In the 2007 paper “Waterlogging Tolerance of Kentucky Bluegrass Cultivars,” by Kehua Wang and Yiwei Jiang [HortScience 42(2):386–390], the authors report the following corrections:
In the “Materials and Methods,” the correct experimental design was a split plot design with waterlogging treatment as the whole unit and cultivar as the subunit. Turf quality (TQ) was arcsine transformed and shoot water-soluble carbohydrate content (SWSC), root water-soluble carbohydrate content (RWSC), root electrolyte leakage (REL), and root dry weight (RDW) were log transformed. Mean separation tests were performed on means of transformed data; however, the results were presented using untransformed means. The corrected Table 1 (ANOVA) is as follows:
Analysis of variance of arcsine transformed turf quality (ArcTQ), chlorophyll content (Chl), root oxidase activity (ROA), log transformed shoot water-soluble carbohydrate content (LogSWSC), root water-soluble carbohydrate content (LogRWSC), root electrolyte leakage (LogREL), and root dry weight (LogRDW) under 10 and 30 d of waterlogging in Kentucky bluegrass.
In Table 2, the WL × C effect was nonsignificant for Chl, REL, and RDW. For ROA, significance should be compared for cultivar Eagleton, not Awesome; and no significance should be compared for cultivar Julia under variable of SWSC. In Table 3, the WL × C effect was nonsignificant for Chl, SWSC, RWSC, REL, ROA, and RDW. Therefore, comparing these means is inappropriate and should be ignored.