The first computer at a nutritional research institute: how its use has evolved.

The background to the installation of a digital computer at the Rowett Research Institute is that there has been a Biometry Department since 1948. The function of this department has evolved naturally over the years until before the acquisition of the computer when it consisted of two scientific officers, one experimental officer and three machine operators. Over the last 15 years or so it has been customary for consultation to take place between research workers and the Biometry Department on the rationale and design characteristics of projected experiments. Within the Biometry Department a formal design is prepared for each agreed experiment along with specific questions which the experiment is intended to answer. The experiment is identified by number, and a copy of the design retained in Biometry pending the completion of the experiment. On completion the data are analysed by the Biometry Department and a report is written summarizing the conclusions in terms of the stated aims of the experiment, and giving, in addition, information on various aspects of variability. With the steady increase in numbers of research workers in the Institute, and the development of equipment designed to speed up various analytical measurements the work involved in the Biometry Department has increased over the years. I n addition to the electric calculators in the Biometry Department, seventeen others were sited throughout the Institute and were about 75% fully occupied on various routine calculations. Hand machines were also employed and a survey made in 1965 indicated that, at a very conservative estimate, over 400 man hours weekly were occupied in routine calculations of one kind or another. We recognized that many of these calculations were too slight to warrant programming, punching and checking, but clearly there were many for which the use of a computer would be advantageous. T h e calculations involved in determinations of nitrogen and other entities, in respiration calorimetry and in metabolic trials, to name a few, come into this category. A third factor entered into our decision that a computer was necessary. Over the last 10 years there had been advances in automated analysis-of particular import-