Pneumatic Bones

The pneumatic nature of dinosaur bones was first recognized by Richard Owen in 1856, when he described deep lateral fossae on the neural arches of the theropod Becklespinax altispinax from the Wealden Supergroup of England. Owen wrote that these were "three deep depressions, probably receiving parts of the lungs in the living animal." This constitutes the earliest published inference of postcranial pneumaticity in any non-avian dinosaur. In 1870, Harry Seeley described the sauropod Ornithopsis hulkei from the Wealden and correctly interpreted its enormous lateral foramina and honeycomb-like internal camellae as evidence of a bird-like air sac respiratory system. Seeley inferred "bird-like heart and lungs" for the animal based on these features. During the same period, Edward Cope (1877) and Othniel Marsh (1877) independently recognized pneumatic features in the North American sauropods Camarasaurus and Apatosaurus, respectively. Cope considered the pneumatic interpretation so self-evident that he did not bother to defend it.

Despite these early insights, the significance of skeletal pneumaticity in dinosaurs was largely overlooked for nearly a century. The complex vertebral laminae were employed as taxonomic tools, but their pneumatic implications were seldom discussed. Werner Janensch (1947) provided a notable exception, arguing convincingly for the pneumatic nature of sauropod vertebrae based on his studies of the Tendaguru fauna, though he interpreted the function as buoyancy maintenance for presumed aquatic habits. The modern resurgence of pneumaticity research was catalyzed by Brooks Britt's 1993 doctoral dissertation, which provided the first comprehensive survey of postcranial pneumaticity across Archosauria and established five osteological correlates for recognizing pneumatic bone in fossils: large external foramina, fossae with crenulate texture, thin outer bone walls, smooth or crenulate pneumatic tracks, and internal chambers connected to exterior foramina.

Subsequently, Mathew Wedel's series of publications beginning in 2000 revolutionized the understanding of sauropod pneumaticity. Using CT scanning of vertebrae—including running Sauroposeidon bones through a hospital CT scanner as an undergraduate—Wedel demonstrated that individual sauropod vertebrae could be up to 89% air by volume, classified pneumatic internal morphologies into a systematic framework, and traced their phylogenetic distribution. Patrick O'Connor (2006) further refined the methodology for identifying PSP in fossils, establishing strict criteria that distinguish unambiguous pneumatic features from superficially similar non-pneumatic structures found in crocodilians and other non-pneumatic archosaurs.

Wedel (2003) and Wedel et al. (2000) developed a comprehensive classification system for the internal pneumatic architecture of sauropod vertebrae, which has become the standard framework for describing pneumatic morphology in saurischian dinosaurs.

Acamerate: Pneumatic features limited to external fossae that do not significantly invade the centrum. Found in basal sauropods such as Vulcanodon and Isanosaurus. These fossae represent the simplest expression of pneumatic influence on vertebral morphology.

Procamerate: Deep fossae penetrate to a median septum dividing the centrum, but are not enclosed by bony margins to form true chambers. Haplocanthosaurus exemplifies this condition, representing an intermediate stage between simple fossae and enclosed camerae.

Camerate: Lateral pneumatic foramina open into large internal chambers (camerae) that occupy most of the centrum, with a regular branching pattern producing secondary and occasionally tertiary camerae. Camarasaurus is the prototypical camerate sauropod and the genus for which the term was originally coined.

Polycamerate: Successive bifurcations of pneumatic diverticula produce three or more generations of progressively smaller camerae, with increasing numbers of branches at each generation. Found in derived diplodocids such as Apatosaurus and Diplodocus, where radially arranged camerae fill condyles and cotyles.

Camellate: The entire internal structure of centra and neural spines is composed of numerous small, irregularly arranged camellae with thin bony septa (1–3 mm thick). Sauroposeidon is the most extreme known example; Wedel discovered that its 137-cm cervical vertebrae contained bone walls so thin that the vertebrae reached approximately 89% air by volume.

Somphospondylous: A derived subcategory of camellate morphology in which neural arch laminae are reduced and the neural spine has an inflated appearance. Found in derived titanosaurians and forms the basis for the clade name Somphospondyli.

Within any individual sauropod, there is considerable serial variation: posterior cervical vertebrae tend to exhibit the most complex pneumatic architecture, while dorsal and caudal vertebrae are typically simpler.

In extant birds, postcranial pneumatization is a postnatal ontogenetic process. After hatching, diverticula extending from the air sacs contact and begin to remodel the skeleton. The primary vertebral diverticulum—the diverticulum intertransversalis—follows the brachial plexus and vertebral artery, advancing through the transverse foramina. From these major diverticula on either side of the vertebral column, smaller accessory diverticula enter the bone through existing nutrient foramina and spread throughout the bone to produce an interconnected system of camellae. Eventually, even the neural spines and cervical ribs become filled with pneumatic camellae.

Critically, pneumatization proceeds from two independent directions: the cervical air sacs pneumatize the cervical and anterior thoracic vertebrae, while the abdominal air sacs pneumatize the posterior thoracic vertebrae, synsacrum, and in some species the hindlimb. The anterior and posterior thoracic air sacs lack invasive diverticula and do not pneumatize any bones. Because the two advancing fronts of pneumatization proceed independently, they occasionally fail to meet, leaving a gap of non-pneumatized vertebrae in the middle of the spinal column—a phenomenon Wedel termed a "pneumatic hiatus."

Wedel recognized that if dinosaurs possessed the same dual air sac system, some dinosaur skeletons should display the same gap pattern. By revisiting classic descriptions of sauropods from the early 1900s, he discovered that several specimens showed exactly this pattern: pneumatized vertebrae in the neck and pelvis regions, with solid vertebrae in between. This constituted powerful evidence that saurischian dinosaurs possessed anterior and posterior air sacs homologous to those of modern birds.

In CT scans of juvenile diplodocid vertebrae, Wedel also observed tiny resorption pits (coels) in the bone ahead of developing camerae, consistent with osteoclastic activity driven by the advancing pneumatic epithelium—the same process observed during pneumatization of bird bones.

Mass Reduction: The most widely accepted function of postcranial skeletal pneumaticity is reduction of skeletal mass. This was particularly significant for sauropods, whose vertebral columns bore enormous mechanical loads. Wedel (2003) argued that air sacs and skeletal pneumaticity "probably facilitated the evolution of extremely long necks in some sauropod lineages by overcoming tracheal dead space problems and by lightening the neck." Without pneumatization, the cervical vertebrae of large sauropods would have been prohibitively heavy. In extant birds, large flying species generally show proportionally greater pneumaticity than small ones, and Britannica notes that "the highly pneumatic bones of large flying birds are reinforced with bony struts at points of stress." However, recent work has nuanced this view by demonstrating that avian bones are not necessarily lightweight in absolute terms but are light relative to their structural strength.

Respiratory Efficiency: The presence of pneumatic bones strongly implies the existence of a bird-like air sac system supporting unidirectional airflow through the lungs. In birds, the respiratory system achieves oxygen extraction during both inhalation and exhalation: posterior air sacs draw fresh air through the lungs during inhalation, and anterior air sacs push the same air through the lungs during exhalation. This continuous, one-way flow of air across the gas exchange surfaces (parabronchi) is significantly more efficient than the tidal, bidirectional ventilation of mammalian lungs. The inference that saurischian dinosaurs shared this efficient respiratory mechanism has profound implications for understanding their metabolic capabilities and activity levels.

Thermoregulation: The extensive system of air sacs and pneumatized bones distributed throughout the body may have aided in dissipating metabolic heat. For multi-tonne sauropods, managing core body temperature would have been a critical physiological challenge, and an internal network of air-filled spaces could have facilitated convective heat transfer.

Non-ventilatory Functions: In birds, the air sac system also plays roles in buoyancy (especially in aquatic species), phonation and display, and potentially other functions. These are generally considered exaptations of a primarily respiratory system.

Theropoda: Unambiguous postcranial skeletal pneumaticity first appeared in the Late Triassic (approximately 210 Ma), with clear evidence in coelophysoids such as Coelophysis bauri. PSP became increasingly extensive through theropod evolution, with derived forms such as Tyrannosaurus rex showing pneumatization extending into the pelvis and proximal caudal vertebrae. Wedel estimated that more than 10% of T. rex's body volume may have comprised air space. The 2012 comprehensive review by Benson et al. documented the distribution and physiological implications of pneumatic postcranial bones across Theropoda.

Sauropodomorpha: Basal sauropodomorphs such as Thecodontosaurus caducus show incipient pneumatic features in their vertebrae. Postcranial pneumaticity became progressively more elaborate through sauropod evolution, from simple fossae in basal forms to the extraordinary camellate vertebrae of derived titanosauriforms. This escalating complexity correlates with increasing body size and neck length. Crucially, a 2022 study by Aureliano et al. using μCT scanning of the oldest known dinosaurs from Brazil (~233 Ma)—Buriolestes, Pampadromaeus, and Gnathovorax—found no unambiguous evidence of PSP, suggesting that invasive air sac systems evolved independently in multiple archosaur lineages rather than being inherited from a single common ancestor.

Pterosauria: Pterosaurs also possessed pneumatic postcranial bones, with evidence documented in specimens dating back to the Late Triassic. Butler et al. (2009) showed that even the earliest pterosaurs exhibited PSP, indicating it was a fundamental feature of the clade likely linked to the demands of flight.

Ornithischia: Unambiguous postcranial skeletal pneumaticity has not been documented in any ornithischian dinosaur. Butler et al. (2012) proposed that pulmonary air sacs may have been present in the common ancestor of Ornithodira but were subsequently lost or reduced in ornithischians. Alternative hypotheses suggest that PSP may have evolved independently in Saurischia and Pterosauria without ever being present in the ornithischian lineage.

Humans and Other Mammals: Humans possess pneumatic bones in the skull—the frontal, maxillary, ethmoid, and sphenoid bones that house the paranasal sinuses. However, these are cranial pneumatic structures associated with the nasal cavity, fundamentally different in origin and mechanism from the postcranial pulmonary pneumaticity of birds and dinosaurs. No extant mammal exhibits postcranial skeletal pneumaticity.

O'Connor (2006) established the rigorous criteria now standard for recognizing unambiguous postcranial skeletal pneumaticity in fossil archosaurs. The key insight was that many external features previously used to infer pneumaticity—including fossae, foramina, and vertebral laminae—are also present to varying degrees in non-pneumatic taxa such as extant crocodilians, where they are associated with adipose deposits or vascular structures rather than air sacs.

Unambiguous evidence of PSP requires the co-occurrence of three features: (1) well-developed fossae and laminae on the external vertebral surface; (2) foramina within these fossae; and (3) connection of these foramina to internal pneumatic chambers (camerae or camellae). Internal chambers that lack connection to exterior foramina are considered non-pneumatic and were likely filled with marrow or fat in life.

Modern research increasingly employs μCT (micro-computed tomography) to non-destructively examine the internal architecture of fossil vertebrae at high resolution. This technology has been instrumental in resolving longstanding questions about the distribution and evolution of PSP, including the recent demonstration by Aureliano et al. (2022) that the earliest dinosaurs lacked invasive air sac systems.

Pneumatic bones represent one of the most compelling lines of anatomical evidence linking non-avian dinosaurs to modern birds. The shared pattern of dual-origin pneumatization—with anterior structures pneumatized by cervical air sacs and posterior structures by abdominal air sacs—demonstrates that this respiratory architecture was established well before the evolution of flight. Wedel's discovery of pneumatic hiatuses in dinosaur vertebral columns, mirroring the same developmental phenomenon seen in modern birds, provided what he called "the last stake in the coffin" for the hypothesis that saurischian dinosaurs possessed bird-like air sacs.

The evolutionary origin of postcranial skeletal pneumaticity remains an active area of research. Whether PSP evolved once in the common ancestor of Ornithodira (the clade containing dinosaurs and pterosaurs) or arose independently in multiple lineages has been debated. The 2022 μCT study of the earliest known dinosaurs from Brazil supports the multiple-origins hypothesis, suggesting that the invasive air sac system and associated skeletal pneumaticity may have evolved at least twice—once in pterosaurs and once (or more) within Saurischia—rather than being inherited from a single ornithodiran ancestor. Butler et al. (2012) proposed an alternative scenario in which non-invasive air sacs were primitively present in Ornithodira, with invasive pneumatization of the skeleton evolving separately in different lineages as body size increased and respiratory demands intensified.