A Classification Problem Arising from Pneumonia Cases
Frederick Griffith worked in the British Ministry of Health's pathological laboratory, studying the relationship between pneumococcal types, changes in virulence, and the spread of infectious disease. His 1928 paper in the Journal of Hygiene first organized bacterial findings from pneumonia patients' sputum collected over several years, then discussed experimentally induced changes. Its starting point was not a search for DNA, but an effort to explain why the same bacterium could occur in different types and what those changes meant for observations of disease.
The capsule on a pneumococcus's surface has type-specific properties distinguishable through reactions with the appropriate antisera. Researchers also called common forms smooth or rough, abbreviated S and R, according to the appearance of their colonies. Typical encapsulated smooth forms often had greater virulence, while rough variants that had lost their capsules were generally less virulent. These classifications did not coincide perfectly, however, and the original paper recorded exceptional strains; all rough bacteria cannot mechanically be described as harmless.
Comparing Living Bacteria, Dead Bacteria, and Mixtures
Griffith compared the effects of different materials in mice. Some living encapsulated bacteria could cause infection; corresponding material thoroughly killed by heating could not by itself produce the same infection with living bacteria. Particular living rough strains used alone were also unlikely to produce the outcome associated with the original highly virulent form. The surprise was that mice could become ill when living rough bacteria were combined with heat-killed encapsulated bacteria, and living encapsulated bacteria could then be isolated from them.
Controlled comparisons gave substance to the conclusion that the mixture produced a different result. Researchers had to check that material described as dead bacteria contained no surviving organisms, while also considering whether rough forms might spontaneously regain their original traits. The paper compared numerous strains and types. The familiar four-group diagram is a compressed representation, not evidence that the original research consisted of one brief experiment or yielded uniform results.
Why a New Type Mattered More than Simple Reversion
If a rough form merely reverted to its original smooth type, the result could still be explained as the reappearance of an existing trait. A stronger observation was that rough bacteria derived from one capsular type could, after exposure to dead material of another type, develop a new type corresponding to the source of that material. For example, rough bacteria derived from Type II could acquire the capsular characteristics of Type III. The direction of transformation was specifically linked to the material added.
Griffith called this change transformation. Newly isolated bacteria could continue multiplying while retaining the corresponding type, so the phenomenon could not be understood merely as dead bacteria adhering to living ones or as mice temporarily becoming more susceptible to disease. Culture morphology, serological classification, and effects in animals supplied related observations, but each answered a different question. Crucially, together they pointed to a stable change in living bacteria, rather than temporary harm caused by a mixture of materials.
Observing Transformation Did Not Yet Identify DNA
Griffith thought that some component of dead bacteria might help rough forms reconstitute encapsulated, type-specific characteristics, and he sought explanations involving bacterial nutrition and changes in antigenic structure. He did not isolate a pure transforming substance or establish its chemical identity using nucleases or similar methods. To say that the 1928 experiments had already proved DNA to be the genetic material would therefore wrongly project later findings into his paper.
His public-health interests should likewise not be obscured by the discovery's later fame in molecular biology. The paper discussed strains isolated from the same patient at different stages and changes in the proportions of particular types among local cases. Transformation led researchers to reconsider whether bacterial types were always fixed. But demonstrating a change in the laboratory did not prove that every natural epidemic arose through the same process; explanations of events in the field still required appropriate evidence.
From an Animal Phenomenon to a Chemical Investigation
In 1931, Dawson and Sia published research on transformation in vitro, showing that the change did not depend entirely on the environment inside a mouse and could be observed in a suitable culture system. Moving the phenomenon outside the animal reduced the complexities introduced by its body and facilitated further treatment and comparison of dead bacterial material. This was an important methodological shift rather than a simple repetition of Griffith's discovery: the focus began moving from the entire infection process toward the component capable of inducing change.
In 1944, Avery, MacLeod, and McCarty at the Rockefeller Institute for Medical Research in New York linked the transforming substance to DNA through purification, chemical analysis, and evidence from enzyme action. Establishing the phenomenon in 1928, developing an in vitro system in 1931, and investigating the substance in 1944 each answered different questions. The lasting value of Griffith's experiments was to make stable changes in bacterial type a fact demanding explanation, allowing subsequent researchers to ask what transmitted that change.