How researchers test reconstructions of dinosaur organs

Reconstructing an internal organ begins with a specific question: does a fossil preserve tissue, the space once occupied by tissue, or a skeletal feature associated with that tissue in living animals? Those are different kinds of evidence. A mineralised intestinal segment can retain anatomical detail. A braincase records a surrounding cavity. A pneumatic vertebra can indicate respiratory diverticula without preserving the lung itself. Researchers need to establish which kind of record they are examining before proposing how the organ worked.

What survives after the soft anatomy disappears

Most viscera decay before a skeleton is buried deeply enough for long-term preservation. Exceptional deposits can capture tissues through early mineral replacement, films, impressions or chemically altered residues. Even an unusually complete specimen has passed through decomposition, compaction and mineral change. A preserved patch therefore needs a geological explanation as well as an anatomical one: how was its shape retained, and what happened to the surrounding sediment?

The Early Cretaceous Italian dinosaur Scipionyx, described by Cristiano Dal Sasso and Marco Signore in 1998, demonstrates how much this exceptional record can reveal. Its tiny skeleton retains intestinal and other soft-tissue structures in a flattened limestone slab. Their location and continuity provide information unavailable from an isolated bone. They do not supply an undistorted, complete set of organs, and a juvenile of one species cannot serve as the internal blueprint for every dinosaur.

CT reveals structure before interpretation

Computed tomography measures how X-rays pass through an object from many directions and reconstructs a volume of small three-dimensional units called voxels. Researchers can inspect virtual slices through a skull, follow a canal or isolate a hollow space without cutting through the specimen. The method is especially valuable where a delicate fossil remains partly enclosed in rock. Its success depends on contrast between materials and the resolution of the scan.

The coloured surfaces seen in a published model usually result from segmentation: a researcher or algorithm separates regions interpreted as bone, matrix or cavities. Boundaries can be uncertain when minerals have similar X-ray properties, cracks intersect an anatomical feature, or a structure is smaller than the scan can resolve. A smooth digital rendering is consequently a visualisation of selected data, not an independent confirmation that the selected region was an organ.

Brain cavities are not preserved brains

An endocast reconstructs the interior of a braincase. It may help identify the positions of olfactory regions, nerve exits, blood vessels and parts of the inner ear. The cavity also contained membranes, blood vessels and other tissues, and the brain did not fill it in the same proportion in all animals. Converting endocranial volume into brain volume requires a comparative assumption; converting that estimate into a neuron count requires another.

This distinction has practical consequences. A 2024 reassessment by Kai Caspar and colleagues challenged claims of primate-like cognition in large theropods derived from estimated neuron numbers. It showed how assumptions about brain occupancy and comparisons with living animals change the result. Endocasts remain useful anatomical evidence, but they do not record an animal's learning ability or a behavioural score. The broader guide to dinosaur internal anatomy explains how these partial records fit together.

Reading respiratory traces in bones

In many sauropodomorphs and theropods, openings lead into internal chambers with patterns comparable to skeletal pneumaticity in birds. These features support the presence of air-filled diverticula extending from the respiratory system. Their distribution helps constrain where such extensions reached along the neck, trunk and tail. An external pit alone is weaker evidence than an opening that demonstrably communicates with an internal chamber.

Air sacs, lungs and airflow should nevertheless be distinguished. Gas exchange occurs in the lung; an air sac is not a second gas-exchanging lung. Experiments reported by Farmer and Sanders in 2010 documented unidirectional airflow in alligator lungs despite the absence of avian air sacs. Living archosaurs therefore show that airflow direction and skeletal pneumaticity need not be a single inseparable package. Pneumatic bones cannot, by themselves, establish oxygen consumption, exercise capacity or body temperature.

Digestive evidence has several scales

Recognisable food remains in the abdominal region can identify a meal, provided their relationship with the skeleton rules out accidental overlap during burial. Coprolites preserve material that passed through a digestive system, although the producer may be uncertain. Neither source automatically reveals intestinal length, digestion time or the whole lifetime diet. The arrangement of ribs can constrain abdominal space, but it does not measure the soft organs that once occupied it.

Stones need similar scrutiny. Position, concentration, rock type and burial context help assess whether stones near a skeleton were swallowed. Experiments and comparisons with ostriches, published by Oliver Wings and Martin Sander in 2007, challenged the idea that sauropods routinely possessed an effective bird-like gastric mill. A smooth stone is consequently insufficient evidence for a grinding stomach, and plant eating does not require all dinosaur lineages to share the same digestive mechanism.

A rejected heart claim and a useful comparison

A mineral mass inside the Thescelosaurus specimen nicknamed Willo was interpreted as a heart in 2000. A 2011 reexamination by Timothy Cleland, Michael Stoskopf and Mary Schweitzer used histological, chemical and morphological tests and did not substantiate a preserved heart. The case illustrates why a plausible location and organ-like outline are insufficient when internal structure and composition disagree with the identification.

Birds and crocodilians still provide a strong comparative basis for reconstructing a four-chambered dinosaur heart. That is an inference from evolutionary relationships, rather than a fossilised cardiac muscle. Detailed pumping performance, vessel elasticity and blood pressure require additional assumptions. The same distinction applies throughout internal anatomy: a conclusion may be well supported without being directly preserved, provided its evidence and remaining uncertainty are identified.