ABSTRACT In 1833 John Herschel published a graphical method for determining the orbits of double stars. He argued that his method, which depended on human judgment rather than mathematical analysis, gave better results than computation, given the uncertainty in the data. Herschel found that astronomy and terrestrial physics were especially suitable for graphical treatment, and he expected that graphs would soon become important in all areas of science. He argued with William Whewell and James D. Forbes over the process of (...) induction, over the application of probability, and over the moral content of science. Graphs entered into all these debates; but because they constituted a method, not a metaphysics, they were acceptable to most practicing scientists and became increasingly popular throughout the nineteenth century. (shrink)