This paper deals with simultaneous estimation of supply and demand functions for money in the United States. It gives special attention to supply formulations relating the money stock to maximum possible money stocks and to demand functions incorporating the product of national income and the rate of interest. The estimations test and attempt to improve on these formulations, and they provide evidence on the effects of changing measurement techniques of economic time series. Specifically, substantial differences emerge between estimates using quarterly averages of daily data on stock and flow variables and similar estimates using one-day end-of-quarter figures to characterize a series over a quarter. In the context of this investigation, quarterly average data appear superior in describing the true economic series, as at least some a priori judgments would suggest.