Relict_duck-billed_dinosaurs

bioRxiv preprint doi: https://doi.org/10.1101/2023.03.04.531097; this version posted March 6, 2023.
The copyright holder for this preprint

(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
available under aCC-BY 4.0 International license.
1 Relict duck-billed dinosaurs survived into the last age of the dinosaurs in subantarctic
2 Chile
3
4 Authors
5
6 Jhonatan Alarcón-Muñoz1,2*, Alexander O. Vargas1*, Hans Püschel3, Sergio Soto-
7 Acuña1,4, Leslie Manríquez5, Marcelo Leppe6, Jonatan Kaluza7, Verónica Milla1,8,
8 Carolina Simon-Gutstein1,9, José Palma-Liberona10, Wolfgang Stinnesbeck11, Eberhard
9 Frey12, Juan Pablo Pino1,2, Dániel Bajor1,2, Elaine Núñez1,2, Héctor Ortiz1,13, Héctor
10 Mansilla1,6, David Rubilar-Rogers14 and Penélope Cruzado-Caballero15,16
11
12
13 Affiliations
14
15
1
Red Paleontológica U-Chile. Departamento de Biología, Facultad de Ciencias,
16 Universidad de Chile. 2Departamento de Ecología, Facultad de Ciencias, Universidad de
17 Chile. 3School of GeoSciences, University of Edinburgh, Grant Institute, Edinburgh,UK.
4
18 KayTreng Consultores SpA, Ñuñoa, Santiago, Chile. 5Universidad de Vale do Rio do
19 Sinos, Brazil. 6Laboratorio de Paleobiología, Instituto Nacional Antártico Chileno.
7
20 Fundación Félix de Azara, Argentina, CONICET, Argentina. 8Universidad de
21 Concepción, Concepción, Chile. 9Paleo Consultores, Huelen 165c, Providencia,
22 Chile. 10Escuela de Medicina Veterinaria, Pontificia Universidad Católica de Chile,
23 Santiago, Chile. 11Institut für Geowissenschaften, Ruprecht-Karls-Universität
24 Heidelberg, Im Neuenheimer Feld 234-236, 69120 Heidelberg, Germany. 12Staatliches
25 Museum für Naturkunde Karlsruhe (SMNK), Erbprinzenstraße 13, 76133 Karlsruhe,
26 Germany. 13Universidad de Magallanes, Punta Arenas, Chile. 14Área Paleontología,
27 Museo Nacional de Historia Natural de Chile. 15Área de Paleontología y Área de
28 Petrología y Geoquímica, Departamento de Biología Animal, Edafología y Geología
29 Universidad de La Laguna.16Grupo Aragosaurus-IUCA, Facultad de Ciencias,
30 Universidad de Zaragoza, Zaragoza, Spain.
31
32 (*) jhoalarc@gmail.cl, alexvargas@uchile.cl
33
34
35 Abstract
36
37 In the dusk of the dinosaur era, the advanced duck-billed dinosaurs (Family
38 Hadrosauridae) are thought to have outcompeted other herbivores, making ecosystems
39 less diverse and more vulnerable to the Cretaceous-Paleogene asteroid impact. They were
40 also among the first terrestrial organisms to disperse from North America into South
41 America. Here, we present the first new species of subantarctic duck-billed dinosaur,
42 CPAP 3054, of early Maastrichtian age in Magallanes, Chile. Surprisingly, unlike
43 duckbills further north in Patagonia, CPAP 3054 is not an advanced duckbill, but
44 descends from North American forms that were transitional to Hadrosauridae, diverging
45 shortly before the origin of this family. In North America, these forms were replaced by
46 hadrosaurids in the late Campanian. The survival into the Maastrichtian of a pre-
bioRxiv preprint doi: https://doi.org/10.1101/2023.03.04.531097; this version posted March 6, 2023. The copyright holder for this preprint
47 hadrosaurid lineage suggests the ancestors of CPAP 3054 arrived earlier in South
48 America than the hadrosaurids, reaching further south before the Cretaceous-Paleogene
49 mass extinction, where they avoided competition from hadrosaurids.
50
51
52 Additional note
53 This work contains a new biological name. New names in preprints are not considered
54 available by the ICZN. To avoid ambiguity, the new biological name is not included in
55 this preprint, and the holotype specimen number CPAP 3054 is used as a placeholder.
56 Paratypes described in this preprint are also used in the diagnosis.
57
58
59 Introduction
60
61 Duck-billed dinosaurs (Hadrosauroidea) are among the most successful herbivores in the
62 history of life on earth. Their tooth batteries with hundreds of teeth are arguably the most
63 complex in vertebrate evolution (1), and were capable of crushing, grinding, and
64 shearing, allowing them to exploit a broad range of plant resources that they could further
65 reach at both short and high distances above ground (2). The duck-billed dinosaurs
66 attained their maximum historical diversity upon the radiation of advanced forms of the
67 family Hadrosauridae (the “true” duckbills), during the Campanian-Maastrichtian. In this
68 time period, they are thought to have outcompeted other herbivores, contributing to
69 overall declines of dinosaur biodiversity in North America and Central China previous to
70 the asteroid impact of the Cretaceous-Paleogene (K-Pg) mass extinction (2). Also in this
71 time period, Hadrosaurids were among the very few dinosaurs from Laurasia that are
72 known to have colonized Gondwanan continents (3,4). One species has been described
73 from northern Africa (5) and 5 from South America, all within central and northern
74 Patagonia (Chubut and Neuquén) (4, 6, 7,8, 9). There is also evidence that duck-billed
75 dinosaurs reached further into southern Patagonia, and even the Antarctic peninsula
76 (10,11,12). These regions at the time retained close geographic proximity with
77 intermittent formation of land bridges and intercontinental biotic exchanges (13); further,
78 they may have been at least partially isolated from the rest of South America due to
79 marine transgressions in the latest Cretaceous such as the Kawas Sea (14,15,16,17).
80 Dinosaur faunas in these far southern regions are poorly known, and duck-billed
81 dinosaurs are only documented by few and partial unnamed remains (11,12,13), which
82 are currently assumed to belong to hadrosaurids like those that inhabited central and
83 north Patagonia. Although Antarctica remained close to Australia and New Zealand, and
84 often exchanged biotic components (10), the absence of any confirmed remains of duck-
85 billed dinosaurs in the latter two suggests duckbills did not have enough time to arrive
86 there before the K-Pg mass extinction.
87 In 2013, we initiated excavation of a monodominant bonebed with abundant
88 disarticulated remains from several individuals of a new species of subantarctic duck-
89 billed dinosaur (18,19). The site is of early Maastrichtian age (see Methods) and is
90 located in the Río de Las Chinas valley of the Magallanes Region of subantarctic Chile,
91 at the southernmost tip of southern Patagonia (51ºS, Fig. 1). Here, we will use the term
92 “subantarctic” in the physiographic sense, to refer to the territories that are immediately
93 north of the Antarctic convergence, being roughly south of 46ºS (the southern limit of
94 Chubut, central Patagonia). So far, excavation of about 28 m2 of a layer approximately
95 80 cm thick has yielded over 42 skeletal elements, but other unexcavated outcrops of the
96 bonebed are found in an extension of about 5 km (see Methods). The materials recovered
97 so far are the most informative for any duck-billed dinosaur this far south, yielding
98 unexpected conclusions. Instead of being specifically related to other South American
99 duckbills, the new dinosaur is not a hadrosaurid: instead, it is similar to North American
100 forms that diverged shortly before the origin of that family. In technical terms, it is a non-
101 hadrosaurid hadrosauroid that is closely related to Hadrosauridae. For the sake of
102 simplicity, we will refer to it as a “transitional” duckbill, to emphasize that its
103 morphology is transitional to that of the “true” duckbills of the family Hadrosauridae.
104
105 Results
106
107 Systematic Paleontology
108 Dinosauria Owen, 1842
109 Ornithischia Seeley, 1887
110 Ornithopoda Marsh, 1881
111 Hadrosauroidea Cope, 1870
112 CPAP 3054 (gen. et. sp. nov.)
113 Holotype— CPAP 3054, right ilium.
114
115 CPAP 3054 was discovered among several other bones; all skeletal elements were
116 disarticulated, precluding a confident assignment of multiple elements to specific
117 individuals. We therefore designated CPAP 3054 as the holotype because of evident
118 diagnostic features. The preacetabular process of CPAP 3054 is incomplete but shows a
119 “T” shaped cross-section due to the well-developed lateral and medial crests that
120 characterize adults (20,21). These features allowed us to discard a juvenile status (that is,
121 an animal with no signs of impending maturity, 22).
122
123 Paratypes— Several disarticulated bones were selected as paratypes of the same species
124 (Fig. 2 and Figs. S1-S3) given the lack of evidence for more than one species at the site,
125 and the fact that monodominant associations are common in the fossil record of duck-
126 billed dinosaurs, reflecting social interaction or herding behavior (23,24,25,26,27).
127 Taphonomic aspects such as Hydraulic equivalence and lack of abrasion also suggest that
128 bones at the site underwent minimal transport, discarding water currents as the reason for
129 accumulated remains of numerous individuals (see Taphonomy section in Supplementary
130 Materials and Tables S1 and S2). The bones assigned as paratypes are the following: a
131 right premaxilla (CPAP 5337), the incomplete left maxilla of a juvenile individual (CPAP
132 5338), an incomplete left maxilla (CPAP 5339), an incomplete right maxilla (CPAP
133 5340), an incomplete left postorbital (CPAP 5341), an incomplete right dentary (CPAP
134 5342), a left dentary (CPAP 5370), a fragmentary right dentary (CPAP 5368), a right
135 quadrate (CPAP 5343), a cervical vertebra (CPAP 5344), a cervical centrum (CPAP 5380),
136 two incomplete dorsal vertebrae (CPAP 5345 and CPAP 5346), a dorsal neural arch
137 (CPAP 5377), eight proximal and mid caudal vertebrae (CPAP 5347, CPAP 5348, CPAP
138 5349, CPAP 5350, CPAP 5351, CPAP 5384, CPAP 5385 and CPAP 5386), four
139 incomplete ribs (CPAP 5378, CPAP 5381, CPAP 5382 and CPAP 5383), a right scapula
140 (CPAP 5371), an incomplete left sternum (CPAP 5352), a complete left humerus (CPAP
141 5353), an incomplete left humerus (CPAP 5354), two fragments of a left humerus (CPAP
142 5369), an incomplete left ulna (CPAP 5355), an incomplete right ulna (CPAP 5373), an
143 incomplete right radius (CPAP 5379), a postacetabular process with part of the iliac blade
144 (CPAP 5356), an incomplete right ischium (CPAP 5357), a complete left femur (CPAP
145 5358), an incomplete right femur (CPAP 5359), the distal end of a right femur (CPAP
146 5360), the distal end of the right femur of a juvenile individual (CPAP 5361), a complete
147 left tibia (CPAP 5362), an incomplete left tibia (CPAP 5372), an incomplete right fibula
148 (CPAP 5363), a right metatarsal III (CPAP 5364).
149 Etymology — The words used for the new species are in the language of the Aónik’enk,
150 the indigenous people that inhabited the region where the new species was discovered
151 (28).
152 Type locality and horizon — El Puesto Area (50°42'42"S y 72°32`29" W), Río de las
153 Chinas Valley, Estancia Cerro Guido, Magallanes Region, Chilean Patagonia (51° S).
154 Upper section of the Dorotea Formation (lower Maastrichtian), between 71.7 ± 1.2 and
155 70.5 ± 4.5 million years (29,30) (See Methods).
156 Diagnosis. Small-sized hadrosauroid dinosaur (total body length ~ 4 m) that differs from
157 all other members of this clade in possessing a scapula with an antero-ventrally curved
158 pseudoacromion process, and an ilium with a medioventrally and anteroposteriorly well-
159 developed sacral crest extending posterior to the base of the postacetabular process.
160 CPAP 3054 also possesses a unique combination of characters, differing from
161 Hadrosauridae in a maxilla with a jugal articular surface showing a prominent and
162 caudally projecting dorsal tubercle; a dentary with a short diastema, mandibular
163 symphysis oblique relative to its long axis, and tooth rows with less than 30 tooth
164 positions, that do not extend beyond the coronoid process, and converge anteriorly with
165 the lateral surface of the dentary; quadrate with a medial condyle that is not markedly
166 elevated dorsally compared with the lateral condyle; a deltopectoral crest with a rounded
167 laterodistal end extending less than 48% of the total length of the humerus; ilium with
168 nearly straight dorsal border, and a supraacetabular crest longer than 70% of the length of
169 the iliac blade; and tibia with a cnemial crest extending less than 50% of the total length
170 of the bone. In addition, CPAP 3054 differs from South American Hadrosauridae by a
171 scapula with a mediolaterally narrow coracoidal facet, with a more ventrally curved
172 pseudoacromion process (Fig. S4), and an ilium with a more laterally curved
173 preacetabular process, a less laterally developed supraacetabular process, and a shorter
174 and ventrally oriented pubic peduncle (Fig. S5). CPAP 3054 differs from early-diverging
175 Hadrosauroidea in that it presents derived characters also found in Hadrosauridae: a
176 premaxilla with a “double-layer” morphology; a supraacetabular process anteriorly
177 located with respect to the dorsal tuberculum of the ischial peduncle; the ratio between
178 the base of the preacetabular process and the distance between the dorsal border of the
179 ilium and the pubic peduncle is greater than 0.5; and a laterocaudal condyle of the
180 proximal end of the tibia is more robust than the lateral one. Regarding similar non-
181 hadrosaurids closely related to Hadrosauridae, CPAP 3054 differs from Eotrachodon in
182 that the precircumnarial region of the premaxilla is more anteroposteriorly extended,
183 lacking a caudally everted oral margin; and the maxilla has an articular surface for the
184 jugal located anterior to the dorsal process rather than posterior to it. CPAP 3054 differs
185 from Lophorhothon in that the jugal process of the postorbital is pointing ventrally and
186 the posterior edge of this process is slightly convex, rather than concave; and the ilium is
187 higher, with a laterally less developed supraacetabular crest, located anterior to the dorsal
188 tubercle of the ischial peduncle. CPAP 3054 differs from Huehuecanauhtlus in that the
189 preacetabular process of the ilium in dorsal view is markedly curved laterally, less curved
190 ventrally, and its base is dorsoventrally narrower; the dorsal tubercle of the ischial
191 peduncle of the ilium is less dorsally elevated; and the postacetabular process is
192 approximately rectangular in lateral view, rather than triangular.
193
194 Diagnosis of the genus. As for the type species.
195 Description— The premaxilla has an oval, very deep circumnarial fossa which lacks
196 accessory fossae, foramina, and ridges, unlike Saurolophinae (31,32,33,34,35,36,37,
197 Fig.3). The oral margin is convex and anteroventrally deflected, unlike most
198 Saurolophinae (38,39, Fig. 3). It shows a denticulate margin, separated by a shallow
199 groove from a denticulate border with the palatine ("double layer"), as in Saurolophidae
200 and some non-hadrosaurid hadrosauroids (31,37,40,41,42; Fig. 4). The maxilla is
201 subtriangular in lateral view. The articular surface for the jugal is robust, subquadrangular
202 in shape and has a prominent dorsal tubercle, unlike hadrosaurids (43,44). A large rostral
203 maxillary foramen is present at the lateral side, as in Saurolophinae (31,40, Fig. 4). The
204 medial surface bears anteroposteriorly aligned neurovascular foramina (31,40). The teeth
205 are poorly preserved, making it difficult to determine their morphology and number in the
206 maxilla. The postorbital is T-shaped in lateral view. Unlike most hadrosaurids, the
207 dentary has a short diastema (less than 20% of the length of the alveolar row), and the
208 mandibular symphysis is oblique relative to the long axis of the dentary, as in several
209 non-hadrosaurid duckbills (35,45,46, Fig. 3). The alveolar furrows are parallel to each
210 other and dorsoventrally extended. The most complete dentary (CPAP 5370) shows about
211 25 tooth positions, rather than over 30 as in Hadrosauridae (39,40,47,48, Fig. 3). The
212 dentary tooth row extends further posteriorly than in basal iguanodontians, reaching the
213 coronoid process, but not extending beyond it, unlike hadrosaurids (40,44,48, Fig. 3). The
214 tooth row converges anteriorly with the lateral surface of the dentary, unlike
215 Plesiohadros and Hadrosauridae (37,40,49,50, Fig. 3). The coronoid process is robust,
216 almost straight, and separated from the alveolar row by a groove.
217 The quadrate is dorsoventrally elongated. Its dorsal half curves slightly posteriorly and
218 lacks a squamosal buttress. The quadratojugal notch is ventrally positioned. Unlike most
219 hadrosaurids, the lateral condyle is not markedly offset ventrally with respect to the
220 medial condyle (40,45, Fig. 3). The cervical centrum is longer than dorsoventrally high,
221 and strongly opisthocoelous. The anterior and posterior articular surfaces are oval. The
222 postzygapophyses are well elevated dorsally to the roof of the neural canal and curved
223 posteriorly and laterally, which is common in hadrosaurids (31; Fig. 4). The dorsal centra
224 are elongate, higher than wide, opisthocoelous and have heart-shaped articular faces like
225 hadrosaurids and at least some non-hadrosaurid hadrosauroids (31,33,42, Fig. 4). The
226 caudal centra are amphiplatyan with hexagonal articular contours, which is typical but
227 not exclusive to hadrosaurids (31,33,39,51 Fig. 4).
228 The sternum is hatchet-shaped. The scapula has a dorsally curved blade as is typical of
229 hadrosaurids (40; Fig. 4), with a rectangular posterior end. The deltoid ridge is poorly
230 developed. The glenoid fossa narrows mediolaterally and the pseudoacromion process is
231 prominent and anteroventrally curved, unlike in all other South American duckbills (Fig.
232 S4). In addition, unlike South American hadrosaurids such as Huallasaurus,
233 Lapampasaurus and Kelumapusaura, the coracoidal facet is mediolaterally narrower
234 (Fig. S4).
235 Unlike Hadrosauridae, the deltopectoral crest of the humerus extends over 45% the total
236 length of the bone (40,52). The ratio between length and maximum width of the
237 deltopectoral crest is 3.9, which is higher than in Saurolophidae (20). The laterodistal
238 corner of the deltopectoral crest is widely arcuate, unlike most hadrosaurids (40,52,53,
239 Fig. 3). The ulna has a prominent olecranon process. The iliac peduncle of the ischium is
240 longer than the pubic peduncle. The dorsal border of the ilium is almost straight, unlike
241 Huehuecanauhtlus and hadrosaurids, including South American duckbills
242 (33,35,39,52,54,55,56, Fig. S5). The sacral crest is prominent and more developed than in
243 other South American duckbills (Fig. S5). The preacetabular process is incomplete. This
244 is robust, with a “T” shaped cross-section, and is more laterally curved than in other
245 South American duckbills (Fig. S5). The preserved anterior end of the preacetabular
246 process surpasses the lateral extension of the supraacetabular process, as in
247 Secernosaurus, but not in other South American duckbills (Fig. S5). The ratio between
248 the maximum height of the caudal end of the preacetabular process and the distance
249 between the dorsal margin of the ilium and the ventral edge of the pubic peduncle is 0.58,
250 which is higher than in non-hadrosaurid hadrosauroids (9,40, Fig. 4). The pubic peduncle
251 is robust and triangular in lateral view, unlike the elongated, rod-like pubic peduncle of
252 basal iguanodontoids such as Iguanodon (57). The pubic peduncle is ventrally oriented
253 and shorter than in some South American duckbills (Fig. S5). The ischial peduncle
254 consists of two low and blunt tubercles of similar size. The supraacetabular process
255 projects slightly lateroventrally, with its ventral apex slightly anteriorly located with
256 respect to the dorsal tubercle of the ischial peduncle, as in Hadrosauridae (20,52, Fig. 4).
257 The length of the supraacetabular crest reaches approximately 82.6 % of the length of the
258 central plate of the ilium. This is within the range of non-hadrosaurids: lower than in
259 basal iguanodontoids such as Iguanodon (>85%), and higher than in Hadrosauridae
260 (<70%) (50, Fig. 3). The lateroventral protrusion of the supraacetabular process is less
261 marked than in other South American duckbills except for Secernosaurus (Fig. S5). The
262 postacetabular process is rectangular, slightly dorsally inclined, and twisted. The femur is
263 straight, with a triangular fourth trochanter. The flexor and extensor grooves are deep,
264 with the latter almost closed, and the distal condyles are strongly projected posteriorly,
265 which are features typical of hadrosaurids (31, Fig. 4). The lesser trochanter is small and
266 well-separated from the greater trochanter. The cnemial crest extends to less than half the
267 total length of the bone, which differs from all hadrosaurids except Bonapartesaurus
268 (9,45,56,58,59, Fig. 3). Posterior to the cnemial crest there are two rounded condyles
269 separated by a sulcus. The posteromedial condyle is more robust and longer than the
270 lateral condyle, like Hadrosauridae (60, Fig. 4). The tibia has a torsion of approximately
271 45° between the proximal and distal epiphysis. The lateral malleolus is more extended
272 distally than the medial one. A fragment of a fibula reveals an enlarged proximal end. As
273 in several non-hadrosaurids, the articular surface of the metatarsal III is triangular
274 (33,58); the ratio between the length and mediolateral width of this bone is 4.5.
275 Discussion
276
277 CPAP 3054 is a duck-billed dinosaur that shares some derived traits with “true” duckbills
278 of the family Hadrosauridae (Fig. 4), while several other characters retain their ancestral
279 condition (Fig. 3). This transitional morphology suggests CPAP 3054 is not a
280 hadrosaurid, but is nevertheless closely related to Hadrosauridae, having diverged shortly
281 before the origin of this family. To properly address the relationships of CPAP 3054, we
282 carried out the most comprehensive phylogenetic analysis of duck-billed dinosaurs to
283 date, using a data matrix from a recent study (4) which we modified to correct scorings
284 and include more characters, as well as Gobihadros and Huehuecanauthlus, which are
285 important taxa with a morphology that is transitional to Hadrosauridae (42,54), like
286 CPAP 3054 (Methods). The South American duckbill Lapampasaurus was excluded
287 because it is too fragmentary to confidently establish its affinities (adding this taxon also
288 had no effect; see Methods). Our results for maximum parsimony analysis were
289 unprecedentedly well-resolved, obtaining only 4 most parsimonious trees (1312 steps, CI:
290 0.396; RI: 0.821); further, all differences among these trees were among early-branching
291 taxa at a node distant from CPAP 3054 (Fig. S6 shows the strict consensus of all 4 trees).
292 Parsimony places CPAP 3054 outside of Hadrosauridae, but among transitional duckbills
293 that are closest to this family (Fig. 5 and S6-S8; see Methods for selection and time
294 calibration of this tree). Bayesian analysis also places CPAP 3054 as a close relative of
295 Hadrosauridae, although its relations to other transitional duckbills were less resolved
296 (Figs. S7 and S8). Resolution improved in the undated Bayesian analysis, which
297 coincides with parsimony in that CPAP 3054 and the North American taxa Lophorhothon
298 and Huehuecanhautlus are closer to Hadrosauridae than to other transitional duckbills
299 (Fig. S7). Thus, these 3 taxa are likely the most relevant for understanding the origins of
300 Hadrosauridae. However, both Lophorhothon and Huehuecanhautlus are represented by
301 very partial materials (54,61); currently, provides the best information, which is also
302 bound to increase upon future excavations at the monotypic bonebed.
303 CPAP 3054 is not specially related to other South American duckbills, which are
304 advanced forms of the family Hadrosauridae. Recent work has proposed that all other
305 South American duckbills are saurolophine hadrosaurids, forming a monophyletic group
306 that is the sister to the Kritosaurini, a tribe that inhabited Laramidia in North America (4).
307 The monophyly of these South American hadrosaurids suggests that all members of this
308 group descend from a single species that dispersed into South America. In our own
309 analysis, despite inclusion of CPAP 3054 and some re-coding of South American
310 duckbills, we recovered the same monophyletic group of South American saurolophine
311 hadrosaurids in both Parsimony and Bayesian analyses, which was also recovered as
312 sister to Kritosaurini by parsimony (Figs. 5, S6-S8). Although these South American
313 hadrosaurids are related to Kritosaurini, they cannot be referred to this tribe: both
314 parsimony and bayesian analyses do not support a nested position within the North
315 American forms that are used to define Kritosaurini (62); thus, the current definition of
316 Kritosaurini excludes these South American relatives. Given the need for a different and
317 succinct term for the monophyletic group of “South American saurolophine hadrosaurids
318 closely related to Kritosaurini”, we propose the name Austrokritosauria for the most
319 inclusive clade that contains Huallasaurus, Secernosaurus, Bonapartesaurus and
320 Kelumapusaura, but not Gryposaurus. Importantly, CPAP 3054 does not belong to this
321 clade: placing CPAP 3054 as sister of Austrokritosauria requires a steep 15 extra steps in
322 the parsimony analysis (Table S3). Therefore, CPAP 3054 represents a completely
323 different lineage of duck-billed dinosaurs that independently colonized South America.
324 As the most informative duck-billed dinosaur from far southern regions, CPAP 3054 also
325 prescribes a reinterpretation of the record of partial fossils found in Southern Patagonia
326 and Antarctica. This actually suggests hadrosaurids may have never arrived to these high
327 austral latitudes. Partial remains in these regions can no longer be assumed a priori to
328 belong to hadrosaurids like those of central and northern Patagonia: in fact, none of these
329 remains show characters that are exclusive to Hadrosauridae, especially when we take
330 into account the hadrosaurid-like features of CPAP 3054. In particular, two incomplete
331 caudal centra from deposits of the Chorrillos Formation (Santa Cruz Province, Argentina)
332 have been attributed to Hadrosauridae based on the hexagonal contour of their articular
333 surfaces (11), but this trait is also present in CPAP 3054 (Fig. 4). An isolated and
334 incomplete tooth from the Maastrichtian of Vega Island in the Antarctic Peninsula was
335 attributed to Hadrosauridae, and tentatively to Hadrosaurinae (=Saurolophinae) because
336 of its relatively symmetrical crown, with a single and nearly straight medial keel, and
337 poorly developed denticles (13). Teeth are currently unknown for CPAP 3054 and most
338 non-hadrosaurid duckbills, but the above characters were likely ancestral for
339 Hadrosauridae, and may have been present in taxa close to this family; indeed, they are
340 present in some maxillary teeth of Eotrachodon (44) and were likely present from before
341 in evolution, since they are also observed in maxillary teeth of Eolambia, a much earlier
342 non-hadrosaurid from North America (39). Also from the Antarctic peninsula, a
343 metatarsal fragment found in Maastrichtian rocks of Seymour Island was attributed to
344 Hadrosauridae (12), but its morphology has already been pointed out to be shared by
345 other ornithopods, including those that inhabited these austral regions such as the
346 Elasmaria (63) and now, non-hadrosaurid duckbills. In synthesis, southwards from
347 central Patagonia (Chubut, 46ºS), no remains can be reliably attributed to Hadrosauridae.
348 This may add to other faunistic differences that have been proposed when comparing the
349 fossil record of southern Patagonia and northern Patagonia (64), which could also relate
350 to the presence of an archipelago at the time in southern South America, as a result of the
351 marine transgressions that produced inland seas such as the Kawas Sea (14,15,16,17).
352 CPAP 3054 is the first non-hadrosaurid ever found in Gondwana, and its presence so far
353 south poses a challenging biogeographic enigma. Any explanation implies remarkably
354 long routes, large gaps in between with no records of non-hadrosaurid duckbills, and
355 marine barriers that stopped most Laurasian dinosaurs; most likely, these marine barriers
356 would have been breached by chains of islands rather than continuous land bridges (65),
357 such that any dispersal of terrestrial animals must have involved swimming or “rafting”
358 on debris (Fig. 6). The first American biotic exchange (between North America and
359 South America) is proposed to have occurred during the latest Cretaceous (Campanian-
360 Maastrichtian) (66). Mammals of North American origin were already diverse in the
361 earliest Paleocene of South America, suggesting they must have crossed over earlier
362 (67,68). However, actual fossil evidence of exchanges during the latest Cretaceous is
363 only provided by dinosaurs: the Austrokritosauria and the patagonian nodosaurid
364 Patagopelta had North American ancestors (4,69), while large titanosaurs in North
365 America likely had ancestors in South America (70,71).
366 Another possibility is that non-hadrosaurids from Europe made it into Africa, and from
367 there into South America (the "Atlantogean" route, 71; See Fig.5). This is a longer route
368 that also implies crossing two marine barriers (rather than one), but exchanges through
369 this route have been argued to be better supported than those between the Americas,
370 based on numerous shared biotic components between Europe and South America during
371 the latest Cretaceous (71). Whichever the case, duck-billed dinosaurs had the highest
372 vagility (dispersal capacity) documented for any dinosaurs, with the greatest number of
373 dispersal events that likely crossed marine barriers, namely: between Asia and Laramidia
374 (9,70); Appalachia and Laramidia (twice) (32,72); Europe and Appalachia (50); Asia and
375 Europe (twice) (50,73); Europe and North Africa (5); Laramidia and South America (4);
376 and South America and Antarctica (63). Duck-billed dinosaurs are also most often
377 preserved near or within coastal environments (31; also the case for CPAP 3054; see
378 Methods) and have been suggested to be apt swimmers, or even semi-aquatic (5,51).
379 To formally test hypotheses on the biogeographical history of CPAP 3054, we
380 reconstructed likely ancestral areas for all internal nodes of the phylogeny obtained from
381 parsimony analysis, using BioGeoBEARS (74) as well as statistical Dispersal-Vicariance
382 Analysis (s-DIVA) (75,76,77,78). The best-fit model in BioGeoBEARS was the
383 DIVALIKE + j model (AICc: 182.21), which supports a strong role for events with
384 founder effects (5, Table S4). This model provides clear-cut support for a Laramidian
385 origin for the last common ancestor shared by CPAP 3054 and Hadrosauridae, as well as
386 for the last ancestor shared with Eotrachodon, at the previous node (Figs. 5 and S12). The
387 results of s-DIVA also support a North American origin, despite some differences, and
388 decreased resolution: the last common ancestor shared by CPAP 3054 with
389 Hadrosauridae is inferred in broad areas that combine Laramidia and/or Appalachia with
390 South America, and the previous last ancestor shared with Eotrachodon is inferred as
391 Appalachian (Fig. S13). As for Austrokritosauria, the results of both s-DIVA and
392 BioGeoBEARS support a Laramidian origin. This is the first time that formal
393 biogeographical statistical analysis is shown to support the occurrence during the
394 Cretaceous of the first American biotic exchange, for CPAP 3054 and Austrokritosauria;
395 similar studies may further confirm dispersal of nodosaurid ankylosaurs into South
396 America, and of titanosaurian sauropods into North America.
397 Although Biogeographic statistics do not support an Atlantogean route (from Europe via
398 Africa) for the ancestors of CPAP 3054, this could be a result of a sampling bias in Late
399 Cretaceous faunas, which are very poorly known in Africa (5,79,80). Non-hadrosaurids
400 were present in Europe in the Late Cretaceous (35,50), when lambeosaurine hadrosaurids
401 managed to disperse from Europe into Africa (5), which suggests non-hadrosaurids could
402 have done the same. To explore this, we added a hypothetical African taxon of a non-
403 hadrosaurid duckbill to our BioGeoBEARS and s-DIVA analyses, placing it between
404 CPAP 3054 and Eotrachodon, or as sister to CPAP 3054. A North American origin
405 continued to be best supported, suggesting that this conclusion will not be easily
406 overturned, even if such discoveries are made in Africa. (Figures S12-S15, Tables S5 and
407 S6, see Methods and Supplementary Material for details).
408 Some dinosaur clades that are present in both North and South America have been shown
409 to reflect ancient cosmopolitan origins, before the final separation of Gondwana and
410 Laurasia (81,82,83,84,85,86,87). Although CPAP 3054 is not a hadrosaurid, such a
411 scenario would require for a much earlier divergence near the origin of duck-billed
412 dinosaurs (Hadrosauroidea) in the late Hauterivian (130 mya), when biotic exchanges
413 between Europe and Gondwana were still possible (2012). Such a basalmost position for
414 CPAP 3054 would require a steep 45 extra steps (Table S3). Given the actual
415 phylogenetic position of CPAP 3054, to insist on ancient cosmopolitan roots would imply
416 that duckbills close to Hadrosauridae had already originated in the Hauterivian, despite a
417 gap of about 45 million years in their worldwide fossil record (Fig. 5).
418 We conclude that CPAP 3054 is very likely descended from North American non-
419 hadrosaurids that are transitional to Hadrosauridae, of the likes of Eotrachodon,
420 Lophorhothon and Huehuecanhautlus. These were also the last non-hadrosaurids of
421 North America, where they were replaced by hadrosaurids: Eotrachodon and
422 Huehuecanhautlus are of Santonian age; only Lophorothon may have lived in the
423 Campanian, given its large chronostratigraphic uncertainty, ranging from the latest
424 Santonian (83,6 mya) to the late Campanian (75,8 mya) (32,54,61). Non-hadrosaurids are
425 absent in the abundant and well-studied record of North American dinosaurs of the latest
426 Campanian and Maastrichtian, strongly suggesting they had become locally extinct
427 (41,88). However, at some point before that, the non-hadrosaurid ancestors of CPAP
428 3054 managed to leave North America, surviving into the Maastrichtian as distant relict
429 populations in Subantarctic Chile.
430 Given that CPAP 3054 is found so far south, and that no hadrosaurid remains can be
431 confirmed from Subantarctic and Antarctic regions, this pattern suggests the ancestors of
432 CPAP 3054 arrived earlier than those of Austrokritosauria, giving the former a head-start
433 in reaching more austral regions. Hadrosaurids may have not had enough time to reach
434 this far south before the K-Pg mass extinction. This interpretation is allowed by the
435 divergence times in our time-calibrated tree: CPAP 3054 would have last shared an
436 ancestor with North american forms around 91 mya (Turonian), while the
437 Austrokritosauria would have shared their last ancestor with the North American
438 Kritosaurini around 85 mya (Santonian; see Methods for calibration details).
439 Another reason to suspect that hadrosaurids may have never arrived into Subantarctic and
440 Antarctic regions is that they tended to replace non-hadrosaurids in those regions where
441 they co-existed. The niches of hadrosaurids likely overlapped onto those of non-
442 hadrosaurids, while non-hadrosaurids like CPAP 3054 were often smaller-sized, and had
443 smaller tooth plates with fewer tooth positions, taking up a smaller proportion of the jaw
444 (39,40,47,48). Given the importance of the tooth-jaw apparatus, this could explain why in
445 the Maastrichtian non-hadrosaurids had disappeared in North America (41,88), while a
446 single species is known from Asia (89,90). Non-hadrosaurids like Telmatosaurus
447 persisted into the Maastrichtian of Europe, but this is argued to have resulted from their
448 geographic isolation in islands (35,50). This is also a feasible explanation for CPAP
449 3054, given marine transgression events at the time in southern South America.
450 Although the radiation of the Hadrosauridae in North America and Central China
451 coincides with a decline in dinosaur biodiversity previous to the K-Pg asteroid impact,
452 the presence of Hadrosauridae cannot be confirmed in southern Patagonia and Antarctica,
453 while New Zealand and Australia may have had no duckbills at all (hadrosaurid or not),
454 with exclusively endemic gondwanan faunas right until the K-Pg mass extinction (91). It
455 remains to be seen if declines in dinosaur biodiversity were a global phenomenon; if so,
456 they did not correlate everywhere with the radiation of the Hadrosauridae. Alternatively,
457 if declines in dinosaur biodiversity did not occur in austral regions, their ecosystems were
458 nevertheless equally vulnerable to mass extinction upon the asteroid impact, and would
459 suggest a greater role of the impact itself as a kill mechanism (92).
460
461
462 Materials and Methods
463
464 Experimental model and subject details
465
466 The anatomical information used for both our comparisons and phylogenetic analyses
467 were obtained from the literature and from direct observation of specimens housed in
468 public repositories. The specimens mentioned below were observed directly:
469 -Huallasaurus australis, Natural History Museum Bernardino Rivadavia, Buenos Aires,
470 Argentina (MACN-RN 02, MACN-RN 142, MACN-RN 142B, MACN-RN 143, MACN-
471 RN 144, MACN-RN 145, MACN-RN 146, MACN-RN 826).
472 -Secernosaurus koerneri, FMNH, The Field Museum, Chicago, USA (FMNH PP13423).
473 -Bonapartesaurus rionegrensis, Vertebrate Paleontology Collection at the National
474 University of Tucuman and the Fundación Miguel Lillo (Tucuman, Argentina), (MPCA-
475 Pv SM2).
476
477
478 Specimen provenance and geological Setting
479
480 The specimens studied here were found in the Río de las Chinas Valley located 350 km to
481 the north of the city of Punta Arenas, Ultima Esperanza Province (Magallanes Region,
482 Chile). The studied outcrops extend along the valley in N-S direction, near to the
483 international border with Argentina (Fig.1). The fossil-bearing levels are assigned to the
484 Dorotea Formation (Upper Campanian to Danian; 29,93,94), which corresponds to the
485 upper section of the continental-marine sedimentary succession that fills the
486 Magallanes/Austral Basin. This basin was formed during the break-up of Gondwana and
487 the opening of the Atlantic Ocean (95,96,97). This succession conformably overlies the
488 Campanian to lower Maastrichtian Tres Pasos Formation (98,99) and is unconformably
489 overlain by the Paleogene sequences of the Man Aike/Río Turbio formations of middle to
490 late Eocene age (100,101). The Dorotea Formation represents a transition between
491 shallow marine and continental environments
492 (102,103,104,105,106,107,108,109,110,111,112). Specifically, it has been interpreted as
493 a transition from a shallow marine shelf-edge to tide-dominated delta and fluvial systems
494 (106,110,111,112,113,114).
495 The Dorotea Formation mainly comprises greenish-gray and reddish-brown sandstones
496 with abundant conglomerate and siltstone lenses, thin beds of sandy calcareous
497 concretions and mudstones of 900 to 1200 m thickness (98,110,115). Along the
498 succession there are fossil-bearing levels with bivalves, ammonites, gastropods, sharks,
499 plesiosaurs, mosasaurs, frogs, turtles, dinosaurs, mammals, fossil wood and leaves
500 (30,87,98,110,111,116,117,118,119,120,121,122).
501 The material here studied comes from the upper section of the Dorotea Formation in the
502 Río de las Chinas Valley, consisting of brown sandy mudstone (60 cm thickness) with
503 coal lenses and reddish-brown fine-grained sandstone (20 cm thickness) levels (CHA-2,
504 Fig. 1). Numerous small to medium -sized hadrosauroid bones were discovered in the
505 sandy mudstone level and larger bones in the sandstones. The bones are preserved three-
506 dimensionally and show linear and angular fragmentation. In Addition, a 0.5 m thick
507 silicified trunk was discovered in the surface debris on a topographic slope. Apparently,
508 this layer with hadrosaur bones extends laterally for about 5 km to the northeast,
509 exposing itself at several points in the valley. The most notable differences have to do
510 with the grain size at the different points where this layer is exposed. Further stratigraphic
511 studies are needed to determine whether this entire layer originated from a single
512 depositional event, and whether bones exposed at other points also correspond to CPAP
513 3054. The depositional environment of these levels with hadrosauroid bones are
514 interpreted as the distal section of a floodplain in a continental environment (fluvial
515 channels), with lower energy. An age estimation using U-Pb detrital zircon for the levels
516 above the hadrosauroid horizons provided values of 71.7 ± 1.2 Mya and 70.5 ± 4.5 Mya
517 (29,30), supporting an early Maastrichtian age (Late Cretaceous) for the fossil-bearing
518 levels.
519
520
521 Extraction and technical preparation
522
523 After the initial finding of fossil bones in situ at the exposed surface, we identified the
524 bearing-bone layer. The level was excavated in detail and several more specimens
525 appeared randomly distributed. Bone remains were extracted individually or as a set of
526 bones, using plaster jackets for protection and transport. Some bones were prepared in the
527 Paleobiology Laboratory of the Chilean Antarctic Institute (INACH), located in Punta
528 Arenas, Magallanes Region, while most of them were prepared in the Paleontology
529 Laboratory of the Faculty of Sciences of the University of Chile, located in Santiago de
530 Chile. The bones were prepared using air scribes, helped with tips and needles, in some
531 cases under a binocular microscope. Glues such as paraloid (B-72) and cyanoacrylate
532 were used to stabilize the samples. Finally, each element was numbered and conditioned
533 for their preservation in the collection of INACH with the acronym CPAP.
534
535 Character Dataset and taxa
536
537 For the phylogenetic analyses (and subsequent biogeographical analyses), we used a
538 modified version of the matrix of a recent phylogenetic analysis of all duck-billed
539 dinosaurs (Hadrosauroidea) by (4), with the most extensive character and taxon sample
540 published to date. We rescored/redefined 20-character states:
541
542 1) Secernosaurus koerneri (character 258). Changed from 1 to ?, since the distal half of
543 the scapula is not preserved.
544 2) Secernosaurus koerneri (character 259). Changed from 0 to ?, since the scapula is
545 incomplete.
546 3) Secernosaurus koerneri (character 278). Changed from ? to 1, based on fig. 12 of
547 (123).
548 4) Secernosaurus koerneri (character 290). Changed from 0 to 1, based on calculated
549 proportions.
550 5) Secernosaurus koerneri (character 293). Changed from 2 to 1 since the
551 supraacetabular process does not extend far enough ventrally to reach the middle of the
552 iliac blade.
553 6) Secernosaurus koerneri (character 303). Changed from 0 to 1, as the thickening of the
554 postacetabular process is due to its dorsomedial rotation.
555 7) Secernosaurus koerneri (character 312). Changed from 1 to 0, since in our opinion, the
556 pubic peduncle of the pubis is shorter and more robust compared with other
557 hadrosauroids, see (47).
558 8) Bonapartesaurus rionegrensis (character 247). Changed from ? to 0, since the
559 chevrons are shorter than the caudal neural spines.
560 9) Bonapartesaurus rionegrensis (character 291). Changed from 0 to 1, since the apex of
561 the supraacetabular process is in an anterodorsal position with respect to the caudal
562 tuberosity of the ischial peduncle of the ilium.
563 10) Bonapartesaurus rionegrensis (character 335). Changed from 1 to 0, based on the
564 original description given in (9).
565 11) Kelumapusaura machi (character 291). Changed from 0 to 1 since the apex of the
566 supraacetabular process is in an anterodorsal position with respect to the caudal
567 tuberosity of the ischiadic peduncle of the ilium.
568 12) Huallasaurus australis (character 287). Changed from 0 to 1 based on our
569 observations.
570 13) CPAP 3054 (character 263). Redefinition (new character state added). A fourth state
571 (3) was added for the pseudoacromion process, which refers to the ventral curvature of
572 this structure.
573 14) Laiyangosaurus youngi (character 17). Changed from 2 (undefined character in
574 character list) to 1. Based on the description of (124).
575 15) Parasaurolophus walkeri (character 89). Changed from 2 (undefined character in
576 character list) to 1 based on (125).
577 16) Olorotitan arharensis (character 89). Changed from 2 (undefined character in
579 17) Hypacrosaurus altispinus (character 89). Changed from 2 (undefined character in
581 18) Lambeosaurus lambei (character 91). Changed from 2 (undefined character in
583 19) Lambeosaurus magnicristatus (character 91). Changed from 2 (undefined character
584 in character list) to 1 based on (125).
585 20) Hypacrosaurus altispinus (character 91). Changed from 2 (undefined character in
587
588 In addition, we added 5 new characters and asides from CPAP 3054:
589 Character 361 – Maxilla, ectopterygoid ridge: continuous with jugal tubercle (0) or
590 discontinuous (1). Added from (5), Character 244.
591 Character 362 – Maxilla, ectopterygoid ridge: ridge extends to ventral jugal tubercle (0)
592 main part of shelf lies distinctly below it (1). Added from (5), Character 245.
593 Character 363 – Maxilla, neurovascular foramina: a single very large foramen below the
594 jugal, with a second, smaller foramen below it: absent (0); present (1). Added from (5),
595 Character 247.
596 Character 364 – Maxilla, posterior dentigerous process: dentigerous margin straight (0);
597 posterior downturn (1) strong posterior downturn, i.e., "boomerang" shape (2). Added
598 from (5), Character 340.
599 Character 365 – Humerus, shape of the deltopectoral crest apex: well rounded (0);
600 extending abruptly to produce a prominent angular profile (1). Added from (126),
601 Character 37; following (40), Character 221.
602
603 We also added Gobihadros mongoliensis and Huehuecanauhtlus tiquichensis. As in
604 previous work (4), Lapampasaurus was left out from the results because its extremely
605 partial remains (only 11 of 365 characters) preclude any significant inference of its
606 phylogenetic affinities. Inclusion in our analysis did not alter the topologies of the 4
607 MPT’s and placed this south american taxon within Saurolophinae, but retrieved it
608 closest to the japanese Kamuysaurus.
609 The character distribution was modified with Mesquite 2.75 (127). The characters were
610 coded based on bones belonging to adult or subadult individuals of CPAP 3054. The
611 character states of juvenile individuals were only coded for characters that do not show
612 significant ontogenetic variation. We modified the coding of some characters of
613 Huallasaurus australis based on direct observation by two of the co-authors (SSA and
614 PC-C). We also modified the scoring of some characters of Secernosaurus koerneri and
615 Bonapartesaurus rionegrensis based on published images and direct observation by AOV
616 (Secernosaurus koerneri) and PC-C (Bonapartesaurus rionegrensis). The resulting
617 matrix included 77 species-level taxonomic units (four non-hadrosauroid iguanodontians
618 and 73 hadrosauroids) scored for 365 equally weighted characters.
619
620 Parsimony analysis
621
622 The maximum parsimony analysis was carried out in TNT version 1.5 (128). The non-
623 hadrosauroid iguanodontian Ouranosaurus nigeriensis was selected as the outgroup. A
624 heuristic search of 1000 replicates of Wagner trees using random additional sequences
625 was performed, followed by branch swapping by tree-bisection-reconnection, retaining
626 100 trees per replicate. All characters were treated as equally weighted and unordered.
627 For all the trees, Bremer support (129) was calculated for each node to assess its
628 robustness by computing decay indices in TNT using the script Bremer.run (130).
629 Bootstrap proportions (131) were also calculated with TNT, setting the analysis for 5000
630 replicates using heuristic searches (see Fig. S6).
631 Templeton tests
632
633 We used a Templeton test to assess the significance of different hypotheses where CPAP
634 3054 was forced into alternative phylogenetic positions, evaluating changes in the
635 number of steps. We randomly used MPTs obtained from the unconstrained analysis to
636 compare with the forced topologies, applying a one-sided Wilcoxon signed-rank test to
637 the differences in character transformations between trees. We assessed the forced
638 positions of CPAP 3054 as the basalmost hadrosauroid (1), and as a member of the
639 Austrokritosaurini (2). (Table S3) We assessed whether the difference in steps between
640 the original and forced hypotheses was significant using Templeton Tests (132),
641 implemented with a script in TNT 1.5 (133).
642
643 Selection and Time calibration of trees obtained from Parsimony analysis
644
645 Only 4 trees were obtained from parsimony analysis with the standard option to collapse
646 zero-length branches. Close relationships of CPAP 3054 were fully resolved: differences
647 among trees were all in taxa distant from CPAP 3054, represented in a single polytomy at
648 a basal node (see Fig. S6). We selected one of the 4 trees to show the relationships of
649 CPAP 3054, and for biogeographic analyses with BioGeoBEARS, which require a fully
650 resolved tree (see below). We picked the only tree with no zero-length branches (no
651 polytomies), which is shown in Fig. 5 (tree number 0 in TNT). To time-calibrate this tree,
652 we ran a tip-dated Bayesian analysis with the entire topology constrained to this tree; we
653 kept the same priors and MCMC parameters than in the unconstrained Bayesian Analysis
654 (see below) and used the resulting maximum clade compatibility tree (MCCT). To obtain
655 the MCCT tree we used the posterior files of MrBayes and the
656 obtainDatedPosteriorTreesMrB() function of the R package paleotree v3.4.4 (134). Table
657 below shows FADs (First Appearance Data) and LADs (Last Appearance Data), which
658 were used for the time calibration of the phylogenetic tree according to
659 chronostratigraphic uncertainty ranges (oldest possible age and youngest possible age).
660
661 FADs (First Appearance Data) and LADs (Last Appearance Data):
662
Taxon FAD LAD
Ouranosaurus nigeriensis 129.4 100.5
Iguanodon bernissartensis 125.4 117.2
Mantellisaurus atherfieldensis 121.4 117.2
Jinzhousaurus yangi 127.4 122.0
Equijubus normani 113.0 110.5
Sirindhorna khoratensis 121.4 113.0
Xuwulong yueluni 121.4 100.5
Probactrosaurus gobiensis 129.4 100.5
Zuoyunlong huangi 100.5 93.9
Yunganglong dathongensis 100.5 93.9
Eolambia caroljonesa 98.46 98.32
Protohadros byrdi 100.5 93.9
Jintasaurus meniscus 113.0 100.5
Levnesovia transoxiana 91.86 89.5
Tanius sinensis 77.85 69.05
Bactrosaurus johnsoni 77.85 71.9
Gilmoreosaurus mongoliensis 77.85 71.9
Claosaurus agilis 88.05 85.8

Nanningosaurus dashiensis 100.5 66.0
Adynomosaurus arcanus 71.449 69.269
Telmatosaurus transsylvanicus 72.3 69.05
Tethyshadros insularis 77.85 69.05
Eotrachodon orientalis 84.95 83.4
Zhangenglong yangchengensis 86.8 83.4
Plesiohadros djadokhtaensis 77.85 71.9
Lophorhothon atopus 83.8 75.85
Hadrosaurus foulkii 80.5 78.5
Yamatosaurus izanagii 71.94 71.69
Acristavus gagslarsoni 80.6 79.8
Wulagasaurus dongi 72.3 68.0
Maiasaura peeblesorum 76.4 77.2
Brachylophosaurus canadensis 77.85 71.9
Probrachylophosaurus bergei 83.8 77.85
Kritosaurus navajovius 73.49 73.83
Huallasaurus australis 77.85 69.05
Gryposaurus latidens 81.08 80.6
Rhinorex condrupus 75.88 75.15
Gryposaurus monumentensis 77.85 71.9
Gryposaurus notabilis 76.5 74.5
Prosaurolophus maximus 77.03 75.46
Saurolophus angustirostris 77.85 69.05
Saurolophus osborni 72.3 69.05
Shantungosaurus giganteus 77.85 71.9
Edmontosaurus annectens 67.5 66.0
Edmontosaurus regalis 74.9 71.9
Kamuysaurus japonicus 72.3 69.05
Kerberosaurus manakini 69.05 66.0
Laiyangosaurus youngi 73.5 72.9
Aralosaurus tuberiferus 84.95 77.85
Canardia garonnensis 69.05 66.0
Nipponosaurus sachalinensis 84.95 77.85
Blasisaurus canudoi 69.05 66.0
Pararhabdodon isonensis 66.2 66.0
Tsintaosaurus spinorhinus 83.8 77.85
Jaxartosaurus aralensis 86.8 83.4
Parasaurolophus cyrtocristatus 76.439 74.893
Parasaurolophus walkeri 77.03 76.39
Charonosaurus jiayinensis 69.05 66.0
Parasaurolophus tubicen 77.83 73.49
Arenysaurus ardevoli 69.05 66.0
Olorotitan ararhensis 69.05 66.0
Amurosaurus riabinini 69.05 66.0
Sahaliyania elunchunorum 69.05 66.0

Lambeosaurus lambei 76.39 76.1
Lambeosaurus magnicristatus 77.8 72.1
Magnapaulia laticaudus 75.21 74.55
Velafrons coahuilensis 72.5 71.4
Corythosaurus casuarius 77.03 76.39
Corythosaurus intermedius 77.03 75.46
Hypacrosaurus altispinus 70.9 69.97
Hypacrosaurus stebingeri 79.7 72.1
Kelumapusaura machi 72.3 66.0
Bonapartesaurus rionegrensis 77.85 69.05
Secernosaurus koerneri 69.05 66.0
Gobihadros mongoliensis 100.5 83.4
Huehuecanauhtlus tiquichensis 86.8 83.4
CPAP 3054 72.9 66.0
663
664
665 Bayesian phylogenetic analyses (undated and tip-dated)
666
667 We conducted an undated Bayesian analysis in MrBayes v3.27a (135). As in the
668 parsimony analysis, the outgroup was the non-hadrosauroid iguanodontian Ouranosaurus
669 nigeriensis. We used the Mkv + Γ model of morphological evolution with ascertainment
670 bias correction (136), setting the coding as variable and with six rate categories for the
671 gamma distribution. We used two independent runs for the analysis with 40 million
672 Markov chain Monte Carlo (MCMC) generations, sampling every 1000 generations and
673 discarding 50% of the samples as burn-in. The posterior distributions for each parameter
674 were checked to confirm that the effective sample size was > 200 and the deviation of
675 split frequencies was below 0.05.
676 In addition, we performed a tip-dated Bayesian analysis in MrBayes v3.27a (135) using
677 the same settings for the model of morphological evolution (see above). The root of the
678 tree was calibrated using an exponential distribution with a minimum age of 129.4 mya
679 and a mean of 139.4 mya. The minimum age of the root was based on the maximum age
680 of the non-hadrosauroid iguanodontian Ouranosaurus nigeriensis. For calibrating the tree
681 tips, we used a uniform distribution based on the radiometric age uncertainty of the
682 fossils scored in the morphological matrix (137)
683 We used a normal distributed clock rate prior, with a mean of 0.02253 and standard
684 deviation of 0.00839. This prior was extracted from the previously obtained undated
685 Bayesian consensus tree and by using the packages ape (138) and fitdistrplus (139) from
686 R 2022.07.2+576 (140) following the methodology described in (141). The best model
687 following the Bayesian information criterion was the gamma distributed clock rate (shape
688 = 5.92176, rate = 262.84887). However, preliminary analyses of the data showed
689 problems of convergence between the two independent runs using this model, therefore
690 the normal distributed model was preferred for our analysis. We used the FBD model
691 using an exponential net diversification prior with rate 1, a beta fossil sampling
692 proportion prior with shape parameters α = 1 and β = 1, a beta turnover prior with shape
693 parameters α = 1 and β = 1, and an extant sampling proportion of 1. We modeled branch
694 rate variation using the Independent Gamma Rate (IGR) relaxed clock model with an
695 exponential distribution of rate 10. We used diffuse priors for the FBD model and the
696 clock variance, which reflect our prior uncertainty in the distribution of these parameters.
697 We used two independent runs for the analysis with 40 million Markov chain Monte
698 Carlo (MCMC) generations, sampling every 1000 generations and discarding 50% of the
699 samples as burn-in. The posterior distributions for each parameter were checked to
700 confirm that the effective sample size was > 200 and the deviation of split frequencies
701 was below 0.05.
702 In the 50% majority-rule consensus tree of the tip-dated Bayesian analysis, CPAP 3054 is
703 placed in a polytomy with Eotrachodon, Zhangenlong and Plesiohadros (Fig. S8).
704 Resolution improved in the undated Bayesian analysis, when geological ages were not
705 allowed to influence the topology; as in the parsimony analysis, CPAP 3054 was
706 retrieved closer to the clade that includes Lophorhothon, Huehuecanauthlus and
707 Hadrosauridae than to Eotrachodon, Zhangenlong or Plesiohadros (Fig. S7). This
708 difference between tip-dated and undated analysis may result from the fact that the ages
709 of the nearby Nanningosaurus and Gobihadros are very uncertain, with large possible
710 temporal ranges. Regardless, in both approaches, many polytomies persisted regarding
711 the close relationships of CPAP 3054. Even in the maximum posterior probability trees
712 (not shown), most relationships were recovered in less than 50% of trees. We therefore
713 decided not to select any tree resulting from Bayesian phylogenetic analysis for the
714 purposes of biogeographic analyses (which require fully resolved trees).
715
716 Statistical Biogeographic analyses
717
718 We implemented a statistical Dispersal-Vicariance Analysis (s-DIVA75,76,77,78) in
719 RASP 4.2 (78) to reconstruct the ancestral areas for all internal nodes of the phylogeny
720 obtained in our parsimony analysis. Six general areas were considered: Asia (A), Europe
721 (B), Laramidia (C), Appalachia (D), South America (E) and Africa (F). Prior to the s-
722 DIVA, we performed a new phylogenetic analysis in TNT 1.5 to obtain the phylogenetic
723 trees that were used in this analysis. The phylogenetic analysis recovered 6 most
724 parsimonious trees without collapsing zero-length branches (CI: 0.369, RI: 0.821), to
725 obtain fully resolved trees. The 6 trees were loaded in RASP 4.2, as well as the strict
726 consensus tree, which was used for the graphical representation of the s-DIVA results.
727 The "Allow Reconstruction (Slow)" option was selected to implement the method of
728 calculating the frequency of ancestral states for each node (142). The Supplementary
729 Material contains the details of the results.
730 We also used the R package BioGeoBEARS v.1.12 (74) as a second approach for
731 estimating the biogeographic history of duck-billed dinosaurs and the ancestral areas for
732 CPAP 3054. This method evaluates alternative models for extinction, dispersal,
733 cladogenesis and founder effect (143). BioGeoBears requires fully resolved time-
734 calibrated trees. We did not use any tree obtained from Bayesian inference since most
735 relationships of CPAP 3054 remained unresolved in the 50% majority rule tree (144).
736 Therefore, we used the only tree with no zero-length branches obtained from parsimony
737 analysis, which we time-calibrated by running a tip-dated Bayesian analysis with the
738 entire topology constrained to this tree (see above). The terminal taxa were coded for the
739 following biogeographical provinces: Asia, Europe, Laramidia, Appalachia, South
740 America, and Africa. We used a dispersal matrix that considers different dispersal
741 probabilities between land masses. The dispersal weights between geographical areas
742 were obtained from (5). We evaluated six different models: DIVALIKE and DIVALIKE
743 + j (75), DEC and DEC + j (145), and BAYAREALIKE and BAYAREALIKE + j (146).
744 We permitted only four simultaneous areas. These models were compared using their
745 AICc (Corrected Akaike Information Criterion). Values of log-Likelihood (lnL),
746 Dispersal (d), Extinction (e), Founder effect (j), Corrected Akaike Information Criterion
747 (AICc), and AICc Weight (AICc wt) scores from each model implemented. The model
748 with the lowest AICc value (and higher AICc + wt) was selected as the model with the
749 best fit to the data.
750
751
752
753 References
754
755 1) Weishampel, D. B. Evolution of jaw mechanisms in ornithopod dinosaurs. Advances in
756 anatomy, embryology, and cell biology, 87, 1-109 (1984).
757
758 2) Condamine, F.L. et al. Dinosaur biodiversity declined well before the asteroid impact,
759 influenced by ecological and environmental pressures. Nat. Commun 12, 3833 (2021).
760 https://doi.org/10.1038/s41467-021-23754-0
761
762 3) Viñola-López, L. W., Cerda, I. A., Correa-Narvaez, J., Codorniú, L., Borges-Sellén, C. R.,
763 Arano-Ruiz, A. F., & Ceballos-Izquierdo, Y. New fossils shed light on the Late Cretaceous
764 terrestrial community in the Caribbean and the First American Biotic Interchange. Cretac.
765 Res., 130, 105067 (2021).
766
767 4) Rozadilla, S., Brissón-Egli, F.; Agnolín, F.L.; Aranciaga-Rolando, A.M.; Novas, F.E. A
768 new hadrosaurid (Dinosauria: Ornithischia) from the Late Cretaceous of northern Patagonia
769 and the radiation of South American hadrosauroids. J. Syst. Palaeontol. 19(17), 1207–1235
770 (2022).
771
772 5) Longrich, N.R., Pereda Suberbiola, X., Pyron, A.R. & Nour-Eddine, J. The first duckbill
773 dinosaur (Hadrosauridae: Lambeosaurinae) from Africa and the role of the oceanic dispersal
774 in dinosaur biogeography. Cretac. Res. 120, 104678 (2020).
775
776 6) Brett-Surman, M.K. Phylogeny and palaeobiogeography of hadrosaurian dinosaurs.
777 Nature 277, 560–562 (1979).
778
779 7) Bonaparte, J. F., Franchi, M.R., Powell, J.E., Sepúlveda, E.G. La Formación Alamitos
780 (Campaniano-Maastrichtiano) del sudeste del Río Negro, con descripción de Kritosaurus
781 australis n. sp. (Hadrosauridae). Significado paleogeográfico de los vertebrados. Revista de
782 la Asociación Geológica Argentina 39, 284–299 (1984).
783
784 8) Coria, R. A, González Riga, B., Casadío, S. Un nuevo hadrosáurido (Dinosauria,

785 Ornithopoda) de la Formación Allen, Provincia de La Pampa, Argentina. Ameghiniana 49,
786 552–572 (2012).
787
788 9) Cruzado-Caballero, P., Powell, J. Bonapartesaurus rionegrensis, a new hadrosaurine
789 dinosaur from South America: implications for phylogenetic and biogeographic relations
790 with North America. J. Vertebr. Paleontol. 37, e1289381 (2017).
791
792 10) Reguero, M.A., Goin, F.J. Paleogeography and biogeography of the Gondwanan final
793 breakup and its terrestrial vertebrates: New insights from southern South America and the
794 “double Noah’s ark” Antarctic Peninsula. J. of South American Earth Sciences 108, 103358
795 (2021).
796
797 11) Rozadilla, S., Agnolín, F., Manabe, M., Tsuihiji, T. & Novas, F.E. Ornithischian remains
798 from the Chorrillo Formation (Upper Cretaceous), southern Patagonia, Argentina, and their
799 implications on ornithischian paleobiogeography in the Southern Hemisphere. Cretac. Res.
800 125, 104881 (2021).
801
802 12) Rich, T.H., Vickers-Rich, P., Fernández, M. & Santillana, S. A probable hadrosaur from
803 Seymour Island, Antarctic Peninsula. Proceedings of the Second Gondwanan Dinosaur
804 Symposium 15, 219–222 (1999).
805
806 13) Case, J. A. et al. The first duck-billed dinosaur (Family Hadrosauridae) from Antarctica.
807 J. Vertebr. Paleontol. 20, 612–614 (2000).
808
809 14) Casamiquela,R.M. La zona litoral de la transgresión Maastrichtiense en el norte de la
810 Patagonia. Aspectos ecológicos. Ameghiniana 15, pp. 137–148 (1978).
811
812 15) Aguirre-Urreta, B., Tunik, M., Naipauer, M., Pazos, P., Ottone, E., Fanning, M.,
813 Ramos,V.A. Malargüe Group (Maastrichtian–Danian) deposits in the Neuquén Andes,
814 Argentina: Implications for the onset of the first Atlantic transgression related to Western
815 Gondwana break-up. Gond. Res., 19, 482–494 (2011).
816
817 16) Malumián, N., Náñez, C. The Late Cretaceous – Cenozoic transgressions in Patagonia
818 and the Fuegian Andes: foraminifera, palaeoecology and palaeogeography Biological Journal
819 of the Linnean Society, 103 (2011), pp. 269-288
820
821 17) Gasparini, Z., Sterli, J., Parras, A., O'Gorman, J.P., Salgado, L., Varela, J., Pol, D. Late
822 Cretaceous reptilian biota of the La Colonia Formation, central Patagonia, Argentina:
823 Occurrences, preservation and paleoenvironments. Cret. Res., 54, pp. 154–168 (2015).
824
825 18) Jujihara, T., Soto-Acuña, S., Vargas, A., Stinnesbeck, W., Vogt, M., Rubilar-Rogers, D.,
826 Leppe, M. The southermost dinosaurs of South America. In: XXIII Latin American
827 Colloquium. Heidelberg, 88 (2014).
828
829 19) Soto-Acuña,S., Jujihara, T., Novas, F.E., Leppe, M., González, E., Stinnesbeck, W.,
830 Isasi, M.P., Rubilar-Rogers, D., Vargas, A.O. Hadrosaurios (Ornithopoda: Hadrosauridae) en
831 el Cretácico Superior del extremo austral de América del Sur. In: Actas del IV Simposio
832 Paleontología en Chile. Valdivia, 73 (2014).
833
834 20) Brett-Surman, M.K.,Wagner, J.R. in Horns and Beaks: Ceratopsian and Ornithopod
835 Dinosaurs (ed. Carpenter, K.) 135–169 (Indiana University Press, 2007).
836
837 21) Prieto-Márquez, A., Wagner, J.R. Pararhabdodon isonensis and Tsintaosaurus
838 spinorhinus: a new clade of lambeosaurine hadrosaurids from Eurasia. Cretac. Res. 30,
839 1238–1246 (2009).
840
841 22) Hone, D. W. E., Farke, A. & Wedel, M. J. Ontogeny and the fossil record: What, if
842 anything, is an adult dinosaur?. Biol. Lett. 12, 20150947 (2016).
843
844 23) Godefroit, P., Bolotsky, Y. L. & Van Itterbeeck, J. The lambeosaurine dinosaur
845 Amurosaurus riabinini, from the Maastrichtian of Far Eastern Russia. Acta Palaeontol. Pol.
846 49, 585–618 (2004a).
847
848 24) Pereda-Suberbiola, X. et al. The last hadrosaurid dinosaurs of Europe: a new
849 lambeosaurine from the uppermost Cretaceous of Arén (Huesca, Spain). Comptes Rendus
850 Palevol 8, 559–572 (2009).
851
852 25) Brink, K. S., Zelenitsky, D. K., Evans, D. C., Therrien, F., Horner, J. R. A sub-adult skull
853 of Hypacrosaurus stebingeri (Ornithischia: Lambeosaurinae): anatomy and comparison. Hist.
854 Biol. 23, 63–72 (2011).
855
856 26) Rogers, R.R. Taphonomy of three dinosaur bone beds in the Upper Cretaceous Two
857 Medicine Formation of northwestern Montana: evidence for drought−related mortality.
858 Palaios 5, 394–413 (1990).
859
860 27) Holland, B. et al. Taphonomy and taxonomy of a juvenile lambeosaurine (Ornithischia:
861 Hadrosauridae) bonebed from the late Campanian Wapiti Formation of northwestern Alberta,
862 Canada. PeerJ 9, e11290 (2021).
863
864 28) Fernández Garay, A. Diccionario Tehuelche-Español/Índice Español-Tehuelche. Países
865 Bajos: Escuela de Investigación de Estudios Asiáticos, Africanos y Amerindios (CNWS),
866 (Universidad de Leiden, 2004).
867
868 29) Schwartz, T. M., Fosdick, J. C., Graham, S. A. Using detrital zircon U-Pb ages to
869 calculate Late Cretaceous sedimentation rates in the Magallanes-Austral basin, Patagonia.
870 Basin Research 29(6), 725–746 (2016).
871
872 30) Gutiérrez, N. M. et al. Tectonic events reflected by palaeocurrents, zircon
873 geochronology, and palaeobotany in the Sierra Baguales of Chilean Patagonia.
874 Tectonophysics 695, 76–99 (2017).
875
876 31) Horner, J.R., Weishampel, D.B., Foster, C.A. In The Dinosauria (eds. Weishampel, D.B.,
877 Dodson, P., Osmólska, H.) 438–463 (Univ. California Press, 2004).
878
879 32) Prieto-Márquez, A., Erickson, G.M., Ebersole, J.A. A primitive hadrosaurid from
880 southeastern North America and the origin and early evolution of ‘duck-billed’ dinosaurs. J.
881 Vertebr. Paleontol. 36, e1054495 (2016a).
882
883 33) Godefroit, P., Dong, Z.-M., Bultynck, P., Li, H. & Feng, L. Sino-Belgian Cooperative
884 Program. Cretaceous dinosaurs and mammals: 1) New Bactrosaurus (Dinosauria:
885 Hadrosauroidea) material from Iren Dabasu (Inner Mongolia, P. R. China). Bulletin de
886 l’Institut royal des sciences naturelles de Belgique 68, 1–70 (1998).
887
888 34) Wagner, J. R. Hard-tissue homologies and their consequences for interpretation of the
889 cranial crests of lambeosaurine dinosaurs (Dinosauria: Hadrosauria). J. Vertebr. Paleontol.
890 24, 125A–126A (2004).
891
892 35) Dalla Vecchia, F. M. Tethyshadros insularis, a new hadrosauroid dinosaur (Ornithischia)
893 from the Upper Cretaceous of Italy. J. Vertebr. Paleontol. 29, 1100–1116 (2009).
894
895 36) Prieto-Márquez, A. Cranial and appendicular ontogeny of Bactrosaurus johnsoni, a
896 hadrosauroid dinosaur from the Late Cretaceous of northern China. Palaeontology 54, 773–
897 792 (2011a).
898
899 37) Tsogtbaatar, K., Weishampel, D.B., Evans, D.C. & Watanabe, M. in Hadrosaurs (eds.
900 Eberth, D. & Evans, C.), 108–135 (Indiana University Press, 2014).
901
902 38) Bell, P. R. Cranial osteology and ontogeny of Saurolophus angustirostris from the Late
903 Cretaceous of Mongolia with comments on Saurolophus osborni from Canada. Acta
904 Palaeontol. Pol. 56, 703–722 (2011).
905
906 39) McDonald, A.T., Bird, J., Kirkland, J.I., Dodson, P. Osteology of the basal hadrosauroid
907 Eolambia caroljonesa (Dinosauria: Ornithopoda) from the Cedar Mountain Formation of
908 Utah. PLoS ONE 7, e45712 (2012).
909
910 40) Prieto-Márquez, A. Global phylogeny of Hadrosauridae (Dinosauria: Ornithopoda) using
911 parsimony and Bayesian methods. Zoo. J. Linnean Soc. 159, 435–502 (2010b).
912
913 41) Campione, N.E., Brink, K.S., Freedman, E.A., McGarrity, C.T. & Evans, D.C.
914 ‘Glishades ericksoni’, and indeterminate juvenile hadrosaurid from Two Medicine Formation
915 of Montana: implications for hadrosauroid diversity in the latest Cretaceous (Campanian-
916 Maastrichtian) of western North America. Palaeobio Palaeoenv 93, 65–75 (2012).
917
918 42) Tsogtbaatar, K., Weishampel, D.B., Evans, D.C., Watabe, M. A new hadrosauroid
919 (Dinosauria: Ornithopoda) from the Late Cretaceous Baynshire Formation of the Gobi Desert
920 (Mongolia). PLoS ONE 14, e0208480 (2019).
921
922 43) Sues, H.-D., Averianov, A. A new basal hadrosauroid dinosaur from the Late Cretaceous
923 of Uzbekistan and the early radiation of duck-billed dinosaurs. Proc. R. Soc. B 276, 2549–
924 2555 (2009).
925
926 44) Prieto-Márquez, A., Erickson, G.M. & Ebersole, J.A. Anatomy and osteohistology of the
927 basal hadrosaurid dinosaur Eotrachodon from the uppermost Santonian (Cretaceous) of the
928 southern Appalachia. PeerJ 4, e1842 (2016b).
929
930 45) Prieto-Márquez, A., Norell, M. A. Anatomy and Relationships of Gilmoreosaurus
931 mongoliensis (Dinosauria: Hadrosauroidea) from the Late Cretaceous of Central Asia.
932 American Museum Novitates 3694, 1–49 (2010).
933
934 46) Prieto-Márquez, A.; Gaete, R.; Ribas, G.; Galobart, A.; Boada, M. Hadrosauroid
935 dinosaurs from Western Europe: Pararhabdodon isonensis and Koutalisaurus kohlerorum n.
936 gen. et. sp. J. Vert. Paleontol. 26, 929–943 (2006).
937
938 47) Prieto-Márquez, A. Phylogeny and historical biogeography of hadrosaurid dinosaurs.
939 Ph.D. Thesis, Florida State University, 936 pp (2008).
940
941 48) Xing, H. et al. A new basal hadrosauroid dinosaur (Dinosauria: Ornithopoda) with
942 transitional features from the Late Cretaceous of Henan Province, China. PLoS ONE 9,
943 e98821 (2014).
944
945 49) Prieto-Márquez, A., Fondevilla, V., Sellés, A.G., Wagner, J.R., Galobart, À.
946 Adynomosaurus arcanus, a new lambeosaurine dinosaur from the Late Cretaceous Ibero-
947 Armorican Island of the European archipelago. Cretac. Res. 96, 19–37 (2019).
948
949 50) Prieto-Márquez, A., Carrera Farias, M.Á. A new late-surviving early diverging Ibero-
950 Armorican duck-billed dinosaur and the role of the Late Cretaceous European Archipelago in
951 hadrosauroid biogeography. Acta Palaeontol. Pol. 66, 425–435 (2021).
952
953 51) Weishampel, D. B., Horner, J.R. in The Dinosauria (eds. Weishampel, D.B. et al.), 534–
954 561 (Univ. California Press, 1990).
955
956 52) Prieto-Márquez, A. Revised diagnoses of Hadrosaurus foulkii Leidy, 1858 (the type
957 genus and species of Hadrosauridae Cope, 1869) and Claosaurus agilis Marsh, 1872
958 (Dinosauria: Ornithopoda) from the Late Cretaceous of North America. Zootaxa 2765, 61–68
959 (2011b).
960
961 53) Prieto-Márquez, A., Weishampel, D.B., Horner, J.R. The hadrosaurid dinosaur
962 Hadrosaurus foulkii from the Campanian of the East coast of North America, with a review
963 of the genus. Acta Palaeontol. Pol. 51, 77–98 (2006).
964
965 54) Ramírez-Velasco, A.A., Benammi, M., Prieto-Márquez, A., Ortega, J.A., Hernández-
966 Rivera, R. Huehuecanauhtlus tiquichensis, a new hadrosauroid dinosaur (Ornithischia:
967 Ornithopoda) from the Santonian (Late Cretaceous) of Michoacán, Mexico. Can. J. Earth
968 Sci. 49, 379–395 (2012).
969
970 55) Wang, R.-F. et al. A second hadrosauroid dinosaur from the early Late Cretaceous of
971 Zuoyun, Shanxi Province, China. Hist. Biol. 29, 17–24 (2015).
972
973 56) Gates, T.A., Tsogtbaatar, K., Zanno, L.E., Chinzorig, T., Watabe, M. A new
974 iguanodontian (Dinosauria: Ornithopoda) from the Early Cretaceous of Mongolia. PeerJ 6,
975 e5300 (2018).
976
977 57) Norman, D.B. Anatomy of Iguanodon atherfieldensis (Ornithischia: Ornithopoda). Bull.
978 Inst. r. Sci. nat. Belg.: Sciences de la Terre 56, 281-372 (1986).
979
980 58) Norman, D. B. On Asian ornithopods (Dinosauria: Ornithischia). 4. Probactrosaurus
981 Rozhdestvensky, 1966. Zoo. J. of the Linnean Soc. 136, 113–144 (2002).
982
983 59) Wang, R-F. et al. A New Hadrosauroid Dinosaur from the Early Late Cretaceous of
984 Shanxi Province, China. PLoS ONE 8, e77058 (2013).
985
986 60) Qing-Wei, T., Hai, X., Yong-Guo, H., Lin, T., Xing, X. New hadrosauroid material from
987 the Upper Cretaceous Majiacun Formation of Hubei Province, central China. Vertebrata
988 PalAsiatica 53, 245–264 (2015).
989
990 61) Gates, T.A., Lamb, J.P. 2021. A redescription of Lophorhothon atopus (Ornithopoda:
991 Dinosauria) from the Late Cretaceous of Alabama based on new material. Canadian Journal
992 of Earth Sciences 58(9), https://doi.org/10.1139/cjes-2020-0173 (2021).
993
994 62) Prieto-Márquez, A. Skeletal morphology of Kritosaurus navajovius (Dinosauria:
995 Hadrosauridae) from the Late Cretaceous of the North American south-west, with an
996 evaluation of the phylogenetic systematics and biogeography of Kritosaurini. J. Syst.
997 Palaeontol., 12, 2, 133-175 (2014b).
998
999 63) Lamanna, M.C. et al. Late Cretaceous non-avian dinosaurs from the James Ross Basin,
1000 Antarctica: description of new material, updated synthesis, biostratigraphy, and
1001 paleobiogeography. Advances in Polar Sciences 30(3), 228–250 (2019)
1002
1003 64) Lamanna, M.C, Casal, G.A., Ibiricu, L.M., Martínez, R.D.F. Non-Avian dinosaur faunas
1004 of the Upper Cretaceous Bajo Barreal and Lago Colhué Huapi formations of Central
1005 Patagonia, Argentina: Paleobiogeographic and Biostratigraphic implications. Actas del XXI
1006 Congreso Geológico Argentino, Sesión Técnica VI, Paleontología, Paleoecología y
1007 Bioestratigrafía, Puerto Madryn, Chubut, 259–260 (2022).
1008
1009 65) Scotese, C.R. Atlas of Late Cretaceous Maps, PALEOMAP Atlas for ArcGIS, volume 2,
1010 The Cretaceous, Maps 16 – 22, Mollweide Projection, PALEOMAP Project, Evanston, IL
1011 (2014).
1012
1013 66) Goin, F. J., Gelfo, J. N., Chornogubsky, L., Woodburne, M. O., Martin, T. Origins,
1014 radiations, and distribution of South American mammals: from greenhouse to icehouse
1015 worlds. Pp. 20–50 in B. D. Patterson & L. P. Costa (eds) Bones, clones, and biomes: an 80-
1016 million year history of recent Neotropical mammals. University of Chicago Press, Chicago.
1017 doi:10.7208/9780226649214-003 (2021).
1018
1019 67) Goin, F. J., Woodburne, M. O., Zimicz, A. N., Martin, G. M., Chornogubsky, L.
1020 Dispersal of vertebrates from between the Americas, Antarctica, and Australia in the Late
1021 Cretaceous and Early Cenozoic. In A brief history of South American Metatherians (pp. 77–
1022 124). Springer, Dordrecht (2016).
1023
1024 68) Carneiro,L.M., Oliveira, É.V., Goin, F.J. Autrapediomys marshalli gen. et. sp., a new
1025 Pediomyoidea (Mammalia, Methateria) from the Paleogene of Brazil: paleobiogeographic
1026 implications. Revista Brasileira de Paleontologia 21(2), 120–131 (2018).
1027
1028 69) Riguetti, F., Pereda-Suberbiola, X., Ponce, D., Salgado, L., Apesteguía, S., Rozadilla, S.,
1029 Arbour,V. A new small-bodied ankylosaurian dinosaur from the Upper Cretaceous of North
1030 Patagonia (Río Negro Province, Argentina), Journal of Systematic Palaeontology, 20:1,
1031 2137441, DOI:10.1080/14772019.2022.2137441 (2022).
1032
1033 70) Tykoski,R.S., Fiorillo, A.R. An articulated cervical series of Alamosaurus sanjuanensis
1034 Gilmore, 1922 (Dinosauria, Sauropoda) from Texas: new perspective on the relationships of
1035 North America's last giant sauropod. Journal of Systematic Palaeontology 15, 5, 339–364
1036 (2017).
1037
1038 71) Ezcurra, M.D., Agnolín, F.L. A new Palaeobiogeographical Model for the Late Mesozoic
1039 and Early Tertiary. Sys.Biol.64(4), 553–566 (2012).
1040
1041 72) Prieto-Márquez, A. Global historical biogeography of hadrosaurid dinosaurs. Zoo. J.
1042 Linnean Soc. 159, 503-525 (2010a).
1043
1044 73) Cruzado-Caballero, P., Canudo, J. I., Moreno-Azanza, M., Ruiz-Omeñaca, J. I. New
1045 material and phylogenetic position of Arenysaurus ardevoli, a lambeosaurine dinosaur from
1046 the late Maastrichtian of Arén (northern Spain). Journal of Vertebrate Paleontology 33(6),
1047 1367–1384 (2013).
1048
1049 74) Matzke, N. BioGeoBEARS: BioGeography with Bayesian (and likelihood) evolutionary
1050 analysis in R Scripts. CRAN: The Comprehensive R Archive Network, Vienna, Austria.
1051 http://cran.r-project.org/package¼BioGeoBEARS (2013).
1052
1053 75) Ronquist, F. Dispersal-vicariance analysis: a new approach to the quantification of
1054 historical biogeography. Syst. Biol. 46, 195–203 (1997).
1055
1056 76) Nylander, J. A. A., Olsson, U., Alström, P., Sanmartín, I. Accounting for phylogenetic
1057 uncertainty in biogeography: a Bayesian approach to Dispersal-Vicariance Analysis of the
1058 thrushes (Aves: Turdus). Syst. Biol. 57, 257–268 (2008).
1059
1060 77) Yu, Y., Harris, A.J., He, X. S-DIVA (Statistical Dispersal-Vicariance Analysis): A tool
1061 for inferring biogeographic histories. Molecular Phylogenetics and Evolution 56, 848–850
1062 (2010).
1063
1064 78) Yu, Y., Harris, A.J., Blair, C., He, X. RASP (Reconstruct Ancestral State in
1065 Phylogenies): A tool for historical biogeography. Molecular Phylogenetics and Evolution 87,
1066 46–49 (2015).
1067
1068 79) Mateus, O. Cretaceous amniotes from Angola: dinosaurs, pterosaurs, mosasaurs,
1069 plesiosaurs, turtles. Actas de V Jornadas Internacionales sobre Paleontología de Dinosaurios
1070 y su Entorno, Salas de los Infantes, Burgos, 71–105 (2012).
1071
1072 80) Buffetaut, E., Hartman, A-F, Al-Kindi, M., Schulp, A.S. Hadrosauroid Dinosaurs from
1073 the Late Cretaceous of the Sultanate of Oman. PLoS ONE 10, e0142692 (2015).
1074
1075 81) Novas, F. E., Pol, D. New evidence on deinonychosaurian dinosaurs from the Late
1076 Cretaceous of Patagonia. Nature 3285, 858–861 (2005).
1077
1078 82) Coria, R. A., Salgado, L. South American ankylosaurs. In: Carpenter, K. (Ed.) The
1079 armored dinosaurs. Indiana University Press, Bloomington, Pp. 159–168 (2001).
1080
1081 83) Arbour, V. M. and Currie, P. J. Systematics, phylogeny and palaeobiogeography of the
1082 ankylosaurid dinosaurs. Journal of Systematic Palaeontology 14(5), 385–444 (2015).
1083
1084 84) Pereda-Suberbiola, X., Díaz-Martínez, I., Salgado, L., de Valais, S. Síntesis del registro
1085 fósil de dinosaurios tireóforos en Gondwana. In: Fernández, M. and Herrera, Y. (Eds.)
1086 Reptiles Extintos – Volumen en Homenaje a Zulma Gasparini. Publicación electrónica de la
1087 Asociación Paleontológica Argentina, 15(1), 90–107 (2015).
1088
1089 85) Makovicky, P.J., Apesteguía, S., Agnolín, F.L. The earliest dromaeosaurid theropod of
1090 South America. Nature 437, 1007–1011 (2005).
1091
1092 86) Ding, A. et al. The biogeography of coelurosaurian theropods and its impact on their
1093 evolutionary history. 117–157. In Pittman, M. and Xu, X. (eds) Pennaraptoran theropod
1094 dinosaurs: Past progress and new frontiers. Bull. Am. Mus. Nat. Hist. 440 (2020).
1095
1096 87) Soto-Acuña, S. et al. Bizarre tail weaponry in a transitional ankylosaur from subantarctic
1097 Chile. Nature 600, 259–263 (2021).
1098
1099 88) Evans, D. C., Barrett, P.M., Seymour, K.L. Revised identification of a reported
1100 Iguanodon-grade ornithopod tooth from the Scollard Formation, Alberta, Canada. Cretaceous
1101 Research 33, 11–14 (2011).
1102
1103 89) Borinder, N.H., Poropat, S.F., Campione, N.E., Wigren, T., Kear, B.P. Postcranial
1104 osteology of the basally branching hadrosauroid dinosaur Tanius sinensis from the Upper
1105 Cretaceous Wangshi Group of Shandong, China. J. Vertebr. Paleontol. 41, e1914642 (2021).
1106
1107 90) Xing, L., Niu, K., Yang, T-R., Wang, D., Miyashita, T., Mallon, J.C. Hadrosauroid eggs
1108 and embryos from the Upper Cretaceous (Maastrichtian) of Jangxi Province, China. BMC
1109 Ecology and Evolution 22, 60, https://doi.org/10.1186/s12862-022-02012-x (2022).
1110
1111 91) Kubo, T. Biogeographical network analysis of Cretaceous Australian dinosaurs.
1112 Gondwana Research 82, 39–47 (2020).
1113
1114 92) Chiarenza, A.A., Fansworth,A., Mannion, P.D., Allison, P.A. Asteroid impact, not
1115 volcanism, caused the end-Cretaceous dinosaur extinction. Proc. Natl. Acad. Sci. U.S.A. 117,
1116 17084–17093 (2020).
1117
1118 93) Fosdick, J. C. et al. Detrital thermochronologic record of burial heating and sediment
1119 recycling in the Magallanes foreland basin, Patagonian Andes. Basin Research 27(4), 546–
1120 572 (2015).
1121
1122 94) George, S. W. et al. Chronology of deposition and unconformity development across the
1123 Cretaceous–Paleogene boundary, Magallanes-Austral Basin, Patagonian Andes. Journal of
1124 South American Earth Sciences 97, 102237 (2020).
1125
1126 95) Biddle, K. T., Uliana, M. A., Mitchum, R. M. Jr., Fitzgerald, M. G., Wright, R. C. The
1127 stratigraphic and structural evolution of the central and eastern Magallanes Basin, southern
1128 South America. Foreland Basins, Special Publication of the International Association of
1129 Sedimentologists 8, 41–61 (1986).
1130
1131 96) Mella, P. Control tectónico en la evolución de la Cuenca de antepaís de Magallanes, XII
1132 Región, Chile. Memoria para optar al título de Geólogo. Universidad de Concepción,
1133 Departamento de Ciencias de la Tierra (inédito), 149 pp. (2001).
1134
1135 97) Mpodozis, C., Mella, P., Padva, D. Estratigrafía y Megasecuencias sedimentarias en la
1136 Cuenca Austral-Magallanes, Argentina y Chile. VII Congreso de Exploración y Desarrollo
1137 de Hidrocarburos, Simposio Cuencas Argentinas (2011).
1138
1139 98) Katz, H. R. Revision of Cretaceous stratigraphy in Patagonian cordillera of Última
1140 Esperanza, Magallanes Province, Chile. American Association of Petroleum Geologist
1141 Bulletin 47, 506–524 (1963).
1142
1143 99) Daniels, B. G. et al. Timing of deep-water slope evolution constrained by large-n detrital
1144 and volcanic ash zircon geochronology, Cretaceous Magallanes Basin, Chile. GSA Bulletin
1145 130(3-4), 438–454 (2018).
1146
1147 100) Malumián, N., Caramés, A. Upper Campanian-Paleogene from the Río Turbio coal
1148 measures in southern Argentina: micropaleontology and the Paleocene/Eocene boundary.
1149 Journal of South American Earth Sciences 10(2),189–201 (1997).
1150
1151 101) Camacho, H. H., Chiesa, J. O., Parma, S. G. Relaciones estratigráficas entre
1152 formaciones terciarias en el occidente de la Provincia de Santa Cruz. Revista de la
1153 Asociación Geológica Argentina 53(2), 273–281 (1998).
1154
1155 102) Macellari, C. E., Barrio, C. A., Manassero, M. J. Upper Cretaceous to Paleocene
1156 depositional sequences and sandstone petrography of southwestern Patagonia (Argentina and
1157 Chile). Journal of South American Earth Sciences 2(3), 223–239 (1989).
1158
1159 103) Wilson, T. Transition from back arc to foreland basin development in the southern
1160 Andes: stratigraphic record from the Ultima Esperanza District, Chile. Geological Society of
1161 America Bulletin 103, 98–111 (1991).
1162
1163 104) Shultz, M. R., Fildani, A., Cope, T. D., Graham, S. A. Deposition and stratigraphic
1164 architecture of an outcropping ancient slope system: Tres Pasos Formation, Magallanes
1165 Basin, southern Chile. Geol. Soc. Lond. Spec. Publ. 244, 27–50 (2005).
1166
1167 105) Fildani, A., Romans, B. W., Hubbard, S. M., Fosdick, J. C. Tectonic setting of the
1168 Magallanes Basin, Chile. SEPM Field 10, 1–7 (2009).
1169
1170 106) Hubbard, S. M., Fildani, A., Romans, B. W., Covault, J. A., Mchargue, T. R. High-
1171 relief slope clinoform development: Insights from outcrop, Magallanes Basin, Chile. Journal
1172 of Sedimentary Research 80, 357–375 (2010).
1173
1174 107) Romans, B. W. et al. Evolution of deep-water stratigraphic architecture, Magallanes
1175 Basin, Chile. Marine and Petroleum Geology 28(3), 612–628 (2011).
1176
1177 108) McAtamney, J. et al. Alongstrike variability of backarc basin collapse and the
1178 initiation of sedimentation in the Magallanes foreland basin, southernmost Andes (53–
1179 54.5°S). Tectonics 30(5), TC5001 (2011).
1180
1181 109) González, E. J. Estratigrafía secuencial y sedimentología de la Formación Dorotea
1182 (Maastrichtiano), sector Río de las Chinas, Región de Magallanes y Antártica Chilena, Chile
1183 (50°S). BSc thesis, Universidad de Chile (2015)
1184
1185 110) Manríquez, L. M., Lavina, E. L., Fernández, R. A., Trevisan, C., Leppe, M. A.
1186 Campanian-Maastrichtian and Eocene stratigraphic architecture, facies analysis, and
1187 paleoenvironmental evolution of the northern Magallanes Basin (Chilean Patagonia). Journal
1188 of South American Earth Sciences 93, 102–118 (2019).
1189
1190 111) Manríquez, L.M.E, Lavina, E. L., Netto, R., Horodyski, R., Leppe, M. Evolution of a
1191 high latitude high-energy beach system (Maastrichtian–Eocene, Magallanes/Austral Basin,
1192 Chilean Patagonia). Sedimentary Geology 426, 106026.

1193 https://doi.org/10.1016/j.sedgeo.2021.106026 (2021).
1194
1195 112) Rivera, H. A. et al. Tectonic controls on the Maastrichtian-Danian transgression in the
1196 Magallanes-Austral foreland basin (Chile): Implications for the growth of the Southern
1197 Patagonian Andes. Sedimentary Geology 403, 105645 (2020).
1198
1199 113) Covault, J. A., Romans, B. W. & Graham, S. A. Outcrop expression of a continental-
1200 margin-scale shelf-edge delta from the Cretaceous Magallanes Basin, Chile. Journal of
1201 Sedimentary Research 79(7), 523–539 (2009).
1202
1203 114) Schwartz, T. M., Graham, S. A. Stratigraphic architecture of a tideinfluenced
1204 shelfedge delta, Upper Cretaceous Dorotea Formation, MagallanesAustral Basin,
1205 Patagonia. Sedimentology 62(4), 1039–1077 (2015).
1206
1207 115) Cecioni, G. Etá della flora del Cerro Guido estratigrafia del Departamento Ultima
1208 Esperanza, provincia di Magellano, Cile. Estratto dal Bollettino della Societa Geologica
1209 Italiana 76, 1–15 (1957).
1210
1211 116) Cortés, R. 1964. Informe Geológico del área Río Las Chinas – Río Bandurrias (Última
1212 Esperanza). Empresa Nacional del Petróleo, Magallanes, Punta Arenas, Chile, 33 pp.
1213
1214 117) Suazo Lara, F. et al. Primer registro de Calyptocephalellidae (Anura, Australobatrachia)
1215 en el Cretácico Superior de Chile. Libro de Resúmenes de la Reunión de Paleontología de
1216 Vertebrados de Chile 1, 1–17 (Santiago, 2017).
1217
1218 118) Alarcón-Muñoz, J. et al. Freshwater turtles (Testudines: Pleurodira) in the Upper
1219 Cretaceous of Chilean Patagonia. Journal of South American Earth Sciences 102, 102652
1220 (2020).
1221
1222 119) Goin, F. J. et al. First Mesozoic mammal from Chile: The Southernmost record of a
1223 Late Cretaceous Gondwanatherian. Boletín del Museo Nacional de Historia Natural, Chile
1224 69(1), 5–31 (2020).
1225
1226 120) Trevisan, C., Dutra, T., Wilberger, T., Leppe, M., Manríquez, L. An austral fern
1227 assemblage from the Upper Cretaceous (Campanian) beds of Cerro Guido, Magallanes
1228 Basin, Chilean Patagonia. Cretac. Res. 106, 104215 (2020).
1229
1230 121) Martinelli, A.G. et al. New cladotherian mammal from southern Chile and the evolution
1231 of mesungulatid meridiolestidans at the dusk of the Mesozoic era. Scientific Reports 11, 7594
1232 (2021).
1233
1234 122) Suazo Lara, F., Gómez, R.O. In the shadow of dinosaurs: Late Cretaceous frogs are
1235 distinct components of a widespread tetrapod assemblage across Argentinian and Chilean
1236 Patagonia. Cretac. Res. 131, 105085 (2022).
1237
1238 123) Prieto-Márquez, A., Salinas, G.C. A re-evaluation of Secernosaurus koerneri and
1239 ‘Kritosaurus’ australis (Dinosauria, Hadrosauridae) from the Late Cretaceous of Argentina.
1240 J. Vertebr. Paleontol. 30, 813–837 (2010).
1241
1242 124) Zhang,J., Wang,X.,Wang,Q., Jiang,S.,Cheng,X.,Qiu,R. A new saurolophine hadrosaurid
1243 (Dinosauria:Ornithopoda) from the Upper Cretaceous of Shandong, China. Anais da
1244 Academia Brasileira de Ciências, http://dx.doi.org/10.1590/0001-3765201720160920 (2017)
1245
1246 125) Evans, D.C., Reisz, R.R. Anatomy and relationships of Lambeosaurus magnicristatus, a
1247 crested hadrosaurid dinosaur (Ornithischia) from The Dinosaur Park Formation, Alberta.
1248 Journal of Vertebrate Paleontology 27(2), 373–393 (2007).
1249
1250 126) Weishampel, D. B, D. B. Norman, Grigorescu, D. Telmatosaurus transsylvanicus from
1251 the Late Cretaceous of Romania: the most basal hadrosaurid dinosaur. Palaeontology 36,
1252 361–385 (1993).
1253
1254 127) Maddison, W. P., Maddison, D. R. Mesquite: a Modular System for Evolutionary
1255 Analysis. Version 3.51. http://www.mesquiteproject.org (2018).
1256
1257 128) Goloboff, P.A., Catalano, S.A. TNT version 1.5, including a full implementation of
1258 phylogenetic morphometrics. Cladistics 3, 221–238 (2016).
1259
1260 129) Bremer, K. 1988. The limits of amino acid sequence data in angiosperm phylogenetic
1261 reconstruction. Evolution 42, 795–803.
1262
1263 130) Goloboff, P., Farris, J.C., Nixon, K.C. TNT, a free program for phylogenetic analysis.
1264 Cladistics 24, 774–786 (2008).
1265
1266 131) Felsenstein, J. Phylogenies and the comparative method. The American Naturalist 1, 1–
1267 15 (1985).
1268
1269 132) Templeton, A.R. Phylogenetic inference from restriction endonuclease cleavage site
1270 maps with particular reference to the evolution of humans and the apes. Evolution 37, 221–
1271 244 (1983).
1272
1273 133) Schmidt-Lebuhn, A.N. TNT script for the Templeton Test. DOI:
1274 10.13140/RG.2.1.3484.9522 (2016).
1275
1276 134) Bapst, David W. paleotree: an R package for paleontological and phylogenetic analyses
1277 of evolution. Methods Ecol. Evol. 3, 803–807 (2012).
1278
1279 135) Ronquist, F. et al. Mrbayes 3.2: efficient Bayesian phylogenetic inference and model
1280 choice across a large model space. Syst. Biol. 61, 539–542 (2012).
1281
1282 136) Lewis, P. A likelihood approach to estimating phylogeny from discrete morphological
1283 character data. Soc. Syst. Biol. 50, 913–925 (2001).
1284
1285 137) Püschel, H. P., O’Reilly, J. E., Pisani, D. & Donoghue, P. C. J. The impact of fossil
1286 stratigraphic ranges on tip-calibration, and the accuracy and precisión of divergence time
1287 estimates. Palaeontology 63, 67–83 (2020).
1288
1289 138) Paradis, E., Claude, J., Strimmer, K. APE: analyses of phylogenetics and evolution in R
1290 language. Bioinformatics 20, 289–290 (2004).
1291
1292 139) Delignette-Muller, M. L., Dutang, C. fitdistrplus: An R Package for Fitting
1293 Distributions. Journal of Statistical Software, 64(4), 1–34 (2015).
1294
1295 140) R. Core Team R: A Language and Environment for Statistical Computing (R
1296 Foundation for Statistical Computing, 2019).
1297
1298 141) Gunnell, G.F. et al. Fossil lemurs from Egypt and Kenya suggest an African origin for
1299 Madagascar’s aye-aye. Nat. Comm. 9, 3193, https://doi.org/10.1038/s41467-018-05648-w
1300 (2018).
1301
1302 142) Harris, A.J., Xiang, Q-Y. Estimating ancestral distributions of lineages with uncertain
1303 sister groups: a statistical approach to Dispersal-Vicariance Analysis and a case using
1304 Aesculus L (Sapindaceae) including fossils. J. Syst. Palaeontol. 47, 349–368 (2009).
1305
1306 143) Matzke, N.J. Model selection in historical biogeography reveals that founderevent
1307 speciation is a crucial process in island clades. Syst. Biol. 63, 951–970 (2014).
1308
1309 144) O’Reilly, J. E., & Donoghue, P. C. The efficacy of consensus tree methods for
1310 summarizing phylogenetic relationships from a posterior sample of trees estimated from
1311 morphological data. Syst. biol. 67(2), 354–362 (2018).
1312
1313 145) Ree, R.H., Smith, S.A. Maximum likelihood inference of geographic range evolution by
1314 dispersal, local extinction, and cladogenesis. Syst. Biol. 57, 4e14 (2008).
1315
1316 146) Landis, M.J., Matzke, N.J., Moore, B.R., Huelsenbeck, J.P. Bayesian analysis of
1317 biogeography when the number of areas is large. Syst. Biol. 62, 789–804 (2013).
1318
1319 Supplementary References
1320
1321 147) Behrensmeyer, A.K. Terrestrial vertebrate accumulations. In Taphonomy: Releasing the
1322 Data Locked in the Fossil Record. Edited by A. Allison and D.E.G. Briggs. Plenum Press,
1323 New York, NY, 291–335 (1991).
1324
1325 148) Rogers, R.R., Kidwell, S.M. A conceptual framework for the genesis and analysis of
1326 vertebrate skeletal concentrations. In Bonebeds: Genesis, Analysis, and Paleobiological
1327 Significance. Edited by R.R. Rogers, D.A. Eberth, and A.R. Fiorillo. University of Chicago
1328 Press, Chicago, 1–64 (2007).
1329
1330 149) Voorhies, M.R. 1969. Taphonomy and Population Dynamics of an Early Pliocene
1331 Vertebrate Fauna, Knox County, Nebraska. Contributions to Geology, Special Paper No. 1,
1332 69 PP.
1333
1334 150) Behrensmeyer, A.K. The taphonomy and paleoecology of Plio-Pleistocene vertebrate
1335 assemblages east of Lake Rudolf, Kenya. Bulletin of the Museum of Comparative Zoology, v.
1336 146, p. 473–578 (1975).
1337
1338 151) Ryan, M.J., Russell, A.P., Eberth, D.A., Currie, P.J. The Taphonomy of a Centrosaurus
1339 (Ornithischia: Ceratopsidae) bone bed from the Dinosaur Park Formation (upper
1340 Campanian), Alberta, Canada, with comments on cranial ontogeny. PALAIOS, 16, 482–506
1341 (2001).
1342
1343 152) Bell, P., Campione, N.E. Taphonomy of the Danek Bonebed: a monodominant
1344 Edmontosaurus (Hadrosauridae) bonebed from the Horseshoe Canyon Formation, Alberta:
1345 Canadian Journal of Earth Sciences, v. 51, p. 1–15 (2014).
1346
1347 153) Loughney, K., Badgley, C. The influence of depositional environment and basin history
1348 on the taphonomy of mammalian assemblages from The Barstow Formation (middle
1349 Miocene), California. PALAIOS, 2020, v. 35, 175–190 (2020).
1350
1351 154) Fiorillo, A.R. Taphonomy of Hazard Homestead Quarry (Ogallala Group), Hitchcock
1352 County, Nebraska: Contributions to Geology, University of Wyoming, v. 26, 57–97 (1988).
1353
1354 155) Longrich, N. R. 2016. A ceratopsian dinosaur from the Late Cretaceous of eastern North
1355 America, and implications for dinosaur biogeography. Cretac. Res., 57, 199–207
1356
1357 Acknowledgments
1358
1359 To Edwin González, Sebastián Jiménez, Juan Pablo Venegas, Roy Fernández, Juan Pablo
1360 Guevara, María Jesús Ortuya, Sebastían Garrido, Viviana Lobos, Roberto Yury,
1361 Alejandra Manríquez, Marcelo Isasi, Marcela Milani, Sebastián Apesteguía, Jared
1362 Amudeo, Valentina Poblete, Cynthia Cabezas, Antonia Atisha, Daniela Flores and
1363 Claudio Bravo for their valuable help in the excavations and/or technical preparation of
1364 the CPAP 3054 bones; to Estancia Cerro Guido and especially the Matetic, Simunovic
1365 and Reyes families for granting access and important logistical support in the field; to the
1366 Consejo de Monumentos Nacionales (National Monuments Council) of the Chilean
1367 Ministry of Culture, Arts and Heritage for fieldwork permits to Marcelo Leppe. We thank
1368 William F. Simpson (Field Museum of Natural History, USA) for sending us photographs
1369 of the scapula of the holotype of Secernosaurus koerneri. To Martín Ezcurra (Museo
1370 Argentino de Ciencias Naturales “Bernardino Rivadavia”, Buenos Aires, Argentina) for
1371 clarifying doubts about calibration of phylogenetic trees. Finally, we thank Agustín
1372 Martinelli (Museo Argentino de Ciencias Naturales “Bernardino Rivadavia”, Buenos
1373 Aires, Argentina) for allowing us access to the materials of Huallasaurus australis.
1374
1375
1376 Funding:
1377
1378 Agencia Nacional de Investigación y Desarrollo ANID (National Agency for Research
1379 and Development) of the Chilean Ministry of Science, Technology, Knowledge and
1380 Innovation through grants PIA Anillo ACT172099 (A.O.V.),
1381 FONDECYT 1230713 and REDES 190190 (A.O.V.).
1382 FONDECYT 1151389 (to M.L.).
1383 ANID scholarships for PhD studies in Chile (J.A.-M., S.S-A., J.P.-L., J.P.P. and D.B.
1384 BMBF project CHL 10/A09 (W.S., E.F.)
1385 Becas Chile para estudios de Doctorado en el extranjero/2018-72190003, ANID (H.P.)
1386 Jurassic Foundation Grant (J.A.-M.)
1387
1388 Author contributions:
1389
1390 Conceptualization: J.A.M., A.O.V., H.P., P.C.-C.
1391 Methodology: J.A.M., A.O.V., H.P., S.S.-A., J.P.-L., L.M., V.M., C.S.-G., P.C.-C
1392 Investigation: J.A.M., A.O.V., H.P., S.S.-A., L.M., J.P.P., D.B., V.M., C.S.-G., P.C.-C.
1393 Resources: A.O.V., M.L., P.C.-C.
1394 Data Curation: J.A.M., H.P., J.P.-L.
1395 Formal Analysis: J.A.M., H.P., J.P.-L.
1396 Visualization: J.A.M., A.O.V., H.P., L.M., V.M.
1397 Writing—original draft: J.A.M., A.O.V., H.P., P.C.C.
1398 Writing—review & editing: J.A.M., A.O.V., H.P., S.S.-A., L.M., J.K., V.M., C.S.-G.,
1399 J.P.-L., J.P.P., D.B., E.N., H.O., H.M., D.R.-R., P.C.-C
1400 Project administration: A.O.V., M.L.
1401 Funding acquisition: A.O.V., M.L., W.S., E.F.
1402
1403
1404 Competing interests: Authors declare that they have no competing interests.
1405
1406 Data and materials availability: The authors declare that all CPAP 3054 specimens
1407 used in this study are available in the Paleontological Collection of Antarctica and
1408 Patagonia, Instituto Antártico Chileno (INACH), Punta Arenas, Chile. In addition, the
1409 authors declare that all data generated or analyzed during this study are included in this
1410 manuscript, in the Supplemental Material, and as Auxiliary files associated with this
1411 research.
1412
1413
1414
1415
1416
1417
1418
1419
1420
1421
1422 Figures and Tables
1423
1424 Fig. 1. Stratigraphic location of CPAP 3054 and geographic distribution of South American
1425 duck-billed dinosaurs. A, Stratigraphic section of the Dorotea Formation, in which the location
1426 of the bones of CPAP 3054 is indicated. B, Location of the duck-billed dinosaurs from South
1427 America: 1, Lapampasaurus cholinoi (Islas Malvinas locality, La Pampa Province, late
1428 Campanian-early Maastrichtian, Allen Formation); 2, Kelumapusaura machi (Matadero Hill,
1429 General Roca city, Río Negro Province, middle Campanian-early Maastrichtian, Allen
1430 Formation); 3, Bonapartesaurus rionegrensis (Salitral Moreno, General Roca Department, Río
1431 Negro Province, middleCampanian-early Maastrichtian, Allen Formation); 4, Huallasaurus
1432 australis (Arroyo Verde Puelén Departament, Río Negro Province, late Campanian-early
1433 Maastrichtian, Los Alamitos Formation); 5, Secernosaurus koerneri (Río Chico, east of Lake
1434 Colhué Huapi, Chubut Province, Maastrichtian, Lago Colhué Huapi Formation); 6, CPAP 3054
1435 (Río de las Chinas Valley, Magallanes Region, late Campanian-early Maastrichtian, Dorotea
1436 Formation). C, The quarry from which CPAP 3054 bones were excavated (bonebed level); D,
1437 detail of a nearly complete tibia.
1438
1439 Fig. 2. CPAP 3054, skeletal anatomy. A, Bones described in this work (white). Some elements
1440 are indicated specularly to facilitate their representation. B, CPAP 5337, right premaxilla in
1441 lateral view. C, CPAP 5341, incomplete left postorbital in lateral view. D, CPAP 5340,
1442 incomplete right maxilla in lateral view. E, CPAP 5370, left dentary in medial view. F, CPAP
1443 5343, left quadrate in lateral view. G, CPAP 5344, cervical vertebra in anterior view. H, CPAP
1444 5345, dorsal vertebra in anterior view. I, CPAP 5371, right scapula in lateral view. J, CPAP
1445 5352, left sternum in ventral view. K, CPAP 5378, incomplete right rib in anterior view. L,
1446 CPAP 5379, proximal portion of right radius in posterior view. M, CPAP 5355, incomplete left
1447 ulna in anterolateral view. N, CPAP 5353, left humerus in lateral view. O, CPAP 3054
1448 (holotype), right ilium in lateral view. P, CPAP 5356, left postacetabular process in lateral view.
1449 Q, CPAP 5357, proximal portion of left isquion in lateral view. R, CPAP 5363, proximal
1450 portion of right fibula in lateral view. S, CPAP 5360, incomplete right femur in distal view. T,
1451 CPAP 5358, left femur in anterior view. U, CPAP 5362, left tibia in lateral view. V, W, CPAP
1452 5349, caudal vertebra in anterior and lateral views. X, CPAP 5364, right metatarsal III in
1453 anterior view. Scale bars, B-G, J, K-N, Q, S, V-X= 50 mm; H-I, O-P, R, T-U=100 mm.
1454
1455 Fig. 3. CPAP 3054 lacks several derived characters found in Hadrosauridae. A selection of
1456 plesiomorphic characters is shown. A-B, CPAP 5337, right premaxilla in dorsal (A) and lateral
1457 (B) views showing a circumnarial fossa without accessory fossa or foramina (1), and an
1458 anteroventrally deflected oral margin (2). C, CPAP 5343, right quadrate in posterior view. The
1459 lateral condyle is not markedly offset ventrally with respect to the medial condyle, unlike
1460 Hadrosauridae. CPAP 5342, right dentary in medial (D) and dorsal (E) views, and CPAP 5370,
1461 left dentary in medial (F) and dorsal (G) views. Tooth rows do not exceed the coronoid process
1462 (4), the diastema is short (5), there are less than 30 dental positions (6), tooth row converges
1463 anteriorly with the lateral surface of the dentary (7), and the symphysis is oblique (8). CPAP
1464 5353 (H), CPAP 5354 (I), and CPAP 5369 (J), left humeri in posterolateral view, in which the
1465 ratio between length of the deltopectoral crest and the total length of the humerus is lower than
1466 0.48 (9), and the deltopectoral crest is mediolaterally short (10), with a widely arcuate
1467 laterodistal corner (11). K-L, CPAP 3054, right ilium in lateral (K) and dorsal (L) views. The
1468 dorsal border of the ilium is almost straight (12), and the length of the supraacetabular crest is
1469 higher than 70% of the length of the iliac blade (13). M, CPAP 5362, left tibia in lateral view. In
1470 CPAP 3054, the length of the cnemial crest is lower than the 50% of the total length of the bone
1471 (14). Scale bars=50 mm (A-C) and 100 mm (D-M).
1472
1473 Fig. 4. Derived characters of CPAP 3054 shared with Hadrosauridae. G.nanoi supports these
1474 characters evolved previous to the origin of this family. A, CPAP 5337, right premaxilla in
1475 ventral view, which shows a “double-layer” morphology (1). B, CPAP 5340, right maxilla in
1476 lateral view, with a rostral maxillary foramen on the lateral surface of the bone (2). C-D, CPAP
1477 5344, mid-cervical vertebra in anterior (C), and lateral view (D). This cervical vertebra has
1478 elevated and well developed postzygapophyseal processes, which are dorsally arched (3). E,
1479 CPAP 5346, dorsal vertebra whose centrum has a “heart-shaped”articular surface (4). F, CPAP
1480 5347, mid-caudal vertebra with a hexagonal articular surface (5). G, CPAP 5371, right scapula
1481 in medial view, which is dorsally arched (6). H, CPAP 3054, right ilium in lateral view. The
1482 ratio between the base of the preacetabular process (H1) and the distance between the dorsal
1483 border of the ilium and the pubic peduncle (H2) is greater than 0.5 (8). Furthermore, the
1484 supraacetabular process is anteriorly located with respect to the dorsal tuberculum of the ischial
1485 peduncle. I, CPAP 5360, distal end of the right femur in distal view, showing a deep
1486 intercondylar extensor groove, in which the condyles nearly meet anteriorly (9). In addition, the
1487 condyles are posteriorly well-developed (10). J, CPAP 5362, left tibia in proximal view, in
1488 which the posteromedial condyle is wider than the lateral condyle. Scale bars=50 mm (A, F), and
1489 100 mm (B-E, G-J).
1490
1491
1492 Fig. 5. Phylogenetic relationships. Parsimony and Bayesian analyses recovered CPAP 3054 at a
1493 position transitional to Hadrosauridae, diverging shortly before the origin of this family; results
1494 of Parsimony are shown. The DIVA-like + j model (BioGeoBEARS) recovers Laramidia as the
1495 ancestral area for the most recent common ancestor of Eotrachodon and Hadrosauridae (node 1),
1496 as well as for the most recent common ancestor of CPAP 3054 and Hadrosauridae (node 2).
1497 Appalachia or Laramidia are recovered as ancestral areas for Hadrosauridae (node 3), and
1498 Laramidia as the ancestral area for Kritosaurini +Austrokritosauria (node 4). South American
1499 duck-billed dinosaurs are in red.
1500
1501
1502 Fig. 6. Biogeographic history of South American Duck-bill dinosaurs. Biogeographic
1503 statistics suggest that CPAP 3054’s ancestors arrived from North America (in purple). The
1504 alternative hypothesis, that CPAP 3054’s ancestors could have dispersed from Europe into South
1505 America via Africa (Atlantogean hypothesis, in yellow) was not supported by biogeographic
1506 statistics. Hadrosaurids related to the tribe Kritosaurini (in orange) would also have dispersed
1507 from North America, giving rise to the clade Austrokritosauria in South America. These
1508 hadrosaurids may have never reached the same high latitudes as CPAP 3054’s ancestors,
1509 possibly because they arrived later into South America. Paleogeographic drawing is based on
1510 information from the PALEOMAP project of Christopher Scotese and (10).
1511
1512
1513
1514
1515
1516
1517
1518
1519
1520
1521
1522
1523
1524
1525
1526
1527
1528
1529
1530
1531
1532
1533 Supplementary Materials for

1534
1535 Relict duck-billed dinosaurs survived into the last age of the dinosaurs in
1536 subantarctic Chile
1537
1538
1539 Jhonatan Alarcón-Muñoz et al.
1540
1541 *Corresponding author. Email: jhoalarc@gmail.com; alexvargas@uchile.cl
1542
1543
1544
1545
1546
1547 This PDF file includes:
1548
1549 Supplementary Text
1550 Figs. S1 to S15
1551 Tables S1 to S7
1552
1553
1554
1555
1556
1557
1558 Supplementary Text

1559
1560 Taphonomy
1561
1562 Different analyses were carried out including hydraulic equivalence, bone modification and assemblage
1563 data analyses, and Voorhies groups classification. The overall study of taphonomic aspects followed traditional
1564 literature (147,148). Only bones of duck-billed dinosaurs were recovered at the site, as is common in the fossil
1565 record of these dinosaurs, probably related to their gregarious behavior (e.g., 23,27). Most of the bones were found
1566 scattered without superposition, and without any evident articulation. The size of the elements ranges between 4.8
1567 cm and 48.1 cm, the former value corresponding to the length of the smallest caudal vertebral centrum (CPAP 5351)
1568 and the latter to the length of the largest femur (CPAP 5359). 62.5% of the elements are complete and 94% present
1569 some extent of fracture. A minimal number of three individuals (MNI) have been excavated based on the number of
1570 right femora. Relative abundance of elements from each skeletal region is shown in Table S1.
1571
1572 The transport of a bone element in water depends on multiple factors, such as shape, density, size and
1573 degree of articulation, in addition to the flow velocity. Disarticulated bones form distinctive classification groups
1574 (“Voorhies groups”) according to the flow speed to which they are subjected (149,150; for dinosaurs, see 151 and
1575 152). Group 1 corresponds to bones of lower density and relatively low mass, which are affected by light currents
1576 (vertebrae and phalanges, for example); group 2 is formed by elements that are eliminated gradually, mainly due to
1577 traction; and group 3 corresponds to bones that tend to resist transport (due to their greater density and size). We
1578 found a predominance of Group 3 elements, which is consistent with the interpretation of an original deposit with
1579 the influence of low stream current transportation (see Table S2). Fluvial transport is therefore the most likely
1580 explanation for the disarticulation and scattering of the elements within the quarry.
1581
1582 Hydraulic equivalence allows comparing the settling velocity of fossils with that of clasts to estimate
1583 whether they were deposited at similar flow velocities, that is, by a similar process (150). Assemblages that
1584 accumulated due to fluvial processes should have a high proportion of material in hydraulic equivalence with the
1585 sediment matrix, whereas a lack of hydraulic equivalence indicates that fluvial processes were not the dominant
1586 cause of accumulation (although it does not rule out fluvial influence, 153). The average length of (complete) bones
1587 is 22 cm; large bones range between 48 cm and 32 cm and represent 35% of the complete bones. The remaining
1588 65% range between 24.6 cm and 4.8 cm in length. When comparing these dimensions with quartz grains, using the
1589 equivalence table of (150, p. 496), most of the elements are found to be equivalent to gravel-sized grains. This
1590 means that the current velocity that transported these bones should have also been enough to transport gravel-sized
1591 grains. The matrix of the bonebed however consists predominantly of sandy mudstones with coal lenses and fine-
1592 grained sandstones. This discrepancy between the hydraulic equivalence of the bones and the matrix suggests that
1593 flow competition would have been insufficient to transport the bones a long distance, thus ruling out allochthonous
1594 assembly. This result is consistent with the interpretation based on the Voorhies group classification.
1595 Analysis of pre-fossilization and diagenetic weathering are precluded due to current climatic conditions at
1596 the site (i.e., snow cover during most of the year, temperature fluctuations and strong winds during the summer)
1597 which have an important influence on the state of preservation, accelerating weathering and erosion.
1598
1599 Abrasion stages can be represented by a number from 0 to 3 where 0 represents pristine fossil surfaces
1600 without signs of abrasion, and 3 corresponds to fossil bones with extremely well-rounded edges (151,154). The
1601 elements at the bonebed show low levels of abrasion (20.6% with stage 0 and 79.4% with stage 1) that suggests
1602 minimal transport for most specimens. As already mentioned, practically all the bones are incomplete. 93.8% of the
1603 elements present fractures, most of which correspond to transverse and longitudinal fractures as caused by
1604 trampling, or contemporary exposure of the fossil material to temperature variations (climate) that produce the
1605 expansion and contraction of the material. No spiral fractures have been found in the analyzed fossils.
1606
1607 The different analyses above support the elements were either not transported or underwent minimal
1608 transport. This leads us to conclude that the bones are close to the original thanatocoenosis and probably to the
1609 habitat of these animals, being parauthoctonous and synchronic to some degree. The surface of many of the bones is
1610 damaged due to their exposure to current extreme weather conditions, precluding an appropriate bone modification
1611 analysis. The present data rule out death by predation and scavenging since there are no bite marks on the elements
1612 and no predator teeth have been recovered in the quarry. All lines of taphonomic evidence indicate that the reasons
1613 for the death of the individuals are biological. Considering this interpretation, the dominance of elements from
1614 Voorhies Group III is unusual. However, it may be the result of a flooding event, which would have removed
1615 elements from Voorhies Groups 1 and 2 at a time when the skeletons were exposed.
1616
1617
1618 Reconstruction of biogeographic history
1619
1620 The results of analysis with BioGeoBEARS are summarized in Table S4 and Fig. S10. The best-fit model
1621 in BioGeoBEARS was the DIVALIKE + j model (AICc: 182.21), which supports a Laramidian origin for the last
1622 ancestor shared by CPAP 3054 with Hadrosauridae, and for the last ancestor shared with Eotrachodon, at the
1623 previous node (Fig. S10). The + j versions of all models (DEC, BAYERALIKE, and DIVALIKE) showed a better
1624 fit to the data (Table 1) and supported a Laramidian origin for the last ancestor shared by CPAP 3054 with
1625 Hadrosauridae, and for the last ancestor shared with Eotrachodon. The +j models support a strong role of founder
1626 effects (155), which is consistent with discrete events of dispersal across important barriers as proposed for
1627 Hadrosauroidea.
1628
1629 In s-DIVA, the last ancestor shared by CPAP 3054 and Hadrosauridae is recovered to have inhabited either
1630 Laramidia + Appalachia + South America, or Appalachia + South America, in equal probability. In turn, the
1631 previous last ancestor shared by CPAP 3054 with Eotrachodon inhabited Appalachia rather than Laramidia, as
1632 inferred by BioGeoBEARS (Fig. S11). Despite these differences, the results of s-DIVA are generally consistent with
1633 those of BioGeoBEARS in that they support the arrival of the ancestors of CPAP 3054 from North America.
1634 Possible reasons for the differences of s-DIVA with BioGeoBEARS are that only BioGeoBEARS consider the
1635 geological age of taxa in time-calibrated trees, and that BioGeoBEARS includes a dispersal matrix that considers
1636 different dispersal probabilities between land masses.
1637
1638
1639 Conceptual experiments with a hypothetical African taxon
1640
1641 When the hypothetical African taxon was placed between Eotrachodon and CPAP 3054, the best model
1642 (DIVALIKE + j; see Table S5) supported Laramidia + South America as the ancestral area for the last ancestor that
1643 CPAP 3054 shared with Hadrosauridae, and Laramidia as the most likely area for the last ancestor shared with
1644 Eotrachodon (Fig. S12).
1645 When a hypothetical African taxon was added to our BioGeoBEARS analysis as sister taxon to CPAP
1646 3054, the best fitting model (DIVALIKE + j; see Table S6) did not show significant changes and continued to
1647 support Laramidia as the ancestral area for both the last ancestor that CPAP 3054 shared with Hadrosauridae, and
1648 the last ancestor shared with Eotrachodon (Fig. S14).
1649 When a hypothetical African taxon was added to our s-DIVA analysis in between Eotrachodon and CPAP
1650 3054, this led to no changes in the areas supported for the last common ancestor shared by CPAP 3054 with
1651 Hadrosauridae, or the previous ancestor last shared by CPAP 3054 and Eotrachodon (Fig. S13). If the hypothetical
1652 African taxon was placed as sister of CPAP 3054, Africa is now included among poorly resolved areas for the last
1653 ancestor shared by CPAP 3054 with Hadrosauridae, but Europe and Asia remain absent, and are also absent in the
1654 immediately previous ancestor shared with Eotrachodon, which is still retrieved as North American (Fig. S15).
1655 Overall, in all conceptual experiments, the results of both s-DIVA and BioGeoBEARS continued to support the
1656 arrival of CPAP 3054's ancestors from North America.
1657
1658
1659
1660
1661
1662
1663
1664
1665
1666 Fig. S1. Selected cranial bones of CPAP 3054. CPAP 5337, right premaxilla in anterior view (A). CPAP 5340,
1667 right maxilla in lateral (B) and ventral (C) views. CPAP 5339, incomplete left maxilla in lateral (D) and medial (E)
1668 views. CPAP 5338, incomplete left maxilla of a juvenile individual in lateral (F) and medial (G) views. CPAP
1669 5341, incomplete left postorbital in medial view (H). CPAP 5370, left dentary in lateral (I) and ventral (J) views.
1670 CPAP 5342, incomplete right dentary in lateral (K) and ventral (L) views. CPAP 5368, fragment of right
1671 mandibular ramus in medial (M) and dorsal (N) views. CPAP 5343, right quadrate in anterior (O), medial (P),
1672 proximal (Q) and distal (R) views. Abbreviations: af: alveolar foramina, af: angular facet; as: alveolar sulcus; cdp:
1673 caudodorsal process; cvp: caudoventral process; cnf: circumnarial fossa; cp: coronoid process; dp: dorsal process;
1674 dt: dorsal tubercle; d: denticles; eps: ectopterygoid shelf; epr: ectopterygoid ridge; fo: foramina; jp: jugal process;
1675 jw: jugal wing; lc: lateral condyle; ms: meckelian sulcus; mf: meckelian fossa;; mc: medial condyle; orm: orbital
1676 margin; par: palatine ridge; pw: pterygoid wing; qjn: quadratojugal notch; sy: symphysis; to: tooth. Scale bars: A,
1677 H, O-R (50 mm); B-G, I-N (100 mm).
1678
1679
1680 Fig. S2. Selected axial elements of CPAP 3054. CPAP 5380, mid cervical centrum in anterior (A) and lateral (B)
1681 views. CPAP 5344, mid cervical vertebra in dorsal (C) and posterior (D) views. CPAP 5345, dorsal centrum in
1682 anterior (E) and posterior (F) views. CPAP 5346, dorsal vertebra in posterior (G) and lateral (H) views. CPAP
1683 5377, dorsal neural arch in anterior (I) and lateral (J) views. CPAP 5384, proximal caudal centrum in anterior (K)
1684 and lateral (L) views. CPAP 5385, proximal caudal vertebra in anterior (M) and lateral (N) views. CPAP 5386,
1685 proximal caudal centrum in anterior (O) and lateral (P) views. CPAP 5347 (Q-R), CPAP 5348 (S-T), CPAP 5349
1686 (U-V), CPAP 5350 (W-X) and CPAP 5351 (Y-Z), mid-caudal vertebrae in lateral (Q, S, U, W, Y) and anterior (P,
1687 S, U, W, Y) views. CPAP 5381, right rib in anterior view (A’). CPAP 5382, proximal portion of left rib in anterior
1688 view (B’). CPAP 5383, incomplete left rib in anterior view (C’). Abbreviations: ca: capitulum, nc: neural canal; ns:
1689 neural spine; prz: prezygapophysis; poz: postzygapophysis; tp: transverse process; tu: tuberculum; vc: vertebral
1690 centrum. Scale bars: (50 mm).
1691
1692
1693
1694
1695
1696
1697
1698
1699 Fig. S3. Selected appendicular elements of CPAP 3054. CPAP 5352, incomplete left sternum in dorsal view (A).
1700 CPAP 5371, right scapula in anterior (B) and dorsal (C) views. CPAP 3054, right ilium in medial view (D). CPAP
1701 5356, postacetabular process with part of the iliac blade of a left ilium in dorsal view (E). CPAP 5357, proximal
1702 portion of right ischium in lateral (F) view. CPAP 5353, complete left humerus in ventral (G), anteromedial (H),
1703 proximal (I), and distal (J) views.CPAP 5354, incomplete left humerus of an immature individual in ventral (K),
1704 anteromedial (L), and proximal (M) views. CPAP 5369, incomplete left humerus in ventral (N), anteromedial (O),
1705 and distal (P) views. CPAP 5379, proximal portion of right radius in proximal (Q) and anterior (R) views. CPAP
1706 5355, proximal half of a left ulna in anterior (S), and posteromedial (T) views. CPAP 5373, incomplete right ulna in
1707 anterior (U), and posteromedial (V) views. CPAP 5358, left femur in medial (W), posterior (X), and lateral (Y)
1708 views. CPAP 5359, incomplete right femur in anterior (Z), medial (A’), and posterior (B’) views. CPAP 5360,
1709 distal end of the right femur in anterior (C’), lateral (D’), and posterior (E’) views. CPAP 5361, distal portion of the
1710 right femur of a juvenile individual in anterior (F’), lateral (G’), and posterior (H’) views. CPAP 5363, proximal
1711 portion of right fibula in anterior (I’), medial (J’), and proximal (K’) views. CPAP 5362, complete left tibia in
1712 anterior (L’), and posterior (M’) views. CPAP 5372, incomplete left tibia in lateral (N’), and medial (O’) views.
1713 CPAP 5364, right metatarsal III in medial (P’), proximal (Q’), and distal (R’) views. Abbreviations: am: acetabular
1714 margin; asfr: articular surface for radius; asft: articular surface for the tibia; ap: anterior process; asm-III: articular
1715 surface for the metatarsal II; bs: bicipital sulcus; cf: coracoid facet; cw: cranial wing; cc: cnemial crest; dtip: dorsal
1716 tubercle of the ischial peduncle; dc: deltopectoral crest; em: external malleolus; fh: femoral head; ft: fourth
1717 trochanter; gl: glenoid; gt: great trochanter; ha: handle; hh: humeral head; ip: iliac peduncle; it: inner tubercle; ieg:
1718 intercondylar extensor groove; ifg: intercondylar flexor groove; im: internal malleolus, is: intercondylar sulcus; is:
1719 intercondylar sulcus; lp: lateral process; ldc: lateral distal condyle; lt: lesser trochanter; lc: lateral condyle; lc: lateral
1720 condyle; mp: medial process; mdc: medial distal condyle; mc: medial condyle; op: obturator process; ot: outer
1721 tubercle; op: olecranon process; pap: pseudoacromion process; pp: proximal paddle; pop: posterior process; ppe:
1722 pubic peduncle; pmc: posteromedial condyle; pp: posterior process; pp: pubic process; prap: preacetabular process;
1723 poap: postacetabular process; rc: radial condyle; sap: supraacetabular process; sc: sacral crest; su: sulcus; uc: ulnar
1724 condyle; vtip: ventral tubercle of the ischiadic peduncle. Scale bars: A, F, Q-V (50 mm); B-E, G-P, W, X, Y, Z, A’-
1725 Q’ (100 mm).
1726
1727
1728
1729
1730
1731
1732
1733
1734
1735
1736
1737
1738
1739
1740
1741
1742
1743
1744
1745
1746
1747
1748
1749
1750
1751
1752
1753
1754
1755
1756
1757
1758
1759 Fig. S4. Comparison of the scapula of CPAP 3054 to that of other South American hadrosauroids. (A) CPAP
1760 5371, right scapula in lateral and anterior views. (B) Huallasaurus australis, MACN RN-142, left scapula
1761 (reversed) in lateral and anterior views. (C) Lapampasaurus cholinoi, MPHN-Pv-01, anterior portion of left scapula
1762 (reversed) in lateral and anterior views. (D) Kelumapusaura machi, MPCN-PV-810, right scapula in lateral and
1763 anterior view. (E) Secernosaurus koerneri, FMNH P13423, right scapula in lateral and anterior views. (F)
1764 ‘Willinakaqe salitrarensis’, MPCA-Pv SM 2, left scapula (reversed) in lateral and anterior views. Scale bars=100
1765 mm. Abbreviations: cf: coracoid facet; de: distal expansion; df: deltoid fossa; dr: deltoid ridge; gl: glenoid; pap:
1766 pseudoacromion process; sn: scapular neck; vpgl: ventral process of the glenoid.
1767
1768
1769
1770
1771
1772
1773
1774
1775
1776
1777
1778
1779
1780
1781
1782
1783
1784
1785
1786
1787 Fig. S5. Ilia of South American duck-billed dinosaurs. CPAP 3054 (holotype), right ilium in lateral (A), medial
1788 (B) and dorsal views (C). ‘Willinikaqe salitralensis’ MPCA-Pv SM39, left ilium in lateral (D), medial (E) and
1789 dorsal (F). Secernosaurus koerneri, FMNH P13423, right ilium in lateral (G), medial (H) and dorsal (I) views.
1790 Huallasaurus australis, MACN-RN-02, left ilium in lateral (J), medial (K) and dorsal (L) views. Bonapartesaurus
1791 rionegrensis, MPCA-Pv SM2/49, right ilium in lateral (M), medial (N) and dorsal (O) views. Kelumapusaura
1792 machi, MPCN-PV-811, left ilium (reversed) in lateral (P), medial (Q) and dorsal (R) views. Scale bars=100 mm.
1793 Abbreviations: ac: acetabulum; dtip: dorsal tubercle of the ischial peduncle; sc: sacral crest; sap: supraacetabular
1794 process; ri: ridge; sap: supraacetabular process; sc: sacral crest; pp: pubic process; prap: preacetabular process; poap:
1795 postacetabular process; vtip: ventral tubercle of the ischial peduncle.
1796
1797
1798
1799
1800
1801
1802 Fig. S6. Phylogenetic analyses. Strict consensus tree (parsimony analysis) of 4 MPTs of 1312 steps (CI: 0.396; RI:
1803 0.821). Values above nodes are Bremer support. Values beneath nodes are bootstrap proportions (5000
1804 pseudoreplicates).
1805
1806
1807 Fig. S7. 50% Majority Tree obtained by unconstrained undated Bayesian Analysis.
1808
1809
1810
1811
1812
1813
1814
1815 Fig. S8. 50% Majority rule tree obtained from the unconstrained tip-dated Bayesian analysis. Unlike other
1816 analyses, the clade Hadrosaurinae (Saurlophinae + Hadrosaurus) was recovered. Numbers above nodes correspond
1817 to the posterior probabilities. South American hadrosauroids are in blue. AK: Austrokritosauria.
1818
1819 Fig. S9. Chronostratigraphic uncertainty ranges superimposed on a time-calibrated tree obtained by parsimony
1820 analysis (see Methods for selection and time calibration of this tree). South American hadrosauroids are in blue.
1821
1822
1823
1824
1825
1826 Fig. S10. Results of Biogeographical analyses. Time-calibrated tree with BioGeoBEARS results (DIVALIKE + j
1827 model).
1828
1829
1830
1831 Fig. S11. Results of Biogeographical Analyses with s-DIVA. Node 130: Hadrosauridae; Node 133: CPAP 3054 +
1832 Huehuecanauthlus + Lophorhothon + Hadrosauridae.
1833
1834 Fig. S12. Conceptual experiments with BioGeoBEARS with a “Hypothetical African taxon” between
1835 Eotrachodon and CPAP 3054. Time-calibrated tree with BioGeoBEARS results (DIVALIKE + j model),
1836
1837
1838 Fig. S13. Conceptual experiment with s-DIVA with the “Hypothetical African taxon” between Eotrachodon
1839 and CPAP 3054. Node 131: Hadrosauridae; Node 134: CPAP 3054 + Huehuecanauthlus + Lophorhothon +
1840 Hadrosauridae.
1841
1842
1843 Fig. S14. Time-calibrated tree with BioGeoBEARS results (DIVALIKE + j model), with the “Hypothetical African
1844 taxon” as sister taxon of CPAP 3054.
1845
1846
1847
1848
1849 Fig. S15. Results of s-DIVA with a hypothetical African sister taxon of CPAP 3054. Node 131: Hadrosauridae;
1850 Node 134: CPAP 3054 + Huehuecanauthlus + Lophorhothon + Hadrosauridae.
1851
1852
Representation by skeletal region
Region Element representation
Skull 9 26%
vertebrae 8 24%
pectoral girdle 2 5%
Forelimb 4 12%
Pelvic girdle 3 9%
Hindlimb 8 24%
1853
1854 Table S1. Distribution of skeletal regions and series.
1855
1856
1857
1858
1859
1860
1861
1862
1863
1864
1865
1866
1867
1868
1869
1870
1871
1872
1873
1874
1875
1876
Voorhies groups
Caudal vertebrae 5
I
Cervical vertebrae 1 23.5%
Isquion 1
Metatarsal 1
Ulna 1
II Fibula 1 14.7%
Dorsal vertebrae 2
Sternum 1
Skull elements* 6
Dentaries 3
61.8%
III Humerus 3
Femur 4
Tibia 2
Scapula 1
Ilium 2
1877
1878 Table S2. Relative frequency of elements according to Voorhies groups (based on Ryan et al. 2001). *Skull
1879 elements maybe over-represented because of the disarticulated skull, nevertheless it does not change the group III
1880 predominance, even if a single skull element is considered).
1881
1882
1883
1884
Original As the most basal As basalmost
position hadrosauroid Austrokritosauria
N° Trees 4 12 11
N° Steps 1312 1357 1327
Difference in number of 0 45 15
steps
Templeton Tests N/A significant significant
P value N/A <0.01 <0.05

1885
1886 Table S3. Summary of the results of phylogenetic analyses forcing CPAP 3054 into different positions.
1887 Number of trees, number of steps, and differences in steps are indicated, in addition to Templeton Tests results to
1888 assess its significance.
1889
1890
1891
1892
Model lnL d e j AICc AICc + wt
-2 -12
DEC -129.92 1.57 x 10 1.00 x 10 0.00 263.99 1.20 x 10-18
DEC + j -88.98 1.00 x 10-12 1.00 x 10-12 0.20 184.29 2.43 x 10-1
DIVALIKE -133.33 3.04 x 10-2 1.46 x 10-2 0.00 270.83 3.93 x 10-20
DIVALIKE + j -87.94 1.00 x 10-12 1.00 x 10-12 0.19 182.21 6.88 x 10-1
BAYAREALIKE -164.52 2.46 x 10-2 7.45 x 10-2 0.00 333.19 1.13 x 10-33
BAYAREALIKE + j -90.25 1.00 x 10-7 1.00 x 10-7 0.18 186.82 6.86 x 10-2
1893
1894 Table S4. Models and parameters for each of the analyses with BioGeoBears. Values of log-Likelihood (lnL),
1895 Dispersal (d), Extinction (e), Founder effect (j), Corrected Akaike Information Criterion (AICc), and AICc Weight
1896 (AICc wt) scores from each model implemented. The best model (DIVALIKE + j) was selected based on the lowest
1897 AICc value (and higher AICc + wt).
1898
1899
1900
1901
DEC -137.65 1.85 x 10-2 5.82 x 10-3 0.00 279.47 3.27 x 10-17
DEC + j -101.93 8.72 x 10-4 6.15 x 10-4 0.17 210.19 2.62 x 10-02
DIVALIKE -140.25 3.18 x 10-2 1.75 x 10-2 0.00 284.65 3.45 x 10-18
DIVALIKE + j -98.66 1.38 x 10-3 1,00 x 10-12 0.19 203.65 9.53 x 10-1
BAYAREALIKE -169.42 2.61 x 10-2 7.48 x 10-2 0.00 342.99 5.26 x 10-31
BAYAREALIKE + j -103.13 1.11 x 10-3 1.56 x 10-3 0.19 212.58 1.09 x 10-2
1902
1903 Table S5. Experiment 1, “Hypothetical African Taxon between Eotrachodon and CPAP 3054. Models and
1904 parameters for each of the analyses with BioGeoBears. Values of log-Likelihood (lnL), Dispersal (d), Extinction (e),
1905 Founder effect (j), Corrected Akaike Information Criterion (AICc), and AICc Weight (AICc wt) scores from each
1906 model implemented. The best model (DIVALIKE + j) was selected based on the lowest AICc value (and higher
1907 AICc + wt).
1908

DEC -136.01 1.84 x 10-2 5.61 x 10-3 0.00 276.18 6.65 x 10-20
DEC + j -92.17 1.00 x 10-12 1.00 x 10-12 0.21 190.67 2.46 x 10-1
DIVALIKE -139.03 3.13 x 10-2 1.47 x 10-2 0.00 282.22 3.24 x 10-21
DIVALIKE + j -91.15 1.00 x 10-12 1.00 x 10-12 0.19 188.63 6.83 x 10-1
BAYAREALIKE -169.47 2.63 x 10-2 7.49 x 10-2 0.00 343.10 1.95 x 10-34
-7 -7
BAYAREALIKE + j -93.41 1.00 x 10 1.00 x 10 0.19 193.14 7.16 x 10-2
1909
1910 Table S6. Experiment 2, “Hypothetical African taxon” as sister taxon of CPAP 3054. Models and parameters
1911 for each of the analyses with BioGeoBears. Values of log-Likelihood (lnL), Dispersal (d), Extinction (e), Founder
1912 effect (j), Corrected Akaike Information Criterion (AICc), and AICc Weight (AICc wt) scores from each model
1913 implemented. The best model (DIVALIKE + j) was selected based on the lowest AICc value (and higher AICc +
1914 wt).
1915
1916
1917
Specimen (CPAP) Measured structure Measurement (mm)
Premaxilla (CPAP 5337) Maximum anteroposterior length of the 98.2

precircummnarial area
Maximum width 86.1
Maxilla (CPAP 5338) Anteroposterior length 136.2*
Maximum dorsoventral height 54.8
Postorbital (CPAP 5341) Dorsoventral height 86.1
Anteroposterior width of the base of the jugal process 56.3
Dentary (CPAP 5342) Anteroposterior length of the alveolar surface 148.8*
Anteroposterior width of the alveolar sulcus ~3.5
Anteroposterior width of the coronoid process 36.6
Dentary (CPAP 5368) Anteroposterior width of the alveolar sulcus ~5.4
Dentary (CPAP 5370) Length of the proximal edentulous margin 34.5
Anteroposterior length of the alveolar surface 149.3
Length of the symphyseal process 68.2
Anteroposterior width of the alveolar sulcus ~4.6
Anteroposterior width of the coronoid process 33.1

Quadrate (CPAP 5343) Dorsoventral height 168.9
Mediolateral width of the distal end 50.5
cervical centrum (CPAP Anteroposterior length 43.0

5380)
Mediolateral width 58.0
cervical vertebra (CPAP Anteroposterior length of the centrum 53.4

5344)
Mediolateral width of the anterior face 53.8
Mediolateral width of the posterior face 57.0
Anteroposterior length of the postzygapophyses (r/l) 31.7*/35.5*
Dorsal centrum (CPAP 5345) Anteroposterior length of the centrum 53.9
Dorsoventral height of the centrum 39.5
Dorsal vertebra (CPAP 5346) Anteroposterior length of the centrum 50.0
Proximal caudal centrum Anteroposterior length of the centrum 26.0

(CPAP 5384)
Proximal caudal vertebra Anteroposterior length of the centrum 37.4

(CPAP 5385)
Mid caudal vertebra (CPAP Anteroposterior length of the centrum 43.3

5347)

5348)

5349)

5350)

Sternum (CPAP 5352) Preserved length 137.3
Scapula (CPAP 5371) Total length 357.1
Maximum dorsoventral height of the proximal end 104.3
Dorsoventral height of the scapular blade in the 70.7

midpoint
Dorsoventral height of the distal end 89.7
Humerus (CPAP 5353) Total length 306.8
Length of the deltopectoral crest 134.4
Dorsoventral width of the distal end 88.8
Humerus (CPAP 5354) Total length 259.4*
Length of the deltopectoral crest 129.1
Humerus (CPAP 5369) Dorsoventral width of the distal end 78.7
Ulna (CPAP 5355) Length 152.3
Ulna (CPAP 5373) Length 194.2*
Radius (CPAP 5379) Mediolateral width of the proximal end 47.0
Ilium (CPAP 3054) Length of the preacetabular process 134.2*
Height of the base of the preacetabular process 5.7
Anteroposterior length of the iliac blade 150.3
Height of the iliac blade between the dorsal surface 96.4

and the pubic process
Anteroposterior length of the supraacetabular process 124.1
Incomplete ilium (CPAP Anteroposterior length of the supraacetabular process 96.3

5356)
anteroposterior length of the postacetabular process 97.8
Ischium (CPAP 5357) Height of the iliac peduncle 67.1
Height of the pubic peduncle 45.1
Femur (CPAP 5358) Total length 485.1
Femur (CPAP 5359) Total length 475.2

Femur (CPAP 5360) Mediolateral width of the distal end 84.8
Femur (CPAP 5361) Mediolateral width of the distal end 74.3
Tibia (CPAP 5362) Total length 437
Length of the cnemial crest 164.9
Tibia (CPAP 5372) Total length 447.1*
Fibula (CPAP 5363) Anteroposterior length of the proximal end 53.4*
Metatarsal III (CPAP 5364) Length 169.2
Mediolateral width of the shaft 41.4

1918
1919 Table 7. Measurements of bones of CPAP 3054. *Measurement based on preserved element.

Relict_duck-billed_dinosaurs

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Relict_duck-billed_dinosaurs

Uploaded by

Copyright:

Available Formats

bioRxiv preprint doi: https://doi.org/10.1101/2023.03.04.531097; this version posted March 6, 2023.

The copyright holder for this preprint

Claosaurus agilis 88.05 85.8

Sahaliyania elunchunorum 69.05 66.0

784 8) Coria, R. A, González Riga, B., Casadío, S. Un nuevo hadrosáurido (Dinosauria,

1192 Chilean Patagonia). Sedimentary Geology 426, 106026.

1422 Figures and Tables

1533 Supplementary Materials for

1558 Supplementary Text

Region Element representation

position hadrosauroid Austrokritosauria

N° Steps 1312 1357 1327

Templeton Tests N/A significant significant

P value N/A <0.01 <0.05

Model lnL d e j AICc AICc + wt

Premaxilla (CPAP 5337) Maximum anteroposterior length of the 98.2

Maximum width 86.1

Maxilla (CPAP 5338) Anteroposterior length 136.2*

Maximum dorsoventral height 54.8

Maxilla (CPAP 5339) Anteroposterior length 99.1*

Maxilla (CPAP 5340) Anteroposterior length 162.1*

Postorbital (CPAP 5341) Dorsoventral height 86.1

Anteroposterior width of the base of the jugal process 56.3

Dentary (CPAP 5342) Anteroposterior length of the alveolar surface 148.8*

Anteroposterior width of the alveolar sulcus ~3.5

Anteroposterior width of the coronoid process 36.6

Dentary (CPAP 5368) Anteroposterior width of the alveolar sulcus ~5.4

Dentary (CPAP 5370) Length of the proximal edentulous margin 34.5

Anteroposterior length of the alveolar surface 149.3

Length of the symphyseal process 68.2

Anteroposterior width of the alveolar sulcus ~4.6

Anteroposterior width of the coronoid process 33.1

Quadrate (CPAP 5343) Dorsoventral height 168.9

Mediolateral width of the distal end 50.5

cervical centrum (CPAP Anteroposterior length 43.0

Mediolateral width 58.0

cervical vertebra (CPAP Anteroposterior length of the centrum 53.4

Mediolateral width of the anterior face 53.8

Mediolateral width of the posterior face 57.0

Anteroposterior length of the postzygapophyses (r/l) 31.7*/35.5*

Dorsal centrum (CPAP 5345) Anteroposterior length of the centrum 53.9

Dorsoventral height of the centrum 39.5

Dorsal vertebra (CPAP 5346) Anteroposterior length of the centrum 50.0

Dorsoventral height of the centrum 54.1

Proximal caudal centrum Anteroposterior length of the centrum 26.0

Dorsoventral height of the centrum 58.0

Proximal caudal vertebra Anteroposterior length of the centrum 37.4

Dorsoventral height of the centrum 56.3

Mid caudal vertebra (CPAP Anteroposterior length of the centrum 43.3

Dorsoventral height of the centrum 40.5

Mid caudal vertebra (CPAP Anteroposterior length of the centrum 41.0

Dorsoventral height of the centrum 43.7

Mid caudal vertebra (CPAP Anteroposterior length of the centrum 39.3

Dorsoventral height of the centrum 37.6

Mid caudal vertebra (CPAP Anteroposterior length of the centrum 42.7

Dorsoventral height of the centrum 44.7

Sternum (CPAP 5352) Preserved length 137.3

Scapula (CPAP 5371) Total length 357.1

Maximum dorsoventral height of the proximal end 104.3

Dorsoventral height of the scapular blade in the 70.7

Dorsoventral height of the distal end 89.7

Humerus (CPAP 5353) Total length 306.8

Length of the deltopectoral crest 134.4

Dorsoventral width of the distal end 88.8

Anteroposterior length of the postzygapophyses (r/l) 31.7/35.5