Virus-like symptoms (due to banana streak virus, cucumber mosaic virus, or both) have been observed in plants of Musa hybrids (TMPx) and local landraces included in multilocational trials in sub-Saharan Africa. Virus-like symptom incidence in these multilocational trials was analyzed using the additive main effect multiplicative interaction (AMMI) model. There were significant differences in virus-like symptom incidence among environments, which was highest in the cool, rainy season (14% to 42%) and lowest in the warm, dry season (<10%). Genotypes showed significantly different responses to virus(es), which depended on the environment. There were no plants of AA and AAA bananas showing virus-like symptoms (0% incidence), whereas ABB cooking bananas and a cooking banana hybrid (ABB × AA) seldom showed virus-like symptoms (<2% incidence). The AAB French plantains appeared to have a similar genotypic response to virus(es) (about 10% virus-like symptom incidence) and were regarded as less susceptible than the False Horn plantain `Agbagba', which showed virus-like symptoms in most of the environments (average 21% incidence). Hence, `Agbagba' should be considered a susceptible indicator host because it has a stable susceptible host response to Musa virus(es). Plantain hybrids (AAB × AA) showed virus-like symptoms; however, there were significant differences in genotypic response to the virus(es) among various hybrids (11% to 60%). Epistasis due to transgressive segregation may control the susceptibility of TMPx germplasm to Musa virus(es). The AMMI1 model revealed that an increase in clonal susceptibility resulted in a more unstable response to the virus. Similarly, phenotypic instability was associated with an increase in clonal resistance. Environments with very low (dry season) or very high (rainy season) incidence of virus-like symptoms had unstable virus expression. Scoring virus symptoms in cool environments with low rainfall and low potential evapotranspiration provided an unbiased assessment of genotypic response to Musa virus(es). The AMMI2 model showed that seasonal rather than locational diversity accounted for most of the interaction patterns. This finding may indicate a low level of strain differentiation in the region.
There is a genuine need within a plantain and banana (Musa spp.) breeding program to assess thoroughly the experimental materials through a sequence of trials. This will result in the selection of promising clones as potential new cultivars in the targeted agroecozone. Stability analyses and the additive main effects and multiplicative interaction (AMMI) model provide together a means for the identification of clones with 1) homeostatic responses to environmental changes, 2) a genotypic response to environmental changes, and 3) adaptation to specific niches. Fourteen polyploid clones (10 tetraploid hybrids and 4 triploid cultivars) were evaluated in a broad range of environments in sub-Saharan Africa to determine the value of stability and AMMI analyses in Musa trials. The interpretation of the results, especially those concerning the genotype × environment interaction, was facilitated by the combination of stability and AMMI analyses. Tetraploid hybrids combining heavy and stable bunch mass were identified. The results also suggested that a clone should be assessed in the ratoon cycle because plantain and banana are perennial crops. Likewise, high yielding clones with specific adaptation should be selected in environments showing the respective environmental or biotic stress.
Rodomiro Ortiz and Dirk Vuylsteke
Rodomiro Ortiz and Juan Izquierdo
The environment substantially affects the performance of tomato (Lycopersicon esculentum Mill.) genotypes in Latin America and the Caribbean (LAC). Therefore, stability analysis can be used to select stable, high-yielding genotypes. Nine openpollinated and six hybrid tomato genotypes and the most representative local tomato cultivar were evaluated at 20 LAC locations. Each cultivar's yield stability was quantified using the regression of individual genotype's yield on the environmental index, which was measured by the mean of all the genotypes grown in an environment. A high-yielding and stable tomato cultivar had a mean yield higher than the general mean, b1 (coefficient of regression) = 1, s2 d (deviation from linearity) = 0, and r 2 (coefficient of determination) >0.50. `Narita' (hybrid) and `Dina RPs' (open-pollinated) were the most stable genotypes for marketable-fruit yield in LAC. `Flora Dade', an open-pollinated genotype that is grown widely in LAC had unstable marketable-fruit yield. Neither heterogenous composition of an open-pollinated genotype nor heterozygosity per se of the hybrids could explain the yield stability achievement across environments. Therefore, alleles that confer broader adaptation might be required to achieve tomato yield stability across environments. Hence, it is possible to select for yield stability in tomato.
Rodomiro Ortiz and Dirk Vuylsteke
Two multilocational trials, one comprising 18 Musa clones in three locations and another of 20 genotypes across 11 locations, were set up in 1991 and 1992, respectively, to assess the genotype-by-environment interaction (GxE) for important traits and to select stable high-yielding and black sigatoka (BS)-resistant genotypes. Combined ANOVAs showed significant differences among environments and among genotypes for all traits. GxE affected all growth and yield parameters, except fruit girth. Host response to BS disease also showed significant GxE, but there was no cross-order season-by-year interaction. Hence, genotypic response to BS can be assessed in 1 year during the rainy season, when disease pressure is highest. Genotype-by-location effects were more important than the nonsignificant genotype-by-year effects, supporting the need for multilocational trials. Stability analysis showed that full-sib plantain hybrids (TMPx1) exhibited different host responses to BS as well as different interaction patterns, suggesting that selection for stable BS resistance is possible. The BS-resistant TMPx genotypes had higher yields than the plantain landraces, but showed differences in yield stability. TMPx 1658-4, 2796-5, 5511-2, and 6930-1 have been selected as stable high-yielding hybrids, while the initial best selections (TMPx 548-4 and 548-9) were top yielders only in good environments. [Vuylsteke, D., R. Swennen, and R. Ortiz. 1993. Registration of 14 improved tropical Musa plantain hybrids with black sigatoka resistance. HortScience 28:957–959.]
Dirk Vuylsteke and Rodomiro Ortiz
Dirk R. Vuylsteke and Rodomiro Ortiz
In vitro-propagated plants of plantain (Musa spp., AAB group) did not manifest consistently superior horticultural performance compared to conventional propagules. Tissue culture plants grew vigorously and taller than sucker-propagated plants, but higher yield was not obtained, probably because of severe disease and suboptimal husbandry input. Phenotypic variation was higher in tissue culture plants, although this increase was not always statistically significant. There were no other detrimental effects of in vitro propagation on field performance. Botanical seed set rates for the two types of propagules were similar. The advantages of tissue-culture-derived plants as improved planting material would be most relevant for establishing field nurseries for further clean, conventional propagation of newly bred or selected genotypes.
Boniface B. Dumpe and Rodomiro Ortiz
Current efforts to produce improved genotypes of plantain and banana (Musa spp.) depend on crossing female-fertile clones with accessions that produce viable pollen. Musa accessions (168) were screened for production of viable pollen based on staining with acetocarmine glycerol jelly. Diploid hybrids and landraces produced significantly more pollen than triploids and tetraploids, suggesting more successful crosses when using diploid accessions as male parents. There was a positive correlation between the amount of pollen produced and the level of viability in both hybrids (r = 0.65, P≤ 0.01) and landraces (r = 0.61, P≤ 0.01). This finding suggests that closely associated genetic factors determine these characteristics in Musa, while environmental conditions also may influence the quality and quantity of pollen produced. Pollen production at anthesis was absent in 28 accessions. Of the 140 accessions with pollen, 67 were sufficiently fertile for use as male parents in the breeding program.
Rodomiro Ortiz and Dirk R. Vuylsteke
Apical dominance, i.e., the inhibition of lateral bud growth due to growth substances released by the terminal bud, has been considered as a limiting factor for the perennial productivity of plantains (Musa spp., AAB group). Segregation ratios in F1 and F2 plantain-banana hybrids suggest that inheritance of apical dominance is controlled by a major recessive gene, ad. The dominant Ad allele improved the suckering of plantain-banana hybrids, as measured by the height of the tallest sucker at flowering and harvest. At harvest, the ratoon crop of the diploid and tetraploid hybrids had completed 70% to 100% of its vegetative development, whereas the ratoon of the plantain parents, due to high apical dominance, was only at 50% of total pseudostem growth. Sucker growth rates are generally the result of gibberellic acid (GA3) levels, and it is suggested that the Ad gene regulates GA3 production. However, the Ad gene has incomplete penetrance, genetic specificity, and variable expressivity. Increased frequency of the Ad gene and a commensurate improvement in the suckering behavior of the diploid populations may be achieved by phenotypic recurrent selection.
Rodomiro Ortiz and Dirk R. Vuylsteke
Few genetic markers are available in Musa spp. as a result of a lack of inheritance studies. Full-sib diploid (2n = 2x = 22) plantain-banana hybrids of the International Institute of Tropical Agriculture were selfed or outcrossed with other diploid bananas, one of which is an improved selection from Central America. Three populations having albinos (complete lack of chlorophyll in any plant tissue) were produced. The segregation ratios for albinism suggested that this deleterious trait is controlled by one or two recessive alleles. The small sample sizes (a problem inherent in the low reproductive fertility of cultivated parthenocarpic Musa) in two of these three populations did not allow for conclusiveness between the one or two genes model. However, a distinction was possible with the third population, consisting of 64 seedlings, of which four were albinos. The segregation ratio for albinism fit the 15:1 ratio (χ2 = 0.07, P = 0.79) and not the 3:1 ratio (χ2 = 11.02, P < 0.01), suggesting that albinism in Musa spp. is under the genetic control of at least two independent recessive alleles with complementary gene action. This finding also demonstrates that deleterious recessive alleles are present in the cultivated AAB plantain gene pool and in cultivated and advanced AA banana breeding populations. The latter suggests that population improvement through phenotypic recurrent selection for agronomic traits might be based on the elimination of deleterious recessive genes.