摘要:We utilize Classification and Regression Tree analysis to categorize the return of extending theseason for upland. High, medium, and low values for cost of water, lint prices, and qualitydiscount/premiums were applied to the lint yield and quality differentials realized from 198irrigation termination experiments conducted in central Arizona for the crop years of 1991, 1992,1994, 1996, 1996, 1997, and 2000. The progression of each crop was analyzed using heat units(86/55 F). The relative ranking results of CART, where the most important variable isnormalized on 100, were: variety (100), additional heat units after IT-1 (94), yield of IT-1 (93),crop year (83), micronaire associated with IT-1 (68), HUAP for IT-1 (67), lint price (5), watercost (2), and the quality discount/premium year (.09). Significant yield variation from extendingthe season was found. In addition, agronomic signals for predicting micronaire changes werevery important for determining profitability since quality is impacted for both base and anyadditional yields attained. These are the primary reasons why agronomic factors were found tobe more influential than the economic factors considered for explaining the profitability ofextending the season for upland.