The Most Beautiful Experiment in Biology

In one line: In 1958 Matthew Meselson and Franklin Stahl proved that DNA copies itself semiconservatively — each new double helix keeps one old strand and one new — by doing nothing more than making old DNA heavier than new DNA and spinning it until the two settled into separate bands.

---

The world then

Five years earlier, Watson and Crick had published the structure of DNA: a double helix of two complementary strands (1953). In a famous understatement they noted that the pairing "immediately suggests a possible copying mechanism" — unzip the helix, and each strand could template a new partner. But that was a suggestion. By 1958 no one had shown how the two original strands are actually parcelled out when DNA copies itself.

The question

There were three rival pictures. Semiconservative (Watson–Crick): each daughter helix keeps one old strand and one new. Conservative: the original helix stays intact and an entirely new one is built alongside it. Dispersive (Max Delbrück, 1954): the old DNA is broken up and scattered, old and new interleaved along both strands of both daughters. The three differ only in where the original atoms end up.

The design

Meselson and Stahl made "old" and "new" DNA weigh different amounts. They grew E. coli for many generations on a medium whose only nitrogen was the heavy stable isotope ¹⁵N, so every strand was slightly dense — a density label, not a radioactive one. Then they shifted the culture to ordinary light ¹⁴N: from that instant, every newly made strand was light while every old strand stayed heavy. To read the weights they used cesium-chloride density-gradient centrifugation — a method Meselson developed with Jerome Vinograd. Spun to equilibrium, the CsCl forms a density gradient, and each DNA molecule bands at the layer matching its own buoyant density: heavy low, light high, half-heavy exactly between. UV-absorption photography made the bands visible.

The result

At the start, all the DNA was heavy: one low band. After one generation in ¹⁴N, the heavy band was gone — replaced by a single band at intermediate density, every molecule now half-heavy. That alone killed the conservative model, which required two bands (one heavy, one light). But a single half-heavy band was equally consistent with semiconservative and dispersive — so they ran a second generation. After two generations the band split in two: half still half-heavy, half now fully light. That killed the dispersive model, which predicts a single band drifting steadily lighter, never splitting. Only semiconservative survived. As a clincher, heating the half-heavy DNA separated it into one heavy strand and one light strand — direct proof that the hybrid molecule is one old strand paired with one new.

What we learned, and why it's beautiful

DNA replication is semiconservative: every time a cell divides, the helix unzips and each strand templates a fresh partner, so each daughter molecule carries one strand inherited from the past and one made new — the mechanism Watson and Crick had only guessed at. The beauty is the economy. Three sweeping hypotheses about the nature of inheritance, decided by one clean measurement — no radioactive tracer to interpret, no model assumed, just heavy versus light and bands in a tube you could photograph. The historian Frederic L. Holmes built a whole book around the phrase John Cairns gave it: the most beautiful experiment in biology.

Sources

Full claim-by-claim evidence is in references.md. Primary anchors:

• Meselson, M. & Stahl, F. W. "The Replication of DNA in Escherichia coli." PNAS 44(7), 671–682 (1958).
• Meselson, M., Stahl, F. W. & Vinograd, J. "Equilibrium Sedimentation of Macromolecules in Density Gradients." PNAS 43(7), 581–588 (1957).
• Watson, J. D. & Crick, F. H. C. "Molecular Structure of Nucleic Acids." Nature 171, 737–738 (1953).
• Delbrück, M. "On the Replication of Desoxyribonucleic Acid (DNA)." PNAS 40(9), 783–788 (1954).
• Holmes, F. L. Meselson, Stahl, and the Replication of DNA: A History of "The Most Beautiful Experiment in Biology." Yale University Press (2001).

Accuracy notes: ¹⁵N is a non-radioactive density label (nitrogen, not carbon), distinct from the radioactive tracers of Hershey–Chase. The single half-heavy band after one generation rules out only the conservative model; the dispersive model is excluded only by the two-band result after the second generation. The experiment showed a conserved subunit that is single-stranded; identifying that subunit with the Watson–Crick polynucleotide chain rests on the structural model. "Most beautiful experiment in biology" is John Cairns's remark, popularised by Judson and Holmes — not the authors' own words.