4.2. Variation in response across cultivation status, genetic groups and location
Our findings indicate that there is a clear genotypic variation in performance (RGRA) both under ample and restricted-water conditions (Fig. 1 and Appendix Fig. A.3.). The variation in RGRA was larger (more than two-fold) under ample-water than restricted-water conditions (Table 3, Fig. 1 and Appendix Fig. A.3.). The different phenotypic responses of genotypes in ample and restricted-water conditions (Fig. 1 and Appendix Fig. A.3.) probably reflects an underlying genetic polymorphism that may drive different phenotypic responses to different environments (Stearns, 1989; Pigliucci, 2005; Forsman, 2015). The observed genotypic variation in our study in both growth and drought tolerance can be utilised for optimizing breeding programs initiatives to develop drought-tolerant varieties with adequate yield capacity (Table 3; Fig. 1 and Appendix Fig. A.3.). Results did not show significant variations in RGRA between genotypes of different cultivation status (wild, feral or cultivated). This probably indicates that Uganda’s breeding efforts to date have not addressed drought tolerance. Breeding efforts have been focusing on other factors e.g. yield and resistance to pests and diseases, in particular generating wilt disease-resistant coffee varieties (Musoli et al. , 2008). Breeding efforts in C. canephora are relatively limited, partially due to the perennial nature of the crop (with an economic lifespan of about 20 years), which suggests that most of the cultivated material is still very similar to the wild genotypes (Thomas, 1935; Montagnon, Eskes and Leroy, 1998; Ngugi and Aluka, 2019). Indeed, Kiwuka et al., (2021) found that Uganda’s cultivated genotypes were genetically similar to wild populations from Malabigambo, Mabira and Kalangala forests.
Across experimental factors we studied, location exhibited the widest range of reductions in RGRA from 7.1% to 36.7% in Kibale and Malabigambo respectively (Table 4). The genetic distinctiveness of Uganda’s wild C. canephora populations across locations as shown in Kiwuka et al., (2021) (Appendix Table A.1.) and their differential phenotypic response to drought (Table 4 and 5; Appendix Fig. A.3.) indicate that Uganda’s C. canephora diversity could be locally adapted to the climatic conditions within the locations. The significant interaction effect between genetic group and water treatment (Table 3) also provides evidence that the localisation of the genetic groups (i.e. Zoka, Itwara, Kibale and Budongo genetic groups from the NW) could be associated with genetic effects and putatively to adaptive potential. The possibility of genotypes being locally adapted is also indicated by an overlap between the genetic group (Fig. 1) and location (Appendix Fig. A.3.) effects on RGRA. However, the strong effect of location on response to drought could also be reflecting local differences in other factors such as soil types that may influence selection for the difference in growth-related traits.