OPULATION CHARACTERISATION AND PROPAGATION STRATEGYFOR WILD Coffea eugenioides IN NYUNGWE AND CYAMUDONGO FORESTS, RWANDA: A field assessment of mother trees and an evidence-based seed-source pathway.

Abstract

This report presents a field assessment of wild Coffea eugenioides, the diploid maternal progenitor of cultivated Arabica coffee, in Nyungwe Forest National Park and the adjoining, isolated Cyamudongo forest fragment in south-western Rwanda. The assessment targeted trees identified as candidate mother trees (individuals bearing fruit and therefore potential seed sources): 59 such trees were georeferenced and measured for structural, phenological and reproductive traits using a structured KoBoToolbox questionnaire (reproduced in Annex 1). The work was undertaken to inform the establishment of a seed orchard for this species under the RTRP-Seed (IKI) and TREPA (GCF) projects implemented by Landscape Alliance (CIFOR & ICRAF in action). On C. eugenioides, the cherry is the fleshy fruit that contains the seed, so cherry load is a direct measure of fruit and seed output. Reproductive output across the assessed trees is low, highly variable and weakly related to tree size. The median tree carried an estimated 29 cherries in the whole canopy, 31% of trees carried 15 cherries or fewer, and the coefficient of variation in cherry load exceeded 110%. Tree size explained little of this variation (stem diameter, the single significant structural predictor, accounted for under 10% of the variance; r = 0.31). Trees in the isolated Cyamudongo fragment carried significantly fewer cherries than those in the main Nyungwe block (medians 15 vs 32 cherries; Mann-Whitney p = 0.021), and every observation of sub-optimal on-tree seed quality occurred in the fragment. Low and asynchronous fruiting, together with the reduced fruit and seed set typical of small, isolated populations, makes large-scale seed collection an unreliable basis for an orchard. The original plan was to collect seed from wild trees and raise a Breeding Seed Orchard (BSO). On the evidence presented here, that plan should be redirected toward a Clonal Seed Orchard (CSO): rather than collecting seed, scions (budwood) are taken from selected wild trees and grafted onto vigorous local Rwandan Coffea arabica rootstock. This captures the full genotype of each selected tree, shortens the time to flowering, secures the germplasm of a threatened wild relative ex situ, and aligns with Rwandas restoration and climate commitments.