Relationships between line per se and testcross performance for agronomic traits in two broad-based populations of maize Goran Bekavac Æ Boz ˇana Purar Æ – Dord¯eJockovic ´ Received: 23 February 2007 / Accepted: 23 July 2007 / Published online: 8 August 2007 Ó Springer Science+Business Media B.V. 2007 Abstract Degree of the association between line per se performance (LP) and testcross performance (TP) is important in breeding programs and simulta- neous improvement of commercial hybrids and their parental lines. This experiment was designed to study genetic variability and genetic correlation for several agronomic traits in two maize (Zea mays L.) broad- based populations (NS12-SG and NS14-SG). Inde- pendent trials with 80 entries of S 1 progenies as well as their testcrosses were arranged according to an incomplete block design with replicates in sets. Grain yield, stay green, anthesis-silking interval, stalk water content and grain moisture were evaluated in four environments. The anthesis-silking interval had the highest genetic variation, followed by stay green. High heritability estimates (>0.50) for all traits, pointed out that further selection would be possible. Genetic correlations between line per se and testcross performance were lowest for grain yield (0.396** and 0.592**, for NS12-SG and NS14-SG, respectively), and highest for grain moisture (0.937** and 0.821**, respectively). High correlations between line per se and their testcrosses for stay green, anthesis-silking interval, stalk water content and grain moisture indicated that additive gene action might be more important than dominance in controlling the expres- sion of these traits. Keywords Maize  Variability  Correlation Introduction Since the pioneer contribution of Shull it has been clear that single-cross hybrids are the best way to exploit heterosis (Gallais 1997). One method used in developing superior hybrids is the parallel evaluation of both the performance of the lines per se and the testcrosses of these lines using an inbred tester. In practice, S 1 or S 2 lines per se and their crosses to an inbred tester are grown in yield tests. Best lines based on the performance per se and their testcross performance undergoes continued inbreeding and testing (Hallauer and Miranda 1988). This method was thought to be appropriate because modern inbreds should have both—good combining ability and high seed yield potential. This is particularly important for female lines because it determines the cost of F 1 seed production. Goulas and Lonnquist (1976) pointed out that combined half-sib and S 1 family selection may be useful for improving simultaneously combining abil- ity and per se value. This method has also been proposed by many researches (Gallais 1997; Moreno- Gonzales and Hallauer 1982; Coors 1988). In a study of crossbred and S 2 family selection in a full-sib G. Bekavac (&)  B. Purar  Ð. Jockovic ´ Institute of Field and Vegetable Crops, M. Gorkog 30, Novi Sad 21000, Serbia e-mail: gbekavac@ifvcns.ns.ac.yu 123 Euphytica (2008) 162:363–369 DOI 10.1007/s10681-007-9527-y