Differential vegetative and reproductive performances among fifteen guinea grass hybrids

The main scope of this work was to detect (Panicum maximum Jacq.) genotype differences as to morphoagronomic and seed quality indices, and to establish character correlations useful for determining vegetative and reproductive trends. Besides the flowering cycle, eight phenological and two seed quality traits were scored in a greenhouse randomized complete block experiment, as follows: plant height (PH), reproductive tiller number/overall tiller number (RTN/OTN), panicle number/reproductive tillers (PN/RT), leaf length (LL), leaf width (LW), panicle length (PL), fresh weight (FW), dry weight (DW), number of seeds/g (NS/G) and seed sample physical purity (SPP). Very-early and early-flowering hybrids consistently showed the highest correlation values among flowering cycle and RTN/OTN (r = -0.59**), PN/RT (r = -0.48**), NS/G (r = -0.88**) and SPP (r = -0.80**) (reproductive parameters) while intermediate and late-flowering hybrids presented the highest values for LL (r = 0.53**), LW (r = 0.60**), PL (r = 0.77**), FW (r = 0.78**) and DW (r = 0.85**) (vegetative traits). The implications of these results for plant breeding and forage management purposes are discussed.


Introduction
Guinea grass (Panicum maximum Jacq.) is a warm-season perennial bunchgrass widely grown as a forage crop in tropical and warm-temperate regions of both hemispheres.Like most tropical grasses, it is a facultative apomictic species, where apospory and pseudogamy occur during the reproductive process.
Most of the above studies were performed under field conditions, not taking into account flowering cycle differences among cultivars/introductions, so making data comparisons difficult.In addition, none has been set up to evaluate vegetative and reproductive functions among genotypes widely variable for several traits.
The main scope of this research has been to detect genotype differences, as to morphoagronomic and seed quality indices and to establish character correlations, useful to determine vegetative and reproductive trends.

Material and Methods
Fifteen guinea grass hybrids derived from artificial crossings among previously selected highly sexual (female) and apomictic materials (male) were evaluated (Table 1).Seeds of each hybrid were sown in germination boxes, filled up with a mixture of topsoil, sand and organic matter (3:1:1 by volume).Fertilizer was added to soil according to soil analysis results.After four weeks, 30 individual seedlings per hybrid were transplanted to plastic bags (10 x 15 cm), filled up with the same soil mixture and placed in a greenhouse in a randomized complete block design with three replications (ten plants per replication).
During the experiment, all plastic bags were watered daily.At the onset of the flowering period, the number of days to flowering was scored for each individual plant.As the hybrids reached full blooming, the panicles were packed together in cloth bags (in the same treatment) to avoid seed shattering.
At harvest, the seeds were picked up from the bags and eight vegetative traits (plant height, reproductive tiller number/overall tiller number, panicle number/reproductive tiller, leaf length, leaf width, panicle length, whole plant fresh and dry weights) were scored in each treatment.Additionally, two seed quality tests (number of seeds/g and seed sample physical purity) were performed, according to International Seed Testing Association (1985).Plant height has been scored from soil surface to the panicle apex.
Measurement data (plant height, leaf length, leaf width, panicle length, fresh and dry weight) were used per se while countings (days-to-flowering, panicle number/reproductive tiller and number of seeds/g) and percentage data (reproductive tiller number/overall tiller number and seed sample physical purity) were transformed to 5 .0 x and arc sin , respectively, before statistical analysis.An ANOVA computer program was used to test differences among hybrids and mean comparisons of different characteristics were made using Duncans multiple range test.Finally, simple correlations were calculated among flowering cycles and all the quantitative traits studied.

Results and Discussion
The hybrids studied were ranked in four different groups, according to their flowering cycle: late, intermediate, early and very early-flowering types, using statistical analysis and/or consideration of (1) SEA: South East African; PI: Plant Introduction; T 1 SL: Tobiatã1 sexual line; T 2 SL: Tobiatã2 sexual line; C 1 SL: Centauro1 sexual line; C 2 SL: Centauro2 sexual line; C 7 SL: Centauro7 sexual line.
Table 1.Pedigree of the 15 guinea grass (Panicum maximum Jacq.)F 1 apomictic hybrids (1) .The genetic materials used revealed wide genetic diversity for all the parameters studied under controlled greenhouse conditions, similar to that observed among cultivars/ecotypes of the same species in field trials (Costa et al., 1989;Alcantara et al., 1991;Segui et al., 1992;Sun & Liddle, 1993).
Based on these results, the best strategy to preserve the variability within the species should take into consideration the flowering cycle of the available genetic materials.So, a rational and comprehensive germplasm bank should present several groups of ecotypes, introductions and cultivars with different Table 2. Phenological and seed quality indices recorded on 15 guinea grass (Panicum maximum Jacq.) hybrids in a randomized complete block experiment (1) .
flowering cycles.As a consequence, its maintenance and even the choice of genetic materials for breeding purposes would be easier.In this case, early stageselection schemes would be feasible (for example, when the selection is aimed at short and narrow leaves, in a wide array of genotypes, the selected types, in most cases, will belong to the early or very early-flowering groups).For forage management purposes, higher pasture persistence should be achieved by using early and very early-flowering hybrids, taking into account their higher seed yielding potentials (high reseeding rates in the field); however, due to their limited forage yields, lower stocking rates should be applied, as compared to those of intermediate and late-flowering cycle genotypes.

Conclusions
1.The genotypes used reveal wide diversity for all the parameters studied.
2. Significant high and positive simple correlations are obtained among flowering cycle and plant height, leaf length, leaf width, panicle length, fresh weight and dry weight.
3. Late and intermediate-flowering hybrids show higher values for vegetative traits; on the other hand, early and very early-flowering hybrids reveal a significant negative correlation with reproductive parameters.