This paper describes results of a three-year field experiment and a simulat
ion study to quantify the magnitude of improvement in yield potential at di
fferent levels of management and climatic variability. In two crop seasons
for crops planted on time, there was a significant increase of 1.0-1.3% per
year in yield of cultivars but it was negligible in the third season when
March was relatively warmer. In late plantings done in one season, the rate
of improvement in yield varied between -0.16 and 0.57% per year. This indi
cates that in warm conditions, particularly during grain-filling period, th
ere was only a negligible difference among the cultivars. It suggests that
new cultivars yield more than the older ones only when the temperatures are
favorable for growth and development. The trends in simulated yields were
similar in direction but lower in magnitude compared to the trends in obser
ved yields in field experiments. The simulated rate of change in yield was
always greater in the potential production conditions compared to the modes
t level of management as is practiced today. Simulation always showed a neg
ative trend in yielding ability of cultivars in seasons when grain-filling
duration was less than 29 days. When the grain-filling duration exceeded 29
days, the trend in most cases was positive varying up to +0.67% per year.
The late-released cultivars used in this study have long vegetative duratio
n that extends grain-filling to high temperatures. Crop seasons that favore
d long vegetative duration or had high March temperatures that shorten grai
n-filling showed a negative trend in yield improvement whereas seasons with
cool March showed a positive improvement. Conclusions based on limited exp
erimentation in field studies, therefore, can be biased depending upon the
agro-environment experienced by the cultivars. Simulation studies can suppl
ement such field studies and assist in reducing this bias. (C) 2000 Elsevie
r Science B.V. All rights reserved.