An ankylosaur larynx provides insights for bird-like vocalization in non-avian dinosaurs

The study addresses how the tetrapod hyolaryngeal apparatus—particularly the larynx—evolved in archosaurs and what this reveals about vocalization in non-avian dinosaurs. Modern archosaurs (crocodilians and birds) differ in both anatomy and function of the hyolaryngeal apparatus: crocodilians use the larynx as a vocal source, whereas in birds the syrinx at the posterior trachea is the primary sound source and the larynx functions as part of the vocal tract. Although hyoid bones are sometimes preserved in fossils, a larynx has not previously been reported for extinct non-avian reptiles, likely due to its cartilaginous nature in tetrapods (except in some birds). The paper examines the exceptionally preserved hyolaryngeal apparatus of Pinacosaurus grangeri (IGM100/3186), providing the first description of a larynx in a non-avian dinosaur and comparing it to modern reptiles and birds to infer function and evolutionary implications.

Prior work shows major functional and anatomical differences in the hyolaryngeal apparatus of crocodilians and birds: crocodilian larynges act as sound sources, while birds possess a syrinx and use the larynx to modulate the vocal tract. Fossil preservation of hyobranchial elements is known across some dinosaurs and birds, but direct laryngeal fossils are exceedingly rare. The oldest previously known fossil laryngeal element (a cricoid) was from the Eocene bird Presbyornis. Various non-avian dinosaurs preserve hyoid elements (e.g., basihyal and ceratobranchials in some theropods and ankylosaurs), but no confirmed laryngeal elements had been reported in non-avian archosaurs. This study re-evaluates previously identified elements in ankylosaurs (e.g., Saichania) and suggests that some are actually laryngeal (cricoid and arytenoid), indicating ossification of the larynx occurred in some non-avian dinosaurs and providing a new data point for the evolution of avian-like vocal modification.

Specimen and context: Pinacosaurus (IGM100/3186) from Ukhaa Tolgod, Mongolia (middle Campanian, Late Cretaceous) was re-prepared for detailed observation and measurement following earlier publications. Computed Tomography (CT) and laser scanning were conducted at the American Museum of Natural History (AMNH) to study morphology. CT data were used to reconstruct the larynx; image data are available on Morphobank (ID: 7201). Comparative datasets: Laryngeal and hyobranchial elements from modern and fossil specimens (including Alligator mississippiensis and various reptile and bird taxa) in AMNH and the National Museum of Nature and Science, Tokyo, were examined. Measurements: Maximum transverse mandible width and arytenoid lengths were taken; cricoid width was also measured. For comparative morphometrics, measurements were standardized (log-transformed where appropriate) and residuals analyzed relative to mandible width. Statistical analyses: Ordinary Least Squares linear regressions assessed correlations between arytenoid length and mandible width, and between cricoid width and mandible width. Group differences were evaluated using Student’s t-tests (standardized arytenoid length: reptiles vs. birds) and Tukey HSD tests (standardized cricoid width among birds vs. crocodilians, squamates, and turtles). Reported sample sizes included, for example, arytenoid analyses with reptiles (n = 31) and birds (n = 58), and cricoid width analyses with reptiles (n = 62) and birds (n = 90). Descriptive morphology was documented for Pinacosaurus laryngeal elements (cricoid, arytenoids) and associated hyoids, including articular surfaces and inferred muscular attachment sites.

• First definitive report of a fossilized larynx (cricoid and arytenoids) in a non-avian dinosaur, Pinacosaurus grangeri (IGM100/3186), representing the oldest known laryngeal fossil in dinosaurs. - Pinacosaurus larynx shows specializations: firm and kinetic cricoid-arytenoid joint, prominent arytenoid process, elongated arytenoids, and an enlarged cricoid. These features align more with a bird-like vocal modifier function rather than a reptile-like vocal source. - Morphometrics: Arytenoids are relatively long in dinosaurs (Pinacosaurus) and birds compared to non-dinosaurian reptiles; arytenoid length positively correlates with mandible width. A Student’s t-test on standardized arytenoid length residuals supports a distinction between vocal-source (reptiles) and vocal-modifier (birds) groups (t = -5.956, df = 87, p < 0.0001). - Cricoid width correlates positively with mandible width. Birds and Pinacosaurus have relatively larger cricoid widths than non-avian reptiles; Tukey HSD on standardized cricoid width shows birds differ from crocodilians (p = 0.025) and from squamates and both vocal and quiet turtles (p < 0.0001). - Additional ankylosaur material (Saichania) likely preserves ossified laryngeal elements (cricoid and arytenoid), implying that laryngeal ossification occurred in some non-avian dinosaurs. - Functional inference: The specialized joint and long arytenoids in Pinacosaurus likely facilitated easy opening of the glottis and fine control of airflow as in birds, suggesting the larynx primarily acted as a sound modifier. The enlarged cricoid, also seen in vocal reptiles, may indicate association with loud, efficient sound radiation. - Despite the absence of a known syrinx in non-avian dinosaurs, the morphology suggests a non-laryngeal sound source with a bird-like, large, kinetic larynx enabling avian-like vocalization.

The discovery of an ossified larynx in Pinacosaurus fills a critical gap in the archosaur fossil record, indicating that non-avian dinosaurs could possess bird-like laryngeal features. Morphological attributes—such as a firm cricoid-arytenoid joint, long arytenoids, and large cricoid—support a function focused on glottal opening and modulation of the vocal tract rather than direct sound generation, paralleling the condition in birds. Morphometric analyses corroborate that dinosaurs (Pinacosaurus) and birds cluster in traits associated with vocal modification rather than vocal-source larynges typical of many non-dinosaurian reptiles. The enlarged cricoid across vocal taxa suggests adaptive significance for sound radiation efficiency. The rare preservation and ossification of laryngeal elements in ankylosaurs (Pinacosaurus, Saichania) also indicate phylogenetically broader experimentation with laryngeal structure within dinosaurs. These findings imply that bird-like vocal strategies—using a non-laryngeal sound source combined with a modulating larynx—may have evolved prior to the emergence of Aves, with potential behavioral implications for communication such as courtship, territoriality, and predator defense.

This study provides the first definitive description of a larynx in a non-avian dinosaur and the oldest laryngeal fossil evidence in dinosaurs. Pinacosaurus exhibits a bird-like, enlarged, and kinetic larynx suited for vocal modulation rather than sound production, implying avian-like vocal strategies in non-avian dinosaurs. The results reshape understanding of archosaur hyolaryngeal evolution, suggesting bird-like vocalization mechanisms emerged before Aves. Future research should target broader sampling across dinosaur clades, refine functional models linking laryngeal morphology to acoustic performance, and seek additional soft-tissue proxies or osteological correlates (including potential syrinx evidence) to test the generality of these findings.

• Mapping of laryngeal function is based on osteological correlates and comparative morphology; soft tissues are not preserved. - Syrinx remains unknown in non-avian dinosaurs; inference of a non-laryngeal sound source is indirect. - Preservation of laryngeal elements is extremely rare, limiting sample size to a few ankylosaur specimens and modern comparative datasets. - It remains unclear whether the observed laryngeal size and configuration generalize across dinosaur clades or represent ankylosaur-specific adaptations. - Some elements in previously described taxa (e.g., Saichania) are reinterpreted; additional verification and broader taxonomic sampling are needed.