quate data, we were unable to capture important quality differences in a heterogenous commodity. On the contrary, the weekly FOB price series used in our analysis were constructed by calculating an FOB price weighted by size, grade, and production district price differentials (Thompson and Lyon, p. 651). The three weighted market-level FOB series thus reflected changes in the volume of undesirable sizes and grades. Because FOB prices for only those sizes of oranges for which retail prices were collected were used to construct the FOB series, our weighted FOB prices conformed closely to the prices reported in our retail price series. Thus, the three price series explicitly captured quality effects. Any potential bias introduced by ignoring grade and size changes through time was thereby avoided. As acknowledged by Powers, this bias cannot be remedied with the aggregate data he employs. This is a crucial point because Powers concludes that deteriorating product quality at the end of the season should cause price spreads to increase. Powers' second criticism suggests that we failed to properly identify weeks during the 1985/86 and 1986/ 87 marketing seasons when the industry also operated without prorate. Our analysis defined the period of prorate suspension as having occurred when the U.S. Department of Agriculture (USDA) unilaterally suspended prorate during the 1984/85 season. We considered the 1984/85 suspension to be fundamentally different from subsequent periods because of the unique circumstances leading up to the suspension. First, market conditions were markedly different. Freezes in 1984/85 in Florida and Texas left California-Arizona (CA) fresh navel oranges with almost no competing supplies during the 1985 marketing period. Second, a majority of the fresh domestic crop had not yet been shipped. When USDA announced the suspension (beginning the first week in February of 1985 and extending for 18 consecutive weeks until the end of the season), only 37.6% of the crop had been shipped under prorate (Navel Orange Administrative Committee [NOAC] Annual Report, 1985). Third, and most important, the prorate suspension in February 1985 was virtually unanticipated. Industry officials admit that the secretary of agriculture's decision to suspend the prorate caught them off guard.' In contrast, removal of prorate during the 1985/ 86 and 1986/87 seasons was fully anticipated, was of shorter duration, and occurred under much different supply conditions. At the beginning of the 1985/ 86 season, USDA notified NOAC on 25 November 1985 that USDA would issue weekly volume regulations based upon criteria established by NOAC. After protracted discussions among committee members, NOAC finally agreed on 23 February 1986 to three criteria upon which to base the decision to suspend the weekly prorate.2 The prorate in the 1985/ 86 season was suspended after the week ending on 13 April when two of the three criteria were met. This suspension was clearly anticipated and nearly two-thirds of the fresh domestic shipments had already been accounted for. Rules regarding suspension of the prorate were even more stringent during the 1986/87 season as NOAC proposed on 30 October 1986 that volume controls be imposed in each week through the week ending on 23 April 1987; these proposed regulations were accepted by USDA, and 79.9% of the season's fresh domestic crop was shipped before prorate was suspended. In both these seasons, supply conditions permitted larger shipments from Texas and Florida than during the 1984/85 season. Thus the suspensions occurring in the 1985/86 and 1986/87 were clearly different from the 1984/85 susp nsion in important ways. Acknowledging these dissimilarities, we isolated the 1984/85 suspension as the cynosure of our analysis, thereby permitting examination of industry and market reaction to this unanticipated supply-side policy shock.