Constraint #1: Insufficient Natural Resource Management and Production Technologies

Research area: Pest and disease control through genetic and agronomic means

Activity #3: Finish CBB co-evolution work

U.S. researchers: Gilbertson, post-doc research associate, Myers
HC researchers: Mabagala, Mkandawire
Methodology: Conduct survey/collection trip in Tanzania to complement CBB bacteria collection already made in Malawi. Leaves with CBB symptoms will be collected for isolation of CBB bacteria and genotyping of bean materials. The presence of brown-pigmented CBB bacteria and genotypes from which the bacteria were isolated will be correlated. Plant inoculations will also be conducted in order to further examine co-evolution.

Anticipated results of activity: Co-evolution study with CBB will be completed with determination made of whether a correlation exists between Mesoamerican materials and brown-pigmented CBB bacteria. This information will assist in the further identification of appropriate donors of CBB resistance for crosses outlined in Activity #4.

Anticipated impact(s)

• With a better understanding of host-pathogen interactions, varieties with long-term resistance may be deployed; The life of the small farmer would be improved by varieties with higher and stable yields.

Major Changes (Tanzania): Lack of Taq Polymerase enzyme in Tanzania delayed initialization of the PCR work on CBB. The previous stock had expired, thus efforts were made to order the material from Denmark using leveraged funds. Furthermore, due to severe drought in the country and the region in general, there were frequent power interruptions which severely interfered with the smooth running of the PCR activities. It is thus, suggested that a stand-by generator be acquired to back-up power supply for the PCR laboratory when power shortages occur. Due to power shortages, typing of CBB isolates by PCR was not completed.

Progress During Past Year (Tanzania): Bean and weed leaf samples containing typical symptoms of CBB and HB were collected from small-holder farmers' fields in various regions (Morogoro, Tanga, Kilimanjaro, Arusha, Mara) in Tanzania during the May-June, 2000 period. The collection trip involved three scientists from US (Myers, Gilbertson and Miles) and two scientists from SUA (Mabagala, Nchimbi-Msolla). Where possible, seed samples were also collected for gene pool typing. Most of the leaf samples collected were taken to UC-Davis for analysis in Dr. Gilbertson's laboratory. Isolations were made on MXP, NA, YCA and KB media and isolates were identified based on colony color and morphology, starch hydrolysis on MXP, production of diffusable brown pigment on YCA, production of greenish fluorescent pigment on KB, pathogenicity tests on tobacco and bean (variety Canadian Wonder), Biolog Computer Package and finally by PCR using Xanthomonas campestris pv. phaseoli (Xcp) and Pseudomonas syringae pv. phaseolicola (Psp) specific primers. Twenty six bean seed samples were also collected during the survey trip as part of the bean germplasm to be stored at SUA and in the US for future use. Of seventy bean tissue samples collected, 13 were positive for Psp (all Andean type); 12 were positive for Xcp, and 31 were positive for Xcpf. Xcp isolates were mainly obtained from large seeded bean types, however, three of the 12 isolates were obtained from small seeded (Meso-American) beans. Xcpf was also isolated from both Andean and Meso-American bean types. These results are equivocal on the possibility of co-evolution between gene pools and CBB variants. Preliminary data suggests that Psp possibly co-evolved with Andean bean genotypes. More studies are needed to support or reject these hypotheses.

UC Davis: Gene Pool Determination - During the 2000 Tanzania collection, symptomatic leaves with putative bacterial blight infection were collected. From these leaves Xcp/Xcpf and Psp bacteria were isolated and plant DNA was extracted for gene pool determination. Previously, both diseased leaves (for pathogen isolation) and seeds (for gene pool determination) had been collected. Plant DNA was extracted using a modified Dellaporta method. To determine the common bean gene pool the bean genomic DNA was then used in the polymerase chain reaction (PCR) with primers J₁d₁ and J₂d₂.

Bacterial Strains and their Identification: All Xcp and Xcpf isolates from the 1997 collection and those sent by Dr. Robert Mabagala were grown out. For each leaf sample, four yellow xanthomonad-like colonies were selected and streaked onto another plate, then classified as either Xcp or Xcpf. To rapidly confirm the identity of these strains as Xcp or Xcpf, a highly sensitive and rapid PCR-based assay was used. This assay is based on the use of a PCR primer pair developed from the partial sequence of a 3.4-kb Xcp plasmid DNA fragment (X4c: 5'-GGC AAC ACC CGA TCC CTA AAC AGG-3' and X4e: 5'-CGC CGG AAG CAC GAT CCT CGA AG-3'). This primer pair directs the amplification of an ~700 bp fragment from CBB bacteria and not from nonpathogenic xanthomonads or other bacteria associated with common bean. Three strains from the 1997 collection, 11 from Dr. Mabagala and 7 from the 2000 collection were nonpathogenic xanthomonad-like bacteria. During the 2000 collection trip, bean leaves with halo blight symptoms also were collected and 21 pseudomonad-like strains were isolated from these symptomatic leaves. For these isolates, the halo blight bacteria-specific primers HB14 (HB14F: 5'-CAA CTC CGA CAC CAG CGA CCG AGC-3' and HB14R: 5'-CCG GTC TGC TCG ACA TCG TGC CAC-3') that specifically direct the amplification of an ~1.4-kb fragment from pathogenic Psp were used to confirm the identity of these bacteria. Four of the isolates were determined not to be Psp.

Repetitive Bacterial Sequence: Polymerase chain reaction (PCR) amplification of repetitive bacterial sequences (rep-PCR) was used to generate fingerprints of bacterial strains in order to further ascertain their identity and establish relationships among the strains. The repetitive extragenic palindromic (REP)-PCR primers, enterobacterial repetitive intergenic consensus (ERIC)-PCR primers, and BOX element 1A (BOX)-PCR primer were used. Xcp isolates from Malawi and Tanzania had two different fingerprints. Interestingly, a strain of Xcp collected from Puerto Rico in 1998 had a different fingerprint. This showed that there are genetically distinct genotypes for Xcp strains. Xcp strains with the fingerprint of the Puerto Rico strain were not found in East Africa based on earlier collections. However, from the 2000 collection three Xcp isolates (8-2, 8-3, 8-4) collected from one site in Arumeru from Andean materials had the same fingerprint as the Puerto Rico strain. We then retrieved a number of Xcp/Xcpf strains previously collected from the New and Old World (from -80oC storage) and determined their fingerprints. All Xcp isolates from the US, Colombia and Brazil had fingerprints that were indistinguishable from the Puerto Rico strain, whereas Xcp strains from Kenya and Uganda (East Africa) had fingerprints similar to one of the two fingerprints previously associated with East Africa strain. Together, these results suggest that some Xcp strains are unique to East Africa and are genetically distinct from the New World. These results also suggest that Africa may be a center of diversity for CBB. All Xcpf from Malawi and Tanzania collections, as well as those provided by Dr. Mabagala, had identical fingerprints. Furthermore, Xcpf strains from the US, Guatemala, Puerto Rico, Kenya, and Uganda also had the same fingerprint. The Xcpf fingerprint was significantly different from Xcp fingerprints, consistent with Xcp and Xcpf being genetically distinct.

For Psp, three fingerprint patterns were detected for the isolates collected from Tanzania during 2000. This agrees with earlier findings of Mabagala (1992), that Psp strains are variable in Tanzania. The strains 15-1, 15-2, 15-3 and 15-4 were all isolated from the weed Neonotonia wightii and had identical fingerprints. Strains previously isolated from this weed were classified as belonging to Race 1. Thus, it is tempting to speculate that this fingerprint might correspond to Race 1 Psp strain. A Psp strain from our -80°C storage had a fingerprint that was different from the three previously identified fingerprints. If these fingerprints correspond to races of P.s. pv. phaseolicola, rep-PCR may represent an easier method for determination of races of this pathogen compared with the use of differential cultivars.

Pathogenicity: In most cases, four susceptible US bean varieties, Topcrop and White Kidney (Andean) and Black Turtle Soup and Sutter Pink (Mesoamerican), were used for pathogenicity tests. Malawian and Tanzanian materials were also used in the evaluation. Inoculation of five plants per genotype was carried out as described above. Symptoms were rated and recorded 10 days post inoculation. The Puerto Rico Xcp strain (P1) was generally more pathogenic than the Malawi Xcp strain, and it was equally pathogenic on materials of both gene pools. In contrast, the Malawian Xcp strain (M11) was highly pathogenic on the Andean genotypes, but was much less pathogenic on the Mesoamerican ones. This suggests that genetic diversity exists among Xcp strains in terms of pathogenicity on common bean gene pools and, possibly, that Malawian Xcp strains have co-evolved (or have been selected) with the more prevalent Andean landraces in Malawi. Both Xcpf strains (M12 and P2) were generally more pathogenic than the M11 Xcp, and they were highly pathogenic on both gene pool materials, just like the P1 Xcp. This response of Xcpf strains is similar to that observed previously. The greater pathogenicity and lack of differentiation between the two gene pools for the Puerto Rico Xcp was also similar to the responses previously observed of the Brazilian strains 29 and 30. As indicated above, it seems that all Xcp strains in the New World are the same 'race' or pathogenicity type because they have the same fingerprint. Similar pathogenicity tests with the Malawian and Tanzanian bean genotypes as well as the four standard genotypes were performed. The response of the US standard bean genotypes was similar to that observed above in that (i) the Puerto Rico Xcp was generally more pathogenic than the Malawian Xcp, (ii) The Malawian Xcp was much less pathogenic on Mesoamerican cultivars, and (iii) the Puerto Rico Xcp was highly pathogenic on cultivars of both gene pools. Interestingly, in this experiment the Malawian Xcp strain was more pathogenic on the Malawian Mesoamerican beans than previously observed. Both Xcpf strains were more pathogenic than the Malawi Xcp and were comparable in pathogenicity to the Puerto Rico Xcp. However, both Xcp and Xcpf strains from Puerto Rico were slightly more pathogenic than the Malawi Xcpf. Inoculation data for Tanzanian Xcp and Xcpf strains from the 2000 collection were compared with those that were earlier provided by Mabagala. . Again, the Xcpf strains were similar in pathogenicity and were more pathogenic than the Xcp isolates. However, the Tanzanian Xcp strain, 8-2, which had a fingerprint similar that of the Puerto Rican Xcp strain, was more pathogenic than Tanzanian Xcp strain T21, particularly on the Mesoamerican cultivars (just like the Puerto Rico Xcp). This was seen in two independent experiments. This again points toward an earlier finding of possibility of some interaction between the pathogen and host, and demonstrates that this is a characteristic of only a subset of East African Xcp strains. Thus, the Xcp strains that are widely extant in the New World and less common in East Africa, had no differential pathogenicity among bean gene pools. The strain/gene pool relationship was detected only for those strains having the other two fingerprints, those found only in East Africa. Generally, all materials from Malawi and Tanzania were highly susceptible. Thus, the fact that all the East African bean genotypes examined are highly susceptible to CBB indicates a need to incorporate CBB resistance. Moreover, the East African bean breeding programs would also need to incorporate resistance against all three types of Xcp.

Pathogenicity tests of four strains of Psp representing the four fingerprint patterns were conducted similarly to Xcp/Xcpf with one additional Andean genotype, Canadian Wonder, from Tanzania. Generally, Psp strains were more pathogenic on Andean U.S. and East African genotypes than they were on the U.S. and East African Mesoamerican genotypes. The strains representing the four fingerprint patterns also reacted differently. Our earlier collected strain E1 and 12-5 were least and most pathogenic, respectively. However, the strain 7-5 was equally pathogenic on the Andean and Mesoamerican genotypes whereas the strain that was collected from Neonotonia wightii (15-1) was more pathogenic on Andean compared to the Mesoamerican genotypes. This strain may be indigenous to East Africa and shows the same type of response on bean gene pools that was observed for Xcp strains that are extant only in Africa. These results clearly differentiate the four strains and confirm the differences observed for the fingerprint patterns. The relationship between Psp and Andean gene pool was further reflected in the fact that all isolates were only collected from that gene pool.

Summary of the CBB Survey: In the Malawi collection, of 1997 a total of 23 isolates were collected (Table 2). Of eleven Xcp isolates, seven were collected from beans of the Andean gene pool and two were collected from the Mesoamerican gene pool (Table 3). The source of the remaining two strains was unidentified because seeds were not collected. Interestingly, of the nine Xcpf isolates collected, most (8) came from Mesoamerican gene pool bean genotypes. These results suggested a possible relationship between the common bean gene pool and Xcp and Xcpf (i.e., Xcp strains were isolated from Andean gene pool materials and Xcpf strains from Mesoamerican gene pool materials). In Malawi, beans of both gene pools are widely grown even as mixtures in the same field. But in Tanzania, where Andean gene pool materials are prevalent and are grown in pure cultures either as monocrops or in intercropping systems, a different picture emerged. The isolates from Dr. Mabagala were collected only from Rose Koko, an Andean landrace bean material, and only Xcpf was isolated from these. Of the 47 isolates collected the 2000 Tanzania survey, only 4 were Xcp (three from the same site, Arumeru). Moreover, about two-thirds of the Xcpf isolates were also collected from Andean gene pool materials and the other third from Mesoamerican gene pool materials, consistent with Xcpf strains being pathogenic on materials of both gene pools.

Our results suggest Xcpf is the major bacterium causing CBB on beans in Tanzania, where most bean materials are actually of Andean gene pool background. Thus, Xcpf strains should be included in a CBB resistance breeding program.

Examination of genetic diversity of common blight bacteria using Restriction Fragment Length Polymorphism (RFLP) analysis: Broth cultures of bacteria from single colonies were grown and DNA was extracted from bacterial cultures at mid- to late-logarithmic growth phase. DNA was digested with EcoR1 or HincII and fragments were separated on ethidium bromide stained gels. The DNA was denatured and transferred to the nylon membrane (Hybond) and subjected to southern hybridization. Clones P2 and P7, which harbor repetitive elements from the Xcp genome were used as hybridization probes.

When probed with P2, all the Xcp strains had one major common restriction fragment when EcoRI was used to digest bacterial total genomic DNA. However, the P2 probe hybridized with numerous EcoR1 fragments from the Xcpf genome. Similar results were obtained when HincII-digested DNA was probed with P2-derived probe. On the other hand, when EcoRI-digested genomic DNA was probed with the P7 probe, polymorphisms were observed from the Xcp strains from Africa, and the strains were divided into two subgroups. M11, M19, and T21 were in one group, whereas M17, M25, and T21 were in another group. These were the same two subgroups revealed by rep-PCR. Interestingly, both the P2 and P7 probes hybridized to multiple EcoR1 fragments of the Puerto Rico Xcp strain (data not shown), indicating that this Xcp is indeed different from the other two Xcp groups above. Apart from the corroboration of the rep-PCR data, a major finding from the RFLP analysis is the fact that there may be diversity in Xcpf as well.

Fatty Acid Analysis: Eighteen bacterial strains from the 1997 collection and Stenotrophomonas maltophilia (control) were streaked on Trypticase Soy Broth Agar (TSBA) in plates according to the recommended 'Quadrant Streak Technique' by the Microbial Identification System (MIS). They were incubated for 24h at 28°C and cells were harvested from the most dilute (third) quadrant showing late log-phase growth. The fatty acid methyl esters (FAMEs) were saponified, methylated, extracted and base washed following the MIDI method of Sasser (1990) and Paisly (1995). FAMEs were separated by a Hewlett Packard 5890 series II gas chromatograph. Individual FAMEs were identified and quantified by the peak naming table component of the MIS. Quantities were expressed as percentages of the total named FAME peak area. Identification of the FAMEs was based on calculation of equivalent chain length (ECL). Stenotrophomonas maltophilia was correctly identified at the level of 72% and 68% relatedness in two replications (>50% recommended). Only those strains that were identified at >50% relatedness were used in the following analysis.

A total of 28 fatty acids was detected in Xcp and Xcpf. The mean percentages of fatty acids making up >0.2% of the total area for these acids were used in the composite profiles. About 80% of the Xcp/Xcpf fatty acid content was made up of 11:0 iso; 13 iso 3OH; 15:0 iso; 15:0 anteiso; 16:0; 17:0 iso and iso 17:1 w9c. This major profile is different from the X.c. pv. phlei and X.c. pv. graminis, where unsaturated fatty acids 16:1 w7c and iso 17:1 w9c apart from the branched 15:0 iso and 15:0 anteiso fatty acids, are part of the 60% composition (Sletten, 1996). The major composition profile is also different from that of X.c. pv. pruni where about 59% of the fatty acids were only the branched fatty acids 15:0 iso and 15:0 anteiso and the unsaturated acid 16:1 w7c (Scortichini et al., 1996). For Xcp and Xcpf, the branched fatty acid 17:0 iso and the unsaturated branched fatty acid iso 17:1 w9c are part of the major profile. Generally, no major differences were observed between the Xcp and Xcpf fatty acid composition. It would seem, therefore, that the fatty acid analysis is not able to differentiate Xcp and Xcpf strains. Out of 19 strains tested in each replicate only eight were detected correctly. It was speculated that this lack of accuracy may have been due to the high polysaccharide contents in Xcp/Xcpf.

Summary: The simple isolation of yellow pigmented bacteria from common bean leaves or seeds on semi-selective medium MXP is not conclusive evidence of the presence of common blight bacteria because of nonpathogenic xanthomonads. Using CBB-specific primers and rep-PCR, CBB isolates were reliably differentiated from nonpathogenic xanthomonads and other yellow bacteria. rDNA from Xcp and Xcpf was amplified using PCR and general primers for this region and, then, the resulting fragments were sequenced. The rDNA sequences of Xcp and Xcpf were nearly identical (99%), which is greater than that for distinct pathovars.

The 'co-evolution' studies with Malawian Xcp and Xcpf strains showed that certain Xcp strains had greater pathogenicity for Andean compared with Mesoamerican materials, but Xcpf strains were highly pathogenic on both centers of origin. In Malawi Xcp was more commonly isolated from Andean genotypes, and Xcpf was more commonly isolated from Mesoamerican materials. This suggested that some level of CBB resistance may be present in some East African Mesoamerican materials and may be useful for the bean. The data also suggested some level of pathogen-host co-evolution (i.e., host plant selection), particularly in regard to the consistent association of Xcp and Xcpf with Andean and Mesoamerican materials, respectively. A model was proposed in which Xcp strains might represent the indigenous CBB evolved with local Andean materials.

Current Status of Research: Studies on the possible mechanisms of co-evolution between CBB and the bean plant were initiated in 1997. Data obtained from Malawi, Tanzania, Uganda and Kenya do not generally seem to strongly support the theory of pathogen-host coevolution of Xcp with Andean genotypes. These findings may be due to the fact that East and Southern Africa is not a center of origin for beans, but it is a recent introduction to the region (Secondary Center). This means that in order to get more conclusive data on this theory, more work will be needed in the Andean and Meso-American regions, which are centers of origin for Meso- and Andean bean genotypes. There is also a need to conduct further research with Psp because East Africa is thought to be a center of diversity and possibly origin for halo blight.

Activity #4: Continue advance of Temple/Bokosi materials; incorporate CBB resistance (collaborate with LAC)

U.S. researchers: Temple
HC researchers: Bokosi, Mkandawire
Methodology: The objective of our breeding activities is to produce acceptable high yielding bean varieties with general resistance to BCMV and ALS while conserving genetic diversity. Crosses were performed at UCD by Temple using A-lines and MCM lines from CIAT as donor parents for ALS and bc-3 genes for resistance, respectively, and selected Malawi preferred bean lines.

To develop improved disease resistant varieties, advanced materials will continue to be evaluated and screened. This will involve repeating some second backcrosses with group #1 priority materials and screening F₈ progeny of the BC₁ of group #2 priority materials and continuing selections of priority #2 materials. To incorporate CBB resistance into selected improved ALS/BCMV resistant materials, appropriate donors of CBB resistance will be determined based in part upon results of the CBB co-evolution study and based on discussions with LAC cooperators. Presently there appears to be two major options: (i) using one or more VAX lines, which are small-seeded lines that combine CBB resistance from both P. acutifolius and P. coccineus or (ii) Wilkinson 2, a large-seeded material with CBB resistance. CBB screening will be done in greenhouses at UCD and in the field in Malawi/Tanzania.

The new bean-breeding scheme has been operative for the past five years. From crossing blocks the materials are systematically evaluated through a series of trials at three major locations, viz.: Dedza, Bunda and Champhira. These yield trials start with the Observation Nursery and from that trial materials are moved into the Preliminary Yield Trial, Advanced Yield Trial and finally the National Yield Trial, which is conducted at two further sites, viz.: Matapwata, Thyolo and Ng'onga, Rumphi. Each one of these trials is split into two growth habits, viz., determinate and indeterminate types. The Temple materials were produced to combine resistance to ALS and BCMV in most preferred backgrounds. During FY 00 these materials will start to be evaluated for yielding ability in this breeding scheme.

Anticipated results of activity: Advanced lines with multiple disease resistance (ALS, BCMV, and CBB) will be generated for testing in on-farm trials and, ultimately, for release to farmers.

Anticipated impact(s):

• Farmers will receive higher yields consistently from materials with resistance to these major diseases; Among farmers adopting these varieties there will be a reduction in poverty and improvement in nutrition as a result of these materials.

Progress During Past Year: UC-Davis: Following the loss in 1999 of F2 BC3 progenies, a vigorous program was initiated to repeat the backcross and return quickly to our goal of generating an array of improved "component" local landrace genotypes that would be readily accepted and incorporated into the widely grown mixtures. In order to concentrate on progenies of the highest possible merit, male backcross parents were limited to four elite F5 BC2 lines that had performed well in multiple tests using the NL3 strain of BCMNV. The recurrent Malawian parents were traditional cultivars like Nasaka and Nyazembe, but also included several lines selected by the National Program, that represent improved versions of favorite types like Sapalekedwa and Nanyati. A total of 15 hybrid combinations were successfully made, and the F1 was planted in the UC Davis greenhouse (with parental types) to guarantee the greatest quantity and high quality of F2 BC3 seed. An average of 1400 seeds per cross are being prepared for planting and selection in Malawi during the 2000-2001 season. Especially promising is the large array of grain types that are very close to the Sapalekedwa (a type of Dark Red Kidney) and Nanyati (similar to Cranberry) landrace groups. It is reasonable to hope that F4-F6 selections from this BC3 effort can enter replicated yield trials designed to identify elite lines for increase and release. Enough F2 BC2 seed is available to select in more than one Malawian environment. Field selections from these F2 populations must then be screened/tested for specific reaction to BCMNV, Angular Leaf Spot, and other high priority traits.

Screening breeding lines for CBB resistance: We obtained 10 CBB breeding lines with pyramided resistance from P. acutifolius and P. coccineus. from Dr. Shree Singh of University of Idaho. These, together with the four susceptible US bean varieties (Topcrop, White Kidney, Black Turtle Soup, and Sutter Pink), were evaluated for resistance using one Xcp (M11) and one Xcpf (M12) strain. The first fully expanded trifoliolate leaves of five plants of each genotype at two weeks of age were inoculated using the razor blade method. Symptoms were rated and recorded 10 days post inoculation. Generally, M12 (Xcpf) was more pathogenic than M11 (Xcp) as we had observed in previous pathogenicity tests. M11 was more pathogenic on the Andean U.S. susceptible genotypes in both trials than on the susceptible Mesoamerican genotypes. M12 was equally pathogenic on genotypes from either gene pool. All the breeding lines were rated resistant to these strains. The most resistant genotypes were VAX5 and VAX6; black- and red-seeded Mesoamerican materials, respectively. VAX 6 is preferred because of its seed color and is being used as a parent in crosses for CBB resistance improvement in East Africa.

Malawi: Screening F2BC bean plants for disease resistance: The F2BC seeds were received from Dr Steve temple and planted at Bunda and Dedza in 1999/2000 cropping season for evaluation. Dedza turned out to be a better screening site because of heavy prevalence of ALS, BCMV and CBB. The most promising plants, in terms of disease resistance, come out of B105, B116-4, B117-6, B122 and B126-5. The materials have been multiplied (Winter crop) and should be available for further screening in the coming cropping season (2000/2001)

Selections of promising materials continued. The disease pressures at Dedza allowed the differentiation of susceptible from resistant materials for diseases such as ALS, BCMV and CBB. The BCMV resistant lines B2 and B7 have been used as parents in a crossing program to develop BCMV resistant progenies with the preferred seed types. There are segregating materials in various stages (F2, F3, F4 and F5). These will also be evaluated and further selected in the coming season.

Progeny lines from the UCD Temple crosses were evaluated at Dedza and Bunda. F7 lines obtained from single plant selections from the 1998 and 1999 evaluations were grown in plant-to-row arrangement with a BCMV spreader material (Bunda 93) grown every sixth row in unreplicated trials. Only selections with crosses 2, 72A, 66, 60, 21, 54, 51, 95, 75, 89, 99, 129, 134, 22, 25, 27, 34, 56 72B and 93 were carried forward for this evaluation. Some lines apparently had BCMV resistance conferred by bc-3, as well as ALS resistance. A total of 53 test lines were planted in single 2m rows for evaluation. The spreader of BCMV (Bunda 93) effectively produced the disease at expected levels (scores of 5-8 on a 9-point scale). Some lines within the crosses 72A, 60, 22, 25, and 93 succumbed to BCMV, but most materials were resistant to this disease. Lines from the crosses 66, 21, 54, 105, 95, 75, 89, 99, 134, 25, 27, 34 and 56 were still segregating for the I-gene from the A-lines and so they were scored for black root incidence. Bunda 93 was not as effective for establishment of ALS but screening for this disease was achieved. Some lines within the crosses 54, 95, 89, 27, 34, and 56 were susceptible to ALS. All lines from the crosses 2, 66, 54, 105, 51, 99, 129 and 72B were resistant to both BCMV and ALS. Interestingly, lines from the crosses 2, 66, 105, 51, 99 and 129 were also resistant to field-spread CBB whereas those from the 72B cross were fairly susceptible. Most of these materials are in the large seed preferred background and are in final stages of evaluation for release.

Observation nursery: These lines and materials from other crosses and programs are moving through our breeding scheme where initial evaluation is conducted in an unreplicated observation trials at Bunda and a drought hot spot (Dedza this year). Beans are intercropped with maize, which is planted at 30cm apart in the same row with beans. Five-hundred entries were screened. Single plant selections have been made from some segregating materials and these have been increased under irrigation (winter) to increase seed for further evaluation during the 2000/2001 cropping season. Ten entries will be moved into preliminary yield trial.

Preliminary yield trial: Promising materials from the observation trial are evaluated in the Preliminary Bean Yield Trials (PBYTs), located at Bunda, Dedza and Champhira, the latter two sites being in farmers' fields closer to Bunda amongst bean trial sites for easier supervision. Dedza and Champhira are hot spots for ALS and CBB, respectively whereas Bunda usually has both ALS and CBB. The PBYT is divided into two groups depending on plant growth habit. Each of the two trials (determinate and indeterminate) is comprised of 33 test entries and three checks in two replications. Three checks are included; a recent and an old release and the farmer's own material at that site. Released materials from other programs may be included in this trial. Data for Dedza and PT II are being processed. Entries DC 95-118 and PC490-D8 gave the highest and second highest yield at Bunda, respectively. Both entries beat the check variety (Nasaka). Entries 387 and PC 480-D5 had similar yields at Champhira. Interestingly, PC 490-D8 and DC 95-118 also performed well at Champhira, indicating their superiority in performance across the two sites. These two entries also yielded higher than the two control varieties (Nasaka and Kalima). Yields were better in Champhira than at Bunda.

Advanced yield trial: Best entries from PBYTs are advanced to the Advanced Bean Yield Trials (ABYTs), which is comprised of 22 materials evaluated against an old and newly-released variety and a local material. The advanced trials were planted at the sites, Bunda Champhira, and Dedza. However, data from Dedza are still being processed. Entries DC 96-95 from Uganda and PC 490-D8 from the republic of South Africa were the highest yielding lines at Bunda and Champhira, respectively. These two entries have similar maturity duration. The yields at Champhira was better than at Bunda. Entry ZPV 906 was tested in both Advanced I and this year (2000). However, this variety takes 104 minutes to cook to 100%. Therefore further crosses are needed to develop a fast cooking type in this background.

In the Advanced Bean Yield Trial II, IZ238-1 and AFR 88 gave the highest yields at Bunda and Champhira, respectively. Entry 4G/7, a Bunda cross, was the only entry found in the top five at both sites. It is a good candidate for National Bean Yield Trial. Namajengo and Kanzama, two released varieties, performed relatively well at Bunda but not Champhira. Entry 14K/2 appeared in both Advanced Bean Yield Trial II and National Bean Yield Trial II.

National bean yield trial: Better entries from ABYTs are advance to the National Bean Yield Trials (NBYTs), which are comprised of 13 materials evaluated against an old and newly-released variety and a local material. Two of the NBYTs are located at Matapwata and Ng'onga, sites furthest from Bunda, to better capture variability due to locations before better materials are selected for On-Farm evaluation. The National Bean Yield Trial 9NBYT I and NBYTII) were located at 5 sites Bunda, Champhira, Dedza, Matapwata, and Ng'onga. Out of 16 entries tested in NBYTI, 6 were from the National Bean Programme (Chitedze Research Station). Table 4 gives the five top yielding entries of NBYT I Bunda, Champhira, Matapwata and Ng'onga. Lyamungu 90 was the highest yielding entry at Bunda. PC512-B4 gave the highest yield at Champhira. LRK ranked first in yield at Matapwata but was also within the five top yielding entries at both Bunda and Champhira. All the top four entries at Ng'onga were from Chitedze. Entry ZPV 906 was the only variety that performed in the top five at all sites. Bunda and Champhira were favorable environments for yield, while Matapwata and Ng'onga were less favorable. Drought conditions were observed at Ng'onga and high rainfall at Matapwata favored diseases, especially web blight. Weeding and cultivating also was a major problem at Matapwata because of the heavy rains. Both Lyamungu 90 and ZPV 906 have been earmarked for release this year (2000)

Yield performance of NBYTII entries have been analyzed for four locations, but is still being analyzed for the Dedza location. The results show that Namajengo, Kanzama, DC86-244, were the highest yielding entries at Bunda, Champhira, Matapwata, and Ng'onga respectively. Two of the entries, Namajengo (released check) and 2G/Z (Bunda cross) performed well at all four sites, giving yields within the top five in all sites. The small red type did well at all sites, and most notably, held up well under drought conditions at Ng'onga. Their apparent tolerance to drought stress warrants further investigation.

Champhira was the most favorable environment for yield, while Matapwata was the least favorable. Heavy rains at Matapwata favored web blight. Entries 15P/8 and 2 G/2 have wide adaptation and have performed consistently well over the past five year. They have now been earmarked for release this year (2000).

From 2000, the evaluation of NBYT is done jointly with CIAT/National Bean Program both of which jointly contribute test materials. The release of materials will now be from one joint Malawi Bean Program and will together be multiplied and distributed by non-governmental organizations (NGOs).

Current Status of Research: Preliminary evaluation to confirm resistance of the Temple materials indicates that a number of materials were able to stand pressure of ALS in Malawi and BCMV in Davis, CA. As a result it is hoped that this resistance will still be shown in the field trials and more importantly, that the resistant materials will also have the yielding ability advantage over the susceptible materials. The later group currently comprises most of the materials in the breeding scheme. Four potential varieties, two of which are determinate and other pair indeterminants have been earmarked for release this year. The two determinate types are Lyamungu 90, a Kalima kidney type with medium seed type (45g per 100 seeds) and has yield potential of 2000kg/ha. The other is ZPV906, a white kidney type with medium seed weight (42g per 100 seed). The two indeterminate types include 15P/8 and 2G/2. The first variety is a product of crosses made at Bunda between AB136 and 13-3. It is a small red type (20g per 100 seeds) and has a yield potential close to 2000 kg/ha when grown as a sole crop. The other variety, 2G/2, is also a small red type and a product of a cross between C76-4 x 6-5. It has similar yield potential to 15P/8