Grain yields and yield components in certain brazilian-grown wheats

Abstract

Yield components - number of fertile spikes, number of seeds per spike, and average seed weight - were determined in each of 30 varieties of wheat grown as a part of a regional trial near Pelotas, Rio Grande do Sul, Brazil, in 1966. In addition, average plant height and average number of seeds per spikelet were also determined. From the yield component data, a theoretical yield was calculated and compared with the observed yield. Average plant height ranged from only 88 cm for the variety Pel A 509-64 to 133 cm for the variety S-23. The average height of all 30 varieties was 112 cm. The average number of fertile spikes was 278 and the range was from 213 (S-31) to 340 (Cotiporã). Number of seeds per spike averaged 22.37 for all 30 varieties with a range from 16.71 (S-15) to 28.57 (S-23). Average number of seeds per spikelet ranged from 1.30 (IAS 20-Iassul) to 2.11 (S-23 and S-11). Average seed weight for all 30 varieties was 33.77 and ranged from 26.30 mg (Pel A 509-64) to 43.35 mg (IAS 20-Iassul). The observed yield for the whole plot showed nearly the same correlation (0.44 and 0.43) between number of fertile spikes and number of seeds per spike. As expected, the correlation between observed yield of the whole plot and the total number of seeds (number of seeds per spike x number of fertile spikes) was higher (0.56). A tridimensional graph (Fig. 1) shows the relationship between calculated number of seeds, seed weights and yield. Ten of the 12 possible classes are shown with observed yield data. Of the five combinations with above-average yield, all are in the expected top seed number-seed weight classes. The same situation exists for the five combinations with below-average yield. Literature is reviewed in the light of the research reported herein as well as that of other proposals or systems in breeding for higher yield per se.