Interplot competition in spring wheat

Abstract

Research in Europe and the United States showed that interplot competition in variety trials can result in distortion of yield and misleading conclusions from yield comparisons. Field trials were conducted at Swift Current and Saskatoon in 1993 and 1994 to find out if interplot competition occurs in the wheat growing region of Saskatchewan. Four spring wheat (Triticum aestivum L.) cultivars, Biggar, Oslo, Genesis and Glenlea, were grown in four-row plots in a diallel competition treatment design, with the competition treatments applied randomly to each of eight replications in a randomized complete block design. Generally, yield was distorted from pure stand yield when adjacent plots differed for height ($-$0.34% cm$\sp{-1}).$ Plots with Biggar or Oslo flanked by a tall cultivar, Genesis or Glenlea, yielded less than when flanked by themselves, and the reverse occurred when the tall cultivars were flanked by the short cultivars. Interplot competition also affected spike density in the same fashion, did not significantly affect height, and was inconclusive with regards to kernel weight. Simulation indicated that a correlation between height and yield would not affect the covariance adjustment with average height difference. Sixty-five historical data sets were tested to see if an adjustment for interplot competition with average height difference as a covariate would be effective. The negative relationship between yield and the height covariate was significant for 25 of the data sets. At Saskatoon in 1995, row directions were compared and interplot competition was significant in the north/south rows and not significant in the east/west rows. An experiment at Regina and Swift Current in 1995 indicated that less interplot competition occurred when plots were separated by spring planted winter wheat than when plots were 23 cm apart. It was concluded that when plots differ for height, yield distortion can occur in the Saskatchewan wheat growing region.