Obituaries

Peter William Meredith John passed away on January 22nd, 2015, at the age of 91years. He will be remembered for his many contributions to the design of experiments, with applications in a variety of fields. Peter John was born on August 20th, 1923, in Porthcawl, Wales. He attended local schools through middle school, and then in 1937 won a scholarship to Hereford Cathedral School. Because the local Welsh schools did not offer Greek, he was not qualified for the classical curriculum at Hereford, and so he entered the science curriculum. The Hereford mathematics master was outstanding, and in 1941 John won a scholarship in mathematics to Jesus College, Oxford. After 2 years at Oxford reading for a wartime degree, he enlisted in the Royal Air Force in1943 as a university student. In light of his mathematical background, he was assigned to spend several more months at Oxford intensively learning advanced physics; he then used that knowledge to work on technical problems in support of the war effort. He completed his Bachelors degree in 1944 and began full-time military service. After ending military service in 1946, he continued at Oxford utilizing his veteran’s benefits, receiving his Master of Arts degree in mathematics in 1948. The only jobs available for mathematicians in the UK in 1948 were as school masters, which did not interest him at the time. So he used his remaining year of veteran’s benefits at Oxford to pursue a new post-graduate diploma programme (now called the Master of Science) in statistics. Unfortunately, the UK job market was still no better in 1949 than in 1948, so as an adventure he took a temporary job as instructor in the USA at the University of Oklahoma, teaching calculus to veterans. After a year as an instructor, he decided to stay at Oklahoma for further graduate study in mathematics, specializing in probability. Because of his statistics background, he was offered a position one summer interviewing for a door-to-door survey. One door was opened by a bright, attractive history graduate student, Elizabeth Ann Harper. A courtship ensued, and the couple married in 1954. Peter John received his doctorate in 1955, writing a dissertation on birth-and-death processes. Elizabeth finished her doctorate in 1957 and became a well-respected scholar of American Indian and Spanish history in the American south-west as well as Peter John’s wife and companion of more than 60 years. After receiving his doctorate Peter John accepted a position as Assistant Professor at the University of New Mexico, teaching statistics, but he left after 2 years to accept a position as Research Statistician at the Chevron Research Corporation in the San Francisco Bay area. One attraction was that Henry Scheffé was consultant there. In effect, John had 4 years of post-doctoral study in statistics with Scheffé, supplemented by the summer Gordon conferences. In the last three of his years at Chevron, John was also a visiting Assistant Professor of Statistics at the University of California at Berkeley. Those were exciting times, including programming the then new computers to do regression, expanding the methodology of analysis of variance and experimental design, and learning about response surfaces. Next followed a 6-year tenured position in the mathematics department at the University of California at Davis, which was a campus prominent in agricultural research. Just as he had worked closely with chemists and engineers at Chevron, John worked closely with agronomists, geneticists and food scientists at Davis. Their problems required experimental designs that were entirely different from those needed in the oil industry—and the challenge of translating statistics methodology from one applied field to another, seemingly unrelated, field fascinated him. In 1967, John accepted a professorship in the mathematics department at the University of Texas at Austin, where some of the world’s finest research archives for Elizabeth’s academic speciality were located. His book Statistical Design and Analysis of Experiments was first published by Macmillan in 1971. It was one of the first in the field to use matrix methods and notation, and was republished in 1998 in the Society of Industrial and Applied Mathematics’s series ‘Classics in applied mathematics’. He continued working with the Gordon conferences and pursued research in incomplete-block designs, publishing a book on that topic in 1980. In the 1980s, his talents with experimental design led to work with the semiconductor industry and the quality assurance movement, resulting in another book, Statistical Methods in Engineering and Quality Assurance, in 1990, as well as a fruitful and engaging collaboration with the Sematech research consortium, which had located in Austin. He continued to publish through 2003 (his Erdos number is 2) and served as Associate Editor of several statistical journals. Peter John was well recognized for his contributions to statistics, being elected a Fellow of the Royal Statistical Society in 1949, a Fellow of the American Statistical Association (1976) and of the Institute of Mathematical Statistics (1977). In 1991, he was awarded the Shewell Prize of the American Society for Quality Control. In 1995, he received the Don Owen Award from the San Antonio Chapter of the American Statistical Association. In 2003, he was the honoree and keynote speaker at the Quality and Productivity Research Conference. While at the University of Texas, he developed a reputation as an excellent teacher. His courses were sprinkled with fascinating anecdotes about the development of the topics that he taught (as well as their developers). He supervised 11 doctoral students at Texas (as well as one at the University of California at Davis) and more than 40 Masters students. In 1999, he received the University of Texas’s Award for Outstanding Teaching in the Graduate School. After their children were grown, Peter and Elizabeth visited Spain for a month or more each spring for nearly 25 years, both so that she could work in Spanish archives, and so that they could together enjoy slowly travelling throughout the country they grew increasingly fond of. In April 2004, shortly before he taught his last class at the University of Texas, friends, colleagues and former students honored him with a reception, for which he was required to pay an entrance fee: a talk on his life as a statistician. After retirement at 81 years of age, he took good advantage of modern technology to follow cricket and Oxford rowing on line, to solve Sudoku (which is just a special type of Latin square, after all), to research the history of the Royal Welch Fusiliers (his father’s regiment in World War I) and to relearn Welsh. With Elizabeth, he continued to share his passion for opera, symphony, politics and the burgeoning Austin restaurant scene. He is survived by his wife of 61 years, a daughter and a son, and two grandchildren. For over 40 years George Gettinby provided scientific leadership to the animal health industries and the veterinary profession in all statistical matters. He also nurtured successive generations of young scientists who looked on him as a friend, mentor and father figure. George was not a tall man, but his scientific stature, reputation, commitment and generosity dwarf most in the sector and his death in June 2014 left a gap in many lives that will remain unfilled, such was his unique contribution. An Ulsterman born and raised in Larne, George was a mathematician by first degree and doctoral training at Queen’s University Belfast (1971) and the New University of Ulster (1974) respectively. His doctoral interests in the mathematical modelling of cattle parasites continued throughout his career, though George came to the world of statistics from the start of his professional life, a career spent entirely as a member of academic staff at the University of Strathclyde in Glasgow. A ubiquitously popular educator, George served as Head of Department of Statistics and Modelling Science as well as Vice Dean of the Science Faculty. However, it was in the application of statistics that George found his real niche. In Glasgow, he established an immediate and lifelong association with the veterinary world through contacts with Sir James Armour and Max Murray, as part of a long-term collaboration with the Glasgow School of Veterinary Medicine and the International Livestock Research Institute (previously the International Laboratory for Research on Animal Diseases). His attention focused on modelling tick and insect-borne diseases, mainly East Coast Fever and tsetse-borne African trypanosomiasis. As ever, his research spawned training seminars that were universally popular, leaving standing room only during his afternoon lectures and tutorials at the International Laboratory for Research on Animal Diseases in Nairobi. His devotion to animal health was profound—whether through research projects, his commitment to regulatory issues at the Veterinary Products Committee, where he served an unprecedented three terms, the Veterinary Medicines Directorate, the Department for Environment, Food and Rural Affairs or as a consultant to industry. It is clear that simply applying quantitative techniques was not enough for him, as his activities were driven by a desire to make a difference. It was this that led him to bring the new discipline of informatics to veterinary epidemiology which he saw could provide the platform for delivering statistical and mathematical models to support those who are responsible for disease management. Embedding expert systems and statistical inference algorithms in the emerging Web environment as early as 1992 put George and his research team at the forefront of their discipline contributing to equine health, tropical diseases in production animals and endemic disease in UK cattle as well as in developing algorithms for clinical laboratory diagnostic devices such as the globally successful VetTest 8008. Maintaining his close links with animal health policy, George contributed to the debate on control of bovine tuberculosis in the UK through his membership of the Independent Scientific Group with fellow statistician Sir David Cox, and for over 20 years he worked closely with the aquaculture and fish farming industry as they sought expertise in data mining and modelling in the control of pathogens such as sea lice. This work took George’s reputation to the Nordic countries and to the Americas, in particular Canada, where a new community of friends and collaborators became established. For all his commitment to disease in animals, it could be argued that his greatest and most lasting contribution will be in the field of human ocular health. His awareness of the power of ‘big data’ and his epidemiological skills were to shape a large multicentre trial in human cataract surgery, which ultimately led to the changes in best practices that transformed surgical success rates. Modesty was one of George’s defining attributes and it was for others to ensure that he was recognized for his work. His Fellowships of the Royal Society of Edinburgh and the Royal Statistical Society, his prize lectureships such as the Weipers Memorial Lecture, his Honorary Associateship of the Royal College of Veterinary Surgeons and the Dieter Lutticken Award all bear testament to the esteem in which he was held. A man of extraordinary patience and with sublime communication skills, there are very many who are privileged to call him friend. But, above all, his love was his family; Ruth, sons Michael and Peter, and his beloved grandson, Jack. We share but can only guess at their immeasurable loss.

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.