The Oxford Penicillin Work, 1938–1941: From Mould Broth to the First Patients

In 1928 Alexander Fleming noticed that a stray mould on one of his culture plates at St Mary’s Hospital in London had stopped the bacteria around it from growing. He described the effect in 1929, but the active substance was never isolated, and for a decade penicillin stayed a laboratory curiosity. It became a medicine between 1938 and 1941 in a cramped pathology department in Oxford, the Sir William Dunn School of Pathology, where the Australian pathologist Howard Florey and the German-born biochemist Ernst Chain led a small team that grew the mould in improvised vessels, coaxed a fragile powder out of its broth, showed that it saved infected mice, and then gave it to the first human patients.

This page tells that story step by step: the survey of natural antibacterial substances that led the team to Fleming’s mould; the struggle to grow enough of it; the chemistry of Chain and Edward Abraham and the ingenious apparatus of Norman Heatley; the mouse experiment of 25 May 1940 and the Lancet paper that followed; and the first patients of 1941, among them the police constable Albert Alexander, whose treatment showed both what penicillin could do and how little of it there was. The later story — mass production in the United States and the antibiotic era — and the science of the mould itself are told on the other pages of this wing.

Table of Contents

  1. Lysozyme and a Survey of Natural Antibacterials
  2. Choosing Fleming’s Forgotten Mould Juice
  3. Growing the Mould: Flasks, Bedpans and Ceramic Vessels
  4. Chain, Abraham and the Problem of Purification
  5. Norman Heatley’s Apparatus and Assay
  6. The Mouse Experiment of 25 May 1940
  7. The Lancet Paper of August 1940
  8. The First Human Test and Albert Alexander
  9. Recovering Penicillin from Urine
  10. The 1941 Patients and “Further Observations on Penicillin”
  11. Key Research Papers
  12. Connections

1. Lysozyme and a Survey of Natural Antibacterials

The road to penicillin at Oxford began with a different natural substance. In the early 1920s Fleming had discovered lysozyme, an enzyme found in tears, saliva, mucus and egg white that dissolves the walls of certain bacteria. It was a striking example of the body’s own chemical defences, and Howard Florey, a physiologist turned pathologist, had studied it before he took the chair of pathology at Oxford in 1935.

Florey understood that the next questions about lysozyme — what exactly it attacked in the bacterial wall and how — were chemical ones, and he needed a biochemist. He recruited Ernst Chain, a chemistry graduate of Berlin who had left Germany in 1933 and had been working on phospholipids at Cambridge. According to Chain’s Nobel biography, between 1935 and 1939 his work at the Dunn School included snake venoms, the metabolism of tumours, methods of microanalysis and the mechanism of lysozyme action.

From one enzyme to a whole class of substances

The lysozyme work led naturally to a broader idea. Bacteria and moulds compete with one another in soil, water and decaying matter, and microbiologists had long recorded cases in which one microbe stopped another from growing. In 1938–39, according to the Australian Dictionary of Biography, Florey and Chain began a systematic investigation of the antibacterial substances made by bacteria and fungi. (The two men’s own Nobel biographies differ by a year: Florey’s gives 1938, Chain’s 1939.) The Science History Institute notes that at first they expected these substances to be enzymes, like lysozyme — a natural guess for a team that had spent years on an enzyme, and one that turned out to be wrong for penicillin.

Money for the work

Research of this kind needed equipment, chemicals and assistants that a university pathology department in the late 1930s could barely afford. Florey’s Nobel biography records that the work was supported by a grant from the Rockefeller Foundation, the American philanthropy that had already funded his travels in the United States in the 1920s. That grant paid for much of the penicillin programme in its first, uncertain years, before anyone knew whether the substance would work in a living animal.

Back to Table of Contents

2. Choosing Fleming’s Forgotten Mould Juice

Among the antibacterial substances described in the literature, Fleming’s was an obvious candidate. His 1929 paper in the British Journal of Experimental Pathology, “On the antibacterial action of cultures of a penicillium, with special reference to their use in the isolation of B. influenzae”, reported that the liquid in which the mould had grown — the filtered broth — stopped the growth of certain disease-causing bacteria while leaving others untouched; as the title says, Fleming put it to use in the laboratory for isolating a bacterium it did not affect. What he had not done was isolate the active substance.

For the Oxford team, penicillin had two attractions. It was active against exactly the kinds of bacteria that caused wound infections, blood poisoning and many deaths in hospitals, and it was a chemical puzzle: nobody knew what it was. Florey and Edward Abraham later wrote their own history of what followed, “The work on penicillin at Oxford”, published in the Journal of the History of Medicine and Allied Sciences in 1951.

A natural product from a common mould

The source was a Penicillium mould, a relative of the blue-green moulds found on old bread, citrus fruit and damp walls. The question of what species it really was, and how the mould makes penicillin, belongs to the science of the mould and is told on the page The Penicillium Mould and the Science of Penicillin. For the Oxford story what mattered was a practical fact: the mould made very little penicillin, and it made it slowly.

Back to Table of Contents

3. Growing the Mould: Flasks, Bedpans and Ceramic Vessels

The first attempts to grow the mould at Oxford were made from 1939 by Chain together with Leslie Epstein (later Leslie Falk), an American Rhodes Scholar working in the department, as the historian Ronald Bentley describes in a 2007 biographical note in the Journal of Medical Biography. They grew Penicillium notatum, as Fleming’s mould was then called, on liquid broth and tested how much antibacterial activity the broth developed.

A mould that needs air

Penicillium grows as a felt-like mat on the surface of a liquid; it needs air, and in the 1930s nobody knew how to grow it any other way. That meant surface culture: shallow layers of broth in as many flat vessels as possible, each topped by a mat of mould, each yielding a small volume of weak penicillin solution after a week or more. The Science History Institute’s account traces how the team moved from ordinary laboratory flasks to anything shallow, sterile and stackable — including hospital bedpans.

The ceramic culture vessels

Bedpans were not enough. The team had hundreds of specially made ceramic culture vessels produced, flat rectangular containers designed to hold a shallow layer of broth and to be stacked, filled, drained and refilled. The Australian Dictionary of Biography speaks of the “‘Heath Robinson’ contraptions” of the Oxford work — the British phrase for elaborate improvised machinery — and the description fits the whole of the early production line, which was built in wartime from whatever could be found or made.

The “penicillin girls”

To tend the vessels the department took on young women as technical assistants, remembered in the histories as the “penicillin girls”. According to the Science History Institute there were six of them. They inoculated the broth with mould spores, watched the cultures, drew off the broth when the mould had made its penicillin and refilled the vessels, in what amounted to a small factory housed inside a university pathology department.

Back to Table of Contents

4. Chain, Abraham and the Problem of Purification

Growing the mould was only half the problem. The broth contained a tiny amount of penicillin mixed with salts, sugars and the mould’s other products, and penicillin itself turned out to be unstable: it lost its activity when it was heated, when the liquid was too acid or too alkaline, and simply with time. The active substance in Fleming’s broth had never been isolated.

A biochemist’s approach

Chain and the young chemist Edward Abraham worked out how to concentrate and purify penicillin without destroying it. The Science History Institute summarises their method in three parts: careful control of the pH (the acidity of the solution), keeping the material cold, and repeated evaporation under conditions that amounted in effect to freeze-drying, so that water could be removed without heating the penicillin. The result was a powder that could be stored and tested — still far from pure, but concentrated enough for animal experiments.

Not an enzyme after all

The purification work also settled an early assumption. The team had expected the antibacterial substances in their survey to be enzymes like lysozyme. Penicillin proved not to be one. Its chemical structure was a separate story, worked out only after the war; it is told on the page about the science of penicillin.

An enzyme that destroys penicillin

In December 1940 Abraham and Chain reported in Nature that a strain of the gut bacterium Escherichia coli produced an enzyme able to destroy penicillin — later called penicillinase. The note was short, but it showed that some bacteria already possessed a defence against the new substance, a finding whose importance grew with every later decade of antibiotic use.

Back to Table of Contents

5. Norman Heatley’s Apparatus and Assay

Norman Heatley was the team member whose name the public story of penicillin long left out. Born in 1911 in Woodbridge, Suffolk, and educated at Tonbridge School and St John’s College, Cambridge, he took a Cambridge doctorate in biochemistry and came to the Dunn School in 1936. Carol Moberg’s 1991 article in Science called him “penicillin’s forgotten man”, and the Oxfordshire Blue Plaques scheme, which commemorates him in the city, credits him with devising the methods by which penicillin was first extracted and produced in quantity.

The extraction machine

Heatley’s gift was for practical apparatus built out of almost nothing. According to the Oxfordshire Blue Plaques account and the Science History Institute, he built a countercurrent extraction machine and a multi-stage process from improvised vessels and tubing, which drew the weak broth from the culture vessels and passed it against a flowing solvent so that the penicillin moved out of the broth and into the solvent. Historians widely credit him with the method sometimes called back-extraction: moving penicillin from the watery broth into an organic solvent under acid conditions, then back into water when the acidity was changed, leaving many impurities behind at each step. Lobanovska and Pilla, in their 2017 review, describe Heatley as designing extraction methods together with Chain. It turned the output of the culture vessels into a concentrated solution that Chain and Abraham could dry.

Measuring something no one could weigh

A substance that cannot yet be purified cannot be weighed, so the team needed a way of measuring how much penicillin was present in any given sample. Heatley devised one. His method, published in the Biochemical Journal in 1944 as “A method for the assay of penicillin”, became known as the cylinder-plate assay: small cylinders were placed on an agar plate seeded with bacteria and filled with the test solution, and the size of the clear zone where the bacteria failed to grow showed how strong the solution was. Measured against a standard, it gave the team a way of comparing batches and doses in “units” of activity before penicillin could be measured by weight.

Recognition

Heatley received no share of the 1945 Nobel Prize. He was appointed OBE in 1978, and in 1990 Oxford awarded him an honorary Doctor of Medicine; web summaries describe this as the first honorary DM the university had given to a non-medical scientist in its long history, a claim this page reports as theirs. Heatley himself wrote one of the memorial accounts of the work, “In memoriam, H.W. Florey: an episode”, published in the Journal of General Microbiology in 1970.

Back to Table of Contents

6. The Mouse Experiment of 25 May 1940

By the spring of 1940 the team had enough of the brown powder to ask the decisive question: was it safe, and would it work inside a living body? Many substances that kill bacteria in a dish are poisonous to animals, or are destroyed by blood and tissue before they can act.

A first safety test

The Science History Institute records that in March 1940 two mice were injected with the Chain–Abraham extract and tolerated it without apparent harm. That was the first evidence that the substance was not a poison to animals at the amounts the team could make, and it cleared the way for a test of whether it could cure an infection.

Eight mice

The protection experiment was carried out on Saturday 25 May 1940. Eight mice were infected with a lethal dose of virulent streptococci — the kind of bacteria responsible for many fatal wound and blood infections at the time. Four were then given penicillin; four were left untreated as controls. According to the Science History Institute, sixteen and a half hours later the four untreated mice were dead and the treated mice were alive; Lobanovska and Pilla describe all the control mice as dying within 17 hours. Their account also notes that some of the treated mice received a single dose and others repeated doses; the published accounts vary in their details, and this page gives no dose figures.

The experiment was small and the result was stark. It showed for the first time that the substance in Fleming’s mould broth, concentrated and partly purified, could protect a living animal from an otherwise fatal bacterial infection.

A date sometimes misprinted

One modern review prints the date of the experiment as 25 May 1939. Every other source used for this page, and the timing of the Lancet report a few months later, place it in 1940.

Back to Table of Contents

7. The Lancet Paper of August 1940

Through the summer of 1940 the team repeated and extended the experiment on hundreds of mice, while the war came closer: France fell in June, and the Battle of Britain was fought overhead in July and August. The results were published in the Lancet on 24 August 1940 under the title “Penicillin as a chemotherapeutic agent”.

Seven authors

The paper was signed by Ernst Chain, Howard Florey, A. D. Gardner, Norman Heatley, Margaret Jennings, J. Orr-Ewing and A. G. Sanders — chemists, a bacteriologist and experimental pathologists, reflecting how many hands the work required. Margaret Jennings, a long-time colleague in Florey’s laboratory, became his second wife in 1967.

What it reported

The paper was short — three pages in the journal — and reported that penicillin protected infected mice. It made the case that penicillin was a chemotherapeutic agent: a substance that could be given to a living body to attack bacteria inside it, as the synthetic sulfonamide drugs of the 1930s did, but drawn from a natural source.

The problem of quantity

The paper also made the next difficulty plain. A mouse weighs about twenty grams; a person weighs several thousand times as much. Treating a human being would need far more penicillin than the team had ever made, and the Dunn School’s production line — the ceramic vessels, the penicillin girls and Heatley’s extraction machine — was pushed to its limit through the autumn and winter of 1940 to build up a supply.

Back to Table of Contents

8. The First Human Test and Albert Alexander

A volunteer, January 1941

The first human to receive penicillin was not treated for an infection. In January 1941, according to the Science History Institute and the Oxfordshire Blue Plaques account of the first treatment, a woman who was dying of cancer volunteered for a test of whether penicillin was safe to inject into a person. She developed fever and shivering. Edward Abraham traced the reaction to impurities in the preparation rather than to penicillin itself, and the team improved its purification before treating anyone with an infection.

The first patient treated for an infection

On 12 February 1941 penicillin was given at the Radcliffe Infirmary in Oxford to Albert Alexander, a reserve police constable in his forties, who was gravely ill with a severe bacterial infection. The injections were given by the Radcliffe physician Charles Fletcher, who later wrote his own account, “First clinical use of penicillin”, in the British Medical Journal in 1984.

Two accounts of how he was hurt

How Alexander came by his infection is told two ways. The account found in most popular histories is that it began with a scratch from a rose thorn in his garden. Newbury Town Council, which unveiled a blue heritage plaque to him, gives a different account: that he was injured in an air raid while serving on duty in Southampton during the Blitz. The sources do not settle the question, and this page reports both accounts as accounts.

Improvement

Within days of the first injections Alexander improved markedly; the Oxfordshire Blue Plaques account and the historical reviews describe him beginning to recover from an infection that had seemed certain to kill him. It was the first time anyone had seen penicillin act on a human infection, and it confirmed in a patient what the mice had shown the year before.

Back to Table of Contents

9. Recovering Penicillin from Urine

The difficulty was quantity. The whole of the Dunn School’s painstakingly gathered supply was not enough to finish Alexander’s treatment. The team knew that much of an injected dose of penicillin passes out of the body unchanged in the urine, so they did something that captures the improvised character of the whole project: they collected the patient’s urine and extracted the penicillin from it to give it again.

The Oxfordshire Blue Plaques account and the Science History Institute record that Charles Fletcher carried the urine by bicycle from the Radcliffe Infirmary to the Dunn School, where the penicillin was recovered and purified, and that it was then re-injected. The recovery stretched the supply but could not replace what was lost, and in the end the penicillin ran out.

Relapse and death

Without further treatment Alexander’s infection returned. He relapsed and died on 15 March 1941. His case showed two things at once: penicillin worked in a human being, and the amounts that could be made in a university laboratory were far too small. The Science History Institute records that Florey resolved thereafter always to have enough penicillin in hand to complete a treatment before starting one, and the next patients were chosen with that in mind.

Back to Table of Contents

10. The 1941 Patients and “Further Observations on Penicillin”

After Alexander the team treated further patients at the Radcliffe Infirmary, choosing cases that the limited supply could see through to the end. According to the Oxfordshire Blue Plaques account of the first treatment, six patients were treated in that first series, and four made a complete recovery.

The August 1941 report

The results were published in the Lancet on 16 August 1941 as “Further observations on penicillin”, signed by Edward Abraham, Ernst Chain, Charles Fletcher, A. D. Gardner, Norman Heatley, Margaret Jennings and Howard Florey. It was a long paper, running to thirteen pages, and it brought together the team’s further work on penicillin with the first human cases. Where the 1940 paper had shown that penicillin protected mice, the 1941 paper showed that it could be given to people and could overcome serious bacterial infections in them.

What the first cases established

Taken together, the Oxford reports of 1940 and 1941 established that penicillin was effective against the staphylococci and streptococci that caused so many deadly infections, that the early reaction in the first volunteer came from impurities, and that much of an injected dose left the body in the urine. They also established the central obstacle: a single patient needed more penicillin than a laboratory could easily make in weeks. Florey’s wife, the physician Mary Ethel Florey, went on to conduct clinical trials of penicillin at the Radcliffe Infirmary, in military hospitals and at the Birmingham Accident Hospital, and later wrote a four-volume work on the clinical use of antibiotics.

Where the story goes next

By the summer of 1941 Britain, under bombing and stretched by the war, could not take on large-scale production. How Florey and Heatley carried the work to the United States, where deep-tank fermentation made penicillin in industrial quantities, is told on the page Wartime Mass Production and the Antibiotic Era. The lives of the two men who led the Oxford team are told on Howard Florey and Ernst Chain: Lives and Careers.

Back to Table of Contents

Key Research Papers

  1. Chain E, Florey HW, Gardner AD, Heatley NG, Jennings MA, Orr-Ewing J, Sanders AG. Penicillin as a chemotherapeutic agent. The Lancet. 1940;236(6104):226-228. DOI: 10.1016/S0140-6736(01)08728-1
  2. Abraham EP, Chain E, Fletcher CM, Gardner AD, Heatley NG, Jennings MA, Florey HW. Further observations on penicillin. The Lancet. 1941;238(6155):177-189. DOI: 10.1016/S0140-6736(00)72122-2
  3. Florey HW, Abraham EP. The work on penicillin at Oxford. J Hist Med Allied Sci. 1951;6(3):302-17. PubMed PMID: 14873930
  4. Fleming A. On the antibacterial action of cultures of a penicillium, with special reference to their use in the isolation of B. influenzae. 1929. Bull World Health Organ. 2001;79(8):780-90. (Reprint of the 1929 paper.) PubMed PMID: 11545337
  5. Bentley R. Leslie A (Epstein) Falk (1915-2004) and penicillin production at Oxford. J Med Biogr. 2007;15(2):93-6. PubMed PMID: 17551607
  6. Abraham EP, Chain E. An enzyme from bacteria able to destroy penicillin. Nature. 1940;146(3713):837. DOI: 10.1038/146837a0
  7. Heatley NG. A method for the assay of penicillin. Biochem J. 1944;38(1):61-5. PubMed PMID: 16747749
  8. Heatley NG. In memoriam, H.W. Florey: an episode. J Gen Microbiol. 1970;61(3):289-99. PubMed PMID: 4922763
  9. Moberg CL. Penicillin's forgotten man: Norman Heatley. Science. 1991;253(5021):734-5. PubMed PMID: 1876832
  10. Hamilton-Miller JM. Appreciation. Dr. Norman Heatley. J Antimicrob Chemother. 2004;53(5):691-2. PubMed PMID: 15154415
  11. Fletcher C. First clinical use of penicillin. Br Med J (Clin Res Ed). 1984;289(6460):1721-3. PubMed PMID: 6440620
  12. Lobanovska M, Pilla G. Penicillin's Discovery and Antibiotic Resistance: Lessons for the Future? Yale J Biol Med. 2017;90(1):135-145. PubMed PMID: 28356901

PubMed Topic Searches

  1. Penicillin history at Oxford
  2. Papers by H. W. Florey
  3. Papers by E. B. Chain on penicillin
  4. Papers by N. G. Heatley

Further Reading

Back to Table of Contents

Connections

Back to Table of Contents