Direct and Correlated Responses to Selection for Rate and Duration of Grain Fill in Maize

Grain filling duration and dry matter accumulation rate are associated with maize (Zea mays L.) yield potential and might be used for indirect yield selection. Additional physiological and genetic information, however, are needed to verify selection procedures and potential gains. Phenotypic recurrent selection was used to develop subpopulations for the physiological components of kernel growth: long EFPD (effective filling period duration), short EFPD, high DMAR (dry matter accumulation rate), and low DMAR. Our objectives were to measure selection progress for these yield components and to observe correlated responses. The original population (CO) and synthetics from three selection cycles (C1-C3) were evaluated for physiological traits at one location and for agronomic traits at three locations in lattice designs. Regressions and realized heritabilities were calculated. Changes per cycle for the long and short EFPD subpopulations were +4.1 and −3.1% with heritabilities of 0.19 and 0.15, respectively. High and low DMAR selection caused changes per cycle of +4.8 and −8.0% with heritabilities of 0.20 and 0.44, respectively. Concurrent half-sib yield selection lengthened EFPD. Long EFPD C3 yield was not different and high DMAR C3 was 13% greater than CO. Stalk lodging, harvest moisture, and plant height were affected. Kernel number per plant was influenced by DMAR but not EFPD. Although kernel wt. and number were influenced by DMAR and kernel wt. by EFPD, yield per unit area was changed only by the DMAR and half-sib selections. Failure of the long EFPD selection, which also reduced DMAR, to increase yield per unit area was due to marginally reduced kernel size and number. Selection for long EFPD and high DMAR is possible, but improvement of yield may be dependent on concurrent control of other yield components. Journal Article no. 87-3-92 published with the approval of the director of the Kentucky Agric. Exp. Stn.