Opinions differ among economists as to how effective statistical analyses using time series data can be in identifying factors affecting demand and in measuring their influences. Although this presentation may not materially modify these opinions, it should at least succeed in making even the most skeptical aware of some of the problems involved in analyses employing time series data. It is not the purpose of this paper to make a survey or review of previous demand studies. Instead, it concentrates on certain methodological approaches and what implications they may have in helping the analyst to measure demand. But with the many problems that face the statistical analyst in this task, perhaps good luck is what he needs most. Measurement of demand in the final analysis has no meaning unless it helps us answer some of the practical questions of economic life : Helping farmers to predict the expected price associated with given (or assumed) levels of production and consumer income, helping a Congressman to estimate the expected change in consumption if prices to farmers are raised when all other factors are left unchanged. This article is based upon a paper presented at the National Symposium on Dairy Market Development sponsored by the American Dairy Association in Chicago last November. Although the practical examples draw heavily on the dairy industry, the conclusions generally apply to all agricultural commodities. The author gratefully acknowledges helpful suggestions from Arthur Harlow and Hyman Weingarten.