Obituaries: Frank Anscombe; Bernard Benjamin; David Champernowne; Rob Kempton; Derek Maunder; David Wishart

Frank Anscombe was born in Sussex, read Mathematics at Cambridge and during World War 2 worked in the Ministry of Supply, partly on ordnance problems and partly with SR17 (a group within the Ministry of Supply dealing with statistical issues particularly in the context of quality control). His first published paper was on sequential rectifying inspection, developed out of G. A. Barnard's pioneering work at SR17 on sequential sampling. He returned to this theme later. In 1946 Anscombe went for 2 years to Rothamsted Experimental Station, working in the department that was headed by F. Yates. In 1948 followed appointment as a Lecturer in Mathematics, based in the newly formed Statistical Laboratory, headed by J. Wishart, Reader in Agricultural Statistics. There Anscombe joined H. E. Daniels; shortly afterwards D. V. Lindley was appointed to the Laboratory and 2 years later the present writer.

During this period Anscombe was very active in research. Among other material, he published definitive accounts of estimation in the negative binomial distribution and an important paper on sequential estimation (as contrasted with sequential testing) and made a valiant effort to synthesize the model-based and randomization-based analyses of standard experimental designs. At seminars he could be counted on to ask penetrating questions on a broad front.

In 1953 he went on sabbatical leave to Princeton and worked with J. W. Tukey. Shortly before his return to Cambridge he married Phyllis Rapp, in fact Tukey's sister-in-law.

After 2 years back in Cambridge the couple returned to Princeton and then a few years later moved to Yale, where Anscombe spent the rest of his career. On his return to Princeton he published on outliers and in 1961 a definitive account of the formal properties of residuals in normal theory linear regression. This paper, published in the Proceedings of the Berkeley Symposium, is probably his best-known work. He also published a critique of sequential trials, especially in a medical context, partly because of conflict with a likelihood principle, and suggested a decision-oriented approach in which the objective is to treat optimally as many patients as possible. Anscombe then developed a major interest in statistical computing and in particular in APL, a package which attracted quite widespread enthusiasm. Although APL has disappeared into the fossil record, it had appreciable influence for a period. Unfortunately Anscombe's very scholarly book on the topic probably appeared too late to have much impact.

Soon after appointment to Cambridge, Anscombe became internal examiner for the Diploma in Mathematical Statistics. This was under a system in which the examiner set the questions without consulting the relevant lecturers and the questions were at least as important in establishing the work programme of future students as in an assessment of current students. He decided, correctly, that the previous questions were pedestrian and that he must make the new ones `interesting'. He was extremely successful in this although at the cost of some stress to students and in subsequent years to supervisors. Indeed I believe that it was many years before a collective of students proved, what had been widely believed, that one of the questions was indeed totally insoluble. Harmless eccentricities were perhaps more tolerated (or even expected) of university teachers in those days than now. Anscombe had the endearing habit of showing his pleasure at the completion of a proof or line of argument by clapping his hands and saying enthusiastically `splendid, splendid'. From time to time students kept data on the number of times that he did this and gratifyingly high counts were sometimes achieved in a 50-minute lecture.

During his period at Cambridge, four doctoral students completed their doctoral theses under his supervision. By a spectacular piece of maladministration by the university one of the theses was rejected, even though it was widely thought an impressive piece of work. The candidate in question continued to become a successful and much respected figure in our field.

To my regret I rarely saw Anscombe after the early 1960s. It will be clear from the above, based largely on my contact with him from 1950 onwards, that he left a clear impression of a lively and enthusiastic colleague, with very wide interests in research and beyond, and who was unfailingly considerate and helpful to his colleagues.

He is survived by Phyllis and their four children.

D. R. Cox

Bernard Benjamin had a distinguished career as a statistician, actuary and demographer in public service, achieving distinction and the highest levels of recognition and honour in each of these fields.

Benjamin was born in 1910, the youngest of eight children. He was educated at Colfe's Grammar School in Lewisham and went on to study physics part time at Sir John Cass College (then affiliated to the University of London), graduating in 1933. He started work in 1928 as an actuarial assistant to the London County Council pension fund and qualified as a Fellow of the Institute of Actuaries in 1941. From 1936, he worked as a statistician in the public health sections of London County Council until 1943 when he joined the Royal Air Force to con-tinue his work as a statistician. After the war, he returned to London County Council and public health and undertook his doctorate (also part time) on the analysis of tuberculosis mortality.

In 1952, Benjamin was appointed Chief Statistician at the General Register Office, marking his move into demography and from technical analysis to management and leadership of a major public sector department. After 11 years at the General Register Office, he was appointed Director of Statistics at the Ministry of Health in 1963 and then in 1966 became the Greater London Council's first Director of the Intelligence Unit. The Unit was set up as part of local government reorganization in London and its task was to make sure that the Greater London Council had `economic and other information at the right time in the right way', as Benjamin was quoted in the Daily Telegraph. In this role, he brought together the entire planning and transportation research staff into a cohesive and effective unit. But this reorganization did not survive without Benjamin's leadership and drive after his retirement in 1970.

A strong theme of Benjamin's later working life was a series of retirements followed by new beginnings. Thus, in 1970, he became Director of Statistical Studies at the newly established Civil Service College. And then in 1973 he retired and joined City University as the Foundation Professor of Actuarial Science and established and designed the first Bachelor of Science programme in actuarial science in the country. The first intake of 10 was in 1974, but numbers soon reached a healthy level. Although Benjamin enjoyed teaching undergraduates, he took particular pleasure in the personal nature of supervising doctoral students and, over the next decade or so, he supervised a steady stream of students working on statistical methods applied to demography and actuarial science. One of his great interests was the modelling of non-life-insurance claims and (with the late R. E. Beard) he was one of the UK's pioneers in this field.

In each of these roles, Benjamin was concerned with the collection and analysis of statistics and the presentation of results for practical use in public health or demographic or insurance management. He was particularly adept at conveying statistical ideas in a clear manner, without recourse to jargon or indeed mathematical notation.

Benjamin retired in 1975 from City University but continued as a Visiting Professor. In the early 1980s he recognized that, as a discipline, actuarial science needed to build strong links not only with statistics but also with business and finance. With the support of what is now the Association of British Insurers, he set up the Centre of Research, Insurance and Investment within City University's Business School. This was an imaginative idea. Although the original timing was not propitious, this development came to fruition a few months after his death when City's Department of Actuarial Science and Statistics finally joined the Cass Business School in August 2002. On final departure from City University, he was appointed Emeritus Professor and awarded an honorary Doctor of Science degree for his contributions to education and research in statistics and actuarial science.

Benjamin's scientific work and interests were extensive. He published over 100 papers in statistical, actuarial and demographic journals over almost 40 years. A notable achievement was the team that he led which produced the 1954 report on the growth of pension rights and their effect on the national economy, which became the actuarial profession's principal evidence to the `Phillips Committee' on the economic and financial problems for the provision for old age. When the actuarial profession sought to update this landmark report in the mid-1980s (at a time when the Government was questioning the role of public pension provision), it was to Benjamin that they turned to lead the group that produced `Pensions: the problems of today and tomorrow' in 1987.

Benjamin could write concisely and interestingly and his first drafts were almost of final draft quality. This talent contributed to a series of successful text-books and monographs: Elements of Vital Statistics (1959), Social and Economic Differentials in Fertility (1965), Social and Economic Factors affecting Mortality (1965), Health and Vital Statistics (1968), Demographic Analysis (1968) and Population Statistics (1989). The text-books Analysis of Mortality and other Actuarial Statistics (1970, with follow-up editions in 1980 and 1993) and General Insurance (1977) have become seminal works internationally in the actuarial field. His last book, Mortality on the Move (1993), appeared at a time of accelerating decline in mortality in many industrialized countries and is now widely cited.

Among his honours and achievements, Benjamin was UK representative on the United Nations Population Commission from 1955 to 1963, Honorary Consultant in Medical Statistics to the army, a member of the Statistics Committee of the Social Science Research Council, Secretary General of the International Union for the Scientific Study of Population from 1962 to 1963, Vice-President (1963–1966) and President of the Institute of Actuaries (1966–1968), Honorary Secretary (1956–1963) and President of the Royal Statistical Society (1970–1972). As well as being Presidents of both the Institute of Actuaries and the Royal Statistical Society, he was also awarded the highest honours of both bodies—respectively the Gold Medal (1975) and Guy Medal in Gold (1986).

At a personal level, Benjamin was modest and self-effacing, yet he was a determined and clear-sighted manager and leader who inspired respect and loyalty. As a colleague, he was both encouraging and supportive, qualities which made him both an excellent doctoral supervisor and research collaborator.

He married Mary in 1937 and is survived by their two daughters: Anne and Margaret. He had many hobbies before his sight failed. He would describe himself as an `amateur' pianist and painter. The latter activity gave him much pleasure and his water-colours were of a higher standard than he would admit to.

Steven Haberman

David Champernowne, who died on August 22nd, 2000, of bronchial pneumonia, was Professor of Economics and Statistics in the University of Cambridge and contributed substantially to the literature on mathematics and on theoretical and applied economics as well as on statistics.

Champernowne was born on July 9th, 1912, in Oxford, son of Francis Gawayne and Isabel Mary Champernowne. He went to Winchester College as a scholar and in 1931 won a scholarship in mathematics to King's College, Cambridge. He completed the mathematics tripos in 2 years, obtaining double first-class Honours. With the encouragement of John Maynard Keynes he then switched to the economics tripos and again obtained a first-class degree.

As an undergraduate he published a paper on `normal numbers', in which he was the first to produce an example of such a number in base 10. This example has come to be known as `Champernowne's constant'; it is obtained by concatenating positive integers from 1 upwards and interpreting them as decimal digits to the right of the decimal point: 0.123456789101112….

After graduating, Champernowne was Assistant Lecturer at the London School of Economics (1936–1938) and then university Lecturer in Statistics at Cambridge (1938–1940). During this time his applied research concentrated on the determinants of the distribution of UK unemployment during the slump. However, his major intellectual contribution arose from theoretical work focusing on the size distribution of incomes among people. Champernowne's path breaking dissertation on the income-generating process showed how the evolution of an income and wealth distribution could be appropriately represented by a Markovian model of income mobility and why the equilibrium distribution would conform to a shape that is characterized by a specific functional form related to the Pareto distribution. The work resulted in a prize fellowship at King's College in 1937. The model was eventually published in the Economic Journal in 1953 although the full version of the fellowship dissertation was not published for another 20 years. It laid the foundations for the application of stochastic process models to the analysis of income distributions.

In the Second World War Champernowne served with Lindemann as Assistant in the statistical section of the Prime Minister's office (1940–1941) and then worked with Jewkes at the Ministry of Aircraft Production's Department of Statistics and Programming. In 1945 he returned to academia and became a Fellow of Nuffield College, Oxford, and Director of the Oxford Insti- tute of Statistics. He was appointed Professor of Statistics in 1948 and in the same year married Wilhelmina Dullaert who worked at the Institute.

While at Oxford he pursued his pre-war interest in Frank Ramsey's theory of probability: this led Champernowne to work on the application of Bayesian analysis to autoregressive series at a time when the Bayesian approach was intellectually unfashionable. This line of research resulted in several papers in the Journal of the Royal Statistical Society and elsewhere and, in some respects, anticipated the work of Granger and Newbold in the 1970s. It also led to Uncertainty and Estimation in Economics; this trilogy, described by reviewers as a `monumental work', addressed the meaning, interpretation and effects of uncertainty within the social sciences.

Champernowne moved to a Readership in Economics at Cambridge in 1959 and was promoted to a personal chair in 1970, the year that he was made a Fellow of the British Academy. He also acted as one of three co-editors of The Economic Journal from 1971 to 1975. His research in economics included contributions to the theory of capital, models of multisector growth, the methodology of national income estimation and the measurement of economic inequality.

After retirement in 1978 he completed the monograph Economic Inequality and Income Distribution (with Frank Cowell of the London School of Economics) that brought together several of his lifelong intellectual concerns: the questioning of fundamental assumptions underlying market-oriented theories of production and distribution, the application of mathematical modelling to the analysis of income distribution, the desire for theoretical and statistical rigour in explaining the fundamentals of economic inequality, the application of numerical methods to solve analytically intractable problems and the concern for distributive justice.

In 1995 he and his wife moved to Budleigh Salterton where he died in 2000.

F. A. Cowell

Rob Kempton, Director of Biomathemtics and Statistics Scotland since its inception in 1987, died of a heart attack while on a cycling holiday in North Yorkshire, on Sunday, May 11th, 2003, at 56 years of age. He was President-Elect of the International Biometric Society (IBS) and a Fellow of the Royal Society of Edinburgh.

Rodney Alistair Kempton, or Rob as he was known by all, was born in Isleworth, Middlesex, in 1946. After attending Chislehurst and Sidcup Grammar School, he read mathematics at Wadham College, Oxford, graduating in 1968. He followed this by a Bachelor of Philosophy degree at Oxford in applied statistics, an almost unique qualification as the course was discontinued after just 1 year.

Rob's first job was as a statistician at Rothamsted Experimental Station. Here he laid the foundations of his lifelong enthusiasm for biometry and the opportunity that it provided for his involvement in a wide range of applications in the life sciences. He established successful collaborations in entomology and nematology, and published a series of papers on the diversity of species. Later this led to a book, written jointly with Pete Digby, on Multivariate Analysis of Ecological Communities.

Rob was appointed Head of Statistics at the Plant Breeding Institute in Cambridge in 1976. His subsequent contributions to the design and analysis of experiments with spatial trends and treatment carry-over effects were stimulated by his observations of plant breeding trials. He played a key role in developing spatial methods for analysing such experiments, including a landmark paper with Julian Besag in Biometrics in 1986. The book that he edited with P. N. Fox, Statistical Methods for Plant Variety Evaluation, encapsulated, in Rob's typically clear and concise style, many of the statistical good practices in plant breeding.

In 1986, Rob moved to Edinburgh as founding director of the Scottish Agricultural Statistics Service. This brought together a network of statisticians supporting agricultural research organizations in Aberdeen, Ayr, Dundee and Edinburgh. He worked energetically to ensure that the benefit of this sizable collection of specialists was fully realized. These were difficult times for public sector scientific research, the need for which seemed constantly to be questioned by Government, and for statisticians, in particular, as research priorities moved away from subjects with which statisticians were traditionally associated. Yet Rob's argument, that scientific research needed to be underpinned by research level statisticians, was accepted. He saw the changing priorities for scientific research as an opportunity rather than a threat. His vision took the Scottish Agricultural Statistics Service from being a purely statistical organization to encompassing the disciplines of mathematical modelling and bioinformatics, with a broadening of application areas from agriculture to the environment, food, health and risk. Associated with this change in activities came, in 1995, a change of name to Biomathematics and Statistics Scotland. The resulting organization is internationally regarded as a model for how to harness the potential of statistics and mathematics to improve the quality and effectiveness of scientific research.

Rob continued to pursue his research interests in experimental design and published jointly with doctoral students. He also established himself as an expert in the statistical analysis of risk and led a group studying methods for assessing potential health risks from food, including pesticides, microbial organisms and genetically modified organisms. He was on the Food Standards Agency's Working Group on Risks from Mixtures of Pesticides, external reviewer of the Agency's food safety programme and a member of the Risk Analysis Committee of the International Statistical Institute.

Rob also served on many other committees and review groups, including the Royal Statistical Society's Council. He was a strong supporter of the IBS and gave many years of service to its work, as British Region Secretary and President, and on IBS committees. His proudest achievement was a scheme that he initiated in the late 1980s to support East Africans with membership of the IBS. This helped to establish IBS groups in Africa and led to the formation of the Sub-Saharan African Network of Biometricians. He had recently been elected Vice-President of the IBS, to serve as President 2004–2005, and was already starting to plan his Presidential address, to have been delivered at the International Biometric Conference in 2004.

Rob married Annelise Sorensen in 1972, and they have a daughter and twin sons. He was a founder member of the Edinburgh branch of Woodcraft Folk, a youth movement that is dedicated to building a world based on equality, friendship and peace. He was a keen walker and cyclist who loved the Scottish hills, and he enjoyed collecting antiquarian books, particularly of Robert Louis Stevenson.

Rob was a modest, generous and kind man, who focused on people's strengths and brought out the best in them. He enjoyed life and he particularly enjoyed being a biometrician. As he recently wrote for the Royal Statistical Society's careers Web site

Just a week before his unexpected death, Rob bought a book of John Clare's poetry, who's final poem, `I am!', he read aloud to his son Ben, whom he was visiting in London. Ben read the final verse again at Rob's funeral service:

Chris Glasbey, David Elston and Mike Talbot

Professor Derek Maunder, who died in April 2003 in Pietermaritzburg, was Professor of Economic and Social Statistics at the University of Exeter from 1971 until his retirement in 1981.

He graduated in economics from the London School of Economics in 1950, with statistics as his special subject. After a period at the University of Southampton he moved to the Institute of Social and Economic Research at the University College of the West Indies, where he was awarded his doctorate in economic statistics by the University of London; his thesis on `Employment in an economically underdeveloped area' was later published. He became a Lecturer, then a Senior Lecturer, at the University of Hong Kong. He returned to the London School of Economics in 1963, publishing books on Hong Kong Urban Rents and A Bibliography on Index Numbers. The growth of statistics and econometrics in the Department of Economics at Exeter led to the establishment of the Chair to which he was appointed.

His interest in economics developed from a lifelong commitment to economics and statistics as means for the improvement of people's lives and living conditions rather than from enthusiasm for highly abstract economic or econometric theory, however mathematically satisfying in itself. He was notably clear sighted in his early recognition of the importance of computing as an integral part of teaching and learning statistics in economic contexts.

He appreciated the essential role of statistics in assessing economic problems, in tackling them and in reviewing the success of economic policies. Although statistical methods depend on reliable data, there was then no useful guide to their availability. Earlier work by Maurice Kendall was quite out of date. A few standard statistical sources—the Monthly Digest, the Annual Abstract, the `Blue book', etc.—were of undoubted value but failed to comprehend the full resources that were available to the statistician. He proved to be the ideal person to edit the remarkable series of 26 volumes, many covering more than one fairly narrow topic, of lengthy and discursive Reviews of UK Statistical Sources published under the aegis of the Royal Statistical Society and the Economic and Social Research Council; the first three volumes were awarded the McColvin Medal of the Library Association in 1974. Reviewing volume XXII in this journal, P. Curwen rightly wrote that

These reviews continued until funding for them ceased. Though themselves now obsolete, their achievement continues through the periodic publication by the Central Statistical Office, now the Office for National Statistics, of the Guide to Official Statistics covering a wider range than its title indicates.

His gentle kindness, sympathy and patience towards his Exeter colleagues and students were legendary and reflected his own happy family life. This was shattered with the murder of his daughter, in especially terrible circumstances, from which he never fully recovered. He took early retirement, though continuing to edit the Reviews of Statistical Sources.

His concern for world poverty led him to establish a trust fund which helps the education of young children in developing countries. He is survived by his wife and five children.

Frank Oliver

Dr D. M. G. Wishart spent his academic career in mathematics and statistics at the University of Birmingham—where, coincidentally, his father had been a Lecturer in Mechanical Engineering. A great raconteur, he loved teaching. In his lectures, the main stream of explanation was filled out with links to other parts of mathematics, and to the world beyond. It is such links that give strength and depth to understanding; the students, particularly the more able, appreciated this. Though primarily a statistician, David could and did teach courses across mathematics, pure and applied. His breadth of knowledge and vision were recognized when he was the first non-professor to be appointed Chair of the School of Mathematical Sciences at Birmingham, bridging the disparate specialist areas with understanding and diplomacy.

David Wishart was born in 1928 in Stockton-on-Tees. His father, William, was then an engineer with Imperial Chemical Industries. His mother, Nelly, was one of the first women to be awarded a doctorate in Scotland, in physical chemistry. David went from Oundle to St Andrew's to study chemistry but changed to mathematics. He won an English Speaking Union scholarship to Princeton, studying with David Kendall; his doctoral thesis is on applications of probability theory to queuing. The range and quality of conversation that he met at Princeton broadened the already wide interests that were his main characteristic. Quotes from William Feller were a favourite reminiscence.

We live in the age of the research assessment exercises, where the prime academic currency is short original papers that each add a brick to the edifice of understanding in a specialist field. David, although he showed that he could do this, was an academic of a different type—a scholar, perhaps, rather than a researcher. This made him an ideal Editor in the Journal of the Royal Statistical Society team and a perceptive reviewer in Russian for Mathematical Reviews. But his concern was broader than statistics, broader than mathematics. He was knowledgeable across the whole span of culture.

He was a specialist in the prints of John Hamilton Mortimer, a student of Egyptian hieroglyphics, an active member of the Wynkyn de Worde Society of printers and publishers and yet, centrally, a scientist with an encyclopaedic knowledge of mathematics. He was a citizen, active in politics without ever becoming a politician. 18th Century Man is a devalued label—people are called `polymaths' who do not even know `maths', let alone much of science, or philosophy, or literature, or art or music. Even in the 18th century no-one could encompass the whole realm of human knowledge but there were people who could converse with substantial knowledge across this range. David Wishart was one of these in our time. Living in the Midlands, he would have been at home in the Lunar Society, talking with Erasmus Darwin, Matthew Boulton, Joseph Priestley and their friends.

Because of his range, David helped me, a mathematical physicist, to create a new kind of course—on `Broad spectrum applied mathematics'. In Britain, thanks to Newton, applied mathematics has been a constant element in all our mathematics courses, but one which was long dominated by Newton's laws of motion—looking at bouncing balls, gyroscopes, the lift and drag of aircraft wings and such things. We thought that our students should have a chance to see something of mathematical modelling across science, engineering, medicine, economics, linguistics and politics. They could then choose to study specialized final year courses in some of those areas. There was only one problem—who could teach such a course? I knew that I could not do it on my own; I could cope with perhaps a third of the field, on the science and engineering side. Who could cover the rest of such a large expanse? David filled the frame with a zest that fired the students to embrace our quite unreasonable demands, and to bring them to success.

In recognition of his long and varied service to the Royal Statistical Society, particularly as `read' papers editor for Series B, he was awarded the 1989 Chambers Medal of the Society.

His editing work led him into printing itself. On a visit to the printing works producing the Society's journals, he became intrigued by the processes. In 1966, he bought his first press and some type. Over time, and particularly in his `retirement', this developed into the Hayloft Press—a collection of seven printing machines, with more than 100 founts. There he printed elegant editions of 14 books and pamphlets, including a reprint of Tom Paine's The Last Crisis and Elizabeth Craik's new translation of Stobaeus's The Seven Deadly Sins in parallel Greek and English. Wynkyn de Worde, Caxton's assistant who first made printing commercial, would have been proud, particularly of the `rubrication', the large red initial letters at the beginning of each section. When Oxford University Press completed the move to computer typesetting, they sold and David bought the complete set of type for the Egyptian hieroglyphs. This led to a fascination with Egyptology—typically, he learned enough to be invited to lecture on it.

His interest in art, literature, music and politics fed his many contributions to the local community. He was on committees for the Birmingham Chamber Music Society, the Birmingham Bibliographical Society and the Birmingham Consumer Group. He served on various community health councils. He and his wife Eva, who helped to make this rich life possible, have two children, Adam and Benita, and two grandchildren. At home, David built a remarkable library. He was an active patron of musicians and local artists, attending their concerts and buying their paintings.

The range of the man was remarkable; the conversation unceasingly illuminating. The boom of his laugh echoes on.

Hugh Burkhardt