MacTutor
R A Fisher
Times obituary
MATHEMATICAL BIOLOGY
Sir Ronald Aylmer Fisher, F.R.S., one of the outstanding mathematical biologists of his time, died in Adelaide on Sunday at the age of 72.
Born on February 17, 1890, the seventh child of G. Fisher, of Robinson & Fisher, auctioneers in St. James's, he was educated at Harrow and Gonville and Caius College, Cambridge, where he graduated as a wrangler in the mathematical tripos in 1913. For two years he was a statistician for the Mercantile & General Investment Company and then, being excluded from military service by his extremely short sight, he was engaged in teaching at Rugby throughout the First World War. In 1919 he joined the staff of the Rothamsted Experimental Station as head of the statistical department, where he remained until 1933.
Here Fisher began the remarkable series of statistical investigations which led to the techniques described in Statistical Methods for Research Workers (1925, 10th edition 1946), The Design of Experiments (1935, fifth edition 1949), and Statistical Tables (published with F. Yates, 1938, third edition 1947). These works revolutionized agricultural research; they described the methods, now used all over the world, for evaluating the results of small sample experiments and for laying out our experimental trials as to minimize the disturbances due to heterogeneity of soils and the unavoidable irregularities of biological material.
In 1933, Fisher was appointed to the Galton Chair of Eugenics at University College London. As an undergraduate, he had been attracted to genetics and biometry and had shown in 1918 that biometric correlations between relatives were to be expected based on the Mendelian theory; and during his time at Rothamsted, he had carried out genetic experiments as a hobby. Bringing his great mathematical ability to bear on these two sciences of biometry and genetics, which in the past had been largely in conflict, perhaps played the chief part in bringing about their present unity. Already in 1930, he had published The Genetic Theory of Natural Selection, which went far toward reconciling the Darwinian ideas of natural selection with Mendelian theory.
His mathematical treatment of natural selection has provided the basis for most modern studies of populations. His theory of the evolution of dominance introduced the notion of the modification of gene action by selection. This marriage of Darwinism with genetics, for which Fisher was largely responsible, is sometimes referred to as "neo-Darwinism".
Fisher's mathematical approach proved particularly valuable in the difficult field of human genetics. Among many other notable contributions, he was responsible for the theory that provides a basis for understanding the Rhesus blood groups in man, which attracted so much attention during the Second World War. In 1943, Fisher accepted an appointment as Arthur Balfour Professor of Genetics at Cambridge, a post which he held until his retirement in 1957.
FIELD CROP EXPERIMENTS
The rise of quantitative biology, which has been such a noteworthy feature of the past 40 years, has been due above all to the work of R. A. Fisher. He combined mathematical skill with biological insight; and the theoretical and practical methods which he developed for the planning of experiments with field crops in agriculture have spread far and wide. The younger biologists of today all make use of such methods for every kind of experimentation: indeed, their influence now extends far beyond the bounds of biology.
As a penetrating thinker, Fisher was outstanding, but his writings are difficult for many readers. Indeed, some of his teachings have been most effectively conveyed by the books of others who have been able to simplify their expression. As a lecturer, also, Fisher was too difficult for the average student; his classes would rapidly fall away until only two or three students who could stand the spectacle remained as fascinated disciples. Nor was he particularly successful as an administrator; he perhaps failed to appreciate the intellectual limitations of the ordinary man. But with his broad interests and penetrating mind, he was a most stimulating and sympathetic conversationalist.
Fisher had been a Fellow of Gonville and Caius College from 1921 to 1927. He was reelected on his return to Cambridge in 1943 and was elected President of his college in 1956. He was elected Fellow of the Royal Society in 1929, receiving a Royal Medal in 1938, the Darwin Medal in 1948, and the Copley Medal, the highest award in the society's gift, in 1956. He also received the Guy Medal in gold from the Royal Statistical Society. He was knighted in 1952. Many foreign universities conferred honors on him; he received honorary doctorates from the universities of Glasgow, Calcutta, Ames, Harvard, London, and Chicago, and he was an honorary foreign member and fellow of the American Academy of Arts and Sciences, the American Philosophical Society, the United States National Academy of Science, the Royal Swedish Academy of Science, the Royal Danish Academy of Science, and the Royal Danish Academy of Science. In 1957, he was honorary president of the International Statistical Institute.
In 1959 he retired as President of his college, and it was in that year that he toured Australia as the guest of the Commonwealth Scientific and Industrial Organization. He decided that Adelaide would be a pleasant place to live, and he returned there the following January. Since then, he had been attached to the C.S.I.R. division of mathematical statistics and had lectured at Adelaide University.
He married Ruth Eileen, daughter of H. Gratten Guinness, in 1917 and had two sons (the elder was killed in action in 1943) and six daughters.
_________________________________________
Professor G. A. Barnard writes:
Sir Ronald Fisher was not only a great mathematical biologist. He made important contributions to physics. But above all, as he himself rightly foresaw, his work on the design and analysis of scientific experiment has application to other sciences, natural and social, and to engineering. If statistics now occupies a central place in the modern applications of mathematics, this is in large measure due to his influence.
He insisted that while scientific inference must of necessity be uncertain, it need not for that reason lack rigour, since the uncertainty itself may be rigorously specifiable. And by this insistence, and the precise mathematical theory he developed to embody it, he rescued statistical inference from the state into which it had fallen, in which it was regarded with suspicion, if not contempt. By given to habits of thinking, those lifetimes of exact templation of various forms of uncertainty and their measurement enabled him to penetrate more deeply than any before him to the nature of this protean concept. And from this insight he was able to formulate principles (especially his "principle of sufficiency") which rank with the contributions of Bernoulli, Bayes, and Boole to scientific logic.
His character and personality were so strong that something must be said of these also. He was capable of tremendous charm, and warmth in friendship. But also he was the victim, as he himself recognized, of an uncontrollable temper; and his devotion to scientific truth as he saw it being literally passionate, he was an im- placable enemy of those whom he judged guilty of propagating error. The very reverse of the cool, calculating scientist of the popular image, his friends, and his enemies, will miss him. Part of the scientific landscape has disappeared, and left the world much duller.
___________________________________________________
Sir John Russell writes:
It would be difficult to find a better example of the help that abstruse sciences, apparently wholly remote from reality, can give in solving important practical problems than is afforded by the work of Ronald Fisher at Rothamsted during his 14 years there.
I had long been anxious to have the masses of data relating to our unique field experiments examined by modern statistical methods to extract information which I felt sure they contained but which our crude methods had missed. Neither Oxford nor Cambridge could then supply a young mathematician able and willing to undertake the work, but I heard of Fisher and found him ready to do so. His tutor's rather lukewarm opinion of him was that "if he had stuck to the ropes he would have made a first-class mathematician, but he would not"—which suggested that he was the man we wanted, and so it soon appeared,
MONUMENTAL INVESTIGATION
He began with some theoretical studies on the development of suitable methods for dealing with agricultural experiments, among them a monumental investigation on the mathematical foundations of theoretical statistics, which was published by the Royal Society and quickly brought him scientific fame. He used his new methods for studying field data and also for improving various researches proceeding in the laboratories, including the estimation of numbers of microorganisms in soils, the counting of aphids, and others.
His now designs for field experiments have proved particularly helpful. It had always been a weakness of the old designs that no good estimate could be made of the validity of the results: no two plots of land provide identical conditions for plant growth, and it was always uncertain how much of a particular result was due to differences of this kind. In the earlier years, this had not mattered much; qualitative effects were often of sufficient interest to justify the work. But by the 1920s, when these investigations were made, much more definite information was required, and the new methods provided it. Simple designs could be used if a moderate amount of uncertainty was permissible, and more complexions where higher accuracy was desired.
A NEW PRINCIPLE
Fisher also introduced a new principle. It had always been assumed that a field experiment should normally deal with one variant only. Fisher showed, however, that better results were obtained by combining two or more variants in the same experiment; this made the design more complex, and practical considerations set limits to the extent to which this "confounding" could go.
Younger colleagues worked out practical methods which are now widely used. They have also led to a marked improvement in the presentation of results, not only in this country but in tropical Africa and elsewhere, and it was with great thrill that we heard Sir Harold Jeffreys declare on an important occasion that, thanks to Fisher's work, "the standard of presentation of results in agriculture is better than in any of the so-called exact sciences—a state of affairs that physicists should cease to tolerate."
Fisher had a wide-ranging mind. He was equally at ease in talking to my small son at tea in our house and in discussing highly abstruse genetic problems with a distinguished scientist.
________________________________________________
Dr. C. I. Bliss writes:
The death of Sir Ronald Fisher removes from the world of science one of its most brilliant minds and colorful personalities. The discovery of a research tool can open the door to new and previously inaccessible areas This is a striking feature of the statistical methods that Fisher devised, which have strongly influenced developments throughout the natural and social sciences. Biological science, faced with the inherent variation in living organisms, presented the kind of problem that started him upon his career. His impact was greater because of contact with real data and he never lost count of the time he spent innumerable hours at his desk calculator testing and checking his ideas.
Despite, or perhaps because of, his mathematical genius, Fisher was not too interested in presenting detailed mathematical proofs of his discoveries, more to the despair of mathematicians than of biologists. I can remember the dismay of his first mathematical assistant at the Galton Laboratory, who was assigned the initial task of familiarizing himself with Fisher's Statistical Methods for Research Workers. Time and again he would come upon statements that seemed based entirely on Fisher's intuition, and he would diligently cover page after page with mathematical derivation, only to end with "Fisher was right".
MATHEMATICAL INTUITION
Fisher's later years were devoted to the more general aspects of inductive reasoning, the subject of his last and most provocative book, Statistical Methods and Scientific Inference. Its final, theoretical chapter is so compactly written that six lectures were needed in a graduate course in mathematics at Yale University to bridge the gaps which Fisher had jumped through his mathematical intuition. His extraordinary insight and originality were by no means restricted to scientific problems, as was at once apparent in conversation. When, for example, a difficult crossword puzzle in The Times resisted the combined efforts of two of his associates in the Senior Combination Room at Caius College, they would appeal to Fisher, who would effort effortlessly supply the missing word.
One consequence of the Fisherian revolution in research methodology, and testimony to its world-wide impact, was the formation 15 years ago of the International Biometric Society, devoted to the mathematical and statistical aspects of biology. As its first president, he watched it grow to its present 2,000 members in 50 different countries, comprising mathematicians, statisticians, biologists in all biological specialties, and representatives of the physical and social sciences. An association of individuals, it crosses the boundaries between nations and between scientific disciplines in living testimony to the unifying concepts that we owe to Sir Ronald Fisher.
Fisher had a wide-ranging mind. He was equally at ease in talking to my small son at tea in our house and in discussing highly abstruse genetic problems with a distinguished scientist.
MATHEMATICAL BIOLOGY
Sir Ronald Aylmer Fisher, F.R.S., one of the outstanding mathematical biologists of his time, died in Adelaide on Sunday at the age of 72.
Born on February 17, 1890, the seventh child of G. Fisher, of Robinson & Fisher, auctioneers in St. James's, he was educated at Harrow and Gonville and Caius College, Cambridge, where he graduated as a wrangler in the mathematical tripos in 1913. For two years he was a statistician for the Mercantile & General Investment Company and then, being excluded from military service by his extremely short sight, he was engaged in teaching at Rugby throughout the First World War. In 1919 he joined the staff of the Rothamsted Experimental Station as head of the statistical department, where he remained until 1933.
Here Fisher began the remarkable series of statistical investigations which led to the techniques described in Statistical Methods for Research Workers (1925, 10th edition 1946), The Design of Experiments (1935, fifth edition 1949), and Statistical Tables (published with F. Yates, 1938, third edition 1947). These works revolutionized agricultural research; they described the methods, now used all over the world, for evaluating the results of small sample experiments and for laying out our experimental trials as to minimize the disturbances due to heterogeneity of soils and the unavoidable irregularities of biological material.
In 1933, Fisher was appointed to the Galton Chair of Eugenics at University College London. As an undergraduate, he had been attracted to genetics and biometry and had shown in 1918 that biometric correlations between relatives were to be expected based on the Mendelian theory; and during his time at Rothamsted, he had carried out genetic experiments as a hobby. Bringing his great mathematical ability to bear on these two sciences of biometry and genetics, which in the past had been largely in conflict, perhaps played the chief part in bringing about their present unity. Already in 1930, he had published The Genetic Theory of Natural Selection, which went far toward reconciling the Darwinian ideas of natural selection with Mendelian theory.
His mathematical treatment of natural selection has provided the basis for most modern studies of populations. His theory of the evolution of dominance introduced the notion of the modification of gene action by selection. This marriage of Darwinism with genetics, for which Fisher was largely responsible, is sometimes referred to as "neo-Darwinism".
Fisher's mathematical approach proved particularly valuable in the difficult field of human genetics. Among many other notable contributions, he was responsible for the theory that provides a basis for understanding the Rhesus blood groups in man, which attracted so much attention during the Second World War. In 1943, Fisher accepted an appointment as Arthur Balfour Professor of Genetics at Cambridge, a post which he held until his retirement in 1957.
FIELD CROP EXPERIMENTS
The rise of quantitative biology, which has been such a noteworthy feature of the past 40 years, has been due above all to the work of R. A. Fisher. He combined mathematical skill with biological insight; and the theoretical and practical methods which he developed for the planning of experiments with field crops in agriculture have spread far and wide. The younger biologists of today all make use of such methods for every kind of experimentation: indeed, their influence now extends far beyond the bounds of biology.
As a penetrating thinker, Fisher was outstanding, but his writings are difficult for many readers. Indeed, some of his teachings have been most effectively conveyed by the books of others who have been able to simplify their expression. As a lecturer, also, Fisher was too difficult for the average student; his classes would rapidly fall away until only two or three students who could stand the spectacle remained as fascinated disciples. Nor was he particularly successful as an administrator; he perhaps failed to appreciate the intellectual limitations of the ordinary man. But with his broad interests and penetrating mind, he was a most stimulating and sympathetic conversationalist.
Fisher had been a Fellow of Gonville and Caius College from 1921 to 1927. He was reelected on his return to Cambridge in 1943 and was elected President of his college in 1956. He was elected Fellow of the Royal Society in 1929, receiving a Royal Medal in 1938, the Darwin Medal in 1948, and the Copley Medal, the highest award in the society's gift, in 1956. He also received the Guy Medal in gold from the Royal Statistical Society. He was knighted in 1952. Many foreign universities conferred honors on him; he received honorary doctorates from the universities of Glasgow, Calcutta, Ames, Harvard, London, and Chicago, and he was an honorary foreign member and fellow of the American Academy of Arts and Sciences, the American Philosophical Society, the United States National Academy of Science, the Royal Swedish Academy of Science, the Royal Danish Academy of Science, and the Royal Danish Academy of Science. In 1957, he was honorary president of the International Statistical Institute.
In 1959 he retired as President of his college, and it was in that year that he toured Australia as the guest of the Commonwealth Scientific and Industrial Organization. He decided that Adelaide would be a pleasant place to live, and he returned there the following January. Since then, he had been attached to the C.S.I.R. division of mathematical statistics and had lectured at Adelaide University.
He married Ruth Eileen, daughter of H. Gratten Guinness, in 1917 and had two sons (the elder was killed in action in 1943) and six daughters.
_________________________________________
Professor G. A. Barnard writes:
Sir Ronald Fisher was not only a great mathematical biologist. He made important contributions to physics. But above all, as he himself rightly foresaw, his work on the design and analysis of scientific experiment has application to other sciences, natural and social, and to engineering. If statistics now occupies a central place in the modern applications of mathematics, this is in large measure due to his influence.
He insisted that while scientific inference must of necessity be uncertain, it need not for that reason lack rigour, since the uncertainty itself may be rigorously specifiable. And by this insistence, and the precise mathematical theory he developed to embody it, he rescued statistical inference from the state into which it had fallen, in which it was regarded with suspicion, if not contempt. By given to habits of thinking, those lifetimes of exact templation of various forms of uncertainty and their measurement enabled him to penetrate more deeply than any before him to the nature of this protean concept. And from this insight he was able to formulate principles (especially his "principle of sufficiency") which rank with the contributions of Bernoulli, Bayes, and Boole to scientific logic.
His character and personality were so strong that something must be said of these also. He was capable of tremendous charm, and warmth in friendship. But also he was the victim, as he himself recognized, of an uncontrollable temper; and his devotion to scientific truth as he saw it being literally passionate, he was an im- placable enemy of those whom he judged guilty of propagating error. The very reverse of the cool, calculating scientist of the popular image, his friends, and his enemies, will miss him. Part of the scientific landscape has disappeared, and left the world much duller.
___________________________________________________
Sir John Russell writes:
It would be difficult to find a better example of the help that abstruse sciences, apparently wholly remote from reality, can give in solving important practical problems than is afforded by the work of Ronald Fisher at Rothamsted during his 14 years there.
I had long been anxious to have the masses of data relating to our unique field experiments examined by modern statistical methods to extract information which I felt sure they contained but which our crude methods had missed. Neither Oxford nor Cambridge could then supply a young mathematician able and willing to undertake the work, but I heard of Fisher and found him ready to do so. His tutor's rather lukewarm opinion of him was that "if he had stuck to the ropes he would have made a first-class mathematician, but he would not"—which suggested that he was the man we wanted, and so it soon appeared,
MONUMENTAL INVESTIGATION
He began with some theoretical studies on the development of suitable methods for dealing with agricultural experiments, among them a monumental investigation on the mathematical foundations of theoretical statistics, which was published by the Royal Society and quickly brought him scientific fame. He used his new methods for studying field data and also for improving various researches proceeding in the laboratories, including the estimation of numbers of microorganisms in soils, the counting of aphids, and others.
His now designs for field experiments have proved particularly helpful. It had always been a weakness of the old designs that no good estimate could be made of the validity of the results: no two plots of land provide identical conditions for plant growth, and it was always uncertain how much of a particular result was due to differences of this kind. In the earlier years, this had not mattered much; qualitative effects were often of sufficient interest to justify the work. But by the 1920s, when these investigations were made, much more definite information was required, and the new methods provided it. Simple designs could be used if a moderate amount of uncertainty was permissible, and more complexions where higher accuracy was desired.
A NEW PRINCIPLE
Fisher also introduced a new principle. It had always been assumed that a field experiment should normally deal with one variant only. Fisher showed, however, that better results were obtained by combining two or more variants in the same experiment; this made the design more complex, and practical considerations set limits to the extent to which this "confounding" could go.
Younger colleagues worked out practical methods which are now widely used. They have also led to a marked improvement in the presentation of results, not only in this country but in tropical Africa and elsewhere, and it was with great thrill that we heard Sir Harold Jeffreys declare on an important occasion that, thanks to Fisher's work, "the standard of presentation of results in agriculture is better than in any of the so-called exact sciences—a state of affairs that physicists should cease to tolerate."
Fisher had a wide-ranging mind. He was equally at ease in talking to my small son at tea in our house and in discussing highly abstruse genetic problems with a distinguished scientist.
________________________________________________
Dr. C. I. Bliss writes:
The death of Sir Ronald Fisher removes from the world of science one of its most brilliant minds and colorful personalities. The discovery of a research tool can open the door to new and previously inaccessible areas This is a striking feature of the statistical methods that Fisher devised, which have strongly influenced developments throughout the natural and social sciences. Biological science, faced with the inherent variation in living organisms, presented the kind of problem that started him upon his career. His impact was greater because of contact with real data and he never lost count of the time he spent innumerable hours at his desk calculator testing and checking his ideas.
Despite, or perhaps because of, his mathematical genius, Fisher was not too interested in presenting detailed mathematical proofs of his discoveries, more to the despair of mathematicians than of biologists. I can remember the dismay of his first mathematical assistant at the Galton Laboratory, who was assigned the initial task of familiarizing himself with Fisher's Statistical Methods for Research Workers. Time and again he would come upon statements that seemed based entirely on Fisher's intuition, and he would diligently cover page after page with mathematical derivation, only to end with "Fisher was right".
MATHEMATICAL INTUITION
Fisher's later years were devoted to the more general aspects of inductive reasoning, the subject of his last and most provocative book, Statistical Methods and Scientific Inference. Its final, theoretical chapter is so compactly written that six lectures were needed in a graduate course in mathematics at Yale University to bridge the gaps which Fisher had jumped through his mathematical intuition. His extraordinary insight and originality were by no means restricted to scientific problems, as was at once apparent in conversation. When, for example, a difficult crossword puzzle in The Times resisted the combined efforts of two of his associates in the Senior Combination Room at Caius College, they would appeal to Fisher, who would effort effortlessly supply the missing word.
One consequence of the Fisherian revolution in research methodology, and testimony to its world-wide impact, was the formation 15 years ago of the International Biometric Society, devoted to the mathematical and statistical aspects of biology. As its first president, he watched it grow to its present 2,000 members in 50 different countries, comprising mathematicians, statisticians, biologists in all biological specialties, and representatives of the physical and social sciences. An association of individuals, it crosses the boundaries between nations and between scientific disciplines in living testimony to the unifying concepts that we owe to Sir Ronald Fisher.
Fisher had a wide-ranging mind. He was equally at ease in talking to my small son at tea in our house and in discussing highly abstruse genetic problems with a distinguished scientist.