Incompleteresistancetocoffeeleafrust (Hemileia vastatrix)

Incmpleteresistancetcffeeleafrust (Hemileia vastatrix) Aanmijnuders ParaHelisa,Chantra e Nanda CENTRALE LANDBOUWCATALOGUS LAHDBOUW W.h.;. «SCHOOL QQQQ QQQQ g 315

Prmtr : dr.ir. J. Sneep, emeritus hgleraar in deleer van deplantenveredeling C-prattr: dr. J.C. Zadks, hgleraarin defytpathlgie

A.B. ESKES fouo ^!,1S3 INCOMPLETE RESISTANCE TOCOFFEE LEAF RUST (Hemileia vas tat rix) Prefschrift terverkrijging van de graad van dctr in de landbuwwetenschappen, p gezag van de rectr magnificus, dr. C.C. Csterlee, hgleraarin de veeteeltwetenschap, in het penbaar te verdedigen cp vrijdag25 maart 1983 desnamiddagste vier uur in de aula van de Landbuwhgeschl te Wageningen \^>-W = \ 1 r lr5^-o r i>

STELLINGEN jj/810\;«?33 Selectie p plygeen verervende incmplete resistentie tegen ziekteverwekkers, met als del het verkrijgen van duurzame resistentie, is veelal te ingewikkeld vr de praktische veredeling. 2. Bij het streven naar verhging van de landbuwprduktiviteit in ntwikkelingslanden wrdt het belang van teeltkundige aspecten nderschat en het belang van de veredeling verschat. De p zich interessante hypthese van Nelsn et al. (1970) dat genen vr 'hrizntale' en 'vertikale' resistentie identiek zuden zijn, is in feite gebaseerd p een njuiste uitleg van hun eigen nderzeksresultaten en van het begrip 'hrizntale' resistentie. Nelsn, E.R., D.R. MacKenzie & G.L. Scheifele, 1970. Interactin f genes fr pathgenicity and virulence in Tri chômetasphaeria turcica with different number f genes fr vertical resistance in Zea mays. Phytpathlgy 60: 1250-1254. 4. De term 'hrizntale' resistentie blijft aanleiding geven tt verwarring. De term i s nndig en zu daarm n i e t verder gebruikt dienen te wrden. Dit prefschrift. 5. De dr Ou (1980) vermeende instabiliteit van fysi's van Pyricularia ryzas berust grtendeels p een verkeerde interpretatie van het hetergene reactietype. Ou, S.H., 1980. Pathgen variability and hst resistance in rice blast disease. Ann. Rev. Phytpath. 18: 167-187. Het 'interactiemdel' beschreven dr Parlevliet en Zadks (1977) verklaart niet de variatie in pathgeniteit waargenmen p waardplanten znder resistentiegenen. Parlevliet, J.E. & J.C. Zadks, 1977. The integrated cncept f disease resistance; a new view including hrizntal and vertical resistance in plants. Euphytica 26: 5-21. 7. Resistentie verkregen als gevlg van transgressie in nakmelingschappen van vatbare uders kan berusten p additief werkende 'minr' genen, zals gesuggereerd is.a. dr Krupinsky en Sharp (1979), f p 'majr' genen waarvan de expressie beïnvled wrdt dr de genetische achtergrnd. In het kader van het verkrijgen van duurzame resistentie is nderzek naar beide mgelijkheden gewenst. Krupinsky, J.M. & E.L. Sharp, 1979. Reselectin fr imprved resistance f wheat t stripe rust. Phytpathlgy 69: 400-404. Dit prefschrift. BIBLIO i,i KH IJKK LAMDBOU"»K<.n,I- SCHOOL VAGÉmBiGEH

8. De schattingen dr Skvmand et al. (1978) vande 'heritability en van het aantal genen betrkken bij de 'slw-rusting' eigenschap vr zwarte rest bij tarwe zijn njuist enmisleidend. Skvmand, B., R.D. Wilcxsn,B.L.Shearer &R.E.Stucker,1978. Inheritancefslwrustingtstemrustin wheat. Euphytica27: 95-107. Vegetatievevermeerdering 'in vitr', antherencultuur en smatische hybridisatie van prtplasten biedenperspectief bij de kffieveredeling. 10. Het prichten van een internatinaal nderzekcentrum vr verblijvende trpische gewassenverdient ernstige verweging. 11. Het 'tnisch' effect van bespuitingen met fungiciden, zals aangetnd bij dekffieteelt in Kenya nder andere dr Van dervssen (1982), dientnader nderzcht te wrden inde kffieprducerende landen van Amerika. VanderVssen, H.A.M.,1982. Cnsequencesfphyttniceffectsf fungicidetbreedingfrdiseaseresistance, yieldandqualityin CffeaarabicaL.,J.Hrt. Sei. 57: 321-329. 12. Heteffect van ziekte p pbrengst is redelijkbekend, terwijl ermtrent het effect vanpbrengst p ziekte veel te weinig bekend is. 13. Het weglatenvande dertiende verdieping in Amerikaanse flatgebuwen vergrt de instabiliteit vandeveertiende verdieping. 14. Het pneren van een stelling is belangrijker dande juistheid ervan. 15. Er iséénuitzndering p de regel 'geenregel znderuitzndering',namelijk deze regel zelf,hetgeen de algemenetepasbaarheid vandeze regel bevestigt. Stellingenbij het prefschrift van A.B. Eskes,getiteld 'Incmplete resistance t cffee leaf rust (Hemileia vastatrix)' te verdedigen p 25 maart1983 in de Aula vande Landbuwhgeschl, Wageningen.

CONTENTS page Intrductin 1 Chapters: 4-122 1 Assessmentmethds frresistance t cffee leaf rust (Hemileia vastatrix Berk. & Br.). 4 A.B.Eskes & M. Tma-Braghini. PI. Prt. Bull. FAO 29:56-66 2 The use f leaf disk inculatins in assessing resistance t cffee leaf rust (Hemileia vastatrix). 15 A.B. Eskes. Neth.J. PI. Path. 88:127-141 3 The effect f light intensity n incmplete resistance f cffee t Hemileia vastatrix. 33 A.B. Eskes. Neth.J. PI. Path. 88: 191-202 4 The effect f leaf age n incmplete resistance f cffee t Hemileia vastatrix. 49 A.B. Eskes & M. Tma-Braghini. Neth.J. PI. Path. 88: 215-226 5 Qualitative and quantitative variatin in pathgenicity f races f cffee leaf rust (Hemileia vastatrix) detected in the State f Sa Paul, Brazil. 63 A.B. Eskes. Neth.J. PI. Path. 89:In press. 6 Variatin fr incmplete resistance t Hemileia vastatrix in Cffea arabica. 82 A.B.Eskes & A. Carvalh. Euphytica 32:In press.

Chapters: page 7 Characterizatin f incmplete resistance t Hemileia vastatrix in Cffea canephra cv. Kuillu. 98 A.B.Eskes. Euphytica 32:In press. 8 Characterizatin f incmplete resistance t Hemileia vastatrix in the Icatu cffee ppulatin. Ill A.B. Eskes & W.M. da Csta. Euphytica 32:In press. General discussin 123 References 133 Samenvatting 135 Resum 138

INTRODUCTION The twcffee species fecnmic imprtance are Cffea arabica L. and C. canephra Pierre. The firstspecies isrespnsible fr abut 80 % fthewrld cffeeprductin (Carvalh et al., 1969). Itis analltetraplid autgamus species, grwing bestintrpicalhighlands.c. canephra is adiplid, allgamus species and grwsbest intrpical lwlands. Cffee leafrustrrange cffee rust (Hemileia vastatrix Berk, et Br.)is a majr disease fcffee. The rustarrived inbrazil in 1970 andhas recently spread t nearly allcffeeprducingcuntries insuthand CentralAmerica.Allcultivars (C. arabica) grwn in thesecuntries are susceptible trace II, thecmmn racef H. vastatrix. Yield lsses in Brazilhave beenestimated t be abut 30%, if n cntrlmeasures aretaken (Mnac, 1977). Fr Brazil, this wuld represent abut500 millin U.S.dllars annually in exprt revenues. The rust can becntrlled by chemicals. IntheState fsä Paul, Brazil, 4t 5applicatins peryear are recmmended (Maritt et al., 1979). Thecsts fthe chemicals needed fr full prtectin fthecffee crp in Brazil isyearly abut200millin U.S. dllars (Mnac, 1977), which represents 10t 20% fthettal prductin csts.chemical cntrl is nt prfitable frplantatins with lw yields. Only asmallpercentage fthebrazilian farmers fllws therecmmended spraying schedule. The abve bservatins indicate thatchemical cntrl isrelatively expensive andthatits value inthe cntrl fcffee leaf rustisrestricted. Thisemphasizesthe imprtance fgenetical cntrl fcffee leaf rust. Breeding frdisease resistance facesthe challenge f btaining durable resistance. Mstpathgens shwgreat variability fr pathgenicity, especially bitrphic pathgens such asrust fungi. Cffee leafrustthas shwn cnsiderable variatin in pathgenicity. Mre than30 physilgic raceshave been identified (Rdrigues, 1975). Fivemajr resistance genes present inc. arabica have nt given an efficient cntrl fcffee leafrust (Eskes, 1981 a ). Many ther majr genes are presentinc. canephra, whichhave nly been partly identified. Presently, in breeding frresistance t cffee leaf rust muchattentin isgiventthehybridppulatins 'Icatu' and 'Catimr',derived frm crssesbetweenc. arabica and C. canephra. Theseppulatins cntain gentypes with cmplete resistance and thers with varying levels f incmplete resistance. Durable resistance isf special imprtance fr perennial crps

like cffee. Thebreedingprcess isslwandcultivars whichhave lstresistance cannt berapidly replaced. Thebasis fdurable resistance has been a subjectf active scientific debate during the past 15years (Vanderplank, 1968and 1982, Simns, 1972, Rbinsn, 1976,Jhnsn andtaylr, 1976, Parlevliet and Zadks, 1977, Nelsn, 1978, Ellingbe, 1981). The histry f resistance breeding suggests that plygenically inherited incmplete resistance isgenerallymre durable thanhigh levels fresistance based n majr genes (Caldwell,1968, Simns, 1972). Vanderplank (1968) suggested thatdurable andtemprary resistance are based ntw distinct types fresistance named byhim 'hrizntal'and 'vertical' resistance. Hrizntal resistance was said t be race-nn-specific. Subsequent debate inthe literature has invlvedbasic questins such as : a) des race-nn-specific resistance ccur in nature, b)can plygenically inherited race-specific resistance be durable, c) candurable resistance be identified in a breeding prgramme, and d) can majr resistance genes be f value inbtaining durable resistance? Thebjective fthe presentresearch was tcharacterize incmplete resistance f cffeet H. vastatrix andt assess its pssiblevalue inbtaining durable resistance. The present research includes all types fresistance whichd nt fully inhibitthe reprductin f a certain strainfthe pathgen (incmplete resistance). The term incmpletewaspreferred t 'partial' because the latterhas beenused mre specifically t indicate incmplete resistancewith a high (susceptible) reactintype (Parlevliet, 1978 ). Theresearchhas resulted frm ajint prject f the 'Institut Agrnômicde Campinas' (I.A.C.), S.P., Brazil and the Fdand Agriculture Organizatin fthe UnitedNatins (F.A.O.), at the initiative f drs. A. Carvalh andl.c. Mnac (I.A.C.) and dr. R.A. Rbinsn (F.A.O.). The researchwas carried ut frm 1976t 1981 attheexperimental Statin fthe I.A.C.at Campinas,and receivedpartial supprt frm the 'InstitutBrasilier decafe (IBC)'. In1981/82 the Agricultural University f Wageningenhas supprted the research, facilitating theelabratin fthe results andthe preparatin fthe manuscripts. The authrhas received c-peratin frm several researchers mentined asc-authrs in fivefthe eight publicatins. Dr.

AleidesCarvalh, head fthe cffeebreedingprgramme atthe I.A.C., gave scientific stimulus andprvided datanprductivity fthecffeegentypes. Eng Agr a Waldir M. dacstahas assisted in the research nthe Icatu ppulatin. Mrs. MasakTma-Braghini served asanassistingbilgisttthe prject frm 1979 thrugh 1981. Acknwledgements frthec-peratin received by many thersare made atthe endf each chapter. Inadditin, Iliket thank Dr. ir.i. Bs and Mr.G.Heemstra fr helpwith the statistical elabratin ftheresults and Mr.J.S.deBlck fr crrectins made inthe manuscript. Als thanks gt Riet Hilhrst andwalter de Millian, whsemre general supprtisgreatly appreciated. (Frreferences see page 132)

1 ASSESSMENT METHODS FOR RESISTANCE TO COFFEE LEAF RUST (HEMILEIA VASTATRIX BERK. &BR.) Summary Assessment scales are prpsed fr quantitative recrding f reactin type and infectin frequency f cffee leaf rust. AH scales run frm 0 t 9. The scale fr reactin type includes hetergeneus reactins that can be frequently fund in the derivatives f Cffea canephra. The scales designed fr infectin frequency may help the cffee breeder in making a quick assessment f partial resistance t cffee leaf rust. The advantages f a labratry methd, using leaf disc inculatins fr resistance assessment, are discussed. The use f genetic resistance t cffee leaf rust cnstitutes the ideal cntrl methd f this very imprtant cffee disease. One f the basic prblems fr resistance breeding is that new races f the fungus, capable f vercming resistance, may appear. This type f resistance is called vertical resistance, accrding t Van der Plank (1968), and its durability is ften related t the extent f cmmercial use f the resistant cultivar. Prblems with vertical resistance als exist in relatin t cffee leaf rust. The five traditinally used resistance genes f Cffea arabica have nt prvided efficient cntrl when used in cultivars. Hemiteia vastatrix pssesses cnsiderable variability fr virulence (Rdrigues, Jr, Bettencurt and Rij, 1975). This was demnstrated in Brazil, where, within 10 years f rust presence, 10 different races culd be detected in breeding plts f the Institut Agrnmic f Campinas (Eskes, 1981). Therefre, the search fr a mre durable type f resistance fr the perennial cffee crp shuld be given high pririty. MR ESKES has been with FAO and is at present with the Institute fr Plant Breeding f the Agricultural University f Wageningen, the Netherlands. Ms TOMA- BRAGHINI is with the Genetics Department f the Institut Agrnmic (IAC), Campinas, Brazil. The authrs wish t acknwledge the encuragement and useful cmments given by Dr J.C. Zadks f the Agricultural University in Wageningen, the Netherlands. This article was prepared as part f the FAO/ IAC prject n durable resistance t cffee leaf rust under executin at the Institut Agrnmic f Campinas, SP, Brazil. Mre durable resistance may be btained by applying adequate breeding strategies fr vertical resistance, such as gene. deplyment, pyramiding f resistance genes in ne cultivar r using multiline cultivars (Nelsn, 1973). Hwever, mst f these strategies are difficult t implement, especially when applied t perennial crps. An alternative may be the search fr hrizntal resistance, which by definitin shuld be equally efficient against all races f the pathgen (Rbinsn, 1976; Van der Plank, 1968). The cncept f this type f resistance is still cntrversial and it cntinues t be widely discussed in wrld literature (Nelsn, 1973; Parlevliet and Zadks, 1977; Van der Plank, 1978). Hwever, a general agreement between researchers seems t exist that plygenically inherited partial resistance cnfers mre durable prtectin t crp pathgens than the mngenic hypersensitive type f resistance (Parlevliet and Zadks, 1977; Parlevliet, 1979; Simns, 1972). Since 1976, FAO has been cllabrating with the Institut Agrnmic f Campinas, Brazil, t develp a prgramme n durable resistance t cffee leaf rust. One imprtant aspect f this prgramme was t develp an adequate methdlgy fr quantitative measurement f the level f resistance, including the establishment f quick assessment methds fr rutine resistance screening. In this paper, a series f illustrated assessment scales, frm 0 t 9, are presented. The 0 t 9 scale was prpsed by McNeal et ai, (1971) fr cereal rust research, and has been

widely adpted fr epidemilgical and resistance bservatins (Zadks and Schein, 1979). This scale has the advantage that the data are cmputable. A unifrm recrding system is necessary fr cmparing research data frm different lcatins r experiments. REACTION TYPE THE TRADITIONALLY USED SCALE FOR REACTION TYPE The Centre fr Cffee Leaf Rust Research, CIFC, in Oeiras, Prtugal, has adpted a recrding system fr disease reactin type that is similar t the traditinally used system fr cereal rusts (d'oliveira, 1954-57). This system includes the fllwing reactin types: /' (immune), fl (flecks), t (tumefactins), ; (necrsis), (chlrsis), I (rare sprulatin), 2 (small pustules), 3 (medium t large pustules), 4 (large pustules), and X fr the hetergeneus reactin. It has served t differentiate the mre than 30 rust races knwn and t chse 24 differential cffee gentypes (Rdrigues, Jr, Bettencurt and Rij, 1975). Often the classificatin is simplified int fur types: R (resistant), MR (mderately resistant), MS (mderately susceptible) and S (susceptible). The fllwing cmments n reactin types f cffee leaf rust shuld be brne in mind: a) Reactin type 2, with small-size pustules, is nly rarely fund in cffee (d'oliveira, 1954-57; persnal bservatins). It shuld preferably be cnsidered as an intermediate reactin type between / and 3. The 1 t 4 reactin types then present a cntinuus scale with an increasing rate f sprulatin in relatin t the diseased r chlrtic area. b) Reactin type t (tumefactin), which is frequently bserved in Rubiaceae, cnsists f swllen spngy tissue cells (Rij, 1972). In cffee, it is ften assciated with incmpatible cmbinatins and has been called a hypersensitive reactin type. The authrs have bserved, hwever, that tumefactins can be fund als in cmpatible cmbinatins. Recent studies in Prtugal cnfirm that the t reactin is nt a specific resistance reactin. It can be induced by heat treatment f cmpatible reactins (C.J. Rdrigues, Jr, persnal cmmunicatin) and by wash-uts f uredspres f a cmpatible race (Rdrigues, Jr, Rij and Medeirs, 1981). c) Whatever the reactin type, necrsis is rarely bserved in yung lesins, but it may ccur when the lesins are ageing. Envirnmental as well as genetic factrs may induce early necrsis. Therefre, necrsis des nt seem t be an entirely suitable criterin fr determining reactin types f cffee leaf rust. d) As necrsis and tumefactin are less apprpriate t indicate resistance, the basic criteria fr the reactin type f individual lesins shuld be the size f the disclured area (flecks, 0,0 + reactin types), and the intensity f sprulatin in prprtin t the extensin f the chlrtic area (/ t 4 reactin types). e) In derivatives f C. canephra, a mixture f susceptible and resistant reactin types is ften fund n the same leaf. This ccurs generally n plants with an intermediate resistance level. Results indicate that this reactin may nt be a result f variatin in the fungus, but is rather a respnse f incmplete resistance that can be affected by light intensity and leaf age (Eskes, Rivera and Da Csta, 1978; Eskes, Krn and Van de Weg, 1980; Eskes, 1981). Cnsequently, fr the descriptin f the resistance f a certain cffee gentype tward a certain rust gentype, it will be necessary t include hetergeneus reactin types in the assessment scale. ƒ) At CIFC, inculatins are generally made nly n yung tender leaves, as it is much mre difficult t prduce disease n lder leaves (d'oliveira, 1954-57). Wrk in Campinas shwed that sme types f resistance are better expressed in the yung r adult leaves, as ld leaves are mre susceptible (Eskes, Rivera and Da Csta, 1978; Eskes, Krn and Van de Weg, 1980). Fr a cmplete appreciatin f the resistance f a plant it wuld be necessary t bserve the reactin type f leaves f varying ages.

A 0 TO 9 SCALE FOR REACTION TYPES OF COFFEE LEAF RUST The adptin f a 0 t 9 scale fr reactin type is prpsed, fllwing the internatinal effrts n unifrm reprting f research data (McNeal et al., 1971). The prpsed scale reflects the afrementined bservatins n reactin types. The scale is made t be used fr resistance determinatins f individual leaves r f entire plants. It includes a wide range fr hetergeneus types f resistance, in view f the frequent ccurrence f these types, especially in interspecific hybrids. A descriptin f the prpsed 0 t 9 scale is given in the table belw. Index value CIFC classificatin fr individual lesins fl.-tfl, t, fi, t,, + 4 fi, t,, 1 ft, t, -2 fi, t, -3 Descriptin f reactin types fr leaves r entire plants Immunity, n visible reactin. Minute "chlrtic spts" ften assciated with small tumefactins. Smetimes nly visible with a hand lens r when hlding the leaf against the light. Larger chlrtic spts, ften assciated with tumefactins. N uredspre prductin. A mixture f varius sizes f chlrtic spts, including very large chlrtic areas. Fewer tumefactins. N uredspre prductin. A mixture f chlrtic spts f varius sizes, with sme uredspre frmatin n large and chlrtic lesins. Sprulatin f less than 25 percent f all lesins. Few tumefactins may ccur. Early necrsis f lesins is smetitr.es bserved. As in 4, but with mre uredspre frmatin. Sprulatin f less than 50 percent f all lesins. As in 5, but with increased uredspre prductin. Sprulatin f less than 75 percent f all lesins. 7 fl, I, -< 9 4 t, 2-4 As in 6, but with abundant uredspre prductin. Sprulatin f up t 95 percent f all lesins. A mixture f lesins with a varying degree f sprulatin, smetimes assciated with a few tumefactins. Only lesins with abundant sprulatin, withut marked chlrsis at the lesin brder. Reactin types 4 t 7 indicate intermediate types in which increasing sprulatin intensity per lesin is generally assciated with an increasing prprtin f sprulating lesins. When n cmputing f the data is needed, the ntatin can be simplified by using the symbls R, MR, MS and 5 t indicate the fllwing reactin types: R = MR = MS = S = 1, 2 and 3 4 and 5 6 and 7 8 and 9 ASSESSMENT OF RESISTANCE DETERMINATION OF THE COMPONENTS OF PARTIAL RESISTANCE All cmpnents f partial resistance behave cumulatively t decrease the reprductive capacity f the pathgen. The individual cmpnents, hwever, may have different effects n the epidemilgical develpment f the disease in the field (Parlevliet, 1979; Parlevliet et ai, 1980; Zadks and Schein, 1979). Cmpnent analysis is therefre helpful in determining efficient selectin criteria fr practical breeding wrk (Parlevliet, 1976; Zadks and Schein, 1979). It is als imprtant in achieving a better understanding f the underlying resistance mechanisms. In inculatin experiments with cffee leaf rust, significant variatin has been fund fr all cmpnents f partial resistance, i.e., infectin frequency, latency

perid, sprulatin intensity, and duratin f sprulatin r sprulatin perid. Infectin frequency is the number f lesins btained per unit f leaf area. A unifrm inculatin methd shuld be used t determine genetic differences fr infectin frequency. The inculatin methd f CIFC, by which dry uredspres are applied first and the leaf is wetted afterwards (d'oliveira, 1954-57), is nt suitable t measure differences in infectin frequency, as the quantity f spres applied is nt cntrlled. A quantitative inculatin methd, using unifrm spre suspensins in water, is t be preferred. Latency perid is denned here as the number f days frm inculatin until 50 percent f the lesins sprulate. If all the lesins d nt sprulate then the latency perid shuld be calculated in relatin t the final number f sprulating lesins, and nt in relatin t the ttal number f all visible infectin sites. Sprulatin intensity is the number f spres prduced per infectin unit r per leaf area unit. Sprulatin perid is the time interval frm cmpletin f the latency perid until cessatin f sprulatin (necrsis f lesins r leaf shedding). THE USE OF ASSESSMENT SCALES FOR PARTIAL RESISTANCE Observatins n the individual cmpnents f partial resistance are time cnsuming and wuld nt be practical fr rutine resistance screening. Therefre, assessment scales are prpsed that serve the breeder in carrying ut bservatins mre quickly. All scales are made fr determinatins n disease incidence (number f lesins per leaf area unit) r disease severity (prprtin f diseased leaf tissue). These scales will measure the cmbined actin f all resistance cmpnents when applied t plycyclic experiments with successive natural disease cycles in the field r nursery. When applied t mncyclic inculatin experiments, the scales will nly assess the infectin frequency. If infectin frequency is well crrelated t the ther cmpnents f resistance, the prpsed assessment scales may als be applied t measure the verall resistance level in mncyclic experiments. In specific inculatin experiments, it may be f interest t determine the individual cmpnents f resistance separately. N assessment scale can be designed fr measuring latency perid. Instead f actually cunting all sprulating lesins ne by ne, the bservatins may be simplified by estimating the day n which 50 percent f the apparently "susceptible type" lesins that will prbably sprulate actually d s. Aleaf r whle plant is used as a reading unit. Sprulatin intensity may be assessed by the scale fr reactin type, if large differences in sprulatin are present. If smaller differences in sprulatin intensity need t be bserved, spre cllecting and cunting shuld be carried ut. The duratin f sprulatin is determined by leaf shedding r lesin necrsis. Necrsis can be induced by the envirnment (extreme temperatures, lw relative humidity), by the cffee gentype r simply by ageing f the lesins. In certain C. arabica cffee types, leaf retentin after inculatin may be a very imprtant factr in determining field resistance (Eskes et al., 1979). This parameter can be assessed in artificial inculatin experiments by applying an arbitrary scale running frm 0 t 9 (see sectin n greenhuse and nursery bservatins). FIELD OBSERVATIONS Epidemilgical parameters. Incidence f cffee leaf rust has been generally estimated by the fllwing parameters: 1 = number f lesins per leaf 2 = percentage f diseased leaves 3 = number f lesins per diseased leaf. These parameters crrelate well. Observatins n 100 C. arabica cv. Catuai plants shwed the fllwing crrelatin cefficients (/ values), all significant at P = 0.01. lgi (parameter 1) parameter 2 parameter 2 parameter 3.83.56

These data indicate that the number f lesins per leaf is the mst reliable parameter fr measuring disease incidence in the field (see Fig. 1). y = 0.30xr 2-0.87 0 1 2 3 4 5 6 7 8 9 Scre f Scale I Figure 1. Crrelatin f the lgarithm f real disease incidence (n. f lesins per leaf) with the disease assessment values f Scale 1 *The lg number f lesinsper leaf (independantvariable) shuldhave been pltted n theadsissa and the scresf scale 1 nthe rdinate. Thefrmula fr linear regressin becmesthen y =3.33 X +5.53. Assessment Scales I and II. Fr measuring resistance f large field ppulatins, tw types f disease-assessment scales have been designed. Scale I (see Fig. 2) is based n the whle tree as a reading unit, whereas fr Scale II a branch has been taken as a reading unit (see Fig 3). Bth scales have been designed t give cmparable results. Scale I is especially suited fr Brazilian grwing cnditins, in which the epidemic builds up frm the lwer part f the tree. Each pint n the 0 t 9 scale shuld indicate equal differences in resistance. The relatinship between the scale pints and actual disease incidence shuld fllw the lgit [In (xjl x)\ curve (Van der Plank, 1963, 1968; Zadks and Figure 2. Disease assessment Scale I: field incidence f cffee leaf rust

Schein, 1979). This curve crrespnds t the nrmal expnential curve fr lw values f x. As the breeder is especially interested in the lwer part f the epidemilgical curve, the number f dts chsen fr each scale value fllw a 2 n expnential curve. Fr scale value 9, nly 192 dts were placed instead f 256, as a massive leaf fall is nrmally induced, which limits the expnential increase f the lesin number per leaf, at this level f disease incidence. When using the scales, nly the quantity f sprulating lesins shuld be cnsidered. The bservatin can be made either n ne side f the cffee shrub r by tw readers simultaneusly n bth sides. In Brazil, the sutheastern side f the cffee shrub nrmally shws a higher degree f attack than the nrthwestern side, prbably due t differences in micrclimate. Descriptin f Scale I. The bservatin unit f this field scale is the entire cffee shrub. Each dt n the drawing represents the number f diseased branches seen n the shrub at first lk (see Fig. 2). Mre detailed bservatins are nly made t cnfirm the 0 and 1 readings. Scale value Descriptin 0 Absence f sprulating lesins; cnfirmatin necessary by pushing aside branches f the lwer canpy and mre detailed bservatin f the branches inside the shrub. / Presence f ne diseased branch at first lk n any side f the shrub. Pushing aside f lwer canpy branches may be necessary fr cnfirmatin. On detailed bservatin, mre than ne diseased branch may be fund. 2 t 8 These values represent a gradual increase in the number f diseased branches per tree, seen at first lk. Value 7 crrespnds with an average attack n susceptible cffee cultivars in Brazil at harvest time. 9 This scale value indicates the maximum f disease incidence, crrespnding t the level f disease f the very susceptible Harar cffee type in Brazil at harvest time. Nrmally, susceptible cultivars may drp leaves befre reaching this level. Therefre, an S-value assciated with intensive leaf shedding shuld be scred as a 9. Descriptin f Scale II. The bservatin unit here is the average disease incidence per diseased branch (see Fig. 3). Readings shuld nt invlve very detailed bservatins. Each small dt represents a sprulating lesin. Bigger dts represent a calescence f tw lesins. As, in the field, lder leaves are generally mre severely infected than yunger nes, mre dts have been placed n the frmer. The number f lesins fr each scale value and the percentage f leaf area they ccupy are given belw. 0 1 2 3 4 5 6 7 8 9 N. f lesins 0 1 2 4 8 16 32 64 128 192 Descriptin Percentage f leaf area diseased 0 0.06 0.12 0.24 0.47 0.94 1.89 3.78 7.56 11.51 0 Absence f sprulating lesins. Mre detailed bservatin, by remval f lwer branches, needed fr cnfirmatin. 1 Average number f lesins per diseased branch is abut 1. 2-8 Increasing number f lesins per diseased branch. Value 7 is cmmnly fund fr nrmally susceptible cultivars in Brazil at harvest time. 9 This value indicates maximum disease incidence, ccurring frequently n the Harar cffee type in Brazil at harvest time. An 8- value that is assciated with intensive leaf shedding shuld be classified as a 9.

Experimental results. Scales I and II were tested during June 1981 n 75 cffee plants with a varying degree f disease intensity. Readings were made by tw trained bservers. The relatinships between the average scres, using Scale I, f the tw bservers were pltted against the lgarithm f the disease intensity as determined by detailed assessment (see Fig. D- Fr the individual bservers the r 2 values were 0.86 and 0.85. The degree f agreement between tw bservers was high, with equal scres btained in 67 percent f the bservatins, ne scale value difference in 30 percent and tw scale values differences in 3 percent. The readings using Scale II were very similar t thse using Scale I, with the fllwing parameters f regressin btained: y = 0.31* 1.69 and r 2 = 0.88, with x being the scre f Scale II and y the lgarithm f the number f lesins per leaf. Here als the degree f agreement between the bservers was high. In 64 percent f the bservatins, scring was equal. In 34 percent f the bservatins, ne scale value, and in 2 percent f the bservatins tw scale values difference were btained. Scale I is recmmended fr Brazilian cnditins. It is less time-cnsuming than Scale II. The rate f bservatins using the frmer is abut 200 readings per hur fr an experienced bserver. AN ASSESSMENT SCALE FOR GREENHOUSE AND NURSERY OBSERVATIONS Greenhuse and nursery bservatins are generally made n ne-year-ld cffee seedlings. The degree f natural infectin can be assessed by Scale II, if the drawing is held upside dwn. Anther scale has been designed t measure Figure 3. Disease assessment Scale II: field incidence f cffee leaf rust 10

infectin frequency btained in artificial inculatin experiments (see Fig. 4). Fr this scale, a deviatin frm the 2 n expnential relatinship had t be included, t crrect fr the verlapping f lesins. The crrectin was made by applying the fllwing frmula (Garwd, 1947; Justesen and Tammes, 1960): r = (1 slm)", where r = disease-free area, s = surface f ne lesin, m = surface f the leaf, and n = number f lesins per leaf. Furthermre, the leaf area vulnerable t infectin is cnsidered t be nly 90 percent f the ttal area. N lesins develp n the veins and few lesins at the tip and the brder f the leaf. Taking these assumptins int accunt, the fllwing number f lesins and percentage f diseased area have been calculated fr each scale value: 0 I 2 3 4 5 6 7 S 9 N. f lesins 0 1 2 4 8 15 29 52 84 118 Percentage f diseased area 0 0.6 1.3 2.5 5.2 9.7 18.5 33.1 54.1 75.6 Scale III will be useful when screening a large number f plants. When few bservatins are t be made, it will be mre accurate t cunt Figure 4. Disease assessment Scale III: infectin frequency f cffee leaf rust n individual leaves 11

the actual number f lesins per unit leaf area. In additin t assessments f infectin frequency, it may be f interest t bserve leaf retentin after inculatin. This parameter can be imprtant in determining field susceptibility in sme C. arabica cffee types (Eskes et al., 1979). An arbitrary 0 t 9 scale can be used, giving a 5 value fr the cmmercial cffee cultivar. Higher values wuld indicate a lnger leaf retentin, lwer values a shrter leaf retentin. Bth very high and lw values may nt be desirable. High values will give an inculum build-up and, thus, a high apparent susceptibility, whereas lw values may be crrelated with lw prductin capacity. LABORATORY INOCULATION TESTS Leaf disc inculatins. In the past, leaf disc inculatins have been spradically used t measure vertical resistance (Narasimhaswamy, Narayanan Nambiar and Sreenivasan, 1961; CIFC, persnal cmmunicatin). Recently, this methd has been extensively tested at the Institute Agrnmic f Campinas (Eskes, 1977, 1978; Eskes, Rivera and Da Csta, 1978; Eskes and Tma-Braghini, 1979). The inculatin methd used cnsists f applying a 0.025-ml drplet f a unifrm uredpre suspensin t the centre f cffee leaf discs 1.7 cm in diameter. The discs are incubated in the dark fr 24 h, after which the drplets are allwed t dry. The discs are then maintained at 100 percent humidity, a temperature f 22 ± 2 C and mderate light intensity. The methd has given reliable results fr different types f experiments, i.e., determinatin f vertical resistance and race differentiatin, determinatin f partial resistance f field plants, determinatin f partial resistance f cffee lines grwn in the nursery, determinatin f envirnmental effects n the expressin f resistance, and determinatin f the effect f leaf age n the expressin f resistance. The advantages f the leaf disc inculatin methd, cmpared with inculatins f intact plants, are as fllws: very little inculum and space are needed; the inculatin cnditins are cntrlled, allwing fr cmparisns between lcatins and different inculatin data; plants selected in the breeding prgramme that cme frm distant lcatins can be tested by taking leaves t a central testing labratry; resistance f field plants may be better determined in leaf discs than by field inculatins, as a randm sample f leaves can be used fr the inculatins ; hst/pathgen interactins can be studied very precisely by using discs f the same leaves fr each cmbinatin; and in ne experiment a large number f hst/pathgen cmbinatins can be studied under unifrm cnditins. Assessment scale fr disease severity n leaf discs. The bservatin unit fr this scale is a grup f 10 leaf discs. Scale IV (see Fig. 5) has been designed fr measuring disease severity, using the same mathematical principles as thse applied fr Scale III. The size f ne leaf disc is taken t be 2.54 cm 2, with the area f the infectin drplet being 0.78 cm 2 (1 cm in diameter). Each dt n the scale represents a diseased area f abut 3.5 mm 2. Observatins can be made nce r repeatedly during the develpment f the lesins. Readings may include ttal area f diseased tissue r ttal sprulating area. The assessment scale has given gd crrelatins with ther parameters fr disease severity. Fr the number f lesins per disc, the crrelatin cefficient was r = 0.86 (n = 84) and, fr the percentage f diseased discs, r = 0.84 (n = 96). The degree f agreement between tw trained bservers was gd. N difference in readings was btained fr 68 percent f the bservatins, ne pint difference fr 30 percent and tw pints difference fr 2 percent. FACTORS AFFECTING THE RESISTANCE LEVEL Partial resistance t cffee leaf rust may be affected by envirnmental as well as physilgical factrs. Experiments have shwn a marked effect n resistance f prductivity, leaf age, and light intensity (Eskes, 1977, 1978; Eskes, Rivera and Da Csta, 1978; Eskes and 12

Tma-Braghini, 1979; Eskes et al, 1979; Eskes, Krn and Van de Weg, 1980). These factrs cmbined may prduce substantial seasnal differences in resistance, and will affect the efficiency f selectin fr partial resistance t cffee leaf rust. The level f natural disease in the field is the final test fr partial resistance. Hwever, bservatins shuld be repeated ver several years, taking int accunt the biannual bearing habit f the cffee tree. Plants will shw a higher degree f attack in a prductin year than in a nn-prductin year. Selected plants shuld have a gd yield and a lw disease intensity. In the C. canephra cv. Kuillu and in derivatives f the C. canephra x C. arabica crss, a marked effect f leaf age n resistance was fund fr several plants (Eskes, Rivera and Da Csta, 1978; Eskes, 1981). Generally, ld leaves becme mre susceptible t cffee leaf rust. Inculatins using yung leaves nly may nt be sufficient t indicate the real level f resistance f the plants. Natural disease ccurring in the nursery may be very valuable in determining the resistance f cffee ppulatins. Hwever, plants shuld be placed under mderate shade levels. Full sunlight may increase susceptibility t cffee leaf rust, but it may als kill the yung lesins by verheating f the leaf (Eskes, Krn and Van de Weg, 1980). The envirnmental effect n disease severity is cnsiderable in greenhuse inculatin experiments. The cefficient f variatin f a Figure 5. Disease assessment Scale IV: disease severity f cffee leaf rust n leaf discs 13

genetically hmgeneus ppulatin f C. arabica plants may be ver 100 percent. This variatin can be reduced by predispsing the cffee seedlings t similar light cnditins and by chsing seedlings f similar height. Hwever, even then, the efficiency f individual plant selectin fr disease severity is cnsidered lw in greenhuse r nursery inculatin tests. References D'OUVEIRA, B. As ferrugens d cafeeir. Centr de 1954-57 Investigaçadas Ferrugensd Cafeeir,Oeiras, Prtugal. Revista d Café Prtugués, Separata n. 3. 61 p. ESKES, A.B. Us de discs de flhas para avaliar a 1977 resistencia d cafeeir a Hemileia vastatrix: efeit da luminsidade e cncentraçà de inócul. (Use f leaf discs t determine resistance f cffee t Hemileia vastatrix: effect f light intensity and inculum cncentratin.) Prceedings V Brazilian Cffee Research Cngress, Guarapari, ES, Brazil, Octber 1977, p. 85-86. ESKES, A.B. Avaliaçà d us de discs de flhas 1978 para medir nivel de resistencia d cafeeir a Hemileia vastatrix. Prceedings VI Brazilian Cffee Research Cngress, Ribeirä Pret, SP, Brazil, Octber 1978, p. 114-115. ESKES, A.B., RIVERA, F.A. & DA COSTA, W.M. Efeit 1978 da idade da flha e luminsidade n camp sbre a resistencia d cafeeir a H. vastatrix medid em discs de flhas. (Effect f leaf age and light intensity in the field n the level f resistance t H. vastatrix, measured in leaf discs.) Prceedings XXX Annual Meeting f the Brazilian Sciety fr the Prgress f Science, Sä Paul, Brazil, July 1978, p. 12. ESKES, A.B. & TOMA-BRAGHINI, M. USO de discs de 1979 flha d cafeeir para estimar nlvel de resistencia a Hemileia vastatrix. (Use f cffee leaf discs t estimate the level f resistance t Hemileia vastatrix.) Prceedings VII Brazilian Cffee Research Cngress, Araxd, MG, Brazil, December 1979, p. 128-129. ESKES, A.B., TOMA-BRAGHINI, M., KROON, K. & HOOG- 1979 STRATEN, J. Parametrs para medir grau de suscetibilidade a Hemileia vastatrix em cultivares e ppulaçôes de Cffea arabica. (Parameters t measure the level f susceptibility t Hemileia vastatrix in C. arabica cultivars and ppulatins.) Prceedings VII Brazilian Cffee Research Cngress, Araxd, MG, Brazil, December 1979, p. 70-71. ESKES, A.B., KROON, C. & VAN DE WEG, W.E. Efeit 1980 da luminsidade sbre nivel de resistencia a H. vastatrix. (Effect f light intensity n the resistance t H. vastatrix.) Prceedings VIII Brazilian Cffee Research Cngress, Camps d Jrdä, SP, Brazil, Nvember 1980, p. 83-86. ESKES, A.B. Partial resistance f cffee t Hemileia 1981 vastatrix. Prceedings Internatinal Cnference n Trpical Crp Prtectin, Lyn, France, July 1981, p. 57. GARWOOD, F. The variance f verlap f gemetrical 1947 figures with reference t a bmbing prblem. Bimetrica, 34: 1-18. JUSTESEN, S.H. & TAMMES, P.M.L. Studies n yield 1960 lsses. I. The self limiting effect f injurius r cmpetitive rganisms n crp yield. T. PI. Ziekten, 66: 281-87. MCNEAL, F.H., KONZAK, CF., SMITH, E.P., TATE, W.S. 1971 & RUSSELL, T.S. A unifrm system fr recrding and prcessing cereal research data. Washingtn, DC, USDA, Agricultural Research Service. Dcument ARS/M71/34-121, 42 p. NARASIMHASWAMY, R.L., NARAYANAN NAMBIAR, K.K. 1961 & SREENIVASAN, M.S. Reprt n wrk f testing races f leaf disease fungus n cffee selectin at Cffee Research Statin, Balehnnur. Indian Cffee, 25(11): 333-336. NELSON, R.R., ed. Breeding plants fr disease resistance 1973 cncepts and applicatins. Pennsylvania State University Press. 401 p. PARLEVLIET, J.E. Screening fr partial resistance in- 1976 barley t Puccinia hrdei Otth. Prceedings 4th Eurpean and Mediterranean Cereal Rusts Cnference, Interlaken, Switzerland, 1976, p. 153-155. PARLEVLIET, J.E. & ZADOKS, J.C. The integrated cn- 1977 cept f diseases resistance, a new view including hrizntal and vertical resistance in plants. Euphytica, 26: 5-21. PARLEVLIET, J.E. Cmpnents f resistance that reduce 1979 the rate f epidemic develpment. Annual Rev. Phytpathlgy, 17: 203-222. PARLEVLIET, J.E., LINDHOUT, W.H., VAN OMMEREN, A. 1980 & KUIPER, H.J. Level f partial resistance t leaf rust, Püccinia hrdei, and hw t select fr it. Euphytica, 29: 1-8. RIJO, L. Histpathlgy f the hypersensitive reactin 1972 T (tumefactin) induced n Cffea spp. by Hemileia vastatrix Berk. & Br. Agrnmia Lusitana, 33: 427-431. ROBINSON, R.A. Plantpathsystems. Berlin, Heidelberg, 1976 New Yrk, Springer Verlag. 184 p. RODRIGUES, JR, C.J., BETTENCOURT, A.J. & RIJO, L. 1975 Races f the pathgen and resistance t cffee rust. Annual Rev. Phytpathlgy, 13: 49-70. RODRIGUES, JR, C.J., RIJO, L. & MEDEIROS, E. Induc- 1981 tin f flecks and tumefactins n cffee leaves by washuts f cffee rust uredspres. Prceedings Internatinal Cnference n Trpical Crp Prtectin, Lyn, France, July 1981, p. 58. SIMONS, M.D. Plygenic resistance t plant diseases 1972 and its use in breeding resistant cultivars. J.env. Qual., 1: 232-239. VAN DER PLANK, J.E. Plant diseases: epidemics and 1963 cntrl. New Yrk, Lndn, Academic Press. 349 p. VAN DER PLANK, J.E. Disease resistance in plants. 1968 New Yrk, Lndn, Academic Press. 206 p. VAN DER PLANK, J.E. Genetic and mlecular basis f 1978 plant pathgenesis. Berlin, Heidelberg, New Yrk, Springer Verlag. 167 p. ZADOKS, J.C. & SCHEIN, R.D. Epidemilgy and plant 1979 disease management. New Yrk, Oxfrd University Press. 429 p. 14

2 THE USE OF LEAF DISK INOCULATIONS IN ASSESSING RESISTANCE TO COFFEE LEAF RUST (HEMILEIA VASTATRIX) A.B.Eskes InstitutAgrnômic fcampinas, Campinas, SP, Brazil ABSTRACT The suitability finculatins fleafdisks f1.8 cm diameter in determining resistance fcffeet Hemileia vastatrix, the causal agentfcffee leafrust, was studied.results btained by this methdwere similar tthsebtainedby greenhuse testswith respecttreactin typesfcffeeplants with cmplete and/rmajr gene resistance. Theefficacy fthe methd inassessing incmplete resistance was tested n19 plants f Cffea canephra cv Kuillu, which varied inlevel fdisease inthe field.fur series f inculatin were carried utinfurdifferent mnths f the year,and sixcmpnents fresistancewere assessed. The analysis fmultiple crrelatin, applied tthe average data fthe fur series, indicated that 79% fthevariatin indisease inthe field culd be explained by thebservatins inthe leafdisk test.fr the individual series thispercentage varied frm 58t 70. The cefficients fcrrelatinbetween the six cmpnentswere significant and high. Thepercentage f leafdisks with sprulating lesins was fund t be the mst suitable cmpnent fr assessing incmplete resistance. Thenumber f lesinsper leafdisk was affected substantially by thehur ftheday f leafcllectin andby light intensity inthe field.n effect was bserved fthe sizefthe disks (1t 2cm in diameter)and fthe leaf wetnessperid after inculatin (24and 48 h). Results weremre cnsistent whenthe inculumwas applied in drplets f0.025 ml than whenthe inculum was sprayed nt the disks. N gentype x treatment interactins were bserved frthe hurfleafcllectin, frthe size f the leafdisk, frthe inculatinmethd r frthe leaf wetness perid. Itiscncluded that the leafdiskmethd, ifadequately standardized, can be a very useful tl in breeding frcffee leafrust resistance and als in basic research nthe cffee - H. vastatrix relatinship. 15

Additinal keywrds: labratry tests, greenhuse tests, cmpnents fresistance, light intensity, leaf wetnessperid, uredispre density. INTRODUCTION Cffee leafrustrrange cffee rust, Hemileia vastatrix Berk, et Br., hasrecently invadedmany Suth andcentralamerican cuntries. All Cffea arabica cultivars grwn in America aresusceptible tthe cmmnrace II f H. vastatrix. Yield lsses in Brazil wingtthe disease have beenestimated at 30%, if n cntrlmeasures aretaken (Mnac, 1977). Breeding frresistance has beenhampered bythe appearance f new rust races. Within threeyears f rust presence in Brazil theresistance genes S l, S 2and S 4lsttheireffectivetl ri rl ness (Mnac, 1977). Bythe jint effrtfthe InstitutAgrnômic f Campinas, Brazil, and fthefd andagricultural Organizatin ftheunited Natins, a prgram was started in1976t investigate incmplete resistance tcffee rust, with the aimt btaindurable resistance. Thepresentstudyresulting frm this effrt was carried utatthe GeneticsDepartment fthe InstitutAgrnmic. Methdstdetermine incmplete resistance aretimeand space cnsuming and results mayvary accrding tthe envirnmental cnditins. Inculatins f leaf disks, perfrmed under cntrlledcnditins, may reducethese incnveniences. Detached plant parts have lng beenused tdetermine resistance t plant pathgens (Yarwd, 1946). Incmplete resistance t bitrphicpathgens has beendetermined with thismethd, reprtedly with success (Hdgsn, 1961; Umaerus and Lihnell, 1976; Vermaand Pétrie, 1978; Shain and Crnelius, 1979), althugh susceptibility ratingsmay behigher frdetached plant parts than fr intact plants (Atifand Wilcxsn, 1975; Zadks, 1963). Retardants f senescence such as benzimidazles have been applied tkeepthe tissues in agdstate. Inculatins fdetached leaves r leaf parts fcffee have ccassinally beenused t study the bilgy f H. vastatrix (Ward, 1882; Nutman and Rberts, 1963;Saccas and Charpentier, 1971)andt assess majr generesistance (Mayne,1932; B.d'Oliveira,persnal cmmunicatin; Narasimhswamy et al., 1961). Cffee leafdisks were incubated byplacing them n mist filter paper r by flating them n water.csta et al. (1978)fund that leafdisksculd be easier kept in agd state thanentire detached leaves. Cnservatin f 16

the disks, when placed n mist filter paper, was satisfactry fr mre than100days and nextra advantage f theuse f benzimidazle was bserved. The presentstudy was undertaken tevaluate the leafdiskmethd fr itsefficacy in determining cmplete as well as incmplete resistance t H. vastatrix. Als studies arereprted n factrs which may influence theresults fthe leafdisk test. MATERIALS ANDMETHOD Definitins. 'Cmplete resistance' iscnsidered here t be afrm fresistance that fully inhibits the reprductin fthe pathgen. 'Majrgeneresistance' results frm the actin f ner afew, generally dminant genes, which individually have a greateffectn resistance, but is ntnecessarily cmplete resistance. 'Incmplete resistance' isdefined here as afrmfresistance that allws fr atleastsme reprductin fthe pathgen. Cffee gentypes. Plants f C. arabica cv Mund Nv, C. canephra cvkuillu and fthe Icatu hybridppulatin were used.the 'Kuillu' field plantswere abut 45years ld andthe Icatu plants wereplanted inthe field in 1970.Icatucnsists f backcrsses f anartificial hybridbetween C. arabica (4 x = 44) and C. canephra (2 x = 22) with cultivars f C. arabica (Mnac, 1977). Mstfthe Icatu andkuilluplants used shwedmedium r lw levels f attack fcffee rustinthe field, in cmparisn tcv Mund Nv. Frtheevaluatin f majr generesistance thecffeedifferentialscarryingthe S l, S 2, S 3 and S 4 genes fthecffee Rusts Research Center (CIFC)in Oeiras, Prtugal, wereused, and als tw plants f Icatu andkuillu withcmplete race-specific resistance (Eskeset al., 1981). Rust material. Unless stated therwise,greenhuse r field islatesfrace II (v5) f H. vastatrix were used. These islates were maintained n living plants in thegreenhuse.spres were stred in therefrigeratr at 52% relative humidity.germinatin percentages ftheuredispres were checked befre eachexperiment by placing small quantities f uredispres n drplets f distilled water. Incubatin was dne indarkness at22 ± 2 Cand germinatin cunts were made after 14 hurs. Only spre batches with mre than 10% germinatin wereused in experiments. 17

Standard methd f leaf disk inculatin. Randmsamplesffullgrwn leaves were btained frm field,nursery rgreenhuse plants, theuse fld anddamaged leaves being avided. Leaves cllected between 08.00 and 10.00 AMwere temprarily stred in mistplastic bags.disks were cut with acrk brer f1.8 cmdiameter early in the afternn.thesewereplaced in plastic bxes fsize 48 x 64 x 10cm, with the upper leafsidedwn,n a plastic fam layer f 2cm thickness saturated with tapwater. Thediskswere kept mist by spraying distilled wateruntil inculatin time,attheend ftheafternn. Unifrm inculum was btained by suspending uredispres in distilled water and shaking r stirring during fivet ten minutes. Thereafter the suspensin was keptin mtin by a magnetic stirrer, t avid depsitin fthe spres.fr inculatin, samplescntaining10 mlwere taken frmthe suspensinwith a plyethylene micrpipette (Becktn and Dickinsnnr 5688 dispsable pipette). One drpletfabut0.025 mlwasplaced nteach disk. Due tthe mist surface fthediskthedrplets spread tcver anarea f abut0.5-0.7 cm 2. The cefficient f variatin frthenumber f uredispres perdrplet,asdetermined in 48 drplets, was 10.5 percent. Preparatin f the inculum and inculatinwas always carried ut under lw light intensities. Uredispre densities applied inthe experiments ncmplete and/ r majr gene resistancevaried frm 0.8 t1.2 mg ml", and inthe experiments nincmplete resistance frm0.1 t0.5 mg ml", accrdingtthegerminatinpercentage fthe spre batch.the number f spres presentin ne mg is abut1.5 x 10 5. After inculatin the bxeswereclsed with glass lids and incubated inthe dark, at22 ± 2 C, during20 hurs. Thenthelids were remved inrder t allwthe inculum drplets tevaprate. In anair-cnditined rm this was cmpleted within 3t 5 hurs. Carewastakentavid dryingutfthe disks. After evapratin fthedrplets,thedisks were slightlywetted again andthe clsed bxes were placedundermderate lightcnditins (flurescent r indirectday lightf500-1000 lux) with daily a12 hur dark perid. During incubatin the waterlevel in theplastic bxes was checkedweekly and thedisksmistened slightly. It was avided that theedgewuld cmeint cntact withliquid water. Temperature was ntcntrlled inthe experiments f which theresults are shwn in Table 2, 3, 5and7. In the therexperiments temperature was maintained at 22 ± 2 C by theusef anair-cnditiner. This imprved lngevity f thedisks andwas adpted as a standard, althugh ade- 18