Rainfall amounts and the distribution across the landscape are critical to decision-making and evaluation of hydrological models. Spatial variation in rainfall has been observed through anecdotal evidence and limited studies; however, there is little quantitative evidence that can be used to assess rainfall variation within a watershed on a daily, monthly, or yearly temporal scale. This study was conducted to quantify the spatial variation within a watershed in central Iowa and to determine if there were consistent differences among rain gages for the period from 1991 through 1998. The study was conducted within Walnut Creek watershed located south of Ames, Iowa on the Des Moines Lobe Landform region. The topography of this 5130ha watershed is characterized by gently rolling fields with a narrow area of steeper land along the stream in the lower part of the watershed. Twenty-two tipping bucket rain gages were placed throughout the watershed and rainfall was recorded as 5 minute totals and then aggregated into daily totals. Accumulation of errors of the 5 minute values into the daily totals were considered to be random. There was a large coefficient of variation in the average daily totals; however, there was no consistent pattern of variation among rain gages, and coefficient of variation decreased with amount of rain. Each rain gage had an equal chance of receiving the lowest or highest rainfall total for any given storm event. When the daily average was computed over the year, there were no differences among rain gages. Monthly and yearly totals showed a decreased coefficient of variation compared to daily totals. There was no consistent pattern of spottiness within the watershed and if daily rainfall amounts are required for a decision, then direct measurements may be required.