Most economic analyses indicate that one of the more important factors to obtain profitable returns from growing cotton is to have high yields. Moreover, in 1977 (17) and again in 1982 (10) lower yield problems were highlighted at this conference. For example, from 1936 to 1960, U.S.A. lint yields increased about 9 lbs/acre/year, but in the next 20 years yields decreased at a rate of 1 lb/acre/year (10). This loss plateauing of yield was in spite of increasing economic inputs, research information and technology improvements. For example, variety improvements during this 20 year period remained at 9 lbs/acre/year (Table 1), but Mississippi state average yields were decreasing at the rate of about 6 lbs acre/year. Since varietal yield gains and other technology improvements had been made, why weren't state or national averages progressing at least at a similar rate? This suggested to some (10) that we were experiencing a yield decline. The yield decline is economically disastrous, generally controversial, and needed to be investigated from a scientific viewpoint. The yield decline analyses of 1982 (10) resulted in the following observations: (1) Maximum yields were reached in the mid-1960's followed by a yield decline. (2) The decline was not due to genetic factors (Table 1). (3) The linear decline with years suggests a cumulative effect of negative production factors on yield. (4) The decline was over a wide geographical area. (5) The decline was caused by either long term weather changes or by unknown effects of cotton production practices.

Because of the economic loss, controversy over causes, and absence of scientific explanation of the yield decline, The Cotton Foundation formed a committee to study the possible causes(s) of the yield decline. This committee consisted of about 20 persons from industry, State Agricultural Experiment Stations, Agricultural Research Service, Cotton Incorporated, and the National Cotton Council. Implied, but not directly stated in the previous analyses (10) was that years per se do not cause yield declines, but that it is physical, cultural or biological factors associated with years that is causing the yield decline. To attack the unknowns associated with the yield decline, weather and management data were obtained for 20 years of variety tests from five states, South Carolina, Mississippi, Texas, Arizona, and California. These data were input a computer so that GOSSYM simulations could be compared with actual field data. While closeness of predictions and observed field results is important, it is more important in the yield decline studies to associate trends simulated by GOSSYM with observed trends. The basic idea of the GOSSYM model is expressed elsewhere (1), but in general, basic cotton physiology research from many disciplines are incorporated into GOSSYM to integrate all of the processes into one growth response. The factors studied were weather, compaction, herbicides, insect control, and management inputs. It's well known that all of these factors can influence yield, but the important test is whether any of these factors are responsible for the long term trend of decreasing yields. Examples of the more important conclusions reached by the committee will be given. The results from three locations Stoneville, Mississippi; Lubbock, Texas; and Fresno, California will be used to summarize these studies. More detail and technical reports will be presented elsewhere at this conference and a more complete technical documentation is being prepared. In some cases where conventional research results on this subject are available, actual research studies will be used in connection with GOSSYM simulations to reach conclusions.