Titanosaurs are the only sauropod dinosaurs known to bear a dermal armor. Their osteoderms are relatively rare finds, with few more than a hundred specimens recovered worldwide. Also, little is known about their intra-individual, intra-specific or inter-specific variability. The macrovertebrate site of Lo Hueco (Upper Cretaceous; Cuenca, Spain) has yielded several complete specimens of osteoderms, some associated with fairly articulated specimens. They are all variations of the morphotype known as bulb and root. The presence of only this morphotype in Europe, which is considered as the primitive condition among titanosaurs, seems to indicate that the known Upper Cretaceous Laurasian titanosaurs only bore these referred bulb and root osteoderms. An eliptic Fourier analysis on the outline of complete specimens from this morphotype reveals: i) that they truly are part of a morphological cline; and ii) the existence of a consistent correlation between the outline and the morphology of the bulb. Such variation along a cline is more consistent with intra-individual rather than inter-specific variation. The osteoderms associated with a single titanosaur individual from Lo Hueco reinforce this hypothesis.
Citation: Vidal D, Ortega F, Sanz JL (2014) Titanosaur Osteoderms from the Upper Cretaceous of Lo Hueco (Spain) and Their Implications on the Armor of Laurasian Titanosaurs. PLoS ONE 9(8): e102488. https://doi.org/10.1371/journal.pone.0102488
Editor: Peter Dodson, University of Pennsylvania, United States of America
Received: March 21, 2014; Accepted: June 18, 2014; Published: August 13, 2014
Copyright: © 2014 Vidal et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: The authors confirm that all data underlying the findings are fully available without restriction. Relevant data are included within the Supporting Information files.
Funding: Daniel Vidal is recipient of a Postgraduate Grant “Beca de posgrado” from the Universidad Autónoma de Madrid. Research funding was provided by project CGL2012-35199 of the Spanish Ministerio de Economía y Competitividad. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
The first evidence of an armored titanosaur was an osteoderm from the Upper Cretaceous found in Madagascar . Piveteau reinterpreted the evidence as an ornithischian osteoderm , and the idea of armored titanosaurs was abandoned for many years. However, Bonaparte & Powell reported the first unambiguous evidence of an armored titanosaur  and, since then, titanosaur osteoderms have been found in South America, Africa, Madagascar, India, Pakistan and Europe , –. Despite all the evidence, most sites yield at most a handful of specimens and not in anatomical connection to the other skeletal remains , even those associated with fairly complete skeletons .
The Upper Cretaceous site of Lo Hueco (Cuenca, Spain) has yielded several specimens of titanosaurs, in various degrees of articulation and completeness, being the most common tetrapods in the assemblage . Preliminary comparisons indicate the presence of, at least, two titanosaur species in Lo Hueco based upon two distinct types of cranial morphologies, appendicular bones and teeth. Up to now, there is no clear relationship between the remains, and therefore it is not possible to correlate the distinct types of teeth with the postcranial sets of bones and the isolated braincases at the moment. Some cranial elements present affinities with the French titanosaur Ampelosaurus  and others with the Indian Jainosaurus . The recognized tooth morphotypes can be respectively correlated with the robust spatulated teeth of the titanosaur from Massecaps  and the teeth of Atsinganosaurus .
The sets of axial and appendicular skeletons available fall into two morphotypes containing autapomorphies respect to the three titanosaurs currently known in the Ibero-Armorican domain (Ampelosaurus, Atsinganosaurus and, particularly the Iberian Lirainosaurus). Although the analysis is still in progress, it is considered preliminarily that the remains of the two titanosaurs from Lo Hueco correspond to new taxa to be described. Moreover, it is not possible to establish the minimum number of individuals attributable to each taxa. An approximation based on the more than seventy femora collected states that the minimum number of titanosaur individuals represented in the association is above forty. However, the number of partially articulated individuals is twenty. Among them, one titanosaur specimen was found in close association with two osteoderms, although up to eighteen isolated osteoderms attributable to titanosaur sauropods have also been recovered in the site.
In this paper we describe the titanosaur osteoderms from Lo Hueco, comparing them to the worlwide known osteoderm record attributed to titanosaurs and discuss the previously published classification systems. We also evaluate their possible body arrangement based on locality data, an eliptic Fourier analysis on the outline of the osteoderms and a cluster analysis.
The eighteen osteoderms retrieved are housed at Museo de las Ciencias de Castilla-La Mancha, Cuenca (Spain). Their specimen numbers are listed in Table 1. All the osteoderms from Lo Hueco are identified as titanosaurian on the basis of the absence of an attachment area for tendons  and the presence of internal keel/s, a feature that is postulated as a synapomorphic trait of armored titanosaurs.
Terminology and classification
The term osteoderm is used for dermal bones bigger than 1 cm and ossicles for dermal bones smaller than 1 cm, as in Arbour et al. . We use the anatomical terms of internal and external according to D'Emic et al. , for referring to the surfaces of the osteoderm towards and away from the body, respectively. The term lateral refers to the sideviews. The terms bulb and root refer to two distinct regions of the osteoderm in external and lateral views, well delimited by a cingulum (Figs. 1–4).
A - HUE-00561 in external (top), lateral (middle) and internal (bottom) views. B - HUE-00590 in external (top), lateral (middle) and internal (bottom) views. C - HUE-02029 in external (top), lateral (middle) and internal (bottom) views. b = bulb; c = cingulum; f = foramina; k = keel; r = root. Scale = 10 cm.
A - HUE-04505 in external (top), lateral (middle) and internal (bottom) views. B - HUE-00913 in external (top), lateral (middle) and internal (bottom) views. C - HUE-02900 in external (top), lateral (middle) and internal (bottom) views. Scale = 10 cm.
A - HUE-02326 in external (top), lateral (middle) and internal (bottom) views. B - HUE-02000 in external (top), lateral (middle) and internal (bottom) views. Scale = 10 cm.
A - HUE-01330 in external (top), lateral (middle) and internal (bottom) views. B - HUE-00950 in external (top), lateral (middle) and internal (bottom) views. Scale = 10 cm.
As for the classification, we use that of D'Emic et al. , with slight variations (see discussion). All the institutional abbreviations from previously published specimens of all morphotypes discussed here are listed on Table S2.
Every osteoderm retrieved from Lo Hueco belongs to the bulb and root morphotype : they have an amygdaloid outline, a convex external side and a concavo-convex internal side, with two distinct, well differentiated regions in lateral and external views: the bulb and the root (Figs. 1–4). In external or internal views, most specimens have a straight side whereas the other is more curved. The bulb is displaced toward the straightest side, giving the specimens a peculiar asymmetry. Some specimens are almost symmetrical (i.e. HUE-00561, Fig. 1 A), with both sides straight and without the bulb displaced.
All the available fragmentary specimens can be identified as parts of bulbs, roots or both. Their measurements are summarized in Table 1. Their internal anatomy can be preliminarily explored in broken specimens. These specimens reveal that some osteoderms had large cavities inside while some were compact with or without longitudinal canals (Fig. 5 A–B). However, a more detailed analysis using radiographies, CAT scans or histological slices remain to be performed.
A - Top: HUE-08738 in lateral view. Bottom: Section of HUE-08738, revealing a large cavity filled with gypsum. B - Section of UPUAM-13952, revealing longitudinal channels. C - The largest foramen found in the internal surface of bulb and root osteoderms, right under the bulb (filled with sediment). bo = bone; f = foramen; g = gypsum deposit; lc = longitudinal canals. Scale = 10 cm (A and B), 1 cm (C).
In internal view, they range from flat to slightly concave and display a woven texture: a pattern of intersecting, criss-crossed bone fibers, common in archosaur osteoderms . All of them present up to three internal keels, which are slightly displaced toward the straightest side of the osteoderm in the asymmetric specimens. They present nutrient foramina of different sizes as well, with the largest, up to 2 cm in diameter, always located under the bulb (Fig. 5 C).
In external view, the bulb is a sub-circular region, well delimited by a cingulum. It has a regular pattern of radiating fibers or pits. The root limits with the bulb, reaching its maximum width and tapering away from the bulb. Its length is variable, ranging from twice to up to five times the length of the bulb. Its texture is more irregular, with nodules, fibers and grooves.
In lateral view, the bulb presents different morphologies : spines (convex or conical), flat or concave. The root reaches its maximum thickness at the cingulum, tapering away from the bulb. For the present study we have decided upon using the terms convex, flat and concave to describe the surface of the bulb in lateral view. The specimens from Lo Hueco show all of the aforementioned bulb morphologies:
Externally, their outline is amygdaloid and quite rounded, symmetrical in HUE-00561 (Fig. 1A), slightly asymmetric in HUE-02029 (Fig. 1C) and quite asymmetric in HUE-00590 (Fig. 1D). HUE-00913, 02900 and 04505 (Fig. 2) are broken, but the former two appear to have been asymmetric as one side is quite straight while the other is curved, whereas HUE-04505 is very symmetric. In lateral view, HUE-00561, 00590, 00913 and 04505 are flattened whereas HUE-02029 and 02900 are quite thicker. All the specimens present a pattern of radiating foramina but not fibers in the bulb. In the root, HUE-00561, 00590, 02029 and 02900 present a pattern of intersecting fibers and channels but not nodules while HUE-00913 and 04505 present fibers, nodules and grooves. HUE-00561 and 00590 present three shallow keels in internal view, displaced to the straightest side on HUE-00590. HUE-00913 has a shallow bilobate keel. HUE-02029 has a thick, bilobate keel, being its apex the most internal point of the osteoderm. The internal surface of HUE-00913 and 02900 cannot be observed due they are broken.
HUE-02029 is broken in the root region, revealing its internal structure, which is quite compact, without any duct or large openings. The bone is spongy with the microcavities filled with gypsum.
Externally their outline is amygdaloid and ranges from slightly to quite elongated, all of them asymmetric, having a straight and a curved side. The bulb presents a radial ornamentation of fibers in HUE-02326 (Fig. 3. A) and 08738, and foramina in HUE-02000 (Fig. 3 B). In lateral view, the height of the bulb is greater than that of the root in HUE-02326 and 08738, and about the same height in HUE-02000. Internally, HUE-02000 and 02326 present 3 keels. HUE-08738 is broken and has not preserved most of the root or the internal side. The fracture reveals a large gypsum deposit, embedded in the large openings in the spongy bone of the osteoderm (Fig. 5 A). Such a cavity seems similar to that reported by Curry-Rogers et al. in an osteoderm with a flat bulb . HUE-08393 was collected one year after the main dig and thus was affected by meteorization. Despite its surface being too eroded, the general morphology of the bulb can be determined as convex.
The most remarkable osteoderms retrieved at Lo Hueco are the very elongated HUE-00950 and HUE-01330 (Fig. 4). The surface of the bulb is concave, totally in HUE-01330 and only in the center of the bulb in HUE-00950, with a radial ornamentation of foramina. The cingulum is well developed in both osteoderms and the root is about five times longer than the bulb. Internally, HUE-01330 presents a thick, wide keel whereas HUE-00950 presents a thick, bilobate keel. Both are broken, with their sections revealing compact bone without any large openings or ducts.
Morphological analysis methods
Morphological analysis of the specimens was carried out using an elliptic Fourier analysis (EFA). Given the high disparity observed in the outlines of the specimens (from rounded to elongated osteoderms), the aim of the analysis was to find out whether bulb and root osteoderms could be splitted in distinct morphotypes or not. In order to limit the amount of speculation in reconstructing the outlines, only the eighteen published bulb and root osteoderms that were complete enough were included in the analysis (Table 2).
To perform the analysis, the data from the outline of every osteoderm was converted to coordinates in a Cartesian System, a hundred per each specimen, using tpsDIG2 freeware (Coordinates available on Table S1).
Following the recommendations of Hammer & Haper , every osteoderm was given the same orientation, with the widest portion (the bulb) always on the left side and the straightest side parallel to a straight line. In order to avoid the influence of the bulb morphology, images in internal view were used whenever possible.
The coordinates were analyzed in PAST , obtaining the first 30 harmonics and the variable “size” of each osteoderm. A principal components analysis (PCA) was performed in order to further reduce the number of variables from the EFA and study the variation in a XY graph.
HUE-00913 (Fig. 2B) and HUE-00950 (Fig. 4B) are both associated to the same titanosaur specimen (HUE-EC-11), and might give an insight on how much of the morphological variation of the osteoderms was present in the same individual. However, HUE-00913 is not complete and cannot be included in the EFA. In order to study those osteoderms which are not complete and contrast the EFA as well, a cluster analysis from qualitative characters (Table 3) from the osteoderms of Lo Hueco was carried out in PAST, with the paired group algorithm and the Jaccard index for similarity measure.
The EFA output shows that, upon their silhouette in internal view, bulb and root osteoderms are evenly distributed in a morphological cline (Fig. 6). The first principal component (PC1) explains more than 99% of the variance, with the rest of components explaining less than 1%. Considering that the variable “size” contributes the most to PC1 and that all the osteoderms were scaled to the same length we interpreted it as the relative width of the osteoderms, and thus as a general measure of their shape: those osteoderms with negative PC1 scores are more elongated and those with more positive scores are more rounded. The third principal component (PC3) was found to correlate well with the symmetry of the specimens in the scatter plot, and was thus interpreted as a symmetry parameter: osteoderms with positive scores are more symmetric than those with more negative scores.
A - Scatter plot with the first and third principal components obtained from the principal components analysis, using the data of an elliptic Fourier analysis on the outline of the osteoderms, with the morphology of the bulb highlighted: triangle = concave bulb; square = convex bulb; circle = flat bulb. B - Bulb and root osteoderm specimens located in the scatter plot. From left to right, HUE-01330, HUE-02000, HUE-02326, MDE-C3-192 and HUE-00590.
Among this morphological cline, three different groups of osteoderms can be identified if the morphology of the bulb is highlighted in the scatter plot (Fig. 6A). Osteoderms with a flat bulb are more rounded, with lengths not much greater than their widths; osteoderms with a convex bulb have greater lengths than widths and those with a concave bulb have enormous roots, up to five times the length of the bulb. There are “transitional” osteoderms between these groups, with slightly convex bulbs yet quite elongated roots (HUE-02000, Fig. 6A).
The osteoderms from Lo Hueco are scattered throughout the graph, thus being representative of the worldwide fossil record of titanosaur bulb and root osteoderms (Fig. 7).
A - Scatter plot with the first and third principal components obtained from the principal components analysis, using the data of an elliptic Fourier analysis on the outline of the osteoderms, with the specimens from Lo Hueco highlighted (as red squares). B - Dendrogram obtained from the cluster analysis on those specimens of Lo Hueco complete enough to have >50% of the characters (Table 3). Highlighted in bold are the osteoderms associated with the titanosaur individual HUE-EC-11.
The cluster analysis separates the osteoderms into 3 main clusters: one includes elongated osteoderms with either concave or convex bulbs (Fig. 7.1), another includes thick osteoderms with flat bulbs (Fig. 7.2), and the third cluster includes flattened osteoderms with flat bulbs (Fig. 7.3). The last two clusters are part of a more inclusive but not so well supported cluster, which would include all the osteoderms with a flat bulb. If compared to the EFA, the distribution and clustering is quite similar (Fig. 7), and thus HUE-00913 and HUE-00950 appear to have displayed different morphologies, suggesting that bulb and root osteoderms of different outlines and bulb morphologies were associated with the same individual.
Classification and identification
The most comprehensive and complete classification of osteoderms attributable to titanosaurs was proposed by D'Emic et al. . This classification established four different morphotypes based on the published fossil record of titanosaur osteoderms: ellipsoid, keeled, cylindrical and mosaic. Curry-Rogers et al. proposed that many osteoderms identified as part of the mosaic morphotype might have been fragments of larger osteoderms . As broken specimens are always susceptible of reinterpretation (Fig. 8), in this paper we suggest a variation in D'Emic et al. system, based upon complete specimens only. Moreover, as noted by Arbour et al. , there should be a distinction between osteoderms (dermal bones greater than 1 cm) and ossicles (dermal bones smaller than 1 cm), as they have been proven to be different both in morphology  and histology . Among the osteoderms, we propose the following classification:
A - UPUAM-13952 in lateral, interpreted as part of a root (top), external (bottom, left) and internal (bottom, right) views. B - UPUAM-13951 in lateral, interpreted as a bulb and a cingulum (top), external (bottom, left) and internal (bottom, right) views. Scale = 10 cm.
Morphotype 1 - bulb and root.
Osteoderms with amygdaloid outline and the general appearance of a barrel vault, with two well differentiated regions in lateral and external views: the bulb and the root, delimited by a cingulum. The bulb has a surface texture made of a radial pattern of fibers or foramina while the root presents a more irregular pattern of ducts and nodules. Their base (internal side) is slightly concave, with up to three ridges or keels, which range from slight to quite thick.
D'Emic et al. stated that some ellipsoid osteoderms are made up only of the “bulb” or the “root” regions . However, some of these specimens seem to be broken, and could be interpreted as parts of a complete bulb and root osteoderm. This would be the case of the Armuña (Segovia, Spain) specimens UPUAM-13951 and UPUAM-13952 (Fig. 8, see , referred respectively as AR-01 and AR-02 in ), FSL 92827, UA 8675 or MN-5013V (see ).
This morphotype would include most of the ellipsoid osteoderms sensu D'Emic et al., MPMA 08-0058-11  and MSM-84-7 and MSM-85-4 . FSL 92827 and UA 8675, both cylindrical osteoderms sensu D'Emic et al., are here interpreted as a bulb and a root, respectively. Bulb and root osteoderms are found in Europe, India, South America, Africa and Madagascar (Fig. 9).
Morphotype 2 - scutes (or keeled sensu D'Emic et al.).
These osteoderms present one region of sub-circular contour with a radial pattern of foramina or fibers, sometimes surrounded by a cingulum or a perlated contour. In lateral view, they might present an external keel, a convex or a flat morphology. They could be interpreted as bulbs that somehow lost the root portion. However, there is not enough evidence to support such homology.
We interpret this group of osteoderms as in D'Emic et al. However, the presence of internal keels is not unique to this morphotype, as it is present in bulb and root osteoderms from Spain, France and Romania as well. Therefore, as the term “keeled” can be misleading, we propose the term “scute” for this morphotype. Aside from all the osteoderms in D'Emic et al. , the more recent findings UFMA 1.10.1958  and the putative QM F6737  would also be scutes. They have been found in Africa, Madagascar, South America and Australia (Fig. 9).
Some osteoderms do not seem to fit in the morphotypes proposed either here or in D'Emic et al. . According to photographs and descriptions, they are elongated, compressed, with an external apex and a convex internal face. However, they do not show any distinct features from a scute, and could either be part of their variability or be a distinct morphotype on their own. More detailed study on the specimens is required to clarify this issue. These osteoderms are those associated with Mendozasaurus  and Aeolosaurus , as well as the “placa dérmica pequeña” described by Powell . They have only been found in Argentina (Fig. 9).
The nature of the Laurasian titanosaur osteoderms
Considering the wide disparity displayed by the fossil record of titanosaur osteoderms, the question whether such disparity represents inter-specific, intra-specific or intra-individual variability remains unclear. The identification of the different morphotypes within the taxa in the different localities around the world is crucial to understand their phylogenetic and biogeographical distribution.
The collection of titanosaur osteoderms so far available from Europe consists of at least a complete osteoderm from Romania ; three complete osteoderms from Massecaps at Cruzy, France; a complete osteoderm from Vitrolles, France; four complete osteoderms from Aude, France ; two fragmentary osteoderms from Laño, Spain , ; five fragmentary osteoderms from Armuña, Spain  and seven complete and eleven fragmentary osteoderms from Lo Hueco, Spain (; this paper).
The complete known specimens from the Laurasian titanosaur record are all bulb and root osteoderms, and the fragments can all be interpreted as portions of complete bulb and roots (Fig. 8). Also, there are no osteoderms from Laurasia interpreted as a scute or dermal ossicle. This could mean that the european titanosaurs only bore bore bulb and root osteoderms, if the fossil record is taken at face value.
The simultaneous presence of both bulb and roots (FMNH PR 2342) and scutes (FMNH PR 2021) occurs only in the Maevarano Formation, Madagascar, where at least two different titanosaurs were found: Rapetosaurus krausei, a non saltasaurine lithostrotian ,  and the Malagasy Taxon B, a saltasaurine . Thus far, only saltasaurine titanosaurs have been associated with scutes, so it would be likely that both titanosaurs from the Upper Cretaceous of Madagascar were armored, and bore different types of armor, thus implying that the type of osteoderms might have taxonomic and systematic value.
The bulb and root condition appears to be primitive in Titanosauria, as the morphotype is more widespread in time and space (Fig. 9). Also, basalmost forms of Lithostrotia such as Malawisaurus are associated with bulb and roots , whereas derived forms (i.e. saltasaurines) are associated with scutes and dermal ossicles, thus possibly representing a derived condition.
Arrangement of osteoderms on the titanosaur body
The anatomical arrangement of titanosaur osteoderms has been debated since Bonaparte & Powell first proposed a tessellated arrangement of dermal ossicles surrounding the larger plates (scutes sensu this paper) . This hypothesis was reinforced by embryonic skin impressions of titanosaurs found at Auca Mahuevo –, where rows and rossette arrangements of scales were found. However, it is not possible to assess the exact position of those patterns in adult titanosaurs, due to the difficulty to assign the skin impressions to a precise body region. Moreover, in extant archosaurs, osteoderms are post-natal ossifications  and, therefore, these impressions may or may not correspond to the adult skin pattern.
The general consensus is that titanosaurs were lightly armored forms , , , with several proposed arrangements, such as osteoderms restricted to the sacro-pelvic region  or osteoderms with flat bulbs in the dorsal and sacral regions and osteoderms with convex bulbs restricted to the scapular region . Salgado suggested a sagittal row for the most symmetric osteoderms and parasagittal rows for the asymmetrical specimens  in the light of two evidences: i) the more asymmetric the osteoderms of Alligator mississippiensis, the further they are located from the median plane  and ii) the existence of median epidermal structures in an undescribed diplodocid sauropod .
We also consider the asymmetrical specimens of bulb and root osteoderms to be part of parasagittal rows because of the orientation of the big nutrient foramina located in the internal side. These nutrient foramina enter obliquely the minor axis of osteoderm (Fig. 5 C), which might evidence a vascular irrigation similar to that of extant archosaurs, with a dorsal median artery that bifurcates to the left and to the right to perfuse a whole transversal row of osteoderms, median to lateral direction . The more symmetric osteoderms might belong to a median arrangement that could represent a transition to a median position.
The bulb and root osteoderms display extreme morphologies, yet continuum exists between those extremes (Figs. 6–7). Such a morphological cline is likely to represent the variability of one individual, as evidenced by two bulb and root osteoderms (HUE-00913 and HUE-00950) found associated with the same titanosaur individual at Lo Hueco (Fig. 10). These osteoderms were found a few meters away from each other and near the dorsal vertebrae of a single, isolated and partially articulated titanosaur specimen (HUE-EC-11), with the same orientation as the rest of long bones and with few elements of other taxa in the same bed. We consider, given the evidence, that HUE-00913 and HUE-00950 belonged to the same individual.
The main skeletal elements are highlighted and identified, as well as the osteoderms. Scale = 1 m.
We interpret the morphological disparity from the EFA analysis as cline of intra-individual variability because i) one of the osteoderms associated with HUE-EC-11 is twice the size of its counterpart (Fig. 10); ii) both are located at different extremes of the morphological cline (Fig. 7); and iii) both belong to the same individual. This intra-individual variability would be similar to that displayed by the Stegosauria, where the size and shape of the dermal plates varies greatly among a single individual. How that intra-individual variation might have been arranged in titanosaurs, however, cannot be assessed from current evidence. Nevertheless, considering that locality data supports that titanosaurs would be lightly armored , a craniocaudal cline in two paramedian rows, as seen in stegosaurs, would be the most conservative conjecture. Alternative arrangements of more than two paramedian rows, each with its own craniocaudal cline could also be possible, but would require more osteoderms per individual. Nevertheless, osteoderms with such elongated roots have only been found at Lo Hueco thus far. It is possible that they may be a shared feature of Iberian forms, and thus other titanosaurs might have had slightly different arrangements.
Cartesian coordinates obtained for the osteoderm outlines employed in the elliptic Fourier analysis.
We thank Fátima Marcos and the “Taller de Empleo de Restauración Paleontológica” of Lo Hueco for the skillful preparation of the specimens. Thanks to Ioannis Sarris and Jesús Marugán-Lobón for their comments and suggestions on the elliptic Fourier analysis. We thank Michael D. D'Emic and Jeffrey A. Wilson for their replies and clarifications on their paper. Thanks to Pedro Mocho, Alejandro Serrano-Martinez, Elena Cuesta, Carlos De Miguel and Ignacio Arribas for their insightful and helpful comments. We also want to thank the reviewers Kristina A. Curry-Rogers and Michael D. D'Emic for their suggestions and comments, which have greatly improved the manuscript.
Conceived and designed the experiments: DV FO JLS. Performed the experiments: DV. Analyzed the data: DV. Contributed reagents/materials/analysis tools: FO JLS. Contributed to the writing of the manuscript: DV FO JLS.
- 1. Curry Rogers KA (2005) Titanosauria. In: Curry Rogers KA, Wilson JA, editors. The Sauropods: Evolution and Paleobiology. Berkeley: University of California Press. pp. 50–103.
- 2. D'Emic MD, Wilson JA, Chatterjee S (2009) The titanosaur (Dinosauria: Sauropoda) osteoderm record: review and first definitive specimen from India. Journal of Vertebrate Paleontology 29(1): 165–177.
- 3. Depéret C (1896) Note sur les dinosauriens sauropodes et théropodes du Crétacé supérieur de Madagascar. Bulletin de la Société Géologique de France 21: 176–194.
- 4. Piveteau J (1926) Contribution à l'étude des formations lagunaires du Nord-Ouest de Madagascar. Bulletin de la Société Géologique de France 4: 33–38.
- 5. Bonaparte J, Powell JE (1980) A continental assemblage of tetrapods from the Upper Cretaceous beds of El Brete, northwestern Argentina (Sauropoda-Coelurosauria-Carnosauria-Aves). Mémoires de la Société Géologique de France, nouvelle séries 139: 19–28.
- 6. Malkani MS (2010) Osteoderms of Pakisauridae and Balochisauridae (Titanosauria, Sauropoda, Dinosauria) in Pakistan. Journal of Earth Sciences 21(1): 198–203.
- 7. Molnar RE (2011) New morphological information about Cretaceous sauropod dinosaurs from the Eromanga Basin, Queensland, Australia. Alcheringa 35(2): 329–339.
- 8. Curry Rogers K, D'Emic MD, Rogers R, Vickaryous M, Cagan A (2011) Sauropod dinosaur osteodermsfrom the Late Cretaceous of Madagascar. Nature communications 2: 564.
- 9. Marinho TS, Iori FV (2011) A large titanosaur (Dinosauria, Sauropoda) osteoderm with possible bite marks from Ibirá, São Paulo state, Brazil. In: Carvalho IS, Srivastava NK, Strohschoen O, Lana Jr CC editors. Paleontologia: Cenários de Vida. Rio de Janeiro: Editora Interciência. pp. 367–377.
- 10. Lindoso RM, Marinho TS, Santucci RM, Medeiros MA, Carvalho IDS (2013) A titanosaur (Dinosauria: Sauropoda) osteoderm from the Alcântara Formation (Cenomanian), São Luís Basin, Northeastern Brazil. Cretaceous Research 45: 43–48.
- 11. Salgado L, Apesteguía S, Heredia S (2005) A new specimen of Neuquensaurus australis, a Late Cretaceous saltasaurine titanosaur from north Patagonia. Journal of Vertebrate Paleontology 25: 623–634.
- 12. Barroso-Barcenilla F, Cambra-Moo O, Escaso F, Ortega F, Pascual A, et al. (2009) New and exceptional discovery in the Upper Cretaceous of the Iberian Peninsula: the palaeontological site of “Lo Hueco”, Cuenca, Spain. Cretaceous Research 30(5): 1268–1278.
- 13. Knoll F, Ridgely RC, Ortega F, Sanz JL, Witmer LM (2013a) Neurocranial Osteology and Neuroanatomy of a Late Cretaceous Titanosaurian Sauropod from Spain (Ampelosaurus sp.). PLoS ONE 8(1): e54991
- 14. Knoll F, Witmer LM, Ridgely RC, Ortega F, Sanz JL (2013b) A new titanosaur sauropod neurocranium from the Late Cretaceous of Spain. Program and abstracts of the 73rd Annual Meeting of the Society of Vertebrate Paleontology. Los Angeles, USA. Supplement to the online Journal of Vertebrate Paleontology. pp. 122.
- 15. Díez Díaz V, Tortosa T, Le Loeuff J (2013) Sauropod diversity in the Late Cretaceous of southwestern Europe: The lessons of odontology. Annales de Paléontologie 99(2): 119–129.
- 16. Díez Díaz V, Ortega F, Sanz JL (in press) Titanosaurian teeth from the Upper Cretaceous of “Lo Hueco” (Cuenca, Spain). Cretaceous Research
- 17. Arbour VM, Burns ME, Bell PR, Currie PJ (2013) Epidermal and Dermal Integumentary Structures of Ankylosaurian Dinosaurs. Journal of Morphology 275(1): 39–50.
- 18. Le Loeuff J, Buffetaut E, Cavin L, Martin M, Martin V, et al. (1994) An armoured titanosaurid sauropod from the Late Cretaceous of southern France and the occurrence of osteoderms in the Titanosauridae. Gaia 10: 155–159.
- 19. Hammer Ø, Harper DAT, Ryan PD (2006) Paleontological data analysis. Wiley-Blackwell, 368 p.
- 20. Hammer Ø, Harper DAT, Ryan PD (2001) PAST: Paleontological statistics software package for education and data analysis. Palaeontologia Electronica 4(1): 1–9 Available: http://palaeo-electronica.org/2001_1/past/issue1_01.htm.
- 21. Cerda IA, Powell JE (2010) Dermal armor histology of Saltasaurus loricatus, an Upper Cretaceous sauropod dinosaur from Northwest Argentina. Acta Palaeontologica Polonica 55(3): 389–398.
- 22. González Riga BJ (2003) A new titanosaur (Dinosauria, Sauropoda) from the upper Cretaceous of Mendoza province, Argentina. Ameghiniana 40: 155–172.
- 23. Salgado L, Coria RA (1993) El genero Aeolosaurus (Sauropoda, Titanosauridae) en la Formaciónn Allen (Campaniano–Maastrichtiano) de la provincia de Río Negro, Argentina. Ameghiniana 30: 119–128.
- 24. Powell JE (2003) Revision of the South American titanosaurid dinosaurs: palaeobiological, palaeobiogeographical, and phylogenetic aspects. Records of the Queen Victoria Museum 111: 1–173.
- 25. Csiki Z (1999) New evidence of armoured titanosaurids in the Late Cretaceous –Magyarosaurus dacus from the Hateg Basin (Romania). Oryctos 2: 93–99.
- 26. Sanz JL, Powell JE, Le Loeuff J, Martínez R, Pereda Suberbiola X (1999) Sauropod remains from the Upper Cretaceous of Laño (Northcentral Spain). Titanosaur phylogenetic relationships. Estudios del Museo de Ciencias Naturales de Álava 14(Número Especial 1): 235–255.
- 27. Díez Díaz V, Pereda Suberbiola X, Sanz JL (2013) Appendicular skeleton and dermal armour of the Late Cretaceous titanosaur Lirainosaurus astibiae (Dinosauria: Sauropoda) from Spain. Palaeontologia Electronica 16(2.19A): 18 palaeo-electronica.org/content/2013/502-titanosaur-skeleton.
- 28. Sanz JL, Buscalioni AD (1987) New evidence of armoured titanosaurs in the Upper Cretaceous of Spain. In Currie PJ, Koster EH editors. Fourth Symposium on Mesozoic Terrestrial Ecosystems, Short Papers. Drumheller: Tyrrell Museum of Palaeontology. pp. 197–202.
- 29. Ortega F, Sanz JL, Díez Díaz V, Escaso F (2012) Armoured titanosaurs from the Upper Cretaceous of Lo Hueco (Cuenca, Spain). Fundamental! 20: 173–174.
- 30. Calvo JO, González-Riga B, Porfiri JD (2007) A new titanosaur sauropod from the Late Cretaceous of Neuquén, Patagonia, Argentina. Arquivos do Museu Nacional, Rio de Janeiro 65(4): 485–504.
- 31. Gomani EM (2005) Sauropod Dinosaurs from the Early Cretaceous of Malawi, Africa,. Palaeontologia Electronica 8(1): 37 Available: http://palaeo-electronica.org/paleo/2005_1/gomani27/issue1_05.htm.
- 32. Chiappe LM, Coria RA, Dingus L, Jackson F, Chinsamy A, et al. (1998) Sauropod dinosaur embryos from the Late Cretaceous of Patagonia. Nature 396: 258–261.
- 33. Coria RA, Chiappe LM (2007) Embryonic skin from Late Cretaceous sauropods (Dinosauria) of Auca Mahuevo, Patagonia, Argentina. Journal of Paleontology 81(6): 1528–1532.
- 34. Salgado L (2003) Considerations on the bony plates assigned to titanosaurs (Dinosauria, Sauropoda). Ameghiniana 40: 441–456.
- 35. Seidel MR (1979) The osteoderms of the American alligator and their functional significance. Herpetologica 35: 375–380.
- 36. Czerkas S (1992) Discovery of dermal spines reveals a new look for sauropod dinosaurs. Geology 20: 1068–1070.