a long, toothless tubular posterior portion of the oral
cavity that may have favored the transfer of food to the posterior part of the mouth by suction, before
swallowing [140]. Such an elongated post-symphyseal portion of the mandible and associated
relatively short tooth row are probably plesiomorphic features, also observed in allodelphinids [121]
and partially in squalodelphinids, but not in platanistids, all having the mandibular symphysis and
tooth row more posteriorly extended and of which the extant representative P. gangetica is considered
a typical raptorial feeder [141,142]. In addition to the suction feeding hypothesis, a relatively
elongated and robust post-symphyseal portion may have also functioned to strengthen the posterior
portion of such narrow and elongated jaws when the mouth was closed. Finally, the long space
between the post-symphyseal mandibles could also be related to sound reception, delimiting the
“gular pathway” observed for example in the extant Ziphius cavirostris [143].
Life 2020, 10, 27 48 of 62
Figure 23. Comparison of the skulls in right lateral views of the platanistoids from lower Miocene of
Chilcatay Fm (Zamaca, Pisco Basin, Peru). Stippled lines and grey shading indicate reconstructed
missing part for main reconstructed bony parts; green shading for a hypothetical reconstruction of
the soft tissue.
In summary, the above observations support the interpretation that E. riveroi captured prey
using suction-assisted raptorial feeding: (1) either after having stunned it with lateral oscillations of
the rostrum; or (2) after having grasped it when laterally sweeping through the water; (3) or after
flushing it out of soft sediment along the bottom as proposed by Lambert et al. [54] for the
longirostrine ziphiid Ninoziphius platyrostris.
Another significant feature of the skull of the holotype of E. riveroi is the right-side torsion of the
rostrum. Unfortunately, E. riveroi is only known from this single skull and consequently it is not
possible to check if this peculiar morphology is an anomaly present in a single individual, or rather
a distinctive character of the species. In support of the “anomaly” hypothesis, a similar torsion has
been observed in some skulls of P. gangetica [113], Inia geoffrensis [144], and Pontoporia blainvillei
Life 2020, 10, 27 49 of 62
[114,145]. Even if the causes of this and other cranial anomalies observed on the skull of extant "river
dolphins" are still unclear, Gerholdt [145] speculated that a slight twisting due to a trauma in a young
individual of P. blainvillei could increase disproportionally in the adult due to the positive allometric
growth of the rostrum. Having a rostrum significantly longer that P. blainvillei, E. rivireoi, may have
suffered an even greater increase of this anomaly during ontogeny. The hypothesis of the rostral
torsion as a distinctive character of the species is instead partially supported by the observation of a
similar twisting in all squalodelphinid species and specimens, although in all these cases the rostral
torsion processes towards the left side (see below).
The cervical vertebrae of E. riveroi are all free, with their centra anteroposteriorly elongated,
suggesting a significant neck elongation similar to that of P. gangetica. The distinct ventral tubercle
on the ventral surface of the atlas and the deep ventral excavations on the centrum of the axis can be
interpreted as surfaces of insertion for a well-developed M. longus colli, responsible for the flexion
of the neck [50,146,147]. The massive neural spine of the axis is anteroposteriorly thick in lateral view;
it constituted the origin of the M. rectus capitis dorsalis major, the M obliqus capitis caudalis and the
M. multifundus cervicalis that extend and rotate the head in extant cetaceans [50,146,147]. In brief,
the morphology of the cervical vertebrae suggests that E. riveroi had a relatively long neck, possibly
with an even greater flexibility than in P. gangetica, potentially useful for rotational movements of the
snout when searching and/or catching prey. An even more elongated and slender neck was described
in allodelphinids [121].
6.2. aff. Araeodelphis
The fragmentary MUSM 631 specimen here assigned to aff. Araeodelphis consists of a rostrum
and associated mandibles, both being narrow, elongated, dorsoventrally flattened, and bearing ca 40
small single-rooted teeth for each quadrant. An extant analogue of this platanistoid, which is smaller
with a shorter rostrum than E. riveroi, is Pontoporia blainvillei, having a roughly similar rostrum
elongation and a slightly higher tooth count per quadrant (51–58) [148]. P. blanvillei feeds mainly
upon bottom fish and, secondarily, squid and shrimps [148] and we like to imagine a similar feeding
behavior for this small platanistid from Chilcatay Fm.
6.3. Furcacetus flexirostrum
The squalodelphinid Furcacetus flexirostrum had a body length in the range of the extant Inia
geoffrensis (estimated TBL = 2.34 m). Its rostrum is moderately elongated (RI = 0.67), anteriorly tapered
in dorsal view, and dorsoventrally flattened; the transverse width at rostrum mid-length equals
2.92% of BZW. It bears about 25 teeth in each quadrant. The temporal fossa of F. flexirostrum is
moderately voluminous. The most peculiar feature of this new species is its curved, sinusoidal
rostrum in lateral view, combined with large and proportionally larger, probably procumbent, upper
incisors. A similar sinusoidal rostrum, but not associated with large and procumbent anterior teeth,
has been observed in some skulls of Pontoporia blainvillei [114]. Large, but not procumbent anterior
teeth, sometimes combined with an upward but not sinusoidal curvature of the rostrum are also
present in Platanista gangetica [23,149]. Pilleri [141] reported that captive P. gangetica individuals catch
their prey by first securing it in the distal third of the jaws, then maneuvering it toward the throat,
and swallowing it head first. A similar feeding behavior has been observed for I. geoffrensis, taking
fish with the anterior teeth to be transferred to the posterior teeth [150]. Procumbent and large
anterior teeth are also present in the squalodelphinid Squalodelphis fabianii [151] and in several other
extinct odontocetes, including the longirostrine squalodontids (e.g., Squalodon, Eosqualodon, and
especially Kelloggia) and waipatiids (e.g., [99,130,152]). Squalodontids were moderately heterodont
odontocetes that probably used part of their conical and elongate post-apical teeth to grasp and
restrain prey and their triangular, laterally compressed denticulated posterior teeth to slice and shear
[153]. This probably corresponds to the plesiomorphic feeding strategy among odontocetes, also
proposed for the basilosaurids [146]. On the other hand, the anteriormost incisors of squalodontids,
as well as of some kentriodontids [154], are almost horizontal and anteriorly directed. Due to this
Life 2020, 10, 27 50 of 62
peculiar orientation these teeth are unlikely to be used for grabbing prey items. Fordyce [99]
speculated that similar delicate, procumbent, and elongated anteriormost incisors of Waipatia were
used for display, rather than for predation. However, there is no evidence that the first incisor of F.
flexirostrum was horizontal as in the squalodontids, Waipatia and some kentriodontids. Instead, the
combination of a sinusoidal rostrum with procumbent large premaxillary teeth is a feature unique,
within the cetaceans, to F. flexirostrum. It is reminiscent of the rosette structure observed in the African
slender snouted crocodile (Mecistops cataphractus), the Indian gharial (Gavialis gangeticus), some
Muraenosocidae anguilliform fishes, and the spinosaurid Baryonyx [155,156]. This rostral
morphology has been interpreted as a biomechanical adaptation for biting and grabbing elusive prey
items [156].
In summary, considering the moderately wide temporal fossa and the dorsoventrally flat and
delicate sigmoidal rostrum bearing procumbent and large incisors, we speculate that F. flexirostrum
may have fed near the bottom, grasping with quick bites small and elusive prey such as shrimps and
small fishes.
6.4. Notocetus vanbenedeni
Notocetus vanbenedeni is currently the best known squalodelphinid species, thanks to five well-
preserved skulls and other fragmentary material from the Chilcatay Fm and the Argentinian Monte
León Formation. N. vanbenedeni resembles Furcacetus flexirostrum for the body size (estimated TBL:
2.37–2.55 m) and the anteriorly tapered and moderately elongated rostrum (RI = 0.62–0.68). However,
N. vanbenedeni differs from F. flexirostrum in the generally more robust skull (including the rostrum),
the lower tooth count for each quadrant (18–23), the larger teeth (transverse width at rostrum mid-
length = 4% of the BZW), and a more voluminous temporal fossa, being anteroposteriorly more
elongated and posterodorsally defined by a developed temporal crest. The interlocking teeth of N.
vanbenedeni are roughly vertical and their crowns show occlusal wear surfaces and embrasure pits
are observed between and laterally to the alveoli.
In summary, N. vanbenedeni probably captured with quick grasps larger and harder prey as
compared to F. flexirostrum, possibly using its interlocking posterior teeth with low, triangular, and
carinated crowns to cut in smaller pieces the prey before swallowing it, in a similar way to that
observed in Inia geoffrensis [150].
As already mentioned above, all squalodelphinids exhibit a left-side torsion of the rostrum.
Being present in the five skulls of N. vanbenedeni currently known, this character could be considered
a distinctive character of this family rather than an individual ‘anomaly’ as interpreted in some extant
longirostrine odontocetes [113,114,144,145]. Although a similar twist of the rostrum is observed in
protocetids and basilosaurids [48,157], none of the extant cetaceans consistently shows such a
character. It is therefore not easy to associate this shape to a specific function. With a more speculative
approach we can propose that the rostral torsion is linked to: (1) either side swimming, a behavior
observed in P. gangetica in very shallow waters [141,158]; (2) or the leftward shift of the larynx, which
would allow homodont odontocetes to swallow entire prey without suffocating [159]; (3) or
improved directional hearing, for the reception of high frequency sounds [157,160]. These and other
hypotheses should be tested with rigorous morphofunctional analyses in the future, aiming at
understanding the evolutionary pressure having generated such a peculiar rostral torsion.
Interestingly, this rostral torsion is associated in squalodelphinids to a marked left-side shift of the
facial region, more so than in other closely related platanistoids.
6.5. Macrosqualodelphis ukupachai
Macrosqualodelphis ukupachai is the largest squalodelphinid and platanistoid of the Chilcatay
assemblage (estimated TBL: 3.5 m). It further differs from the other platanistoids from the Chilcatay
Fm in the more robust rostrum, larger teeth (transverse width of alveoli at rostrum mid-length = 4.2%
of the BZW), more voluminous temporal fossa, and well-developed temporal and nuchal crests [26].
Life 2020, 10, 27 51 of 62
All these features suggest M. ukupachai had the role of a macropredator within the odontocete
Chilcatay paleocommunity, targeting larger prey.
6.6. Huaridelphis raimondii
Huaridelphis raimondii is a diminutive squalodelphinid, having an estimated TBL of 2.05 m, a
value reached by adult males of P. gangetica [23]. Its skull is relatively more gracile than those of the
other squalodelphinids, having a slender and more pointed rostrum and a less voluminous temporal
fossa [24]. The rostrum is moderately elongated (RI = 0.67). The tooth count for each quadrant (28–
30) is greater than in all other squalodelphinids, whereas the teeth are rather small (transverse width
of alveoli at rostrum mid-length = 2.6% of the BZW).
It is likely that the feeding strategy of H. raimondii was not far from that of N. vanbenedeni, but
this diminutive squalodelphinid certainly fed on smaller prey.
7. Concluding Perspectives on Trophic Partitioning Among the Platanistoids of the Chilcatay
Assemblage
The rich sample of platanistoids from the Chilcatay Fm collected in the localities of Ullujaya and
Zamaca represents a unique opportunity to reconstruct the ecological roles for a significant portion
of a fossil cetacean community that lived in a well-defined and limited space and time. In fact, all
platanistoids from the Chilcatay Fm are restricted to an 18–19 Ma time interval and, with the only
exception of Macrosqualodelphis ukupachai, are precisely positioned along the well-described
stratigraphical sequence of a sedimentary basin for which environmental conditions and vertical and
horizontal variations are well known [20–22,89,90]. Moreover, the platanistoid specimens from
Ulluyaja and Zamaca here examined represent a significant part of the entire cetacean assemblage
recorded by us (more than 180 specimens of cetaceans, although a significant portion has been
referred to Odontoceti indet.).
Expanding on previous, preliminary investigations of this assemblage [24–26], the range of sizes
and the morphological disparity observed at the level of the oral apparatus for the taxa discussed
here suggest that at least part of the trophic partitioning for the platanistoids of the Chilcatay Fm
could be related to: (1) different prey sizes (related to the predator's size and the size of its teeth), (2)
different prey types (from fish to other marine tetrapods and from cephalopods to harder prey items
like benthic crustaceans and shelled mollusks), and (3) different feeding strategies and associated
behaviors (raptorial feeding including lateral snapping, stunning of prey, prey probing in soft sediment,
suction-assisted raptorial feeding.; see above and [161]. Furthermore, these various foraging behaviors
could happen at different depths (from sea surface to seafloor). It should also be expected that these
different platanistoid taxa did not all feed at the same distance from the coast (shoreface versus offshore,
as the deposits of the Chilcatay Fm record different environments [20-22]). However, such a parameter
is even more difficult to test, considering that the carcasses of these cetaceans could have drifted for
some time with surface currents before sinking to the seafloor for final burial [20,162]. More
quantitative methods (e.g., [17,137,163], together with stable isotope analyses on teeth (e.g., [164,165],
would be needed to further investigate morphological disparity versus ecological parameters in this
species-rich odontocete assemblage from a critical time of cetacean evolutionary history [1,166].
Author Contributions: Conceptualization, methodology and formal analysis, G.B., O.L., C.M.; fossil discovery
and collection, M.U.; figure preparation, G.B.; writing—original draft preparation, G.B.; writing—review and
editing, G.B., O.L., C.M. All authors have read and agreed to the published version of the manuscript.
Funding: This research was supported by grants from the Italian Ministero dell’Istruzione, dell’Università e
della Ricerca (MIUR) to Bianucci (PRIN Project, 2012YJSBMK EAR- 9317031); National Geographic Society
Committee for Research Exploration to Lambert (grant number 9410-13); and the University of Pisa to Bianucci
(PRA_2017_0032).
Acknowledgments: We thank W. Aguirre for the careful preparation of the fossil specimens studied here; W.
Aguirre, J.A. Chauca Luyo, and E. Díaz, for field assistance; R. Salas-Gimondi and R. Varas-Malca for the access
to the laboratory and facilities at MUSM; A. Benites-Palomino for sending us some photographs and
Life 2020, 10, 27 52 of 62
measurements of MUSM specimens; C. Di Celma, G. Bosio, A. Collareta, A. Gioncada, W. Landini, E.
Malinverno, G. Molli, P.P. Pierantoni, K. Post, G. Sarti, and T. DeVries, for the helpful discussions about the
stratigraphy, biostratigraphy, geochronology, geology, and paleontology of the Chilcatay Fm; D.J. Bohaska and
N.D Pyenson (USNM), S. Bruaux, G. Lenglet, and O. Pauwels (RBINS), S. Farina (MSNUP), L. Del Favero and
M. Fornasiero (MGP), Z.N. Gasparini and L.H. Pomi (MLP) for the access to the collections under their care.
We also thank two anonymous reviewers and the editor Veronica Wang for their constructive comments and
suggestions.
Conflicts of Interest: The authors declare no conflict of interest. The funders had no role in the design of the
study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to
publish the results.
Appendix A
List of Characters for the Phylogenetic Analysis
Characters are polarized with respect to the basilosaurid Zygorhiza as the outgroup
1. Rostrum elongation ([167], modified): short, ratio between rostrum length and CBL < 0.75 (0);
elongated, ratio > 0.75 (1).
2. Premaxilla at the apex of the rostrum ([168], modified): apex of the rostrum constituted only by
the premaxillae on less than 10 per cent of its total length (0); apex of the rostrum constituted
only by the premaxillae on more than 10 per cent of its total length and lacking alveoli (1): apex
of the rostrum constituted both of the premaxillae and the maxillae (2).
3. Lateral rostral suture between premaxilla and maxilla deeply grooved [99]: no (0); yes (1).
4. Asymmetry of the premaxillae on the rostrum, at some distance anterior to the premaxillary
foramina, with the right premaxilla distinctly narrower than the left in dorsal view ([24],
modified): absent (0); present (1).
5. Widening of the premaxillae at the rostrum base [24]: narrow premaxillae, ratio between the
width of the rostrum and the transverse width of the premaxillae at the antorbital notch < 0.60
(0); wide premaxillae, ratio between 0.60 and 0.75 (1); extremely wide premaxillae nearly
reaching the lateral margin of the rostrum, ratio > 0.75 (2).
6. Dorsal opening of the mesorostral groove anterior to the rostrum base ([169], modified):
narrower than the premaxilla (0); wider than the premaxilla (1).
7. Deep, V-shaped, left antorbital notch, related to an anteriorly pointed antorbital process [24]: no
(0); yes (1).
8. Elevated antorbital region, distinctly higher than the dorsal margin of the rostrum base in lateral
view [24]: no (0); yes (1).
9. Distinct prominent dorsal crest in the antorbital-supraorbital region forming an acute angle in
cross section ([24], modified): no (0); yes (1).
10. Thickening of the antorbital process of the frontal, quantified as a ratio between the height of
this process measured in lateral view perpendicular to the maxilla-frontal suture and the vertical
distance from the lower margin of the occipital condyles to the vertex of the skull [24]; absent,
ratio < 0.25 (0); present, ratio > 0.30 (1).
11. Widening of the neurocranium [24]: neurocranium roughly as long as wide or longer than wide
with ratio between neurocranium length (longitudinal, from occipital condyles to level of
antorbital notches) and postorbital width > 0.90 (0); neurocranium distinctly shorter than wide
with ratio < 0.90 (1).
12. Posterior infraorbital foramen(ina) along the vertex more medial than the lateralmost margin of
the premaxilla in the cranium [24]: no (0); yes (1).
13. Deep fossa in the frontal on orbit roof, at the level of the frontal groove [24]: no (0); yes (1).
14. Vertex distinctly shifted to the left compared to the sagittal plane of the skull [24]: no (0); yes (1).
15. Transverse premaxillary crest on the vertex [168]: absent (0), present (1).
Life 2020, 10, 27 53 of 62
16. Ventral exposure of the palatine ([92], modified): palatine exposed and joined sagittally anterior
to the pterygoids (0); palatines lose sagittal contact and are displaced dorsolaterally but visible
on the palate (1); palatines displaced dorsolaterally and totally covered by pterygoids (2).
17. Hamular fossa of the pterygoid sinus [54]: short, not reaching anteriorly the level of the
antorbital notch (0); long, extending anteriorly on the palatal surface of the rostrum (1).
18. Thickening of the zygomatic process of the squamosal [24]; absent, ratio between the maximum
distance from the anteroventral margin of the zygomatic process to the posterodorsal margin, in
lateral view, and the vertical distance from the lower margin of the occipital condyles to the
vertex of the skull < 0.35 (0); present, ratio > 0.35 (1).
19. Dorsal outline of the zygomatic process of the squamosal in lateral view ([24], modified): almost
rectilinear (0); clearly dorsally convex forming a regular arc (1); dorsally convex with a with a
abruptly ventral sloping in its posterior portion (2).
20. Articular rim on the lateral surface of the periotic ([92], modified): absent (0); present (1);
present, forming a hook-like process (2).
21. Pars cochlearis of the periotic square-shaped in ventral view [92]: no (0); yes (1).
22. Aperture of the cochlear aqueduct of the periotic ([92], modified): small (0); very small (1); large
and thin-edged (2).
23. Aperture of the cochlear aqueduct of the periotic ([92], modified): faces mediodorsally (0); faces
dorsally (1).
24. Transverse thickening of the anterior process of the periotic [92]: no (0); yes (1).)
25. Internal auditory meatus of the periotic oval, with the dorsal opening for the facial canal lateral
to the spiral cribriform tract [170]: no (0); yes (1).
26. Separate ossicle at the apex of the anterior process of the periotic [170]: no (0); yes (1).
27. Elongated anterior spine on the tympanic bulla, associated with a marked anterolateral
convexity ([92], modified): no (0); moderate elongation (> 25% total length of the bulla) (1);
extreme elongation (> 25% total length of the bulla) (2).
28. Ventral groove of the tympanic affecting the whole length of the bone, including the anterior
spine [92]: no (0); yes (1).
29. Extent of the inner and outer posterior prominences of the tympanic [24]: both prominences with
approximately the same posterior extent (0); outer posterior prominence posteriorly longer than
the inner posterior prominence (1); outer posterior prominence posteriorly shorter than the inner
posterior prominence (2).
30. Dorsal margin of the involucrum of the tympanic cut by a median indentation, in medial view
[168]: absent (0), present (1).
31. Apical extension of the manubrium of the malleus [92]: no (0); yes (1).
32. Loss of double-rooted posterior teeth: [92]: no (0); yes (1).
33. Retention of accessory denticles on posterior teeth ([92], modified): yes (0); no (1).
34. Tooth count per upper or lower row ([24], modified): < 25 (0); > 25 and < 33 (1); > 33 (2).
35. Strong development of the dorsal transverse process of the atlas and extreme reduction of its
ventral process [92]: no (0); yes (1).
36. Great reduction of coracoid process of the scapula [92]: no (0); yes (1).
37. Great reduction or loss of supraspinatus fossa, with acromion located on anterior edge of
scapula [92]: no (0); yes (1).
38. Deep lateral groove on mandible [171]: no (0); yes (1).
39. Medial margin of the antorbital notch made of a thin plate [108]: (0) no, robust lateral margin of
the rostrum at base; yes (1).
40. Dorsal surface of vertex [108]: flat (0); markedly transversely and longitudinally convex (1).
41. Vertex strongly transversely pinched [108]: absent (0); present, maxillae converging markedly
posterior to bony nares (1).
42. Ventral edge of zygomatic process of squamosal in lateral view ([104], modified): concave (0);
almost straight or convex (1).
Life 2020, 10, 27 54 of 62
43. Space between the apex of the zygomatic process of the squamosal and the posterior margin of
the postorbital process of frontal in lateral view: zygomatic process distinctly posterior to the
postorbital process and consequently there is space between the apex of the zygomatic process
and the posterior margin of the postorbital process (0); apex of the zygomatic process exceeds
the posterior margin of postorbital process and both processes are strictly in contact together (1).
44. Transverse width of the temporal fossae: temporal fossae wide transversally, ratio between the
transverse width of the right + left temporal fossae and the BZW > 0.50 (0); temporal fossae
moderately narrow transversally, ratio > 25 and < 50 (1); temporal fossae markedly narrow
transversally < 25 (2).
45. Left-side torsion of the rostrum with longitudinal axis of the neurocranium forming an angle of
about 5 degree with the main axis of the rostrum in dorsal view generating asymmetry of the
posterior portion of the rostrum (right side transversally wider than the left side) and of the
antorbital notches (left antorbital notch clearly more anteriorly placed and narrower than the
right antorbital notch): absent(0); present (1).
46. Size of alveoli at the middle of the rostrum (greatest transverse diameter expressed as
percentage of the BZW): > 3 % (0); <3 and >2 (1); < 2% (2).
47. Elongation of mandibular symphysis and angle between the mandibula rami; symphysis < 65 %
of the total length of the mandibles and angle between the mandibles < 50 degrees (0);
symphysis > 65 % of the total length of the mandibles and angle between the mandibles > 50
degrees (1).
48. Orientation of the head of the humerus: posteroproximally (0); posterolaterally (1).
Appendix B
Table B1. Data matrix of 48 characters for one outgroup (Zygorhiza), 18 patanistoids and other
possibly related odontocetes (Squalodon, Waipatia and the eurhinodelphinids Eurhinodelphis,
Xiphiacetus and Ziphiodelphis). All characters are treated as unordered; 0, primitive state; 1, 2, derived
states; a, variable between 0 and 1; b, variable between 1 and 2; ?, missing character.
5 10 15 20 25 30 35 40 45 48
Zygorhiza 00000 00000 00000 00000 00000 00000 00000 000?0 ?0000 000
Squalodon 00000 10000 00000 00010 00000 00000 00000 11000 00010 00?
Waipatia 00001 10000 00000 00000 00000 00000 00000 ??000 00000 00?
Xiphiacetus 11101 00100 10000 000a0 00000 00011 ?1020 00100 00110 20?
Eurhinodelphis 11101 00000 10001 00000 00000 00011 01?20 ???00 00110 0??
Ziphiodelphis 11101 00100 10001 00000 00000 00011 01020 ??100 00110 20?
Zarhinocetus 12100 10100 10010 0100? 00000 0???0 ?1?2? ??100 10120 200
Allodelphis 12100 10000 10000 01001 10000 01020 ?1?20 ??a0 10120 200
Ninjadelphis 12??0 a1000 10000 0100? 10000 01020 ?1?20 10100 10120 200
Ensidelphis 10011 0?1?? 11?10 b111? ????? ?202? ?1?20 ??110 01100 20?
MUSM 603 ??011 00100 1??10 ?1111 0??0? 02020 01??? ???1? 01110 20?
Macrosqualodelphis ??011 00100 1?110 1111? ????? ????? ?1?01 ???10 01101 0?1
Huaridelphis 00011 01100 11110 11111 12100 0???0 ?1011 ??000 01111 1??
Notocetus 00011 01100 11110 11111 12100 01100 11001 11000 01111 001
Squalodelphis 00012 11100 1?110 ?11?1 12100 01100 1100? ??01? 0?11? 10?
Phocageneus ????? ????? ????? ????1 12100 01100 110?1 ????? ??? ?? ?0?
Medocinia ??012 11101 1?110 1111? ????? ????? ?1??? ???00 011?1 ???
Furcacetus 00011 0?100 11110 b1111 1??0? 0???? ?1?0? ????0 0?111 1??
Dilophodelphis 00011 01101 11110 11111 1??0? 0???? ?1?2? ???11 01111 2??
Araeodelphis 00001 00100 11110 ???1? ????? ????? ?102? ??111 01110 21?
Zarhachis 12111 00111 11110 11122 00010 11010 01120 ??111 01110 21?
Pomatodelphis 12111 00111 11110 11122 00010 01010 01120 ??110 01110 21?
MUSM 1611 ????? ????? ????? ????2 01001 1???? ????? ????? ????? ???
Platanista 02111 00010 11110 21112 01001 11010 01110 11111 11100 210
Life 2020, 10, 27 55 of 62
References
1. Marx, F.G.; Lambert, O.; Uhen, M.D. Cetacean Paleobiology; John Wiley & Sons, Ltd.: Chichester, UK, 2016.
2. Fordyce, R.E. Cetacean Evolution. In Encyclopedia of Marine Mammals, 3rd ed.; Würsig, B., Thewissen,
J.G.M., Kovacs, K.M., Eds.; Academic Press: London, UK, 2018; pp. 180–185.
3. de Muizon, C. L’origine et l’histoire évolutive des Cétacés. Comptes Rendus Palevol 2009, 8, 295–309.
4. Gingerich, P.D. Evolution of whales from land to sea. Proc. Am. Philos. Soc. 2012, 156, 309–323.
5. Uhen, M.D. The origin(s) of whales. Annu. Rev. Earth Planet. Sci. 2010, 38, 189–219.
6. Thewissen, J.G.M. The Walking Whales: From Land to Water in Eight Million Years, 1st ed.; University of
California Press: Oakland, CA, USA, 2014; 245p.
7. Marx, F.G.; Fordyce, R.E. Baleen boom and bust: A synthesis of mysticete phylogeny, diversity and
disparity. R. Soc. Open Sci. 2015, 2, 140434.
8. Tsai, C.-H.; Fordyce, R.E. The earliest gulp-feeding mysticete (Cetacea: Mysticeti) from the Oligocene of
New Zealand. J. Mamm. Evol. 2015, 22, 535–560.
9. Geisler, J.H.; Boessenecker, R.W.; Brown, M.; Beatty, B.L. The origin of filter feeding in whales. Curr. Biol.
2017, 27, 2036–2042.e2.
10. Lambert, O.; Martínez-Cáceres, M.; Bianucci, G.; Di Celma, C.; Salas-Gismondi, R.; Steurbaut, E.; Urbina,
M.; de Muizon, C. Earliest mysticete from the late Eocene of Peru sheds new light on the origin of baleen
whales. Curr. Biol. 2017, 27, 1535–1541.e2.
11. Slater, G.J.; Goldbogen, J.A.; Pyenson, N.D. Independent evolution of baleen whale gigantism linked to
Plio-Pleistocene ocean dynamics. Proc. R. Soc. B Biol. Sci. 2017, 284, 20170546.
12. Bianucci, G.; Marx, F.G.; Collareta, A.; Di Stefano, A.; Landini, W.; Morigi, C.; Varola, A. Rise of the titans:
Baleen whales became giants earlier than thought. Biol. Lett. 2019, 15, 20190175.
13. Geisler, J.H.; Colbert, M.W.; Carew, J.L. A new fossil species supports an early origin for toothed whale
echolocation. Nature 2014, 508, 383–386.
14. Churchill, M.; Martinez-Caceres, M.; de Muizon, C.; Mnieckowski, J.; Geisler, J.H. The origin of high-
frequency hearing in whales. Curr. Biol. 2016, 26, 2144–2149.
15. Park, T.; Fitzgerald, E.M.G.; Evans, A.R. Ultrasonic hearing and echolocation in the earliest toothed whales.
Biol. Lett. 2016, 12, 20160060.
16. Boessenecker, R.W.; Fraser, D.; Churchill, M.; Geisler, J.H. A toothless dwarf dolphin (Odontoceti:
Xenorophidae) points to explosive feeding diversification of modern whales (Neoceti). Proc. R. Soc. B Biol.
Sci. 2017, 284, 20170531.
17. McCurry, M.R.; Pyenson, N.D. Hyper-longirostry and kinematic disparity in extinct toothed whales.
Paleobiology 2019, 45, 21–29.
18. Marx, F.G.; Uhen, M.D. Climate, critters, and cetaceans: Cenozoic drivers of the evolution of modern
whales. Science 2010, 327, 993–996.
19. Steeman, M.E.; Hebsgaard, M.B.; Fordyce, R.E.; Ho, S.Y.W.; Rabosky, D.L.; Nielsen, R.; Rahbek, C.; Glenner,
H.; Sørensen, M.V.; Willerslev, E. Radiation of extant cetaceans driven by restructuring of the oceans. Syst.
Biol. 2009, 58, 573–585.
20. Bianucci, G.; Collareta, A.; Bosio, G.; Landini, W.; Gariboldi, K.; Gioncada, A.; Lambert, O.; Malinverno, E.;
de Muizon, C.; Varas-Malca, R.; et al. Taphonomy and palaeoecology of the lower Miocene marine
vertebrate assemblage of Ullujaya (Chilcatay Formation, East Pisco Basin, southern Peru). Palaeogeogr.
Palaeoclimatol. Palaeoecol. 2018, 511, 256–279.
21. Di Celma, C.; Malinverno, E.; Collareta, A.; Bosio, G.; Gariboldi, K.; Lambert, O.; Landini, W.; Pierantoni,
P.P.; Gioncada, A.; Villa, I.M.; et al. Facies analysis, stratigraphy and marine vertebrate assemblage of the
lower Miocene Chilcatay Formation at Ullujaya (Pisco basin, Peru). J. Maps 2018, 14, 257–268.
22. Di Celma, C.; Pierantoni, P.P.; Malinverno, E.; Collareta, A.; Lambert, O.; Landini, W.; Bosio, G.; Gariboldi,
K.; Gioncada, A.; de Muizon, C.; et al. Allostratigraphy and paleontology of the lower Miocene Chilcatay
Formation in the Zamaca area, East Pisco basin, southern Peru. J. Maps 2019, 15, 393–405.
23. Reeves, R.R.; Brownell, H.R. Susu Platanista gangetica (Roxburgh, 1801) and Platanista minor Owen, 1853. In
Handbook of Marine Mammals: River Dolphins and the Larger Toothed Whales; Ridgway, S.H., Harrison, R.L.J.,
Eds.; Academic Press: London, UK, 1989; Volume 4, pp. 69–95.
24. Lambert, O.; Bianucci, G.; Urbina, M. Huaridelphis raimondii, a new early Miocene Squalodelphinidae
(Cetacea, Odontoceti) from the Chilcatay Formation, Peru. J. Vertebr. Paleontol. 2014, 34, 987–1004.
Life 2020, 10, 27 56 of 62
25. Bianucci, G.; Urbina, M.; Lambert, O. A new record of Notocetus vanbenedeni (Squalodelphinidae,
Odontoceti, Cetacea) from the Early Miocene of Peru. Comptes Rendus Palevol 2015, 14, 5–13.
26. Bianucci, G.; Bosio, G.; Malinverno, E.; de Muizon, C.; Villa, I.M.; Urbina, M.; Lambert, O. A new large
squalodelphinid (Cetacea, Odontoceti) from Peru sheds light on the Early Miocene platanistoid disparity
and ecology. R. Soc. Open Sci. 2018, 5, 172302.
27. Mead, J.G.; Fordyce, R.E. The therian skull: A lexicon with emphasis on the odontocetes. Smithson. Contrib.
Zool. 2009, 627, 1–248.
28. Evans, H.E.; de Lahunta, A. Miller’s Anatomy of the Dog, 4th ed.; Saunders: St. Louis, MO, USA, 2012.
29. Swofford, D.L. PAUP*. Phylogenetic Analysis Using Parsimony (*and Other Methods). Version 4b10; Sinauer
Associates: Sunderland, MA, USA, 2001.
30. Thornburg, T.; Kulm, L.D. Sedimentary basins of the Peru continental margin: Structure, stratigraphy, and
Cenozoic tectonics from 6°S to 16°S latitude. In Nazca Plate: Crustal Formation and Andean Convergence; Kulm,
L.D., Dymond, J., Dasch, E.J., Hussong, D.M., Eds.; Memoir: Geological Society of America: Boulder, CO,
USA, 1981; Volume 154, pp. 393–422.
31. Dunbar, R.B.; Marty, R.C.; Baker, P.A. Cenozoic marine sedimentation in the Sechura and Pisco basins,
Peru. Palaeogeogr. Palaeoclimatol. Palaeoecol. 1990, 77, 235–261.
32. León Lecaros, W.R.; Rosell Solís, W.; Alemán, R., A.M.; Torres Bazán, V.R.; De la Cruz Matos, O.
Estratigrafía, sedimentología y evolución tectónica de la cuenca Pisco Oriental. Bol. INGEMMET Ser. D
2008, 27, 1–154.
33. Cobbing, E.J.; Pitcher, W.S. The coastal batholith of central Peru. J. Geol. Soc. 1972, 128, 421–454.
34. Shackleton, R.M.; Ries, A.C.; Coward, M.P.; Cobbold, P.R. Structure, metamorphism an geochronology of
the Arequipa Massif of coastal Peru. J. Geol. Soc. 1979, 136, 195–214.
35. Mukasa, S.B.; Henry, D.J. The San Nicolás batholith of coastal Peru: Early Palaeozoic continental arc or
continental rift magmatism? J. Geol. Soc. 1990, 147, 27–39.
36. DeVries, T.J. Oligocene deposition and Cenozoic sequence boundaries in the Pisco Basin (Peru). J. S. Am.
Earth Sci. 1998, 11, 217–231.
37. DeVries, T.J.; Urbina, M.; Jud, N.A. The Eocene-Oligocene Otuma depositional sequence (East Pisco Basin,
Peru): Paleogeographic and paleoceanographic implications of new data. Bol. Soc. Geol. Perú 2017, 112, 14–38.
38. Di Celma, C.; Malinverno, E.; Bosio, G.; Collareta, A.; Gariboldi, K.; Gioncada, A.; Molli, G.; Basso, D.;
Varas-Malca, R.M.; Pierantoni, P.P.; et al. Sequence stratigraphy and paleontology of the upper Miocene
Pisco Formation along the western side of the lower Ica Valley (Ica desert, Peru). Riv. Ital. Paleontol. Stratigr.
2017, 123, 255–27320.
39. DeVries, T.J.; Jud, N.A. Lithofacies patterns and paleogeography of the Miocene Chilcatay and lower Pisco
depositional sequences (East Pisco Basin, Peru). Bol. Soc. Geol. Perú 2018, 8, 124–167.
40. Brand, L.R.; Esperante, R.; Chadwick, A.V.; Porras, O.P.; Alomía, M. Fossil whale preservation implies high
diatom accumulation rate in the Miocene–Pliocene Pisco Formation of Peru. Geology 2004, 32, 165–168.
41. Esperante, R.; Brand, L.; Nick, K.E.; Poma, O.; Urbina, M. Exceptional occurrence of fossil baleen in shallow
marine sediments of the Neogene Pisco Formation, southern Peru. Palaeogeogr. Palaeoclimatol. Palaeoecol.
2008, 257, 344–360.
42. Esperante, R.; Brand, L.R.; Chadwick, A.V.; Poma, O. Taphonomy and paleoenvironmental conditions of
deposition of fossil whales in the diatomaceous sediments of the Miocene/Pliocene Pisco Formation,
southern Peru—A new fossil-lagerstätte. Palaeogeogr. Palaeoclimatol. Palaeoecol. 2015, 417, 337–370.
43. Gariboldi, K.; Gioncada, A.; Bosio, G.; Malinverno, E.; Di Celma, C.; Tinelli, C.; Cantalamessa, G.; Landini,
W.; Urbina, M.; Bianucci, G. The dolomite nodules enclosing fossil marine vertebrates in the East Pisco
Basin, Peru: Field and petrographic insights into the Lagerstätte formation. Palaeogeogr. Palaeoclimatol.
Palaeoecol. 2015, 438, 81–95.
44. Gioncada, A.; Collareta, A.; Gariboldi, K.; Lambert, O.; Di Celma, C.; Bonaccorsi, E.; Urbina, M.; Bianucci,
G. Inside baleen: Exceptional microstructure preservation in a late Miocene whale skeleton from Peru.
Geology 2016, 44, 839–842.
45. Bianucci, G.; Di Celma, C.; Landini, W.; Post, K.; Tinelli, C.; de Muizon, C.; Gariboldi, K.; Malinverno, E.;
Cantalamessa, G.; Gioncada, A.; et al. Distribution of fossil marine vertebrates in Cerro Colorado, the type
locality of the giant raptorial sperm whale Livyatan melvillei (Miocene, Pisco Formation, Peru). J. Maps 2016,
12, 543–557.
Life 2020, 10, 27 57 of 62
46. Bianucci, G.; Di Celma, C.; Collareta, A.; Landini, W.; Post, K.; Tinelli, C.; Muizon, C. de; Bosio, G.;
Gariboldi, K.; Gioncada, A.; et al. Fossil marine vertebrates of Cerro Los Quesos: Distribution of cetaceans,
seals, crocodiles, seabirds, sharks, and bony fish in a late Miocene locality of the Pisco Basin, Peru. J. Maps
2016, 12, 1037–1046.
47. Uhen, M.D.; Pyenson, N.D.; Devries, T.J.; Urbina, M.; Renne, P.R. New middle Eocene whales from the
Pisco Basin of Peru. J. Paleontol. 2011, 85, 955–969.
48. Martínez-Cáceres, M.; Lambert, O.; de Muizon, C. The anatomy and phylogenetic affinities of Cynthiacetus
peruvianus, a large Dorudon-like basilosaurid (Cetacea, Mammalia) from the late Eocene of Peru.
Geodiversitas 2017, 39, 7–163.
49. Lambert, O.; Bianucci, G.; Salas-Gismondi, R.; Di Celma, C.; Steurbaut, E.; Urbina, M.; de Muizon, C. An
amphibious whale from the middle Eocene of Peru reveals early south pacific dispersal of quadrupedal
cetaceans. Curr. Biol. 2019, 29, 1352–1359.e3.
50. de Muizon, C. Les vertébrés fossiles de la Formation Pisco (Péru). Deuxième partie: Les Odontocetés
(Cetacea, Mammalia) du Pliocène inférieur de Sud-Sacaco. Trav. Inst. Fr. Etud. And. 1984, 27, 1–188.
51. de Muizon, C. Les vertébrés fossiles de la Formation Pisco (Pérou). Troisième partie: Les odontocètes
(Cetacea, Mammalia) Miocènes. Trav. Inst. Fr. Etud. And. 1988, 42, 1–244.
52. de Muizon, C. Walrus-like feeding adaptation in a new cetacean from the Pliocene of Peru. Nature 1993,
365, 745–748.
53. Lambert, O.; Bianucci, G.; Post, K.; de Muizon, C.; Salas-Gismondi, R.; Urbina, M.; Reumer, J. The giant bite
of a new raptorial sperm whale from the Miocene epoch of Peru. Nature 2010, 466, 105–108.
54. Lambert, O.; de Muizon, C.; Bianucci, G. The most basal beaked whale Ninoziphius platyrostris, 1983: Clues on
the evolutionary history of the family Ziphiidae (Cetacea: Odontoceti). Zool. J. Linn. Soc. 2013, 167, 569–598.
55. Lambert, O.; Bianucci, G.; Beatty, B.L. Bony outgrowths on the jaws of an extinct sperm whale support
macroraptorial feeding in several stem physeteroids. Naturwissenschaften 2014, 101, 517–521.
56. Lambert, O.; Collareta, A.; Landini, W.; Post, K.; Ramassamy, B.; Di Celma, C.; Urbina, M.; Bianucci, G. No
deep diving: Evidence of predation on epipelagic fish for a stem beaked whale from the Late Miocene of
Peru. Proc. R. Soc. B Biol. Sci. 2015, 282, 20151530.
57. Lambert, O.; de Muizon, C.; Bianucci, G. A new archaic homodont toothed cetacean (Mammalia, Cetacea,
Odontoceti) from the early Miocene of Peru. Geodiversitas 2015, 37, 79–108.
58. Lambert, O.; Bianucci, G.; de Muizon, C. Macroraptorial sperm whales (Cetacea, Odontoceti,
Physeteroidea) from the Miocene of Peru. Zool. J. Linn. Soc. 2016, 179, 404–474.
59. Lambert, O.; Bianucci, G.; Urbina, M. A new inioid (Cetacea, Odontoceti, Delphinida) from the Miocene of
Peru and the origin of modern dolphin and porpoise families. Zool. J. Linn. Soc. 2017, 179, 919–946.
60. Lambert, O.; de Muizon, C.; Malinverno, E.; Celma, C.D.; Urbina, M.; Bianucci, G. A new odontocete
(toothed cetacean) from the Early Miocene of Peru expands the morphological disparity of extinct
heterodont dolphins. J. Syst. Palaeontol. 2018, 16, 981–1016.
61. Bianucci, G.; Di Celma, C.; Urbina, M.; Lambert, O. New beaked whales from the late Miocene of Peru and
evidence for convergent evolution in stem and crown Ziphiidae (Cetacea, Odontoceti). PeerJ 2016, 4, e2479.
62. Collareta, A.; Lambert, O.; de Muizon, C.; Urbina, M.; Bianucci, G. Koristocetus pescei gen. et sp. nov., a
diminutive sperm whale (Cetacea: Odontoceti: Kogiidae) from the late Miocene of Peru. Foss. Rec. 2017, 20,
259–278.
63. Bouetel, V.; de Muizon, C. The anatomy and relationships of Piscobalaena nana (Cetacea, Mysticeti), a
Cetotheriidae s.s. from the early Pliocene of Peru. Geodiversitas 2006, 28, 319–395.
64. Collareta, A.; Landini, W.; Lambert, O.; Post, K.; Tinelli, C.; Di Celma, C.; Panetta, D.; Tripodi, M.; Salvadori,
P.A.; Caramella, D.; et al. Piscivory in a Miocene Cetotheriidae of Peru: First record of fossilized stomach
content for an extinct baleen-bearing whale. Sci. Nat. 2015, 102, 70.
65. Marx, F.G.; Kohno, N. A new Miocene baleen whale from the Peruvian desert. R. Soc. Open Sci. 2016, 3,
160542.
66. Marx, F.G.; Lambert, O.; de Muizon, C. A new Miocene baleen whale from Peru deciphers the dawn of
cetotheriids. R. Soc. Open Sci. 2017, 4, 170560.
67. Marx, F.G.; Collareta, A.; Gioncada, A.; Post, K.; Lambert, O.; Bonaccorsi, E.; Urbina, M.; Bianucci, G. How
whales used to filter: Exceptionally preserved baleen in a Miocene cetotheriid. J. Anat. 2017, 231, 212–220.
68. de Muizon, C.; Bianucci, G.; Martínez-Cáceres, M.; Lambert, O. Mystacodon selenensis, the earliest known
toothed mysticete (Cetacea, Mammalia) from the late Eocene of Peru. Geodiversitas 2019, 41, 401–499.
Life 2020, 10, 27 58 of 62
69. de Muizon, C. Les vertébrés fossiles de la Formation Pisco (Pérou). Première partie: Deux nouveaux
Monachinae (Phocidae, Mammalia) du Pliocène inférieur de Sud-Sacaco. Trav. Inst. Fr. Etud. And. 1981, 6,
1–150.
70. Amson, E.; de Muizon, C. A new durophagous phocid (Mammalia: Carnivora) from the late Neogene of
Peru and considerations on monachine seals phylogeny. J. Syst. Palaeontol. 2014, 12, 523–548.
71. Clarke, J.A.; Ksepka, D.T.; Stucchi, M.; Urbina, M.; Giannini, N.; Bertelli, S.; Narvaez, Y.; Boyd, C.A.
Paleogene equatorial penguins challenge the proposed relationship between biogeography, diversity, and
Cenozoic climate change. Proc. Natl. Acad. Sci. USA 2007, 104, 11545–11550.
72. Clarke, J.A.; Ksepka, D.T.; Salas-Gismondi, R.; Altamirano, A.J.; Shawkey, M.D.; D’Alba, L.; Vinther, J.;
DeVries, T.J.; Baby, P. Fossil evidence for evolution of the shape and color of penguin feathers. Science 2010,
330, 954–957.
73. Ksepka, D.T.; Clarke, J.A.; DeVries, T.J.; Urbina, M. Osteology of Icadyptes salasi, a giant penguin from the
Eocene of Peru. J. Anat. 2008, 213, 131–147.
74. Stucchi, M.; Varas-Malca, R.; Urbina-Schmitt, M. New Miocene sulid birds from Peru and considerations
on their Neogene fossil record in the Eastern Pacific Ocean. Acta Palaeontol. Pol. 2016, 61, 417–427.
75. Parham, J.F.; Pyenson, N.D. New sea turtle from the Miocene of Peru and the iterative evolution of feeding
ecomorphologies since the Cretaceous. J. Paleontol. 2010, 84, 231–247.
76. de Muizon, C.; McDonald, H.G. An aquatic sloth from the Pliocene of Peru. Nature 1995, 375, 224–227.
77. de Muizon, C.; Mcdonald, H.G.; Salas, R.; Urbina, M. A new early species of the aquatic sloth Thalassocnus
(Mammalia, Xenarthra) from the Late Miocene of Peru. J. Vertebr. Paleontol. 2003, 23, 886–894.
78. de Muizon, C.; Mcdonald, H.G.; Salas, R.; Urbina, M. The youngest species of the aquatic sloth Thalassocnus
and a reassessment of the relationships of the nothrothere sloths (Mammalia: Xenarthra). J. Vertebr.
Paleontol. 2004, 24, 387–397.
79. Amson, E.; Argot, C.; McDonald, H.G.; de Muizon, C. Osteology and functional morphology of the forelimb
of the marine sloth Thalassocnus (Mammalia, Tardigrada). J. Mamm. Evol. 2015, 22, 169–242.
80. Amson, E.; Argot, C.; McDonald, H.G.; de Muizon, C. Osteology and functional morphology of the hind
limb of the marine sloth Thalassocnus (Mammalia, Tardigrada). J. Mamm. Evol. 2015, 22, 355–419.
81. Amson, E.; Argot, C.; McDonald, H.G.; de Muizon, C. Osteology and functional morphology of the axial
postcranium of the marine sloth Thalassocnus (Mammalia, Tardigrada) with paleobiological implications. J.
Mamm. Evol. 2015, 22, 473–518.
82. Ehret, D.J.; Hubbell, G.; Macfadden, B.J. Exceptional preservation of the white shark Carcharodon
(Lamniformes, Lamnidae) from the early Pliocene of Peru. J. Vertebr. Paleontol. 2009, 29, 1–13.
83. Ehret, D.J.; Macfadden, B.J.; Jones, D.S.; Devries, T.J.; Foster, D.A.; Salas-Gismondi, R. Origin of the white
shark Carcharodon (Lamniformes: Lamnidae) based on recalibration of the upper Neogene Pisco Formation
of Peru. Palaeontology 2012, 55, 1139–1153.
84. Landini, W.; Altamirano-Sierra, A.; Collareta, A.; Di Celma, C.; Urbina, M.; Bianucci, G. The late Miocene
elasmobranch assemblage from Cerro Colorado (Pisco Formation, Peru). J. S. Am. Earth Sci. 2017, 73, 168–190.
85. Landini, W.; Collareta, A.; Pesci, F.; Di Celma, C.; Urbina, M.; Bianucci, G. A secondary nursery area for the
copper shark Carcharhinus brachyurus from the late Miocene of Peru. J. S. Am. Earth Sci. 2017, 78, 164–174.
86. Landini, W.; Collareta, A.; Di Celma, C.; Malinverno, E.; Urbina, M.; Bianucci, G. The early Miocene
elasmobranch assemblage from Zamaca (Chilcatay Formation, Peru). J. S. Am. Earth Sci. 2019, 91, 352–371.
87. Collareta, A.; Lambert, O.; Landini, W.; Di Celma, C.; Malinverno, E.; Varas-Malca, R.; Urbina, M.; Bianucci,
G. Did the giant extinct shark Carcharocles megalodon target small prey? Bite marks on marine mammal
remains from the late Miocene of Peru. Palaeogeogr. Palaeoclimatol. Palaeoecol. 2017, 469, 84–91.
88. Collareta, A.; Landini, W.; Chacaltana, C.; Valdivia, W.; Altamirano-Sierra, A.; Urbina-Schmitt, M.;
Bianucci, G. A well preserved skeleton of the fossil shark Cosmopolitodus hastalis from the late Miocene of
Peru, featuring fish remains as fossilized stomach contents. Riv. Ital. Paleontol. Stratigr. 2017, 123, 11–22.