Donald M. (Mike) Titterington sadly passed away last April after suffering for many years from Parkinson’s disease; he had spent his entire academic career at the University of Glasgow until his retirement in 2011. Mike was born in Marple, England, but moved at an early age to Stirling, Scotland and graduated in 1967 from the University of Edinburgh, before heading south of the Border to start a PhD advised by Peter Whittle. He was a genuine Scotsman, if not in the SNP sense, and he would spend summer breaks with his family looking after Robert Burns’ house in Dumfries and sailing around the coast of Scotland.

Mike was a hugely significant figure in the fields of statistics, classification, and machine learning, and a ground-breaker in many ways, innovating in an incredible range of topics, from mixtures and hidden Markov models, to data generative processes, variational Bayes, image analysis, mean-field approximations, and belief propagation. Among his books, Statistical Analysis of Finite Mixture Distributions, written jointly with Adrian Smith and Udi Makov in the early 80s, was to become a standard reference for decades (and was CPR’s first exposure to mixtures a few years later). He also edited Statistics and Neural Networks: Advances at the Interface with Jim Kay in 2000, which provided strong insights on the incoming neural network modelling revolution, and the beautiful Biometrika: One Hundred Years, which brought together landmark papers published in that journal since its creation by Francis Galton, Karl Pearson, and Raphael Weldon in 1901.

Mike indeed made many outstanding research contributions over the years, calling on both his deep scholarship and on a keen eye for new opportunities. Sequential design, mixture analysis, image and signal analysis, missing data, neural networks and high-dimensional data were among the key cross-cutting topics in which his work was highly regarded and impactful, all areas with both mathematical depth and broad scientific and societal impact. Among the PhD students he advised let us mention Adrian Bowman (Glasgow), Ernest Fokoue (Rochester Institute of Technology), Nial Friel (UCD), Clare McGrory (QUT), and Jing-Hao Xue (UCL).

However, we also want Mike’s unique personality to be remembered here: he was one of the kindest people in the discipline, tolerant to a fault and generous with his time and advice; CPR and many others, notably Peter Hall, greatly enjoyed their repeated visits to Glasgow to work with him (sometimes escaping to the hills).

An example of this dedication to the community was the time and effort he invested as the (sole) editor of Biometrika for 11 years. To this day we remain amazed at the workload he mastered, remembering his arrival at his office on Byres Road with printouts of accepted papers covered in red with suggestions for style and contents, taking advantage of his bus commute from and to his (country) home to act as a managing editor as well. In these days when statistics journals had not yet adopted web interfaces, he would write (in his rather terrible handwriting) to every author, rather than delegating to managing editors. The Biometrika style guide (a substantial eight-page document) sets out his vision of what a Biometrika paper should look like. Before the Biometrika era, he was editor of JRSS Series B (1986–89) and AE for many top journals. He also was the co-chair of the AISTATS 2010 conference, which is a hugely demanding if short-term responsibility, as the time constraints for handling numerous contributions are extreme. Positions like these, and others such as his service in the governance of the Royal Statistical Society, and in chairing UK Research Council panels, are a testament to the trust and respect that he enjoyed in the research community.

Mike spent several long-term visits in India (ISI, Kolkata), Australia (Melbourne), and France (Grenoble). The trips to Australia were associated with his long friendship with Peter Hall (1951–2016), for whom he wrote a strong and touching obituary in 2016 for Significance. His conclusion that Peter Hall “was held in deep affection and admiration by the many friends, colleagues and passing acquaintances (…) Many have expressed gratitude for assistance provided freely in collaborations. He was kind, generous and humble, and is already sorely missed” applies equally well to himself.

Written by Peter J. Green (Bristol) and Christian P. Robert (Paris)