Fruit thinning and shade improve bean characteristics and beverage quality of coffee (Coffea arabica L.) under optimal conditions

Journal of the Science of Food and Agriculture J Sci Food Agric 86:197 204 (2006) DOI: 10.1002/jsfa.2338 Fruit thinning and shade improve bean characteristics and beverage quality of coffee (Coffea arabica L.) under optimal conditions Philippe Vaast, 1,2 Benoit Bertrand, 2 Jean-Jacques Perriot, 2 Bernard Guyot 2 and Michel Génard 3 1 Centro Agronómico Tropical de Investigación y Enseñanza (CATIE), Apdo 3, 7170 Turrialba, Costa Rica 2 Centre de Coopération Internationale en Recherche Agronomique pour le Développement (CIRAD), 2477 Avenue du Val de Montferrand, BP 5035, F-34032 Montpellier, France 3 Institut National de la Recherche Agronomique (INRA), Plantes et Systèmes de Cultures Horticoles, Domaine Saint-Paul Agroparc, F-84914 Avignon Cedex 9, France Abstract: Under two contrasting light regimes (full sun and 45% shade) and the optimal coffee-growing conditions of the central valley of Costa Rica, production pattern, bean characteristics and beverage quality were assessed over two production cycles on dwarf coffee (Coffea arabica L. cv. Costa Rica 95) trees with varying fruit loads (quarter, half and full loads) imposed by manual fruit thinning. Shade decreased coffee tree productivity by 18% but reduced alternate bearing. Shade positively affected bean size and composition as well as beverage quality by delaying berry flesh ripening by up to 1 month. Higher sucrose, chlorogenic acid and trigonelline contents in sun-grown beans pointed towards incomplete bean maturation and explained the higher bitterness and astringency of the coffee beverage. Higher fruit loads reduced bean size owing to carbohydrate competition among berries during bean filling. These results have important implications in terms of agricultural management (shade, fruit thinning, tree pruning) to help farmers increase coffee plantation sustainability, produce coffee beans of larger size and higher quality and ultimately improve their revenues, especially during times of world overproduction. 2005 Society of Chemical Industry Keywords: bearing pattern; bean biochemical content; berry maturation; coffee beverage quality; fruit load; shade INTRODUCTION Until recently the emphasis in coffee research has focused mainly on orchard management practices to increase coffee productivity and on breeding programmes to enhance resistance to pests and diseases. Owing to the current low market prices caused by world overproduction, there is a strong interest in producing and marketing coffee of higher quality to alleviate financial difficulties encountered by coffee farmers. Coffee quality is mainly assessed through the physical aspects of coffee beans such as bean colour, size, density and percentage of physical defects in producing countries, whereas cup quality is the main criterion in consuming countries. Numerous factors affect coffee quality, 1 including soil water status, 2 climatic conditions, 3,4 maturity of coffee berries at harvest and bean processing (fermentation, washing, drying, storage, roasting, beverage preparation), agricultural management (shade, pruning, fertilisation) and genetic properties of cultivars. 5,6 Arabica coffee (Coffea arabica L.) is a cash crop of major economic importance in Central American countries, which have a long-lasting reputation for producing coffee of high quality. In this region, Arabica coffee was traditionally grown under shade in complex agroforestry systems with up to three storeys of vegetation. 7 In the late 1970s, however, the rapid development of leaf rust disease (Hemileia vastatrix)led to the planting of a new generation of resistant, dwarf cultivars. Compared with traditional ones, these dwarf cultivars have a more compact canopy and shorter branches, allowing higher planting densities under their own dense shade and hence increased production per hectare. 8 This has radically modified agricultural practices, especially pruning and fertilisation regimes, and often resulted in the complete elimination of shade trees. However, these full-sun and intensively managed coffee systems are not recommended in the absence of optimal ecological conditions. Furthermore, these systems are not only more economically risky but Correspondence to: Philippe Vaast, Centro Agronómico Tropical de Investigación y Enseñanza (CATIE), Apdo 3, 7170 Turrialba, Costa Rica E-mail: pvaast@catie.ac.cr Contract/grant sponsor: European Commission; contract/grant number: CASCA ICA4-CT-2001-10071 Contract/grant sponsor: Science and Cultural Cooperation Centre, French Embassy, Costa Rica (Received 23 March 2004; revised version received 7 December 2004; accepted 25 May 2005) Published online 10 October 2005 2005 Society of Chemical Industry. J Sci Food Agric 0022 5142/2005/$30.00 197

P Vaast et al. also less ecologically sustainable. The presence of shade trees is known to improve soil organic matter content, alleviate high solar irradiance, buffer detrimental diurnal changes in air temperature and humidity 9 and reduce nutrient leaching, especially nitrate which contaminates aquifers. 10 Additionally, products derived from associated timber and fruit trees help farmers to diversify their income. Finally, shade trees play an important role in the region owing to the valuable impact of coffee agroforestry systems on the environment and natural resources (preservation of biodiversity, soil conservation, water quality, carbon sequestration). Recent studies in Guatemala 11 and Costa Rica 12 have demonstrated that elevation and shade improved coffee quality owing to cooler climatic conditions and probably a longer ripening period of coffee berries. In 1999 a collaborative research effort was developed in Central America to compensate low coffee market prices by promoting coffee agroforestry systems to improve coffee farmers incomes through diversification (timber production), production and commercialisation of high-quality coffee and payment of incentives for environmental services provided by these ecologically sound coffee systems. Within this research framework, several scientific investigations have been undertaken to determine the importance of factors such as microclimatic conditions, tree productivity, berry position within the canopy, shade management and fertilisation regimes on coffee tree physiology and beverage quality. The present study focuses on the effects of tree productivity, manipulated by fruit thinning, and light regime on the size and biochemical composition of coffee beans and their impact on beverage quality. MATERIALS AND METHODS Experimental location and design The study was carried out over three production cycles, from March 1999 to February 2002, on Arabica coffee (C. arabica L.) trees of the dwarf cultivar Costa Rica 95 planted under the optimal coffee-growing conditions of the Coffee Research Centre (CICAFE), Heredia, Costa Rica at 1180 m elevation, with an annual average temperature of 20.5 C, on an Andosol soil and with 2200 mm of annual rainfall. The plot was initiated in 1997 without shade. Costa Rica 95 has a maximum height, after four to five years, not exceeding 2.5 3.0 m. Plant spacing was 2 m between rows and 1 m within rows. Plants received 144 kg N, 24 kg P, 80 kg K, 64 kg Mg and 4 kg B ha 1 annually, split into two applications in May and August, and an additional application of 96 kg N ha 1 in early October. In March 1999, four groups of 24 adjacent trees were selected within the plot in adjacent rows; two groups of trees were maintained in full sun (sun) and the other two groups were shaded (shade) with a net allowing the passage of only 55% of the photosynthetic photon flux density (PPFD). Within each group, three fruit loads (F, full initial fruit load; 1/2, half of initial fruit load; 1/4, quarter of initial fruit load) were imposed on the coffee trees, resulting in six combinations of light levels and fruit loads with 16 replicates per combination. Fruit thinning was performed manually a few days after flowering at the beginning of the 1999 and 2000 production cycles. In 2001, fruit thinning was not undertaken, to test the effects of previous fruit loads on tree productivity. Furthermore, only coffee production was assessed for this third production cycle (2001). Leaf area, leaf-to-fruit ratio, leaf temperature and irradiance measurements The area of individual leaves was measured with a LICOR 1800 area meter (LICOR, Lincoln, NE, USA). Measurements were performed on three branches (branch positions 10, 17 and 25 as seen from the top of the coffee tree) positioned in the main producing zone of the canopy. From these measurements at these three levels (U, upper; M, middle; L, low) in the canopy an average branch leaf area was computed, a total plant leaf area was estimated and leaf-to-fruit ratios were derived for each tree after counting the number of coffee berries. To estimate light availability and temperature within the canopy, instantaneous measurements of PPFD and temperature were taken at leaf level in September 2000 under clear sky conditions (ca 2000 µmol m 2 s 1 )using the quantum and temperature sensors of a DeltaT AP4 porometer (DeltaT, Burwell, UK). Measurements were performed on the right leaf of leaf pairs 3, 6, 12, 15, 18, 21, 24 and 27 on the three selected branches, with leaf pair 3 being the closest to the tree trunk inside the canopy and leaf pairs 18 27 the outermost ones. Values presented for each leaf position within a branch are means of the 16 spot readings, all taken between 10:00 and 12:00 (i.e. solar noon ±1h). Fruit parameters Only fully ripe coffee berries, as determined by the bright red colour of their skin, were harvested from individual trees. Five to six harvests were necessary to collect all the berries over a period of 3 months. The fresh weight of berries was recorded at each harvest. Coffee samples were prepared by the wet processing method (wet de-pulping, anaerobic fermentation for 24 h, sun drying, de-husking) to obtain ready-to-beroasted coffee beans (commonly named green coffee or green beans). For the production cycles of 1999 and 2000, bean size was assessed for each tree and at each harvest with a series of sieves after sun drying beans to a water content of 120 g kg 1. The % of green beans with larger sizes (bean diameter >6.75 mm) was calculated. At each harvest, green beans of four trees per treatment were combined and a 50 g sample was analysed for caffeine, trigonelline, chlorogenic acid, fat and sucrose contents by near-infrared reflectance spectrometry (NIRS) based on calibration curves established for each compound. 13 These compounds are considered important precursors for coffee aroma 198 J Sci Food Agric 86:197 204 (2006)

Effects of productivity and shade on coffee quality and organoleptic properties of coffee beverage upon degradation during roasting through the Maillard reaction. 14 These analyses were performed on an NIRS model 6500 spectrometer (NIRS System Inc., Silver Spring, MD, USA) based on the reflectance of ground (<0.5mm) greencoffee. TheNIRS systemwas driven by NIRS2 (4.0) software (Intrasoft Int., Port Matilda, PA, USA). Data of harvests were pooled per treatment for the purposes of this study. Beverage quality assessment After eliminating most defective beans, 150 g samples of green coffee were roasted for 7 8 min at 220 Cin a Probat BRZ2 laboratory roaster (Probat, Emmerich, Germany). Cup quality tests were performed on an infusion prepared with 12 g of roasted and ground coffee. A panel of ten judges tasted three cups of 120 ml of infusion for each sample. The four main beverage attributes (acidity, bitterness, astringency and body) were estimated using scales ranging from 0 to 5, where 0 = null, 1 = very light, 2 = light, 3 = regular, 4 = strong and 5 = very strong. An additional preference score was also used: 0 = not good for drinking, 1 = bad, 2 = regular, 3 = good and 4 = very good. The tests were repeated three times and values are means of three sessions. Data analysis Statistica software (version 5, Statsoft Inc., Tulsa, OK, USA) was used to perform all statistical analyses. Data were analysed per year (only production for 2001) according to a factorial design with PPFD level (shade and sun) and fruit load (F, 1/2 and 1/4) as the two factors. RESULTS Effects of shade and fruit load on coffee productivity and alternate bearing Shade significantly decreased coffee production under the optimal coffee-growing conditions of the site (Table 1). With fruit loads combined, coffee production was 14% lower in shade than in full sun for the 1999 production, 6% for the second production and 29% for the third production, resulting in a reduction of 18% for the cumulative production over the three years. Fruit load also significantly affected coffee production. In 1999, fruit thinning did not result in a proportional decrease in production (Table 1). Coffee production of trees with 1/4 load was only reduced by 50% in comparison with that of trees with full load. In the same manner, coffee production of trees with 1/2 load was only 25% lower than that of trees with full load. In 2000, the trend was completely reversed: trees with full load produced about 33% less coffee than trees with 1/2 load and about 50% less than trees with 1/4 load (Table 1). Cumulative production of years 1999 and 2000, with imposed fruit thinning, was not different for full and 1/2 loads. However, trees with 1/4 load produced 25% less coffee. Without any fruit thinning in 2001, coffee production was not significantly different between trees that had had different fruit loads during the previous two production cycles (Table 1). Changes in productivity of trees with full load between the two production cycles (1999 and 2000) illustrated perfectly the alternate bearing pattern of coffee, as they produced the most during the first year and the least during the second year. Interestingly, fruit thinning to a quarter of the production potential resulted in the most balanced production pattern, but led to the highest productivity without fruit thinning during the third year. Shade also reduced alternate bearing (Table 1). The 2000 production was equivalent to 75% of the previous production (1999) and 80% of the following one (2001) in shade conditions. In sun conditions, variation in productivity from one year to the next was much higher, as the 2000 production was equivalent to 69% of the previous production (1999) and only 60% of the 2001 production. Effects of fruit load and shade on light availability, leaf temperature and leaf-to-fruit ratio Shade considerably affected the microenvironment in which fruits were developing. Shade had a very strong effect on light availability within the coffee Table 1. Effects of PPFD regime (45% shade or full sun) and fruit load (full (F), half (1/2) or quarter (1/4) of initial fruit load) on coffee production, branch leaf area, individual leaf area, leaf-to-fruit ratio and % of beans with larger sizes Yield (g berries per plant) Branch leaf area (cm 2 ) Average leaf area (cm 2 ) Leaf-to-fruit ratio (cm 2 per fruit) % of beans with larger sizes 1999 2000 2001 a 1999 2000 1999 2000 1999 2000 1999 2000 Shade 2340 1770 2220 730 590 36 42 18 18 63.5 72.1 Sun 2700 1880 3130 650 380 30 28 14 12 62.5 65.6 P 0.03 NS 0.001 0.001 0.001 0.001 0.001 0.05 0.001 NS 0.001 F 3463 928 2917 620 397 33 31 8 19 56.2 67.1 1/2 3210 1457 3074 730 549 34 38 11 17 61.7 70.7 1/4 1698 1889 3453 730 503 33 38 20 12 67.0 71.8 P 0.001 0.001 NS 0.05 0.05 NS NS 0.05 0.05 0.02 NS P interact NS NS NS 0.05 0.05 NS NS 0.05 0.05 NS NS a Without fruit thinning. NS, non-significant (P > 0.05). J Sci Food Agric 86:197 204 (2006) 199

P Vaast et al. PPFD (µmol m -2 s -1 ) PPFD (µmol m -2 s -1 ) 1600 1400 1200 1000 800 600 400 200 0 1600 1400 1200 1000 800 600 400 200 0 3 6 9 12 15 18 21 3 6 9 12 15 18 21 24 3 6 9 12 15 18 21 24 27 U M L Leaf position within branch 3 6 9 12 15 18 21 3 6 9 1215182124 3 6 9 12 15 18 21 24 27 U M L Leaf position within branch Figure 1. Measurements of PPFD levels at different leaf positions (from inner leaf position 3 close to tree trunk to outermost positions) on branches in upper (U), middle (M) and lower (L) parts of coffee canopy of (A) sun-grown and (B) shade-grown trees at bean-filling stage during 2000 production cycle. canopy (Fig. 1). With the exception of the outermost leaves, PPFD levels were below 200 µmol m 2 s 1 for all leaves at all branch levels within the canopy of shade trees. In contrast, almost all leaves of sun trees received PPFD levels well above 200 µmol m 2 s 1. Furthermore, leaves on the outer parts (leaf positions 12 18) of the upper and middle branches, where most of the coffee berries were produced, received PPFD levels above 500 µmol m 2 s 1. Leaf temperature differences were also noticeable between shade and sun trees. Differences of 4 C for inner leaves (up to leaf 6) and 2 C for outer leaves were observed between shade and sun trees. Shade and fruit load considerably altered leaf-tofruit ratios. Although coffee branch development is continuous and therefore leaf-to-fruit ratio evolves along the production cycle, the ratios presented here (Table 1) are the ones registered at the peak of bean filling, the most critical time in terms of carbohydrate competition between developing vegetative parts and coffee berries. In 1999, shade increased the average area of individual leaves, as values for shade and sun leaves were 36 and 30 cm 2 respectively. This trend was even more pronounced in 2000, with values of 42 and 28 cm 2 for shade and sun leaves respectively. Furthermore, shade increased leaf lifespan by up to 2 months (data not presented). This resulted in a higher average leaf area per branch, with a value of 650 cm 2 for sun versus 730 cm 2 for shade in 1999 and a value of 380 cm 2 for sun versus 590 cm 2 for shade in 2000. Consequently, leaf-to-fruit ratios were also strongly affected by shade, with a value of 14 cm 2 per A B fruit for sun versus 18 cm 2 per fruit for shade in 1999 and a value of 12 cm 2 perfruitforsunversus 18 cm 2 per fruit for shade in 2000. As expected, fruit thinning also resulted in strong alterations of the leaf-to-fruit ratio. At the bean-filling stage in 1999, this ratio was 8cm 2 per fruit for full load versus 11 and 20 cm 2 for 1/2 and 1/4 loads respectively. In 2000, owing to lower production caused by alternate bearing, the trend was reversed and the leaf-to-fruit ratios were almost twice as large for full and 1/2 loads compared with 1/4 load, with values of 19, 17 and 12 cm 2 per fruit respectively. Incidentally, these higher leaf areas of shade trees also explained the low light availability of shade trees in comparison with sun trees as previously presented in Fig. 1. Effects of fruit load and shade on coffee berry ripening Shade and fruit load significantly affected the berry-ripening process (Fig. 2). In a warmer microenvironment with high irradiance, coffee berries ripened faster in full sun than in shade. Therefore the harvest peak was delayed by about 1 month owing to shade. During the first production cycle (1999) and by the fourth harvest (19/01/2000), more than 85% of the coffee berries were already harvested in full sun compared with 65% in shade (Fig. 2). A clear trend could also be observed with fruit load: the lower the fruit load, the faster was the ripening of the coffee berries, irrespective of whether the coffee plants were managed under sun or shade. By the third harvest (29/12/1999), 65% of the berries were ripened at 1/4 load in full sun, whereas only 42% were ready for harvest at full load. During the second production cycle (2000) the observations confirmed the delaying effect of shade on coffee berry ripening, as 82% of berries were already harvested at the third harvest (07/01/2001) in full sun and only 60% in shade (Fig. 2). The hastening effect of fruit load on berry ripening was not observed in 2000. This is due to the fact that the overall productivity of trees was lower during this second cycle and therefore fruit thinning did not result in fruit load differences large enough to significantly affect coffee ripening as observed during the previous production cycle. Effects of fruit load and shade on coffee bean size Fruit load and shade had significant effects on coffee bean size (Table 1). In 1999, it could clearly be observed that decreasing fruit load resulted in an increasing % of beans with larger sizes (bean diameter >6.75 mm). Owing to an overall lower tree productivity caused by the alternate bearing pattern during the second year (2000), fruit thinning did not result in fruit load differences large enough to significantly affect bean size. On the other hand, shade enhanced bean size during this 2000 cycle. 200 J Sci Food Agric 86:197 204 (2006)

Effects of productivity and shade on coffee quality Harvests 1999 in Shade Harvests 1999 in Sun 24/02/00 10/02/00 19/01/00 24/02/00 10/02/00 19/01/00 100% 29/12/99 05/12/99 22/11/99 100% 29/12/99 05/12/99 22/11/99 80% 80% 60% 60% 40% 40% 20% 20% 0% F 1/2 1/4 0% F 1/2 1/4 Harvests 2000 in Shade Harvests 2000 in Sun 100% 22/02/01 07/02/01 07/01/01 07/12/00 23/11/00 100% 22/02/01 07/02/01 07/01/01 07/12/00 23/11/00 80% 80% 60% 60% 40% 40% 20% 20% 0% F 1/2 1/4 0% F 1/2 1/4 Figure 2. Effects of PPFD regime (45% shade or full sun) and fruit load (full (F), half (1/2) or quarter (1/4) of initial fruit load) on distribution of harvests during 1999 and 2000 production cycles. Effects of shade, fruit load and year of production on coffee bean composition In 1999, shade had significant effects on the biochemical composition of coffee beans (Table 2). Caffeine and fat contents were higher in beans of shade-grown plants, whereas sucrose, chlorogenic acid and trigonelline contents were higher in beans of sungrown plants. In 2000, the same trends could be observed, but only significantly for caffeine, fat and trigonelline. Fruit load had no significant effect on coffee bean composition (Table 2). Effects of shade, fruit load and year of production on beverage quality Shade significantly affected beverage quality (Table 3). Negative attributes such as bitterness and astringency were higher for coffee beverage prepared from sungrown beans than for that prepared from shade-grown beans during the two consecutive production cycles. Furthermore, positive attributes such as beverage acidity and preference were significantly higher for shade-grown beans. Fruit load had a significant effect on beverage quality, with a trend indicating higher Table 2. Effects of PPFD regime (45% shade or full sun), fruit load (full (F), half (1/2) or quarter (1/4) of initial fruit load) and year of production on coffee bean biochemical composition (g kg 1 bean dry weight) Caffeine Fat Sucrose Chlorogenic acids Trigonelline 1999 2000 1999 2000 1999 2000 1999 2000 1999 2000 Shade 14.8 14.1 131 117 82 77.3 76.2 82.1 9.9 9.7 Sun 14.2 13.6 122 115 84 78.4 77.1 82.6 10.7 10.1 P 0.001 0.001 0.001 0.05 0.001 NS 0.001 NS 0.001 0.001 F 14.3 14.1 127 119 84 76.7 76.6 82.3 10.0 9.9 1/2 14.5 14.2 127 119 83 76.6 76.7 82.2 10.3 9.9 1/4 14.6 14.1 125 116 83 77.8 76.6 82.6 10.4 10.0 P 0.001 NS NS NS NS NS NS NS 0.01 NS P interact NS NS NS NS NS NS NS NS 0.007 NS NS, non-significant (P > 0.05). J Sci Food Agric 86:197 204 (2006) 201

P Vaast et al. Table 3. Effects of PPFD regime (45% shade or full sun), fruit load (full (F), half (1/2) or quarter (1/4) of initial fruit load) and year of production on beverage characteristics Acidity a Bitterness a Astringency a Body a Preference b 1999 2000 1999 2000 1999 2000 1999 2000 1999 2000 Shade 2.27 2.45 2.65 2.65 1.68 0.35 2.78 2.50 2.57 2.80 Sun 1.67 2.21 2.95 2.88 1.86 0.41 2.91 2.67 2.29 2.58 P 0.001 0.04 0.002 0.01 0.02 NS 0.05 0.05 0.01 0.02 F 1.91 2.47 2.86 2.83 1.82 0.46 2.92 2.66 2.42 2.76 1/2 2.02 2.41 2.75 2.73 1.80 0.36 2.89 2.53 2.64 2.70 1/4 2.13 2.27 2.75 2.75 1.79 0.34 2.72 2.66 2.73 2.74 P 0.03 0.05 NS NS NS NS 0.05 NS 0.001 NS P interact 0.09 NS NS NS NS NS NS NS 0.001 NS a The scores for acidity, bitterness, astringency and body were based on a scale of 0 5, where 0 = null, 1 = very light, 2 = light, 3 = regular, 4 = strong and 5 = very strong. b Overall preference was based on a scale of 0 4, where 0 = not good for drinking, 1 = bad, 2 = regular, 3 = good and 4 = very good. Each value is the average score of ten judges during three tasting sessions. NS, non-significant (P > 0.05). preference with decreasing fruit load, especially in 1999 (Table 3). Interestingly, the overall beverage quality (higher acidity, lower astringency and higher preference) was higher in 2000, when production was around 30% lower (Table 1), than in 1999 (Table 3). DISCUSSION The results of this study show the importance of light regime on coffee productivity and bearing pattern. In the optimal ecological conditions for coffee growing of the present study (the central valley of Costa Rica is one of the world s coffee-producing regions with the highest commercial productivity, with up to 3000 kg coffee beans ha 1 year 1 ), a rather dense shade level of 45% reduced by only 18% the productivity of trees over three consecutive production cycles. These results confirm previous studies showing that artificial shade 12,15,16 or shade trees 9 reduce coffee fruit load through their effects on coffee physiology, such as longer internodes, fewer fruiting nodes and lower flower induction. In sun conditions, biannual variation in productivity was much higher than under shade owing to the fact that sun-grown trees had higher fruit loads and that coffee trees prioritized the allocation of carbohydrates to berries at the detriment of young vegetative branch parts, thus conditioning the production level of the following year. 4,5 Therefore the present results support the idea that shade decreases the magnitude of the alternate bearing of coffee trees as often hypothesised in the literature. 5,9,12 This is particularly important for individual coffee growers or cooperatives, as it limits the need to invest in processing facilities that are oversized every other year and allows them to market a more constant volume and quality of coffee to buyers. The present results confirmed the importance of light regime regarding bean biochemical composition and quality of the coffee beverage. 11,12,17 For the first time they documented the large differences in light exposure and air temperature that are experienced by coffee berries under shade and sun conditions. Many authors 18 24 have highlighted the strong effect of light exposure of the fruit on its maturation, particularly its skin colour and flesh ripening. Although Montavon et al. 25 have recently emphasised the fact that berry maturation clearly favours the development of highquality flavour in the coffee brew, bean maturation is certainly more critical than that of its surrounding flesh for coffee as for other nut trees. 26 A delay in ripening between the berry pulp and the bean has already been documented 27 and proposed as one of the reasons explaining observed differences in beverage quality between shade-grown and sun-grown coffee. 11 In the present study, caffeine and fat contents were highest in beans of shade-grown plants, whereas sucrose, chlorogenic acid and trigonelline contents were highest in beans of sun-grown plants. A similar negative relationship between fat and sucrose contents has been reported for C. canephora 28 and C. arabica 29 coffee beans. Sucrose is a precursor of polysaccharides and fat compounds in coffee beans. 30 Therefore a high sucrose content coupled with a low fat content in beans of sun-grown coffee could very well indicate that bean filling and fat synthesis were not fully achieved. Higher chlorogenic acid and trigonelline contents in beans produced by sun-grown plants could also point towards incomplete bean maturation and explain the higher bitterness and astringency of the coffee beverage. To decrease the number of hand pickings or to allow mechanical harvesting, experimental spraying with ethylene-based products to group maturation of coffee berries has resulted in lower beverage quality. 31,32 This was attributed to the fact that coffee berry flesh maturation was artificially hastened, whereas that of the bean was not. The present results demonstrate the beneficial synchronising effect of shade through a decrease by half in the light intensity and by several degrees (up to 4 C) in the temperature around the berries, which slows down the ripening process of coffee berry flesh and allows more time for complete bean filling. This indicates that this 202 J Sci Food Agric 86:197 204 (2006)

Effects of productivity and shade on coffee quality buffering effect of shade mimics that of increasing altitude, as it is generally accepted in the tropics that each increase of 100 m in altitude decreases the average daily temperature by 1 C. Therefore shade management can partially compensate the altitude deficit of many low-altitude coffee-producing zones of Central America and should permit coffee production of good quality. Nowadays, time of harvest is guesstimated on the basis of the number of weeks after flowering and accumulated experience of local conditions by coffee growers. It would be particularly worth undertaking an investigation on the sum of degree days as an indicator of coffee bean maturation, as commonly used for many other fruit trees. The present results illustrate the antagonistic relationships between coffee tree productivity and fruit size and quality as already documented for many other fruit tree species. 18,33 35 Cannell 5 has already pointed out the competition for carbohydrates among coffee berries during heavy production cycles and its negative impact on bean size and beverage quality. The present results confirm this competitive effect among coffee berries regarding bean size and beverage quality in view of the effects of fruit thinning, shade and year of production. In 1999, the year of the highest overall tree productivity, decreasing fruit load resulted in a larger bean size and higher coffee quality. Irrespective of the light regime, trees produced coffee of higher quality in 2000 when the overall tree productivity was 30% lower than during the previous year. Through its effects on limiting flowering intensity and hence tree productivity, shade management consistently resulted in higher beverage quality over the two consecutive production cycles. These results demonstrate that fruit thinning enhanced bean size and hastened the maturation process. Palmer et al. 36 have also shown for apple trees that lighter crop loads resulted in earlier fruit maturity and enhanced quality compared with heavy crop loads. In the present experiment the shorter duration of the bean-filling period was counterbalanced by an ample carbohydrate supply to growing fruits due to high leaf-to-fruit ratios in their vicinity. As leaf-tofruit ratios varied greatly within the coffee canopy, more investigation is under way to test the effect of coffee berry position and yield distribution on bean characteristics and beverage quality. Fruit thinning is a common practice to improve fruit growth and quality for many fruit trees such as peach, kiwi and apple. 18,35 37 However, this management practice (via chemical spraying or manual thinning) is rarely implemented by coffee growers worldwide despite the fact that it has been long documented that coffee trees experience a biannual bearing pattern and overbearing branch die-back due to competition for carbohydrates between coffee berries and developing young branch parts, thus conditioning the production level of the following year. 5 This is due to the fact that neither chemical spraying nor manual thinning has been intensively studied in the case of coffee, because past research was mostly focused on enhancing productivity rather than coffee quality, tree longevity and plantation sustainability. Nonetheless, alternative techniques do exist. Through its effect on decreasing flower intensity, shade management is certainly a management practice worth promoting to coffee farmers in order to overcome the inability of dwarf coffee cultivars to regulate their productivity. Selective pruning of heavybearing branches is also a management technique that deserves more attention. Recent works on the breeding of dwarf cultivars with traditional ones also appear to be promising, as hybrids are more vigorous and exhibit lower biannual bearing patterns. 29 In any case, these results highlight the need for further investigation on carbohydrate production and allocation between vegetative and reproductive parts in different regions of the coffee canopy to refine the existing model predicting fruit growth at the coffee branch level, 4 extend its applicability to the whole plant level and explore new management practices. CONCLUSION It appears clearly that in full sun the dwarf coffee cultivar Costa Rica 95 poorly self-regulates its productivity, produces beans of lower quality and experiences stronger alternate bearing than in shade. As dwarf cultivars (caturra, catuai and catimors) are now predominant in coffee plantations of Central America, this greatly advocates for the promotion of agricultural practices that improve plantation sustainability and coffee quality to secure higher revenues to farmers in the long term. Indeed, the present differences due to shade and fruit load are important enough to have economic significance for coffee farmers, since the premium paid for high coffee quality can amount to more than 100% of the market price of standard coffee quality. Although the present observations need to be confirmed under different producing conditions, they should result in recommendations in terms of coffee management. Certainly, agricultural practices (particularly shade management and pruning) limiting fruit load, lowering tree stress, better balancing leaf-to-fruit ratios and favouring slow ripening of coffee berry pulp and adequate bean filling should help produce coffee of higher quality and larger bean size. Bean size is particularly important, as it is often the main criterion along with bean colour and % of physical defects on which the exportability of coffee in producing countries is based. Since coffee unfit for export and consumed locally is often sold at less than half the international market price in Central America, there is considerable economic interest in producing the highest proportion possible of exportable beans of large size. ACKNOWLEDGEMENTS The authors thank ICAFE (Coffee Institute of Costa Rica) for maintenance of the experimental J Sci Food Agric 86:197 204 (2006) 203

P Vaast et al. plot and usage of their processing facilities; Fabrice Davrieux, a researcher at the CIRAD laboratory of coffee technology, for the NIRS analyses; and the European Commission (INCO-DEV project CASCA ICA4-CT-2001-10071) and the Science and Cultural Cooperation Centre, French Embassy, Costa Rica for supporting part of the operational costs of this research. REFERENCES 1 Clifford NM and Wilson KC (eds), Coffee: Botany, Biochemistry and Production of Beans and Beverage. Croom Helm, London (1985). 2 Carr MKV, Review paper: the water relations and irrigation requirements of coffee. Exp Agric 37:1 36 (2001). 3 Maestri M and Barros RS, Coffee, in Ecophysiology of Tropical Crops, ed. by Alvim PT and Kozlowski TT. Academic Press, New York, pp. 249 278 (1977). 4 Vaast P, Génard M and Dauzat J, Modeling the effects of fruit load, shade and plant water status on coffee berry growth and carbon partitioning at the branch level. Acta Hort 584:57 62 (2002). 5 Cannell MGR, Physiology of the coffee crop, in Coffee: Botany, Biochemistry and Production of Beans and Beverage, ed. by Clifford NM and Wilson KC. Croom Helm, London, pp. 108 134 (1985). 6 Bertrand B, Guyot B, Anthony P and Lashermes P, Impact of the Coffea canephora gene introgression on beverage quality of C. arabica. Theor Appl Genet 107:387 395 (2003). 7 Soto-Pinto L, Perfecto Y, Castilio-Hernandez J and Caballero- Nieto J, Shade effect on coffee production at the northern Tzeltal zone of the State of Chiapas, Mexico. Agric Ecosyst Environ 80:61 69 (2000). 8 Bertrand B, Aguilar G, Santacrea R and Anzueto F, El mejoramiento genético en América Central, in Desafíos de la Caficultura en Centroamérica, ed. by Bertrand B and Rapidel B. Editorial Agroamérica, San José, pp. 407 453 (1999). 9 Beer J, Muschler R, Kass D and Somarriba E, Shade management in coffee and cacao plantations. Agroforest Syst 38:139 164 (1998). 10 Babbar LJ and Zak DR, Nitrogen loss from coffee agroecosystems in Costa Rica: leaching and denitrification in the presence and absence of shade trees. J Environ Qual 24:227 233 (1995). 11 Guyot B, Manez JC, Perriot JJ, Giron J and Villain J, Influence de l altitude et de l ombrage sur la qualité des cafés arabica. Plant Rech Dévelop 3:272 280 (1996). 12 Muschler R, Shade improves coffee quality in a sub-optimal coffee zone of Costa Rica. Agroforest Syst 51:131 139 (2001). 13 Guyot B, Davrieux F, Manez JC and Vincent JC, Détermination de la caféine et de la matière sèche par spectrométrie proche infrarouge. Applications aux cafés verts et aux cafés torréfiés. Café Cacao Thé 37:53 64 (1993). 14 Dart SK and Nursten HE, Volatile components, in Coffee, Vol. I, Chemistry, ed. by Clarke RJ and Macrae R. Elsevier Applied Science, London, pp. 223 265 (1985). 15 Alvim PT, Factors affecting flowering of coffee. JCoffeeRes 7:15 25 (1977). 16 Boyer J, Influence de l ombrage artificiel sur la croissance végétative, la floraison et la fructification des caféiers Robusta. Café Cacao Thé 12:302 319 (1968). 17 Decazy F, Avelino J, Guyot B, Perriot JJ, Pineda C and Cilas C, Quality of different Honduran coffees in relation to several environments. J Food Sci 68:2356 2361 (2003). 18 Corelli-Grappadelli L and Coston DC, Thinning pattern and light environment in peach tree canopies influence fruit quality. Hort Sci 26:1464 1466 (1991). 19 Barritt BH, Rom CR, Konishi BJ and Dilley MA, Light level influences spur quality and canopy development and light interception influences fruit production in apple. Hort Sci 26:993 999 (1991). 20 Dann IR and Jerie PH, Gradients in maturity and sugar levels of fruit within peach trees. JAmSocHortSci113:27 31 (1988). 21 Marini RP, Sowers DL and Marini MC, Peach fruit quality is affected by shade during final swell of fruit growth. JAmSoc Hort Sci 116:383 389 (1991). 22 Génard M and Bruchou C, Multivariate analysis of within-tree factors accounting for the variation of peach fruit quality. Sci Hort 52:37 51 (1992). 23 Robinson TL, Seeley EJ and Barritt BH, Effect of light environment and spur age on Delicious apple fruit size and quality. JAmSocHortSci108:855 861 (1983). 24 Warrington IJ, Stanley CJ, Tustin DS, Hirst PM and Cashmore WM, Light transmission, yield distribution, and fruit quality in six tree canopy forms of Granny Smith apple. J Tree Fruit Prod 1:27 54 (1996). 25 Montavon P, Duruz E, Rumo G and Pratz G, Evolution of green coffee protein profiles with maturation and relationship to coffee cup quality. J Agric Food Chem 51:2328 2334 (2003). 26 Nzima M, Martin GC and Nishijima C, Leaf development, dry matter accumulation, and distribution within branches of alternate-bearing Kerman pistachio trees. JAmSocHortSci 122:31 37 (1997). 27 Guyot B, Petnga E and Vincent JC, Analyse qualitative d un café Coffea canephora var. Robusta en fonction de la maturité. I. Evolution des caractéristiques physiques, chimiques et organoleptiques. Café Cacao Thé 32:127 140 (1988). 28 Montagnon C, Guyot B, Cilas C and Leroy T, Genetic parameters of several biochemical compounds from green coffee, Coffea canephora. Plant Breed 117:576 578 (1998). 29 Bertrand B, L amélioration génétique de Coffea arabica L. en Amérique Centrale par la voie hybride F1, Thèse de Doctorat, ENSAM, Montpellier (2002). 30 Bradbury AGW and Halliday DJ, Chemical structures of green coffee bean polysaccharides. J Agric Food Chem 38:389 392 (1990). 31 Oyebade T, Influence of pre-harvest sprays of ethrel on ripening and abscission of coffee berries. Turrialba 26:86 89 (1976). 32 Winston EC, Hoult M, Howitt CJ and Shepherd RK, Ethyleneinduced fruit ripening in arabica coffee (Coffea arabica L.). Aust J Exp Agric 32:401 408 (1992). 33 De Silva HH and Ball RD, Mixed model analysis of withinplant variation of fruit weight and implications for sampling kiwifruit vines. Ann Bot 79:411 418 (1997). 34 Palmer JW, Effects of varying crop load on photosynthesis, dry matter production and partitioning of Crispin/M.27 apple tress. Tree Physiol 11:19 33 (1992). 35 Smith GS, Gravett IM, Edwards CM, Curtis JP and Buwalda JG, Spatial analysis of the canopy of kiwifruit vines as it relates to the physical chemical and post harvest attributes of the fruit. Ann Bot 73:99 111 (1992). 36 Palmer JW, Guiliani R and Adams HM, Effect of crop load on fruiting and leaf photosynthesis of Braeburn /M.26 apple trees. Tree Physiol 17:741 746 (1997). 37 Souty M, Génard M, Reich M and Albagnac G, Influence de la fourniture en assimilats sur la maturation et la qualité de la pêche (Prunus persica L. Suncrest ). Can J Plant Sci 79:259 268 (1999). 204 J Sci Food Agric 86:197 204 (2006)