CROP YIELDS RANDOM IN RECENT YEARS? Kuang-hsing

After three to four decades of substantial inlittle evidence of a trend, it is considered to crease, yields per acre of major U.S. crops aphave plateaued. The second objective is to pear to have tapered off. The average annual investigate whether a crop yield series, if rate of growth in crop yields' decreased trendless, is random or oscillatory and, if nonmarkedly from 2.7 percent for the 1950s to 1.8 stationary, whether cycles are present. An percent for the 1960s, and finally fell to 0.1 peroscillatory series is defined as a trendless cent for the first half of the 1970s [2]. Scienseries dominated by cycles; a random series is tists commissioned by the National Academy one lacking trend and cyclical elements. It is of Sciences to investigate agriculture producnot the authors' intention to quantitatively tion efficiency first discovered the problem [6]. evaluate factors underlying the level and Concern about the situation was expressed in changes in the level of individual crop yields. the literature early this year by both a plant physiologist [11] and agricultural economists [1, 2]. If a crop yield series has reached and reFIGURE 1. LASSFICATION OF DATA mains on a plateau, it would conform with what is expected in a stationary or trendless Ti 'eries series. Therefore, examination of the leveling of crop yields can be accomplished by testing (Tred significnce test)