Draft Copy

FACTORS INFLUENCING CUP QUALITY IN COFFEE

Photo Courtesy of SPREAD, Rwanda

Prepared for the Global Coffee Quality Research Initiative Brian Howard

Global Coffee Quality Research Initiative Review 1 Draft Copy

quality on a number of levels. At the ex‐ FACTORS IN- porter or importer level coffee quality is linked to bean size, number of defects, FLUENCING CUP regularity of provisioning, tonnage avail‐ able, and physical characteristics. At the QUALITY IN roaster level coffee quality depends on moisture content, characteristic stability, COFFEE origin, organoleptic (taste and smell) quali‐ ties and biochemical compounds. At the consumer level coffee quality is about taste and lavor, effects on health and alertness, Introduction geographical origin, and environmental and sociological considerations. At every link in the supply chain there is the consid‐ eration of price. In 2004 the International Organization for Standardization (IOS) de‐ ined a standard for green coffee quality which entails defects, moisture content, size, and some chemical compounds of beans as well as standardization of prepa‐ ration of a sample from which to perform cup tasting. According to Bertrand, Arabica coffee production makes up 70% of the world total. Consequently this review will deal mainly with C. Arabica with a few no‐ Photo Courtesy SPREAD Rwanda table exceptions in the sections on genetics Cup quality in coffee is affected by a great and crossbreeding for disease resistance. number of factors; agronomic, genetic and production related. In this review the author seeks to summarize the major ind‐ Agronomy: ings of the research that has been con‐ ducted that is speciically related to cup quality and how it is affected by the envi‐ Soil Nutrition ronment in which the coffee tree is grown, the genetic makeup of the coffee plant itself Coffee can be cultivated on a wide variety and the manner in which coffee is prepared of soil types, provided these are at least 2 for consumption. More than 800 aromatic meters deep, free‐draining loams with a compounds combine to give acidity, body good water retention capacity and a pH of and aroma to a cup of coffee. These three 5‐6, fertile and contain no less than 2% or‐ descriptors will serve as the parameters ganic matter. High quality, acidic Arabica around which cup quality is described in coffees tend to be produced on soils of vol‐ this document. canic origin.

The quality of coffee is extensive in its dei‐ Van Der Vossen, 2005 expresses concern nition. Leroy et al, 2006 deines coffee that, “to sustain economically viable yield levels, 1 ton green coffee per hectare (4.5

Global Coffee Quality Research Initiative Review 2 Draft Copy acres) per year, large additional amounts of composted organic matter will have to come from external sources to meet nutri‐ ent requirements, especially nitrogen & po‐ tassium. The majority of small land hold‐ Fertilizer ers will not be able to acquire the neces‐ sary quantities and will be confronted with Organic vs Inorganic Fertilizer declining yields. Organic farming does not necessarily prevent disease or pests below Organic production of coffee is often economically harmful thresholds and the thought preferable due to the strong poten‐ humid conditions of heavily shaded coffee tial of negative environmental impacts may actually stimulate the outbreak of oth‐ from fertilizer leaching into surface waters ers. and groundwater. However, any produc‐ tion crop signiicantly depletes its soils Vaast et al, 1998 found that total uptake of ability to replenish key nutrients and hu‐ nitrate (N03) and ammonium (NH4), key mic matter taken from it in the form of nutrients for plant growth and develop‐ produce. Inorganic fertilizer is often ap‐ ment and the limiting nutrient in Arabica plied at rates approaching 100 to 300 kilo‐ Coffee, at any ratio was higher than that of grams per hectare at signiicant expense to plants fed solely with nitrate or ammonium producers. (Carvajal, 1959) Because of the alone. Anaerobic, lack of oxygen, soil con‐ preference for organically produced pro‐ ditions reduced nitrate and ammonium up‐ duce, especially in the specialty coffee take by 50% and 30% respectively and the market, solutions for such a deicit must be presence of dinitrophenol almost com‐ found and implemented in regionally ap‐ pletely inhibited N uptake in any form. propriate ways. Vaast suggests that these results indicate that Arabica coffee is well adapted to acidic In the shaded Indian coffee terrior of Kar‐ soil conditions and can effectively utilize nataka, India Nagaraj et al., 2006 found the seasonally available forms of inorganic that the addition of inorganic potassium in nitrogen (N). These observations can help the form of muriat of potash and sulphate to optimize coffee nitrogen nutrition by of potash had the effect of increased coffee suggesting agricultural practices that main‐ yields over the period of four years at a tain root systems in the temperature range rate of approximately 15%. The difference that is optimum for both ammonium and between the two treatment methods out‐ nitrate uptake. Vaast found that both ni‐ lined by Nagaraj not being statistically sig‐ trate and ammonium uptake peaked when niicant. It should be noted that the soils in root systems were maintained at 34 de‐ which this coffee were planted were receiv‐ grees Celsius. Below this temperature ing approximately 40 to 60 kg of potassium plant color indicated a loss of vigor. There‐ per hectare per year in leaf fall. The study fore both nitrate and nitrite availability in indicated that no consistent trend could be soi, as well as the coffee trees capacity for observed in the cup evaluation report for uptake through ideal temperature regimes, three years. Cup quality of Arabica coffee can be maximized. Van Der Vossen, 2009 was found to be similar in both MOP and notes that excessive calcium and potassium SOP treated plots but that there was a in soils produce a hard and bitter tasting modest improvement in the cup quality of liquor. robusta coffee in the sulphate potash

Global Coffee Quality Research Initiative Review 3 Draft Copy treated plots compared to muriate of pot‐ in 1 to 3 year old plots of C. Arabica and I. ash applications in the second and third oerstediana. In these young plots it was years. It should be noted that there has not found that coffee trees were deriving ap‐ been any evidence of changes occurring in proximately 20% of their nitrogen from the the lavour compounds due to agronomic biological nitrogen ixation occurring via use of sulphur or otherwise. (Krishnamur‐ symbiosis with I. oerstediana. No estimate thy Rao, 1989) Studies conducted in Kenya could be derived for plots between 5 and 7 by (Njoroge and Mwakha, 1985) did not years. It is a reasonable assumption to note any difference in liquor quality of cof‐ make that greater availability and uptake of fee between NPK fertilized plots and con‐ soil nitrogen has a strong positive correla‐ trol treatment over eleven years of re‐ tion to cup quality via plant health and search. bean size.

Cup quality differences have been found in studies contrasting organic and inorganic fertilization. In a 2008 study undertaken by Malta, et al. no signiicant differences Environmental Factors were observed on the cup quality among beans from conventional and organic Shade vs Sun plants in the irst year. However in the sec‐ ond year, cup quality of some organic It has been shown that on the most appro‐ treated plants was superior when com‐ priate sites, with intensive management, pared to conventionally treated plants. A self‐shading coffee monocultures can give 2 positive effect on sensorial attributes was and 3 fold increases over more traditional observed using cattle manure, either alone shaded systems. (Beer, 1987) Shade tends or associated with coffee straw and green to reduce photosynthesis, rates of transpi‐ manure. ration, plant metabolism and thus demand on soil nutrients. Due to lower nutrient In Hawaii, Youkhana & Idol, 2009 found needs a crop could potentially be obtained that the addition of mulch from shade tree on more marginal soils with lower fertility. pruning signiicantly offset net nitrogen In areas where regular fertilization cannot and carbon losses from coffee cultivation. be guaranteed it is recommended that Improved carbon and nitrogen sequestra‐ some shading trees be retained as a hedge tion in soil was measured over two years against uncertain future soil inputs. Dr. and it was found that soil bulk density did John Beer of CATIE stresses, “The funda‐ not decline in mulched plots as opposed to mental question, when planning the reno‐ signiicant changes in bulk density for un‐ vation or establishment of a coffee and ca‐ mulched plots. cao plantation is whether the owner has the site, education and resources to main‐ Grossman et al., 2006 found that organic tain the crops without shade. Coffee under production standards are being met while shade will survive setbacks far better than available Nitrogen in soil is supplemented monocultures of the crop.” (Beer, 1987) As by nitrogen ixing shade trees. Biological important as the question “could im‐ Nitrogen Fixation is facilitated through the provements in coffee quality through shade use of leguminous shade trees in the genus be made” in a given location, the question Inga with the most signiicant results found of whether a farmer “should” seek to use

Global Coffee Quality Research Initiative Review 4 Draft Copy shade in his plantation becomes equally (6) New crop varieties are often bred for important. Dr. Beer suggests a number of monocultural conditions possible advantages and disadvantages of (7) Reduced air movement and humidity such a decision. may encourage fungal growth (8) Shade trees can be other potential Advantages of Shading Include: hosts for pests and diseases (1) The Suppression of Weed Growth (9) Reduction of photosynthetically avail‐ (2) Potential for Product Diversiication able sunlight such as fruits and timber alongside cof‐ (10) Shade tree root competition of nutri‐ fee ents (3) Greater control over crop phenology (11) Rainfall redistribution could cause such as fruit setting and maturation spot erosion or reduce overall soil (4) Potential improvement of crop quality moisture (5) Reduction of evapotranspiration rate of (12) Harvesting of wood or fruit from the shaded crop shade trees is an additional drain of soil (6) Removal of excess soil moisture by nutrients transpiration of a heavy shade tree cover Desirable Characteristics for Shade (7) Potential for increased moisture input Trees Include: through horizontal interception of mist or clouds (1) The ability to ix nitrogen from the at‐ (8) Extension of the productive life of the mosphere crop (2) Compatibility with the crop in terms of (9) Reduction of temperature extremes facilitating minimal competition for wa‐ (10) Reduction of damage caused by hail or ter, nutrients and growing space heavy rain and winds (3) Strong rooting systems as shade trees (11) Better soil drainage and aeriation are more exposed to adverse climatic (12) The provision of soil mulch from tree conditions throughfall (4) Ability to extract soil nutrients which (13) Reduction of erosion on slopes, ie soil are not trapped by the crop conservation (5) A light crown that provides regular (14) Recycling of nutrients not accessible mottled shade rather than uniform to the crop shadow and poor light quality (15) Nitrogen ixation by shade trees (6) Flexible branches and stems (7) High biomass productivity of recycled Disadvantages of Shading Include: material in soil (1) Throughfall damage to understory crop (8) Absence of major disease susceptibility (2) Sudden defoliation of shade trees could that could lead to rapid defoliation cause a shock to the crop (9) Small leaves to reduce rain drop coales‐ (3) Additional labor necessary to maintain cence the same acreage (10) Not an alternative host for insects or (4) Mechanization of the crop may be hin‐ pathogens which could endanger the dered crop (5) Erosion control structures like terrac‐ ing would be hampered once shade Bosselmann et al., 2009 in Huila, Colombia trees are planted found that sensory attributes were inlu‐

Global Coffee Quality Research Initiative Review 5 Draft Copy enced negatively by shade, and that physi‐ grown beans. A noteworthy aspect of this cal attributes were inluenced positively by study was that the overall beverage quality, altitude. In higher altitudes (approxi‐ higher acidity, lower astringency and mately 1700 meters above sea level) shade higher preference, was found to be higher had a negative effect on fragrance, acidity, in the year 2000 when production was body, sweetness and preference of the bev‐ around 30% lower than in 1999. erage, while no effect was found on the physical quality of the bean. At lower alti‐ Geromel et al, 2008 builds his study on the tudes, shade did not have a signiicant ef‐ effects of shade on the development and fect on sensorial attributes, but signii‐ sugar metabolism of Coffea Arabica L. on cantly reduced the number of small beans. the premise that coffee fruits grown in At high altitudes with low temperatures shade are characterized by larger bean size and no nutrient or water deicits, shade than those grown under full sun condi‐ trees may have a partly adverse effect on C. tions. Bean size, as noted, is a strong con‐ Arabica cv Caturra resulting in reduced tributor to cup quality. He found that sensory quality. The occurrence of berry shade led to a signiicant reduction in su‐ borer (Hypothenemus hampei) was lower crose content and to an increase in reduc‐ at high altitudes and higher under shade. ing sugars. In pericarp and perisperm tis‐ Bosselman goes on to suggest that future sues, higher activities of sucrose synthase studies on shade and coffee quality should and sucrose phosphate synthase were de‐ focus on the interaction between physical tected at maturation in the shade com‐ and chemical characteristics of beans. pared with full sun and that both enzymes also had higher peaks of activities in devel‐ A study was done in Costa Rica by Vaast et oping endosperm under shade than in full al., 2005 contrasting light regimes and op‐ sun. Geromel went on to suggest that timal coffee growing conditions on dwarf metabolic pathways for sucrose needed coffee, Coffea Arabica. Shade was found to further study for identiication. Van Der decrease coffee tree productivity by 18% Vossen concurs that shade has a positive but reduced alternate bearing. Shade also effect on coffee quality, particularly at me‐ positively affected bean size and composi‐ dium altitudes but also reduces yields. He tion as well as cup quality due to a delay in also found that at altitudes above 1800 me‐ berry lesh ripening by up to a month. ters shade did not improve cup quality. Higher levels of sucrose, chlorogenic acid and trigonelline in sun grown beans indi‐ Muschler, 2001 found across the board im‐ cated incomplete bean maturation and re‐ provement in organoliptic parameters as sulted in higher bitterness and astringency shading increased. A blind tasting experi‐ in cup quality. Higher fruit loads, which ment showed highly consistent shade in‐ can be mitigated through branch thinning, duced improvements in both green and reduced bean size owing to carbohydrate roasted coffee. Improvements were made competition among berries during bean in both the acidity and body of the bever‐ illing. Higher taste preferences were age for both Caturra and Catimor arabicas. linked to lower fruit load. Shade was also The one mildly negative effect of shade was found to mitigate negative attributes in cof‐ found to be on the aroma of the beverage fee quality like bitterness and astringency for Catimor. It is thought that in the sub‐ while positive attributes like acidity were optimal coffee zone that was studied, the found to be signiicantly higher in shade shade promoted slower and more balanced

Global Coffee Quality Research Initiative Review 6 Draft Copy

illing and more uniform ripening of ber‐ ditions after harvesting. On the other end ries that yielded a better cup quality than of the spectrum in drought years shoot the unshaded coffee plants. (Muschler, dieback and premature ripening of the ber‐ 2001) ries can result in light beans producing a liquor with immature and astringent notes.

Da Silva et al, 2005 investigated the inlu‐ ence of environmental conditions and irri‐ gation on the chemical composition of green coffee beans and the relationships of those parameters to the quality of the bev‐ erage according to both sensorial and elec‐ tronic analysis. He found that irrigation was not a major factor affecting chemical composition since there were few differ‐ ences in relation to non irrigated coffee plants. He found the production site tem‐ perature differentials to be the main inlu‐ encing factor on biochemical composition. The study was undertaken near Sao Paulo, Brazil and the major inding was that cup quality decreased as air temperature rose to about 3.5 degrees above the optimum limit for coffee cultivation at 18 to 21 de‐ Above: Quality attributes of green beans of two grees Celsius. Similar indings are reported varieties of Coffea arabica grown without shade by Decazi et al., 2003 in Argentina. and with dense shade (80% shade) of Erythrina poeppigiana at a low elevation. Beans were har- In a study done in Australia between 2000‐ vested at full maturity. (This figure courtesy of 2002 by David Peasley of Rural Industries CATIE) Research, irrigation was found to signii‐ cantly increase bean yield as well as pro‐ duce the following results under interna‐ Rainfall & Irrigation tionally recognized SCAA cupping forms. The non irrigated control crop scored a 69 Van Der Vossen states that rainfall re‐ with the description of “low acidity and quirements for Arabica coffee production mild smoky lavour and thin body.” The are at least 1200mm per year with a low water stressed irrigated treatment maximum of 2500mm. He contends that scored 73 with the description of “dull coffee plants grow and yield better if ex‐ bakey aroma, nice acidity, sour, green apple posed to alternate cycles of wet and dry lavor.” The medium water stressed irri‐ seasons and that a period of water deicit is gated treatment scored 75.5 with the de‐ helpful in synchronizing lower bud differ‐ scription of “faint but sweet aroma, juicy, entiation. Areas with precipitation in ex‐ citrus lavour, OK body.” Each of the three cess of 2500mm have the tendency to pro‐ areas received 1634mm of rainfall with the duce lower quality coffee due to irregular low water stressed irrigated area having cherry ripening and poor bean drying con‐ 2100mm of irrigation applied to it and the

Global Coffee Quality Research Initiative Review 7 Draft Copy highly water stressed irrigated area having tions at high elevations. The analysis con‐ 647mm of irrigation applied to it. irmed homeostasis of the hybrids for which bean biochemical composition was only slightly affected by elevation than that Temperature & Altitude of traditional varieties. The cupping per‐ formed on samples originating from high Decazy et al, 2003 found that Honduran elevations showed no signiicant differ‐ coffees of superior quality came from high ences between Arabica hybrids and tradi‐ altitudes, above 1000m, where rainfall re‐ tional lines. (Bertrand et al., 2006) mains relatively low, that is to say below 1500mm per year. It was found that a Slope strong inverse relation between rainfall and fat content exists and that this relation East facing slopes were found by Avelino et needs to be considered in relation to alti‐ al., 2005 to yield beverages with generally tude because in the sampled regions in superior attributes, probably because of Honduras, rainfall and altitude were found superior exposure to morning sunlight. to be inversely correlated. High altitude The beverages from east facing slopes were green coffee beans had a higher fat content mainly more acid, with a score of 2.73 out than lower altitude beans and gave a better of 5, in the higher quality terriors, as op‐ cup quality. Van Der Vassen stresses that posed to 2.36 out of 5 for other exposures. high altitudes are critical for the successful In addition a positive relation was found production of high quality Arabica coffees between altitude and taster preferences. in equatorial regions. Lower temperatures, and their longer daily amplitudes, tend to Laderach & Vaast et al. in their “Geographi‐ induce slower growth and more uniform cal Analyses to Explore Interactions be‐ ripening of the berries, and produce larger tween Inherent Coffee Quality and Produc‐ and denser beans. Bean size and density is tion Environment” state that increasing often correlated to aroma, lavor and supe‐ slope negatively inluenced the inal score rior beverage quality. Altitude also tends of cup quality from terriores on two test to have a positive effect on acidity while sites in Columbia and Nicaragua. reducing bitterness.

While for traditional cultivars elevation has tended to play a signiicant role in bean biochemistry and organoleptic qualities, Genetics: with chlorogenic acid and fat concentration increasing with increasing elevation, this is With total economic damage to coffee not necessarily the case with Arabica hy‐ crops mounting to an estimated US $2‐2.5 brid cultivars like the Caturra. For Arabica billion annually, coffee leaf rust and coffee hybrids little correlation is found in varia‐ berry disease affect a signiicant portion of tion in chorogenic acid concentration and the supply chain. In addition to increased none of the variation in fat can be corre‐ scarcity due to disease and fungus signii‐ lated to elevation.. However, Arabica hy‐ cant environmental hazards exist due to brids were found to have 10‐20% higher the copper based fungicides used to ight fat concentrations than traditional carieties Coffee Berry Disease and Coffee Leaf Rust at low elevations and similar fat concentra‐ chemically. (Van Der Vossen, 2009) Ac‐

Global Coffee Quality Research Initiative Review 8 Draft Copy cording to Walyaro, 1997 the aim of most In C. Arabica and C. Canephora bean devel‐ genetic improvement programs is disease opment is a long process that is character‐ resistance and quality. Determination of ized by fundamental shifts in the makeup berry and bean characteristics using plant of the beans’ tissue. For economically vi‐ breeding is relatively simple, aroma and able species ,C. Arabica & C. Canephora, 6 lavor attributes present signiicantly to 8 months and 9‐11 months are respec‐ greater dificulties due to their chemical tively required for beans to reach maturity. complexity and susceptibility to agronomic Fruit development does not occur at the variability. Walyaro goes on to postulate same time with varying proportions of dif‐ that the development of reliable lab proce‐ ferent fruit sizes occuring on the same dures which relate individual chemical plant. Despite variation in fruit growth compounds to cup quality could have im‐ timing and the differences in the length of portant bearing on genetic improvement of each species reproductive cycle, the pri‐ cup quality in coffee. It has been shown mary steps of seed development are that resistance to disease and nematodes thought to be identical between commer‐ can be increased through genetic exchange cial species. (De Castro et al., 2006) between C. Arabic and C. canephora. (Ber‐ trand et al., 2005, Dessalegn et al., 2008, After fertilization and up to mid develop‐ Leroy et al., 2006) ment the fruit is primarily made up of peri‐ carp and perisperm tissue. In the latter The Coffea genus encompasses about 100 phases of development the perisperm different species. Within those species, C. gradually disappears and is replaced by the Arabica and C. Canephora make up ap‐ endosperm, also known as the “true seed” proximatley 70 and 30% of the total coffee which was initially present in a liquid state. market. At a genetic level, Arabica is a tet‐ The endosperm hardens as it ripens during raploid (2n=4x=44 chromosomes), which the maturation phase due to accumulation have their origin in a natural cross between of storage proteins, sucrose and complex the species C. Canephora and C. Eugenoides polysaccharides which represent the seeds according to Lashermes et al., 1999, 2000. primary energy reserves. The inal phase Arabica is a self pollinating species which of bean maturation is the dehydration of partially explains its narrow genetic base the endosperm and the color change of the and the ensuing dificulty with incorporat‐ pericarp to a dark red color. Key changes ing new traits. Robusta, however, is diploid accompany the development of the coffee (2n=22 chromosomes) and is not self cherry which require both quantitave and compatible because of a gametophytic sys‐ qualitative description in order to fully de‐ tem of incompatibility controlled by a sin‐ scribe the growth and development of the gle gene with multiple S‐alleles which ex‐ bean in relation to it’s eventual cup quality. plains the higher genetic diversity within (De Castro et al., 2006) the species and ensuing ease of genetic im‐ provement by conventional breeding tech‐ niques. Castro et al., 2006)

Coffee Fruit Development

Global Coffee Quality Research Initiative Review 9 Draft Copy

C: 120‐150 days after lowering: The re‐ maining folded outer perisperm layer is enclosing the completely milky endosperm.

D: 230 – 240 days after lowering: A ma‐ Below: Phases of coffee fruit development (Taken from De Castro et al., 2006): ture cherry fruit has been formed showing two developed mature seeds enclosing one cotyledonary mature embryo inside of the solid endosperm. (Decastro, 2006)

Plant Resistance

According to De Nardi et al., Coffee plants ight off pathogenic infections through a number of strategies, including the strengthening of cell walls, activating or synthesizing antimicrobial compounds (an‐ Schematic representation of the tissues pre- tibiotics) and expressing defense‐ sent at the mature stage (220-250 days after associated proteins like pathogenesis re‐ fertilization). Photo Courtesy of Nestle Inc. lated (PR) proteins. Often, plants will acti‐ vate a hypersensitive response that is char‐ acterized by the rejection and death of in‐ fected cells as shown by Greenburg, 1997. A localized response can trigger longer standing systemic responses known as SAR ‐ systematic acquired resistance. This re‐ sponse prepares the plant for resistance against a large spectrum of pathogens. Plants express large arrays of resistance genes (R genes) that guard against patho‐ gens. These genes encode putative recep‐ tors that respond to the products of aviru‐ A: 0‐60 days after lowering: Ovary after lence genes expressed by pathogens during anthesis. The growing perispern tissue, the the infection of a plant. The recognition of integuments and the young embryo sac avirulence genes initiates a downstream that will further develop in the endosperm signaling that can render the attempted are visible. infection unsuccessful. Systematic ac‐ quired resistance (SAR) inducing chemi‐ B: 60 – 120 days after lowering: The peri‐ cals provide a unique opportunity to view carp and liquid endosperm tissue or “true induced resistance mechanisms in plants in seed” which grows by absorbing the inner the absence of a biological model system. perisperm tissue, become clearly visible in Speciically, 2 chemicals, 2,6‐ crossection. dichloroisonicotinic acid (INA) and benzo (1,2,3)thiadiazole‐7‐carbothionic acid s‐ methyl ester (BTH), have been demon‐

Global Coffee Quality Research Initiative Review 10 Draft Copy strated to be effective inducers of system‐ Coffee leaf rust (CLR) is caused by a atic acquired resistance in a variety of pathogen of the leaf called Hemileia vas‐ plants. These compounds work to protect tarix and is characterized by orange rust plants from pathogenic infections without postules on the under side of the leaf. This having direct microbial activity. Despite pathogen causes signiicant losses as a re‐ acting downstream of salicylic acid in SAR sult of loss of leaf area and the correspond‐ signaling, they activate the SAR signal ing loss of photosynthesis and leaf drop. transduction pathway through the same Coffee leaf rust has now spread through all signaling cascade thus imitating an infec‐ Arabica coffee producing countries in the tion and improving the plant’s natural de‐ world making it a signiicant issue for the fenses, making it more resistant to infec‐ coffee industry as a whole in terms of sup‐ tion in the future. De Nardi went on to ply susceptibility. conirm C arabica showed typical SAR re‐ sponses to BTH treatment. Both roots and Coffee berry disease (CBD) is caused by leaves responded with a shift in metabo‐ Colletotrichum kahawae and is a fungus lism from normal housekeeping activity that causes dark lesions on the green and defensive activity. Unexpectedly, the root ripening berries. CBD is unique in that reaction was found to be different than the crop losses due to the fungus can be severe. leaf reaction. The primary response in the According to the Coffee Research Founda‐ leaf was an increase in physical and chemi‐ tion’ s annual report for 1987‐1988 in Af‐ cal barriers, along with repression of the rica as a whole losses can range between genes involved in photosynthesis and nor‐ 30% and 50% during very high precipita‐ mal metabolic pathways. The enhance‐ tion years even with chemical treatment. ment of cell walls appeared to be the main response in root systems and the normal Fusarium (Giberella) stilboides (Fusar‐ anti‐microbial peptides, in the form of ium bark disease) is a pathogen that causes chitinases and peroxidases, were not ob‐ bark lesions which are a result of damage served. It is propounded by De Nardi that to the vascular system. Vascular wilt often this approach for mimicking a plant dis‐ results in the death of the entire tree. Fu‐ ease or a pest attack is an approach that sarium is reported to have almost killed the can be applied to the problem of producing entire coffee industry in Malawi in the late quality coffee without the use of microbi‐ 1970’s according to Siddiqi, 1980. cides, insecticides and fungicides while safeguarding crops from pests and dis‐ Root Knot Nematode (Meloidogyne ex­ eases. (Di Nardi et al., 2006) igua) has been shown to be a huge threat to Arabica growing areas in Latin America. The greater debate and perhaps of more Coffea Arabicas’ resistance to M. exigua is importance to the topic of cup quality is controlled by an single inherited major whether this genetic resistance will lower gene, Mex‐1. Alipzar et al., compared resis‐ overall cup quality by necessity and in the tant and susceptible pure line cultivars end decrease consumer experience. with clones of hybrid cultivars derived from crosses between resistant and sus‐ ceptible lines. The results show that re‐ Major Diseases: production of the nematode was signii‐ cantly higher on the hybrid cultivars than on resistant pure line cultivars, however

Global Coffee Quality Research Initiative Review 11 Draft Copy resistance was much lower than on a sus‐ characteristics such as sucrose content and ceptible pure line cultivar. Individual galls beverage acidity. on the root systems of hybrid cultivars were found to be much smaller than on Introgression processes that began in the susceptible cultivars. After a 4 year moni‐ 1980’s to combat the spread of coffee leaf toring period Alpizar et al. reports that M. rust and root nematodes was not restricted exigua populations in a test ield were mul‐ to resistance traits but also included genes tiplied by a factor of 14 on susceptible that coffee producers widely contend af‐ plants buy only by a factor of 1.9 on hybrid fected cup quality. Most coffee buyers cultivars. They come to the conclusion that claim that new introgressed varieties have Mex‐1 could have incomplete dominant ex‐ a poorer cup quality than the standard va‐ pression that allowed nematode penetra‐ rieties, this inding was conirmed by Ber‐ tion, but inhibited the durable reproduc‐ trand et al., 2003 in the same study in tion of the namatode. (Alpizar et al., 2006) which he showed that some introgressed lines were similar to the control and that Genetic modiications have in some cases cup quality was not necessarily related to affected cup quality adversely but in many the amount of introgressed genome of Ro‐ cases have not adversely affected cup qual‐ busta. Since the late 1990’s, CIRAD and ity and increased genetic resistance to cof‐ partners have worked to develop a genera‐ fee leaf rust and M. exigua by introducing tion of F1 hybrid varieties that are crosses genetic material into the Arabica plant F1 hybrids of C. Arabica with 7 to 22% ge‐ from C Canephora via the Timor hybrid. netic material from C. canephora have been (Leroy et al., 2006) There are regional shown by Bertrand et al., 2006 to produce variations in resistance to disease that can good cup quality under ideal conditions be exploited while maintaining or improv‐ and in about half of the tested strains. ing beverage quality as is noted by Anzueto et al., 2001. Anzueto remarks that Ethio‐ In order to avoid damage to roots from pian origins provide resistance to nema‐ nematodes, C. Arabica is commonly inter‐ todes and partial resistance to leaf rust and grafted onto C canephora. The perform‐ likely improve beverage quality. ance of such grafts was evaluated over 5 years in Costa Rica by Bertrand et al., 2001. In higher quality C. Arabica stocks the main The grafting did not have an appreciable goal seems to be in the area of improve‐ effect on caffeine, fat and sucrose contents. ment of resistances to pathogens and an However the C. Liberica rootstocks did sig‐ increasing of yield. LeRoy et al., contends niicantly reduce aroma and the size of the that introgression of alien genetic material, bean produced. These deiciencies were material from C. canephora, does not seem partially explained however by tissue in‐ to be linked directly to variation in cup compatibility at the graft level. quality. (Leroy et al., 2006) Bertrand et al., 2003 came to similar conclusions when he With approximately a 10% market share of stated that selection can avoid accompany‐ the total coffee consumed world wide, de‐ ing the introgression of resistance genes caffeinated coffee is being considered as a with a drop in beverage quality due to posi‐ genetic trait. Consideration of genetic di‐ tive results in approximately half of the tri‐ versity and the correlation of caffeine con‐ als that he carried out dealing with taste tent in relation to cup quality was looked at speciically by Dessalegn et al., 2008. Des‐

Global Coffee Quality Research Initiative Review 12 Draft Copy salegn found that Ethiopian genotypes of extremely reduced genetic diversity in low caffeine content typically showed a Coffea Arabica L. in comparison to C lower cup quality but that there were no‐ Canephora. Wild accessions found in table exceptions. Consequently he con‐ Ethiopia seem to constitute a key gene res‐ cludes that simultaneous selection for low ervoir for Arabica stocks. The use of Mo‐ caffeine content and good cup quality is lecular Assisted Selection in coffee breed‐ possible given that there are sources of de‐ ing promises to drastically increase the ef‐ sirable genes in terms of cup quality with iciency of breeding programs as a com‐ relatively low caffeine content that can be plete genetic linkage map of C. Canephora utilized for resistance breeding. was reported and additional linkage maps are under construction for C. arabica. Eco‐ Bertrand et al., 2006 in Central America nomically crucial genes related to the caf‐ found homeostasis, stable equilibrium, in feine biosynthetic pathway or genes encod‐ taste characteristics of Arabica hybrids for ing for seed storage proteins have been iso‐ which bean biochemical composition was lated. The high performance already less affected by elevation than that of the achieved in the in vitro propagation proc‐ traditional varieties. The organoleptic ess by somatic embryogenesis offers the evaluation of hybrids, which was per‐ possibility to mass propagate superior hy‐ formed on samples originating from high brids in different regions for both C arabica elevation, showed no signiicant differ‐ and C. Canephora. (Etienne et al., 2001) ences between Arabica hybrids and tradi‐ Initial production by somatic embryogene‐ tional cultivars. Bertrand goes on to note sis is now permitting preparation for direct that “new hybrid varieties with high bever‐ commercial application. Seed cryopreser‐ age quality and productivity potential vation enables a routine use for long‐term should act as a catalyst in increasing the conservation of coffee genetic resources. economic viability of coffee agroforestry systems being developed in Central Amer‐ Bertrand et al., in his study of C. Arabica ica. hybrid performance isolated three key quality indicators; yield, fertility and Van Der Vossen, 2009 points out that tradi‐ weight and sought to improve plant per‐ tional cultivars of Arabica coffee are sus‐ formance. This was done by breeding F1 ceptible to coffee leaf rust and coffee berry hybrids derived from South American and disease. CLR being of worldwide impor‐ wild Sudanese ‐ Ethiopian stock. The Afri‐ tance while CBD remains restricted to Af‐ can stock is thought to improve genetic di‐ rica. Van Der Vossen contends that there is versity. Sudan‐Ethiopian genetic material a mounting volume of scientiic evidence carries signiicant resistance to root nema‐ accumulated over many years showing todes and leaf rust and possibly even that, given optimum environmental factors, brings better beverage quality to the table. disease resistant cultivars can in fact pro‐ Bertrand’s main objective was to increase duce coffee of equal quality to those from bean quality, plant productivity and to in‐ the best traditional varieties. troduce more variation into a very narrow genetic base. His study found that hybrid Etienne et al., 2001 states that the molecu‐ populations yielded 22‐47% more than the lar phylogeny of Coffea species has been maternal lines but that hybrids showed established using DNA sequencing. Mo‐ signiicantly more sterility than parental lecular markers were found that reveal an control lines. Selection in the hybrid popu‐

Global Coffee Quality Research Initiative Review 13 Draft Copy lations using the aforementioned traits re‐ sulted in signiicant genetic gain for yield Quality Evaluation and dry weight of 100 beans with insignii‐ Methods cant gains in fertility. When selected on the basis of fertility alone, increase in yield and The successful integration of genetic traits, 100‐bean weight were not obtained within which add positive taste characteristics as the hybrid populations. By applying selec‐ well as contribute to robustness against tion on yield and 100‐bean weight, the hy‐ disease and pests, can yield signiicant re‐ brids produced 11‐47% higher yields than sults in terms of supply chain security and the best line with signiicantly higher or cup quality. DNA introgression of alien ge‐ identical 100‐bean weight and the same netic material is being carried out world fertility rates. wide with varying and highly disputed re‐ sults. Discussion of various quality evalua‐ Pautigny et al., 2010 propounds the idea tion methods to determine the extent of that comparative genomics provides the gene introgression success is a key compo‐ opportunity to leverage genetic informa‐ nent of genetic research. Coffee identiica‐ tion between species using comparative tion and classiication serves as a means to genetic mapping. He goes on to state that avoid coffee adulteration. This is essential pholgenetically, the model species most due to the high variability of sale prices closely related to coffee, for which signii‐ which are largely dependent on coffee ori‐ cant genetics and genomic resources are gin and variety. Prices of pure Arabica available, is the tomato. Both coffee and achieve prices upwards of 25% over ro‐ tomato belong to the Euasterid I clade of busta coffees taken as a whole and prices lowering plants and are likely descended for the inest specialty coffees can soar to from a single ancestral species with a hap‐ upwards of $50 per kilo green for rarities loid number of x=11 or 12. They go on to like the Panamanian Esmerelda. produce the irst comparative map for cof‐ fee and link it to the genetic and genomic resources available on tomato and other Near Infrared Relectance Solanaceae species. Near Infrared Relectance (NIRS) is based on the absorption of electromagnetic radia‐ Photo Courtesy of SREAD, Rwanda tion by matter. This method of analysis al‐ lows for the extraction of a large amount of information concerning biochemical com‐ position and is used extensively in a num‐ ber of crops. The ability to quickly extract a great deal of information makes NIR a highly cost effective source of information for researchers, coffee buyers and roasters.

In today’s marketplace coffee identiication and classiication is as crucial to cup qual‐ ity as it is to consumer requirement for origin and species speciication. In order to obtain top market prices, methods of ef‐

Global Coffee Quality Research Initiative Review 14 Draft Copy

icient, inexpensive, and highly accurate sociated with the presence of defective identiication of coffee origins and charac‐ beans. Acidity increased and pH levels de‐ teristics are paramount. creased as cup quality decreased likely due to the effect of defective beans which had NIRS is the method that Bertrand et al., undergone fermentation. After roasting, 2005 & Posada et al., 2009 have shown to the rio sample presented higher density be eficient for determining whether a trigonelline levels, indicating that it had green coffee comes from an Arabica tree not roasted to the same degree as the other that has been introgressed with C. samples tested. Canephora genes. Spectra taken from near‐infrared relectance of green coffee “Each of the main reserve compounds of were capable, by principal component and the bean (parietal polysaccharids, lipids, factorial discrimination, of correctly classi‐ proteins, sucrose) as well as secondary me‐ fying beans into categories of introgressed tabolites (chlorogenic acids, caffein, or non‐introgressed with degrees of accu‐ trigoneline) play a central role in chemical racy from 92.3% to 94.87%. This type of reactions during roasting. We believe that analysis can serve coffee buyers or roasters deciphering the correspondent metabolic as they seek to distinguish between non pathways are the key to better understand‐ introgressed Arabicas and genotypes car‐ ing quality and the use of biomarkers for rying chromosome fragments of C. breeding. On the other hand, volatile com‐ Canephora genetic material which could ponents, mainly from phenylpropanoids produce negative affects on cup quality. and isopropenoids, are synthesized during Posada et al., 2009 concurs that a near in‐ bean maturation. Even very low quantities frared spectroscopy signature that has (nano‐mole) might strongly inluence cup been acquired over a set of harvests can in quality. We have two ongoing preliminary fact effectively characterize a coffee variety. works (to be published) in which we have Posada hypothesizes that the spectral sig‐ been studying the inluence of environ‐ nature is affected by annual environmental ment and genetics on cup quality proiles factors but that through multiple harvest (metabolic ingerprints) using the SPME calibration data can be made useful for GC‐MS technology.” practical application to breeding. ‐‐Dr. Christophe Montagnon of CIRAD, per‐ sonal correspondence

Joet et al, 2010 examined the inluences of The Science of Taste environmental factors and wet processing on the lipid, chlorogenic acid, sugar and Chemical composition in relation to pre‐ caffeine content of green Arabica beans. established cup quality: Each of these biochemical markers repre‐ senting key components of cup quality. He Attributes of green coffee beans, both bean found that chlorogenic acids and fatty acids density and volume, were higher for softer in the bean were controlled by the average tasting samples as opposed to the rio off air temperature during bean development. lavored samples according to a study done However, total lipid, total soluble sugar, to‐ by Franca et al., 2004. The rio sample pre‐ tal polysaccharide and total chlorogenic sented lower lipid contents, most likely as‐ acid contents were not all inluenced by the

Global Coffee Quality Research Initiative Review 15 Draft Copy climate in which beans were produced. levels of caffeoylquinic acids, feruloylquinic Glucose content was positively affected by acids and their oxidation products were altitude and sorbitol content after wet associated with poor cup quality and the processing was directly dependent on glu‐ Rio‐off‐lavor. Similar correlations be‐ cose content in fresh beans. tween cup quality and chemical attributes were observed in green and light roasted samples which indicates that chemical Molecular Markers analysis of green beans may be used as an additional tool for coffee quality evalua‐ Molecular markers are widely used, tion. Lindinger et al., 2009 draws a corre‐ through traditional chemical analysis, to lation between the fermented off‐note and investigate canephora and liberica gene ethyl formate. introgression into Arabica lines as a resis‐ tance booster to pests and diseases. Coffee ampliied fragment length, polymorphism and simple sequence repeats have been used to analyze the introgressions men‐ Harvesting and Post tioned above. Villareal et al., 2009 found fatty acids in particular have proven effec‐ Harvest Handling tive for the discrimination of Arabica varie‐ ties and speciic growing terriors. Crop to crop environmental factors where found to have a signiicant impact on fatty acid con‐ tent and thus limit discrimination to mod‐ erate eficiency across a number of years. Posada et al., 2009 also found that correct classiication and discrimination among different varieties of introgressed genetic material was possible through traditional chemical analysis to the tune of 79% accu‐ racy. He also found in the same study that using spectral signatures in green beans Photo of Washing Station Courtesy SPREAD, Rwanda provided 100% correct differentiation among varieties. Ideal conditions for coffee production such Farah et al., 2005 investigated Brazilian as the agronomic factors of soil nutrition, green and roasted coffee beans for correla‐ shading, watering and superior genetics tions between cup quality and levels of su‐ will not yield high cup quality without op‐ crose, carreine, trigonelline and chloro‐ timal harvesting, processing, storage and genic acids as determinded by HPLC analy‐ brewing. According to Van Der Vossen, sis. They found that trigonelline and 2009 only freshly harvested and fully ripe 3.4‐dicaffeoylquinic acid levels in green berries should be used in any of the three and roasted coffee correlated strongly with methods of primary processing. Those high quality. To some extent caffeine levels methods include washed, semi‐dry and dry were found to be associated with good processing. Hand picking coffee beans is quality. The amount of defective beans, the

Global Coffee Quality Research Initiative Review 16 Draft Copy one method for accomplishing such dis‐ well as the fact that only the fully ripe cof‐ tinction but there are methods of mecha‐ fee cherries are typically used for wet nized picking that separate the immature processing as opposed to dry processing green from the ripe cherry before process‐ where fruits of varying stages of ripeness ing. Unripened coffee beans tend to pro‐ are commonly used. In his description of duce astringent, bitter and “off” lavored the “Inluence of processing on the content beverages but unripened beans can be of sugars in green arabica coffee beans” sorted before processing to mitigate nega‐ There is a close correlation between the tive effects on batch quality. Delays in type of post harvest processing and the depulping and prolonged fermentation of‐ content of fructose and glucose in the cof‐ ten lead to onion lavor or unpleasant fee bean. While in washed coffee beans smells. Wet fermentation may improve la‐ only a small amount of either hexos were vor marginally as may soaking under water present, those in unwashed coffees were for 24 hours after mucilage removal and signiicantly higher. It has been revealed washing. that low levels of both fructose and glucose are a result of decrease in the wet process. Below: A Survey on the low molecular sugars Fructose and glucose levels stayed near pre in a green coffee. The circle is representative processing levels throughout dry process‐ of the low molecular sugars, totaling 7.79% ing according to Knopp. DM. Knopp concludes that the decrease in glu‐ According to Bytof et al., 2007 the speciic cose and fructose in wet processed coffee is ambient conditions of any type of post har‐ as a result of a fermentation enhanced glu‐ vest processing can have signiicant im‐ cose turnover from anaerobic fermentation pacts on the time course of the metabolic in the coffee endosperm. reactions that occur during that processing (Knopp et al., 2005) period. The extent and the time courses of germination in various coffee beans were found to be signiicantly different between dry and wet processing styles. The highest germination activity was found to occur 2 days after the onset of wet processing and approximately 1 week after the onset of dry processing. Bytof goes on to conclude that the substantial differences in lavor between wet and dry processed coffees are the result of the differential expression of germination processes, in other words, they are the result of differences in the metabolic activities that take place in each Harvesting type of processing. (Bytof et al., 2007) When the coffee bean is ready for harvest‐ Knopp et al., 2005 states that lavor differ‐ ing it turns a dark berry color. This will ences in part have to be attributed to dif‐ typically take place between September ferences in the thoroughness applied to ei‐ and March in the Northern Hemisphere ther method of post harvest processing as and between April and May in the Southern

Global Coffee Quality Research Initiative Review 17 Draft Copy

Hemisphere. In some countries where tasting coffees and is used mostly in West‐ there is little clear distinction between dry ern Africa and Brazil. Harvested berries and wet seasons a major and minor crop are sorted and cleaned to remove dirt, may be able to be harvested annually. twigs and leaves by hand . Berries are then Countries on the equator are able to har‐ spread out in the sun and raked regularly vest year round. Harvesting of superior to keep fermentation at bay. In the dry coffees tends to be done by hand as method, coffee beans were dried as a mechanized picking can gather over ripe whole with pulp and mucilage in the cherry berries and an acrid taste may affect cup state. Dry processing is slow and can lead quality according to the FAO ofice for Asia to the translocation of chemical constitu‐ and the Paciic. (Alastair Hicks) Of the ents from the pulp to the inner bean as well various methods in practice, selective hand as chemical transformation that depends picking best ensures that only fully ripe on ambient conditions. beans are taken. However a cost beneit analysis will inevitably be undertaken by It is noted by Clark, 1985 that naturally, coffee growers as to whether the increased dry, processed coffee has a better body due level of quality found in picking only the to the fact that the bean was in contact ripest cherries is worth the extra expense with its mucilage through a greater part of of undertaking multiple harvests in the the processing phase. same season. Cantergiani et al., in 2001, found that Approximatley 20‐25% of all Brazilian cof‐ earthy/moldy off‐lavors that were pro‐ fee production is affected by the Rio‐off de‐ duced in Mexican coffees were likely as a fect. This defect is characterized by a result of dry post harvest treatment. The strong medicinal or iodine or musty and compounds geosmin, 2‐methylisoborneol cedar‐like character. Occasionally this de‐ (MIB), 2,4,6‐trichloroanisole (TCA) were fect has been observed in other countries. found to be the main contributors to the Spadone et al. contend that this defect is a tainted lavor though at the time of Canter‐ result of wet and humid conditions during giani’s study the moldy /earthy defect had harvesting and is the consequence of a type not yet been fully characterized. The con‐ of fermentation and bacterial growth and a centrations found in samples exhibiting the high level of cell structure degradation that moldy/earthy characteristic were between ensues. The principle compound associ‐ 100 and 1000ppm, 5‐8 times more than in ated with the Rio‐off lavor was found to be the reference sample. Cantergiani also 2‐3‐6 Trichloroanisole and is only percep‐ identiied alkyl methoxy pyrazines as hav‐ tible at concentrations at or above 8ppm. ing less signiicant contributions to the ef‐ (Spadone et al., 1990) fect. 2‐methoxy‐3‐isopropyl pyrazine, 2 methoxy‐3‐sec‐butyl pyrazine and 2‐ methoxy‐3‐isobutyl pyrazine (MiPP, MsBP Dry Processing and MiBP respectively) were found in both reference and tainted samples with the Post harvest processes include both dry concentration being only 1‐2 times higher and wet methods used to process green in tainted samples than in the reference cherry coffees. Dry processing is the sim‐ sample. These chemicals evoke strong plest and least expensive method of coffee earthy, green and bell pepper notes and processing. It tends to produce “natural” MsBP was detected in this study for the

Global Coffee Quality Research Initiative Review 18 Draft Copy

irst time in coffee. (Cantergiani et al., begin to become active. The yeasts go on to 2001) convert sugar to alcohol but are also me‐ tabolizing the solid parts of the mucilage Spadone identiied 2,4,6‐trichloroanisole resulting in aroma qualities that can have a as primarily responsible for the Rio off‐ negative impact on cup quality. This taste / lavor in coffee. Vitzthum et al., propounds smell characteristic is sometimes referred MIB as a key substance responsible for the to as “fruity coffee.” When coffee continues earthy tone in Robusta coffee. Rougue et in this state even longer under reducing al., notes his view that the effect of MIB in and acidic conditions, the yeast will con‐ Arabica coffees can be mitigated using vert sugars into acids as opposed to alcohol steam‐treatment and roasting. resulting in sour tasting coffee. (Tea & Cof‐

Wet Processing Below: The concentration of low molecular sugars in variously processed green coffees. In the wet method coffee beans are pulped, fruit and skin are removed, or pulped and demucilated and mucilaginous mesocarp is removed under fermentation. In the wet method fermentation occurs in water at controlled temperatures which produce lower levels of undesirable lavors. For this reason the wet method is often associ‐ ated with better cup quality. Gonzales‐Rios fee, November – December, 2005) Rec‐ et al., 2007 claim that the quality of green ommendations for avoidance of fruity and and roasted coffee, measured by aroma, sour lavors include washing of beans as was better after conventional fermentation soon as fermentation has inished, when all than after mechanical mucilage removal. mucilage has been liqueied. There is no across the board time frame for develop‐ Wet coffee readily takes up smells and ment of fruity and sour lavors as tempera‐ aromas. Oil constitutes a major component ture and altitude play signiicant roles in of the coffee bean’s composition and is able those processes. It is noted that in general to take up and store smells and lavors be‐ the best way to avoid these cup defects is to fore releasing them during roasting greatly wash the parchment coffee as soon as fer‐ affecting cup characteristics. According to mentation has inished and parchment the tea and coffee trade journal most cup feels rough to the touch. defects are caused by wrong fermentation. In fermentation there are primarily two Clifford, 1985 proposes that Wet processed microorganisms at work to shape the even‐ Arabica tends to be aromatic with a ine tual cup quality, bacteria and yeasts. Dur‐ acidity but some astringency while dry ing proper fermentation the bacteria feed processed Arabica tends to be less aro‐ on sugars in the mucilage. As soon as the matic but produce greater body. This is sugars have been digested and the muci‐ largely due to the formation of acids in un‐ lage has been liqueied, the pH in the fer‐ der water fermentation. mentation tank begins to decrease. It is at this point of lowering pH that the yeasts

Global Coffee Quality Research Initiative Review 19 Draft Copy

Gonzales‐Rios et al., 2006 states that a comparison of green coffees from the dif‐ ferent treatments (wet and dry processing) Storage reveals the importance of mucilage re‐ moval in water to obtain coffees with bet‐ Signiicant defects can also arise as a result ter aroma quality. These wet processed of insuficient drying and / or storage con‐ coffees are in fact characterized by pleas‐ ditions as it is the drying process that pre‐ ant and fruity aroma characteristics pares beans for processing later on as well whereas those obtained after a mechanical as storage. When beans are insuficiently mucilage removal used in a more “ecolo‐ or unevenly dried a decrease in cup quality cial” process were characterized by less can occur much more rapidly than with pleasant aromatic notes. beans that have undergone an ideal drying process. Stirling, 1974 shows a rapid de‐ Kleinwachter and Selmar contend that crease in cup quality level with increasing when wet processed coffee beans are dried, storage time from 6 to 18 months given there ensues a decrease in the water poten‐ various moisture contents. The decline in tial which causes a number of metabolic cup quality in wet coffee is due to mold and responses within the coffee bean. They bacteria as molds and bacteria grow best in found that the content of fructose and glu‐ moisture rich environments and cup char‐ cose decreased signiicantly within the irst acteristics change as a result of bacteria day of drying. This decrease in sugar con‐ and mould utilization of sugars in the cof‐ tent, they contend, proves that the lower fee bean for metabolism. Tea & Coffee rec‐ contents of glucose and fructose that are ommend a bean moisture content of 10 to generally found in wet processed coffees 12% before packaging and storage. are, at least in part, due metabolic proc‐ esses within the bean and are not the result One key aspect of coffee storage is bean of leaching of sugars into the water used respiration. Every 24 hours an average of for processing. (Kleinwachter and Selmar, 4.4 milligrams of CO2 are produced by 100 2010) grams of coffee beans and the 96 calories of heat produced by the 4.4 mg of CO2 will In 2001 Mendez et al. found that signiicant raise the temperature of the beans .25o Cel‐ improvements could be made in customer sius. A high respiration rate, in combina‐ receptivity to robusta coffees given high tion with heat generation, can cause a loss quality processing. The study revealed that of weight and dry material in the bean as when roasted at optimum ranges of 225 ‐ well as bean fat decomposition which plays 230* C for 22 to 28 minutes and wet proc‐ a key role in the aroma of the cup. (Sivetz essed, acceptance scores increased to and Dosrosier, 1979). “liked slightly” & “liked moderately” on the hedonic scale for both aroma and lavor. “Stink coffee” can be produced as a result Mendez goes on to make the argument that of excessive fermentation from the normal with higher lavor and aroma scores a microbes that are at work in coffee proc‐ higher percentage of robusta may be used essing. It is recommended that factory in blended coffees without adverse effects tanks and machinery be cleaned daily to on customer satisfaction. (Mendez et al., ensure that old beans caught in crevices of 2001) machinery do not contaminate a later batch of coffee. Extreme over‐fermentation

Global Coffee Quality Research Initiative Review 20 Draft Copy can germinate the coffee seed which dies in fresh wet washed parchment which quickly and leaves a hollow pit in the end shows up the color of the silverskin under‐ of the bean. The dead bean then very neath. One solution is to dun dry the coffee quickly develops a cheese smelling texture when weather conditions permit as ultra‐ which is highly distinctive when the bean is violet light can bleach out the greenness in broken or cut. A single bean can contami‐ the silverskin. A slight degree of green nate and spoil an entire batch of perfectly color will often fade over time, making it good coffee. (Coffee & Tea, 2005) undetectable at the inal destination, but a strong degree of unripeness will facilitate Declining of cup quality is inevitable to chemical absorption back into the oil frac‐ some degree during storage. The reason tion of the inal product. for this decline, or aging, is surface oxida‐ tion from beans aided by microorganisms. The aging effect can be minimized by stor‐ Below: Aroma profiles of unstored raw coffee ing parchment and process beans in low beans (gray line) and coffee bean with a water content of 13.5% stored for 9 months at 40*C temperature, low oxygen and low humidity under an atmosphere of 20% oxygen. conditions in order to dissuade bacterial, yeast and mold activity. Damaged beans tend to be much more subject to agind as the oils in the bean are extruding from the bean and provide a good growing ground for mold and bacteria. Molds have the po‐ tential to grow on dry surfaces and extract needed moisture from the air in storage rooms making humidity a key issue in stor‐ age facility design, maintenance and repair. One potential mitigation of the aging proc‐ ess on coffee beans is to utilize hermeti‐ cally sealed containers or storage silos as opposed to bags. Such units could mini‐ mize oxygen levels as well as moisture con‐ Onion lavor is another potential result of tact metabolic rates of microorganisms and faulty post processing. Onion lavor occurs prolonging the amount of time in storage when the ration of soluble sugars to proto‐ that would cause minimal affects on cup pectins (contained in the mucilage) be‐ quality. Little information is available comes too low. The primary rapid buildup about the effect of CO2 in such a system on of fermentation bacteria is fuelled by the the cup quality of the coffee stored. relatively high level of sugar present within the ripened mucilage. If excessive fresh water is used in pre‐washing of cherries Green beans can produce a “grassy” or and during the pulping process, most of harsh lavor cause by picking and process‐ these soluble sugars are leached out before ing immature cherries. Late in the picking normal fermentation takes place. Conse‐ season many cherries loose their green quently, the beneicial soft rot bacteria can color but do not turn completely red. be overwhelmed later in the fermentation These unripe cherries will pulp easily but process not only by the yeasts but also by are full of chlorophyll. This is readily seen bacteria which convert acetic and lactic ac‐

Global Coffee Quality Research Initiative Review 21 Draft Copy ids to propionic and butyric acids which they are more susceptible to quality dete‐ cause the onion lavor. It has been found rioration. Hermetic enclosures manufac‐ that this fault can be minimized by recy‐ tured by Grainpro Inc. in Concord, Massa‐ cling of pulping water as maintaining a chusetts, are being used to store coffee. high level of sugars and enzymes in the wa‐ These enclosures use ultra violet light re‐ ter will facilitate normal bacterial action. sistance PVC to provide an environment that maintains a very low moisture (hu‐ Scheidig et al., 2007 studied the atypical midity), low oxygen, high CO2 environment. odors that develop during the storage of This type of environment is highly condu‐ raw coffee beans and eventually inluence cive to maintaining low levels of bacterial the aroma of the coffee beverage. His re‐ life, mold, insects and even prevents the sults showed strong increases in damasce‐ formation of alatoxins as a byproduct of none (cooked apple‐like odor), 2methoxy‐ mold development. Temperature and hu‐ 4‐vinyphenol (clove‐like odor), and methyl midity ranges found in the hermetically 2‐methyl and methyl‐3methylbutanoate sealed enclosures were also signiicantly (fruity odor), with an increased water con‐ narrower than for control groups of beans tent and a storage time of 9 months at 40 stored in silos or in bags in a warehouse. *C. It was also found that earthy smelling Cupping tests were done monthly on a ive 3‐isopropyl‐2‐methoxypyrazine as well as point scale and found that after two 2‐phenylethanol and 3‐methozyphenol months there was a signiicant change in remained unchanged during the storage quality in beans stored in sacks and in si‐ period. It was also found that a previously los, from 4.0 to 3.8 while cup quality stayed unknown odor in coffee, the intense smoky the same for beans stored in the hermetic lavor caused by production of 2‐methozy‐ enclosure. After a storage of ive months 5‐vinylphenol, increased signiicantly due cup quality for sacked coffee and coffee to storage. In conclusion, Scheidig recom‐ stored in a silo had decreased a full point to mends a reduction in water content of 3.0 and were described as “Slight old lavor green beans in combination with lower perceptible in cup, slight harshness, storage temperatures to avoid aroma tainted.” The cup quality from beans changes in raw coffee beans over long stor‐ stored hermetically was noted as “Very age periods. (Scheidig et al, 2007) good lavor despite being from the previ‐ ous harvest. Slight loral lavor.” Cupping was done by Café Britt’s cupper Carmen Café Britt of Costa Rica, in conjunction with Lidia Chavarria as well as the cupper for the Mesoamerican Development Institute, the Costa Rican Consortium, COOCAFE, is experiencing positive results in the de‐ Jimmy Bonilla. velopment of a hermetic storage unit. This unit facilitates long term storage of coffee in it’s parchment state without the use of Roasting pesticides, the degradation of cup quality, aroma, or appearance for a period of ive Roasting has been called the most impor‐ months or more. Coffee beans, even when tant step in the production of aroma and properly dried can reabsorb fungus and taste by Juerg Baggenstoss with the insti‐ bacteria encouraging moisture over time tute of Food Science and Nutrition in Zu‐ from the atmosphere. The storage of green rich. As bean color turns to brown or even beans can be even more problematic as black in the roasting process brittleness is

Global Coffee Quality Research Initiative Review 22 Draft Copy greatly increased which facilitates the of volatiles with the notable exceptions of grinding and extraction processes. High haxanal, pyridine, and dimethyl trisulide, temperatures, above 200* C, are necessary whose concentrations increased continu‐ for the roasting process to occur. Green ally during over roasting. When roasters beans are especially hard due to unusually were operated in the “temperature proile” robust cell walls and a lack of intracellular mode, along the identical development of space inherent in the coffee bean. As a re‐ coffee bean temperature over roasting sult, coffee beans are regarded as an aggre‐ time, the kinetics of aroma generation were gate of micro reactor units that sustain a similar in both drum roaster and luidizing considerable pressure buildup due to the bed roaster processes. heat of roasting with theoretical values of up to 16 bars being produced within the Franca et al., 2009 concludes that caffeine cell walls with no evidence of the cell wall and pyridine were the main discriminating disruption being observed in scanning compounds for coffees roasted at high electron microscope investigation. This is temperatures and more intense degrees of thought to be due to the fact that at high roast. Although discrimination of coffee temperature coffee cell walls change from a samples roasted to different degrees was hard glassy state to a more elastic state attained by analyzing the effect of hun‐ which facilitates considerable volume in‐ dreds of compounds together, the com‐ crease without rupturing. It has been gen‐ pounds that presented the most pro‐ erally noted that convective heat transfer is nounced effect on the discrimination of the most effective for uniform roasting. roasting degrees and temperatures were those generated right after the onset of py‐ The state of roasted coffee is inluenced by rolysis (decomposition brought about by the roasting conditions in terms of degree high temperatures). The results presented of roast. A number of options exist for de‐ in this study reveal that color and weight scription of degree of roast; color devel‐ loss in and of themselves are not entirely opment, roast loss, organic roast loss and reliable as roasting degree assessment cri‐ water content. Of these, the color of the teria and the temperature at which beans coffee bean or ground coffee is the most are roasted must also be taken into consid‐ commonly used indicator of roast as bean eration. color intensity has been correlated to roasting temperature. This however, is not Brewed Coffee a direct relationship as color does vary as a function of raw material and process condi‐ Once coffee is brewed the decline of pH and tions. (Baggenstoss et al., 2008) the quality score were correlated at a number of storage temperatures by Rosa et Baggenstoss went on to explore the effect al, 1989. Roas’s sensory analysis allowed of roasting on aroma formation under dif‐ deinition of a lower limit of pH at which ferent time‐temperature conditions. When coffee’s shelf life ended. Sivetz and Desro‐ compared to low temperature‐long time sier, 1979 showed that the decline in pH roasting, high temperature‐short time comes as s result of the formation of caffein roasting results in relevant differences in and quinic acids as breakdown products of the physical properties of aroma formation. chlorogenic acid. This process was found It was found that excessive roasting typi‐ to accelerate with increasing storage tem‐ cally lead to decreasing or stable amounts perature. Results show that coffee quality

Global Coffee Quality Research Initiative Review 23 Draft Copy remains high as long as pH remains high, in the 5.2 range, after brewing and that the The Impact of ideal temperature regime for storage is near freezing at approximately 4oC. Climate Change

Global climate change in the coming years will have signiicant impact on the coffee producing regions of the world. Current correlations between altitude, tempera‐ ture, rainfall and cup quality as well as yield will all be modiied to some lesser or greater extent. It is generally recognized that higher mean temperatures and changes in precipitation regimes will typify the climate changes that we are likely to see across the globe in the next 50 years according to a number of global circulation models (GCM’s).

Coffee systems are characterized by long lead times necessary for both farmers and their business partners to make accommo‐ dation for change in growing patterns, with an average of 8‐15 years being required from decision to fruition. (Laderach et al., 2009) Laderach et al., 2008 predicts that climate change will shift the prime altitude range to higher elevations over time. He estimates that in Central America a shift will occur from 1200m for optimal plant health and cup quality to 1400m in 2020 and 1600m in 2050.

Overall climate will become more seasonal in terms of variability throught the year in terms of temperature and precipitation on the whole with temperature increasing in South America on the order of 2* C by 2050. Consequently, suitability for coffee will move upwards on the altitudinal gra‐ dient with climate change. Lower altitude areas will lose suitability due to their higher temperatures.

Global Coffee Quality Research Initiative Review 24 Draft Copy

The areas that will be highly suitable for but signiicant tradeoffs exist due to the coffee production in 2050 are likely to be possibility of decreasing quality of coffees those areas that are highly suitable today, it picked and processed mechanically at one is the large marginal areas that will proba‐ time as opposed to hand picked beans bly lose their capacity for coffee produc‐ picked at the height of their maturity. Post tion. Altitude and temperature have a ixed harvest separation of ripe from unripe relation called the lapse rate which is equal beans remains a key quality control that to .6*C per 100m of altitude. Precipitation has the potential to mitigate this issue if in the wettest month of the year and mean properly applied, as pointed out by Dr. Car‐ temperature of the coldest quarter of the los Brando. year tend to be the most decisive factors in the suitability of an area for coffee produc‐ At this point major tradeoffs exist for stor‐ tion as stated by Laderach, 2009. Assum‐ age as well. Increased storage times have ing that these indings are widely applica‐ traditionally meant supply chain stability ble, global climate change will require cof‐ at the expense of cup quality. However, fee producers to implement signiicant technologies are being developed that strategic planning to adjust their crops and could change even this paradigm in the production areas in order to mitigate the near future. negative affects of changing climate pat‐ terns. The future of coffee is bright given the con‐ tinued and expanded dissemination and implementation of cutting edge research at the farm level. Working in concert with one another as an industry to implement such research, there is no limit to our com‐ bined potential for producing coffee that is Conclusions as excellent as it is unique in economically viable quantity. While the data support few universal equa‐ tions for coffee quality, a few key tradeoffs are evident from literature. Firstly, that there is a tradeoff in perception, if not in fact, between disease resistance and cup quality. Further study has the potential to uncover thresholds in productivity and cup quality tradeoffs in this ield. Soil nutrition will also be a key area of focus in the com‐ ing years as decreasing soil quality due to consistent high yield output could cause further constriction of the supply chain through decreasing yields. This is largely due to the inability of small land holder farmers to provide the inputs necessary for consistent and long term quality harvests. Larger scale inputs could be facilitated through larger scale farming techniques

Global Coffee Quality Research Initiative Review 25 Draft Copy

Journal of Agricultural and Food Chemistry References 56(14): 5836‐5846.

Alpizar, E., Etienne, H. &Bertrand, B. Bertrand, B., Vaast, P., Edgardo, A., Etienne, (2007). Intermediate resistance to Meloi‐ H., Davrieux, F., Charmetant, P. (2006) dogyne exigua root‐knot nematode in Cof‐ Comparison of bean biochemical composi‐ fea arabica. Crop Protection 26(7): 903‐ tion and beverage quality of Arabica hy‐ 910. brids involving Sudanese ‐Ethiopian ori‐ gins with traditional varieties at various Amorim, A. C. L., Hovell, A. M. C., Pinto, A. C., elevations in Central America. Tree Physi‐ Eberlin, M. N., Arruda, N. P., Pereira, E. J., ology 26, p. 1239‐1248. Bizzo, H. R., Catharino, R. R., Morais, Z. B. &Rezende, C. M. (2009). Green and Roasted Bertrand, B., Etienne, H., Lashermes, P., Gu‐ Arabica Coffees Differentiated by Ripeness, yot, B. & Davrieux, F. (2005). Can near‐ Process and Cup Quality via Electrospray infrared relectance of green coffee be used Ionization Mass Spectrometry Fingerprint‐ to detect introgression in Coffea arabica ing. Journal of the Brazilian Chemical Soci‐ cultivars? Journal of the Science of Food ety 20(2): 313‐321. and Agriculture 85(6): 955‐962.

Anthony, F., Bertrand, B., Quiros, O., Bertrand, B., Guyot, B., Anthony, F. &Lash‐ Wilches, A., Lashermes, P., Berthaud, J. & ermes, P. (2003). Impact of the Coffea Charrier, A. (2001). Genetic diversity of canephora gene introgression on beverage wild coffee (Coffea arabica L.) using mo‐ quality of <i>C. arabica</i>. TAG lecular markers. Euphytica 118(1): 53‐65. Theoretical and Applied Genetics 107(3): 387‐394. Anonymous, 1988. Coffee Research Foun‐ dation Annual Report, 1987/88 Bosselmann, A. S., Dons, K., Oberthur, T., Olsen, C. S., Ræbild, A. &Usma, H. (2009). Avelino, J., Barboza, B., Araya, J. C., Fonseca, The inluence of shade trees on coffee qual‐ C., Davrieux, F., Guyot, B. & Cilas, C. (2005). ity in small holder coffee agroforestry sys‐ Effects of slope exposure, altitude and yield tems in Southern Colombia. Agriculture, on coffee quality in two altitude terroirs of Ecosystems & Environment 129(1‐3): 253‐ Costa Rica, Orosi and Santa María de Dota. 260. Journal of the Science of Food and Agricul‐ ture 85(11): 1869‐1876. Cantergiani, Brevard, Krebs, Y., Feria, M., Amadò &Yeretzian (2001). Characterisa‐ Beer, J. (1987). Advantages, disadvantages tion of the aroma of green Mexican coffee and desirable characteristics of shade and identiication of mouldy/earthy defect. trees for coffee, cacao and tea. Agrofor‐ European Food Research and Technology estry Systems 5(1): 3‐13. 212(6): 648‐657.

Baggenstoss, J., Poisson, L., Kaegi, R., Per‐ Carvahal, J.F., 1959. Mineral nutrition and ren, R. &Escher, F. (2008). Coffee Roasting crop requirements of the coffee crop. Min‐ and Aroma Formation: Application of Dif‐ isterio de Agriculture y Ganaderia (MAG) – ferent Time−Temperature Conditions. Servicio Tecnico Interamericanode Coop‐

Global Coffee Quality Research Initiative Review 26 Draft Copy eracin Agricola, Technical Bulletin No. 9. Farah, A., Monteiro, MC., Calado, V., Franca, San Jose, Costa Rica 1959. 16pp AS., Trugo, LC. (2006) Correlation between cup quality and chemical attributes of Bra‐ Clarke R.J and K.C. and R.Macrae.1988. Cof‐ zilian coffee. Food Chemistry 98 p.373‐380. fee: Agronomy.Vol.4. Franca, A. S., Mendonça, J. C. F. &Oliveira, S. Clifford, M.N., 1985. Chemical and physical D. (2005). Composition of green and aspects of green coffee and coffee products. roasted coffees of different cup qualities. In: M.N. Clifford and K.C. Willson (Eds.), LWT ‐ Food Science and Technology 38(7): Coffee botany, biochemistry, and produc‐ 709‐715. tion of beans and beverage, pp. 305‐374. Croom Helm, London Geromel, C., Ferreira, L., Davrieux, F., Guyot, B., (2008). Effects of shade on the devel‐ Clifford, M. N. and K.C. Willson. 1985. Cof‐ opment and sugar metabolism of coffee fee: Botany, Biochemistry and Production fruits, Plant Physiology and Biochemistry of Beans and Beverages. Croom Helm, Lon‐ 46 p.569‐579. don and Sydney. Gonzalez‐Rios, O., Suarez‐Quiroz, M. L., De Castro, R. D. & Marraccini, P. (2006). Cy‐ Boulanger, R., Barel, M., Guyot, B., Guiraud, tology, biochemistry and molecular J.‐P. &Schorr‐Galindo, S. (2007). Impact of changes during coffee fruit development. "ecological" post‐harvest processing on Brazilian Journal of Plant Physiology 18: coffee aroma: II. Roasted coffee. Journal of 175‐199. Food Composition and Analysis 20(3‐4): 297‐307. Decazy, F., Avelino, J., Guyot, B., Perriot, J., Pineda, C. &Cilas, C. (2003). Quality of Dif‐ Grossman, J. M., Sheaffer, C., Wyse, D., Buc‐ ferent Honduran Coffees in Relation to Sev‐ ciarelli, B., Vance, C. &Graham, P. H. (2006). eral Environments. Journal of Food Science An assessment of nodulation and nitrogen 68(7): 2356‐2361. ixation in inoculated Inga oerstediana, a nitrogen‐ixing tree shading organically De Nardi, B., Dreos, R., Del Terra, L., Martel‐ grown coffee in Chiapas, Mexico. Soil Biol‐ lossi, C., Asquini, E., Tornincasa, P., Gasper‐ ogy and Biochemistry 38(4): 769‐784. ini, D., Pacchioni, B., Rathinavelu, R., Pal‐ lavicini, A. &Graziosi, G. (2006). Differential Hicks, Alastair (2002). Post Harvest Proc‐ responses of Coffea arabica L. leaves and essing and Quality Assurance for Specialty/ roots to chemically induced systemic ac‐ Organic Coffee Products. FAO Regional Of‐ quired resistance. Genome / National Re­ ice for Asia and the Paciic search Council Canada = Génome / Conseil National De Recherches Canada 49(12): Joët, T., Laffargue, A., Descroix, F., Doulbeau, 1594‐1605. S., Bertrand, B., kochko, A. d. & Dussert, S. (2010). Inluence of environmental factors, Esteban‐Díez, I., González‐Sáiz, J. M., Sáenz‐ wet processing and their interactions on González, C. & Pizarro, C. (2007). Coffee va‐ the biochemical composition of green Ara‐ rietal differentiation based on near infra‐ bica coffee beans. Food Chemistry 118(3): red spectroscopy. Talanta 71(1): 221‐229. 693‐701.

Global Coffee Quality Research Initiative Review 27 Draft Copy

Joët, T., Laffargue, A., Salmona, J., Doulbeau, Lindinger, C., Pollien, P., de Vos, R. C. H., S., Descroix, F., Bertrand, B., De Kochko, A. Tikunov, Y., Hageman, J. A., Lambot, C., Dussert, S. (2009). Metabolic pathways in Fumeaux, R., Voirol‐Baliguet, E. &Blank, I. tropical dicotyledonous albuminous seeds: (2009). Identiication of Ethyl Formate as a Coffea arabica as a case study. New Phy‐ Quality Marker of the Fermented Off‐note tologist 182(1): 146‐162. in Coffee by a Nontargeted Chemometric Approach. Journal of Agricultural and Food Kemsley, E. K., Ruault, S. & Wilson, R. H. Chemistry 57(21): 9972‐9978. (1995). Discrimination between Coffea arabica and Coffea canephora variant ro‐ Lyon, S., Bezaury, J. A. & Mutersbaugh, T. busta beans using infrared spectroscopy. (2010). Gender equity in fairtrade‐organic Food Chemistry 54(3): 321‐326. coffee producer organizations: Cases from Mesoamerica. Geoforum 41(1): 93‐103. Knopp, S., Bytof, G. & Selmar, D. (2006). In‐ luence of processing on the content of Malta, MR. (2008). Cup Quality of Tradi‐ sugars in green Arabica coffee beans. Euro‐ tional Crop Coffee Converted to Organic pean Food Research and Technology System. Bragantia, Campinas, v.67, n.3 223(2): 195‐201. p.775‐783.

Lefebvre‐Pautigny, F., Wu, F., Philippot, M., McDowell, J. M. &Woffenden, B. J. (2003). Rigoreau, M., Priyono, Zouine, M., Frasse, Plant disease resistance genes: recent in‐ P., Bouzayen, M., Broun, P., Pétiard, V., sights and potential applications. Trends in Tanksley, S. &Crouzillat, D. (2009). High Biotechnology 21(4): 178‐183. resolution synteny maps allowing direct comparisons between the coffee and to‐ Montavon, P., Duruz, E., Rumo, G. &Pratz, G. mato genomes. Tree Genetics & Ge‐ (2003). Evolution of Green Coffee Protein nomes 6(4): 565‐577. Proiles with Maturation and Relationship to Coffee Cup Quality. Journal of Agricul‐ Peter Laderach, M. L., Andy Jarvis, Julian tural and Food Chemistry 51(8): 2328‐ Ramirez, Emiliano Portilla, Kathleen 2334. Schepp (2009). Predicted Impact of Cli‐ mate Change on Coffee‐Supply Chains. Muschler, R. (2001). Shade improves coffee CIAT. quality in a sub‐optimal coffee‐zone of Costa Rica. Agroforestry Systems 51(2): Leroy, T., Ribeyre, F., Bertrand, B., Charme‐ 131‐139. tant, P., Dufour, M., Montagnon, C., Marrac‐ cini, P., Pot, D. (2006) Genetics in Coffee Posada H., Ferrand M., Davrieux F., Lasher‐ Quality. Brazilian Journal of Plant Physiol‐ mes P., Bertrand B. (2009) Stability across ogy, 18:229‐242. environments of the coffee variety near in‐ frared spectral signature. Heredity, 102 (2) Luderach, P., Vaast, P., Oberthur, T., O’Brien, : 113‐119. R. (Date Unknown). Geographical Analyses to Explore Interactions between Inherent Rouge, F., Gretsch, C., Christensen, K., Kiar‐ Coffee Quality and Production Environ‐ don, R. (1993). Thermal stability of 2‐ ment. methylisoberneol in Robusta coffee. Collo‐

Global Coffee Quality Research Initiative Review 28 Draft Copy que Scientiique International. Paris, pp Vaast, P., Bertrand, B., Perriot, J. J., Guyot, B. 866‐868 &Génard, M. (2006). Fruit thinning and shade improve bean characteristics and Rosa, M. D., Barbanti, D. & Lerici, C. R. beverage quality of coffee (Coffea arabica (1990). Changes in coffee brews in relation L.) under optimal conditions. Journal of the to storage temperature. Journal of the Sci‐ Science of Food and Agriculture 86(2): ence of Food and Agriculture 50(2): 227‐ 197‐204. 235. Vaast, P., Zasoski, R. J. &Bledsoe, C. S. Salmona, J., Dussert, S., Descroix, F., de Ko‐ (1998). Effects of solution pH, temperature, chko, A., Bertrand, B. &Joët, T. (2008). De‐ nitrate/ammonium ratios, and inhibitors ciphering transcriptional networks that on ammonium and nitrate uptake by Ara‐ govern <i>Coffea arabica</i> bica coffee in short‐term solution culture. seed development using combined cDNA Journal of Plant Nutrition 21(7): 1551 ‐ array and real‐time RT‐PCR approaches. 1564. Plant Molecular Biology 66(1): 105‐124. Villarreal, D., Laffargue, A., Posada, H., Ber‐ Scheidig, C., Czerny, M. &Schieberle, P. trand, B., Lashermes, P. &Dussert, S. (2009). (2007). Changes in Key Odorants of Raw Genotypic and Environmental Effects on Coffee Beans during Storage under De‐ Coffee (Coffea arabica L.) Bean Fatty Acid ined Conditions. Journal of Agricultural Proile: Impact on Variety and Origin Che‐ and Food Chemistry 55(14): 5768‐5775. mometric Determination. Journal of Agri‐ cultural and Food Chemistry 57(23): Selmar, D., Bytof, G., Knopp, S. E. & Breiten‐ 11321‐11327. stein, B. (2006). Germination of Coffee Seeds and its Signiicance for Coffee Qual‐ Vitzthum, OG., Weiseman, C., Becker, R., ity. Plant Biology 8(2): 260‐264. Kohler, HS., (1990) Identiication of an aroma key compound in Robusta Coffees. Cafe, Cacao, The 34(1):27‐36 Siddiqi, M.A., 1980. The selection of Ara‐ bica coffee for Fusarium bark disease resis‐ Wintegens, J.N.2004.Coffee: Growing, proc‐ tance in Bvumbwe, Kenya Coffee, 45:55‐59 essing, sustainable production, a guide book for growers, processors, traders, and Sivetz M. and N.W. Desrosier. (1979). Coffee researchers, WILEY‐VCH Verlag GmbH & Technology. The AVI Pub. Co.KGaA, Weinheium. Com.Westport, Connecticut. Youkhana, A. & Idol, T. (2009). Tree prun‐ Sivetz. (1972). How acidity affects coffee ing mulch increases soil C and N in a lavor. Food Technology Champaign 26, shaded coffee agroecosystem in Hawaii. p70‐77. Soil Biology and Biochemistry 41(12): 2527‐2534. Spadone, J. C., Takeoka, G. & Liardon, R. (1990). Analytical investigation of Rio off‐ (2007). Excellence in a cup. Economist lavor in green coffee. Journal of Agricul­ 382(8513): 66‐66. tural and Food Chemistry 38(1): 226‐233.

Global Coffee Quality Research Initiative Review 29