Ontogeny reveals function and evolution of the hadrosaurid dinosaur dental battery
- 4.6k Downloads
Hadrosaurid dinosaurs, dominant Late Cretaceous herbivores, possessed complex dental batteries with up to 300 teeth in each jaw ramus. Despite extensive interest in the adaptive significance of the dental battery, surprisingly little is known about how the battery evolved from the ancestral dinosaurian dentition, or how it functioned in the living organism. We undertook the first comprehensive, tissue-level study of dental ontogeny in hadrosaurids using several intact maxillary and dentary batteries and compared them to sections of other archosaurs and mammals. We used these comparisons to pinpoint shifts in the ancestral reptilian pattern of tooth ontogeny that allowed hadrosaurids to form complex dental batteries.
Comparisons of hadrosaurid dental ontogeny with that of other amniotes reveals that the ability to halt normal tooth replacement and functionalize the tooth root into the occlusal surface was key to the evolution of dental batteries. The retention of older generations of teeth was driven by acceleration in the timing and rate of dental tissue formation. The hadrosaurid dental battery is a highly modified form of the typical dinosaurian gomphosis with a unique tooth-to-tooth attachment that permitted constant and perfectly timed tooth eruption along the whole battery.
We demonstrate that each battery was a highly dynamic, integrated matrix of living replacement and, remarkably, dead grinding teeth connected by a network of ligaments that permitted fine scale flexibility within the battery. The hadrosaurid dental battery, the most complex in vertebrate evolution, conforms to a surprisingly simple evolutionary model in which ancestral reptilian tissue types were redeployed in a unique manner. The hadrosaurid dental battery thus allows us to follow in great detail the development and extended life history of a particularly complex food processing system, providing novel insights into how tooth development can be altered to produce complex dentitions, the likes of which do not exist in any living vertebrate.
KeywordsPeriodontal Ligament Enamel Surface Root Resorption Tooth Wear Tooth Position
Histological thin sections of several amniote taxa (Additional file 1: Table S1) were prepared by first embedding specimens in Castolite AP or Castolite AC polyester resin and placing them under vacuum. One specimen (MOR 559) was embedded in Buehler Epothin resin. Embedded materials were then cut using the Buehler Isomet slow-speed wafer blade saw and the cut surfaces were polished using 600-grit silicon carbide powder. For two specimens (MOR 548, 559), thin wafers were cut using a Buehler Isomet 1000 high-speed wafer blade saw. Specimens were later mounted to frosted plexiglass slides using cyanoacrylate and cut using the Isomet saw. Specimens were then ground down using a Hillquist or a Buehler Ecomet grinding machine and further polished using progressively finer grits of silicon carbide and aluminium oxide powders. The ROM thin sections were imaged using a Nikon DS-Fi camera mounted to a Nikon AZ-100 microscope with NIS Elements BR imaging software registered to D. C. Evans or R. R. Reisz.
Histology of a hadrosaurid tooth
Tooth ontogeny in hadrosaurids
Tooth ontogeny in hadrosaurids can be easily reconstructed using coronal histological sections through a dental battery, which preserve teeth at successive ontogenetic stages (Figs. 3a–e and 4). These sections revealed that, unlike most vertebrates, hadrosaurids did not shed their teeth, and the prevention of normal tooth replacement was a key factor in retaining multiple generations of teeth at each locus. The earliest-forming teeth we sectioned were located next to the “special foramina” in the maxilla and dentary, which represent the sites at which the odontogenic organ, the dental lamina begins forming new teeth at each tooth position  (Additional file 4: Figure S3). These teeth were clearly at the earliest stages of dental tissue mineralization, given their proximity to these foramina and their comparable relative size and morphology to the earliest-staged teeth examined by Horner . Teeth at their earliest ontogenetic stages consisted of thin bands of enamel and primary orthodentine (hereafter simply referred to as dentine). These teeth already began forming the attachment tissues at this early stage, as indicated by the presence of typical acellular cementum [16, 17, 18, 19, 20, 21] and the first layers of cellular cementum along the surface opposite of the enamel (Fig. 3d, e). Replacement teeth at comparable stages in Alligator (Fig. 3f–i; Additional file 5: Figure S4) and theropod dinosaurs (Additional file 6: Figure S5) only consisted of an enamel cap and underlying dentine, and no cementum was present until they neared eruption. The subsequent stage of hadrosaurid tooth ontogeny occurred deep within the jaw and was characterized by extensive deposition of dentine, with the pulp cavity at the tip of the tooth becoming partially closed off (Figs. 3a, c and 4a), and blood vessels being enveloped by the rapid pulpward formation of dentine. This unusual form of dentine development was matched by precocious growth of cellular cementum with Sharpey’s fibers, indicating that the hadrosaurid tooth was already anchored by periodontal ligament (Fig. 4a, d). In other archosaurs and mammals, even nearly erupted teeth have open pulp cavities and lack a ligamentous attachment to the socket (Fig. 3f; Additional file 5: Figure S4, Additional file 6: Figure S5), the latter normally forming during eruption into the oral cavity [11, 16, 22, 23].
The lingual surfaces of unerupted hadrosaurid teeth are partially resorbed by the formation of the subsequent teeth. However, a dentine wall always separates the pulp cavity of an older tooth from the developing tooth underneath (Figs. 3a–e and 4), thus maintaining the vital pulp of the older tooth. In strong contrast, in other vertebrates this process leads to the shedding of old teeth, because the pulp cavity becomes breached by the advancing resorption front of the replacement tooth, which disrupts the vascular supply to the pulp . Prior to tooth eruption in hadrosaurids, the pulp cavity became completely enclosed by rapid deposition of dentine (Figs. 3a, b and 4c–f), first at the occlusal end of the tooth and continuing apically in a zipper-like fashion through the tooth. The advancing walls of dentine converged near the midline of each tooth, creating a line of isolated pockets of the remaining pulp cavity (Figs. 3a–c and 4a). Although it has been commonly thought that the exposed teeth in dental batteries of hadrosaurids were pushed outward from beneath by their younger successors, and that no tooth resorption occurred [2, 7, 24], we found that root resorption was extensive in hadrosaurid teeth (Figs. 1d, f, 3a and 4a, c, e, f; Additional file 4: Figure S3) and instead provided the mechanism through which the teeth became tightly interlocked. The extent of root resorption is demarcated by irregular, jagged surfaces, which are typical of the Howship’s lacunae left by osteoclasts, which form a reversal line (Figs. 1f, g and 4c, f). The presence of extensive root resorption indicates that another mechanism was responsible for continuous tooth eruption in hadrosaurid dental batteries. It is more likely that the periodontal ligament, along with root elongation, served as the agents through which teeth were able to continuously erupt into the oral cavity, similar to the condition in ever-growing teeth in mammals . At the time of eruption into the oral cavity, the teeth of hadrosaurids had nearly completely plugged pulp cavities and thus were probably no longer vital. Sharpey’s fibers were also abundant within the thickened layers of cellular cementum, indicating continued attachment of the tooth to the surrounding socket by periodontal ligament (Fig. 4d). After eruption, each tooth was worn down completely, including the root, instead of being shed (Fig. 4a, e).
Tooth attachment in the hadrosaurid dental battery
The evolution of the hadrosaurid tooth
Our ontogenetic perspective reveals that the histological complexity and the elaborate life cycle of hadrosaurid teeth did not involve the evolution of novel tissues, as previously suggested . Hadrosaurid teeth are composed of homologous tissues to those in other archosaurs, and other amniotes in general [9, 13, 16, 17, 18, 19, 23, 27, 28, 33] (Additional file 2: Figure S1). Hadrosaurid teeth are unique among reptilian dentitions in that the arrangements of homologous tissues forming each tooth have been radically altered from the ancestral dinosaurian condition (Fig. 2). Whereas the enamel of the typical reptilian tooth crown forms a cap over the apex of a tooth (thus forming the anatomical crown ), the enamel of hadrosaurid teeth is restricted to one side of the tooth . As a result, the junction between cementum, typically a root tissue , and the crown tissue enamel has shifted to the apex of the hadrosaurid tooth (Fig. 2h). This shift in the cemento-enamel junction means that each tooth begins forming enamel and cementum simultaneously in the early stages of dental development. This also means that the enamel-covered face of the tooth does not contribute to tooth attachment, which is anatomically more similar to the condition in mammalian ever-growing incisors [12, 34] than to the complex grinding teeth of modern horses [35, 36]. From evolutionary and developmental perspectives, the asymmetrical deposition of enamel in hadrosaurid teeth should not be confused with the evolution and development of coronal cementum in mammals. Hadrosaurids did not independently evolve coronal cementum (contra [5, 8]), because coronal cementum occurs as an external covering of the enamel and is the result of a different developmental process [16, 37] (Additional file 8: Figure S7). The cellular cementum in hadrosaurid teeth never overlaps the enamel as it does in many ungulate mammals. Instead, typical root cementum extends to the tooth apex in hadrosaurids and the cementum becomes incorporated into the occlusal surface through a combination of its unique orientation along the tooth, constant eruption, and wear of the whole tooth (Fig. 4).
Heterochrony and the evolution of the hadrosaurid dental battery
Our comparison of hadrosaurid dental ontogeny to that of other amniotes also provides clear evidence of an accelerated rate of dentine and cementum formation in hadrosaurids. We invoke heterochrony, namely the earlier onset and increased rate of tissue formation as the evolutionary mechanism through which hadrosaurids were able to evolve complex dental batteries (Fig. 3). The vascular nature of the dentine and cementum in hadrosaurid teeth is the hallmark of an accelerated rate of tissue formation that entombed blood vessels of the pulp and the periodontium, not the evolution of new types of tissues as previously suggested . When compared to other amniotes, hadrosaurid teeth were not only forming more rapidly , but also matured earlier. Dentine was deposited so precociously and extensively that the soft tissue within the pulp cavity was completely obliterated by the time the tooth erupted (Figs. 2, 3 and 4; Additional file 4: Figure S3). The infilling of the pulp with dentine created not only a wear surface for grinding plant material , but also served the more critical function of preventing a breach of the pulp cavity. Hadrosaurid teeth were ground down over the course of their use [2, 5, 7], and a breached pulp cavity, caused by attrition or root resorption, could normally lead to infection and severe pain, given the density of blood vessels and nerves in the dental pulp . Cementum in most amniotes is typically avascular [11, 16], but a similar condition occurs in modern horses, where the entombed vasculature of the coronal cementum recedes as the tooth erupts into the oral cavity . In addition, the rapid closure of the pulp cavity would have initially prevented tooth replacement, but once completed, the tooth would have also been non-vital, given the lack of nourishment to the living tissues of the tooth. Filling the pulp cavity with dentine thus provided a remarkable solution to the challenge of tooth wear: during their ontogeny, hadrosaurids were able to grind down an entire tooth because it was completely infilled with dentine while still retaining a connection to its neighbouring teeth, thus allowing dead teeth to remain functional as grinding surfaces. Hadrosaurid teeth were never shed as they are in other vertebrates, but were worn away completely, allowing hadrosaurids to grind away at multiple teeth at once in a single tooth position [1, 39]. This functionalization of the tooth root into the grinding surface was permitted by the connections between teeth within the dental battery, and represents a true evolutionary novelty among vertebrates.
Accelerated cementum formation indicates extensive attachment of hadrosaurid teeth to their surroundings, but in a unique fashion. Previous suggestions that cementum in hadrosaurids served to fuse the teeth together into a single grinding pavement , are incorrect. In thin section, none of the teeth within a battery contact each other (Figs. 1, 4, 5 and 6; Additional file 4: Figure S3, Additional file 7: Figure S6). Cementum was deposited comparatively early in dental ontogeny, at the same time as enamel and dentine (Fig. 3), anchoring the teeth by ligament to the alveoli. We find further support for a ligamentous attachment of hadrosaurid teeth based on taphonomy. Fossil and modern skeletonized jaws of mammals and crocodilians are typically edentulous due to post-mortem tooth loss, which is rare in taxa that have teeth fused to the jaws [16, 40]. This phenomenon occurs because the soft tissues of the periodontal ligaments decay after death, severing the connection between the teeth and the alveoli . The mode of preservation of hadrosaurid dental batteries is inconsistent with the model that hadrosaurid teeth were fused together by hard tissue. As we and others  have shown, hadrosaurid teeth were not shed, but ground down completely, indicating that isolated hadrosaurid teeth are actually the remains of dissociated dental batteries. Many Late Cretaceous vertebrate microsites are replete with isolated hadrosaurid teeth [41, 42, 43] and many dental batteries preserve clusters of teeth that have become dissociated from the rest of the battery [39, 44] (Fig. 5f), which further support a soft tissue attachment between individual teeth. This re-interpretation of hadrosaurid tooth attachment not only more accurately explains the mechanisms underlying hadrosaurid dental taphonomy, but also has important implications for the evolution and function of the hadrosaurid dentition.
Whereas all dinosaurs studied to date possessed a ligamentous tooth attachment system [9, 45, 46], the hadrosaurid dental battery is unique in vertebrate evolution because the entire assembly, as well as individual teeth, were suspended within the jaw by ligaments. Even more surprisingly, individual teeth were suspended to their neighbors within the battery by periodontal ligaments from the onset of tooth development, thus allowing the whole battery to continuously erupt and respond to the compressive forces of chewing as a complex unit. The periodontal ligament permits teeth to erupt into the oral cavity in mammals with continuously-erupting teeth [25, 47, 48] and hadrosaurids exhibit an analogous condition using the same dental tissues. The unique form of tooth-to-tooth attachment in hadrosaurids also provides clear evidence of the modular nature of developing teeth [49, 50], because each developing tooth bud in hadrosaurids formed its own periodontal tissues to which it was attached by a ligament. This would have provided a tremendous mechanical advantage to the dental battery, probably even greater than in mammals. In mammals, the periodontal ligament serves as a shock absorber to dissipate the forces of dental occlusion  between individual teeth aligned in a single row mesiodistally. Given the grinding motions that hadrosaurs employed to consume plant material , a grinding battery possessing several hundred small teeth that were individually suspended by ligaments would have been extremely advantageous. The sheer number and small sizes of these interconnected teeth (much smaller than in ornithopods that did not possess dental batteries ), and their sophisticated ontogeny demonstrate that these dinosaurs evolved a more complex dental system than herbivorous mammals. The differences between hadrosaurid dental batteries and the grinding teeth of mammalian herbivores show convergent solutions to the problem of constant tooth wear. The major differences between ungulate mammals and hadrosaurids relate to the fact that mammals have lost continuous tooth replacement and thus individual teeth are extensively modified for efficient grinding, whereas hadrosaurid dental batteries take advantage of reptilian polyphyodonty.
We also found the same tissue arrangements and ligamentous tooth attachments in nestling (MOR 548) and embryonic (MOR 559) hadrosaurid jaws (Fig. 6). Each of these specimens shows the development of multiple generations of teeth at a single locus, erupted teeth with plugged pulp cavities, partial root resorption of neighbouring teeth, and clear evidence of ligamentous tooth-to-tooth attachments across tooth positions (Fig. 6). These data suggest that the teeth of embryonic and neonatal hadrosaurids were similarly advanced in their degree of tooth tissue formation compared to adults. Late-stage embryos of modern crocodilians also develop, and even replace, multiple generations of teeth prior to hatching [51, 52, 53], but the presence of multiple generations of interconnected teeth in an embryonic and neonatal hadrosaurid suggests that the formation of a dental battery began in ovo and may have been functional immediately after hatching.
Dental batteries have evolved independently in ornithischian (Neoceratopsia, Hadrosauridae) and saurischian (Rebbachisauridae) dinosaurs [2, 54, 55, 56], however, the present study is the first to examine dental battery development and evolution at the histological level. Therefore, the uniqueness of hadrosaurid dental ontogeny and histology depends on further comparisons with the dentitions of neoceratopsians and rebbachisaurid sauropods. One key prediction that can be made from this work is that future studies in the aforementioned taxa will uncover adaptations to avoiding typical amniote tooth replacement either by accelerated tooth development (as in hadrosaurids) or by some other means of spatially separating older teeth from the dental lamina [57, 58]. A brief survey of descriptions of sauropod dentitions shows that they consist of numerous generations of replacement teeth that are spatially separated from one another and gradually migrate to their functional positions through ontogeny [54, 55, 59]. By comparison, the stacked teeth of the neoceratopsian dental battery consist of multiple generations of teeth that are closely packed, apparently more so than in hadrosaurids . Each developing neoceratopsian tooth appears to be nested within the pulp cavity of its predecessor , a phenomenon that does not occur in hadrosaurid teeth. These brief comparisons suggest that studying dental battery evolution in dinosaurs will reveal different ways in which dental ontogeny has been modified to produce novel dental systems.
The evolutionary and ontogenetic model we have proposed here may also help explain other aspects of hadrosaurid dental histology. Although the enamel of hadrosaurid teeth is clearly homologous to that of other amniotes, the enamel surfaces of hadrosaurid teeth possess a rough surface micromorphology consisting of microscopic enamel globules [60, 61, 62]. Sander  hypothesized that the globular surface texture of hadrosaurid teeth, which results in a dull luster to the enamel surface (Fig. 1c), may have evolved as a peculiarity of dental battery development, given that it is only found in hadrosaurid and neoceratopsian teeth. Sander  proposed that the micrometer-scale enamel globules along the surfaces of hadrosaurid and neoceratopsian teeth may be the byproduct of crowding of the inner dental epithelium during amelogenesis, but this was probably not the case, given that hadrosaurid teeth resorbed significant portions of adjacent teeth in order to accommodate their development (Figs. 4c and 5). If this feature is not an adaptation to extensive tooth wear, then another plausible explanation is that it is related to the relatively rapid deposition of enamel, similar to the rapid rates of dentine and cementum deposition. However, further detailed examinations of enamel surface microstructure at several ontogenetic stages using Scanning Electron Microscopy are needed in order to test this hypothesis, as this surface morphology would be predicted to be present at all ontogenetic stages of enamel formation.
The first ornithopod taxa to retain more than one generation of replacement teeth were basal hadrosauroids that appeared in the Early Cretaceous, with the highly integrated dental batteries evolving in tandem with further tooth proliferation and miniaturization in Late Cretaceous Hadrosauridae [1, 39]. Our findings show that the novelty of the hadrosaurid dental system lies in how ancestral reptilian dental tissues were used in the interaction between individual teeth and between families of teeth. Each tooth is analogous to a single scale on medieval armor, or a single denticle in shark skin, where each scale or denticle is a rigid structure, whereas the interconnecting material provides flexibility . Despite the structural complexity of their batteries, dental development in hadrosaurids was constrained by the same processes that govern tooth formation in all other amniotes, given that each tooth is composed of homologous tissues to those in stem and crown Amniota [16, 49, 50]. We not only attribute the differences between hadrosaurid dental development and other amniotes to a shift in the orientation of typical crown and root tissues, but also to heterochronic acceleration , which allowed hadrosaurids to halt tooth replacement, co-opt the tooth roots into the grinding surface, and maintain multiple generations of functional teeth at the same locus [7, 39, 65]. Heterochrony in dental development was therefore a key evolutionary event that promoted the rapid diversification of hadrosaurid dinosaurs, the dominant herbivores in many Late Cretaceous communities [1, 39].
CMN, Canadian Museum of Nature, Ottawa, Ontario, Canada; MOR, Museum of the Rockies, Bozeman, Montana, U. S. A.; ROM, Royal Ontario Museum, Toronto, Ontario, Canada; UMNH, Natural History Museum of Utah, Salt Lake City, Utah, U. S. A.; USNM, National Museum of Natural History, Washington, D. C
We thank K. Seymour (ROM) and B. Iwama (ROM), T. Cullen (ROM), K. Brink (ROM), M. Silcox (University of Toronto), K. Shepherd (CMN), J. Horner (MOR) for assistance in obtaining material for sectioning and imaging. We are also indebted to D. Dufault for preparing the illustration in Fig. 3 and K. Chiba for photographing the specimen in Fig. 5f. Lastly, we thank two anonymous reviewers for their constructive comments and suggestions. Funding for this project came from a Queen Elizabeth II/Dr. F. M. Hill Graduate Scholarship in Science and Technology to A. LeBlanc and an NSERC Discovery Grant to R. Reisz.
Funding was provided by Queen Elizabeth II/Dr. F. M. Hill and Pfizer Graduate Scholarships in Science and Technology awarded to A. LeBlanc and an NSERC Discovery Grant to R. Reisz.
Availability of data and material
The data supporting the conclusions of this article are included within the article (and its additional files).
ARHL, RRR, DCE conceived the project, ARHL, AMB collected the data, ARHL, RRR, DCE, AMB wrote the manuscript. All authors read and approved the final manuscript.
The authors declare that they have no competing interests.
Consent for publication
Ethics approval and consent to participate
All specimens in this study were sectioned at the Royal Ontario Museum or the Museum of the Rockies with the approval for histological sampling from the appropriate institutions (Canadian Museum of Nature, Museum of the Rockies, Natural History Museum of Utah, National Museum of Natural History, and the Royal Ontario Museum).
- 2.Edmund AG. Tooth replacement phenomena in the lower vertebrates. R Ont Mus Life Sci Div Contrib. 1960;52:1–190.Google Scholar
- 4.Weishampel DB. Hadrosaurid jaw mechanics. Acta Palaeontol Pol. 1983;28:271–80.Google Scholar
- 7.Edmund AG. On the foramina in the jaws of many ornithischian dinosaurs. Contrib R Ont Mus Div Zool Palaeontol. 1957;48:1–14.Google Scholar
- 8.Erickson GM, Zelenitsky DK. Osteohistology and occlusal morphology of Hypacrosaurus stebengeri teeth throughout ontogeny with comments on wear-induced form and function. In: Eberth DA, Evans DC, editors. Hadrosaurs. Bloomington: Indiana University Press; 2014. p. 422–32.Google Scholar
- 10.Fong RKM, LeBlanc ARH, Berman DS, Reisz RR. Dental histology of Coelophysis bauri and the evolution of tooth attachment tissues in early dinosaurs: Dinosaur Dental Histology. J. Morphol. 2016; doi: 10.1002/jmor.20545
- 11.Nanci A. Ten Cate’s Oral Histology: development, structure, and function. 7th ed. St. Louis: Mosby; 2007.Google Scholar
- 24.Ostrom JH. Cranial morphology of the hadrosaurian dinosaurs of North America. Bull Am Mus Nat Hist. 1961;122:1–186.Google Scholar
- 26.Cope ED. On the characters of the skull in the Hadrosauridae. Proc. Nat. Sci. Phila. 1883;35:97–107.Google Scholar
- 28.Pretto F, Cabreira SF, Schultz CL. Tooth microstructure of the Early Permian aquatic predator Stereosternum tumidum and paleobiological implications. Acta Palaeontol Pol. 2014;59:125–33.Google Scholar
- 36.Dixon PM. The gross, histological, and ultrastructural anatomy of equine teeth and their relationship to disease. Proc Annu Conv AAEP. 2002;48:421–37.Google Scholar
- 40.Brand LR, Hussey M, Taylor J. Decay and disarticulation of small vertebrates in controlled experiments. J Taphon. 2003;1:69–95.Google Scholar
- 41.Coombs Jr WP. The status of the dinosaurian genus Diclonius and the taxonomic utility of hadrosaurian teeth. J Paleontol. 1988;62:812–7.Google Scholar
- 44.Young CC. The dinosaurian remains of Laiyang, Shantung. Palaeontol Sin New Ser. 1958;142:1–144.Google Scholar
- 45.Reid REH. Bone histology of the Cleveland-Lloyd dinosaurs and of dinosaurs in general, Part I: Introduction: Introduction to bone tissues. Brigh Young Univ Geol Stud. 1996;41:25–72.Google Scholar
- 46.Martin LD, Stewart JD. Implantation and replacement of bird teeth. Smithson Contrib Paleobiol. 1999;89:295–300.Google Scholar
- 54.Sereno PC, Wilson JA. Structure and evolution of a sauropod tooth battery. In: Curry Rogers KA, Wilson JA, editors. The Sauropods: Evolution and Paleobiology. Berkeley, California: University of California Press; 2005. p. 157–177.Google Scholar
- 61.Sander PM. The microstructure of reptilian tooth enamel: terminology, function, and phylogeny. Münch Geowiss Abh. 1999;38:1–102.Google Scholar
- 63.Ehrlich H. Materials Design Principles of Fish Scales and Armor. Biol. Mater. Mar. Orig. Dordrecht: Springer Netherlands; 2015; doi 10.1007/978-94-007-5730-1_5 p. 237–62.
Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.