4.3. Slow growth as a strategy to cope with drought stress and evidence of a trade-off between growth and drought tolerance.
Genotypes that had low RGRA TNL, TL and TLDW values in ample-water conditions were comparatively less affected by drought, a scenario which indicates a trade-off between growth and drought tolerance across the study populations (Fig. 3). This finding concurs with an established ecological paradigm that there is a trade-off between the capacity of plants to grow fast when resources are abundant and their capacity to tolerate resource shortages (Bazzaz and Bazzaz, 1996; Aerts and Chapin, 1999; Grime, 2006). The trade-off between growth and tolerance has been linked to a conservative resource-use strategy in which slow growth results in slow tissue turnover (i.e. conservative use of resources) and subsequently less dependency on the environment for acquisition of new resources. On the contrary fast growth is associated with high resource turnover rates, intensive resource acquisition, high dependency on the environment and ultimately shorter lifespan (Chapin, 1980; Chapin III, Autumn and Pugnaire, 1993; Grime et al. , 1997; Reich et al. , 2003; F. J. Sterck, Poorter and Schieving, 2006; Sterck et al. , 2011). Ecologically, slow growth has been reported as an adaptive strategy for plants in resource limiting conditions. Poorter, (1989) studied the ecological consequences of the interspecific variation in relative growth rate (RGR) of plants and concluded that differences in potential RGR between species were habitat-related whereby fast-growing species were found in resource-rich habitats while slow growers could be found in any adverse environmental condition. In response to drought, a growth-tolerance trade-off could be expected because several traits and mechanisms that confer tolerance in dry conditions (e.g. low specific leaf area, low stomatal size or number) reduce water loss but also reduce rates of net photosynthesis per unit area, which, in turn, results into slower growth under favourable water availability (Lambers, Chapin and Pons, 2008; Stercket al. , 2011).
Although the growth-tolerance trade-off has been widely studied and established across species (interspecific), including tropical forest trees (Poorter and Jong, 1999; F J Sterck, Poorter and Schieving, 2006; Sterck et al. , 2011; Amissah et al. , 2018) much fewer studies (Pallardy and Kozlowski, 1981; Silva et al. , 2013; Menezes-Silva et al. , 2015) have been conducted to explore the intraspecific variation of tropical trees to drought and the manifestation of the growth-tolerance trade-off. Pallardy and Kozlowski, (1981) revealed a probable growth-tolerance trade-off amongPopulus clones: fast-growing clones had a larger initial rate of decline in leaf water potential with transpirational flux density but reduced the rate of decline more than slow-growing clones as the transpirational flux density increased. Similarly, Menezes-Silva et al., (2015) and Silva et al., (2013) studied eight clones of cultivatedC. canephora (variety Conilon) and found that wood density, a trait that partially influences the plant’s water-conducting capacity, was higher in drought-tolerant clones, and was associated with greater resistance to cavitation. This adaptation however could limit growth under favourable water conditions as dense wood is more costly to produce and the associated smaller xylem have lower maximum water conductance (Silva et al. , 2013; Menezes-Silva et al. , 2015).
In our study, the relatively low RGRA and high tolerance of genotypes from Kibale, Itwara and Zoka locations (Table 4; Appendix Fig. A.3.) suggests that those populations employ a more conservative resource-use strategy, while genotypes from Mabira, Malabigambo, Kalangala and Budongo employ a more rapid resource-acquisition strategy. Similar to our results, Silva et al.,(2013) and Menezes-Silva et al.,(2015) also found that across a set of cultivated C. canephora clones, the most drought-tolerant ones tended to be slow growers. Slow growth in stressful conditions could in the long term be more adaptive than fast growth because fast growth results in larger and more resource-demanding plants that could eventually die off if the resource demand is not met. Here, we showed the existence of a growth-tolerance trade-off across a large set of wild accessions of a perennial crop species, suggesting that intraspecific variation in tolerance may be related to selection in natural environments. Evidence of a growth-tolerance trade-off in our study is further corroborated in our related experiment by Kiwuka (2020) where we studied fewer (15) genotypes with more response traits and found that slow-growing genotypes were more drought tolerant and less plastic for most of the response traits.
In interpreting our findings, it should be noted that our experiment was a pot experiment and pots have limited volume. Firstly, this could cause a so-called pot-binding effect (Poorter et al. , 2012; Sinclairet al. , 2017); pots holding insufficient water to support transpiration and therefore growth. This could be more severe for fast-growing plants than for slow-growing ones. However, in our set up we accounted for this effect as we determined the relationship between water consumption and plant size and adjusted the amount of water gift in restricted water treatment to correct for larger plants consuming more water (see section 2.4). Therefore, we are confident that larger plants did not suffer greater drought stress than smaller ones in the water-restricted treatment and that the pot-binding effect was minimised as seen in Appendix Plate. A.1 and Appendix Plate. A.2. Secondly, in the field, rooting depth can be a drought adaptive trait as it allows access to deeper moister soil layers. This effect evidently could not be mimicked in pots. Association of rooting depth with growth potential could be either positive (fast-growth facilitating deeper roots) or negative (larger deeper root systems imposing greater metabolic costs and therefore, slowing growth). Altogether, it is important to determine whether the drought-tolerance trade-off found in our study also occurs in the field. If the observed growth-tolerance trade-off occurs under field conditions, it would pose a dilemma for breeding on what to select for if one cannot have both. For instance, selecting fast growth could result in low drought tolerance which poses a challenge especially for small scale coffee farmers who may not have irrigation facilities to deal with drought spells. Therefore, to sustain C. canephoraproduction in drought-prone environments, breeders should break the negative correlation between poor performance and tolerance (Table 4; Fig. 3). This proposition agrees with Damatta et al., (2018) who suggested that breeding for drought tolerance in coffee should aim at developing tolerant genotypes with ”acceptable yields”. Despite the adaptive advantage of slow growth (conservative resource-use strategy), its positive association with low performance is also a challenge as farmers are interested in good yields. Selection for either slow or fast-growing genotypes should therefore be done in consideration of whether the intended production is in stressful or optimal conditions.