Based on the cost-of-carry model of future prices, a number of studies have estimated nonlinear autoregressive models for the basis at different frequencies (see, e.g., Dwyer GP, Locke, P, & Yu, W, 1996; Monoyios M and Sarno L, 2002; Taylor N, van Dijk D, Franses PH, & Lucas A, 2000). The structure of the models and the speed of adjustment to shocks reported are radically different. In this paper we examine the implications of systematic sampling. The results obtained show that regular sampling of the process seems important in attempting to explain the apparently contradictory results reported on the speed of adjustment to shocks in the cost-of-carry model.