R A Fisher: the life of a scientist Preface
Sir Ronald Fisher was my father. This might seem to disqualify me as his biographer, and I certainly had no notion of becoming one when, 5 years after his death, I began to ask his friends and colleagues for their recollections of him. I thought only to preserve the personal record before it was irretrievable. Then, just as Fisher's discourse used to fascinate his hearers in his lifetime, so the flow of his ideas began to fascinate me. His life took form as one continuous discourse from Fisher's lips. It had the vitality of his immense pleasure in the process of thinking, the play of ideas, the solution of puzzles. I recognized that, having known the man, I knew Fisher even as an undergraduate, when he stated the thesis of his life. From this starting point, I began to follow what, being his, was a logical and coherent argument, full of complexities, subtleties, and intriguing subplots. His conversation was not always easy to follow, but it was always fresh and stimulating. It is my hope that the reader of this biography also may take pleasure in the unfolding of a brilliant mind.
Fisher was an idealist, committed to the establishment of truth and the advancement of mankind. Born in 1890 in East Finchley, London, his mathematical ability was quickly apparent. His biological interests became important during his schooldays at Harrow. He accepted Charles Darwin's theory of evolution by natural selection, and he shared Francis Galton's concern that selective effects in society should tend to improve the biological inheritance of Man. However, the scientific facts were not well established: the nature of human inheritance was a matter of debate between geneticists and biometricians; Darwin's evolutionary theory had largely given way to theories involving genetic mutation, and most biologists assumed the effects of selection to be negligible. Above all, the means of establishing the truth of such matters on a firm, quantitative basis were practically nonexistent. So, while still an undergraduate, Fisher began his lifework on the quantification of the variable phenomena of biology; he began his services to the eugenic movement and his involvement with the elucidation of genetical and evolutionary questions.
Rejected for national service in World War I because of poor eyesight, he taught in public schools for 5 years. During this period he made two scientific advances, notable both for the method and the result: using an argument from n-dimensional geometry, he found the distribution of the correlation coefficient; and, using variance components, he demonstrated that human inheritance was entirely consistent with Mendelian principles. In 1919 he was hired as statistician at Rothamsted Experimental Station, where the work he had already done was immediately applicable. He continued investigations into distributions of statistics which, like the correlation coefficient, were in common use; with small experimental samples, it was essential to know the theoretical distribution in assessing the actual results. At the same time, he responded to the stimulus of actual problems in the field by inventing statistical methods that used all the information in the data to draw correct inferences, and he conceived the whole idea of the design of experiments, by which to gain fuller and more precise information for a given amount of experimental effort. He laid the foundations, coined the language, and developed the methodology of modern biometry.
During the 1930s the influence of his statistical work spread throughout the English-speaking world, and often beyond, so that, after World War II, with the formation of the International Biometric Society in 1947, a new field of scientific research was recognized. As a result of the introduction of quantitative methods and the refinements of inductive reasoning for which Fisher had been primarily responsible, a transformation had taken place in the very way a biologist conceived of his research.
Meanwhile, in 1917, Fisher was married to Ruth E Guinness, younger daughter of Dr Henry Grattan Guinness, a preacher and evangelist who, at the time of his death in 1915, had been head of the missionary training college and the mission his father had started. The Fishers read together in the history of civilizations, noting evidences that the decay of past civilizations had resulted from the pattern of social selection arising in a moneyed economy. Fisher advocated practical measures by which to reduce this deleterious selection in contemporary society and thus save Western Civilization from the fate of its predecessors. Eugenic practice accorded with theory and, for the Fishers, started with the investment of their own physical and economic resources in the next generation. Eventually they raised a family of two sons and six daughters.
In addition, at home Fisher pursued work in population genetics that led to his formulation of The Genetical Theory of Natural Selection (1930) and to his election in 1933 to the Chair of Eugenics at University College, London, where he initiated genetical researches, in particular, in human blood group serology. In 1943 he moved to Cambridge University as Balfour Professor of Genetics. Combining genetical work in the laboratory with evolutionary field work, in collaboration with E B Ford, he demonstrated the reality of the effects of natural selection and developed statistical methods by which they could be measured. He also enjoyed the more theoretical aspects of his work, the theory of inbreeding arising from his work with mice, and the combinatorial aspects of his work with polyploid plants. Logical difficulties in the theory of inductive inference continued to engage him after his retirement in 1957. He died in Adelaide in 1962.
Given its subject, this biography is necessarily scientific. However, I have not assumed that the reader is familiar with any of the sciences involved, and I have avoided mathematical elaboration. The symbols of mathematics are used mainly to clarify logical connections necessary to the argument. I hope that any reader prepared to follow through what is sometimes fairly stiff argument may understand its substance without difficulty. At the same time I have tried to make the account technically correct.
My ignorance of mathematics and statistics would have prevented me from attempting the present work without the help of my husband, George E P Box. Patiently, he has expounded and discussed statistical matters with me and interpreted the mathematics. He has read and corrected several drafts of the manuscript. The errors are mine; much of the clarity of the statistical discussion I owe to him.
I have greatly benefited from the publication of the Collected Papers of R A Fisher (Volumes 1-5), J H Bennett, Ed., University of Adelaide, 1971-1973. These comprise 294 of Fisher's contributions to the scientific literature between 1912 and 1962, with amendments or prefaces Fisher had made in his lifetime, the biographical memoir written by K Mather and F Yates, and a very complete bibliography (included in this volume). Nearly half the collected papers are mainly mathematical or statistical, and most of the others concern topics in genetics, evolution, or eugenics. The broad scope of many of the papers, however, makes any grouping according to subject matter unsatisfactory, and the papers have been arranged chronologically according to year of publication, with the more statistical ones placed ahead of the more genetical. In his Preface, the editor writes, "it is hoped that with the material presented in this way, readers will be better able to appreciate the extraordinary diversity of Fisher's contributions to science and to trace their overall development and interrelationships." The Collected Papers constitute a resource no student of Fisher's work can afford to be without. In this volume, therefore, references in the text are given numerically as they appear in the Collected Papers.
I am deeply indebted to a number of Fisher's friends for their help. Prof J Henry Bennett gave me access to the Fisher papers at Adelaide University and put every facility at my disposal during my 3 months with the Genetics Department there in 1970. He has read and commented on the manuscript, has checked quotations from the papers in Adelaide, and brought source material to my attention. Professor E B Ford recorded for me a priceless account of his long and close association with Fisher and has read and commented on relevant portions of the manuscript. Prof G A Barnard arranged a valuable series of seminars on Fisher's statistical papers during our stay at Essex University in the academic year 1970-1971. He has commented on my discussion of scientific inference. Prof D J Finney has read and commented on a draft of the early chapters, and Prof J F Crow on the chapter on the evolution of dominance.
Many have contributed their own recollections, written records, photographs, and even personal correspondence with Fisher. For such generous assistance I wish to express my gratitude to the following: A E Brandt, Sir Edward Bullard, L L Cavalli Sforza, the late E A Cornish, Gertrude M Cox, A W F Edwards, D J Finney, T N Hoblyn, S B Holt, E Irving, Sir Bernard Keen, R T Leslie, the late P C Mahalanobis and Mrs Mahalanobis, Besse B Mauss, J R Morton, E S Pearson, R R Race, Stuart A Rice, P R Rider, W A Roach, S K Runcorn, the late H Fairfield Smith, Sir Gerard Thornton, Helen N Turner and F Yates. I thank Rothamsted Experimental Station, the Agricultural Research Council, the Genetical Society and the Royal Society for supplying information from their records at my request and K Mather for permission to quote from his contribution to the Royal Society's Biographical memoir of R A Fisher.
Special thanks are due to my aunt, Geraldine Heath (Gudruna), for an account she alone could have given of Fisher during the years 1910 - 1920. The clarity of her perceptions and memories illuminates the pages of the long letters, often hand-written, in which in her mid-eighties she recalled Fisher and commented on my account of him in the first two chapters of this volume. Above all, I thank my mother for opening the gates of sometimes painful memories on my behalf and for her full and enthusiastic discussion of my questions.
JOAN FISHER BOX
JOC/EFR March 2006
The URL of this page is: