Author: PEDRO MÁRIO DE ARAÚJO Adviser: Prof. Dr. JOSÉ BRANCO DE MIRANDA FILHO Nínety S2 progenies from each of the maize (Zea mays L.) populations BR 106 and IAPAR 26, were evaluated under the partia! circulant interpopulation diallel (CDPCI) and top-cross (CTC) mating schemes. ln CDPCI progenies were divided into nine sets of ten progenies each; in each set, each progeny was crossed wíth four (s = 4) progenies of the opposite population, thus totaling 40 intermediate hybrids in each set.

more »

... es of variance were performed in each set and then pooled over sets. ln CTC, two triais were designed: TC 1 with cresses between progenies of BR 106 with the population IAPAR 26; and TC2 with cresses between progenies of IAPAR 26 with the population BR 106. Triais were conducted in 97/98 in three locations: Londrina (PR), Ponta Grossa (PR) e Rio Verde (GO). The following traits were analyzed: graín yield (PG: plot total expressed in kg/ha), plant height (AP; cm), ear height (AE; cm), days for female floweríng (FF; days), and prolificacy (PRO; ears per plant). ln the analysís of variance wíthin locatíons for CDPCI scheme, the mean squares for general combíning ability (CGC) showed higher expression than for specific combining ability (CEC) for ali traits. Among traits CEC showed higher expression for PG and PRO. The partition of the sums of squares for CGC for the two populations showed always higher values for BR 106 for all traits and locations. For PG, the estimates of CGC effects for progenies of BR Xl 106 (gi) and IAPAR 26 (gí) varied from -915 to 917 kg/ha and -657 to 883 kg/ha, respectívely. The contrasts of means, ki = Y i• -Y .. e ki = Y ·i -Y .. , calculated from the observations (hybríd means, Yu) varied from -1134 to 854 and from -727 to 930 kg/ha, respectively. The estimates of CEC were in the range of -908 a 1002 kg/ha, indicatíng that thís effect may contribute expressively for higher yield in specific hybrid combinations. The highest yield for an intermediate hybrid was 9063 kg/ha, that is 15.6% higher than the check (P 3041) mean of 7838 kg/ha; and 35, 1 % higher the mean of the parental populations (6707 kg/ha). The overall hybrid mean associated with high positive values of CGC (gi e 9i), as well as CEC (sii), indicate by themselves the high potential of the populations for the exploitation of hybrid crosses from partia! or completely inbred lines. The analysis of variance for PG, with partition of the variation among hybrids into lines and diagonais in the circulant diallel scheme, revealed average estimates of the interpopulation additive variance ( cr i 12 ) and domínance variance ( cr b 12 ) of 113.22 e 57.18 (g/pl) 2 , respectively. For the other traíts, only for PRO the variance ratio was close to the unity, while the dominance variance was always smaller than the additive variance for the remaining traits. The combined analysis over environments indicated that the interaction dominant x environment variance was more expressive than the interaction additive x environment variance, mainly for PG, thus increasing the ln the top cross analysis, there was evidence that the population IAPAR 26 was more efficient for discriminating lines from , than the reverse BR 106. Estimates of combining ability effects showed a higher dispersion in relation to the CDPCI scheme. ln TC1 triai, estimates of combining ability (ti = Yi -Y.) for PG varied from -1420 a 1607, with general mean of 7563 kg/ha. ln the sarne way, estimates of ti= Yi -Y. for PG in TC2 varied from -1211 a 1269, xíi with general mean of 7363 kg/ha. The average estimate of the interpopulation additíve variance ( cri 12 ) was 310.81 (g/pl) 2 . The overall analyses indicated that CDPCI scheme was more effícíent for progeny selectíon and predíction of hybrid combinations than CTC scheme, mainly as a function of the lower deviations in the combíning ability estimates and also for providing additional information on the effects of specific combining ability and on the interpopulatíon domínant varíance. t..J v,