By the 1920s geneticists could map genes without knowing what a gene was. Chromosomes contain two main substances, protein and deoxyribonucleic acid, or DNA. The Swiss chemist Friedrich Miescher had isolated DNA, which he called nuclein, from pus-soaked bandages in 1869. Yet for decades most scientists assumed genes must be made of protein. Proteins are built from twenty kinds of amino acids and take endless forms. Proteins seemed varied enough to carry instructions, while DNA, built from just four units, looked too simple. The influential biochemist Phoebus Levene even proposed that DNA was a monotonous repeat of its four units, which could carry no information at all.

The transforming principle

The first clue came from medicine. In 1928 the British medical officer Frederick Griffith was studying pneumococcus, the bacterium behind a deadly form of pneumonia. One strain had a smooth coat and killed mice; a rough strain was harmless. Heat-killed smooth bacteria were harmless too. But when Griffith injected mice with live harmless bacteria mixed with heat-killed deadly ones, the mice died, and he recovered live smooth bacteria from them. Something from the dead cells had transformed the living ones, and the change was inherited by their descendants.

At the Rockefeller Institute in New York, Oswald Avery, Colin MacLeod and Maclyn McCarty spent years purifying this "transforming principle." In 1944 they reported that the transforming principle was DNA: enzymes that destroyed protein left it active, while an enzyme that destroyed DNA abolished it. Many scientists were unconvinced, suspecting a trace of protein contamination.

Blenders and viruses

In 1952 Alfred Hershey and Martha Chase settled the matter for most biologists using bacteriophages, viruses that infect bacteria. A phage is little more than a protein coat around DNA. Hershey and Chase labeled the protein with radioactive sulfur and the DNA with radioactive phosphorus, let the phages infect bacteria, then knocked the empty coats off in a kitchen blender. Only the DNA of the virus entered the bacteria and turned up in the new viruses; the protein stayed outside.

Chargaff's rules

DNA's four units, called bases, are adenine (A), thymine (T), guanine (G) and cytosine (C). In the late 1940s the biochemist Erwin Chargaff measured them carefully in many species and found two patterns that demolished the monotonous-repeat idea. First, the base composition differs from species to species, so DNA could vary enough to carry information. Second, within any species the amount of A equals the amount of T, and G equals C. Chargaff did not know why.

The double helix

The reason came from structure. At King's College London, Rosalind Franklin, an expert in X-ray crystallography, and her student Raymond Gosling produced sharp images of DNA fibers. One, known as Photo 51 and taken in 1952, showed a clear X-shaped pattern that signaled a helix. Without Franklin's knowledge, her colleague Maurice Wilkins showed the photograph to James Watson, who with Francis Crick at Cambridge was building models. In April 1953 Watson and Crick published their structure: two strands wound around each other, with bases pairing A with T and G with C across the middle like rungs on a twisted ladder. The pairing explained Chargaff's rules at a stroke. It also suggested how DNA copies itself, a point Watson and Crick noted with famous understatement: each strand could serve as a template for building its partner.

Franklin died of ovarian cancer in 1958, aged 37. When Watson, Crick and Wilkins received the Nobel Prize in 1962 she could not be included, because the prize is not awarded after death. Historians now stress that her data were essential to the discovery and were used without her consent.

Copying the code

The copying idea was tested in 1958 by Matthew Meselson and Franklin Stahl. They grew bacteria in food containing heavy nitrogen, then switched them to ordinary nitrogen and tracked the density of their DNA over generations. After one division every DNA molecule was half heavy and half light. That showed replication is semiconservative: each new DNA molecule keeps one old strand and gains one new one. The experiment is often called the most beautiful in biology.

Worked example: using Chargaff's rules

A lab analyzes DNA from a newly discovered bacterium and finds that 30 percent of its bases are adenine. What are the others? Because A pairs with T, thymine must also be 30 percent, so A and T together make 60 percent. The remaining 40 percent is shared equally by G and C, so each is 20 percent. Now suppose a second sample shows 30 percent A but only 22 percent T. That cannot come from ordinary double-stranded DNA. Either the measurement is wrong, or the genetic material is single-stranded, as it is in some viruses. A simple rule from around 1950 thus becomes a quick check on data, which is exactly how Chargaff's ratios helped Watson and Crick rule out wrong models.