Churchill Eisenhart, 81, founder of the National Bureau of Standards (NBS) Statistical Engineering Laboratory and former President of the American Statistical Association, died of cancer at Suburban Hospital in Bethesda, Maryland, on June 25, 1994. Churchill was a visionary. From the start of his professional career, he was acutely aware that the physical sciences had not kept pace with the biological sciences and industry in the handling of uncertainty in their experiments. He is given credit for inventing the idea that only by treating measurement as a production process could one arrive at uncertainty statements that would stand up in court. His idea was that only if experiments were properly designed to account for sources of variability and the measurements were shown to be in a state of statistical control (using procedures developed by R. A. Fisher, W. A. Shewhart, and W. E. Deming) would they withstand cross-examination (i.e., that the announced uncertainty would be accepted). This approach proved to be crucial in the case in which the objectivity and correctness of the NBS results were questioned. He was also a gifted and painstaking expositor. His 1947 paper in Biometrics, on the assumptions underlying the analysis of variance, clarified the distinction between fixed-effects and random-effects models (Eisenhart 1947); for a long time, the random-effects model was routinely referred to as Model II. Careful attention to assumptions also characterizes his paper on realistic evaluation of the precision and accuracy of instrument calibration systems (Eisenhart 1963), where he discusses the justification for using a statistical model and the importance of estimating and interpreting variance components. Churchill Eisenhart was born on March 11, 1913 in Rochester, New York, but was raised from infancy in Princeton, New Jersey. He attended Princeton University and received an A.B. degree (mathematical physics) in 1934 and an A.M. degree (mathematics) in 1935. It was the distinguished physicist E. U. Condon (then on the faculty) who led him to the field of statistics, pointing out the need for physicists to go beyond the 19th century theory of errors approach. At the suggestion of Samuel S. .Wilks, he went to University College, London, to study with the world's leading statisticians, Egon S. Pearson, Jerzy Neyman, and R. A. Fisher. He received his Ph.D. there in 1937. He began his professional career at the University of Wisconsin in 1937 where he was Associate Professor of Mathematics and head of the Biometry section of the Agriculture Experiment Station. During World War II, while on leave from the University, he became a research associate in mathematics at Tufts College, working on the mathematical theory of aerial combat. Later at Columbia, under the direction of the Office of Scientific Research and Development, he organized and directed mathematical and statistical studies on aerial gunnery; and as senior staff member of the Statistical Research Group, he worked on the mathematical phases of the rocket development program. For this work he received the Naval Ordnance Development Award in 1946. He was co-editor of Selected Techniques of Statistical Analysis (Eisenhart, Hastay, and Wallis 1947), a collection of papers on topics for engineering applications that arose from the Group's applied research. In 1945 E. U. Condon became Director of the NBS. To implement his belief that the NBS needed a program in modem statistics, he offered Eisenhart the job of organizing such a program. Churchill eagerly accepted the challenge and left the University of Wisconsin tojoin NBS. He came to the National Bureau of Standards (now the National Institute of Standards and Technology) in 1946; and in July 1947 the Statistical Engineering Laboratory was established. He recruited W. J. Youden, and together they led pioneering work in introducing modem statistical methods in experimental work in the physical sciences. Eisenhart's 1963 paper on Realistic evaluation. . is still one of the most widely quoted references in discussions of uncertainties. His Laboratory's program focused on experiment designs suited to the needs of the physical sciences and of industrial experimentation, extreme value methods, reliability theory, and statistical computing. He was one of the founding organizers of the ASA Section on Statistics in the Physical and Engineering Sciences. The NBS statistics group met its first real test in the battery additive controversy in 1952. NBS had published results on commercial battery additives that had claimed to lengthen battery life-all were found to be ineffective. Joseph M. Cameron is a Consultant at 12502 Gould Road, Wheaton, MD 20906-3822 and Joan R. Rosenblatt is Director of the Computing and Applied Mathematics Laboratory, National Institute of Standards and Technology, Gaithersburg, MD 20899-0001.