Breeding system of Dacryodes edulis (G. Don.) H. J. Lam: implications for cultivars development, selective breeding, and conservation of genetic resources

Abstract

Growing food insecurity, increasing poverty and deteriorating environmental conditions are the main scourges that profoundly affect rural populations in the tropics. To help solve these problems, the World Agroforestry Centre, former International Centre for Research in Agroforestry (ICRAF) encourages integrated and improved management of natural resources through the promotion of agroforestry systems. Nonetheless, success in breeding for higher yield of indigenous fruit trees is still constrained by the lack of availability of improved germplasm. Improvement of indigenous fruit tree cultivars is drawing great attention from breeders. Dacryodes edulis (G.Don) H. J. Lam (Burseraceae) commonly known as African plum is an andromonoic and oleaginous fruit tree species, endemic to the Gulf of Guinea. It is currently under domestication and widely commercialized in West and Central Africa. In efforts to enhance the species’ genetic conservation and utilization, it was identified as one of the top ten agroforestry tree species for future crop diversification in West and Central Africa. Despite the advantages of vegetative propagation techniques which are hastening sexual maturity and giving an exact replicate of the desired mother-tree characters, it is feared that it could severely narrow the genetic diversity and increase inbreeding within the species at farm level ultimately leading to a decline in future production. However, controlled cross-pollination can help to combine a number of desired fruit characters in one tree and increase inter-tree variability between selected superior genotypes so called “plus trees”. By so doing, the species’ genetic basis could be better conserved. Our goal in this research was to beef up Dacryodes edulis tree breeding and genetic resources’ preservation initiated by ICRAF throughout a tree improvement program, especially through cross-controlled pollinations and vegetative propagating techniques being developed in Cameroon. Specifically, we addressed the following questions: (1) Does African plum fruiting efficiency depend on the parents’ provenance and the type of flower which produced the pollen used for fertilization? (2) Do controlled cross-pollination increases quality of the quantitative fruit traits of well-known African plum accessions? Which Cameroonin African plum’s ecological provenance possesses best fruits traits for selection through controlled cross-pollination? (3) Do seedlings obtained through controlled cross-pollination of African plum behaviors in the nursery depend on the provenance and the crossing to which they belong? In this perspective, controlled hand cross-pollinations were performed on 25 accessions of D. edulis, using a nested mating design. Conventional methods (ANOVA, MANOVA, Ascending Hierarchical Classification and Principal Component Analysis) were used for data analysis. Output data tested were especially size and weight of fruits and more precisely pulp weight or thickness, which is the principal trait of commercial importance. The pollination experiments were performed during three flowering seasons between January 2010 and March 2012 at two localities, Minkoa-Meyos and Mbalmayo in the Centre region of Cameroon. The experimental design included provenance as a fixed factor, treatment as within-subject (i.e. repeated measures) fixed factor and plant individual as a random factor (subject). Although controlled pollination did not increase fruit size in the studied species as compared to other fruit trees such as avocado and citrus specie, fruit set and fruiting were greatly improved. This technique may help to reduce the phenomenon of fruit drop very current in D. edulis. In addition, six best combinations mainly characterized by high fruit-setting rate and fruiting index (˃70% and ˃50% respectively) and a low fruit-dropping rate (˂20%) were identified. Variation in fruiting index that determines fruit yield of the studied species was highly correlated with the combined action of factors studied ((i) provenance of the male parent: Boumnyebel (BUM29), Makenene (MAK33) and Kekem (KEK02); (ii) type of flower (male or hermaphrodite) which produced pollen used for hand fertilization and (iii) female parent status. In order to improve yield, it was also noticed that it is necessary to control fruit load by exploring ways of reducing fruit set during flowering. Phenotypic variation in the traits of 1604 fruits and seeds were characterized within and between trees, and between provenances. Ascending hierarchical classification (cluster analysis) performed showed that variability in relation between fruit and pulp weight confirms the moderate differences between clusters and may have been driven by both ecological and genetic variation. For the fulfilment of cultivar development, the perceived qualitative variation (in pulp oil content and pulp taste) may be genetically determined and should not be neglected during the characterization of F2 and F3 hybrids (further research). Nevertheless, between provenances variation was found to be relatively high, particularly for fruit length, fruit width, fruit weight, pulp weight and thickness. The results suggest that fruit weight is a good predictor of pulp yield, although its predicting power differed among clusters. This study has helped develop a mathematical model for the choice of fruit and pulp size by breeders for D. edulis breeding. Additionally, control-pollinated seedlings obtained from the present study were considered as improved planting material and submitted to early decapitation test. Such approach may produce young seedlings with multiples branches onto which cuttings could be obtained. This approach may also help to characterize African plum control-pollinated seedlings in an attempt to select improved raw-material for the on-going plant breeding program. Thus, early branching induction by the removal of the apex or the uppermost node of 108 selected F1 hybrids in the net-house produced seedlings with early multiple branches (synchronous branching). This architecture may contribute to yield improvement and also facilitate farmers’ handling experiments. In fact, “apical dominance” effect was negative. There was no significant difference from one sprouted seedling to another, regardless of seedling trait targeted and the parent provenance. F1 hybrids were settled on-station as progeny trials for the on-going breeding program. They will be carefully monitored until the first fruit production during which qualitative agro-morphological parameters such as pulp oil content, taste and color of the mesocarp and epicarp of the fruit, will be evaluated and compared to those of parents or selected mother trees (evaluation of the genetic gain in F2 and F3 generations). F2 hybrids will be established within agro-ecological zones around the country as clonal or progeny trials in the fulfilment of cultivars development. This study has depicted number options for the genetic improvement of D. edulis. In fact, the results of the overall study show that hand pollination can help to combine a number of desired fruit characters of one tree and increase the inter-tree variability between plus-trees, without narrowing its genetic base. Coupled with vegetative propagation techniques (layering and leafy stem cuttings) being developed for D. edulis breeding, controlled hand pollination can enhance the genetic resources’ conservation of African plum leading to new varieties or cultivars development for the enrichment of African plum plantations with high-quality planting stock.