Soft Tissue in Dinosaur Bones

Intact collagen, blood vessels, and cell-like structures in fossils supposedly 65 million years old — when the biochemistry says collagen cannot last that long.

In March 2005, the journal Science published a paper by Mary Schweitzer of North Carolina State University and her co-authors, titled “Soft-Tissue Vessels and Cellular Preservation in Tyrannosaurus rex.”¹ The paper reported something that, by all biochemical expectations, should not exist.

Schweitzer’s team had recovered, from a T. rex femur excavated in the Hell Creek Formation of Montana in 2003, what appeared to be pliable, flexible tissue — soft, transparent vessels with branching structure, fibrous matrix with the texture of muscle, and what looked like intact osteocytes (bone cells) with cellular processes still attached.

The fossil was conventionally dated to about 68 million years old.

This was, to put it mildly, not what anyone expected.

What was actually found

The first paper documented the morphology — the structure of the material. Subsequent papers, by Schweitzer’s group and by independent researchers, pushed further:

• In 2007, mass spectrometry of the soft tissue produced peptide sequences that matched chicken collagen — a real protein, partly sequenced, recovered from a fossil supposedly seventy million years old.²
• In 2009, the analysis was extended to a hadrosaur fossil from the same formation, with confirming results.³
• In 2015, an independent group at Imperial College London reported, in Nature Communications, similar findings in several different dinosaur specimens using completely different techniques (focused ion beam imaging and time-of-flight secondary ion mass spectrometry).⁴ The point of this paper was explicitly to test whether Schweitzer’s findings were an artefact of her methodology. They reproduced.
• Subsequent work has identified additional proteins, putative blood-cell-like structures, and preservation in fossils from multiple formations and across the Mesozoic.⁵

This is, in 2026, no longer a fringe claim. It is in the highest- impact journals in biology and palaeontology. It has been reproduced. It is settled science at the level of the empirical finding.

The dispute is over what the finding means.

The chemistry problem

Collagen is one of the most stable proteins in nature, which is why it preserves better than most. But it is not eternal. Its peptide bonds are vulnerable to hydrolysis (chemical breakdown in the presence of water) and to oxidation. The kinetics of these reactions are well-studied. The Arrhenius equation lets us predict how long the protein survives at any given temperature.

The numbers are, in round figures:

• At 20°C (typical surface temperature): collagen half-life is on the order of 400 years. Effectively complete degradation in ~10,000 years.
• At 10°C (cool burial): half-life ~5,000 years. Complete degradation in ~150,000 years.
• At 0°C (frozen): half-life ~100,000 years. Complete degradation in ~3 million years.
• At -10°C (permafrost, which dinosaurs were not buried in): even longer, but capped by other degradation processes around 4– 10 Ma.

The maximum survival time of collagen, under the most generous preservation conditions imaginable, is on the order of a few million years. This is the consensus of the literature on protein diagenesis.⁶

The dinosaurs are supposed to be sixty-five to two hundred million years old.

That is one to two orders of magnitude longer than the chemistry permits. It is the same scale of impossibility as finding wet ink in a 5,000-year-old Egyptian tomb. The biochemistry does not bend on this.

The mainstream response

Schweitzer is herself, theologically, an old-earth creationist — she has been explicit in interviews that her findings do not imply young-earth chronology. She has worked actively to identify a preservation mechanism that would extend collagen survival into the deep-time range. The leading candidate she has proposed, in a 2013 paper, is iron-mediated preservation — that iron from haemoglobin, released during decomposition, crosslinks and stabilises the proteins against further degradation.⁷

Iron does, in fact, have a preservative effect on proteins. The question is whether it has enough effect.

The 2013 paper reports experiments in which collagen-rich tissue from modern ostrich vessels was incubated in iron-rich solution and showed a slowed degradation rate. The slowdown factor reported in the paper is on the order of two or three. To get from a 400-year half-life to a 68,000,000-year survival, you would need a slowdown factor of 10⁴ to 10⁵. The iron-mediated mechanism does not provide this.

The honest position is that we do not have a chemical mechanism that explains preservation of intact collagen for sixty-five million years. The iron hypothesis is a partial story, useful, but insufficient.

What it should mean

When a chemistry-based prediction is off by four orders of magnitude, the appropriate response is to revisit the assumption that drove the prediction. In any other scientific context, this is what we would do.

The most natural revision is the obvious one: the fossils are not that old. Not 68 million years; thousands. The collagen has only had thousands of years to decay, and the chemistry comfortably permits intact collagen in that interval.

The young-earth reading is not the only response available — one could in principle imagine a yet-unknown preservation mechanism operating in deep time — but it is the response that does not require postulating unknown chemistry. It is the response that takes the existing biochemical literature at face value.

For Schweitzer herself, who is a Christian but not a young-earth creationist, this has been a painful question. In interviews she has been candid about the difficulty: the data, on a straightforward reading, points to young fossils, and the theological pressure to preserve the deep-time chronology is the only thing keeping the alternative reading on the table. Her actual words, from a 2014 interview: “I do think it’s a question worth asking… because the chemistry tells us collagen shouldn’t be there.”⁸

The chemistry tells us collagen shouldn’t be there.

What this gets you

It does not, alone, prove a six-thousand-year-old earth. It does strongly suggest that the assigned ages of these specific fossils are wrong by orders of magnitude. If the ages are wrong by orders of magnitude, the entire dating framework that produced them is under pressure to explain why.

Combined with the lunar recession and ocean salinity arguments, the dinosaur-collagen evidence is a third independent line pointing in the same direction: the present is a much shorter distance from creation than the textbook tells us.

Next: carbon-14 where it shouldn’t be, which is the most direct of the four arguments and the one that most clearly puts a number on the upper bound.