Desempenho de populações segregantes de algodoeiro obtidas por três métodos de hibridação

Abstract

CHAPTER I - The objectives of this research were to make predictions on segregant populations performance on upland cotton as for selection, using the method proposed by Jinks and Pooni (1976), as well as to identify amongst three different crossing methods, which will originate the most promising population for selection purposes. The present work was conducted at the Universidade Federal de Uberlândia during the period 2001 / 2005. The parental genotypes had been selected by means of the results of a diallel cross in which the parentals envolved in the hybrid combinations with higher specific combining ability, were the ones chosen to be bred, according with three different methods: the multiple crossing, biparental simple cross and the convergent cross. During the season 2004/05 thirteen populations originated of those crossings, plus one check variety, were planted according to a complete randomized-block design with four repetitions. Each experimental plot consisted of four lines five meters long with the two central rows used for collecting data. The characters analyzed were: seedcotton yield, lint percent, length uniformity, fiber strength, elongation and fiber strength. The method of Jinks and Pooni (1976) was applied only to the trait seedcotton yield. It was observed that populations MGUFU012, MGUFU013, MGUFU015, MGUFU0110 and MGUFU0111 had greater values for probabilities (47,61%, 46.41%, 49,60%, 42.47% and 38.21% respectively) for selection of superior lines. Thus, those populations must be selected to continue in the improvement program, therefore certainly they will be able to yield superior genotypes upon selection. In contrast, populations MGUFU018 (8%), MGUFU0113 (4,85%), MGUFU016 (11,12%) and MGUFU014 (15,87%) had presented reduced probabilities to originate superior plant selection and are prone to be discarded in the early generation stage. Promising segregant populations were obtained by both multiple crossing and simple biparental crosses. The convergent crossing presented intermediate. CHAPTER II - The constant preoccupation on the generation and maintenance of genetic variability of the species in breeding programs lead breeders to the use of different strategies of obtaining segregant populations. The objectives of this work were: to evaluate the performance of families selected out of populations that had presented better performances according to the methodology proposed by Jinks and Pooni (1976); to prove the efficiency of this method for early selection in upland cotton and to verify which method of hybridization is more efficient in the attainment of populations with greater genetic variability. In the 2004/05 growing season, populations were selected which had presented previously the best performances in accordance with the methodology developed by Jinks and Pooni (1976), MGUFU 012 and MGUFU 015 as well as, in contrast, populations that had presented the poorest performances (MGUFU 018 and MGUFU 0113), together with population MGUFU011, originated by convergent hybridization (genitors crossed two by two to form a heterogametic genitor that was then crossed with a commercial variety). From these five populations nineteen to twenty plants were randomly selected, originating 98 F3:4 families which were field evaluated in the 2005/06 growing season in Uberlândia, MG. The experimental design used was a three-replicate lattice with 100 treatments after the addition of the two commercial varieties IAC 23 and Delta Opal. Sowing was carried out in December 23, 2005. Experimental plots were constituted of two rows 2,5 m length each, spaced 0,90 m apart, with seven plants per meter. The characters evaluated were seed cotton yield (kg ha-1) and lint percent (%). It was evaluated the components of variation and heritability estimates for the traits in question. The results obtained were not able to confirm the efficiency of Jinks and Pooni method as applied to cotton breeding selection. It was observed that the population originated through convergent crosses has produced families with higher potential for selection when compared to the other crossing systems.