Computers, the Second Revolution in Statistics

May I first express my appreciation of the honour our new President of the Royal Society, Professor Blackett, has done Sir Ronald Fisher by consenting to take the Chair tonight. I am particularly pleased that he is presiding at this meeting as not only was he an old friend of Sir Ronald, but it was also he and the then Secretary of the Agricultural Research Council, Sir William Slater, who were mainly responsible for enabling us to get started with a computer at Rothamsted. As this is the first Fisher Memorial Lecture it might be thought that it should deal with some facet of Fisher's work. In a sense it does indeed do so, for to Fisher computing was very much a part of statistics, and a necessary adjunct to the development of mathematical theory. Although Fisher never much concerned himself with electronic computers-I remember him referring to them once in the early days as 'meccano arithmetic'-he later, I think, began to recognise that, rightly used, they had possibilities. Certainly he always greatly valued good computing aids. Soon after he came to Rothamsted he persuaded the station to buy him a Millionaire, regarded at that time as a great extravagance for an agricultural research station. And an excellent machine it was. I still retain it for my own personal use. Our early acquisition of an electronic computer at Rothamsted-the first computer, I believe, solely devoted to statistics-stems directly from that early Millionaire. For Fisher set the tradition that to be a good theoretical statistician one must also compute, and must therefore have the best computing aids. It is a tradition we have firmly held to.