Home > About the Academy > Biographical memoirs

BIOGRAPHICAL MEMOIRS

Frank Macfarlane Burnet 1899-1985

Introduction Early life University education Scientific career The Walter and Eliza Hall Institute, 1924 The Lister Institute, London, 1925-1927 Bacteriologist at the Walter and Eliza Hall Institute, 1928-1931 National Institute of Medical Research, London, 1932-1933 Assistant Director, the Walter and Eliza Hall Institute, 1934-1943 Director of the Walter and Eliza Hall Institute, 1944-1965 University of Melbourne, 1966-1977 Retirement, 1978-1985 Pattern of work Daily and weekly routine Intellectual processes Scientific work Bacteriophages Staphylococcal toxin Animal virology Immunology Cancer Public health Human biology Ageing Books Academy activities Public policy Honours and awards Decorations Membership of learned academies and professional bodies Honorary degrees International lectureships Honorific lectures in Australia Other marks of recognition Acknowledgements Sources of information Notes

By Frank Fenner

This memoir was originally published in Historical Records of Australian Science, vol.7, no.1, 1987.

Numbers in brackets refer to the notes at the end of the text.

Introduction

With the death of Frank Macfarlane Burnet on 31 August 1985, Australia lost its greatest biologist, a man who had spent virtually all of a long working life in Australia. His experimental work on bacteriophages and animal viruses, especially influenza virus, resulted in major discoveries concerning their nature and replication, and he was a pioneer in the application of ecological principles to viral diseases. He proposed two concepts in immunology – acquired immunological tolerance and the clonal selection theory of antibody production – which proved to be of critical importance in stimulating research and led to a more complete understanding of immune processes. In the later stages of his life he lectured and wrote extensively about problems of human biology and human affairs, ageing and cancer. He was a Foundation Fellow and, from 1965 to 1969, President of the Australian Academy of Science.

Early life

Burnet was born in Traralgon, in eastern Victoria, on 3 September 1899. His father, Frank Burnet, was born in 1856 in Langholm, Scotland, and emigrated to Australia as a young man; his paternal grandfather was an architect and factor to the Duke of Buccleuch in Dumfriesshire. His mother, née Hadassah Pollock Mackay, was born in Koroit, in Victoria, in 1872. She also came of Scottish middle-class stock, her father being a Glasgow schoolteacher who had emigrated to Australia in the late 1850s and settled in Koroit.

Macfarlane Burnet, who from childhood and throughout his life was known as 'Mac' to his close friends, was the second of seven children. At the time of his birth, his father was manager of the Traralgon branch of the Colonial Bank. In 1909, he was transferred and moved with the family to Terang, in western Victoria. In both places young Burnet went to the local primary school. As related in his autobiography, Burnet retained vivid memories of his life as a boy in Terang, where he returned for vacations until he was in his twenties. He was a shy boy, but revelled in the opportunities to wander in the nearby countryside, especially near Lake Terang, where he was greatly interested in the variety of wild life to be seen. He became a member of the Boy Scouts in 1910, soon after the movement was founded in Victoria, and enjoyed the associated camping and outdoor activities.

The first evidence of a serious interest in biology began in Terang, where young Burnet became an enthusiastic collector of beetles – an interest which he retained all his life. There were no books on biology in his home, and no ready access to them in Terang, but he read all the biological sections of an old Chambers Encyclopaedia (published in the 1860s), which introduced him to Charles Darwin. His parents bought him Harmsworth's Natural History, which appeared as a fortnightly periodical. He wrote to Melbourne for a book about beetles, and was sent an English translation of Fabre's Souvenirs Entomologique. Later he acquired Froggatt's (1907) Australian Insects, his copy of which shows his intense interest in the Coleoptera – these pages are covered with entries concerning his collecting and his own very creditable drawings of some of the beetles he had found. This interest in beetles led the local Presbyterian minister, the Rev. Samuel Fraser, to note that he was a bright boy and to suggest to his parents that he should have a university education. Always a person to have a realistic view of his own qualities and deficiencies, Burnet notes in his autobiography that his attributes in childhood and adolescence fitted well with the picture drawn by Roe (1965) for a group of eminent research scientists working in America: 'Most...were rather shy, socially late-maturing boys with strong hobbies and noticeable persistence in them. [They] were voracious if unselective readers throughout their childhood. Most regarded their fathers with great respect but felt somewhat distant from them'.

University education

Having completed his primary school education in Terang, Burnet was sent to Geelong College for four years – an experience that he did not greatly enjoy. In his final year he gained scholarships enabling him to proceed to the university, the most important being a residential scholarship at Ormond College, in the University of Melbourne. Choice of a course was not so much because of a desire to be a doctor as a choice of the only kind of professional life that had much of an appeal of the three suggested – Medicine, Law or the Church. His early years at the University were accompanied by the usual wide reading and broadening of horizons, and a sorting out of his ideas on religion, during which he moved from the traditional social pattern in which he had grown up – Sunday school and later church every Sunday – to become consciously agnostic. Charles Darwin was his hero, whose writings exerted a profound influence on his scientific work, and H.G. Wells was an important influence on his views about science and society.

At the end of a medical course that was shortened to five years because of the First World War and the perceived need when the course began to produce medical graduates quickly, Burnet graduated MB,BS in April 1922, coming second in a class that contained four other persons who later achieved fame in science and medicine as Sir Roy Cameron, Professor R.A. Willis, Dame Jean Macnamara and Dame Kate Campbell. After graduation, Burnet proceeded immediately to prepare for the degree of MD by examination, which he gained late in 1924. It was then usual to spend one year as a resident medical officer, so as to gain experience in casualty and medical and surgical wards before going into practice. In the surgical wards he came to know two eminent surgeons, each of whom later served as a chairman of the Board of The Walter and Eliza Hall Institute when he was director: Sir Alan Newton and Sir Victor Hurley. However, his greatest satisfaction at this time was to serve as house physician to Melbourne's leading physician at the time, Dr R.R. (later Sir Richard) Stawell, a neurologist. This experience firmly convinced Burnet that his future career lay in clinical neurology, and he applied for the post of medical registrar as a stepping stone for such a career. However, the medical superintendent of the Melbourne Hospital, who was responsible for making such appointments, judged (correctly) that Burnet's character and personality were more compatible with a career associated with the laboratory than with clinical work, and instead he was appointed pathology registrar, and a few months later, senior resident pathologist.

Scientific career

The Walter and Eliza Hall Institute, 1924

At that time the pathology laboratories of the Melbourne Hospital were operated as part (then the larger part) of The Walter and Eliza Hall Institute, which had been established in 1915. As a medical resident, Burnet had been interested in the attempts of Dr N.H. Fairley (later Sir Neil Hamilton Fairley), then a member of the Institute staff, to treat cases of typhoid fever by intravenous injections of typhoid vaccine – an interest that led to Burnet's first scientific papers and subsequently to his interest in bacteriophages.

In 1924 the Institute was transformed with the arrival from University College, London, of Dr Charles Kellaway (later to become Sir Charles Kellaway (1)) to become the second director of the Institute. Kellaway was not content with a predominantly service role for the Institute and proceeded to establish research activities in physiology, biochemistry and bacteriology.

The Lister Institute, London, 1925-1927

Kellaway saw Burnet as the potential leader of the small bacteriology section, but decided that he should first have overseas training, and Burnet left for England as a ship's surgeon in June 1925. He took a position at the Lister Institute because there was a paid position available there as an assistant to the curator of the National Collection of Type Cultures, which allowed him about two-thirds of his time for research. A few months later he obtained a Beit Fellowship Award and was able to devote himself full-time to research on bacteriophages. Under the supervision of Professor J.G. Ledingham, he gained a PhD degree of the University of London (1928). A measure of the respect his work had already gained is provided by the fact that he was invited to write the chaper on bateriophages for the Medical Research Council's System of Bacteriology. A copy of d'Hérelle's (2) expanded work, Le Bacteriophage, purchased by Burnet in Paris in July 1927, reveals how carefully he read the book and picked up aspects which prompted additional experimental work. While working in London he became engaged to a fellow Australian then resident there, Edith Linda Marston Druce, whom he married on 10 July 1928, after his return to Australia.

Bacteriologist at The Walter and Eliza Hall Institute 1928-1931

Shortly after his return to Australia in 1928, an event called the 'Bundaberg disaster' occurred, in which several children died after receiving inoculations of diphtheria toxin-antitoxin. Kellaway headed the Royal Commission appointed to investigate the tragedy (3) and Burnet carried out the bacteriological investigations, leading to important studies on staphylococcal toxins. At the same time he continued studies on bacteriophages, producing some papers later regarded as classics.

National Institute of Medical Research, London, 1932-l933

In November 1931 Burnet received an offer that changed the course of his scientific life. Sir Henry Dale, Director of the National Institute of Medical Research at Hampstead, had received a generous offer from the Rockefeller Foundation to expand the excellent work on animal virology then in progress at Hampstead, and after consultation with Kellaway, he invited Burnet to participate in this work. This was a period of great activity, for with people like Sir Patrick Laidlaw, Wilson Smith, C.H. Andrewes, W.I. Elford and J.E. Barnard, the Hampstead laboratories were world leaders in research on animal viruses. The excitement caused by Laidlaw's comment 'The ferrets are sneezing' remained with Burnet all his life; it may even have influenced his later decision to concentrate on influenza virus. During this period Burnet developed his work on the use of the chick embryo for the isolation and assay of animal viruses. He also acquired a powerful friend in Sir Henry Dale, who offered him a permanent position at the National Institute. However, Burnet decided to return to Melbourne, where he became Assistant Director of The Walter and Eliza Hall Institute, in charge of the virus section.

Assistant Director, The Walter and Eliza Hall Institute, 1934-1943

Back in Melbourne, Burnet rounded off his work on bacteriophages and continued actively to study the behaviour of a variety of viruses in the developing chick embryo. Seizing opportunities as they arose, he worked on psittacosis, an experience that influenced his thinking in his first book, Biological Aspects of Infectious Disease, recognized a rickettsia to be the cause of Q fever, and carried out studies on poliovirus. However, his major interest after 1939 was influenza virus, prompted by the discovery of methods of growing the virus in the amniotic and allantoic cavities of the chick embryo. With the onset of the Second World War, his attention was focused on methods of immunizing against influenza, in case there should be another epidemic like that of 1918-19. In 1942 he was elected FRS, and in 1944 made his first trip to America, where he delivered the Dunham Lectures at Harvard University and received an attractive offer for a chair at Harvard. This tempted him greatly, and was refused only after much soul-searching, principally out of a feeling of loyalty to Australian science and especially to the Hall Institute.

Director of The Walter and Eliza Hall Institute, 1944-1965

In 1943 Kellaway was appointed Director of the Wellcome Foundation in London; Burnet was appointed Director of The Walter and Eliza Hall Institute in 1944. He had been greatly impressed with what he saw of medical research in the United States in 1944, especially at Harvard University, and set out to achieve something of this pattern in Melbourne. He decided that the future activities of the Institute should be concentrated on animal virology, especially influenza virus, and of those already in the Institute (apart from the Clinical Research Unit), only Gottschalk, a biochemist, continued to work on any other topic. When the enzymic nature of influenza virus action on red blood cells became apparent, Gottschalk also joined the team and unravelled the nature of the viral enzyme (neuraminidase).

Although he personally evinced no desire to become involved in experiments using biochemical and biophysical techniques, Burnet recognized that such an approach was essential if the Institute were to contribute to a comprehensive study of animal viruses. In 1946 he sought and obtained from the Commonwealth Government a special grant of £20,000 (then a very considerable sum) to establish a group equipped to carry out biophysical research on viruses, including electrophoresis, ultracentrifugation, and later electron microscopic studies. For the next decade, the Institute was a Mecca for overseas scientists who came to work on influenza virus under Burnet's guidance.

From 1951 to 1956, Burnet himself concentrated on studies of the genetics of influenza virus. His demonstration of high frequency recombination was received with great scepticism by scientists overseas, since it did not accord with what was found with bacteriophages and therefore with conventional wisdom. The soundness of Burnet's experimental work in this field became apparent when it was demonstrated several years later that influenza virus had a segmented genome (4).

Although he was an expert and assiduous experimentalist, Burnet also found time to write books summarizing his views on animal virology and with W.M. Stanley acted as co-editor of a major compendium on virology.

In parallel with his work on virology, Burnet had always been interested in the immune response, and in 1941 he had produced a monograph analysing the nature of antibody production. In 1948 he re-examined this topic and propounded a new hypothesis on antibody production based on analogies with adaptive enzymes. More important, however, was his enunciation in this book of the hypothesis of acquired immunological tolerance.

Honours, both scientific and civil, began to come his way. In 1947 he received a Royal Medal and in 1959 the Copley Medal of The Royal Society; he was knighted in 1951. In 1958 he was awarded the Order of Merit and in 1960 the Nobel Prize in Physiology or Medicine.

Although never a keen committee man, as Director of the Institute Burnet accepted an increasing number of national and international obligations. Apart from Board meetings and membership of committees of the National Health and Medical Research Council that were an obligation of his position, he served as a member of the Defence Research and Development Policy Committee (1947-52), as Chairman of the Radiation Advisory Committee (1955-59), and as Chairman of the Queen Elizabeth II Fellowship Committee (1963-69). As Chairman of the Papua New Guinea Medical Research Committee (1962-69) he played a major part in the establishment of the Papua New Guinea Institute of Human Biology – a name that he preferred to 'Medical Research' in that it emphasized the importance of demography and population growth in the future of that country.

Internationally, Burnet had the unusual distinction of serving as President of both the International Association of Microbiological Societies (1953-57), and the Third International Congress of Immunology (1977). He served on several committees of the World Health Organization, including the WHO (Global) Medical Research Advisory Committee (1959-63) and on retirement undertook the task of acting as foundation chairman of the Commonwealth Foundation in London (1966-69).

In 1957, at an age when most scientists are thinking of contracting their bench work, Burnet made a revolutionary change in the direction of his own work and that of the Institute. He decided that henceforth he (and all staff in the Institute) would abandon virology and concentrate instead on immunology. The reasons for this decision were complex. He saw that virology would in future demand the use of tissue culture rather than the developing chick embryo, and that it would become more and more 'molecular', and he was loth to undertake either transition. Further, as Lederberg noted (personal communication, 1986), Burnet was at the time 'remarkably uninformed with respect to modern views on the mechanism of protein synthesis, DNA coding, etc.' Also, his long time interest in the theory of antibody production had been stimulated by a paper by Jerne (5) that proposed a 'selective' model for the process, rather than the currently fashionable instructive theory. A few years later, this decision was vindicated by the award of the Nobel Prize not for virology (for which the award would certainly have been merited), but for an immunological discovery, acquired immunological tolerance. By then, Burnet had gone beyond immunological tolerance to formulate what he himself regarded to be his major contribution to science, the clonal selection theory of antibody production.

Burnet's increased prestige and international fame led to a change in his work pattern, so that he had less and less time to spend at the bench. Nevertheless he continued to produce papers on experimental immunology – on graft-versus-host reactions, as described by Simonsen (he was deighted to find that he could use the chorioallantoic membrane to study immunological phenomena), and on autoimmune diseases, using NZB mice as a model.

Burnet had always kept the staff of The Walter and Eliza Hall Institute small, partly, no doubt, so as to maximise his opportunities for research at the bench. But by 1962 he saw that his successor, whoever he was to be, would require more space, and he devoted considerable effort to obtaining money for two more floors, which were completed in 1966 and named the 'Nuffield-Burnet Laboratories' by his successor. In 1965 he retired from directorship of the Institute, and Dr G.J.V. Nossal (now Sir Gustav Nossal) was appointed as Director. To mark the occasion, the Ciba Foundation organized a symposium on 'The Thymus' (6) in Melbourne, and the Governor-General of Australia, Lord Casey, attended his Farewell in the University of Melbourne.

University of Melbourne, 1966-1977

For Burnet, as for most scientists, retirement from an official position did not mean the end of active work. Professor S.D. Rubbo, who had just moved into the newly built School of Microbiology in The University of Melbourne, across the road from the Hall Institute, offered Burnet rooms and organized the provision of a secretary, and Burnet began a new career as a writer and elder statesman of science in Australia. At this time (1965) he accepted the presidency of the Australian Academy of Science, which he had declined eight years earlier because of his wish to devote himself primarily to his work as Director of the Hall Institute. During the twelve years that he was at The University of Melbourne, Burnet produced thirteen books, initially on immunology and subsequently on human biology, ageing and cancer, as well as a fourth edition of his first book.

He continued to receive honours, both scientific and civil. In 1969 and again in 1974 international symposia were organized by Nossal to celebrate his 70th and 75th birthdays. He received a KBE in 1969 and Australia's highest award, Knight of Australia (AK), in 1978. However, in 1973 he suffered a grievous loss when his wife Linda died of lymphoid leukaemia. For a time he went to live again in Ormond College, University of Melbourne, where he had lived as a medical student, and renewed his friendship with the Master, Dr Davis McCaughey, who was later to be appointed Governor of Victoria. In 1976 he married again, to Hazel Jenkin, a widow who had endowed the library in the School of Microbiology to commemorate her only daughter, who had died while still a graduate student.

Retirement, 1978-1985

In 1978 Burnet decided, at the age of 78, that the time had come to slow down somewhat. He left the School of Microbiology and moved to his home, where he produced two more books, and continued to maintain an extensive correspondence and to write articles on general problems such as the future of Australia. In November 1984 he was operated on for cancer on the rectum and appeared to have made a good recovery, but secondary lesions were discovered early in August 1985 and he died on August 31 at his son's home at Port Fairy, near where he had spent his boyhood. He was given a State funeral by the Government of Australia, and was buried at Tower Hill cemetery, near Port Fairy. He was survived by his second wife Hazel, his son Ian, his daughters Elizabeth and Deborah, and eight grandchildren.

Pattern of work

Daily and weekly routine

Before embarking upon an analysis of Burnet's scientific work it may be useful to outline the pattern of his activities during the period 1945-55. After this, his increasing fame led to many other calls on his time and increased absences overseas, which disrupted this pattern somewhat, but when at the Institute Burnet always devoted a substantial part of each day to work in the laboratory. Throughout his life at the bench, he worked alone, except for one or sometimes two graduate assistants and one or two technicians. In consequence, many of his papers on experimental research show Burnet as the sole author and few list a co-author other than his current graduate assistant. He was careful in the selection of his graduate assistants, and had a succession of highly competent and devoted women help him in this capacity: Margo McKie (1928-34), Mavis Freeman (1928-40), Dora Lush (1934-39), Diana Bull (1941-43), Joyce Stone (1940-50), Patricia Lind (1944-65), Margaret Edney (1948-56), Margaret Gilpin (1948-49; 1949-52), Margaret Holmes (1958-65), Deborah Burnet (1960-62; 1963-64), and Susi Ernyei (1962-64).

Burnet's abiding passion was his scientific work. As Director of the Institute, he decided policy, usually after consultation with the Deputy Director, Dr I.J. (later Sir Ian) Wood, and often after discussions with Dr E.V. Keogh, the éminence grise of medical research in Victoria in the 1950s. However, he always took absolute responsibility for all appointments of research staff, graduate students and overseas visitors, in accordance with his policy of ensuring that the Institute should be an elite institution of world standard, small enough to be effectively controlled by one man, himself. He delegated the implementation of policy to the Manager of the Institute, Mr Arthur Hughes, and the Personnel Manager, initially Miss Fanny Williams and after her retirement, Dr Margaret Holmes.

Burnet was very proud of being an Australian, and was determined to show that science of first class quality could be carried out in Australia by Australians. The majority of his research papers were published in Australian journals, notably the Australian Journal of Experimental Biology and Medical Science, and for papers with a medical flavour, The Medical Journal of Australia. It was very fitting, and a source of considerable pride, that he was selected as 'Australian of the Year' in 1961.

Intellectual processes

This account of his daily work shows that Burnet was a dedicated and hard-working scientist. Hundreds of other scientists share these traits – what made Burnet so outstanding? Nossal (7) and Cohn (8) have analysed this question, and some of the answers will emerge from the description of Burnet's scientific work which follows. But it may be useful to attempt a summary here.

Although perhaps better known as a theoretical biologist, Burnet was a first-class experimental scientist, who until well into his sixties spent the greater part of each day working at the laboratory bench. His name never appeared on an experimental paper unless he had participated substantially in the benchwork himself. This involvement in benchwork meant that he was able to notice the unexpected result that might otherwise be dismissed as a technical mistake, and follow it up. His own experiments never made use of apparatus more complex than a microscope, for like many medically trained laboratory workers of that era, he overestimated the difficulties inherent in the use of biochemical and biophysical equipment. He rarely used statistical analysis for the evaluation of his experimental results; they had to be capable of unequivocal interpretation without it. And he found that benchwork was excellent 'occupational therapy', that allowed his mind to wander and wonder while his hands were occupied with pipettes and eggs.

In his experimental work, Burnet was a reductionist; he designed experiments to demonstrate or disprove the 'minute particulars' of his current hypothesis. However, in discussions of his own work, and even more that of his associates, he was quick to relate any new finding to biology as a whole in a most perceptive way. As he says of himself, Burnet was an ecologist, and his capacity to integrate discoveries made in diverse fields of science, which is the hallmark of the ecologist, was one of his great strengths.

A remarkable feature of Burnet's career was that although he worked as a virologist until the age of 57, some 90% of his experimental papers being on virology, the two contributions to science for which he became most renowned were in the field of immunology, on aspects in which he had done little or no experimental work. Such breadth and depth of understanding, and such self-assurance as to allow him to challenge established dogma in a field not his own, is rare in the present era of scientific specialization.

In spite of the fact that he never gave a regular course of lectures to undergraduate or graduate students, Burnet was a great teacher. He had an unforgettable impact on the thinking of the stream of scientists who came to the Hall Institute from Australia and overseas, especially between 1944 and 1965 (Walter and Eliza Hall Institute of Medical Research Annual Review 1978-79). Even at this stage of his life, he was shy and withdrawn, and reacted most effectively with his colleagues when he discussed a paper that they wished to submit for publication. But all who worked in the Institute had no doubt that they were privileged to be working with a man of genius. He influenced an even wider audience though his books, the majority of which were not technical monographs, but were written in 'Scientific American' style, for the physician or biologist who was not a specialist in virology, or immunology, or gerontology.

Burnet had knowledge and intelligence in abundance. He was uncommonly broad in his interests and reading and had an excellent memory. But the great and rare qualities to which his knowledge and intelligence were harnessed were originality and creativity. Burnet had a remarkable intuitive grasp of certain fundamental biological concepts, especially Darwinian evolution. He had courage, optimism and the self-assurance and confidence in his own judgement that allowed him to address questions of fundamental importance in spite of his relative isolation in Australia. Indeed he thought that his isolation was an advantage, since it protected scientists from being too much influenced by fashions in scientific thinking. And he was a lateral thinker with an unparalleled capacity to link apparently unconnected observations. This led him to devote as much mental energy into interpreting the world literature as most people put to interpreting their own work. However, he did not have much interest in other people's theories, except in so far as they helped him to remould his own.

Of course, he had weaknesses. He was very reluctant to accept the 'ultimate reductionism' of DNA, and in both articles and books castigated molecular biology as being potentially dangerous, and unlikely to make a contribution to human health commensurate with the funds and talent that were devoted to it. Even this much-criticized shortcoming had its logic. His comment referred to medical science, not biology in general, and he argued that little could ever be done to prevent afflictions due to genetic errors (germline or somatic), and that little research in molecular biology was needed to control or prevent the diseases due to environmental influences. The major problem, he thought, was to ensure the proper distribution of known methods of preventive and curative medicine, which applied almost exclusively to extrinsic diseases, to all of the world's people.

If one had to nominate 'keywords' to describe Burnet's greatness as a biological scientist, they might include – originality, creativity, biological intuition, high intelligence, discipline, persistence, excellent memory, capacity for lateral thinking, ability to write rapidly and clearly, and self-confidence.

Scientific work

Burnet's first scientific paper was published in 1924 and his last in 1983; his first monograph appeared in 1936 and his thirty-first and last book in 1979. For over two-thirds of the long period during which he was writing, he spent well over half of each working day, on average, at the bench. His work covered a wider range of subjects in biomedical research than that of most scientists, hence it is convenient to arrange it by major topics, in roughly chronological order. There are, of course, some overlaps, as Burnet responded to urgent biomedical problems that occurred when he was involved in other studies, e.g., the Bundaberg disaster in 1928 and the poliomyelitis epidemic in 1937; and, as Director of the Hall Institute, the outbreak of Murray Valley encephalitis in 1951.

Bacteriophages

Although as the pathology registrar at The Walter and Eliza Hall Institute in 1924-25 he was responsible for clinical bacteriology for the Melbourne Hospital, Burnet immediately began to carry out research. In 1924, shortly after beginning work at the Institute, he had acquired a copy of an English translation of Felix d'Hérelle's first book on bacteriophage (9). His fascination with this subject was heightened by the observation, soon afterwards, of bacteriophage plaques in a culture of Escherichia coli grown from the urine of a patient with pyelitis. The study of bacteriophages was to dominate Burnet's research for the next decade, and his 32 papers on the subject, published between 1924 and 1937, include two authoritative reviews on bacteriophages themselves, one review on their immunological reactions, and several papers of seminal importance for what came to be the sciences of molecular biology and microbial genetics.

In contrast to d'Hérelle, who held that the phenomenon of transmissible bacterial lysis was caused by self-reproducing virus particles, many other scientists of the period, including such notable figures as Jules Bordet and André Gratia, maintained that the phenomenon was caused by bacterial enzymes. Burnet was convinced by the logic of d'Hérelle's view of the particulate nature of bacteriophage, but his experience with isolations from human faeces soon led him to believe that d'Hérelle was wrong in insisting that there was only a single, highly variable, species of virus – the bacteriophage. He thought that there were many different species of bacteriophage, and showed that different strains differed greatly in physical and physiological characteristics. In order to establish this point unequivocally, he adopted an approach that was to characterise his later work in animal virology and reflects his childhood interest in collecting and classifying beetles. Taking advantage of the opportunity provided by his brother's dairy farm, he collected specimens from fresh excreta of pigs, cows, horses and chickens, from which he isolated many bacteriophages. Up to this time the principal method of classification was that introduced by Bail (10), viz., study of the resistance patterns of 'smooth' and 'rough' salmonellas to various bacteriophage strains. Burnet decided to employ serology (virus neutralization) for the classification of his collection, and found that using this method, 50 cloned bacteriophage strains could be classified into 12 natural groups. All members of each serological group also produced plaques with the same general structure and showed similar patterns when studied by Bail's method. His observations of physical differences between different strains of bacteriophage was greatly strengthened by the demonstration by Elford & Andrewes (11) that different bacteriophages, mainly from Burnet's collection, differered greatly in size, as judged by filtration through graded collodion membranes.

Although d'Hérelle had made many fundamental observations on bacteriophage and had introduced the basic techniques for its study, viz., the limiting dilution method and the plaque assay, he was principally concerned with its possible use for the therapy of human diseases. As a medical bacteriologist, Burnet had a similar concern and produced three papers exploring such possibilities. However, his major interest was with the nature of bacteriophages and their interactions with bacteria. Several of his contributions were to be of lasting historical importance, notably a paper on techniques for studying bacteriophage multiplication, papers on the nature of lysogeny, and experiments on the inheritance of bacterial resistance to bacteriophages.

Bacteriophage multiplication. In 1926 d'Hérelle had demonstrated that with a highly virulent strain of bacteriophage and highly susceptible bacteria, bacteriophage multiplication caused step-wise increases in titre. However, proponents of the bacterial enzyme hypothesis of bacteriophage action regarded this as a special case. By modifying d'Hérelle's methods, Burnet was able to show that the step-wise increase in titre was a general phenomenon, applicable to all bacteriophages. These experiments provided the basis of the classical experiment of Ellis & Delbrück (12) on the one-step growth experiment, a technical manipulation that was to prove of crucial importance in the use of bacteriophages for the development of molecular biology. Shortly after I had joined the staff of the Hall Institute in 1946, Burnet gave me reprints of the Ellis-Delbrück and subsequent Delbrück papers to read, with the remark that they were scientifically fascinating, but of no practical importance. Parenthetically, Delbrück also had a blind spot; he did not believe in lysogeny but thought that persistence of bacteriophages in some cultures was due to cryptic infections.

Burnet also carried out important experiments on the initial stage in virus multiplication, viz., the attachment of virus particles to the susceptible bacterial cell. He proposed that the initial contact between infecting virus and bacterial cell was a stereo-specific process between complementary structures on virus and cell, analogous to an antigen-antibody reaction, and showed that bacterial extracts could specifically inactivate bacteriophage particles to which the intact cell was sensitive, but that similar extracts of bacteriophage-resistant bacteria could not.

The significance of lysogeny. The phenomenon of lysogeny has played a central role in the formulation of ideas about bacteriophages. Excluding contamination of a partially susceptible bacterial strain with a bacteriophage ('carrier' cultures), certain bacterial strains exhibit lysogeny, i.e., during their multiplication the bacteriophage genetic material is replicated as part of the bacterial genome (prophage), but occasionally certain bacterial cells release viral particles, which can be detected by their effects on susceptible bacteria.

Lysogeny provided Bordet and other critics of d'Hérelle with their most serious objection to the notion that serially transmissible bacterial lysis was caused by a particulate virus, since lysogenic bacteria reproduced the lytic principle during their growth without the viability of the cell being affected – a contradiction to beliefs of d'Hérelle and his contemporaries about the essential nature of viral reproduction. The problem was conclusively solved by the elegant experiments of Lwoff & Gutmann (13), involving the cultivation of individual lysogenic bacteria in microdrops, which led to the notion of 'probacteriophage' (later called 'prophage' and ultimately generalized to 'provirus'). As these authors noted, however, Burnet & McKie (14) had already come close to this view, when they said that permanence of the lysogenic character made it necessary to assume the presence of the bacteriophage or its anlage in every cell of the culture, and drew the conclusion that it was a part of the hereditary constitution of the bacterial strain. In other experiments, Burnet also recognized the difference between resistance of bacteria at the level of absorption of bacteriophage particles and what came later to be called the 'immunity' of lysogenic bacteria to infection by a bacteriophage homologous to that it already carried.

Microbial genetics. Since so many bacteria are lysogenic, the development of bacterial genetics has been inseparable from studies on bacteriophages. To this extent Burnet's contributions to the understanding of lysogeny, just discussed, are an important part of the early history of microbial genetics. Two other papers report pioneering experiments in what came to be the science of bacterial genetics. Many years before Luria and Delbrück (15) published their classical paper on the 'fluctuation test', showing that the occurrence of bacteriophage-resistant bacteria in a culture exposed to bacteriophages was due to the selection of bacterial mutants, Burnet (16) had reached the same conclusions, by selecting resistant mutants by their colonial morphology, without the use of phage as a selective agent. Subsequently, Burnet & Lush (17) wrote the first paper on bacteriophage genetics, when they discovered a bacteriophage whose capacity for being carried in the lysogenic state had been lost permanently, by mutation.

Staphylococcal toxin

Burnet arrived back in Australia from his first sojourn in England in December 1927, filled with enthusiasm to carry on his work with bacteriophages. On 27 January 1928, however, within twelve hours after twenty-one children in Bundaberg, in Queensland, had received injections of a diphtheria toxin-antitoxin mixture (then the accepted method of immunization against diphtheria), eighteen of them had become ill, and twelve died within twenty-five hours. Dr Charles Kellaway, the then Director of the Walter and Eliza Hall Institute, was immediately appointed Chairman of a Royal Commission to investigate the fatalities (18) and Burnet was deputed to carry out the laboratory part of the investigations. He soon showed that Staphylococcus aureus could be recovered from both the fluid in the toxin-antitoxin bottle and the pus in the abscesses of survivors. This led him into a completely new field, and over the period 1928-31 the behaviour of staphylococci and their toxins was the central theme of his research, with bacteriophages taking second place.

Over the next four years Burnet published nine papers on the staphylococcal alpha toxin, which was regarded as the cause of death in these children, and some years later, a paper on staphylococcal bacteriophages. The Bundaberg disaster was important in the history of The Walter and Eliza Hall Institute, because the effective work of its Director, Charles Kellaway, and his staff on a matter of great public interest impressed the name of the Institute and the significance of medical research on the Australian public. Burnet's work on the staphylococcal exotoxin extended a field that had barely been studied before, but the most important aspect of this work for Burnet's future development was an incidental observation on the antibody response of rabbits after intravenous or subcutaneous injection of the toxoid. Despite a very small and slow response to the first injection, a second injection a few weeks later led to an immediate and rapid rise in the antitoxin level, which rose logarithmically over a period of 40-120 hours after the second injection. Burnet' s interpretation of this phenomenon was that something was duplicating itself every twelve hours or so to produce the antibody. His paper on these results and their implications was rejected by the British journal to which it had been sent, but this only stimulated him to collect further information on the topic and to publish it in an Institute monograph, The Production of Antibodies, which was eventually published in 1941 (19). These data also figure prominently in the second edition of the monograph (20), for Burnet saw in this difference between the primary and secondary response, overwhelming evidence that the Haurowitz-Mudd-Pauling 'instructive' hypothesis of antibody production could not be correct.

Animal virology

Although Burnet had already carried out some work with poliomyelitis virus (21), his introduction into animal virology really came with his second, two-year-long visit to England in 1932-33. Before Kellaway came back to Australia in 1923, he had worked with Sir Henry Dale, the Director of the National Institute of Medical Research in Hampstead, England. Since its opening in 1919, the microbiology department of the National Institute had concentrated on virus diseases, and by 1931 Dale had gathered together an active group of workers who had made some well-publicized discoveries and were at that time recognized as world leaders in this field. The Rockefeller Foundation offered Dale substantial support to develop work on animal viruses further, and through Kellaway, Dale asked Burnet to come to the Institute on a two-year appointment, to study animal viruses. After assuring himself that there would be a post at the Hall Institute when he returned, Burnet accepted the offer and started work at Hampstead early in 1932.

At that time animal virology was in its infancy. Apart from smallpox and vaccine virus, which had been studied for many decades, the only viruses of medical importance that had been isolated were those that caused herpes simplex, poliomyelitis, rabies and yellow fever, and very little detailed study had been made of any of these. However, since all these viruses were already being studied by staff members at Hampstead, another virus had to be found for Burnet. The chance came when Kikuth, the German worker who had just discovered the first effective synthetic antimalarial drug, atebrin, asked Dale to help with study of a virus that had caused problems in their testing of antimalarials. With the collaboration of J.E. Barnard and W.J. Elford, Burnet (22) showed that the virus was a large one and correctly identified it as canary-pox virus, related to but different from fowlpox virus. The fact that it was a virus of birds suggested to Burnet that it might be a good candidate for growth on the chorioallantoic membrane – a technique described for fowlpox virus a year before by Woodruff & Goodpasture (23) at Vanderbilt University, USA. Many years later, Burnet was to pay gracious tribute to Goodpasture (24). The initial experiments were successful, and Burnet was introduced to the developing chick embryo – an experimental animal that was to dominate his work in virology, and even in immunology, for the rest of his life at the bench.

At the National Institute of Medical Research, Burnet found a lively group of colleagues of about his own age, including C.H. (later Sir Christopher) Andrewes, with whom he continued to conduct a correspondence ('FMB' to 'CHA', and vice versa) for many years afterwards. Although working with animal viruses, Andrewes and Elford were also interested in bacteriophages, studying their size, as determined by gradocol filtration (25) and the mechanism of neutralization by antibody (26). With this example and stimulus, Burnet himself divided his time between studies of animal viruses (mainly their growth on the chorioallantoic membrane), and further work with bacteriophages, on which he published, from the National Institute of Medical Research, seven experimental papers and a major review (27).

Growth of viruses in the developing chick embryo. Although Goodpasture and his colleagues had shown that fowlpox and vaccinia viruses could be grown on the chorioallantoic membrane, they had always used large inocula and obtained confluent growth. In the work described in his paper on canary-pox virus (28) Burnet also used concentrated inocula. Some time later, however, he noticed that with dilute suspensions, opaque spots of proliferating cells a few millimetres in diameter were produced. Here was a system comparable to plaque assay with bacteriophages, that might be employed for the titration of animal viruses and antisera to them. However, it was not until 1936, after he had returned to Melbourne, that he was to utilize the pock-counting technique for studying the relationship between canary-pox virus and fowlpox virus (29).

Growth on the chorioallantoic membrane. Having found that canary-pox virus grew on the chorioallantoic membrane, Burnet again followed his collecting habits. He studied all the viruses he could obtain, whether from human or animal sources, to study their growth on the chorioallantoic membrane and their effects on the developing chick embryo. In rapid succession, papers appeared on the growth on the chorioallantoic membrane of infectious laryngotracheitis virus (30), fowl plague and Newcastle disease viruses (31), vesicular stomatitis virus (32), influenza virus (33), psittacosis 'virus' (34), louping ill virus (35) and ectromelia virus (36). Initially, he merely tested for growth, by subinoculation into susceptible animals, and studied the macroscopic and histological changes in the membrane and elsewhere in the chick embryo.

In 1936 he published his first paper on the use of the pock-counting technique, with avian laryngotracheitis virus (37), and illustrated the potential of this method for assaying antibodies to the virus. He immediately applied the method to other viruses, and by the time he came to write his monograph on the use of the developing egg in virus research (38), he or workers in his laboratory had shown that a variety of viruses could be assayed in this way – avian poxviruses, vaccinia, ectromelia, herpes simplex, infectious laryngotracheitis and louping ill viruses, and after adaption by serial passage, influenza A virus.

Over the next four years (1936-40) he worked on a variety of viruses and with the chlamydia of psittacosis and the rickettsia of Q fever. A series of eight papers utilized pock-counting of egg-adapted influenza virus for the study of various aspects of influenza; other papers illustrated the use of the pock-counting method for the analysis of the natural history of herpes simplex and the pathogenesis of louping ill (39). A few years later he produced a paper (40) describing in detail the methodology of the pock-counting technique, including ways of minimizing the occurrence of non-specific lesions.

Amniotic inoculation. Following a report that inoculation of meningococci into the amniotic cavity produced infection of the lung and meninges of chick embryos, Burnet demonstrated that unadapted (ferret) as well as egg-adapted strains of influenza virus could be propagated in the chick embryo by amniotic inoculation (41), and that this route of inoculation could be used for titration of influenza virus and antibodies to it (42). It also provided a new, simpler and more sensitive method than ferret inoculation for the recovery of influenza virus directly from human patients (43). Until about 1968, when it was found that the Hong Kong strain would grow directly in the allantoic cavity, amniotic inoculation continued to be the method of choice for the recovery of influenza virus from human and animal sources.

Allantoic inoculation. Although he had previously observed that the allantoic fluid contained large amounts of virus after the amniotic inoculation of influenza virus, Burnet had regarded this as having been derived from the infected lung, and did not test whether influenza virus would grow after direct inoculation into the allantoic cavity. However, after reading that Nigg et al. (44) had found that a high yield of influenza virus could be obtained from membranes of chick embryos inoculated through the chorioallantoic membrane, Burnet tested direct allantoic inoculation, and showed that 2-3 days later all strains tested could be recovered to high titre in the allantoic fluid. He commented that this method might be useful for the production of large amounts of virus for use as vaccine (it is still the preferred method of preparation of influenza vaccine). He also noted that with some strains of influenza virus that multiplied to high titre, the embryo was unaffected and hatched normally, however no antibody to influenza virus was produced. It was not until 1950 that he used the chick embryo to test for acquired immunological tolerance (45).

With the discovery of haemagglutination by influenza virus by Hirst (46), the possibility arose of using allantoic inoculation as a cheap, simple and reliable method of titrating influenza viruses and their antibodies (47). Later he was to concentrate the full resources of the virus group in the Hall Institute on the elucidation of the haemagglutination-elution phenomenon. In 1942 Burnet published his first paper on the genetics of influenza virus, based on differences between viruses that were maintained by amniotic passage (O) or passed in the allantoic sac (D) (48). This was a topic that was to become his major interest in the 1950s.

This phase of Burnet's research was rounded off with the publication in 1946, with his colleague W.I.B. Beveridge, of the second edition of their Medical Research Council monograph (49). In contrast to the first edition, which was concerned only with the results of inoculation on the chorioallantoic membrane, all routes of inoculation – chorioallantoic, amniotic, allantoic, intravenous, intracerebral and yolk sac – were discussed.

Psittacosis and the ecological approach to infectious diseases. Burnet did not carry out much research on psittacosis, since he published only six papers, over a period of eight years, on the topic. Two of these were routine papers for a laboratory-based microbiologist – one the demonstration that the chlamydiae of psittacosis, like many viruses, multiplied on the chorioallantoic membrane, with the production of pocks when dilute suspensions were used, with characteristic Levinthal-Coles-Lillie (LCL) bodies (50); and the other with the production of focal pulmonary lesions after the intranasal inoculation of mice with dilute suspensions and the use of the method for titrating the agent (51). However, his work with psittacosis had some interesting side effects:

1. it brought him into contact with Karl Meyer, a powerful figure in contemporary public health activities in California, leading to a lifelong friendship;
2. his recent use of Castaneda's strain for LCL bodies led to his early recognition of rickettsiae in Q fever material; and
3. his studies of latent psittacosis and an outbreak of lethal disease in Australian wild parrots directly influenced his thinking about the ecology of infectious diseases.

It may be useful to elaborate somewhat on the last of these matters here. In his autobiography (52), Burnet notes that he was '...by temperament an ecologist, a naturalist...'. Until 1934, his naturalist's instincts had been largely directed to beetle collecting, bird watching, and curiosity about the ecology of the bacteriophages of intestinal bacteria. In 1934, in response to a request from the Commonwealth Director-General of Health, he demonstrated that psittacosis was present in apparently healthy parrots obtained from bird dealers in Adelaide and Melbourne (53). Following up this study, he demonstrated that asymptomatic psittacosis was enzootic among Australian parrots in the wild, but could cause disease when parrots were stressed under conditions of confinement by bird dealers (54). Some years later, he was able to investigate outbreaks of fatal psittacosis that occasionally occurred among wild parrots in nature (55).

Lysogeny was, of course, a perfect example of latent, inapparent infection, and experimental work during 1935-36 had impressed Burnet with the frequency with which inapparent infections occurred in laboratory animals deliberately infected with different viruses (56). Psittacosis exemplified a similar situation, and he interpreted data on the epidemiology of poliomyelitis and yellow fever in man as indicating that most infected humans suffered inapparent infections with these viruses (57). Over the next year or so Burnet put these ideas about the ecology of infectious diseases and immunity together as a book, Biological Aspects of Infectious Disease, written 'from the point of view of a biologist as much interested in how the parasite species survives as in how the host resists it'. Written before he had read the only other comparable book at that time – Theobald Smith's Parasitism and Disease, it presented a rather similar point of view. He later acknowledged his debt to Smith, whose ideas, he said, 'filtered through the writings of others long before I read his famous exposition of the ecological approach in medicine'. This first semipopular book of Burnet went through four editions (1940, 1953, 1962, 1972), and was translated into German, Italian, Japanese and Spanish. From my own contacts with scientists in the United States, I know that it and his 1944 Dunham Lectures, Virus as Organism, had a considerable impact on many biochemists and microbiologists, by showing the value of thinking of infectious diseases from the point of view of the survival in nature of the parasite, rather than just as diseases of the vertebrate host.

Burnet was to come back time and again to this ecological point of view (58) hence his great interest in myxomatosis and Murray Valley encephalitis in Australia, two diseases on which he did not carry out any investigations himself, although the work on Murray Valley encephalitis was carried out under his direction.

Q fever. In 1935 physicians in Brisbane became concerned with the sporadic occurrence of a typhoid-like disease among abattoir workers, from which no bacteria could be recovered. Guinea pigs were susceptible, but again no organisms could be recovered (59). It was reasonable to assume that the disease was due to a virus, hence organs from an infected guinea pig were sent for investigation to Burnet, late in 1936. Burnet subjected the material to the usual series of tests in experimental animals, making inoculations in guinea pigs, monkeys, mice, rats and on the chorioallantoic membrane, but soon concentrated on studies in mice, using normal and immune guinea pigs to determine the specificity of the findings (60). In all his studies of the growth of viruses in experimental animals, Burnet used to examine infected organs histologically. Sections of the enlarged mouse spleens showed no inclusion bodies but under high power magnification Burnet noticed a 'vague herringbone pattern', which recalled what he had seen in psittacosis and had read about for rickettsiae. Using Castaneda's stain, he decided that there was no doubt but that the agent was a rickettsia. In an addendum to his first paper on Q fever, Burnet reported that he had been able to recover the organism from the blood of a patient, and that acute and convalescent sera of another patient showed a substantial rise in agglutinating titre against a rickettsial suspension, thus establishing that it was the cause of Q fever.

The next step was the development of a serological test. Mouse spleens contained very high concentrations of the rickettsiae, which could be substantially purified by differential centrifugation and provided a satisfactory agglutinogen (61). Having confirmed by serological tests that the rickettsia that they had isolated was without question the cause of the human disease, further work with the agglutination test devolved on Derrick, who used the method with good effect to unravel the epidemiology of Q fever in Queensland (62).

Burnet's subsequent investigations on Q fever were concerned mainly with determining the relationship of the Q fever organism to other microorganisms. In a rare excursion into tissue culture, he showed that it behaved like the typhus rickettsiae and unlike viruses or chlamydiae in its capacity to continue to multiply in damaged cells (63).

Subsequent studies involved direct comparisons with other known rickettsiae, the upshot of which was to show that there was no serological relationship between the Q fever rickettsia and other known pathogenic rickettsiae (64). However, early experiments (65) suggested and later investigations (66) conclusively demonstrated that it was identical with a rickettsia isolated from ticks in Montana, USA, by Cox (67), except that the American strain was much more virulent for guinea pigs.

Apart from being the first of many laboratory workers to be infected with Q fever (68), the other feature of note in Burnet's association with Q fever is that the causative organism was named after him – first, by Derrick, as Rickettsia burneti and subsequently, when taxonomists split the genus, as Coxiella burnetii. As noted in a recent review (69), 'The papers of Derrick and of Burnet and Freeman remain models of careful investigations, critical analyses, and conclusions'.

Poliomyelitis. During the 1920s and 1930s epidemics of poliomyelitis were common in Melbourne, and as a medical virologist Burnet inevitably became involved in the experimental study of polioviruses. An early study (70) provided the first inkling that there was more than one serotype of poliovirus; monkeys that had recovered from intracerebral inoculation with either the Rockefeller Institute 'MV' strain (now known to be poliovirus type 2) or the local strain (probably type 1) were immune to reinfection with the homologous strain but susceptible to the heterologous strain. A severe epidemic of poliomyelitis occurred in Victoria in 1937-38, with over 1900 paralytic cases, and Burnet was appointed by the State Government to the local Advisory Council on the outbreak, and had his first experience of public affairs when he acted as its spokesman.

He was also asked to undertake experimental investigations into the disease, and over the period 1938-40 he and his colleagues produced seven research papers on poliomyelitis in monkeys. After isolation of the virus causing the epidemic in rhesus monkeys, Burnet and his colleagues (71) developed intraocular inoculation as a preferable alternative to intracerebral inoculation in tests for neutralizing antibodies. Then the supply of rhesus monkeys ran out, because of a six-months-long closed season in India. As an alternative, cynomolgus monkeys were obtained from Singapore. Although some earlier workers had reported that cynomolgus monkeys, unlike rhesus, could be infected by the oral route, Flexner (72), in extensive experiments with the 'MV' strain, had been unable to confirm this result. Burnet and his colleagues (73) found that cynomolgus monkeys were readily infected by all routes of inoculation, including feeding, swabbing the pharynx, and after laparotomy, inoculation directly into the stomach or small intestine. The orthodox view at the time was that, apart from cases after recent tonsillectomy, the only 'natural' route of human infection was via the olfactory bulbs (74). However, the results obtained with cynomolgus monkeys suggested to Burnet that infection of humans with poliovirus might normally occur by oral or pharyngeal routes. Extending this study (75), he found that poliovirus could be recovered from pharyngeal tissue, certain local nerves (vagus, coeliac plexus), and mesenteric lymph nodes of cynomolgus monkeys infected by the oral or intestinal routes, and then went on to carry out the second and last experiment of his career employing tissue culture. Lung, intestine and buccal tissues of a 12-weeks-old human foetus were used to set up 'Rivers-type' tissue cultures and each culture was inoculated with poliovirus. After incubation for three days, the centrifuged supernatant fluids were inoculated intracerebrally into monkeys; those from the intestinal and buccal tissues, but not from the lung tissues, yielded virus. Confirmation of this experiment, the first demonstration of the cultivation of poliovirus in non-nervous tissues, was not possible because 'we have been unable to obtain any other suitable human embryos...so that its implications must be accepted with great reserve'. That paper reported the last of Burnet's experimental work with poliovirus. In a review article (76) published, ironically, in 1949, Burnet 'adopted a wholely defeatist attitude towards the problem of poliomyelitis and...[hoped] that further developments [would] prove [him] wrong'. Yet his last unconfirmed experiments ten years before had left him poised on the edge of the discovery reported in the classical paper of Enders et al. (77), which was to make possible the effective control of the disease.

Herpes simplex. Burnet and his co-workers wrote only six papers on herpes simplex, all of which were published in 1939. They provided him with confirmatory evidence of the value of the ecological approach in virus research. The studies began with the demonstration that herpes simplex virus of man, B virus of monkeys and pseudorabies virus of swine, which Sabin (78) had shown shared many characteristics, grew well on the chorioallantoic membrane (79), which provided an accurate and sensitive method for the titration of antibodies to them (80). Burnet confirmed Sabin's opinion that these three viruses were members of a natural group (now designated as the subfamily Alphaherpesvirinae).

Burnet's principal contribution lay in describing, for the first time, what is now the accepted view of the epidemiology of this ubiquitous human disease (81). After confirming that aphthous stomatitis in infants was usually due to herpes simplex virus, he and his colleagues showed by serial antibody assays that these were primary infections. They suggested that non-specific resistance to primary infection developed in later childhood, except when there was intimate exposure. In adults, there was a sharp distinction between persons with high titre antibody and those without any antibody – intermediate levels of antibody were not found. Further, the presence of antibody was correlated with socio-economic status, being lowest among university graduates and highest among public hospital patients. It was clear from the occurrence of recurrent herpes that virus persisted somewhere in the body, but like others, Bumet failed in efforts to demonstrate it directly, by cultivation of fragments of skin or of Gasserian ganglion. Recurrent herpes simplex occurred in the presence of high levels of antibody and was due to reactivation of the latent virus, by mechanisms then unknown. The epidemiology of pseudorabies in swine and B virus in monkeys, Burnet concluded, was very similar to that of herpes simplex in man, viz., asymptomatic infection, usually in very young animals, with lifelong persistence of both virus and antibody. In animals other than their natural hosts, all three viruses could produce severe disease.

Poxviruses. Burnet's first paper on animal virology was the demonstration that the causative agent of a disease of canaries was a poxvirus (82), a study that led to his lifelong devotion to the use of the developing egg as a laboratory animal. His early studies with infectious ectromelia virus, which had been discovered at the National Institute of Medical Research a few years earlier (83), established that pock-counting was a feasible method for assaying this poxvirus (84), a technique that I was to use extensively a decade later. It was in studies with ectromelia virus in the developing chick embryo that Burnet introduced into virology the concept that the temperature of incubation influenced viral multiplication, later to be extensively developed in poxvirus research by Bedson & Dumbell (85) under the designation of 'ceiling temperature'.

Following the chance observation by Burnet that a suspension of vaccinia virus agglutinated fowl red blood cells, Nagler (86), working at The Walter and Eliza Hall Institute, demonstrated that vaccinia virus agglutinated the red cells of certain fowls only, and that this haemagglutination could be inhibited by anti-vaccinial antibodies. Recalling his experiments with ectromelia virus a decade earlier, Burnet then showed that ectromelia virus would agglutinate cells agglutinable by vaccinia virus and that ectromelia haemagglutination was inhibited by vaccinia-immune serum (87). When he had been working in Hampstead in 1932-33, Burnet had been interested in Topley's studies in experimental epidemiology, especially in those involving ectromelia virus (88). Now that he had shown that ectromelia virus was an Orthopoxvirus (as the genus was later designated), he decided to develop further work in the experimental epidemiology of viral diseases in the Walter and Eliza Hall Institute, based on studies with ectromelia, and in 1946 he appointed me to do this. Burnet himself continued with laboratory studies of vaccinia haemagglutinin, and showed that unlike haemagglutination by influenza virus and the arboviruses, the haemagglutinin of vaccinia and ectromelia virus, as found in extracts of infected egg membranes or rabbit skin, was separable from the virions, and that nonspecific tissue lipids also agglutinated only those red blood cells susceptible to agglutinination by the orthopoxvirus haemagglutinins.

Virus classification. In contrast to his friend C.H. Andrewes and the famous French virologist André Lwoff, Burnet was not deeply interested in the classification and nomenclature of viruses. However, because of his eminence as a virologist and his position as President-elect and then President of the International Association of Microbiological Societies, he unavoidably became involved in discussions about viral taxonomy. His first contribution came at an international conference of which he was chairman, 'Virus and Rickettsial Classification and Nomenclature', held at the New York Academy of Sciences in 1952. In his introductory address (89), he outlined his ideas on criteria for allocation to a genus ('...approximately the same size and appearance in electron micrographs and, at least, one common functional characteristic'), and concluded by suggesting that '...we should go all out to make a start on virus classification...'. This initiative was followed up at the International Congress of Microbiology in Rome in 1953, where Burnet played an important role in developing a compromise between those who wished to introduce a Linnaean binomial nomenclature forthwith, and those who opposed this. He also suggested that for animal viruses the group names should carry the suffix '-virus', an idea that eventually developed into the present system for viral families ('-viridae'), subfamilies ('-virinae'), and genera ('-virus').

Subsequently he was a member of the subcommittees that prepared reports on two virus groups: the 'myxoviruses' (90) – later to be divided into two families, Orthomyxoviridae and Paramyxoviridae – and the poxviruses (91).

Influenza. Between 1934 and 1939, after his return from Hampstead, Burnet's investigations ranged over a wide variety of different animal viruses, and included also the non-viral causative agents of psittacosis and Q fever. Influenza virus was among the viruses for which he was able, with a suitably adapted strain, to develop a pock-counting method of assay. However, this technique was never used by other investigators, and his major contributions to the study of influenza virus began in 1940, with the demonstration that amniotic inoculation of the developing chick embryo provided a method for isolating virus directly from human patients, a method which quickly supplanted intranasal inoculation of ferrets. Continuing his exploration of routes of inoculation of the developing egg, he showed that the allantoic route, while not suitable for isolation of virus from human subjects, could be used for large-scale production of virus that had initially been isolated in the amniotic sac.

By that time the Second World War had begun, the then Director of The Walter and Eliza Hall Institute, Kellaway, was heavily involved as Director of Pathology for the Australian Army, and Burnet had to serve as Acting Director. With memories of the devastation caused by the influenza pandemic that followed the First World War revived by a review of the literature of that disaster (92), Burnet decided that his war effort should be the development of a method of immunization against influenza. In fact, the study of influenza virus became the major focus of his work, and that of The Walter and Eliza Hall Institute, of which he became Director in 1944, until 1957, when he made an historic shift to immunology. It took some two years after that change before papers on virology ceased to appear, and over the period 1942-59 Burnet's name was attached to some 114 papers on influenza virus. Since almost every other independent worker in the Hall Institute at that time, apart from the Clinical Research Unit, was working on influenza virus, the volume of investigations on this topic in which he was involved as an adviser was perhaps three times greater than this. It is not possible to describe here work on influenza virus carried out at this time by his colleagues and students in the Hall Institute – a description of this can be found in the Annual Reports of The Walter and Eliza Hall Institute over the relevant period, or more conveniently in Burnet's history of the Institute (93). However, it should be remembered that all of this work was strongly influenced by Burnet's ideas and perceptions, and often by his advice.

Although his own work covered almost every aspect of the biology of influenza and influenza virus, his major contributions fall into four fields:

1. methods of isolation of influenza virus from human subjects;
2. immunization against influenza;
3. the phenomena of haemagglutination and elution; and
4. influenza virus genetics.

His discovery and development of the amniotic and allantoic routes of inoculation have already been discussed; the next few pages outline in turn Burnet's work on immunization against influenza, haemagglutination and influenza virus genetics.

Immunization. As early as 1937 Burnet had found that egg-passaged influenza virus (after 65 passages on the chorioallantoic membrane) was non-pathogenic for ferrets and mice, but produced an immune response and conferred protection against challenge with virulent virus (94). Taking the view that only a live virus vaccine administered by the natural route was likely to be of any use if there was an influenza epidemic during or after the Second World War that was anything like that experienced in 1918-19, Burnet concentrated his efforts on trying to produce an effective attenuated live virus vaccine. In 1940 he reported the results of spraying various strains of influenza A virus, some attenuated by passage on the chorioallantoic membrane and others fully virulent, into the nose and throat of human volunteers (95). The attenuated strains had no protective effect, whereas the virulent strains caused typical influenza in most subjects who were previously sero-negative, but had no effect or produced subclinical infection (as evidenced by antibody rises) in those who had high antibody levels at the time of challenge.

Saving recovered influenza virus B by amniotic inoculation from human subjects in an epidemic in Melbourne (96), Burnet proceeded to test the efficacy as a vaccine of influenza B virus attenuated by amniotic passage, inoculated in human volunteers by the intra-nasal route (97). Antibody responses were observed only in those with low initial titres, and second inoculations produced a much lower proportion of antibody responses and virus reisolations than the first series of inoculations, suggesting that the vaccine might be protective.

However, any pandemic was likely to be caused by influenza A, and in February 1942 Burnet received permission to test in Australian Army volunteers influenza A virus that had been grown in the allantoic cavity of the chick embryo. Initial experiments were satisfactory, but by the time vaccination got under way on a large scale (20,000 men by June 1942), a natural epidemic of influenza A had already occurred (98); the immunization programme had been launched just too late to test its efficacy adequately. By 1943 experiments in USA with inactivated vaccine (produced by Burnet's method, in the allantoic cavity) had shown good enough results to convince the Australian Army that further experiments with live virus vaccine were not justified. Almost half a century later the position remains unchanged; inactivated influenza vaccines are not very effective, but a satisfactory live virus vaccine has still to be produced.

Haemagglutination. The agglutination of chicken red blood cells by influenza virus was reported independently by Hirst (99) and McClelland & Hare (100). It was a discovery that Burnet conceded that he should have made, for he had been working with influenza virus in developing eggs for much longer, and much more intensively, that anyone else – but he didn't follow up his observation that such clumping occurred. However, immediately after reading Hirst's paper, he saw the value of the method for assaying influenza viruses, and realized that the phenomena of haemagglutination and elution had the makings of a first class scientific problem. As the Institute staff built up after the end of the Second World War he deployed almost all of them on the study of haemagglutination. This work reached its peak in the period that I worked at the Hall Institute (1946-48); I was the only virologist there at that time who was not working on influenza virus and in one way or another on the phenomenon of haemagglutination-elution. Burnet believed that intensive team work was essential if the Institute was to be competitive with what were assumed to be the large teams working on the problem in the USA. In fact, McClelland & Hare did not follow up the discovery, and Hirst, who did, preferred to work alone, and was very conscious of the size and power of the group of scientists that Burnet had assembled.

A practical result of the availability of the haemagglutination test was to make all other methods of assay of influenza virus and antibodies to it obsolete (101), especially as it was directly applicable to untreated allantoic and amniotic fluids. Following his usual practice of testing new discoveries with all available viruses, Burnet soon showed that Newcastle disease virus also exhibited haemagglutination and elution (102). He seized on the demonstration by Levens & Enders (103) that mumps virus also agglutinated fowl red cells to point out its similarity to Newcastle disease virus (104), and demonstrated that vaccinia and ectromelia viruses produced a different kind of haemagglutination.

However, the major focus of interest was the phenomenon of elution. It was shown that cells from which a particular 'myxovirus' (influenza, Newcastle disease virus or mumps virus) had eluted were inagglutinable by that virus but agglutinable by others further down a 'receptor gradient'. Then came one of those feats of biological intuition which were the hallmark of Burnet's genius. Having observed that fowl or human erythrocytes from which 'myxoviruses' had eluted became susceptible to agglutination by normal sera that were without action on normal cells, Burnet recalled the phenomenon of 'panagglutinability' of human red cells described by Thomsen (105) and Friedenrich (106). This was ascribed by them to the action of bacterial enzymes, and Burnet (107) showed that enzymes of Vibrio cholerae, one of the bacterial species that produced panagglutinability, would remove viral receptors from red cells in almost the order of the receptor gradient. Further studies showed that V. cholerae filtrates contained other enzymes of interest – a mucinase and a 'tissue disintegrating enzyme' (108); however his main interest was in what was described as the 'receptor-destroying-enzyme' (109). At the same time Burnet seized on the discovery of Francis (110) that mucins would inhibit influenza virus haemagglutination to show that mucins were a substrate for both bacterial and viral receptor-destroying-enzymes. These two discoveries opened the way for Gottschalk, a skilled carbohydrate biochemist who until then had been outside the virus group, to join it (111), an event which changed Gottschalk's subsequent career and led to his pioneering work on sialic acid and the glycoproteins (112). The immediate result was the definition of receptor-destroying-enzyme as a neuraminidase (113); since then the influenza virus neuraminidase has been crystallized and its sequence and three-dimensional structure determined (114).

Leaving the biochemical work to others, Burnet's interest in haemagglutination and elution was principally in relation to what light it might shed on the initiation of infection by influenza virus, a topic that he reviewed and chose for his Croonian Lecture to the Royal Society (115).

Influenza virus genetics. Like other virologists, Burnet had always been interested in the changes in virus virulence, for particular hosts, that occur after serial passage of a virus in another host – the classical method of 'adaptation' for laboratory use and attenuation for use as a vaccine. With influenza virus, he had observed such changes after passage on the chorioallantoic membrane and after amniotic passage. However, his first explicit discussion of genetic changes in influenza virus came with observations of changes in the haemagglutination behaviour of strains of influenza virus newly isolated in the amniotic sac, and after serial passage (116). Newly isolated virus (O; original) differed from passaged virus (D; derived) in a number of characteristics, notably O virus showed a much higher haemagglutination titre with guinea pig cells than with fowl cells and would not multiply in the allantoic cavity; with D virus the haemagglutination titre was much the same with guinea pig and fowl cells and the virus multiplied readily in the allantoic cavity. Further, since passage in the amniotic cavity at high dilutions maintained the O characteristics whereas passage at low dilutions produced D virus, Burnet concluded that the change from O to D was a 'discontinuous mutation'.

He returned to this problem in 1945 (117), and showed that virus could be maintained in the O form if the inoculum was obtained from embryo lung emulsion purified by absorption with fowl cells, to which such O form virus does not attach. Further observations of sporadic and epidemic cases of influenza (118) supported the concept that in human infections influenza virus always occurred in the O form, and clarified anomalies apparent in earlier work (119). The molecular explanation of the difference between O and D forms emerged 40 years later. They differ in a specific amino acid residue in the cell-binding site at the distal tip of the haemagglutinin molecule, which alters the binding preferences of the virus for glycoprotein receptors with one type of sialic acid linkage to those with another. The mutation also produced an antigenic change that may explain the ineffectiveness of inactivated influenza virus vaccines, all of which are produced from allantoic fluid (120).

At the time, however, others had not been able to confirm Burnet's ideas about the mutational nature of the O-D change, partly, he believed, because of the difficulty inherent in the system. He therefore used a simpler system to establish the same principle, namely the maintenance of the neurotropic character of NWS influenza virus by serial passage in the allantoic sac at limit dilution, and the loss of the neurotropic character when passage was made at low dilutions (121).

Having established to his satisfaction that mutations occurred in influenza virus similar to those observed in bacteriophages, bacteria and higher organisms, Burnet set out to determine whether recombination would occur with mixed infections. He first defined two derivatives of the original WS strain of influenza A virus, WSM and NMW, by a number of very simple 'marker' characteristics – virulence for mouse lung and neurotropism (122). Taking advantage of the phenomenon of viral interference, he found that when a mixture of varying larger amounts of non-neurotropic WSM were mixed with a constant small amount of neutrotropic NWS and inoculated intracerebrally in mice, recombinants occurred at the level at which interference with NWS by WSM was just being overcome (123). Subsequently he extended the system by demonstrating recombination between two strains with different serological characteristics (124).

However, mouse brain inoculation, followed by limiting dilution analysis of the progeny of mixed infections, was a laborious process, and it was natural for Burnet to try to demonstrate recombination after inoculation of viral mixtures into developing eggs. In the first of three papers describing recombination between strains of influenza A virus in the developing egg (125), Burnet observed and described a novel kind of interference. He mentioned the possibility that the observed interference was due to some product of the virus-cell interaction which might modify the susceptibility of the target cells (vascular endothelium) – what would now be interpreted as interferon – but preferred the 'negative' interpretation, viz., that the ongoing viral multiplication in the allantoic membrane led to a deficiency in some plasma component which was needed if the virus was to multiply in and damage the vascular endothelium. Like his failure to discover haemagglutination, this was another 'near miss' – Isaacs, who was later to describe interferon and open a new field in cell biology (126), was working on interference between heat-inactivated and active influenza viruses in Burnet's laboratory at this time.

Subsequent papers (127) demonstrated that reciprocal recombination occurred between two different strains of influenza A virus in first-cycle viral multiplication in the allantoic cavity; back-cross experiments were also positive (128), and provided suggestive evidence for the production of 'heterozygotes' – a matter which was subsequently elaborated. He also showed that recombination would occur between two different strains of influenza B but not between strains of influenza A and influenza B virus (129), and obtained recombinants with a wide range of virulence for the mouse lung, a result which led him to postulate the possibility that the genome of influenza virus 'may fracture and the fragments themselves replicate independently' .

In his earlier writings on influenza virus genetics, Burnet noted with regret that single-cell experiments of the type used in bacteriophage genetics were not then feasible with animal viruses – a deficiency made good a few years later by Lwoff et al (130). However, he tried to simplify the system as much as possible, and turned to the use of de-embryonated eggs. Some of the progeny obtained in such experiments were doubly neutralizable, partly as a result of phenotypic mixing, partly, he thought, because some of them were heterozygous (131).

Over the next three years Burnet explored a number of unusual features of influenza virus multiplication by means of this approach, including the production of 'incomplete' virus (132), which he showed could contribute genetic information in recombination experiments (133), and the reactivation of inactivated influenza virus (134), which he interpreted, correctly, as being due to genetic recombination. He also reinvestigated the significance of heterozygosis (135) and probed further into the genetic control of viral virulence. By this time, however, Burnet realized that he had exploited the purely biological approach to influenza virus genetics as far as it would go. Ada, working in the Hall Institute, had shown that the genome of influenza virus was RNA (136) and that 'incomplete' virus contained less RNA than infectious virus (137). However, it was not until the demonstration by Pons & Hirst (138) that the genome of influenza virus was segmented that Burnet's results, and those of Hirst, fell into place. Until then, many virologists had regarded the 'high frequency recombination' demonstrated by these two workers with great suspicion, since it was so unlike the results obtained with bacteriophages.

Later, long after Burnet had abandoned the field, genetic reassortment, as the process has come to be called, was taken up as a method of producing vaccine strains (139), although in the process the occurrence of the O-D change rendered the vaccine less than ideal. It is now widely accepted, also, that new pandemic strains of influenza A virus arose, and may arise again, by reassortment between animal and human strains of influenza virus (140).

Mumps and Newcastle disease viruses. During his wide ranging examination of other viruses for evidence of haemagglutination, Burnet noticed that Newcastle disease virus behaved very like influenza virus, producing haemagglutination and then eluting from the red cells, although there was no serological relationship between the two viruses. He suggested that influenza, Newcastle disease, and mumps viruses belonged to the same group, and used all three species in experiments with the 'receptor gradient'. However, unlike influenza viruses, mumps and Newcastle disease viruses also lysed red blood cells. His only other contribution with these viruses was that he was himself the subject of the first reported case of human conjunctivitis due to Newcastle disease virus (141).

In 1955 he was one of the three members of a subcommittee which proposed the name 'Myxovirus' group for the influenza, mumps and Newcastle disease viruses, a taxonomic view based on particle morphology and the property of haemagglutination and elution, which had to be discarded when the properties of the genomes of these viruses were discovered (142).

Immunology

Apart from the relatively small specialty of human blood group serology, immunology remained largely the province of the microbiologist until transplantation became a practical measure in the 1950s. Like other microbiologists, Burnet employed serological techniques from the time of his entry into the laboratory (143), and he was an early exponent of serology as a method of bacteriophage classification. His interest in the immune response per se was stimulated by observations on the antibody response to staphylococcus toxoid, which led to an abiding interest in the production of antibodies and the publication of his first monograph on this topic (144).

After starting work on influenza virus in 1935, Burnet had by 1956 'worked out' what could be done with influenza virus genetics without adopting a molecular biological approach (which still lay some years in the future). Tissue culture methods were essential for the study of all other viruses, and he was reluctant to use this technique. On the other hand, his latent interest in immunology had been restimulated by Jerne's (1955) paper describing a 'selective' hypothesis for antibody production. At about the same time Dr Carleton Gajdusek, working in the Hall Institute, had found very high levels of autoantibodies in a patient with an immunoproliferative disease (145), and Simonsen (146) had shown that graft-versus-host reactions could be demonstrated on the chorioallantoic membrane, producing pocks due to cellular proliferation that could be regarded as clonal.

This combination of circumstances led, in 1957, to a decision by Burnet to reorient work at the Hall Institute from virology to immunology, although it took until about 1960 before publications on virology ceased to appear. From 1957 onwards, however, new students, staff and visitors to the Institute worked on immunological problems, Burnet himself being involved in bench work relating to autoimmune diseases and the graft-versus-host reaction, and increasingly in theoretical studies of immunology, immunological surveillance and cancer.

The production of antibodies. During the 1930s Breinl & Haurowitz (147) and Mudd (148) proposed a hypothesis to account for antibody production which was clarified and reformulated by Pauling (149); namely that antibody protein was synthesized, or according to Pauling, folded, in specific ways in spatial contact with the antigenically significant (determinant) parts of the antigen, which acted as a template – an 'instructive' hypothesis. Burnet could not accept this 'chemical' picture of antibody production, for a number of biological aspects of antibody production were incompatible with it. In 1941 he summarized his views in a monograph, in which he reviewed the known facts and developed some ideas on antibody production. Because of the apparently almost infinite variety of antibodies, he accepted an instructive hypothesis, but suggested that the antigen impressed a complementary pattern not on the globulin molecule, but on some cellular component – for 'antibody-producing cells must be capable of giving rise to descendant cells with the same faculty'. The same point of view was developed more forcefully in the second edition of the monograph (150), together with a new hypothesis for the process of antibody production itself based on an analogy with adaptive enzymes.

The more important feature of the second edition, however, was the exposition of a hypothesis concerning the manner in which the body normally failed to make antibodies to its own components – the 'self-marker' concept. In the course of the discussion of this concept, Burnet noted reports in the literature to the effect that mice and calves exposed continuously to antigens during embryonic life (congenital lymphocytic choriomeningitis virus and red cell antigens in some twin births respectively) failed to produce antibodies if exposed to these antigens in adult life. He made the comment: 'If in embryonic life expendable cells from a genetically distinct race are implanted and established, no antibody response should develop against the foreign cell antigen when the animal takes on independent existence'. This prediction was to form the basis for the award of the 1960 Nobel Prize in Physiology or Medicine to Burnet, jointly with Sir Peter Medawar, who had developed an experimental system demonstrating the generality of this phenomenon (151), something that Burnet (152) had attempted to do without success. However, even in 1955 Burnet saw no alternative to an instructive theory to account for the great multiplicity of antibodies that all animals can produce, although on this occasion he invoked the concept of an RNA 'genocopy' to serve as the template.

The revolution in his thinking came in 1956, after reading a paper by Jerne (1955), which developed a 'selective' hypothesis, in which it was postulated that every animal had a large set of natural globulins that had become diversified in some unknown fashion. According to Jerne, the function of an antigen was to combine with those globulins with which it made a chance fit and to transport the selected globulins to antibody-producing cells, which would then make many identical copies of the globulin presented to them. Burnet turned this idea over in his mind for several months, and '...it gradually dawned on me that Jerne's selection theory would make real sense if cells produced a characteristic pattern of globulin for genetic reasons and were stimulated to proliferate by contact with the corresponding antigenic determinant. This would demand a receptor on the cell with the same pattern as antibody...'. (153) Under appropriate conditions, such cells would either liberate antibodies or give rise to descendant cells that would do so.

Just before writing a short paper setting out this hypothesis, he saw Talmage's (154) review, in which somewhat the same idea was suggested. Essentially, Burnet envisaged the problem in terms of the population genetics of mesenchymal cells, with the variety of surface receptors and antibody globulins arising as a result of somatic mutation or 'by some other obscure process occurring during differentiation and development'. He published a paper on the subject (155) in an Australian journal not readily accessible to overseas scientists, for reasons which reveal some aspects of his personality. One was his Australian nationalism; he knew that it was a good idea and he wanted it to see first light in Australia. On the other hand, he had received adverse criticism of theories he had elaborated in a recent book (156), and he thought that by publishing the paper in this way he would have established priority, if it was eventually going to be recognized as important, and if there was something very wrong with it, very few scientists in America or England would have seen it (157). In fact, this short paper, in which he acknowledged Talmage' s contribution, still provides an excellent summary of the theory. Within two years, he had elaborated the concept as a book entitled The Clonal Selection Theory of Acquired lmmunity. He regarded the elaboration of this hypothesis as his most important scientific achievement (158), a view with which many biomedical scientists concur. Two immunologists who were working at the Hall Institute during the 1950s have recently summarized the history of the clonal selection theory. Over the last thirty years, it has led to a vast amount of experimental work, which has provided 'a rich insight into the biologic basis of immunity, and the central, unifying framework underlying this understanding is the clonal selection theory of antibody production' (159).

In 1957 Nossal (Burnet's successor as Director, now Sir Gustav Nossal) was working as a PhD student in Burnet's laboratory and he set out to test the clonal selection theory, by determining whether one antibody-producing cell could make more than one kind of antibody. None of 456 single cells challenged with two antigens produced two antibodies, although 33 were active against one antigen and 29 against the other (160). Further experiments from the Hall Institute (161), together with other evidence, finally provided formal proof of the validity of the clonal selection theory.

As he was increasingly called upon to give honorific lectures or to participate in symposia, Burnet used the clonal selection theory as the central point of his contributions, and it formed the theoretical basis of the major books on cellular immunology that were produced after his retirement. As new discoveries in immunology were made, e.g., the immunological functions of the thymus and the bursa of Fabricius, they were incorporated within the framework of the theory. However, his experimental work, which went on until his retirement from the Institute in 1965, was principally concerned with two other aspects of immunology – graft-versus-host reactions and auto-immune disease.

Graft-versus-host reactions. In 1957 Simonsen showed that when a chick embryo was inoculated intravenously with adult fowl blood, a graft-versus-host reaction occurred. Here was an immunological phenomenon demonstrable in the chick embryo, Burnet's favourite experimental animal. Further, it was amenable to quantitative study by the pock-counting technique. Over the three years 1960-62 the 'Simonsen phenomenon' was the major focus of Burnet's personal laboratory work. He studied the role of major histocompatability antigens (162) and the effects of corticosteroids on the reaction (163), and showed that chickens could be rendered tolerant by prenatal administration of embryonic spleen cells (164). He and his colleagues also continued to explore the roles of the thymus and bursa of Fabricius in the immune responses of the chicken. In a review of the history of the graft-versus-host reaction, Simonsen (165) commented that: '...the most significant use to which Burnet's group put their CAM assay was in their investigations of bursectomized chicks... That work marked the beginning of our understanding of the T and B cell dichotomy in lymphocytes'. By the end of 1962, however, Burnet felt that investigation of the graft-versus-host reaction on the chorioallantoic membrane had yielded as much as it was likely to in his hands, although a few years later it was to form the basis of experimental work that helped to reestablish the 'passenger leukocyte' concept in tissue transplantation (166).

Autoimmune disease. Burnet became interested in autoimmune disease in about 1955, partly because staff of the Clinical Research Unit of the Hall Institute suspected that some aspects of chronic hepatitis appeared to have an autoimmune basis. At that time, the laboratory findings on which ideas about autoimmune disease rested were the Coombs anti-globulin test, the anti-nuclear antibody basis of the lupus erythematosus (LE) cell effect, and autoimmune thyroiditis. The demonstration of LE cells in a patient with active chronic hepatitis (167), and the subsequent observation that such patients, and patients with macroglobulinaemia, had very high levels of antibody to extracts of normal human liver, forced Burnet to face up to the problem of autoimmune disease in the formulation of the clonal selection theory of antibody production (168).

One important aspect of Burnet's elaboration of the clonal selection theory was the notion of 'forbidden clones', which he suggested would provide an explanation for the 'self' 'not-self' conundrum. Autoimmune diseases were seen as aberrations of this mechanism. At 63, Burnet was still a keen experimenter, and he therefore turned to an experimental model which promised to provide an opportunity for the study of autoimmune disease. The model he chose was a strain of mice, 'New Zealand Black' (NZB), of which he heard by chance (169). With Dr. Margaret Holmes, he devoted the last few years of his life at the bench exploring various aspects of the biology and immunopathology of these mice, which spontaneously develop a high incidence of haemolytic anaemia of an autoimmune type, at an early age, and other signs recalling human systemic lupus erythematosus (170). Having shown that the anaemia could be transferred to young isologous mice by transfer of spleen cells from older mice (171), Burnet and Holmes showed that the affected mice developed characteristic thymic lesions (172). Over the next few years they studied the inheritance of autoimmune disease and thymic lesions emphasising the importance of a genetic factor and dismissing the influence of a virus (as suggested by others). The clearcut effect of cyclophosphamide in enhancing survival and abrogating renal disease (173) influenced clinical thinking on the use of immunosuppressive drugs in human autoimmune diseases. However, when Burnet abandoned laboratory work at the end of 1965 he was unsatisfied with the results of these experimental studies: '...without [a] break, the whole field may be deserted within a year or two', he wrote in 1967. In fact, the discovery of other inbred strains of mice that also developed autoimmune disease showed that the phenomenon was not just an idiosyncrasy of the NZB mouse, and murine models of systemic lupus erythematosus continue to be extensively exploited (174). However, mouse models have not proved to be very useful for studying the basic mechanisms of autoimmune disease.

After his retirement from the Hall Institute, Burnet continued to lecture and write on autoimmune diseases, and in 1972 he followed up the earlier technical book on the subject (175), written mainly by Mackay, with a second more general book of which he was sole author (176), designed for the 'physician or biologist', rather than the immunologist. Later he became more and more interested in ageing and diseases associated with it, such as cancer, which he approached as he had approached the biological basis of immunity, i.e., as a biologist interested in the population genetics of the cells of the body. Looking at cancer as an immunologist, he developed the concept of 'immunological surveillance'.

Immunological surveillance. In 1957 Burnet suggested that 'small accumulations of tumour cells may develop and because of their possession of new antigenic potentialities provoke an effective immunological reaction with regression of the tumour and no clinical hint of its existence' (177), a concept for which he later coined the term 'immunological surveillance'. However, he has said that he really developed this concept only after hearing of remarks by Lewis Thomas (178), suggesting that 'perhaps, in short, the phenomenon of homograft rejection will turn out to represent a primary mechanism for natural defense against neoplasia.' This happened after he had abandoned virology and was reorienting his interests, ranging widely over other kinds of human disease in which immune mechanisms might play a role, notably autoimmune diseases and cancer. He did not carry out any experimental work on surveillance, but discussed it in lectures (179) and reviews (180) over the succeeding years, ascribing a major responsibility to cellular immunity, mediated by T lymphocytes. In 1970 his views on the topic were elaborated in a book entitled Immunological Surveillance (181), and the same year saw the first international congress on the topic, which Burnet was unable to attend, but for which he provided a final comment (182). Over the next decade he further refined his views on surveillance in books and lectures, expounding the idea that a self-monitoring system was of major importance in cancers of the lymphoid cells, but accepting the widely held view that immune surveillance was probably much less effective in affecting the development of epithelial tumours (183). Inevitably, Burnet's views on surveillance now look dated, because since 1970 new immunological mechanisms that bear directly on the phenomenon have been discovered, such as suppressor cells, natural killer cells, and MHC restriction of the activity of T lymphocytes. However, it is still regarded by tumour immunologists as a useful concept (184).

Cancer

Burnet's experience on the Australian Radiation Advisory Committee (1955-59) had made him think about the relationship between ionizing radiation and cancers, especially leukaemia, and he spoke about this problem at some length to both Australian and overseas audiences (185). In 1957, in the process of looking at other fields of biomedical science as he moved out of virology, he undertook a survey of cancer as a biological problem, much as he had reviewed the 1918-19 pandemic of influenza in 1941 prior to embarking on attempts to produce an influenza vaccine. In these articles, and subsequently (186), he was highly critical of research in tumour virology, holding that the conditions under which experiments in this field were carried out were so highly selected and artificial that they had no relevance for the understanding of human cancer, its prevention or its cure. Burnet did not foresee how the 'oncogene' hypothesis, proposed in 1969 as a direct outcome of research in tumour virology (187), would change and develop so that by the late 1980s, in a radically different form, it promised to provide 'the final common pathway to tumorigenesis' (188).

His approach to cancer was profoundly influenced by the observation that in all mammals that have been adequately studied, the incidence of cancers increases with increasing age, reaching much the same levels towards the end of the life span, whether this was 2 years, as in the mouse, or 70 years, as in man. He looked for random processes in the renewable cells of the body, the likelihood of which would increase with the passage of time, as the key to the development of the malignant cell, and therefore espoused the somatic mutation hypothesis of cancer causation. As advances in molecular biology revealed the complexity of DNA replication and the role played by various enzymes in 'error repair', Burnet emphasized the importance of random somatic mutations in the genes for such enzymes in relation to both carcinogenesis and ageing. He found support for this concept in certain 'experiments of nature', such as high frequency of skin cancers in patients suffering xeroderma pigmentosum, in which there are congenital defects in these enzymes.

He thought that environmental causes of cancer – cigarette smoke, irradiation, etc. – might greatly enhance the likelihood that relevant sequential mutations might occur, but that even without such influences the error-proneness in the DNA replication process was subject to random mutation – a process that he called 'intrinsic mutagenesis'. In parallel with the increased likelihood with time of the emergence of a series of somatic mutations that might result in the production of a clone of turnout cells, Burnet envisaged that immunological surveillance (see above) diminished in efficiency with increasing age (189). As a corollary, turnouts would be likely to develop earlier in individuals with genetic or acquired immunodeficiencies. His concept of cancer was thus a logical extension of the application of Darwinian principles to the phenomena of disease and the interactions of cells within the body, just as was the clonal selection theory of antibody production.

Compared with the impact of the concepts of clonal selection and immunological tolerance on the field of immunology, Burnet's hypothesis of intrinsic mutagenesis has had little influence on cancer research. However, it illustrates Burnet's penchant for looking at specific problems from a broad biological and evolutionary point of view.

Public health

It was inevitable that as a medically qualified scientist interested in microbiology, Burnet should have been actively interested in public health and preventive medicine. Even in his early days of bacteriophage research, he explored the possibility that bacteriophages might have a role in the treatment of bacillary dysentery (190), and his major virological work, on influenza virus, was always done with one eye on the risks of another pandemic of influenza like that of 1918-19 and the need to develop a satisfactory method of protection against it. Later, when poliovirus vaccines became available, he was active both in Australia (191) and in the World Health Organization in advising on their use.

Perhaps his major contributions to public health, however, were in lucid addresses on the application of science to public health. Many of these were given to Australian audiences and most of them were published. Between 1939 and 1955, when he was still working on viruses at the bench, they dealt with infectious diseases in general, poliomyelitis, rickettsial diseases, influenza, allergic diseases, tuberculosis, staphylococcal infections, and Murray Valley encephalitis. In a more general analysis of infectious diseases (192), he took recent data on mortality statistics in childhood, and by using a log-log scale gave a good graphical illustration of the interactions of changes due to three factors: the inexperience and development of the immune system in early childhood, its over-reaction in early adult life and its decline in old age.

After moving away from virology he lectured on cancer and leukaemia, always with the possibilities of prevention in mind, and in his Presidential Address to the Australian and New Zealand Association for the Advancement of Science, he made a strong and well-publicized attack on the danger of undue exposure to medical and dental ionizing radiation, and on the relation between cigarette smoking and lung cancer. Other lectures with public health importance covered such topics as autoimmune disease, diseases of old age, kuru, and the risks of radiation.

Human biology

'Human biology' receives special though brief mention in this memoir because of Burnet's view of himself primarily as a human biologist, who from about 1940 had repeatedly tried to apply an understanding of biology to human diseases, and subsequently to human affairs. Initially his interest was in the ecology of the infectious diseases of man. His laboratory experience with the population genetics of bacteriophages and later of influenza virus was then applied to the populations of lymphocytes that make up the immune system, leading to the enunciation of the clonal selection theory of acquired immunity. Later, he applied a similar approach to his interpretation of the nature of cancer.

However, it was his experiences just after the Second World War, when as the newly appointed Director of The Walter and Eliza Hall Institute he was asked to serve on a number of official committees concerned with scientific research, that led him to take a serious interest in the major problems confronting the human species – notably war and overpopulation. A naive newcomer to official committees, he was shocked by their lack of interest in anything except short-term approaches to the problems with which they dealt. In an attempt to draw attention to the long-term problems of man as a mammal, in 1947 he wrote a book with the title Dominant Mammal. However, at that time it was rejected by both an English and an Australian publisher, and Burnet forgot about it until after his retirement from the Institute. He then went back to his original manuscript, reduced its formerly over-ambitious coverage, and rewrote the book in conformity with scientific knowledge in the late 1960s. It was published, with the title Dominant Mammal: the Biology of Human Destiny, in 1970 (193). This time, perhaps reflecting Burnet's status as an elder statesman of science, it was a success, being reprinted by Penguin Books and translated into Danish, Japanese and Spanish. Dominant Mammal expresses most clearly Burnet's philosophy of life. He returned to the same subject again in his last two books (194), in which, amongst other things, he examined human aggression as the expression of the genetic make-up of man, selected for during his long evolution as a hunter-gatherer, but totally inappropriate for civilized life.

Burnet's deep concern with human biology, encompassing particularly problems of population growth, was expressed again in his choice of these words, rather than 'medical research', for the title of the research institute established in Papua New Guinea in 1968, when he was chairman of the Papua New Guinea Medical Research Advisory Committee. It is ironic that in 1973, reflecting the greater popular and political interest in short-term medical research than in longer-term demographic problems, the name of the institute was changed to the 'Papua New Guinea Institute of Medical Research'.

Ageing

It was perhaps inevitable that a human biologist with as wide a spectrum of interests as Burnet, who continued actively to read and write well into his eighties, would become interested in the ageing process. It had been implicit in his earlier writings about immunological surveillance and the origin of cancers that both of these processes had a secular component – with increasing age both surveillance and error-correcting mechanisms became less efficient, whereas the likelihood of the occurrence of sequential mutations that might lead to cancer increased with age. In 1970 he specifically examined immunological surveillance in relation to problems of ageing, and a few years later wrote his first paper that dealt explicitly with the concept that the characteristic life span of man and other mammals was genetically determined, and that much of the process of ageing was due to somatic mutations in clonally proliferating cells in the body (195). He suggested that quite apart from the effects of extrinsic mutagens, somatic mutation depended on random errors in copying the DNA message, that mutations in the 'editing' enzymes might increase (or, rarely, decrease) the rate of 'intrinsic mutagenesis'. This was followed, in 1974, by his last referenced, 'technical' book, Intrinsic Mutagenesis: a Genetic Approach to Ageing (196), in which he discussed all aspects of senescence, including ageing of the post-mitotic cells of the brain and the social implications of the biological approach to ageing which he espoused. He accepted the biological necessity for death and was impatient with proposals designed to prolong the human life span. However, he saw '...wide scope for research on the best means of minimizing the depression and misery of pre-death...'. Happily, he and his relatives were spared a long period of dependent pre-death – he died, mentally acute until he lost consciousness, shortly after the onset of his last illness.

Books

This account of Burnet's scientific career mentions incidentally most of the books that he wrote, but this does not give adequate emphasis to his extraordinary productivity. He wrote no fewer than 31 books