Professor Roger Mead, who passed away at home on August 11th, 2015, was for many years a leading researcher in the design and analysis of experiments and applications of statistics to agriculture. He spent 34 years at the University of Reading, where he helped to turn the Department of Applied Statistics into one of the UK's most dynamic statistics research groups, particularly with links to the biological sciences. Roger graduated with a Bachelors degree in mathematics and a Diploma in mathematical statistics from Cambridge. Although he valued his attachment to Cambridge and always retained a very natural feel for mathematical statistics, Roger felt that he only really began to learn statistics when he worked at the National Vegetable Research Station at Wellesbourne, alongside John Nelder, for 5 years from 1961. During this period the now famous Nelder–Mead simplex algorithm for numerical optimization was developed. Their 1965 paper has over 20000 citations and is so widely known that most researchers simply refer to ‘the Nelder–Mead algorithm’ without citing (or reading) the original paper. Roger always seemed slightly surprised when researchers used the method, regarding the paper as ‘one of [his] earliest and certainly least statistical’. In 1966 Roger was appointed as a Lecturer in the Unit of Biometry, led by Robert Curnow and part of the University of Reading Department of Agricultural Botany. The Unit soon expanded to become the Department of Applied Statistics, offering service courses at undergraduate and Masters level to students from many different departments, and running a very successful Master of Science degree in biometry. Roger particularly enjoyed teaching agriculture students and from the late 1960s was a pioneer in developing computer games for learning about curve fitting, as well as designing experiments. The experiment games Tomato and Chick have been used at Reading by many generations of students from agriculture undergraduates to statistics doctoral students and wherever ex-Reading people teach design of experiments. An unusual feature of Reading's department was that it offered a statistical advisory service, with walk-in advice from professional statisticians available to researchers and students from across the University every day. This was particularly appreciated by Faculty of Agriculture and Food researchers and led to many collaborations, as well as improved design and analysis in research projects. However, although some applied statistics staff regarded the advisory service as an enormously useful, if sometimes painful, learning experience, others viewed it with sheer terror. Roger, in contrast, relished his advisory sessions and continued full involvement even when Head of Department. When some of his advisory sessions were video recorded, for use with Master of Science students, one became a classic. His unique ability to put the client at ease and to explain everything at just the right level, so that by the end of the session they were extremely grateful for having been persuaded that the experiment that they had done was essentially useless, was easy to admire if difficult to emulate. Roger was promoted to Senior Lecturer in 1971, to Reader in 1978 and to Professor of Applied Statistics in 1984. He served as Head of the Department of Applied Statistics in 1974–1977, 1986–1990 and 1992–1993. The latter periods saw the department continuing to be successful; Roger's leadership ability was recognized and he became Dean of the Faculty of Agriculture and Food (1993–1996) and Deputy Vice-Chancellor (1996–2000). He retired in 2000 but continued to be active as Emeritus Professor. Roger's research covered both very applied work and general methodology driven by the applications. He made many contributions in agricultural statistics, particularly the design of experiments and the analysis of intercropping data. He co-authored the popular textbook Statistical Methods in Agriculture and Experimental Biology, published in 1983, with Robert Curnow and, for the third edition in 2003, Anne Hasted. This remains popular, particularly for the real examples described, which are taken from the authors’ consulting and advisory experience. Roger's 1988 book The Design of Experiments presented his unique insights into the subject from the bottom up. Roger strongly believed in building the theory on the fundamental statistical principles allowing designs to be developed to meet the objectives of the specific experiment at hand. Although not hiding the mathematical developments needed to apply the statistical theory, Roger firmly believed that they should never be presented as an end in themselves. A new book Statistical Principles for the Design of Experiments, based on the 1988 original, appeared in 2012, co-authored by Steven Gilmour and Andrew Mead (Roger's son). Highlights of Roger's research career included his four papers read to the Royal Statistical Society, ‘A review of statistical ideas relevant to intercropping research’ (1981) with Janet Riley, ‘Continuous second order models of spatial variation with application to the efficiency of field crop experiments’ (1986) with Alan Brewer, ‘The non-orthogonal design of experiments’ (1990) and ‘Statistical analysis of public lending right loans’ (1991) with Anne Hasted, J. Sumsion and Sharon Wheeler. The third of these presented a strong defence of Roger's view of experimental design and generated a very heated discussion. Roger later felt that some annoying numerical mistakes in the paper had skewed the discussion away from his main, almost philosophical, points. Roger was also very committed to statistics for international development and was involved in several projects in the late 1970s and 1980s. He visited Sri Lanka on several occasions, where Reading had strong links with the University of Colombo. He helped to set up a link between Reading and the University of Ege in Turkey and gave lectures on the design and analysis of experiments there. Roger had great influence wherever he visited, and scientists at the International Maize and Wheat Improvement Center in Mexico and the World Agroforestry Centre in Kenya have commented on how he improved their design and analysis work. Although much of Roger's research was in experimental design, one of his unique skills was the ability to extract information just by looking at raw data. His ability to spot patterns which others would take days to find seemed entirely natural; the only way to try to learn this skill was to watch Roger in action, as he never seemed to think that there was anything to explain. After hearing news of Roger's death, several people said that they had been recently thinking about Roger, as they tried to solve a difficult problem by trying to think about it the way Roger would. Roger will be remembered by many colleagues for his kind and cheerful advice, as well as his unique statistical insights. Those who worked closely with Roger found every meeting stimulating and enjoyable. He never seemed to have a bad day and even when he was heavily involved in University management he had time for junior colleagues and students. Roger served as Joint Editor for the Royal Statistical Society's journal Applied Statistics from 1985 to 1987. He was particularly active in the International Biometric Society, of which he was International Secretary from 1985 to 1992. He was made an Honorary Life Member of the International Biometric Society in 2014. Roger was elected a member of the International Statistical Institute in 1988. After retirement, Roger continued to be interested in statistics. As well as working on the new design book, he served on the local research ethics committee. He suffered from a rare form of late onset muscular dystrophy and became increasingly immobile, although his mind remained active until the end. Roger is survived by his wife Nina, his children Rachel, Simon and Andrew and his seven grandchildren. Steven Gilmour