Pricing performance of futures markets for seasonally produced, storable commodities often has been evaluated in part by determining whether or not th~ basis at the delivery point reflects the theoretical factors expected to affect it. 1 These factors include marginal net storage costs (Working, Brennan, and Weymar), market liquidity (Gray 1960 and 1967, Martin andStorey) and market characteristics unique to a specific commodity (Ward and Dasse).Futures markets for these commodities also are used to establish current prices at nondelivery points. Following the argument of Ward and Dasse, if pricing performance at nondelivery points is satisfactory, local basis should reflect market characteristics unique to the local area because it is a measure of the value of a commodity relative to the delivery point.Futures markets also are used to forward price grains at nondelivery points either by hedging or forward contracting. In either case, expectations of basis are important in establishing forward prices. Most extension literature stresses the importance of understanding the local basis and asserts that local basis is affected by local market conditions (e.g., Kenyon and Blakely). Yet for the few cases in which analyses of local basis have been undertaken, relatively naive models have been used to explain basis variations, such as moving averages or basis at the time a hedge is initiated (Kenyon and Blakely, Driscoll and Blakley, and Heifner 1966). The ability to explain nondelivery point basis using variables reflecting local market conditions is important for evaluating pricing performance and for providing local traders and hedgers an understanding of the factors which cause variations in the local basis.In this paper an analysis of nondelivery point basis is undertaken with the objective of providing information for these two purposes. The analysis is carried out by constructing a hypothetical model to explain variations in local basis and then testing it by applying it to basis data for southwestern Ontario.