, I é V ,IJ J Z0f17 ~(\....., ,........,.C"":' l=~, ORGA IZATION AND MANAGEMErn OF crATIS Ac'i¿IE 11" .~ COLECCION HISTORICA 1 f1:iOIU30RT OF NATlONAL BEAN H-1PROVEt~ENT PPOGRA~ - - tJl~IDAD D: 1 ro ,fMLION y IN EASTERN AFRICA l DOCUMENIACIOH Roger A Klrkbi Bean ProductlOn and Research Opportunltles 8iBLfOTECA Beans (Phaseolus vulgarls L ) were lntroduced lnto Aírlca from the Latln Amerlca gene center by Vlest European traders over t:1J last centunes Currently Afrlca 15 the second most 1mportant common bean-prOOUCltlg re010n of the tr0~1($, fol1oWl ng Latm Amen ca The total average annua 1 Afncan prodactlOn, accordl ng to FAO productl0n statlstlcs, amounts to 1 4 ml11lon t per year over the last decade Productlon estlmates vary greatly, as was documented In a workshop ln Malaw1 on the Potentlal for Fleld Beans ln Eastern Afrlca (CIAT ,1981) A large part of the total bean productlOn 15 consumed 10ca11y and lS thus not captured ln the FAO estlmates shown ln Table 1 For example, Kenya and Uganoa report productlon of 467,000 t and 300,000 t, respectlvely vs FAO estlmates of 161,000 t and 175,000 t respectlvely Total bean product1on In Afrlca has lncreased ayer the last decade However, thlS has been achleved through area lncreases, whlle productlvlty has been stagnant and 1S currently at around 500 Kg/ha Productlon lncreases have not kept up wlth populatlon growth rates Per caplta consumptlon 15 therefore fall­ lng, and prlce lncreases have been aboye normal lnflatlon rates ln most countrles Beans play d cr1tlcal role In the human nutntlon of Eastern Afnca, provldlng u;J to 45% of total proteln consumptlon In Burundl and Rwanda (the h1ghest ln the world) and over 10% of proteln consumed In Kenya and Uganda Beans contrlbute nearly as much proteln to average natlonal dlets as al1 amma1 products comblned In Malawl and Uganda and far more ln Burundl and Rwanda Moreover, because beans are cheaper than anlmal products, they are of even greater slgnlflcance ln the dlet of the poor, who are oovlous1y most vulnf>rahlP ta rJalmltntlOn ThlS appl1es especlally to those countnes where based dlets on cassava and banana are assoclated wlth serlOUS proteln deflclencles, such as Uganda Sr"all farmers are the pnnclpal producers of beans 10 most of Afnca The vast IMJorlty of beans are cultlvated In assoclated cropplng systems wlth malze, sorghum or bananas In Kenya for example, only 6% of total productlon 15 estFlat­ ed to be In monoculture M05t bean productlon 15 for subslstence, wlth less _ 1~ a thlrd of the output belng marketed Use of fertl1lZers, pest1cldes and fUi, cldes In bean productlon 15 rure (CIAT, 1981) However large dlfferences In management appl1ed by farmers can be fouod wlthln relatlVe1y short dlstances In EthlOpla, for example, small-seeded whlte beans are grown as a pnmary cash crop In the rlft valley, under a ralnfal1 reglne of 6uO to 800mm, ln a mlnlmal lnpUt monocropplng system - a s1r\gle ploughlng to lncorporate broadcilst seed, wlth no subsequent weedlng In a hlgher ralnfall area of the nearby Sldamo reglon, by Paper presented at Consultatlve Group MeetIng for Eastern and Central Afrlca RegIonal Research on Gra"In Legumes AddlS Ababa, Ethlopla, 8-10 December, 1986 2 Coordlnator and CroPI:l1ng $ystems Agronomlst, CIAr RegIonal Program on Beans ln Eastern Afrlca, P o Box 67, Debre Zelt, Ethl0pla /2 - 2 - Table 1 Afrlcan bean productlon ln the last decade Apparent annual per Annual area and caplta graln legume Annual prooucLlon ylelo lflcrease conSulllf.'L 1011 ('000 tons) 1962-1979 (kg/ caplta) Country 1966-68 1977-79 (%) 1977 -79 Pn nc 1 pa 1 Producers flreól. d Uganda 175 175 6 7 -2 3 29 3 Burundl 133 162 3 6 -O 9 44 O Kenya 133 161 21 O Rwanda 126 174 3 7 O 3 50 6 Tanzama 108 150 2 7 O 9 12 O Other Producers FthlOpla 68 13 -8 8 -o (. 21 2 Somalla 2 8 1 7 Angola 64 64 2 7 -2 1 10 5 South Afncil 50 75 -1 6 4 7 3 I Hadagascar 49 47 O 1 O 4 5 6 Cameroun 24 82 6 7 O 4 131 21mbabwe 23 25 O 4 1 7 5 2 Togo 20 16 1 4 -O 6 10 1 Others 5" 198 8 [) Total 1,027 1,350 3 8 -o 1 11 7 Source rAO Productlon Yearbook, vanous years contrast, 1arge-seeded red beans are grown tWlce per VP?r Tor Subslstence purposes, often lntercropped ln malze, and sown ln rows behlnd the ox plouqh and are weeded by hand subsequently In thlS example, recommendatlOns for ralslng Ylelds from the low-lnput monocropplng system have been developed but, except for change ln vanety, most appear to be unacceptable to farmers because thelr lmplementatlOn would requlre dlverslon of labor from other crops or actlvltles at peak perlods ln the crop calendar (Tllahun Mulatu, 1986) Very llttle research, on the other hand, has been conducted on the lmprovement of the subslstence lntercropplng system, although results aval1able from other countrles ln the reglon lndlcate some promlslng dlrectlons On-farm research lS needed to ldentlfy local prlorltles Dlseases and lnsects, low 5011 fertl1lty, and perlodlC water deflclts form the prlnclpal natural constralnts assoclated wlth the low average ylelds The Reglona1 Bean Workshop ln MalaWl and the 1983 Workshop of Afrlcan bean researchers held at CIAT ldentlf1ed anthracnose (Colletotnchum 11ndemuthlanum), bactenal b11ght J~hoJl1~ phaseoll), angular leaf spot J IsanopS1S gnseo1a), bean common mosalC vlrus, and rust (Uromyces phase_oJll as the most lmportant dlseases !3 - 3 - across countrleS The beanfly (Ophl0mya Spp ) 15 the prlnClpal lnsect problem Although sources of reslstance to the aboye problems have be en ldentlfled, they often occur ln materlals wlth graln types lack1ng consumer acceptance, or are 1n poorly adapted mater1als and are, therefore, not approprate for farmers' C1rcumstances For example, the most lmportant cult1var ln Eth10p18, Mexlcan-142, lS susceptlble to most of the aboye pathogens To reduce dlsease pressure farmers ln many countr les plant at suboptlmal dens1tles and accept a lower potentlal y1eld by plant1ng beans dangerously clase to the next dry season Beans ln Afrlca are mostly cansumed as mature beans, elther drled ar befare the seed drylng process has started Green pods are also lmportant, and young tender leaves are somet1mes consumed as a vegetable Large red, red-and brown-mottled or speckled seed types are preferred 1n many areas but seed color preferences seem less strlngeot thao ln Latln Amerlca Short cooklng tlme, however 15 very 1mportant where, as 10 Rwanda, the flrewood problem 1S acute Beans, belng slower to cook than most other common foodstuffs ln the dlet. largely determ1ne the amount of f1rewood used (erAT, 1986b) Taste lS also 11nportant 10 help1ng to determ1ne accept­ ance of a new va r1 ety, a lthough, here too, genera llZatl on 1 s d1 ffl cult a commerc 1d 1- 1y oon-preferred, sma 11 gra 1n ed varl ety was found to be popul a r W1 th the poores t sector of the rural populatlon of K1rlnyaga ln Kenya, because seed for plantlng was less expenS1ve (Franzel, 1982) There a re many s 1m11 a r1C1 es between the bean productlOn sys tems and productl0n problems 10 Afrlca and Lat1n Amerlca rn both regl0ns, beans are produced prlmdrl­ ly on small farms ln aSsoclat1on wlth malZe, and wlth llttle use of chem1cal 1nputs In both Afr1ca and Latln Amerlca, drought lS an lmportant l1mltlng factor 1n product- 10n, and many of the most lmportant d1seases constltute sev~re problerlls anthracrose, angular leaf spot, bean common mosalC Vlrus ([)Cr~V), rust and bactenal bJ1ght Prospects therefore appear promls1ng that lmportant Dean technology components may be transferable from Latll1 Amerlca to Afrlca Of course, there are also dlfferences between the Afr1ca and Latln Amerlca bean sltuatlons Halo bllght and necrotlc stra1ns of ªCMV are relatlvely more lmportant 1n Afnca than Latln Amenca, nelther the beanfly nor bean scab lS found 111 Latln Amenea SOClOecononllC productlon condltlons also d1ffer between the contlnents For example, ln Afrlca, a hlgher percentage of bean product10n lS used for home consumptl0n and the taste and coaklng quallt1es of local vanetles need to be malntalned 1n new materlal Consequently, transfer of flnlshed technology to Afrlca ean be, at best, a short­ term expedlence and no Substltute for the permanent stre11gthenlng of local research capab111 ty Natl0nal Bean Research Programs Bean research has a long hlstory ln Afr1ca and has contlnued to lncrease ln lmportanee over the past few years Many natlonal programs are backed by a reason­ able 1nfrastructure, although seed storage, transport for on-farm researeh, and fleld equlpment not avallable for purehase ln local currenCles are generally lnadequate for effle1ent deployment of the aval1able research manpower The adequacy of human resourees aV811able to these natlonal research programs var1es greatly Uganda, wlth an estlmated 450,000 hectares of beans grown ln several dlst1l1ct agroecolog1eal zones, has twelve graduate staff of the 1'lln1stry of Agrlculture devot1ng on average 70% of thelr research t1me to the bean crop Not a11 eountnes arf" <;0 I>/éll endowed In recen e years Jsrono",y n semeral aplJears LO have recelVed 1e ss a ttentl on than breed1 ng and crop protect lOn A1 1 cClI,hnes, however, place prlorlty on lmprov1ng the tralnlng of thelr research staff, ooth graduates and technlcal ass1stants These blo pnnc1pal categorles of staff requlre dlfferent types of tralnlng, and lt lS becom1ng apparenc that a concentratlon of tra1nlng opportunltles upon those at the qraduate level does not necassarl1y lead to a 14 - 4 - concomlmtant mprovement 1n the pract1cal Skl11s of tl1e1r aSslstants Exchange of research methodog1es, 11 terature and other results among the VdrlOUS natlonal and 1nternatlonal prograrns has been largely lacklng, unt1l recently A questl onna 1 re survey of a 11 known bean researchers 1n Afn ca, conducted 1n 1985, ldentlf1ed the crltical need for lmproved 1nformatlon and documentatlon access ln the reglon For example, none of the natlOnal sClentlsts who responded was able to subscr1be to a SC1entlflc Journal, for re~sons of costs and fore1gn exchange restrlctlons, and most 11brarles 1n the reglon are deflc1ent and deterloratlng One bnght spot was the emergence three years ago of the Phaseolus Beans Newslette~ for East_Afrlca, compl1ed and publ1shed by Kenya 's natl0nal program at Th1ka Th1S newsletter 15 attractlng contrlbutl0ns from throughout the reglon and lS already rece1ved by one thlrd of the reglon's researchers, accord1ng to a recent survey There was l1ttle exchange of bean germplasm witllln the regl0n unt11 recently, and the present range of var1etles appears qUlte llmlted Eth1opla, for example, has the w1dest altltudlnal range of agrlcultural envlronments ln Afrlca but does not produce beans aboye approxlmately 2000 masl, whereas Latln Amerlcan countrles at a Slm1lar lat1tude grow beans up to about 3200 masl. ObJectlves (lf the CIAT Bean Program ln Afrlca CIAT's actlvltles ln support of natlonal efforts 1n bean research have the follow­ ln9 three broad obJect1ves To lncrease the productlvlty and productlon of food beans by breedlng and selectlng hlgher Y1eldlng genotypes ldentlfled from among a more dlVerse germplasm base, both from lntroductlon and from locally adapted landraces Such cultlvars are 11kely to be selected for yleld stabl11ty, relylng on reslstance to bl0tlC and ablot1c stresses, and for consumer acceptabl11ty 2 To develop more productlve systems of cropplng, utll1z1ng promls1ng new cultlVars and vanecal mlxtures when approprlate, wh11e ensurlng that such lnnOvatlons of cropp1ng system and cu1tlVar remaln acceptable to producers and consumers and do not dlsrupt eXlstlng farnnng systems adversely 3 To asslst scfengch::mng 01' nac1O'1;:¡1 fesearch prcgrams, to a degree that 15 both appropr1ate aod susta1nable nat1onally, through glvlng substantlal ernphasls to tralmng Tralnlng 15 offered to postgraduate SClentlsts at un1versltles wlthln or outslde the reglon, e1ther wlth or wlthout perlods of research Wl th CIAT programs Short-tlffie tra 1n l ng 1s encouraged, Wl thl n Afrlca or outslde the reglOn On-thl:~Job tralnlng 10 ObJectlVe 1 and 2 lS lmportant OrgarllZatlOn of CIAr Bean Actlvltles 1nAfnca In a flrst meetlng of bean researchers ln Afnca, held 1n Lllongwe 1n ~larch, 1980, CIAl was asked by delegates from the chlef bean-produclng countr1es of Eastern Afrlca to mount a reglonal program 1n order to support natlonal bean research ln Afrlca (CIAT, 1981) After sorne years of ne~otlat1on, 1t became apparent that no slngle donor was prepared to support the entlre reglon of Eastern and Southern Afrlca CIAT therefore sought to establlsh proJects ln geographlcal subreglons, ldentlflable on agroecologlcal and/or econom1C grounds, for WhlCh fundlng could be found By 1983 CIAl was ln a pos1tlon to set up a regl0nal program for the Great Lakes countrles of Burundl, ,'Vianda and Za1re wlth SW1SS (SDC) support, and a ful1 reglOnal tea m 15 now ln operatlon from a coordlnatlng center ln Rwanda Towards the end of I~ - 5 - 1984, CIOA and USAID funds(through COA) beca me ava11able for establlshlng operatlons ln the rest of Eastern Afr1ca lncludlnQ Kenya, Uganda, Ethlop1a and Somal1a CIAT placed one bean SC1ent1st at thlka, Kenya 1n September 1984 to start up operat10ns The reg10nal coordlnat10n center for these countrles 1S no\~ establ1shed ln Ethlopla Further fund1ng from CIDA has enabled CIAT to start establlshlng a thlrd reg10nal base 1n Arusha, Tanzan1a to serve the SADCC countr1es, and a reg10nal coord1nator \~as posted there ln July, 1986 Each of the three reglOnal programs 15 located, by agreement wlth the respect1ve nat10nal research 1nst1tutlOn, vl1th a natlOnal bean 1mprovement program The Southern Afrlca program lS a J01nt program of CIAT w1th the Southern Afr1can Center for Cooperat1on 1n Agr1cultural Research (SACCAR) The probable pattern of staff1ng lS 1ndlcated ln Table 2 Table 2 Stafflng and Locatlon of CIAT ReglOna 1 Bean Programs Reglon Posltl0n LocatlOn Great Lakes Breeder/Coordlnator ISAR-Rubona" Rwanda Cropplng Systems Spec1al1st/ 11 !I .. Anthropologlst U 11 " Pathologlst n 11 " C¡Opp1ng Systems Agronom1st n 11 " Nutrl tlonl st 11 U .. Eastern Afrlca Cropplng Systems Agrononnst/ IAR-Nazret/Debre Ze1t, Eth10pla Coord~nator Breeder,Pathologlst (1) MAF-Kawanda, Uganda Agronomlst/Breeder (1) ti !I 11 Econom1 st ( 1) (2) Southern Afrl ca Pathologlst/Coordlnator TARO-Arusha, Tanzanla Breeder 11 11 11 CroPP1ng SY5tems Agronomlst(l) .. .. .. Entomolog1st (1) .. " .. Breeder ( 1 ) Bunda, t·la 1a Wl (1) Poslt10n to be f111ed durlng 1987 (2) Locat1on not yet f1xed One member of each reglonal program acts a150 as coord1nator, and all staff report to the coordlnator of CrAT's Bean Program ln Colombla The dlstrlbutlOn of reglonal staff 15 lntended to comb1ne elements of central1zatlon (conferr1ng advantages of easy 1nterdlsc1plary teamwork and a crltlcal mass) \~lth the advantages of decentral1za­ Clon (dally contact wlth a large number of natl0nal programs and agroecolog1cal zones, and smaller groups of expatrlates less llkely to dom1nate nat10nal program decls1ons) The decentrallzed model 1$ felt to be partlcularly approprlate to Eastern Afr1ca, where nat10nal programs are generally more developed than ln the other two reg10ns Ph.1].?~ophy of Reg10nal Program Oper:at1o~ CIAT's ph1losophy 1n operatlng thesk' programs 15 that the1r Ilr1nC1pal obJectlVe lS to strengthen nat10nal progranls ln such a way that they become fully effectlve, pract1cally onented lnterdlSClpllnary research teams that renlaln effectlVe after /6 - 6 - the wlthdrawal of external support The key to th15 lles, we bel1eve, ln worklng alongslde natlonal bean teams as colleagues wlth1n the research sltes where they work Care 15 taken that reglonal staff gUlde, encourage and supplement (but do not replace) the actlvltles of natlonal 5c1entlsts CIAT's reglonal programs usually do not run separate fleld trlals Instead, they make every effort to ~lve full support to natlonal teams ln conceptual1zatlon, plannlng and fleld executlon Jf the research for Wh1Ch each natlonal team retalns responslblllty and the credlt ,or ltS ach1evements Varletal releases and recommendatlOns of lmproved cultural practlces cnat emergefrom thlS research collaboratlon are made by the natl0nal IJrogram The foregolng "bllateral" component of a reglOnal program requlres careful recrult­ ment of lnternat10nal staff On the one hand, they must be SClentlsts of sufflClent callbre to earn the respect of thelr colleagues In natlOnal programs through thelr work together on a dal1y or frequent basls 00 the other hand, they must also be sympathetlc to the needs and asplratlons of the natlonal programs and sClentlsts, and need to be prepared to assess thelr own ach1evements 1n terms of the performance of nat10nal programs Concern wlth ach1ev1ng rapld research progress ln the short Lerm needs to be balanced by an equal concern for fosterlng long-term sustalnabl11ty of research capaclty As slmllar agroecologlcal zones tend to be found lO several nelghborlng countrles, natlonal programs can beneflt greatly from regIonal collaboratlon Sharlng of 1n forma tlOn and experl ences among cOL.ntrl es on prevlOus and present research actlvltles 15 an essent1al flrst step RegloDal tr1als and technlcal meetlngs have a last1ng effect ln ln1t1atlng personal contact among SClentlsts worklng on related problems Problem-speclflc workshops or monltorlng tours provlde opportunltles both for lndepth dlScusslons among these SClentlsts and for lnJectlng relevant external experlence through partlclpatlon of lnternatl0nal center staff or other speclallsts as consultant5 to the reglon Purposeful collaboratlon among natlonal programs ln solvlng one or more common research problems lS consldered by CIAT to be a further potpnclal of a reglonal progrGf, Not only are llmlted resources usen more efflclently through ,oncentratlon of effort by dlfferent natl0nal programs upon complementary aspects of a problem shared by them, but a150 the plannlng and analytlcal abllltles of natlonal program SClentlsts are enhanced through collaborat­ ¡ve plannlng seSSlons and peer group reVlew of research progress The role of the reglOnal program 10 these "network" actlvltles 15 twofold Flrstly, the program can catalyze collaboratlon among countrles so that thelr understandlng of shared problems and thelr rate of progress ln exploltlng research opportun1t1es are greater than would be llkely through nat10nal research conducted ln lsolatl0n Secondly, a reglonal program should have the techn1cal backJp to be able to feed lnto nat1ona1 programs the new germplasm, research methods and SClentlflc documentat­ lon that lS requlred and requested The full lnterdlsClpllnary nature of CIAT's Bean Program wor1dwlde, WhlCh lncludes economlsts and other 50c1a1 SClentlsts as \lell as the usual blologlcal dlsclpllnes and nutntlon, lS advantageous 15 fostenns dlsclpllnary lntergratlon ln natlOna1 programs r~naqemeot by Re1lonal Steerlng Commlttee Each reglonal pro~la~ 15 monltored by a 5teerlng cOfllWlttee that meets at least ene" p"r )ear(ln~ervJls of SlX to nlOe months ILve oeen faund 1I5t..,ul 1,' the early stages of a program) The comml ttee 1 s composed o r the nat lOna 1 bean resea rch coordlnators or team leauers and the reglOllal coordlnator Donor representatlOn lO an observer capaclty 15 common /7 - 7 - The general functlons of the steenng comnnttee are to gUlde CIAT ln ltS lmplement­ atlon of support funetlons and to set prl0rltles for the reglan Speelflc taples that requlre agreement wlthln the commlttee lnelude strategles and lmplementatlon plans for the follawlng actlvltles - Seleetlon of research prlorltles wlth reglonal appllcatlon - Reglona1 germp1asm movements, nurserles, etc - Reglonal tralnlng program - Organlzatlon of workshops and mon1torlng tours - Ident1flcatlon of reglonal needs for consu1tancy serVlces from eIAT, from wlthln the reglon or from e1sewhere - Annua1 work plan coverlng al1 the aboye (submltted ln draft by the reglonal coordlnator) - A1locatlon of flnanclal resources ~ihere dlscretlOn 15 provlded wlthln the budget, e 9 , for collaboratlve re5earch subproJects and for capltal equlpment for natlonal programs Chalrmans'l1p and venue of the meetlng rotates among countnes The chalrman serves untll the follow1ng meetlng and durlng that perlod may be consulted by the reglonal coordlnator on matters pertalolng cO ch2 reglonal prograr The reglonal coordlnator provldes secretanat serVlces to the steenng comollttee and represents the center Iotegratlon of Independentl Y:fjjnded Reg lOna 1 Programs CIAT encourages meetlngs to be small and lnformal because thlS creates en atmosphere that em.ourages commUnlCatlon among lndlVlduals and ln tlme bUllds the professl0nal trust upon WhlCh the establlshment of a re~10nal network depends In th1S respect, three separately funded reglonal grouplngs can be useful For example, an annual technlcal \/orkshop for a11 bean researchers 15 held lndependently wlthln each reglon Slmllarly, most tralnlng courses accommodate staff from a 51ngle eountry or reglon Nevertheless, the lntentlon femalns to 1ntegrate lIIany of thelr actlvltles lnto a Sl ngl e opera tlOn Methods al ready be1 ng used 1o c 1u de the fo 11 0~1l ng - Interchange of reglOnal sClentlsts although each reglOnal tealll lS multl­ dlsc1pllnary and 1ne1udes at least one agronomlst and one plant breeder, the needs of al1 the three programs are to be met by a slng1e economlst, anthropologlst and entomologlst, each located ln a dlfferent program Expenses lncurred by eaeh SClentlst ln asslstlng another regl0nal team are met by the beoefltLlng progralll - AttendanCf of nelghborlng reglonal and natlonal coordl0ators as observers at steerl0g eommlttee meetlngs reglonal coordlnators have a speclal role to play ln lnterreglonal communlcat1on, both 1n research plaoolng and 10 dlssemlnat10n of results At the more local level, commun1eatlon between researchers responslble for agroecologlcal zones dlssected by a reglonal boundary (e 9 nothern Rwanda and south-western Uganda) lS faCl11tated by the attendance of both natlonal coordlnators at thelr respectlve steerlng comm1ttee meetlngs - Exchange of germplasm The Afrlcan Bean Y1eld anu Adaptatlon Nursery (AFBYAN), assembled by the Great Lakes reglOnal breeder from the most promls1ng materlals avallable from each natl0nal program, 15 belng used both to glVe each country wlder access to germplasnl and to start assesslng systematlcally che agroecologlcal Varlatlon ln bean-growwg areas of Afrlca Rwandan and T;:¡IL~miln JJrcealn.9 maccnals have been selccisd by V1Sl Llng sClentlsLs frol'l Uganda for lncorporatlO>i lil chelr own program - E,chan~e V1S1tS and mon1torlng tours Already three Ugandan SClentlsts ;d - 8 - have made worklng V1SltS to the Rwanda natlona1 program, accompanled a1so by Great Lakes reglona1 staff A three-country monltorlng tour for SClentlsts from Rwanda, Tanzanla and Uganda may faCl1ltate co11aboratlon focussed upon thelr common lnterests ln the lmportant bean-produclng areas located at the Junctlon of the three reglons - Afrlca-Wlde Workshops Specla1lzed technlca1 workshops,for as c esslng the state-of-the-art and for deduclng approprlate research strategy,draw upon SClentlsts wlthout regard for thelr geographlca1 locat1on For examp1e, a bean f1y workshop recent1y drew logether breeders and entor1Olog1sts havlng re1evant experlence Slml1ar workshops may be useful ln the flelds of beao breed1ng, pathology ano agronomy An occas1ona1, broadly-based conference for bean researc~ers ln Afrlca lS a1so under conslderat1on Co11aboratlve Reglona1 Research A reglonal vanety tna1 of the AFBYAN type lS one ~¡ay ln WhlCh lnformatlon can be shared usefully across countrles that ha ve slmllar agroecolog1cal zones or cropplng systems nlches Undoubted1y the purpose and deslgn of these trla1s wl11 change as understandlng of the reglon lmproves CollatlOn, lnterpretatlon and feedback of resu1ts across countrles lS an lmportant functlon for the regl0nal programs There lS an01:her approach to lmprovlng the efflclency wlth WhlCh natlona1 resources are used for overcomlng researchable prob1ems that are shared by several countrles ThlS approach lnvolves the purposefu1 dlvlslon of effort among natlonal programs that choose to co1laborate The complex of wldespread bean dlseases has trlggered thlS form of collaboratlve research, favored by the opportunlty to select for genetlc res1stance ln the best hotspots avallable ln the reglan The follO\~lng collaboratlVe subproJects are currently ln progress among the Great Lakes countrles - Screenlng of gerr.lplJ.sl'1 for' reS1S\.ilnc:: co ascoci1yta (I:;A.3U, Burundl) - Screemng of germp1asm for reslstance to anthracnose (ISAR, Rwanda) - Screenlng of thlS germplasm for reslstance to angular leaf spot (~rogramne r~a t lOna 1 Leguml neu3e, Za 1r e) Eastern Afrlca has provlslnally agreed on the foll0l11ng asslgnment of prlonty research tOpl cs - Screenlng for rust reslstance (Ethlopla) - Screenlng for drought tolerance (Soma1la) - Screenlog for baccerla1 bl1ght and ascochyta reslstance (Uganda) Proposals for collaboratlve research proJects can arlse both from the settlng of reglona1 prlorltles by the steerlng commlttee and from lndependent Jppllcatlons submltted throug~ a natlonal coordlnator by lntercSéed SClentlsts of any research organlzatlon ln hlS country A slmpllfled appllcatlon form 15 aval1ab1e to facllltate the development and assessment of proposals RegIonal funds can be al10cated for successful proposals ln rpconnltlon of thelr reglonal res~onslbllltles The steerlng commlttee conslders each proposa1 on the basls of (1) re1evance of the expected research output to the regl0n as a who1e, and (2) progress made by the proposers (,n the case of renewals) Progress wlll be assessed from reports requlred of the researchers, from presentatlons at r~CJlona1 workshops, and from V1S1tS by one or more members of the steerlng commlttee /9 - 9 - Sean Informatlon SerY1CeS 10 the Reglon CIAT opera tes a Bean Informatlon Center at lts headquarters, utl11z1ng core fund- lng and speclal proJect funds from IDRC 10 addltl0n to publ1ShlOg Abstracts 00 Fleld Beaos, the center has compl1ed and dlstrlbuted three blbllographles on bean research 10 Afrlca (Lopez, 1983, CIAT, 1984, CIAT, 1986) The most recent volume lncludes "fugltlve" llterature obtalned by means of personal V1SltS to bean research­ er throughout the reglan by a coosultant A free monthly serVlce provldes researchers and llbrarles wlth current contents 11Sts for a wlde range of agrlcultural Journals The page charges for photocoples requested by researchers are met by coupons dlstrlbuted by reglooal and natlonal coordlnators Reglonal coordlnators also asslst 10 updatlng the dlstrlbutloo 11StS Results of research emanatlng from profeSSlonal collaboratlon between reglonal and natlonal program SClentlsts lS not reported by the reglonal sClentlst lndependently In the case of collatlon or extrapolatlon of research results across the regl0n by a reglonal sClentlst, results are sent to al1 natlonal bean programs and dlr~ctors of research ConclUS1On The reglonal actlvltles outllned aboye are llkely to be successful ln meetlng all three obJectlves only lf they are undertaken wlth determlnatlon and ln partnershlp The countrles of Southern Afrlca (1 e members of the Southern Afrlcan Development Coordlnatlon Conference) have the advantage of a comman forum for settlng and for lmplementlng thelf own broad reglonal prlOrltles 10 agncultural research However, the steerlng commlttee mechanlsm lS provlng ltS value ln other reglons also, by faCllltatlng collaboratlon among research lnstltutlons and sClentlsts to solve problems that are held to be natlonal prlorltles by several countrles By enabl­ lng lnternatlonal organlzatlons to respond more dlrectly to natlonal 1nterests, thlS mechan1sm a1so reduces the rlsk of paterna11sm The JUd1C10US provls10n of technlcal asslstance needs to be accompanled by equal attentlon to other forms of support to strengthen 10ngterm research capabl11ty Whl1e strategles belng encouraged by CIAT for lmprovement of bean Varletles and cropplng systems \varrant a separate descrlpbon elSellhcrE!, th" r"91Onal approach descrlbed here has lmpl1catlons a1so ln these areas For example, a recent Afrlcan Beanfly .~orkshop recommended tha t beanfly reS·l stance screenlOg work be coordl na ted ln the future from wlthln the reglon rather than from CIAT's headquarters A degree of decentrallzatlon of thlS nature lS lnherent ln the phllosophy Other lr.,portant components of the program 1 nc Juúe a 51.roog con,nnr.lent cO cra 101 ng by all reglonal staff, an lmaglnatlve approach to the selectlon of tralDlnQ modes and locat1Ons anc' thc: provlSlon of 110dest fund1~(1 to n~tlon~1 pro~r¡:ms j"or str'teglc purchase of 511pplleS ~n