A quantitative trait response evaluation to selection in the BS13(S) maize (Zea mays L.) population

Seven cycles of half-sib recurrent selection were conducted in the BSSS maize population followed by ten cycles of S2 progeny recurrent selection. The population under S2 progeny recurrent selection is formally known as BS 13(S). The selection criteria have always placed importance on high grain yield, low grain moisture, and reduced plant lodging. Two evaluations estimating the response of multiple agronomic traits in multiple response units including the population sampled at F IS = 0, FIS = 0.5, and FIS = 0.75 levels of inbreeding, and multiple testcrosses of the population were conducted. The average response of grain yield in the FIS = 0.5 and FIS = 0.75 response units (i.e. inbred response units) of the population are significantly greater than the average responses in both the panmictic population and multiple testcrosses of panmictic population. There is no statistical grain yield response to selection in the FIS = 0 response unit (panmictic population). A testcross of the panmictic population to BS13(S)C0 has an average response greater than the panmictic population indicating that the limited response in the FIS = 0 response unit is not likely due to random genetic drift, however, the allelic frequencies are diverging from the progenitor population. Favorable responses have been observed for root lodging in all of the response units however, a limited response for stalk lodging exist. The response of grain moisture is inconsistent between the two evaluations likely as a result of inconsistent selection practices and the difficulty to select for this trait. We hypothesize that the most important reason for the realized lack of response in the panmictic population is due to a low inbred-outbred correlation and may be caused by an overdominantlike gene action within this population. To avoid the necessity for assumptions about gene action controlling the traits in maize and importance of additive and non-additive effects, the genetic gain equation must be developed in relation to the selection unit and target response unit (i.e. S1s, S 2s, half-sibs, or full-sibs). This dissertation supports both the genetic gain equations reported by Wardyn (2006) and the hypothesis of overdominant-like gene action in this population reported by Edwards and Lamkey (2002).