Statistical analysis of differences in response of beans to ozone

Abstract

The shape of the frequency-distribution response curves of Phaseolus vulgaris L. ‘California Light Red Kidney’ beans exposed to three levels of ozone was examined. A non-parametric statistical test, the Kolmogorov-Smirnov (K-S) test, was useful in determining shape or distributional differences in response to ozone. The analysis indicated that significant differences existed in the distribution of oxidant stipple injury response at different ozone fumigation treatment levels. Significant non-normal response distributions, skewed and truncated toward zero injury, were evident at low ozone fumigation treatment levels. There were near-normal distributions of response to ozone at the higher fumigation treatment level that resulted in moderate injury. Variability in injury increased with higher treatment level. Data for growth parameters indicated normal distribution of response to treatment. Parametric tests showed few differences among chambers or among sections within chambers for the fumigation facility used in this experiment. The experiment demonstrates that parametric statistics may provide inaccurate analysis of injury data with some treatments showing little injury.